OSTERREICHISCHE AKADEMIE DER WISSENSCHAFTEN

PHILOSOPHISCH-HISTORISCI{E KLAS
DENKSCHRIFTEN, 449.8 AND

SE

CATHY CA}.{TWELL, ROBERT MAYER

A h{cble Noose of }rlethods,

The Lotus Garland Synopsis:
A Mahayoga Tantra and its Commentary

oste rre

cn

u.

r,

[*[ii3 g?l

"
der Wissenschaften

ffiAW

CATHY CANTWELL, ROBERT MAYER

A Noble Noose of Methods, the Lotus Garland Synopsis:

A Mahyoga Tantra and its Commentary

STERREICHISCHE AKADEMIE DER WIS SEN SCHAF TEN

PHILOSOPHISCH-HISTORISCHE KLASSE
DENKSCHRIFTEN, 449. B A N D

Beitrge zur Kultur- und Geistesgeschichte Asiens

Nr. 73

Herausgegeben von Helmut Krasser

STERREICHISCHE AKADEMIE DER WIS SEN SCHAF TEN

PHILOSOPHISCH-HISTORISCHE KLASSE
DENKSCHRIFTEN, 449. B A N D

CATHY CANTWELL, ROBERT MAYER

A Noble Noose of Methods,

The Lotus Garland Synopsis:
A Mahyoga Tantra and its Commentary

Vorgelegt von w. M. ERNST STEINKELLNER in der Sitzung am 15. Juni 2012

Die verwendeten Papiersorten sind aus chlorfrei gebleichtem Zellstoff hergestellt,

frei von surebildenden Bestandteilen und alterungsbestndig.

Alle Rechte vorbehalten

ISBN 978-3-7001-7273-4
Copyright 2012 by
sterreichische Akademie der Wissenschaften
Wien
Druck und Bindung: Prime Rate kft., Budapest
Printed and bound in the EU
http://hw.oeaw.ac.at/7273-4
http://verlag.oeaw.ac.at

TABLE OF CONTENTS
Acknowledgments

vii

Note on Transliteration of Tibetan

ix

Abbreviations and List of Sigla used in the Editions

xi

Introduction
Introducing the Textual Sources, and their Significance

The Contents of the Thabs zhags and their Significance for the Historical Study
of the rNying ma pa

The Dunhuang Manuscript as a valuable source from the 'Time of Fragments'

A Summary of the Salient Points of our work on the Thabs zhags Textual Tradition

10

Methodological Issues in the Study of the Thabs zhags Textual Transmissions

13

The Ancient Tantra Collection (NGB)

14

Textual Obscurity and Scribal Corruption in the extant Ancient Tantra Collection

15

Increasing the usability of the Ancient Tantra Collection texts

16

How do we edit Ancient Tantra Collection texts? Can we stemmatise them?

How do we account for their variations?

18

Concluding Reflections on the variations in TZ

20

Textual Analysis
a) The Editions of the Root Text and Commentary

26

b) Features of the Dunhuang Manuscript

32

c) Examination and Assessment of the Differing Boundaries of TZ in Different Editions

35

d) The Stemma of the Root Text

43

e) A Summary of The Commentary on A Noble Noose of Methods, the Lotus Garland Synopsis
('Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa)

68

f) The "Citations" or Attributions of the Teachings in the Thabs zhags Commentary to other Tantras 84
Padmasambhava, the Thabs zhags and its Commentary
Padmasambhava's Pith Instructions on the Garland of Views

87

The Evidence from the Commentary and its Dunhuang witness

91

vi

Table of Contents

Introduction to the Editions

of the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa and its Commentary

99

The Presentation of the Critical Edition of TZ

100

The Presentation of the Edition of TZComm

101

Critical Edition of the Root Text (TZ),

The 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa

103

Edition of the Commentary (TZComm),

The 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

229

Appendix
The Deities of the Peaceful Maala in the Thabs zhags tradition

349

The Deities of the Wrathful Maala in the Thabs zhags tradition

358

Bibliography

363

Index

372

CD Images of the Dunhuang Manuscript IOL Tib J 321 from the Stein Collection in London

vii
ACKNOWLEDGMENTS
We are extremely grateful to the UK's Arts and Humanities Research Council (AHRC), who so
generously provided the funding that enabled both of us to spend a full quarter of our working time each,
between 2006 and 2010, in pursuing the research for this book. We are equally grateful for the patience,
generosity, and intellectual inspiration of our many colleagues and friends at Oxford University, and in
particular in its Oriental Institute, who have been so unfailingly supportive of all our efforts since our arrival
there in 2002. We would also like to thank the John Fell Oxford University Press (OUP) Research Fund,
which provided vital support in the final stages of preparing the publication, and Mr. Dylan Esler, who
helped with the final preparation of the manuscript, including the conversions of transliterated Tibetan into
Tibetan script, and some additional collation in the process of rationalising the presentation of the root text
edition. In the early months of the project, the International Trust for Traditional Medicine in Kalimpong
helped with data input from the Golden Tenjur edition of the commentary.
Some sections of the book began life as conference papers, given in Berkeley, Vancouver, Santa Barbara,
Atlanta, London, Bonn and Lumbini, while other sections started life as invited guest lectures delivered at
the Universities of Chicago, Harvard, CNRS-Paris, SOAS-London, Humboldt-Berlin, and Vienna. We would
like to acknowledge the valuable feedback from colleagues attending those presentations, as well as other
academic exchanges we have had the good fortune to enjoy in the course of this work. In particular, we
would like to mention Chris Beckwith, Jos Cabezn, Brandon Dotson, Lewis Doney, David Germano, Paul
Harrison, Agnieszka Helman-Wazny, Adelheid Herrmann-Pfandt, Nathan Hill, Christian Luczanits, Chris
Minkowski, Karma Phuntsho, Charles Ramble, Geoffrey Samuel, Sam van Schaik, Jonathan Silk, Phillip
Stanley, Ernst Steinkellner, Pter-Dniel Sznt, Tanaka Kimiaki, Tsuguhito Takeuchi, and Vesna Wallace.
Very special thanks are due to Changling Tulku of Shechen Monastery, who so generously offered help
with this text despite having so many other more pressing claims on his time, and similar thanks are also due
to Khenchen Pema Sherab, Lopon P. Ogyan Tanzin, and Lama Kunzang Dorjee.
A large debt of gratitude is owed to Helmut Eimer, Helmut Tauscher and Bruno Laine, who at the final
revision stages drew our attention to the versions of the Thabs zhags found in the so far uncatalogued local
Kanjurs of Bathang and of Hemis, providing us also with copies of these additional witnesses that did so
much to clarify and confirm our picture of the historical transmission of the text. We would also like to
acknowledge the professionalism and efficiency of staff at the British Library's Asian & African Studies
Reading Room, who facilitated our unhindered access to the original Dunhuang manuscript, IOL Tib J 321.
The staff at the Central University of Tibetan Studies in Sarnath, notably P. Dorje and Ngawang Tsepag,
went to quite exceptional lengths to make possible under very difficult technical circumstances our
consultation of their microfilm copy of the Tawang O rgyan gling Kanjur, a particularly important historical
witness for the edition. Finally, we are extremely grateful to Tsymzhit Vanchikova, who so generously took
the trouble to supply us with suitable sample pages of the Ulan Bator manuscript Kanjur.
We must also thank the publisher's anonymous peer reviewers, who gave useful feedback at the final
editing stages.
Following the example of Paul Harrison's edition of the Druma-kinnara-rja-paripcch-stra (1992), we
have presented our critical edition of the root text in Tibetan font, in the hope that it might make our work
more accessible to a Tibetan readership. This was no simple undertaking, and would not have been possible
without the technical help and advice of a number of people. Firstly, we must thank Stefan Hagel, who gave
us ongoing instructions on formatting Tibetan text within his Classical Text Editor software. Chris Fynn
advised on a number of computing issues, and contributed the excellent DDC Uchen font that we use. David
Chapman's WylieWord program enabled us to make the complex conversions between Roman and Tibetan
fonts; and George Cantwell at very short notice wrote a macro for Classical Text Editor which resolved a
difficult presentational problem in the Tibetan text.

viii
Earlier versions of some materials from this volume have been published previously. A precursor to the
section below entitled "Methodological issues in the study of the Thabs zhags Textual Tradition", was
published in the JIATS (Cantwell Mayer 2009), with regards to which we must thank an anonymous peer
reviewer who offered useful comments. Other earlier takes on materials that now contribute to this book
appeared in the OTDO Monograph Series Volume III (Cantwell and Mayer 2011), and in the volume
Tibetan Ritual edited by Jos Cabezn (Cantwell and Mayer 2010a).
We thank the British Library for permission to distribute their images of the Dunhuang manuscript IOL
Tib J 321 with our book, which should enable readers swiftly to consult this important witness of the
commentarial text.
Finally, we must acknowledge that whatever shortcomings there might be in this volume are without
exception the result of our own failings, and for this we request our readers' forbearance.

ix
NOTE ON TRANSLITERATION OF TIBETAN
Transliteration of Tibetan in this work conforms to the internationally widely used system often referred
to as Wylie Conventions,1 although we do not use the single contribution which Wylie proposed, that is, the
capitalisation of the first letter of a word where appropriate. Instead, if necessary in the case of names or
titles,2 we capitalise the root Tibetan letter (or the first Roman letter representing the root letter), since this
conforms more closely to Tibetan conceptions, and has a well-established usage in Western scholarly
writings, from Nebesky-Wojkowitz 1956.3 For Tibetan representations of Sanskrit letters, we use the
generally accepted appropriate Roman letters with diacritical marks.
Conventions used in transcribing the Dunhuang document
In presenting transcriptions of the Dunhuang manuscripts, we have conformed to the usages established
by Tsuguhito Takeuchi in a number of publications on Old Tibetan documents, made in accordance with the
suggestions of A. Delatte and A. Severyns (1938: Emploi des signes critiques, disposition de l'apparat dans
les ditions savantes de textes grecs et latins / conseils et recommandations par J. Bidez et A. B. Drachmann,
Bruxelles : Union acadmique internationale).
We have not needed to use Takeuchi's complete list but have used the following.
From Tsuguhito Takeuchi 1995 Old Tibetan Contracts from Central Asia, Tokyo pp.137-138:
I
reversed gi gu
(abc)
editor's note
[a(/b)]
ambiguous readings
[abc]
our conjectural restorations of letters partly illegible or lost in the original
[abc?]
uncertain readings
[...]
illegible letters, number unknown
[---]
illegible letters, number known, indicated by broken line
[3]
illegible letters, approximate numbers known, indicated by numeral with
] abc
beginning of line lost through damage
abc [
end of line lost through damage
***
blank spaces left by copyist (in the case of IOL Tib J 321, generally due to the string holes)
From Tsuguhito Takeuchi 1997-1998 Old Tibetan Manuscripts from East Turkestan in The Stein
Collection of the British Library, Tokyo and London Vol. 2: Descriptive Catalogue 1998, p.xxxii.
$
page initial sign (mgo yig, siddha)
text deleted in the original manuscript4
abc

Following Turrell Wylie 1959. Wylie adopted in its entirety the system earlier used by Ren de Nebesky-Wojkowitz (1956: xv)
and David Snellgrove (1957: 299-300). See the discussion in David Snellgrove 1987a: xxiv, and our own comments in Cantwell,
Mayer and Fischer 2002: Note on Transliteration: "Not Wylie" Conventions
(http://ngb.csac.anthropology.ac.uk/csac/NGB/Doc/NoteTransliteration.xml). In line with Tibetan understanding and the most
common contemporary scholarly usage, we modify the system by using "w" rather than "v" for the subjoined Tibetan letter, "wa"
(wa zur).
We do not capitalise words at all in representing our Tibetan source documents, but do so within the English language discussion
where necessary.
The root letter (ming gzhi) is the main letter of a syllable and that under which words are ordered in Tibetan dictionaries, so it is
the letter of the syllable to which attention is drawn.
Tsuguhito Takeuchi's preferred usage is now not to include deleted words within the main text, but rather in the Critical
Apparatus, marked as, "cancellavit" (this convention is given in his 1995 list). However, we have modified that list in this case,
since it seems helpful in the case of our texts with only short deleted passages, for the reader immediately to see a transcription
which as closely as possible resembles the original.

x
We have also added one further convention:
:
ornamental punctuation mark, generally marking a section ending and new opening, and
varying in design from two large vertically arranged circles to two dots.

xi
ABBREVIATIONS AND LIST OF SIGLA USED IN THE EDITIONS

Throughout the book, we use the anglicised word, Kanjur, for Tibetan bka' 'gyur, and Tenjur, for Tibetan
bstan 'gyur.
Abbreviations used are:
NGB for rNying ma'i rgyud 'bum (Ancient Tantra Collection)
MTph for Man ngag lta phreng
TZ for 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa
TZComm for 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa'i 'grel pa
We follow to some extent the sigla established by Harrison and Eimer (1997) for the editions of the
Kanjur and Tenjur. This is modified because some of the letters from their lists were already used by our list
for NGB texts. Since the Thabs zhags root text is found in both the NGB and a number of Kanjurs, and its
commentary is found in three Tenjurs, as well as being represented in a Dunhuang document, it seemed most
straightforward to qualify the sigla with a lower case letter, k, for Kanjur texts, and t, for Tenjur texts, when
there might otherwise be confusion. Thus, D stands for the sDe dge NGB, and Dk for the sDe dge Kanjur,
while Qk stands for the Peking Kanjur, and Qt for the Peking Tenjur. Recent work on local Kanjurs has also
established the principle of using more than one letter for these further editions, and thus, we follow Helmut
Eimer in using Ogl for the Tawang Kanjur from O rgyan gling, and Bth (used earlier in Michael
Zimmermann's work on the Tathgatagarbhastra) for the Bathang Kanjur held in the Newark Museum.
See the Bibliography for fully bibliographical references. The local Kanjurs and proto-Kanjurs from
Western Tibet do not yet have established sigla. Pending such establishment by those working on these
collections, principally at the University of Vienna, we use He for the (incomplete) Hemis Kanjur which
contains our text.
We have used Tibetan language sigla for the root text edition, which is presented in Tibetan. We have
tried to present a shortened form of the Tibetan names, mostly using the place names where the edition was
produced or preserved, and we have tried to make them easy to guess, so that it should not be necessary to
check this list repeatedly.
Ms = Dunhuang manuscript; =

J = Lithang or 'Jang sa tham Kanjur; =

Qk = Peking Kanjur; =

Nk = Narthang Kanjur; =
U = Urga Kanjur; =

Dk = sDe dge Kanjur; =

D = sDe dge NGB; =

Hk = lHa sa Kanjur; =

V = Ulan Bator Kanjur; =

Qt = Peking Tenjur; =

Gt = Golden Tenjur; =
Nt = Narthang Tenjur; =

xii
M = mTshams brag NGB; =

G = sGang steng b NGB (G-a is used for sGang steng a NGB); =

Gr = dGra med rtse NGB; =

T = gTing skyes NGB; =

R = Rig 'dzin NGB; =

K = Kathmandu NGB; =

Bth = Bathang Kanjur; =

He = Hemis; =

Ogl = Tawang Orgyan ling; =

INTRODUCTION
Introducing the Textual Sources, and their Significance
This volume represents the outcome of a research project on a famous rNying ma Mahyoga root tantra,
the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa (hereafter abbreviated as TZ), together with
its commentary. The surviving sources have left us no clearly established Sanskrit titles,1 but we translate the
Tibetan title of the root text as A Noble Noose of Methods, the Lotus Garland Synopsis. Versions of TZ were,
as far as we can tell, prominently included in all known editions of the Ancient Tantra Collection (rNying
ma'i rgyud 'bum, hereafter NGB).2 A version has also been transmitted through the editions of the Tshal pa
branch of the Kanjur that contain a special Ancient Tantra (rNying rgyud) section. TZ has additionally
surfaced in the three local Kanjur editions of Hemis, Tawang, and Bathang, that is, local manuscript Kanjur
collections which do not reflect either of the two major transmissional branches of Tshal pa and Them
spangs ma (which between them subsume all the popular printed Kanjur editions).
The Tibetan title of the commentary is, simply, 'Phags pa Thabs kyi zhags pa padma 'phreng gi don
bsdus pa'i 'grel pa (hereafter abbreviated as TZComm). A version of TZComm is shared by three editions of
the Tenjur (bsTan 'gyur), namely Golden, Peking, and Narthang. Unfortunately, this, the sole traditionally
transmitted version of TZComm, has lost around thirty percent of its text, and has also here and there suffered
further transmissional misfortunes. No other traditionally transmitted versions of TZComm appear to have
survived, not even in such suitable rNying ma corpora as the various bKa' ma compilations (see below).
Fortunately however, a probably tenth century manuscript in eighty-five folios preserving an almost
complete TZComm was amongst the famous Dunhuang treasures brought to London by Sir Aurel Stein in the
early 20th century, and is now held at the British Library (IOL Tib J 321).
TZ is admired as a key scripture by the rNying ma pa, and is consequently preserved within a distinct and
prestigious doxographical section of the NGB known as 'The Eighteen Tantras of Mahyoga' (ma ha yo ga'i
1

Out of all the versions of TZ and TZComm, only the Bhutanese NGB edition of TZ attempts a Sanskrit title: rya ka la pa sha
padma m le sang kra ha, perhaps intending * rya-upya-pa-padma-ml-sagraha? Reconstructions of TZComm's
Sanskrit title have been suggested by Chattopadhyaya (1972: 49, *Upya-pa-padma-ml-pirtha-vtti) and HerrmannPfandt (2000: 270-1, * rya-arthasagraha-nma-upyapa-padmvali-vtti for the front title and, *Upyapa-padmamlkalparja-arthasagraha-nma-vtti for the colophonic title). A much earlier attempt to reference the Sanskrit title of TZ is
found in the Dunhuang text PT 849, which Hackin dubbed the Formulaire Sanscrit-Tibtain du XE Sicle. Here we find that a
Tibetan entry rgyud thabs kyi zhags pa is mistakenly rendered a mo ga pa sa tan tra, a confusion between the Amoghapatantra of the Kriytantra genre that was popular at Dunhuang, and our somewhat rarer *Upyapa-tantra of the Mahyoga
genre. Since the references occur within a list of famous Mahyoga texts, and directly next to the rGyud gsang ba'i snying po
or *Guhyagarbha with which TZ is often paired, it is highly likely that *Upyapa was intended by the author of PT 849
[Hackin 1924:6]. For further analysis of PT849, see Kapstein 2006. Excellent digital images of the original scroll of PT849
can now be accessed at The Mellon International Dunhuang Archive <http://www.artstor.org/what-is-artstor/w-html/colmellon-dunhuang.shtml>.
For those who are not familiar with the Ancient Tantra Collection, we should mention that the editions studied so far break
down into three distinct branches, which we have provisionally termed the Bhutanese, South Central Tibetan and Eastern.
These provisional names reflect the empirical data, in that the distribution of NGB editions has so far followed a consistently
regional pattern. Future discoveries might well render these regional names inaccurate, and in due course, we hope to go
beyond such ad hoc regional identifications by discovering the original source of each distinctive redaction, and then more
appropriately naming the various branches after them. For now, we must talk of a Bhutanese branch in forty-six volumes, with
four available manuscript versions, all virtually identical to one another; an Eastern branch in twenty-six volumes, so far
represented only by the conflated single witness sDe dge xylograph edition; and a South Central Tibetan branch with two
subdivisions, represented by four available complete manuscript versions, two with thirty-three volumes and two with thirtyseven volumes. Rig 'dzin and gTing skyes had thirty-three volumes; Nubri and Kathmandu had thirty-seven volumes. See
Cantwell and Mayer 2007: 11, 16-19.

Introduction

rgyud sde bco brgyad). It is interesting to note that a high proportion of the rNying ma tantras titles
mentioned at Dunhuang are found within this grouping, which moreover includes some titles of known Indic
provenance, such as the Guhyasamja, a Buddhasamyoga, and a rparamdya. TZComm, however,
displays some sign of probable authorship in Tibet, or at least, contains some material most probably
composed in Tibetan. Its Chapter Six glosses the Tibetan term for maala, dkyil 'khor, according to its two
halves, giving first an explanation of centre (dkyil), followed by an elaboration on circle ('khor).3 But the
verses of TZ itself do not appear Tibetan in any such obvious way, and TZ was duly accepted as an authentic
Ancient Tradition scripture in the text lists of two early Sa skya masters. Grags pa rgyal mtshan (1147-1216)
selected it as one of the only six rNying ma scriptures included in his tantra catalogue, while his greatnephew Chos rgyal 'Phags pa (1235-1280) followed suit in his own catalogue of 1273 (Eimer 1997: 52).
Later, TZ was included in the special three-volume rNying rgyud section of twenty-four texts added to the
Tshal pa Kanjur, the dkar chag of which was established by the great bKa' brgyud scholar Tshal pa Kun dga'
rdo rje (1309-1364). As we have mentioned above, TZ was also included in at least three local Kanjur
collections. In the two very old local Kanjurs of Hemis and Bathang, as far as we can tell from these still
uncatalogued collections, TZ is located in the midst of several other rNying ma tantras, which might (but
need not) suggest a segregated rNying rgyud section; but in the case of the Tawang Kanjur of 1699, it seems
that it and the other sixty rNying ma texts in that unique collection were not segregated in a separate section,
but included among the other gSar ma tantras in the main rGyud section.4 However, while not explicitly
condemning it, the famous fourteenth century Kanjur compiler Bu ston (1290-1364) failed to endorse TZ as a
valid translation from Sanskrit. Hence it does not occur in Kanjurs of the Them spangs ma branch, which do
not have rNying rgyud sections. The various Kanjur traditions, then, were not in final agreement about the
Indian origins of TZ, and we too remain uncertain of its provenance.
TZ has much to offer the philologist. It is one of only two full-length, complete Ancient Tantra scriptures
recovered from Dunhuang, the other being the Guhyasamja (yet the latter is a text far more used by the
gSar ma pa than the rNying ma pa, who de facto rarely practise Guhyasamja, even while retaining it in their
NGB). TZ is furthermore one amongst that comparatively small band of Ancient Tantras still to have extant
its own word-by-word commentary. TZComm must be old, since it was found at Dunhuang, and it serves
also as our source for the Dunhuang witness of TZ, which comes embedded within TZComm in the shape of
lemmata. Yet TZComm seems to have been comparatively neglected or even forgotten by the later rNying
ma tradition:5 despite the fact that an albeit corrupt and partial version of it survives in three Tenjur editions,
it does not seem to have had a consistent presence in appropriate rNying ma collections such as the rNying
ma bKa' ma,6 and few if any of the highly learned rNying ma lamas we showed it to appeared to have had

It is unlikely that the Sanskrit word, maala, could have been similarly separated into two parts with exactly these
implications. It seems then, that this part of TZComm cannot be an unmediated translation from a Sanskrit original. It is worth
noting that Tibetan commentarial traditions sometimes break the Sanskrit word, maala, into two for the purpose of glossing
its meaning, but the connotations would not correspond neatly to the Tibetan equivalent term. For instance, Mi pham glosses
maal as essence or vital juice, and la as taking or holding, so that maala would mean, to grasp the essence enlightened
qualities. He adds that if the word is taken as a whole, it can also mean, completely circular or entirely surrounded, and hence
is translated as dkyil 'khor (maal ni snying po'am/ bcud dang la ni len cing 'dzin pa ste snying po'i yon tan 'dzin pa'i gzhir
gyur pa'am/ rnam pa gcig tu sgra 'brel mar thad kar bsgyur na kun nas zlum zhing yongs su bskor ba'i don du 'jug pas dkyil
'khor zhes bya ste/, Mi pham rgya mtsho: 136. Thanks to Karma Phuntsho for drawing our attention to this source).
See Jampa Samten 1994. This edition of the Kanjur had been comissioned and copied in the late seventeenth to early
eighteenth centuries at the temple of O rgyan gling (the Sixth Dalai Lama's family temple), on the basis of an earlier gold and
silver illuminated Kanjur (gser chos bka' 'gyur).
We emphasise that this comment applies most especially to the recent past. Our research seems to indicate, however, that the
versions of the root text included within the Bhutanese NGB, and the Tshal pa Kanjur editions, must both separately have been
compiled through extracting the root text lemmata from an edition of the commentary (see below, p.35-42, 44-45).
No version of TZComm was included in Dudjom Rinpoche's bKa' ma collection. We do not know if it was included in earlier
bKa' ma collections. A copy of the Peking Tenjur version has been included in the new bKa' ma shin tu rgyas pa compiled by
Ka thog mkhan po 'jams dbangs (Chengdu 1999: volume 80 Wu: 125-236). It has been copied anew for this collection, but it

The Textual Sources and their Significance

much prior awareness of its existence. Yet one finds citations from TZComm in the works of several early
masters, such as the ninth or tenth century gNubs chen Sangs rgyas ye shes, the eleventh to twelfth century
Rong zom chos kyi bzang po (Rong zom bka' 'bum: 397-398) and the fourteenth century Klong chen pa
(1308-1363) (bDud 'joms bka' ma volume La: 63; see also Dorje 1988: 393; we discuss the citations more
fully just below). The earlier figure, gNubs chen Sangs rgyas ye shes, in Chapter Six on Mahyoga of his
famous bSam gtan mig sgron,7 offered citations from both TZ and TZComm, yet without explicitly
differentiating between the two. gNubs first paraphrases two statements given in TZComm's Chapter One,
relating to engagement in all dharmas, and then, while discussing awareness of the sameness of dharmas,
cites more exactly a single line relating to instantaneous omniscience. Finally, he cites the first two lines of
TZ's Chapter Five, on the level of attainment. Rong zom and Klong chen pa both elected to cite from the
discussion of samayas in Chapter Two (see below, p.4, and TZComm edition, Chapter 2), although they
selected different passages to cite. Rong zom pa's citation is taken almost verbatim from TZComm although
with some words omitted, while Klong chen pa's citation more generally paraphrases the meaning of a
number of the points made in TZComm. As did gNubs before them, both Rong zom and Klong chen pa as
often as not merely indicate that their respective citations are from the Noble Noose of Methods literature
(using words such as 'phags pa thabs kyi zhags pa las), without specifying if they come from the
commentary or the root text. Such a lack of consistently explicit differentiation between the root tantra TZ
and its commentary TZComm, found in all three of these examples from the earlier literature, might prove of
philological interest, in the light of some of our discussion to follow below.
The Contents of the Thabs zhags texts and their Significance for the Historical Study of the rNying ma pa
The contents of the Thabs zhags literature amply demonstrate how historically interesting and worthy of
editing such texts can be. General features of TZ suggest that to some extent it shares historical indicators
with the type of tantric literature represented by the Sarvabuddhasamyogakinjlaavara (a version of
which the rNying ma pas nowadays count as one of the Eighteen Tantras of Mahyoga mentioned above, and
which is also listed in the Dunhuang text PT 849). This Indian tantra is historically intermediate between the
Sarvatathgatatattvasagraha and the Guhyasamja on the one hand, and the full-on Yogin or
Yoganiruttara tantras on the other hand. Sanderson locates the production of such literature from the late
eighth century through the ninth century. Historical indicators which TZ shares with it include (i) Heruka in
terrifying skull-bearing cemetery-dwelling (kplika) appearance as the main deity, with female retinue; (ii)
an abbreviated form of the introductory verses setting the scene (nidna) that were previously standard in
Buddhist scriptural literature, for example, unlike the Guhyasamja, it does not have the formula, "'di skad
bdag gis thos pa dus gcig na"; (iii) the considerable (but not exclusive) use of verse rather than more
balanced verse and prose; (iv) the inclusion of feast rites (gaacakra); (v) hints of sexual yogas; (vi) some
rhetoric of taming aiva deities; (vii) the absence of any inner yogas involving subtle physical veins and
wheels (n, cakra), which first appear only with the subsequent Yoganiruttara tantras. In this text,
however, we have no coded mantra table (sngags btu ba = Skt. mantroddhra), and moreover the rhetoric
of taming aiva deities, although present, is not elaborate.8
TZ and TZComm when taken as a whole present a complex Mahyoga system that arguably equals the
contemporary rNying ma tradition in sophistication and complexity. Vairocana is the expounder of the
tantra, and Vajrasattva his interlocutor. Vairocana and the others of the five family buddhas, together with
their consorts and retinues of bodhisattvas, make up the peaceful deities, and here we can see how TZ is aptly
named a Mah-yoga tantra, in the sense of being a Greater-yogatantra. For its peaceful maala of fifty

does clearly correspond to the Tenjur versions of the text, and an additional colophon identifies its provenance: pe cin bstan
'gyur las bthus (presumably, btus or 'thus intended), "extracted from the Peking Tenjur" (volume Wu: 236.5).
Dylan Esler of the Institut Orientaliste, Universit Catholique de Louvain (UCL), Belgium, has been working on a Ph.D on the
bSam gtan mig sgron, and we would like to thank him for drawing our attention to these citations. The citations are given in
the footnotes to our editions below.
For a brilliant historical analysis, see Alexis Sanderson 2009: 145ff.

Introduction

deities is an adaptation of the thirty-seven deity maala of the Sarvatathgatatattvasagraha that is so

basic to Yogatantra, but in TZ, the male and female figures are now paired together as consorts, and a
number of further female deities are added to complete the set. The central male deity of the wrathful
maala is a ferocious form of r Heruka with nine heads and eighteen arms,9 but we have not yet identified
this exact form in later sources. Nevertheless, very similar and clearly related forms of r Heruka or Mah r
Heruka do still occur widely as the central figures in the wrathful maalas of several important extant
rNying ma pa cycles, such as the Tshogs chen 'dus pa and the sGrub pa bka' brgyad cycles, or in root tantra
sources such as the rGyud gsang ba'i snying po.10 The central wrathful deity is surrounded by the Ten
Wrathful Deities, or Khro bo bcu. The central female is a great fearsome female deity ('Jigs byed chen mo),
specified in TZComm as Ral gcig ma (Ekaja), still to this day the main ma mo or wrathful female deity of
the rNying ma pantheon. These main deities are surrounded by a large entourage of emanations whose
names, ordering and attributes as given in the commentary remain very similar in some transmitted rNying
ma texts, including some modern liturgical texts (for a systematic exposition of the deity maalas, see the
Appendix below).
It is interesting that some later doxographers envisaged TZ, along with the famous root text for all the
Mahyoga tantras, the rGyud gsang ba'i snying po, as the two texts from amongst the Eighteen Tantras of
Mahyoga that expound Mahyoga in general (spyi'i rgyud), rather than merely a single maala, such as
that of Hayagrva or Vajraklaya (Dorje 1988: 33-35). With a similar intention, in his thor bu entitled rGyud
spyi'i dngos po gsal bar byed pa, Rong zom pa likewise singles out TZ for praise as a source for clarifying
the general topics of all tantras.11
TZ and TZComm present versions of Mahyoga theory and practice that in their details bear recognisable
resemblances to the doctrine of the sameness of all dharmas (mnyam pa'i chos) of the rGyud gsang ba'i
snying po. Like the rGyud gsang ba'i snying po, TZ and TZComm also on numerous occasions use
terminology built around the words mnyam pa or mnyam pa nyid or mnyam nyid ('even', 'evenness' or
'sameness'). This famous doctrine of the rGyud gsang ba'i snying po, the basis of which involves realising all
phenomena as primordially pure, is seen by some modern scholars (Karmay 1988: 11) as one of the
historical roots of the rDzogs chen or Great Perfection mysticism of the rNying ma pa. A very similar kind of
thinking pervades the entirety of TZComm, so that the various aspects of tantric ritual are consistently
interpreted from this more inward or mystical viewpoint. In this respect, perhaps the author of TZComm can
be seen as anticipating such views as those of Klong chen pa's famous commentary on the rGyud gsang ba'i
snying po called Phyogs bcu'i mun sel, which interiorises Mahyoga and orients it towards a rDzogs chen
view. In his Phyogs bcu'i mun sel, Klong chen pa refers to TZComm's Chapter Two on the samayas which
need not be guarded, and he additionally cites TZ on a number of further occasions. Rong zom, whose views
are seen by many to anticipate those of Klong chen pa, also offers citations from both TZ and TZComm.12
9
10

11

12

Described in TZComm Chapter Twelve: 47r-49v (see summary below p.74-75 and text in the edition below p.293-296).
For the Tshogs chen 'dus pa, see Dharma r's version in the bDud 'joms bka' ma, volume Pha, p.376 ff; for the sGrub pa bka'
brgyad cycles, see Nyang ral 1979-1980: Volume Ka: 132-134; and for the rGyud gsang ba'i snying po, see its Chapter
Fifteen, mTshams brag edition of the NGB, volume Wa: 200.
See Rong zom, Chos kyi bzang po 1976: 490, where he begins this thor bu with the following statements: rgyud dang kalpa'i
nang nas/ bstan par bya ba'i dngos po ni/ dam tshig dang/ dbang dang/ phrin las dang/ dkyil 'khor dang/ bsgom pa dang/
sngags dang/ phyag rgya dang/ dngos po 'di dgu khong khrar rgyud kun nas kyang 'byung na/ dngos po ming gdags par gsal
ba ni/ tantra kun kyi mjug gi don 'phags pa thabs kyi zhags pa las 'byung ngo/.
Klong chen pa's discussion of Chapter One of the rGyud gsang ba'i snying po (bDud 'joms bka' ma volume La: 63; see also
Dorje 1988: 393) has a citation which paraphrases some of TZComm's Chapter Two (see above, p.3). Klong chen pa's other
references (bDud 'joms bka' ma volume La: 255, 279-280, 445-446, 488-489, 618-619), however, give lexically more precise
citations from TZ's Chapters One, Five, Ten and the final words of Chapter Forty-two (see the notes to the relevant chapters of
our editions of TZ and TZComm below for the quotations). In an incomplete text from his miscellaneous writings (thor bu),
Rong zom (1976: 397-398) offers a citation from Chapter Two, which, like Klong chen pa's citation of a different part of
Chapter Two, also parallels the commentary TZComm rather than the root text TZ. Elsewhere however, Rong zom (1976: 375,

The Textual Sources and their Significance

The evidence of TZComm suggests that in adopting such an outlook in his Phyogs bcu'i mun sel, which
became so definitive for much of the later rNying ma pa, Klong chen pa was not innovating: on the contrary,
such interpretations were current five hundred years before his time.
One example of how such interiorisation works in TZComm is the description of empowerment rites
found in its Chapter Three. Usually, empowerments are described in terms of complex ritual procedures
using various implements. But here, in the verse cited from the root text, Vairocana says to Vajrasattva,
"Great Being, empowerments are obtained through the expressive power of one's own innate awareness." 13
The tiny marginal notes in the Dunhuang manuscript of TZComm observe, "Empowerment can be obtained
both through ritual articles and through awareness.14 Here, (it is) through the expressive power of
awareness."15 TZComm itself explains, "When one is aware of the sameness of all dharmas, (this) is called,
obtaining empowerment through (one's own) natural qualities: that is what is meant."16 After expounding
further on this inward interpretation of empowerment, TZComm concludes with a citation attributed to a
tantra that was to become a famous rNying ma scriptural title, the Rampant Elephant: (Glang po rab 'bog):
'Not indeed from anywhere within the worlds of the ten directions / Can the buddha be found to come;/ Since
the buddha is the aware nature of mind/ Do not seek the buddha anywhere else.'17
Another aspect of this interiorization process is TZComm's exegetical interpretation of all ostensibly
pragmatic tantric rituals towards transcendental rather than mundane goals. Towards the end of TZ, for
example, we find a series of short chapters on the four rites in which homa and phur pa rituals are used to
achieve the apparently this-worldy goals of destroying, captivating, enriching and pacifying. But according
to TZComm's exegesis, these four rites are not simply concerned with the outer performance of burnt
offerings rites and liberating troublesome beings through striking an effigy with a phur pa and so on, but
with the transformative power of the ritual in the path to enlightenment. Each phur pa comes to embody an
aspect of understanding so that it can infuse the object of the rite with the realisation it exemplifies: for
example, the wrathful phur pa is, "a single phur pa of [the nature of] mind",18 and the pacifying phur pa is
"the elemental nature's faultless essential pure awareness, the samdhi phur pa, so it pacifies everything
through its natural qualities".19 At the end of each of the chapters on the four rites, the ritual description is
concluded with a verse further glossing the meanings in unambiguously soteriological terms, attributed, as in
the extract from the Rampant Elephant above, to various named rNying ma tantras. TZComm in this way
cites or refers to a good number of other tantras, including several with titles corresponding to prominent
members of the Eighteen Tantras of Mahyoga. However, we have not located the quoted passages in the
extant scriptures of the same names, and it appears that they may not be intended as exact citations in any
case (see below, Textual Analysis, Section f, p.84-86).
As indicated by our discussion of the empowerment rites above, the Dunhuang manuscript version of
TZComm enjoys the added feature of copious marginal notes in a tiny handwriting, expanding on the
commentary. These anonymous marginal annotations are a valuable source of historical data, and exist
nowhere else, since the Tenjur versions of TZComm did not reproduce them. For example, they mention
ntigarbha on one occasion and three times speak of Sambhava or Padmasambhava, and in this and other

13
14

15
16
17

18
19

408) offers short lemmata from Chapter One, using words that coincide closely with the extant root text TZ. There is also a
further apparent citation (1976: 392-393) we have not located within the extant versions of the Thabs zhags literature.
See our TZ edition for this statement: sems dpa' chen po dbang 'di dag ni rang gi rig pa'i rtsal gyis thob bo/.
Here and below, the annotation uses rigs, seemingly for rig. This is not uncommon in Dunhuang and other old texts. There
are also some instances in the manuscript's main text (eg. below, 11v.6, where the Tenjur version gives rig).
dbang la yang yo byed kyis thob pa dang rigs pa thob pa gnyis la 'dir ni rigs pa'i rtsal gyis (11v.4).
chos thams cad mnyam pa nyid du rig pa na/ dbang rang bzhin gyis thob bo zhes bya ba'i don to/ (11v.5-6).
/glang po las kyang // phyogs bcu 'i 'jig rten gang nas kyang / /sangs rgyas rnyed par yong myi 'gyur/ /rig pa'i sems nyid sangs
rgyas te//sangs rgyas gzhan du ma tshol cig / zhes 'byung ba lta bu 'o/ (12r.6-12v.1).
sems kyi phur pa gcig (64v).
chos nyid ma nor par rig pa nyid ting nge 'dzIn gyi phur pa yin te/ thams cad rang bzhin gyis zhi bar 'gyur (75v-76r).

Introduction

ways (see below p.87, 95-98) help to shed light on the pre-history of the rNying ma tradition. In particular,
they make it unambiguously clear that the Thabs zhags literature was seen by the annotator as specifically
associated with Padmasambhava. We can only estimate that they were copied into the Dunhuang
manuscript20 in the mid to late tenth century, but the dating of Dunhuang texts remains too primitive to
permit any real certainty. In short, the Dunhuang manuscript of TZComm adds considerable weight to the
evidence for substantial representatives of what we now call rNying ma Mahyoga being already present
before the Dunhuang caves were closed, or even earlier: for our stemmatic analysis shows that a Tibetan
archetype of the root text, which we have sought to restore in our edition, must pre-date the Dunhuang
manuscript.21 But such continuity is hardly surprising, since TZ itself still exists within the NGB, and most
of TZComm still survives in the Tenjur, even if somewhat neglected.
However, it is worth pointing out that there are some uncertainties in the exact list of deities in all
surviving versions of the texts. TZ seems to have some inconsistencies between its mantra list for the
peaceful deities in Chapter 9, and its mudr list in Chapter 10, which moreover, appears not to be entirely
complete nor in the most logical order throughout. TZComm to some extent fills in the gaps in the deity lists,
especially in Chapter 7 for the peaceful maala and in Chapters 12 and 13 for the wrathful maala. Yet it
also adds slightly to the confusion, suggesting some names for the peaceful deities which seem at odds to
those given in the mantra and/or mudr lists, and seeming to introduce a few further inconsistencies of its
own (see the Appendix below for tables of the principal deities). In both surviving versions, TZComm's
Chapter 13, which specifies the female deities of the wrathful maala, appears to be rather corrupt in parts,
and includes further female deities without specifying quite how they fit with those presumed to be the main
set (assuming that the usual consorts and female attendants of the ten Wrathful Deities [khro bo bcu] are
intended to constitute the principal wrathful females, an assumption consistent with Chapter 12's list of the
male deities). Furthermore, Chapter 11 which also in other respects does not fit entirely comfortably with
the text as a whole (see e.g. below, p.39 note 7) lists further wrathful female deities not included in Chapter
13. We cannot be sure, but a possible contributory reason for the later rNying ma pas' relatively lower level
of interest in the Thabs zhags tradition, in comparison with, say, that of the rGyud gsang ba'i snying po,
might be these uncertainties and corruptions in the texts inherited by the tradition.22
The Dunhuang Manuscript as a valuable source from the 'Time of Fragments'
Perhaps nowhere in the Buddhist world do questions of the production and reception of texts, of textual
and cultural translation, and of the historical transmission of Buddhism across time and place, appear so
fascinating yet more obscure and less understood, than in the occasion of the early transmission of Tantric
Buddhism to Tibet. This is because for a crucial period of one hundred and fifty years, from the mid-ninth to
the start of the eleventh century, much of the Tibetan historical record was obliterated, within a period of
civil war and the collapse of the Tibetan state.
We know a little more about what happened before that disaster. Although refracted and patchy, we do
possess some historical records of the efforts of Tibetan emperors to introduce Buddhism to their country
between the eighth century and the mid-ninth century, in what is traditionally termed the early diffusion of
Buddhism in Tibet (snga dar). We have sources describing their invitation of famous foreign Buddhist
masters to Tibet, including ntarakita, Kamala la and many others. Likewise we have sources for their
Imperially sponsored Buddhist foundations, their huge official translation and lexicographic projects, and the
20
21

22

We present evidence below (see p. 32-33) which suggests that they were copied from a previous exemplar.
This is because the Dunhuang version already incorporates indicative scribal errors shared by the Tshal pa Kanjur and
Bhutanese NGB versions, but avoided by the texts of the South Central NGB and the local Kanjurs (see Textual Analysis,
Section d below, especially p.50-54).
These uncertainties remain even today, when modern reprographic technologies facilitate such easy comparison of multiple
witnesses, such as the Dunhuang manuscript, and the local Kanjur versions that originate from such far flung regions of Tibet.
TZ's textual problems might have seemed even more intractable in the past, before such technologies existed.

The Textual Sources and their Significance

way they handled some of the controversies following the introduction of the foreign religion to Tibet. Yet
after the mid-ninth century, when the three great empires of the Tang, the Uighurs and the Tibetans
simultaneously suffered political calamity, our historical record in Tibet becomes much thinner. Tibet enters
its notorious 'time of fragments' (sil bu'i dus) and for around one hundred and fifty years we know very
little about what went on.
However, when the historical record picks up again with the so-called later diffusion of Buddhism to
Tibet at the turn of the eleventh century (phyi dar), we seem to encounter a country transformed. The
evidence suggests that Tibet had entered the 'time of fragments' with Buddhism still working to
institutionalise itself at a grass roots level, but it seems to have emerged as a country more or less
predominantly Buddhist at all levels. Tibet entered the 'time of fragments' as a disintegrating empire ruled by
martial aristocratic clans, but emerged one hundred and fifty years later with old and new leading families
alike striving to project themselves as a religious aristocracy. As Samten Karmay has observed (1988: 9), it
had entered the 'time of fragments' as a country where the state actively propagated exoteric Buddhism, but
severely restricted, curtailed and some would say even outlawed the translation and practice of the more
radical types of esoteric tantric Buddhism that used kplika imagery; it emerged from the period as a
country overwhelmingly dominated by such radical forms of tantric Buddhism, with a substantial associated
tantric literature, both indigenous and translated. It had entered the 'time of fragments' with no single
religious specialist identified as the national patron guru; it emerged from the period with a cult and legend
of Padmasambhava as a national patron guru clearly developing. It had entered the 'time of fragments' with
nothing much resembling the clerical Bon religion of today; yet shortly after the period, we find the present
form of the Bon religion beginning to emerge.23
It appears that arguably the most significant transformations in Tibetan history occurred within a one
hundred and fifty year period for which we have only the slenderest record of events. Moreover, the
transformations of this period established fundamental cultural patterns of great importance that still persist:
this was truly a formative period in Tibetan history, out of which came the highly influential rNying ma or
"Ancient Tantra" school, with its hereditary tantric lineages, its powerful cult of Padmasambhava, and the
Mahyoga, Anuyoga and Atiyoga tantric systems that remain so hugely popular to this day. In this context, it
is interesting to note that the social-historical origins of the Yoganiruttara or Yogintantras favoured by the
Tibetan New Translation schools is equally obscure. Although hugely influential to this day, we know very
little about the conditions or circumstances of their production south of the Himalayas at a time not distant
to, and under conditions of political decline not entirely different from the sil bu'i dus in Tibet.24
In many popular discussions of Tibet's conversion to Buddhism, terminological confusion arises through
an unreflective and simplistic use of the Tibetan terms, snga dar and phyi dar. Especially if not analysed too
carefully, Tibetan historians seem to speak of only two main phases of Buddhist dissemination: a late eighth
to mid-ninth century Imperially sponsored snga dar, and a late tenth to eleventh century phyi dar beginning
with the new translations of Smtijnakrti and Rin chen bzang po. In much traditional writing, rNying ma
tantras are primarily linked with Padmasambhava and the associated Imperial period and counted as snga
dar. But this can become an occasional source of confusion, with some voices inaccurately allocating the
first proliferation of the rNying ma tantras to a snga dar understood as late eighth to mid-ninth century, and
some other voices equally inaccurately allocating it to a phyi dar understood as late tenth to eleventh century.
23

24

Jacob Dalton (2011: 5-8) discusses the sil bu'i dus as a culturally creative period which later came to symbolise chaos and
darkness, in contrast to the Imperial glory which preceded it. This characterisation is generally apt, although it is also true that
rNying ma pa historians were not wholly negative about the period, retaining some positive memories of the continuity of their
lineages throughout the era.
This point point is made very clearly by Pter-Dniel Sznt in his description of 'The Dark Ages in India', the opening section
of his article, "Before a Critical Edition of the Sampua", in Cppers, Mayer and Walter forthcoming. The most sustained
attempt so far to investigate or speculate about the social-historical origins of the Yogintantras is Ronald Davidson 2002.

Introduction

The evidence found so far suggests that even though such kplika-style texts did exist in India at the time of
Khri Srong lDe'u btsan (Sanderson 2009: 145ff.) and so might have been then translated into Tibetan in some
restricted manner, the widespread proliferation and popularisation of what we now call rNying ma tantras
came later. It began either towards the very end of empire (we have little direct evidence for this, but the
possibility cannot be discounted), or, more certainly, in the hundred and fifty years after the fall of empire,
for which we have plentiful evidence from Dunhuang. This means it actually occured after the snga dar as
popularly defined, but before the phyi dar as popularly defined. The rNying ma tantra's first proliferation
could be said to be located in the snga dar only if one clearly understood the snga dar to persist in full flood
continuously up to the late tenth century; but some do not interpret it that way, instead implying the real snga
dar to be co-terminous only with the late Empire, and wrongly seeing the post-Imperial century and a half as
a chaotic 'time of fragments' (sil bu'i dus), in which no such major cultural proliferation could have
happened. The mistake here is perhaps a failure to understand that religious culture, and especially tantric
religious culture, can genuinely flourish in politically chaotic conditions. It might have been more felicitous
for some purposes if the Tibetan chos 'byung authors of the past had instead adopted a three-part convention,
counting the rNying ma tantras' first widespread dissemination as a bar dar, a third and culturally distinctive
middle phase of Buddhist expansion falling between snga dar and phyi dar, and which gathered steam
during the sil bu'i dus. But the first priority of Tibetan historians was to connect the rNying ma tantras with
Padmasambhava and other famous personalities of Khri Srong lDe'u btsan's reign, so that they were more
conspicuously concerned with the putative Imperial first seeds of rNying ma tantrism than with its
historically very late-Imperial, or more verifiably post-Imperial, development and proliferation. One should
add, learned traditional historiography could be far more complex than any such simplistic snga darphyi
dar binary. Dudjom Rinpoche, for example, systematically divides the rNying ma propagation into three
periods in his Chos 'byung, in saying that the rNying ma tantric teachings "fell first to gNyags, fell to gNubs
during the intermediate period, and fell to Zur in the end" (Dudjom 1991: 599).25 Several authors even
employed the term bar dar, although with a number of quite different meanings, and perhaps few if any in
the exact sense that we discuss here.
In the last few years, following the digitisation and wider distribution in usable form of a significant
proportion of the archaeologically recovered Dunhuang Tibetan texts, some evidence from the obscure sil
bu'i dus is beginning to become more easily available. Yet this too has inherent limits. Current scholarship
believes that internal textual clues can help locate a few of these documents (such as the Annals) to the
Imperial period, and a few others, such as PT 849, to the late tenth century (Kapstein 2006: 10-17). Yet given
the currently still preliminary nature of Tibetan palaeographic scholarship, it remains beyond our capability
accurately to locate the bulk of the Dunhuang tantric texts to precise decades within the long stretch of the
'time of fragments', so that in the absence of clear evidence, it is hard to draw conclusions about tantric
developments through this important period. It is currently suggested by some scholars that the Dunhuang
manuscript collections came from a storehouse of the Three Realms (Sanjie) Monastery. Rong Xinjiang
reports that in the tenth century, a monk of this monastery named Daozhen collected considerable additions
to his monastery's library stocks, so that a proportion of the Dunhaung texts might originate from Daozhen's
efforts (Rong Xinjiang 2000; see also Takeuchi, forthcoming). However, the store also contained earlier
materials, from the time of the Tibetan occupation of Dunhuang and throughout the intervening period.
Yoshiro Imaeda discusses a number of features of the history of the collection, but like many other scholars,
emphasises how little we know about the provenance of the various documents (Imaeda 2008). The library
cave contained texts from a very long period, from Imperial times until the eleventh century, and moreover,
documents were removed in the early twentieth century without proper records of which part of the cave they
came from, so unfortunately, we can never know. Hence we need to exercise caution in fixing dates for
manuscripts with no clear clues. In addition, even if we were able to date them all accurately, the Dunhuang
25

Dudjom Rinpoche understood gNubs to have lived for 111 years, beginning as a direct student of Padmasambhava, and
continuing 37 years beyond the death of Glang dar ma (Dudjom 1991: 607-614).

The Textual Sources and their Significance

texts at best comprise a partial and possibly unrepresentative sample of the total manuscript corpus of their
time, moreover all taken from a single multi-ethnic location, situated at a geographical and political
extremity of the Tibetan cultural world.
Nevertheless, despite such limitations, the Dunhuang documents are quite extensive, and do offer our best
available sources for understanding an extraordinarily important period in Tibetan history. On the horizon
are other promising sources, the most important of which are the great volume of so far largely unread
contemporaneous Tang dynasty Chinese sources describing Tibet and the Tibetans, as well as the findings of
ongoing largely Chinese archaeological excavations within Tibet.26 Of the Dunhuang sources, however,
there is no doubt that the manuscript of TZComm we examine in this book is an exceptionally fine specimen.
It offers possibly our best and certainly our most sustained window into the early doctrinal world of the
tantric systems that were later to be known as rNying ma pa. Moreover, depending on the date of the original
composition, it potentially offers a comparatively early window into the initial phases of the type of radical
Buddhist Tantrism that was eventually to become dominant in Tibet, in which skull-bearing cemeterydwelling kplika symbolism was prominently employed.

26

The huge quantity of unexplored Tang dynasty sources that could shed some light on Tibet in this period are described by
Bianca Horlemann in her article, "Tang Dynasty (618907). Sources for Tibetan Empire Studies: A Bibliographic Essay", in
Cppers, Mayer and Walter, forthcoming. The extraordinary potential of archaeological excavations for elucidating this
period is made clear in the contributions to the same volume of Guntram Hazod, "The Plundering Of The Tibetan Royal
Tombs: An Analysis of the Event in the Context of the Uprisings in Central Tibet of the 9th/10th Century", and of Amy
Heller, "Observations on Painted Coffin Panels from Tibetan Tombs". A more comprehensive review can be found in the PhD
dissertation of Tao Tong 2008.

A Summary of the Salient Points of our work on the Thabs zhags Textual Tradition
First, scrutiny of the textual variants of the twenty-one different versions of TZ we consulted exposed four
distinctive textual groupings and four unique single witnesses.
One of the textual groupings, and three of the single witnesses, descend unproblematically from an
archetypal TZ text, it would appear quite independently of one another. These direct descendants of an
archetypal TZ text are the three surviving witnesses of the South Central NGB recension, and the local
Kanjurs of Tawang, Hemis and Bathang.
We can deduce that two further textual groupings of TZ descended from a word-by-word commentarial text
that contained the root text as lemmata, that is, they extracted TZ from a version of TZComm that had
contained TZ as lemmata. The first of these comprises the eight Tshal pa Kanjur texts (which here must
include the sDe dge xylograph NGB because it re-used the woodblocks of the sDe dge Kanjur), and the
second comprises the four Bhutanese NGB manuscripts.
A fourth textual grouping and a fourth single witness have still retained their full commentarial character and
are thus witnesses to the root text only through the lemmata they contain: these are the three extant witnesses
to the Tenjur redaction of TZComm, and the single witness Dunhuang manuscript TZComm.
Second, analysis of the patterns of indicative errors within these four textual groupings and four single
witnesses resulted in a stemma codicum that felicitously had more than two lines descending directly from
the archetype, a necessary condition for the application of stemmatic reasoning to arbitrate between
conflicting readings. The analysis also offered clear evidence for the existence of two hypearchetypes, and
weaker evidence for the possible existence of a third. (i) A variety of errors and other readings shared
between the Tenjur, Dunhuang manuscript, Tshal pa Kanjur, and Bhutanese texts indicated that these
witnesses must all descend from a common ancestor not shared by any of the others, which we designated
hypearchetype b. Thus when taken together, these four constitute a major branch of the transmission, united
in error against the South Central NGB and the three local Kanjur texts. (ii) The Dunhuang manuscript, Tshal
pa Kanjur, and Bhutanese NGB versions share indicative errors, notably a significant accidental loss of text
in the final section of Chapter Ten, which does not afflict the Tenjur, nor the South Central NGB, Tawang,
Hemis and Bathang witnesses. The material in question is necessary to TZ, so that even prior to our belated
collations of Tawang, Hemis and Bathang, it was already evident that the South Central redaction could not
have simply added the passage as an expansion. With this and other indicative errors, we can infer
hypearchetype c, common ancestor of the Dunhuang manuscript, Tshal pa Kanjur, and Bhutanese NGB
versions, and itself a descendant of hypearchetype b. (iii) A third possible hypearchetype is d, which, if it
existed, was a version of TZComm from which descended the Tshal pa and Bhutanese versions of TZ.
But the crucial point arising from stemmatic analysis is that we have five separate branches descending
directly and independently from the archetype: hypearchetype b, the South Central NGB editions, Hemis,
Tawang and Bathang. Through their testimony, we can by logical deduction identify many readings with a
high probability of their having been in the archetype.
It is perhaps instructive to reflect on one of the implications of this second point above: contrary to
popular expectation, the Dunhuang manuscript does not represent a pristine ancient version avoiding all the
scribal errors accumulated through centuries of copying since the eleventh century sealing of the Dunhuang
cave library. On the contrary, repeated cycles of copying have also left their damaging mark on the
Dunhuang manuscript itself. It has its own unique errors, such as the accidental loss of passages of root text
lemmata found in all other versions, and the misplacement of other passages, presumably the result of
disarranged folios during a previous copying. Moreover, as noted above, it shares many errors, including a
major omission, with some of the more popular extant witnesses of TZ, yet these errors are avoided by other
more obscure witnesses in the form of the South Central NGB and the local Kanjur texts. So for those who
may doubt the antiquity of the rNying ma tantras, it is worth reflecting not only on the self-evident external
evidence, namely that the existence of the Dunhuang manuscript demonstrates that TZ and TZComm must
date from some point prior to this late tenth or early eleventh century copy. One must in addition reflect on

Salient Points on the Thabs zhags Textual Tradition

11

the historical and also bibliographical implications of the internal evidence, which now shows that a lessfavoured branch of the extant NGB transmission, and some decidedly obscure local Kanjurs, are in many
ways our most reliable surviving witnesses, reproducing readings from a source older still than the
Dunhuang manuscript's two deducible error-bearing ancestors b and c. The ancestor of the South Central
NGB and local Kanjur texts was therefore as yet untainted by the combined transmissional faults of b and c,
which the Dunhuang manuscript shares with the more popular extant editions such as the sDe dge NGB,
Tshal pa Kanjur and Bhutanese NGB. We have no certain way of dating this earlier Tibetan version of TZ
(which is of course our archetype a). Although it is not logically possible to draw temporal conclusions of
any reliable historical significance purely from transmissional evidence, the layering and density of
transmissional errors already accumulated in b and c and hence already reflected in the tenth or eleventh
century Dunhuang manuscript, render it eminently possible (albeit unprovable) that a goes back a long way,
perhaps even as far as the Imperial period. It is worth adding that the South Central NGB may transpire not
only to represent a transmission established in a specific regional area. It is possible that its original source
may have been a rDo rje Brag manuscript, or even one from sMin grol gling, and thus could have been much
more culturally central than might be supposed from the currently known extant witnesses.
Third, it has also been possible to recover significant sections of TZComm lost or misplaced in one of its
two extant versions but found in the other: the Dunhuang manuscript has proven invaluable in restoring long
sections lost from the Tenjur versions, while conversely, some much smaller but still significant omissions
and misplacements in the Dunhuang manuscript can be recovered from the Tenjur.
Fourth, although TZComm incorporates the entirety of TZ as lemmata, it is not always clear or consistent
in marking these lemmata as such. A little unexpectedly to us, it further transpired that this fuzzy
demarcation between lemmata and commentary within the body of TZComm eventually issued into
significant differences between the three currently most popular recensions of TZ itself, namely the Tshal pa
Kanjur (including the sDe dge NGB), the Bhutanese NGB, and the South Central NGB. As we have
intimated above, investigation revealed that both the Tshal pa Kanjur and the Bhutanese NGB derived from
separate efforts to extract or reconstruct TZ from out of TZComm. Yet because the lemmata were not clearly
demarcated, their redactors made different decisions at various points about what was root text and what was
commentary. Thus, the Tshal pa Kanjur version has some erroneously lengthened sections where their
redactor mistook parts of TZComm as TZ. The Bhutanese NGB recension likewise both adds and omits
materials at different points. The South Central NGB recension by contrast is descended from an archetypal
TZ and was never confused in reconstruction from TZComm. The contrasting decisions about which
passages of TZComm constituted the lemmata account for all significant recensional differences between
these three currently popular versions of TZ. Virtually all other variations are transmissional in nature,
purely scribal errors afflicting one version or the other, or minor spelling corrections and so forth.
Furthermore, although we have been able to conclude through stemmatic analysis that it is the South Central
NGB and three local Kanjur recensions which accurately reflect the original boundaries of TZ, it would have
been hard or impossible to adjudicate between the different recensions in this way without stemmatic
analysis: were it not for stemmatics, the different recensions of TZ would all ostensibly offer equally good or
bad claims. Yet such unresolvable indeterminacy in itself offers a valuable clue to understanding the manner
in which the rNying ma textual tradition has existed over the centuries, and how it can accommodate a
certain level of variation within its texts across their different recensions and editions.
Fifth, the statements found in the Dunhuang manuscript suggesting its strong connection with
Padmasambhava invited a comparative study with the Man ngag lta phreng (Pith Instructions on the
Garland of Views), the one work which modern scholars see as credibly attributable to Padmasambhava.
Our comparative study showed no indication that the texts were by the same author. Perhaps more
interestingly, what we do find in the Dunhuang manuscript's mentions of Padmasambhava is an early
association of the figure of Padmasambhava at the highest level of the Mahyoga transmission, not simply as

12

Introduction

a human scholar composing a commentarial work, but rather as a realised being, in some ways recognisable
from the later gter ma traditions.
Finally, the Dunhuang manuscript of TZComm is one of the most valuable items in the British Library's
Stein collection. It is a fortunate and remarkable textual survival, portraying a highly developed Mahyoga
system from post-Imperial Tibet. Its teachings claim to derive from Padmasambhava's tradition. In them, we
find a detailed exposition of tantrism as a path to enlightenment, doctrinally similar to the rGyud gsang ba'i
snying po, and striking in the way it turns even the ostensibly most worldly of rites towards a soteriological
purpose. We hope our edition of it might stimulate some interest in a more popular re-appropriation of this
remarkable text, as happened with the Saghastra following Prof Giotto Canevascini's edition (1993) of
the Gilgit manuscript. We are gratified to hear word of its inclusion in a forthcoming redaction of the rNying
ma bKa' ma.1

As yet we have no confirmation or details, but we understand that the Dunhuang manuscript version of TZComm may be
included in a new bKa' ma apparently being compiled in Eastern Tibet.

Methodological Issues in the Study of the Thabs zhags Textual Transmissions

Here, we reflect briefly on the culture of tolerated scriptural variation that we find in rNying ma
Buddhism, which raises methodological issues for textual scholars of rNying ma tantras such as TZ.1
Although our analysis is specific to rNying ma texts, it might also shed useful comparative light on other
genres. In comparing appropriate methodological approaches for the NGB and for the Kanjur, we make two
main points:
1. While current Kanjur scholarship is, for entirely compelling reasons, in many cases abandoning the
hope of recovering unitary original texts or archetypes through philological analysis, study of the rNying ma
canon suggests the opposite.2 Our admittedly meager analysis so far tends towards the provisional conclusion
that some NGB tantras may have had original redactorial moments, and should, in theory at least, present
archetypes recoverable through philological methods. We certainly do believe we have succeeded to a useful
extent in recovering an archetype of TZ, and present our evidence for this in this volume. However, note
well that neither the terms 'original' nor 'archetype' need always imply in the context of tantric texts
independence from borrowings from earlier texts, nor even freedom from orthographical or grammatical
error! Those are quite different issues.
2. If Kanjur scholarship is currently emphasising the non-unitary and varied nature of texts before their
incorporation into Tibetan canons, our work on the NGB is currently emphasising diversity after
incorporation into Tibetan canons. Hence our second point, a little ironically, slightly devalues the first. Even
if historians of the rNying ma pa might (as we do here to some degree) enjoy the luxury of recovering very
early archetypes or even originals through philological methods, this luxury is of partially limited value
because the rNying ma pa themselves do not operate in quite this way. On the one hand, their tantras have
come to vary over time through scribal error and piecemeal attempts at correction, or, as in this instance,
confusions between root text and commentary. On the other hand, the rNying ma pa have never sought to
establish a centralised authority that could standardise their scriptures. Nor do they systematically identify or
specify in their tantra catalogues the different versions of a text. Of course, there is little problem where one
reading is clearly better than another bad readings can be eliminated without controversy but the
cumulative effect of centuries of dissociated hypercorrections made without recourse to other editions, leads
to an occasional variety of good readings, each equally plausible, each the potential basis of further learned
exegesis. For example, the commentarial tradition on the tantric deity rDo rje Phur pa had to make sense of
two rather different readings within the root verses on the creation of the deity's maala, even though these
root verses are so important that they are shared by all the scriptural texts, and repeated in all major practice
texts.3 Hence, despite in many cases apparently starting out with unitary texts or redactorial moments, the
rNying ma pa are by now no better off than the Kanjur tradition. They have had to accept that in different
editions and in different regions, ostensibly the same versions of important tantras can vary somewhat, at
some points displaying what appear to be equally viable yet different readings. Yet NGB texts tend to be
1

TZ is of course a rNying ma tantra, although it was also adopted into some Kanjurs. Thus, for the purposes of the present
discussion, TZ is to be characterised primarily as an NGB text, and only secondarily as a Kanjur text.
Since Kanjur scholarship is still very much in its infancy, very little can be asserted about it with any real confidence. It might
well transpire that many Kanjur texts do in fact descend from unitary sources.
Generally, commentarial writings will consider only the readings witnessed in the specific liturgical tradition on which they
are commenting, but occasionally, variation is acknowledged. Kong sprul's commentary on the rDo rje phur pa rtsa ba'i
rgyud kyi dum bu (bDud 'joms bka' ma volume Tha: 78-80) explicitly recognises one variant in the first root verse (thig le/
thigs pa), and implicitly acknowledges another (sgor shar/ gor shar), elucidating the text with reference to the alternative
connotations. Kong sprul (bDud 'joms bka' ma volume Tha: 92-93) furthermore draws attention to a contrast between the
reading of the Kanjur's rDo rje phur pa rtsa ba'i rgyud kyi dum bu and of the NGB texts for the second root verse, suggesting
they can be understood as variant translations, and reiterating that the significance of these profound vajra words of the root
tantras is the same in both cases (bsgyur ba gnyis don ni gong du bshad pa dang 'dra'o, volume Tha: 92.6). See the discussion
in Cantwell 1997: 115.

14

Introduction

ritual if Mahyoga or poetic if rDzogs chen, and for such materials, moderate differences in wording do not
have the same impact they might have for philosophical or doctrinal tracts. So long as they all point to the
correct enlightenment, a moderate degree of discrepancy between different versions is often not even
remarked upon. Over the course of three decades of research in which we have worked with a number of
lamas,4 we have several times encountered such variation, and while not particularly welcomed, it has never
so far been rejected as catastrophic. Besides, for the last many centuries, the rNying ma lamas have de facto
had little option other than to live with it, and perhaps partly making a virtue out of necessity, but perhaps
equally because of their ontological beliefs about lamas and tantric texts, they largely take it in their stride. It
seems then that they share a distributive understanding of knowledge, very like the understanding of
Mahyna and tantric scriptures that often prevailed in India, in which each sound and meaningful variant
version can be appreciated for contributing its partial vision of the Buddhas' total authorial intention (bad
readings are of course rejected without hesitation as the mistakes of scribes). The point should not be overstated: lamas will often insist on particular readings established in commentarial traditions in which they
have been taught. Yet they will hesitate to dismiss or criticise alternatives presented by other lamas of
different commentarial traditions, and when pressed, may affirm that the alternative readings are valid for
that other lineage of descent. To approach this within narrowly political terms: if a lama takes too strong a
stand in rejecting one plausible good reading in favour of another plausible good reading, he incurs the risk
of unwittingly challenging some other respected authority. If he could have near-total knowledge of all
previous and present views, then he might be happier to take such a risk, but such complete knowledge is
seldom available. What then, if Rin po che A were to take a radically exclusivist position today, only to find
out tomorrow he had in doing so inadvertently labelled gTer chen B or mKhan chen C as definitely
mistaken? The embarrassment could be considerable. So dogmatism tends to be avoided, and the range of
good readings cautiously tolerated. For most rNying ma pa, the perfect and complete text of a tantra is not an
historical archetype recoverable philologically: it is something that exists only in the tantric heavens or in the
guardianship of the kins, of which extant terrestrial versions, including the philologist's, are in all
likelihood little more than imperfect partial reflections.5 We are not yet certain to what degree and in what
way this partly ideological and partly pragmatic view is peculiar to the rNying ma tantras, and to what degree
it pervades other genres as well. We are aware of the variant versions of texts like the Heart Stra preserved
in the Kanjur (Silk 1994: 31-41), and we are equally aware of the perspective of modern scholars like
Jonathan Silk and Paul Harrison, who understand Indian Mahyna scriptures as works in constant motion
that never indigenously achieved a fixed entity. In addition, we are aware of the works by colleagues in
parallel fields, notably the modern Talmudists, who have clarified the way in which anonymous collective
scriptural authorship may work on the ground. Nevertheless, we do not want to extrapolate from other fields,
but will ensure through careful collations that the rNying ma texts talk to us directly with their own historical
message.
The Ancient Tantra Collection (NGB)
The rNying ma school of Tibetan Buddhism, like the Bon, has the unusual distinction of basing its major
tantric systems upon scriptures largely excluded from the Kanjur. Even TZ was not accepted by Bu ston (see
above p.2). The rNying ma response was to consolidate their tantras within their own compilation of the
NGB, a process that achieved increasing maturity in the fifteenth century. In its fullest editions, the NGB
nowadays includes around one thousand works, in about thirty-five thousand folios, or seventy thousand
4

The lamas we have worked with include in particular Lopon Ogyan Tanzin (the consultant on several of our AHRC projects);
Lama Kunzang Dorjee of Jangsa Monastery, Kalimpong; Alak Zenkar Rinpche; gSang sngags Tulku (while he was at Shechen
Monastery, Kathmandu), and others on a more occasional basis. We have also consulted Dr Karma Phuntsho on this specific
issue.
After writing the above, we noticed David Gray's article in JIATS (2009), which discusses different versions of this mythic
ideal throughout Tibetan Buddhism, including the specifically rNying ma version of the revelation of the tantras of Mahyoga
to King Ja. In conclusion, Gray makes the point that the absence of the ideal texts was a factor in preventing absolute canon
closure.

Methodological Issues

15

pages. The provenance and authenticity of the rNying ma tantras has been questioned in various ways from
the turn of the eleventh century until the present day. Some considered them translated from Sanskrit, hence
authentic; others considered them Tibetan compositions, hence inauthentic. Yet others, including the famous
eleventh century rNying ma sage Rong zom chos kyi bzang po, seemed to accept the possibility that they
were compiled in Tibet, yet nevertheless deemed them authentic.6 Either way, the rNying ma tantras have
had, and continue to have, a very powerful influence on Tibetan religion.
NGB scriptures often become more widely known through references reproducing citations given in
important commentarial works rather than through direct reading of the source texts, although learned lamas
certainly could, did, and still do, have direct recourse to scriptural texts (Cantwell and Mayer 2007: Chapter
2.V). The genre first comes into clear view in the post-Imperial period. Modern academic analysis, including
our own, finds that most rNying ma scriptures studied so far resemble what Davidson has dubbed 'gray' texts
(Davidson 2002: 212). Neither wholly Indian nor wholly Tibetan, they are Tibetan compilations in the style
of Indian tantrism comprising predominantly Indic type materials, with some Tibetan admixture and
localization.7 Our most reliable sources for the early phases of this literature are of course the Dunhuang
tantric texts. Recent research on them reveals a sophisticated and complex tantrism demonstrably continuous
with the rNying ma tantrism of later centuries, although with interesting differences too. More specifically,
detailed comparative examinations of Dunhuang tantric materials with texts from the NGB now shows with
certainty that the NGB does indeed conceal within its vast bulk a great deal of genuinely pre-gSar ma pa
tantric materials.8 For those interested in studying the early history of tantric Buddhism in Tibet, the NGB is
thus a potential treasure trove of information.
Textual Obscurity and Scribal Corruption in the extant Ancient Tantra Collection
Tibetology has known the historical potential of the Ancient Tantras since Rolf Stein's time, yet use of
them has remained slight, because they are so difficult to consult. One major problem is the lack of
commentarial literature: only a very small proportion of NGB texts have extant commentaries, so that the
root verses on their own frequently remain obscure even to the most learned lamas.
Another problem is scribal corruption. We know from comparisons with Dunhuang manuscripts that
much of the material in these texts is around one thousand years old, ample time for scribal errors to appear.
When editing two NGB texts some years ago, we found around one word in three differing between our six
witnesses. If punctuation was included, there was a statistical average of one variant for every three or four
6

Many (although not all) of these tantras are presented with Sanskrit equivalent titles and colophons naming one or more of the
illustrious figures from the traditional historical accounts of the early transmissions as translators. Some later rNying ma
voices argued that these texts were indeed verbatim translations from Sanskrit originals from the reign of Khri Srong lde'u
btsan. By contrast, much (not all) modern scholarship locates them somewhat later, mainly after the collapse of empire around
842. Interestingly, the criterion of Indian provenance used by the compilers of the Kanjur to judge the authenticity of a tantra
was not always fully accepted by rNying ma pa scholars, such as the influential Rong zom pa. As Dorji Wangchuk puts it
(2002: 282), "Rong zom pa's response... does not categorically rule out the possibility of the tantra being a compilation or a
composition by a Tibetan scholar... but rather addresses his opponents from a stance of spiritual ethics, trying to persuade them
that in spite of such a possibility, one should approach the text with reason and respect on the basis of its scriptural
coherency". Wangchuk presents some passages from Rong zom pa's work; perhaps most pertinent is the point that the
buddhas need not be restricted by time or place, or to superior buddha-like bodies, but arise in response to the needs of sentient
beings. Thus (in Wangchuk's translation), "even if tantric treatises are taught with overlaps and so on, and even if it is possible
that they were compiled and composed by [Tibetan] Updhyyas, they should not be considered objects of doubt, for the ways
the blessings of the tathgathas appear are not restricted" (rgyud kyi gzhung ldab bu la sogs par ston pa dang/ gal te mkhan po
rnams kyis bsdus shing sbyar ba srid na yang/ de bzhin bshegs pa'i byin gyis rlabs byung ba la tshul nges pa med pa yin pas
the tshom gyi yul du bya ba ma yin no/, Wangchuk 2002: 283-284).
Nonetheless, as we note above, we have no clear evidence which would prove TZ a "gray" text; it might well be a translation,
despite its lack of a translator's colophon.
Our work on the Dunhuang texts relating to the tantric phur pa practices demonstrates substantial passages in common
between the Dunhuang materials and NGB texts. See especially Cantwell and Mayer 2008: Chapters 5 and 6.

16

Introduction

syllables. In addition to such small-scale variants, there are also larger ones, where longer passages, whole
folios and entire chapters can vary, be lost, misplaced, or otherwise jumbled. Complete chapters can differ
immensely across the different editions, to the extent that the average reader might wonder if they are the
same text at all.
When faced with such textual difficulties, Tibetan lamas, like their Western counterparts, seek out other
editions. Before the Cultural Revolution, there were undoubtedly more versions than there are today. As the
major repository of their scriptural tradition, every major rNying ma monastery once held a copy of the
NGB, and several hundreds must have existed in the 1950's. Yet after the Cultural Revolution, only a handful
are now available to scholars like us (it is to be hoped that more will turn up in due course): four from
Bhutan which are almost identical, a xylograph edition from sDe dge in East Tibet, and four manuscripts
from northern Nepal and Southern Tibet which are related to one another.9 In short, we can currently muster
only nine witnesses, representing three distinct traditions, and with the exception of the sDe dge xylograph,
mainly drawn from the geographical peripheries of the Tibetan cultural region. Many famous manuscript
editions of the past seem to have been lost, such as the collection made by Ratna Gling pa in the fifteenth
century, or 'Jigs med gling pa's eighteenth century edition, and the library copies from major rNying ma
centres like Mindroling (sMin grol gling) and Ka thog have not yet reappeared, and might not have
survived at all.
Increasing the usability of the Ancient Tantra Collection texts
Being so early and so influential, there is clearly an incentive to study NGB texts. The question is, how?
Can we simply start reading them, with no particular regard to which edition, and without comparing the
different editions? Or should we try to critically edit them first, carefully comparing manuscripts and
assessing variant readings? Purists might argue that work on unedited NGB texts is unsafe, yielding little
more than random impressions. Pragmatically speaking, that goes too far: especially at this very early stage
in their study, useful insights can certainly be gained by browsing the texts just as one finds them. But
undoubtedly we get a much finer and more nuanced understanding by editing the texts.
Tibetans scholars certainly made serious efforts to edit these texts, but circumstances conspired against
them. The great bulk of the volumes, their rarity and expense, and the long distances separating the Tibetan
cultural regions meant that rNying ma editors could only rarely assemble a fully representative collection of
their tantras for comparative purposes. Bringing one huge collection from far away was difficult enough;
assembling all editions from everywhere almost impossible. Of necessity, rNying ma editorial techniques for
such huge collections more often relied on comparison with geographically proximate editions, accompanied
by conjecture. We see evidence of this in the way that the nine available editions of the NGB follow
distinctive regional affiliations. For example, the four editions from Bhutan often remain textually almost
identical, slavishly reproducing exactly the same errors, lacunas, folio misplacements and good readings
alike. Much the same can be said for the four editions from South Central Tibet and Northern Nepal,
although across this more dispersed region, the homogeneity is less pronounced, so that, for example, some
individual South Central texts can follow the Bhutanese recension. Likewise the sDe dge edition from far off
East Tibet is entirely different from any of the South Central or Bhutanese editions, even though it is said to
have some readings from Central Tibet. What we have read so far seems mainly to suggest the dominant
influence of its several known Eastern Tibetan exemplars. Such empirical evidence suggests that, perhaps
somewhat more than with the Kanjur, regional traditions grew up, as new copies were made from editions
nearby, with only occasional admixture from far-away editions. However, two words of caution need to be
added here: the regional distinctions might in fact also have a sectarian element, and moreover are more
typical of the times after canonical formation, rather than before. All the South Central texts might be Byang
9

The National Archive in Kathmandu also holds microfilm of a fragmentary NGB from Thini written in dbu med, and other
isolated volumes (available through the Nepal Research Centre), thanks largely to the efforts of the Nepalese-German
Manuscript Cataloguing Project.

Methodological Issues

17

gter we do not yet know and all the Bhutanese ones are now associated with the Padma Gling pa
tradition, although we do not yet know their original provenance. By contrast, the sDe dge xylograph edition
drew on seven different ma dpe making no sectarian distinctions. Only time, and the finding of further
manuscripts, will give a broader picture of how regional and sectarian considerations interacted. Given that
lineage differences within the rNying ma pa school are so fluid and permeable, at this stage we tend to
emphasise regional factors. It must also be emphasised that regional differentiation grew with time, and that
individual text versions could and did travel vast distances far more easily in the pre-canonical period, before
they were thrown together into huge unwieldy collections. Hence we should see, as theory predicts, that the
genuinely early text versions are not yet so affected by such regional factors.
In addition, we find evidence of conjecture in the redaction of NGB texts: one volume of the Rig 'dzin
Tshe dbang nor bu manuscript from South Central Tibet still preserves many emendations made in red ink,
which on close analysis appear to be conjectural, made without systematic reference to other editions.10 This
resembles the editorial styles still used by Tibetan lamas in preparing new editions, in which intensive
conjectural effort is typically invested into correction of orthographic and grammatical errors. However,
where there are serious cruxes and no other editions to consult, the hazard of hypercorrection, as mentioned
above, becomes serious. Aware of such a danger, even fine editions backed by powerful scholarly monastic
institutions have sometimes preferred to reproduce obvious error, rather than attempt audaciously to insert
corrections with no textual support. In fact, it is probable that deliberate conjectural correction was much
less frequent in the transmission of these revered scriptural texts than in compilations of monastic liturgies
and so on in everyday use. The corrections throughout half of one volume of the Rig 'dzin edition were
clearly distinguished from the original text by their red colouring (unfortunately, not so clearly distinguished
for modern users of the microfilm copy), and we cannot know whether these emendations were sanctioned.
Certainly, they stop rather abruptly in the second half of the volume. In the case of the text we examine here,
it is clear that the South Central version did not introduce any significant recensional amendments: it
continues to reflect its descent from the ancient archetype, with only transmissional errors distinguishing it
from earlier readings. In the carefully edited sDe dge NGB edition, the only NGB we know to have been
prepared from a comparatively wide range of ma dpe including some from other regions, occasionally where
uncertainty occurs, alternative readings are noted in marginal annotations.11
Our belief is that at the very least, representative samples of this diverse NGB tradition should be
subjected to the best attentions of modern editorial technique, not only for the sake of modern academic
scholars for whom questions of transmission and the historical reliability of textual evidence are so
important, but also, in the longer run, for those millions of Tibetan Buddhists who revere the NGB as the
ultimate scriptural source of their religion.

10

11

We discuss this feature elsewhere (Cantwell and Mayer 2007: 74-78). These emendations, which appear not to have been
made by the original scribe, cannot be considered a comment on the standards adopted by the edition's editors, yet had a future
copyist adopted them, they would have impacted on the textual tradition. See also our discussion below (p.29, 46) of the
emendations to the sGang steng-a NGB manuscript and their possible relevance to the transmission of the wider Bhutanese
NGB tradition.
In many of the sDe dge NGB edition's texts, there are occasional marginal notes which supply alternative readings. There
were eight such notes in the rDo rje phur bu chos thams cad mya ngan las 'das pa'i rgyud chen po (Myang 'das), and two in
the rDo rje khros pa phur pa rtsa ba'i rgyud (See Cantwell and Mayer 2007; the alternative readings are most easily examined
by browsing the diplomatic editions of the sDe dge texts given on the accompanying cd). In some cases, the notes could be
seen as suggesting better readings or amending a reading, but often the variant merely suggests an alternative or underlines a
textual crux. Generally, the variants are followed by wording which could be translated as, "also/alternatively, we find...",
such as kyang/yang, yin nam, or simply, byung. In one case in the Myang 'das (sDe dge volume Zha f.76v, Cantwell and
Mayer 2007: 214 nb 572), sDe dge's two readings, mgyogs and 'khyog, essentially corresponded to differing alternative
readings of the Bhutanese and the South Central editions.

18

Introduction

How do we edit Ancient Tantra Collection texts? Can we stemmatise them? How do we account for their
variations?
The question then arises, with no precedents to emulate, how does modern scholarship approach the
editing of NGB texts? To start with, we turned to a near example for inspiration. The NGB shares the same
methods of reproduction as the Kanjur and, even if not quite so vast as the Kanjur, is nevertheless of massive
size and of comparable difficulty to transport long distances. Like the Kanjur, its texts are also mostly
considered buddhavacana, the actual speech acts of the enlightened ones. Hence it seemed rational to start
by following the lead of such Kanjur scholars as Helmut Eimer and Paul Harrison, who have attempted to
use classic stemmatic analysis.
Stemmatic analysis is a method developed over recent centuries largely by Classicists and Biblical
scholars. Its methods involve the systematic analysis of indicative errors, to infer lines of textual descent.
From this, one can reconstruct, or partially reconstruct, an earliest ancestor, or archetype, from which all
texts descend. But there are clear limits to what stemmatics can do, as scholars such as Bdier and
Timpanaro have shown.12
The Kanjur scholars quickly ran into exactly such limitations. Stemmatics is based on the premise that the
tradition is closed: in other words, that there is a single ancestor or archetype from which all existing
versions of a text descend. When Helmut Eimer first opened up this field, there were probably many
Tibetologists who hoped that a single translation of a text into Tibetan might in most cases function as such
an archetype. But later research, notably that of Paul Harrison, found that Kanjur traditions can be open.13 In
other words, he found that in some cases, the two main branches of the Kanjur transmission may from the
outset have represented different recensions of a text, differing, for instance, in the Tibetan equivalents of
Sanskrit terms.14 It goes without saying that in an open tradition, the variant versions can interact with one
another at any stage, thereby further confusing the issue. Soon after, Peter Skilling also found that in a
significant number of specific cases, texts from the Tshal pa and Them spangs ma lines cannot descend from
a common source (Skilling 1997: 101). In short, many Kanjur texts underwent re-translations or revisions of
earlier translations, and the versions in extant collections need not always stem from a single translational
effort. Recent scholarly work on further materials, such as texts or text fragments in the proto-Kanjur
collections in Western Tibet, underlines this caution. For instance, Tomabechi discovered that the Tabo
fragments of the Guhyasamja are in parts close to the Dunhuang manuscript readings, perhaps preserving
Rin chen bzang po's (9581055) early translation before the recensional amendments of 'Gos lhas btsas
(c.1050) (Tomabechi 1999: 56, 7678). Under such circumstances, the central premises of classic
Lachmannian stemmatics do not pertain, even if other useful results can still be derived from less ambitious
stemmatic analysis.
Further complications arise due to the well-known historical fluidity of the Indian Buddhist scriptures
themselves, before translation into Tibetan.15 Jonathan Silk, for example, argues that they never had a unique
12

13

14

15

See Cantwell and Mayer 2007: 14, for a discussion of Bdier 1928; for Timpanaro, see especially his essay, "Textual Criticism
and Linguistics, and their Crises at the End of the Nineteenth and in the Twentieth Century" in Timpanaro 2005: 119-138.
See Harrison 1992: xlvi-xlviii. Apart from his important pioneering studies, which began the work of clarifying the textual
relationships between the Kanjur editions, Helmut Eimer has continued to play a central role in the continuing development of
Kanjur and Tenjur studies.
Eimer 1997b: viii. Texts of proto-Kanjur collections found in Western Tibet seem to represent a further line of descent,
independent of the Tshal pa or Them spangs ma branches (Tauscher and Lain 2008: 350, 358). Of this group, in this study
we have been able to consult the Hemis local Kanjur (which Tauscher and Lain, 2008: 354, note as discovered but not yet
catalogued). There is no doubt that Hemis and the other local Kanjurs we have examined are independent from the
mainstream Tshal pa Kanjur transmission of TZ, although in this case, there is no evidence that they represent different
recensions at the outset.
Harrison (1992: xlvi) remarks, "we are already familiar with the same... with regard to Sanskrit texts: this century has seen the
demise of the notion of 'the text',... we are... accustomed to... a textual tradition".

Methodological Issues

19

compositional kernel, nor were ever subjected to a unique redactorial moment, but on the contrary, continued
to change and grow organically throughout their history. Under such circumstances, where the very notion of
an original work is negated, what could stemmatics hope to recover? Silk suggests we loosen our fixation on
quests for original works, and instead adopt the editorial methods developed by Peter Schfer for medieval
rabbinic literature, which, quite unlike the masoretic Bible, is highly diffuse.16
Naturally, we have to ask how such limitations on stemmatics might apply to the NGB texts. Are their
traditions open or closed? Do they derive from multiple revisions, or multiple translations, of possibly
varying originals? Did they ever have a unique compositional kernel, or a unique redactorial moment? Our
previous editions of NGB texts were intended, amongst other things, to test the experimental hypothesis that
since many of them were Tibetan-made compilations of largely Indic fragments, some might have had an
identifiable moment of redaction or compilation, and therefore need not be irrevocably diffuse in their
origins, even if they might have become so in their subsequent transmission. Prior to our new edition of TZ,
we had edited three NGB texts (Mayer 1996, Cantwell and Mayer 2007), and despite their considerable
corruption, and the sometimes very considerable differences between their different witnesses, we did not
find in these cases much evidence that cannot be most easily explained as the outcome of transmissional
error and attempts to correct it, as well as some cases of editorial intervention to standardise spellings of
mantras etc. One text even seemed partially amenable to stemmatic reconstruction, although bifidity stymied
stemmatic analysis for the other two (Cantwell and Mayer 2007: 79). One of the tantras perhaps gave a
greater sense of coherent redactorial vision than the other two.
With all three of these important rNying ma tantras, after attempting to account carefully for the causes of
every individual textual error or variation found within their different versions, we felt we were most likely
dealing with texts formed in an identifiable initial redactorial moment. But these texts were also in many
places demonstrably compiled from pre-existent parts, to the extent that all of them actually shared some
similar text passages at various points. It is entirely possible some such pre-existent parts were already
replete with orthographic and other errors before incorporation into their new locations. A further still
unexplored feature in early rNying ma tantrism is evidence suggesting that the same or very similar text titles
might have served as the basis for quite separate compositions at different occasions, so that several quite
different texts bearing the same or almost the same title seem to have been in circulation. However, bearing
the same title is of course not the same thing as being a variant version of the same composition. While the
issue about titles still needs more research, what we can say with reasonable certainty is that all the three
texts we studied diverged over time, and we reiterated our earlier proposal that the divergences such texts
have acquired over the last thousand years are in many cases no longer resolvable into a single 'original,
correct' text at least not within a traditional framework. In any case, an 'original' text recovered
philologically might well incorporate 'incorrect' features or incoherencies, since, as noted above, these texts
were compiled from pre-existing parts that were quite likely not in themselves error-free. In short, the most
'correct' text is not necessarily the historically earliest or even the archetype. Unsurprisingly, our philological
analysis shows that the learned redactors of the sDe dge xylograph on several occasions and with some
justification corrected problematic readings, most notably mantras, that were quite likely genuinely
inherited from original redactorial moments. As with many other bodies of world literature, textual

16

Unlike the Masoretic Bible with its single fixed form and the effective effacing of all variants, in medieval Jewish Rabbinic
literature, pericopes, long divorced from any original context, are shared between different texts, while texts and pericopes
alike vary endlessly over time and space. Since such characteristics are shared by Indian Mahyna scriptures, Silk suggests
that Schfer's analysis is relevant to Buddhist Studies (Jonathan Silk, Numata Lecture, Oxford May 19th, 2008, "What Can
Students of Indian Buddhist Literature Learn from Biblical Text Criticism?").

20

Introduction

mutability has de facto become accepted and accommodated as an inherent feature of rNying ma tantric
culture, even if reference is still made towards the ideal of a more unitary and more perfect tradition.17
The situation has now developed with our edition of TZ. As we describe below (see Textual Analysis,
Section c, p.35-42), the different extant versions of this tantra can vary quite dramatically, underlining how
prominently textual variation figures within the literary culture of the NGB. While the natural limits of
stemmatic analysis mean that no-one can assert with absolute certainty if TZ's transmission was open or
closed, it is eminently reasonable to propose a higher probability that it was closed. Since we found no
evidence so far to suggest that the extant versions of TZ descended from multiple revisions or translations
that later intermingled, we feel it simpler and more reasonable to conclude that its branches descended from
a single redactorial moment. What is most interesting is that in most of its more prestigious surviving
editions (even though not in some more obscure peripheral versions), this scripture descended in a strikingly
fluid relationship with its own commentary, and it is this that has been the major cause of the very high
degree of variation between its surviving versions today, not the openness of its transmission.
Concluding Reflections on the variations in TZ
In previous studies, we have looked at how what may seem significant variation to modern Western eyes,
is tolerated in different editions of rNying ma tantric texts. For example, we looked at the development of the
bDud 'joms gter gsar sngon 'gro text, from bDud 'joms gling pa's (1835-1904) earlier version to that of the
second Dudjom (bdud 'joms) sprul sku, 'Jigs bral ye shes rdo rje (1904-1987).18 Some years ago, we looked
at two variants within the Vajraklaya root verses, perhaps transmissionally generated, both of which have
attracted an immense degree of commentarial exegesis over the centuries from many great masters.19 Thus
TZ reminds us once again that beyond technical questions of open or closed recensions, a major desideratum
for rNying ma textual scholars is further investigation into its culture of seemingly tolerated textual variation
a topic not so far sufficiently explored.
Institutionally speaking, rNying ma tantric culture was a world of shifting, decentralised religious
authority, where no single body could establish a text definitive for all. Moreover, textual reproduction
through manuscript copies dominated over print culture. sDe dge's late eighteenth century xylographic
edition was the only NGB ever printed, but as has often proven to be the case with Tibetan literature, the
impact of its printing in any case did very little to reduce the variation between editions.20 In common with
17

18

19
20

Of course, apart from a handful of key texts like the rGyud gsang ba'i snying po, most of these texts are not the object of
regular classroom study in monastic curricula. Yet they are read by the very learned, for edification and inspiration.
Understood as the backbone of the rNying ma tradition on which the all-pervading gter mas are textually dependent, they are
held in great esteem.
Cantwell 2006. Rather more radical variation has since been generated within English language sources presenting the bDud
'joms gter gsar sngon 'gro practices, since interpretations of the laconic instructions have greatly varied.
See note 3 above, and on the specific variant of the homophones, sgo/go in the first root verse, see Mayer 1996: 213-216.
For an example of this in dGe lugs pa literature, see van der Kuijp 2010: 448. The popular perception that a manuscript
tradition must necessarily generate more variation than a printed tradition is frequently overstated. The Masoretic Bible was
exclusively a manuscript tradition for many centuries before the age of print, yet famously succeeded in eliminating variation
through the application of ingeniously simple traditional methods of alphabetic and verbal calculations, so that the scribe knew
exactly how many instances of each word and letter should occur in each paragraph. Conversely, as modern textual critics like
Jerome McGann have shown, literary works that are far more recent than the Masoretic text and which only emerged in the
age of print, can vary considerably, for a wide variety of reasons. Hence we must conclude that the unyielding ideological
determination of generations of Jews to keep their Torah absolutely free of variation or error proved more significant than any
limitations of the manuscript medium; while the lack of such determination has allowed some modern printed literary works,
such as Byron's The Giaour, to vary greatly from the outset (McGann 1983: 31-2, 59ff, 105-106). The same has been true of
rNying ma block prints and other printings of religious liturgies in recent history: significant textual variations between
different printed versions of the same popular practice texts are rife in Bhutanese and Tibetan monasteries in India. The mere
fact of printing has failed to eradicate variation or make such rNying ma texts uniform. Thus we can see that in pursuit of
uniformity, ideological commitment clearly trumps technology. One of our major theoretical points is that a Masoretic
Jewish-type absolute commitment to accuracy at any cost has not been evidenced in the NGB tradition, nor generally in

Methodological Issues

21

much traditional Buddhist scholasticism, it was also a world characterised by prodigious feats of memory,
resulting in complex ongoing interactions between written and memorized versions of text, the dynamics of
which so far remain little analysed by international scholarship. More importantly, as with the rest of Tibetan
Buddhism and so typically of many pre-modern cultures, the literary ideal was usually not to author brilliant
entirely original ideas in the modern sense that would be decried as mere rang bzo, or individualistic
contrivance. Rather, the idea was faithfully to pass on existing understandings, often by preserving received
text verbatim. In short, a common ideal was not the 'author' as we moderns know it, but something more akin
to the medieval figure of the 'tradent' that has been so brilliantly analysed in modern Talmudic scholarship.
In such a world of de facto ongoing collective authorship, existing fragments or building blocks of holy
dharma, blessed through their usage by many previous generations of masters, are typically re-inflected, reanthologized and re-arranged by learned lamas to suit the needs of their contemporary audiences. This has
three implications: firstly, the notion of an individual creative author as we have it does not always apply
very well the notion of tradent as part of a collective enterprise descending through the generations often
fits better. Secondly, the notion of a fixed text was not strictly envisaged in all cases: as Jonathan Silk and
Paul Harrison have pointed out for India, some genres of Tibetan literature in a very real sense included an
aspect of collective works in endless progress.21 Thirdly, as we already pointed out above, a moderate
degree of variation in ritual and poetic literature could be considered inconsequential, so long as it leads to
the correct enlightenment.
Such pervasive features of Tibetan literary culture remain desperately under-studied, and we are currently
working on a new research project to make further investigations into them. From what little is understood so
far, it seems that the parameters of process and change might differ between genres. In gter ma, textual
variation often accumulates over time through the complex yet often visible interactions of identifiable
authors, governed by strict cultural norms. For example, our recent work on the 'Chi med srog thig and other
texts within the Dudjom corpus shows how named lamas edit, revise, remix and republish the revelations of
previous gter stons, or else adapt previous revelations into a newer revelation.22 Yet in the NGB, variation

21

22

Mahyna and Vajrayna Buddhism otherwise, as we know from the Jewish example, complete uniformity could technically
have been achieved, even before the advent of print. Nor were techniques for textual uniformity similar to the Masoretic ones
unknown in India and Tibet far from it the stability of the Paninian and Tibetan syllabaries encouraged the development of
methods of alphabetic calculation, so that we find comparable methods routinely applied with complete success in
mantroddhras (see note 24 below). These served to protect the exact spelling of mantras over many centuries against any
possible transmissionally generated variation. There was probably nothing other than the lack of overwhelming ideological
commitment to prevent suchlike or other even more compact and simple techniques of alphabetic and verbal calculation being
applied to the entirety of Buddhist texts, beyond the mantras. Perhaps the very words of the traditional Buddhist nidna
'Thus have I heard: at one time..' leaves open some possibility that someone else heard something slightly different regarding
another occasion when the Buddha taught.
The Tibetan tradition of revelation (gter ma) fits this description very well. To give one extremely simple example, the version
of bDud 'joms gling pa's (1835-1904) fulfillment liturgy (bskang ba) for Vajraklaya which his reincarnation bDud 'joms rin po
che (1904-1987) edited makes one very small yet potent textual change, which has the effect of bringing the group of deities in
the retinue into line with bDud 'joms rin po che's standardised version of bDud 'joms gling pa's revelation, in which earlier
apparent inconsistencies are ironed out. A line which in bDud 'joms gling pa's text referred to the four families of the sras
mchog Klas (sras mchog k la rigs bzhi'i thugs dam bskang:, bDud 'joms gling pa volume Ca: 126), has been amended to five
families (sras mchog k la rigs lnga'i thugs dam bskang:, Dudjom Rinpoche volume Tha: 160).
Dudjom Rinpoche's works in volume Pha of his Collected Works on a tantric longevity practice revealed by one of his root
gurus, Zil gnon nam mkha'i rdo rje (including his Commentary, bsnyen yig: 431-509, Notes on Ritual Procedures, khrigs zin:
193-208 and Ritual Practice Framework Manual sgrub khog: 233-296), reproduce historical/mythological and ritual practice
material from the Fifteenth Karmapa's empowerment texts. They also introduce a number of modifications (many related to
the sMin grol gling practice traditions which he followed) and his own elaborations (we discuss this at length in our
forthcoming book written with Geoffrey Samuel, The Seed of Immortal Life: Contexts and Meanings of a Tibetan Longevity
Practice). His texts on the gNam lcags spu gri cycle of Vajraklaya (volumes Tha and Da of his Collected Works) extensively
re-work and expand bDud 'joms gling pa's gter ma revelation, and incorporate some material from the earlier seventeenth
century lama of the tradition, bDud 'dul rdo rje, whose Vajraklaya revelation was clearly a significant source for bDud 'joms
gling pa's.

22

Introduction

more often takes the form of a mute inscrutable anonymous inheritance from the distant past.23 Our present
question is, how might we begin to approach rNying ma understandings of variation specifically in TZ and in
the wider NGB?
First, we must review what data we have which of course is still small, since most of the NGB remains
unread by modern scholarship. However, we do know that the sDe dge NGB, the only surviving NGB
redaction that was able to consult several geographically diverse editions, carefully reproduces good
alternative readings in its marginal notes (see note 11 above). Taking this further, the Phur pa bcu gnyis even
preserves within this single text two quite variant lists of readings for the entire extended set of Vajraklaya
mantrasone set in a mantra list, the other in a corrupt and inconsistent phonetical rendition encoded in a
mantra tablewhile the sDe dge edition includes a marginal note explicitly drawing attention to the editorial
decision to leave the mantra table unedited, even though it had been scrutinized.24 Thirdly, the two main
variant renderings of the Vajraklaya root verses, quite probably originally the result of accidental
transmissional variation, have over the centuries each generated prodigious quantities of prestigious
commentarial text so that to reduce this variation to a single 'correct' version in modern times has become
unthinkable (see notes 3 and 19 above). Above all, we find the constant and ubiquitous repetition, within
NGB scriptures, of passages of text across the spectrum of being exactly the same (barring transmissional
variants), of remaining close but showing recensional variation, or differing through more creative
rearrangement. This is abundant evidence that both verbatim reproduction and variations on an existing
theme were de riguer for the anonymous compilers of these texts (see notes 22 and 23 above). What does
this imply?
Our impression gained from thirty years of textual studies in consultation with various lamas (see note 4,
p.14 above) is that their response to apparently equally good but variant readings in their NGB scriptures
implicitly resembles the ideas of many modern anthropologists who see knowledge or culture as distributive.
In other words, rNying ma pa lamas, although they might welcome a less varied tradition, de facto operate on
the basis that no single extant form of words from the NGB is necessarily uniquely complete and valid. In
this view, a definitive version, vast in length and perfect in every detail, persists eternally in the tantric pure
realms, but unfortunately, no longer on earth. Instead, pragmatically, they accept that the remaining available
terrestrial sources are varied and distributed, each nevertheless potentially having a valid partial contribution
to make to the total picture.
This is very close to many modern anthropological formulations. In the words of Roger Keesing, a
distributive model of culture, "takes as fundamental the distribution of partial versions of a cultural tradition
among members of a society...[it] must take into account both diversity and commonality...`A Culture' is
therefore seen as a pool of knowledge to which individuals contribute in different ways and degrees"
(Keesing 1981: 71-72).
23

24

We need to distinguish here variation within single texts across different editions, and different versions of the same blocks of
material across different named scriptures. The focus of this discussion has been on variation within one text, but the representation of different versions of the same passages of text in different root tantras is even more striking. See Cantwell and
Mayer 2008 Chapters 5-6 (especially p.76-87 and Appendix to Chapter 6) for an example of a text preserved at Dunhuang
(IOL Tib J 331.III), the contents of which are largely reproduced in varying sequences in different NGB texts. Since
publication, we have further identified a complete and virtually identical version of the Dunhuang text within Nyang ral nyi
ma 'od zer's (1124-1192) bKa' brgyad bde gshegs 'dus pa collection, under a different title. A very different form of variation
is evidenced in the apparently jumbled parallel text within the two NGB texts of the Myang 'das's Chapter 18 and the rDo rje
khros pa phur pa rtsa ba'i rgyud's Chapter 6 (Cantwell and Mayer 2007 Chapter 2.IV).
Such coded mantra tables (sngags btu ba, = Skt. mantroddhra) make possible the preservation of mantras, by using numbers
to correspond to each Sanskrit letter. The decoded mantras in the Phur pa bcu gnyis's mantra table appear to reflect a corrupt
and inconsistent early phonetic rendering in Tibetan, while the mantra list earlier in the text's sDe dge edition instead gives a
more correct representation of Tibetan transliteration of Sanskrit mantras. See the extensive discussion in Mayer 1996: 135146.

Methodological Issues

23

A more recent presentation of the distributive nature of knowledge comes from Fredrik Barth, based on
examples taken from New Guinea, Bali and England, but it seems written while he was staying in Bhutan.
Barth analytically distinguishes three faces or aspects of knowledge. First, any tradition of knowledge must
contain a corpus of substantive assertions and ideas about aspects of the world. Secondly, it must be
instantiated and communicated in one or several media as a series of partial representations in the form of
words, symbols, gestures, or actions. Thirdly, it will be distributed, communicated, employed, and
transmitted within a series of instituted social relations.25
Barth's exposition of a distributive model is underlined in his insistence that the three faces of knowledge
appear only in their particular applications, and not as a generalized abstract entity. We believe that such a
distributive model corresponds quite closely to one important aspect of how rNying ma pa lamas work in
practice with the transmission of their NGB. By understanding their distributive mode of operation, we also
understand how they tolerate such wide discrepancy of good readings in their canonical sources without
excessive dismay or alarm (bad readings are of course rejected by all as scribal error).
To adapt Barth's model to our purposes, we can say that knowledge of the NGB in Tibet had three
aspects:
The first aspect is the actual substantive teachings and doctrines of the NGB: By acknowledging the
importance of these, one avoids the absurdity of an extreme relativism, which would say that the rNying ma
tantric tradition is really just whatever anyone claims it to be. In the parlance of the lamas, this aspect is
symbolically represented by the widespread idea of the complete and perfect versions of the tantras eternally
preserved in transcendent locations such as pure lands, which uniquely represent the complete and full
authorial intention of the Buddhas.
The second aspect is each of the numerous and varying manuscript and xylograph witnesses within
which the NGB has been represented in history, such as the South Central Tibetan text from gTing skyes,
or the sDe dge xylograph: by acknowledging their occasional diversity of good readings, and accepting their
particular differences, one avoids the extreme of positing a monolithic textual uniformity which, as the lamas
are well aware, does not exist.
The third aspect is the instituted social relations through which the NGB is transmitted, taught and
reproduced. By acknowledging this, one accepts that the tradition is inseparable from its human
performance, and thus we counteract the latent tendency of textual scholarship to abstract its materials away
from actual historical realities. In the parlance of the lamas, this corresponds to their notions of transmission
through exalted lineages of enlightened gurus to worthy disciples, sponsored by virtuous donors. [For many
students, direct involvement in the texts might be limited to the receipt of their ritual transmissions, or
occasional public readings for the purpose of making merit and conferring blessings. Their direct study was
an elite activity, while general familiarisation with NGB materials would be effected through the
incorporation of some sections into liturgical practice, and through the medium of commentarial
presentation.]
25

Barth (2002) sees these three faces of knowledge as interconnected, and above all, mutually determining. However, Barth says
that to understand his ideas, we need to invert the way we habitually construct analyses. He emphasises that he is not positing
a highly generalized and abstract unity called knowledge, which then subdivides into the three parts of substantive corpus,
communicative medium, and social organization, which in turn progressively break down further until we finally arrive at the
level of particular human actions and events. On the contrary, Barth proposes that these three faces of knowledge appear
together precisely in the particulars of action in every actual event of the application of knowledge, in every actual transaction
in knowledge, in every actual performance of knowledge. Their mutual determination, says Barth, takes place at those specific
moments when a particular item of substantive knowledge is cast in a particular communicative medium and applied in an
action by an actor positioned in a particular social organization. Their systematic interdependence, he says, arises by virtue of
the constraints in realization that these three aspects impose on each other in the context of every particular application.

24

Introduction

As Barth suggests, these three aspects mutually determine each other, but should not be envisaged as an
abstract unity with three parts. Rather, they appear in the particulars of action in every event of the
application, performance and transmission of knowledge of the NGB. They mutually interact and determine
each other as particulars, in numerous moments distributed through time and space, and it is these particular
events that constitute the actual transmissions of the NGB.
Hence the editorial decisions made in any specific redaction of TZ has depended on the mutually
determining factors of what doctrines TZ is teaching, the readings of the available exemplars or ma dpe, and
on how the presiding editors on that occasion, ideally supported by their extraordinary cognitions, attempted
to ensure the new text's accordance with the eternal doctrines and the understandings of their lineage. In
short, there is no single monolithic or abstract entity that we can call The Transmission of TZ, or of the NGB.
Rather, their transmission is distributed, instantiated in numerous separate events through time and space that
have their own dynamic and individuality. In this way, the range of good readings of the different editions of
the NGB do not as far as we can see cause agony to the lamas simply because they might vary somewhat. A
comfortable degree of latitude, within which variation can be tolerated while still upholding the overall
purity of the scriptural tradition, is gained in the endless mutually determining interplay of the three factors
we have outlined above.
The rNying ma pa's de facto distributive mode of operation over the last many centuries implies that even
if the recovery of a strictly historical single earliest text is the holy grail for philology, it perhaps has
somewhat less absolute significance for them, unless they are to now embrace modern text critical criteria.
Combining pragmatic acceptance of the status quo, a lack of concern about moderate textual variation so
long as the outcome is similar, and a particular set of religious ideas, they tacitly acknowledge that over time,
ongoing repairs to scribal errors can cause texts to present varying good readings, but their response is
almost as much transcendental as historical. To put that another way, it considerably depends on their
important lama-editors, backed by the religious authority of their various lineages. If realised and learned
beings, these great masters must have some degree of direct access to the perfect original meanings of the
tantras and, if gter stons, can even act like Padmasambhava as fountainheads or mediums for new
buddhavacana, over and above mere learned editing. But religious authority in the rNying ma pa has never
been centralised, and on different occasions, the textual decisions of different realised and omniscient lamas
have taken mysteriously different turns, even when faced with exactly the same textual crux or passage. This
means that in different places at different times, a variety of divergent but semantically equally profound
meanings could potentially be generated around a single textual crux or passage; and because of the great
expense of transporting huge NGB collections across large distances, over time such variant readings often
became regional, and only rarely cross referenced with one another. The upshot is that we international
scholars cannot and should not ignore the many transmissionally and recensionally transformed semantically
good readings within the NGB that have appeared through Tibetan history and across its regions. We know,
for example, that the great 'Jam mgon Kong sprul kept a copy of the NGB in his residence at dPal spungs
(Barron 2003: 282, 286), and it also looks quite likely that it was the sDe dge xylograph edition he consulted,
given the shared readings with citations in his compositions.26 We also know that this sDe dge edition seems
to reflect a distinctive East Tibetan tradition, but, unusually for NGB editions, also has some exposure to
editions from distant regions, and is in addition full of recensional changes mainly attributed to famous late
eighteenth century lamas.27 From the point of view of stemmatics, this heavily redacted and conflated edition
is far removed from any original archetype. From the point of view of living Tibetan Buddhism, and hence
of functional Tibetological scholarship, its widely influential readings are crucial. At the time of writing, in
26

27

At least in the specific case of his seminal commentary on rDo rje phur pa, a number of his citations from the Myang 'das
would suggest this (Cantwell and Mayer 2007: 50-52). Kong sprul also refers (Barron 2003: 255) to his receipt of the
transmission of the twenty-five volumes of the NGB, which again would seem to indicate sDe dge's edition.
At the authoritative hands of the great Rig 'dzin Tshe dbang mchog sgrub, who was not a gter ston, but a vastly influential
author of commentaries on gter ma and much else.

Methodological Issues

25

the early twenty-first century, it seems that the rNying ma tradition still follows its time-honoured custom in
privileging the readings sanctified by the living lineage over any merely philological arguments.
This partially transcendental and distributive mode of working, shaped by a shifting and decentralised
religious authority, which we see as an important element in the transmission of the NGB, is thus
considerably at variance to the strictly historical presuppositions of stemmatic analysis. Stemmatics
developed in the West with reference to such cultural models as the monolithically uniform Masoretic bible,
and the notionally fixed and timeless utterances of individual Classical authors. With growing scholarly
awareness of the fluidity of so many Buddhist texts through history, the relevance of stemmatics is
sometimes called into question by contemporary Buddhist scholars, including some important voices within
the major centres of philology, who understand full well that its historicist premises do not fit Buddhist
literature. We feel such a rejection can be overdone. We do recognise the grave limitations inherent in trying
to force onto Buddhist literature a Lachmannian style of stemmatic method, with all its underlying
presuppositions of textual uniformity. Nevertheless we believe we can use stemmatics as a probing device,
an analytic tool, to isolate not merely archetypes such as that of TZ, but also particular versions of an
otherwise fluid scriptural text at interesting junctures in its development and history. For example, we
believe we might have enough evidence to reconstruct for many texts the Tibetan lHa lung hypearchetype
from which all the current Bhutanese manuscripts of the NGB are descended, and we find this a potentially
useful tool. So our approach preserves and combines two perspectives on textual editing otherwise seen as
contradictory. On the one hand, there is the pioneer's optimism of Helmut Eimer, whose initial vision was
that the classic stemmatic methods of Paul Maas could yield dividends in Kanjur analysis. On the other hand,
there is the more pessimistic approach of some recent scholars, who suggest that in a universe of irreducible
textual fluidity, the best we can hope for is the elimination of orthographic errors and other egregious
transmissional accumulations. Our methodology, by contrast, accepts and even celebrates the ongoing
permutations of these texts, but still finds value in stemmatic techniques as a way of recovering both their
original archetypes and also significant moments in their history. In conclusion, we should add how
delightful it would be if further NGB texts turn out to be genuinely amenable to stemmatic analysis, so that
we can reconstruct more early readings from them, to aid our understanding of the doctrines and practices of
the texts, and also historical investigations into the seemingly impossibly obscure period of rNying ma
origins. At the same time, it would be unfortunate if we were to lose ourselves in historically oriented
researches and fail to recognise and explore the de facto distributive realities and the emphasis on the
authority of lineage which constitutes the actual historical existence of the NGB in Tibet and beyond.

TEXTUAL ANALYSIS
a) The Editions of the Root Text and Commentary
The Root Tantra, the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa
As mentioned above, the extant versions of TZ appear to descend in the following five branches:
[a] Four branches of TZ apparently descending independently of one another from a single TZ archetype.
[b] One further more complex branch of TZ ultimately deriving from an old version of TZComm, which
of course contained a complete version of TZ as lemmata. This commentarially-derived branch has four
extant lines: two that retained their original commentarial natures as TZComm, but still witnessing within
their folios complete versions of TZ in the form of lemmata; and two lines descended from versions of TZ
that had at different times been redacted out of TZComm to stand on their own, through extraction and
reproduction of the lemmata as independent works.
The Eight Versions of the Root Text
i. The Four Versions apparently descended independently from the same TZ Archetype
The South Central
Tibetan NGB

The local (incomplete)

Kanjur Collection from
Bathang (held in the
Newark Museum, New
York, and thus sometimes
called, the Newark
Manuscript Kanjur) (Bth)

The local (incomplete)

Kanjur Collection
from Hemis Tshoms
lha khang (He)

The local Kanjur

Collection from
Tawang, originally
from the
O rgyan gling Temple
(Ogl)

gTing skyes (T)

Rig 'dzin Tshe dbang nor
bu (R)
Kathmandu (K)
ii. The two versions of TZ apparently descended from TZComm through the extraction of lemmata
The Kanjur (bka' 'gyur) texts (Tshal pa line)

The Bhutanese NGB

'Jang sa tham or Lithang Kanjur (J)

Peking Kanjur (Qk)
sNar thang Kanjur (Nk)
Urga Kanjur (U)
sDe dge Kanjur (and sDe dge NGB) (Dk, D)
lHa sa Kanjur (Hk)
Ulan Bator Kanjur (V)

mTshams brag (M)

sGang steng-a (G-a)
sGang steng-b (G)
dGra med rtse (Gr)

iii. The two versions of TZComm that witness TZ as lemmata

Lemmata of the root text within the Dunhuang
manuscript of the commentary (Ms)

Lemmata of the root text within the Tenjur (bstan

'gyur) version of the commentary
Peking Tenjur (Qt)
Golden Tenjur (Gt)
sNar thang Tenjur (Nt)

Editions of the Root text and Commentary

27

In studying the text of TZ, we have thus had twenty-one witnesses to work with. This potentially
presented a problem in constructing the edition, because representing so many witnesses could result in an
excessively complex apparatus difficult to navigate, and a tome of excessive length. After careful
examination however, it became apparent that several of the twenty-one witnesses were either repetitive or
descriptive, contributing little or nothing to the stemmatic analysis we employed. Hence a considerable
quantity of the initial collation was eventually removed from the earlier drafts to produce the more
manageable critical apparatus presented here. Nevertheless whatever insights we gained from those
eventually unrepresented collations is of course retained in our analysis, as is made clear at various junctures
in our edition.
[1] A version of TZ is witnessed by the South Central Tibetan NGB manuscript editions of gTing skyes,
Rig 'dzin and Kathmandu. We have not been able to check the TZ witness from the Nubri manuscript of this
grouping, since the relevant folios are missing.1 We have discussed this family of NGB manuscripts
elsewhere, so there is no need for detailed bibliographic or codicological descriptions here.2 In brief, we
should mention as key points that these manuscripts represent a distinctive branch of the NGB transmission
evident in the Nepalese borderlands with South Central Tibet, sometimes connected with Byang gter
establishments and probably dating from the late eighteenth and early nineteenth centuries. In the collation of
some texts, and also doxographically speaking, this family can sometimes be seen to break down into the two
sub-branches of gTing skyes and Rig 'dzin on the one hand, and Nubri and Kathmandu on the other hand.
Since the Nubri witness is missing, we cannot ascertain if that is also the case with TZ. An excellent and
fully detailed catalogue of gTing skyes was made in Roman Wylie transcription by the Japanese scholar
Kaneko in 1982, and an equally detailed catalogue of the Rig 'dzin edition was subsequently made by us and
published in association with the British Library.3 Kaneko's catalogue has subsequently been very usefully
reproduced in an electronic format by the Tibetan and Himalayan Library under the sigla Tk. In the case of
TZ, the South Central Tibetan NGB redaction seems unproblematically to stem directly from the
transmission of an archetypal TZ text, unmediated by any periods embedded within TZComm.
Rather than overburden our edition by fully representing our collations of all three available South
Central NGB editions, we have elected to show only gTing skyes and Kathmandu throughout, merely adding
a short sample of Rig 'dzin for illustrative purposes, or where it helped to clarify or confirm the South
Central readings.
[2] There are three further witnesses of TZ found in the local Kanjurs of Tawang (in modern Arunachal
Pradesh), Hemis (Ladakh) and Bathang (in modern Sichuan), although the Hemis manuscript is not
complete.4 Despite their geographically diverse origins, each of them represents the same basic redaction as
the South Central NGB version, but of course without the South Central NGB group's shared indicative
errors. Although not geographically close to one another, these three local Kanjurs share the characteristic of
being comparatively old manuscripts that in some cases carry much older readings still. The Hemis Tshoms
lha khang Kanjur is currently being studied by Prof Helmut Tauscher and Dr Bruno Laine of the University
of Vienna. Their preliminary research suggests it dates from the early seventeenth century, perhaps shortly
after the foundation of Hemis in the 1630s, since two colophons mention Nam mkha' dpal mgon, a famous
figure of that time, connected with King Seng ge rnam rgyal and sTag tshang ras pa.5 The Bathang Kanjur is
currently being catalogued by Helmut Eimer. Without venturing exact dates, Skilling (2001:74-5) and
Zimmerman (2002a: 24-26; 166-167; 2002b;1998: 35) have already concluded that Bathang is a manuscript
1

2
3
4

Missing from Nubri's Volume Ma. We are not yet sure if the Thini NGB belonged to the South Central grouping, which might
be possible given its geographical provenance, but in any case we found no Thabs zhags text in its surviving volumes.
For bibliographic and codicological information, see Cantwell, C. and R. Mayer 2007: 70-78.
For gTing skyes, see Kaneko, Eiichi. 1982. For Rig 'dzin, see Cantwell, C., R. Mayer and M. Fischer 2002-3.
Thanks to Helmut Eimer, who drew our attention to the Bathang version of the text, kindly making available to us the rnying
rgyud section of his as yet unpublished draft catalogue. Thanks are also due to Helmut Tauscher and Bruno Laine, who have
helped us to consult both the Bathang and Hemis versions. We consulted the Tawang Kanjur copy at the Central University of
Tibetan Studies in Sarnath, and found that it too contained the text.
Helmut Tauscher, Personal communication, 14 December 2010.

28

Textual Analysis

of considerable antiquity which, in the particular texts they consulted, descended from unusually old
redactions, while Dietz (2002: 17) has likewise noted it frequently has independent readings.6 The Tawang
Kanjur has been described by Jampa Samten (1994), who reports it was made in 1699 by a master
calligrapher at the behest of the Desi Sangye Gyatso as an offering to the ancestral temple of the Sixth Dalai
Lama. Out of these three local Kanjur witnesses of TZ, Bathang has its own unique errors and cannot
possibly be a direct ancestor of any of our other versions. Both Tawang and Hemis in contrast would seem
to have few unique errors, most of which are trivial or obvious mistakes that any competent scribe might
correct as he went along, making them unreliable as a basis for stemmatic reasoning. The TZ recensions of
these three local Kanjurs, like that the South Central Tibetan NGB, all descend independently from an
archetypal stand-alone root text, and never passed through a stage as lemmata in a commentary, as we find in
[3] and [4] below. Given the value of their testimony to our stemmatic reconstruction, we represent our
collations of all three of them in full.
[3] Thirdly, several Kanjurs from the Tshal pa line carry TZ in their rNying rgyud sections. As one of the
two main branches of the Kanjur transmission, the production of the famous Tshal pa line and the work of its
redactor Tshal pa kun dga' rdo rje (1309-1364) has been studied by a number of scholars and we have no
need to reproduce their findings here.
There are too many editions of the Tshal pa branch and their variants too slight to justify representing
collations of all of them in the apparatus. Thus, in our edition we have primarily presented the Peking Kanjur
(representing the Peking sub-line of the Tshal pa branch) and the 'Jang sa tham or Lithang Kanjur
(representing the 'Phying ba stag rtse sub-line of the Tshal pa branch), although we have also retained in our
presentation some samples of our collations of the other editions for purely illustrative purposes, or to bring
into play additional evidence where necessary.
The rGyal rtse Them spangs ma manuscript, the progenitor of the other main line of Kanjur descent, did
not have a rNying rgyud section, so TZ is not found in Kanjurs of Them spangs ma descent.7 Hence the four
mixed Kanjurs of sNar thang, sDe dge, lHa sa, and Urga could be predicted also in this case to have relied on
exemplars from the Tshal pa Kanjur branch, and our collations, demonstrating a close relation between all
these Kanjur versions, would support that prediction.
We have not had access to the complete Ulan Bator manuscript Kanjur version of TZ, only its opening
pages. The Ulan Bator Kanjur is said ostensibly to be a Them spangs ma Kanjur, yet it had a rNying rgyud
section added to it, which technically could be taken as grounds for describing it as a mixed Kanjur rather
than purely a Them spangs ma Kanjur. These rNying rgyud texts seem to be identical to those contained in
the Tshal pa editions, so are most likely simply a copy of them. Our examination of the first pages of Ulan

Peter Skilling (2001) reviews features of the Bathang Kanjur, and for our purposes here, makes two important points. First, he
points out that the collection must be a copy of venerable exemplar(s): "Its antiquity may be seen from the orthography
(particularly the transliteration of Sanskrit), the arrangement of contents, and the inclusion of texts excluded from or missing
in the comprehensive Tshal pa edition, which was compiled in CE 1347-51". Secondly, Skilling's specific study of the
Mahstras, "suggests that the Newark Kanjur belongs to an old and independent textual transmission that predates the
compilation of the Tshal pa and Them spangs ma collections" (2001: 74-75). Michael Zimmermann's detailed work on the
different editions of the Tathgathagarbha Stra confirm this picture. Zimmermann makes clear that, "Bth is the only known
representative of a separate, paracanonical translation of the Tathgathagarbha Stra. Judging from its terminology and
syntax, it must have been executed before translations became more standardized following the compilation of compendiums
like the MVy [Mahvyutpatti] and the sGra sbyor bam po gnyis pa, that is, before the early ninth century" (2002a: 166-167).
He argues (1998: 35) that this translation seems to have been based on an identical or very similar Indic original to that used
by the later translation, but that it seems very unlikely that the later translators/editors had access to this earlier work.
Moreover, not only does the Bathang version of the text fail to use the standard translation vocabulary, it also uses
terminology which has parallels in other early translations, while its translation seems to attempt to keep so close to the
Sanskrit source text that it is stilted and partly unintelligible in places (Zimmermann 1998: 46-49). See also Zimmermann
2002a: 24-26, and 2002b. Siglinde Dietz (2002: 17) also attests to the "frequent independent readings" found in the Bathang
Kanjur version of the 'Jig rten gzhag pa.
Paul Harrison commented, "I would say there is virtually no chance that the text was included in the rGyal rtse Them spangs
ma MS Kanjur." Personal communication, August 6th, 2007; see also Helmut Eimer 2002: 66.

Editions of the Root text and Commentary

29

Bator's version of TZ, where its variants are identical to those of the other Tshal pa Kanjurs, would seem to
support this hypothesis.8
The version of TZ in the sDe dge xylograph NGB must also be considered to represent the Tshal pa
Kanjur branch, since it seems to have been prepared using a par yig made from the same blocks as the
slightly earlier sDe dge xylograph Kanjur.9 As a result, it is virtually identical to the sDe dge Kanjur in every
way, with little more than the pagination varying.10 Given the overwhelming significance of the sDe dge
xylograph edition of the NGB in recent rNying ma cultural history, we have elected to present in the edition
our collation of this witness as well.
As we explain below, internal evidence indicates that the Tshal pa redaction of TZ stems from an attempt
to reconstruct and extract the root text TZ from the lemmata contained within its commentarial text TZComm.
The attempt was not entirely successful, because it introduced into this version of TZ extensive passages of
text properly belonging only to TZComm.
[4] A version of TZ is witnessed by all four available Bhutanese NGB manuscripts, namely mTshams
brag, sGang steng-a, sGang steng-b, and dGra med rtse. Regretfully we have not yet found time to make
much further investigations into the external history and descriptive bibliography of the Bhutanese NGB
tradition, beyond what we have already reported previously (Cantwell, C. and R. Mayer 2007: 68-70).
However, we have here for the first time been able to collate all four Bhutanese witnesses to a major text,
and an intriguing situation has revealed itself. Based on the internal evidence of TZ, it seems that the
Bhutanese NGB tradition might have had two sub-branches: one represented by sGang steng-a and dGra med
rtse, the other represented by mTshams brag and sGang steng-b. In many dozens of occasions in TZ, sGang
steng-a and dGra med rtse agree in trivial error against mTshams brag and sGang steng-b, although there are
no major differences between the two groupings. However, sGang steng-a also has numerous corrections to
the text, which emend it to agree with mTshams brag and sGang steng-b.
The unverified oral tradition at sGang steng monastery maintains that their sGang steng-a NGB is older
than their sGang steng-b NGB, and, moreover, that sGang steng-a was the original NGB in all of Bhutan,
taken from a lHo brag lHa lung original. If sGang steng-a is indeed the original Bhutanese NGB edition as
suggested, then it might follow that dGra med rtse was copied from it early on, before the corrections were
made to the manuscript. Later, its manuscript was lightly corrected, and the corrected version subsequently
recopied to make sGang steng-b (which is basically identical to sGang steng-a but with a number of minor
spelling errors and suchlike made good). mTshams brag was then copied either from the corrected
manuscript of sGang steng-a, or, perhaps more likely, from sGang steng-b. It will take further investigation
of both external and internal evidence to test these initial speculations, and we look forward to further
Bhutanese NGB editions becoming available, such as the Pagar manuscript.
As with the Tshal pa Kanjur, the Bhutanese version of TZ seems also to have stemmed from an attempt to
extract the root text from the commentarial text, but in this case, the editors made quite different decisions,
successfully avoiding many of the additional commentarial passages included in the Tshal pa Kanjur. They
did however include some other commentarial text which the Tshal pa Kanjur did not, and also left out some
root text.

9
10

What little we have been able to collate of it seems to show that unlike the Urga, Ulan Bator may belong to the sNar thang /
lHa sa sub-group (see below, p. 45).
Samten G Karmay 2005: 78, describes the process of making the par yig.
We have found one spelling correction: in Chapter 11, at the end of the first line of the sDe dge NGB folio 291v, equivalent to
the sDe dge Kanjur's 608, 'ching zhing is given, for the incorrect 'cing zhing in the sDe dge Kanjur. It is possible that there are
further minor amendments of this type, since we did not exhaustively collate both versions, but there is no doubt that such
amendments are rare, and that the two versions are virtually identical. Such minor changes might have been made with later
printings from the original blocks; Dr. Phillip Stanley notes that there are at least eight printings of the sDe dge Kanjur with
various changes. Stanley also adds in relation to the Karmapa reprint (which we have used): "The Karmapa reprint included
touch-ups to the original which, however, introduced additional errors."

30

Textual Analysis

Rather than overburden the apparatus by representing all four available Bhutanese texts, we have only
presented our collations of sGang steng-b and dGra med rtse, which can adequately represent the two
possible sub-traditions.11
[5] Fifthly, a complete version of TZ can be found within the lemmata of the Dunhuang manuscript of
TZComm. A priceless witness completely untouched, perhaps since the tenth century, we have naturally
represented its entire collation within our edition. We describe this manuscript at length below.
[6] Finally, the entirety of TZ was originally included in the lemmata of the three Tenjur witnesses of
TZComm (see below). These three witnesses belong to the same sub-grouping of Tenjurs and have so few
significant differences that the reader is best served by presenting the collation of only one of them, with the
others represented only in limited specimens to illustrate the close relationship, or to confirm the Tenjur
reading where significant variants occur. We have elected to use the Golden or Gandan (dga' ldan) Tenjur.
One might have expected the two versions of TZ occurring as lemmata within the Dunhuang and Tenjur
versions of TZComm to share their own line of descent, and to have some independence from the other TZ
editions. This does not seem entirely to have been the case, or at least, not in these two extant versions,
which admittedly, have been separated by a considerable measure of time and place. In fact, they frequently
line up against each other, with just one of them supporting readings found in one or another of the root text
editions. Nevertheless, there are some shared readings, and numerous minor shared variants, not least
punctuation and so forth connected to the embedding of the lemmata within the commentary. It is now clear
that one reason for the only limited extent to which the lemmata from the two commentaries agree in
opposition to the other texts, is that the Tshal pa Kanjur and Bhutanese branches of TZ themselves descend
from a commentarial line. Thus, the Tenjur grouping have in fact avoided some of the errors shared by the
Dunhuang, Tshal pa Kanjur and Bhutanese lines.
The Commentary, the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa'i 'grel pa
TZComm survives in only two sources: the Dunhuang manuscript, and a truncated version in the three
Tenjur (bstan 'gyur) editions of Peking, Narthang (snar thang) and Ganden (dga' ldan).12 The Tenjur
versions derive from a single ancestor, as evidenced in numerous shared indicative errors, including the
omission of all text from the middle of Chapter Six until the last lines of Chapter Ten; and from the middle
of Chapter Thirteen until the end of Chapter Seventeen: altogether over thirty percent of the total text.
Without the Dunhuang text, these missing parts would not be recoverable.
The Dunhuang manuscript for its part has its own omissions. Two short chapters, Thirty-one and Thirtytwo, are missing entirely. There is the loss of the passage of root text lemmata at the end of Chapter Ten
mentioned above (p.10), preserved in full only in two of the local Kanjurs and the South Central NGB.
There are a few other omissions, including an omitted passage of text in Chapter Thirteen, as well as a
misplacement of a passage from Chapter Thirteen into Chapter Twelve.
Fortunately, since the main corruptions in the two versions of TZComm generally only occur on one side
of the transmission at a time, it is usually possible to restore the text from the other, although there are a few
parts where both versions appear to be somewhat corrupt. For example, in Chapter Six, just prior to the
Tenjur's major omission, the Dunhuang version does not seem altogether coherent. The Tenjur reads rather
differently, and perhaps a little more coherently, although since the next section of text is lost entirely in the
11

12

sGang steng-b is a marginally better choice than mTshams brag because the digital images available to us have not been
through the recent and potentially distorting process of photo-offset litho reproduction. On the other hand, if the oral tradition
at sGang steng monastery is correct, then sGang steng-a might transpire to have made a marginally better choice than dGra
med rtse. However it carries the so far inadequately examined complication of the numerous corrections to its manuscript, and
in addition we were only able to collate it comparatively late into our project.
It does not, however, survive in the Derge (sde dge) or Con (co ne) Tenjurs. According to Phillip Stanley, there are only two
clear Tenjur groupings, one of the Peking, Golden and sNar thang, and the other of the Derge and Con (D. Phillip Stanley
2005: "Editions of the Tengyur").

Editions of the Root text and Commentary

31

Tenjur, it is slightly hard to judge. Similarly, where the Dunhuang manuscript omits a passage in Chapter
Thirteen, the Tenjur version contains some necessary text, but is not entirely coherent throughout.
The Two Lines of Transmission of TZComm
The Dunhuang manuscript of TZComm (Ms)

The Tenjur (bstan 'gyur) version of TZComm (not

found in the sDe dge and Cone Tenjurs)
Peking Tenjur: xylograph impression prepared in
1724 under the Kangxi emperor, with 224 +1
volumes; a copy is held by the library of Otani
University, Kyoto, from which a modern
photographic reprint has been made. (Qt)
Golden Tenjur: 226 volume manuscript from Ganden
(dga' ldan) in Central Tibet, 1731-1741. Reprinted by
the Chinese Nationality Library in Beijing, 1988. (Gt)
sNar thang Tenjur: 223 volume manuscript prepared
at sNar thang in gTsang, in 1741-1742, plus a
supplementary volume in 1763. (Nt)

Although we have only these two principal lines of transmission, a number of substantial commentarial
passages are also accidentally witnessed within the two branches of TZ which descend from an edition of
TZComm, that is, the Tshal pa Kanjur and the Bhutanese NGB editions. Since both of these TZ editions
inadvertently incorporated some commentarial passages along with the desired lemmata, we have their
further testimony for some parts of TZComm, and whenever they offer readings of value or interest, we
present them in the apparatus of our edition of TZComm. Because they became incorporated into the
currently most popular and widely used editions of TZ, we also represent these extraneous passages in full in
our edition of TZ, albeit clearly demarcated through indentation and explicit notices as more properly
belonging to TZComm.

b) Features of the Dunhuang Manuscript

The Dunhuang manuscript, the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa'i 'grel pa',
IOL Tib J 321
Codicologically speaking, the Dunhuang manuscript is made to high professional specifications, in looseleaf poth format. Judging from its nicely calligraphed dbu med, which is quite close to dbu can, IOL Tib J
321 does not appear to be particularly archaic. Samten Karmay compares the similar style found in the IOL
Tib J 594 and 647 with the later 'bru ma script (Karmay 1988: 43, 59). Palaeographical study of Tibetan is
still at an early stage. In developing a preliminary typology, Sam van Schaik (in press: 2, 16, 18-20) has
discussed IOL Tib J 321's handwriting style as an example of the headless post-Imperial Buddhist
manuscript style, probably dating from the last decades of the tenth century. As van Schaik points out, it is
an elegant calligraphic style which seems not to have been confined to Dunhuang (recent manuscript
discoveries in the Gathang stpa are similar), and which persisted and was developed further in later
centuries. Three paper samples from IOL Tib J 321 have been included in a recent study analysing the paper
of a selection of Dunhuang manuscripts (Helman-Wazny and van Schaik in press: 15, 19-20, 29). The
samples were shown to be composed of rag paper, made from Boehmeria nivea (ramie) and Broussonetia sp.
(paper mulberry), with the possible addition of Cannabis sp. (hemp), indicating that the paper was most
probably produced locally in Dunhuang, or possibly elsewhere in China, but not Tibet. The manuscript is in
such good condition that many of the folios look as though they were written only a few years ago. The
folios are quite small approximately thirty-one centimetres across, by just under ten centimetres in height.1
The sheets of good quality thick paper are fairly even in size and neatly cut, although not cut exactly square,
and they each have two small string holes, around each of which a circle is marked, which the writing avoids
(we have indicated the gaps between the words by repeated asterisks in our edition). There is only slight
fraying and discolouration of the paper, and few blemishes. The ink writing of the main text remains clear,
and this is generally true also of the smaller writing of the annotations, which is slightly lighter and appears
to have been written with a much finer pen. The handwriting style is similar, taking account of adjustments
needed for smaller and slightly more cramped lettering. Although no longer evident, some form of
guidelines might have been used, since the six lines of writing on each page tend to be positioned straight
and evenly spaced, apart from a slight tendency for the writing to drop downwards on the final lines of the
page. Each recto side begins with an opening yig mgo, and the left-hand margins are all marked, ka, with the
pagination beginning at gcig. The letter ka might possibly indicate that the text is the first in a collection,
although the letters of the alphabet are also used in some Dunhuang documents to indicate pagination, the
first hundred pages labelled, ka, the second hundred, kha, and so on (see IOL Tib J 619,
http://idp.bl.uk/database/oo_loader.a4d?pm=IOL Tib J 619, where ka brgya' tham is followed by kha gcig).
There are eighty-five folios, but only eighty-four numbered pages. The verso side of folio 33 appears to
have been intended as a recto side, with an opening yig mgo, but its text content is a repeat of earlier text
(starting at exactly the same point as folio 31r), except that it contains an omission. It is most likely that it
was intended as folio 31r, but when the omission was noticed, the sheet was re-copied. Perhaps in order not
to waste paper, the scribe then made use of the reverse side for 33r. Following 33r, the text continues on the
verso side of an unnumbered folio, which we have described as folio 33a2 and this is then followed by 34r.
The recto side of the unnumbered folio 33a begins with the same three mantras as those which start the verso
side, but the first one has errors, and it may be that the recto side was discarded when the errors were
noticed, the scribe beginning again on the verso side.

The average folio size is 310 x 98mm, ranging from 307 x 97 (but this small size maybe due to fraying since this was the front
sheet), to 311 x 99mm. Note that there was an error in Cantwell and Mayer (2008: 181), where we incorrectly give the height
as 8 cm.
The small modern numerals, 33a, have also been added to the right of the recto side of this folio, presumably by the India
Office Library.

Features of the Dunhuang Manuscript

33

The manuscript is certainly no autograph: there is evidence that it is at the very least a copy of a copy.
Firstly, it omits two short chapters that must have existed in an earlier version, because the text sequence
requires them in this place, further confirmed by their preservation in the canonically transmitted versions of
the text. Its chapters are, however, numbered only in accordance with the chapters that it does include, so
that after the point of omission, the chapter numbering runs two behind the canonically transmitted versions.
This suggests that there was one previous copy that omitted the chapters (or from which the folios of the
chapters were lost), and a further copying during which the jump from Chapter Thirty's title to Chapter
Thirty-three's title was noticed, and the chapters re-numbered accordingly. As we have noted above, the
manuscript in its lemmata also shares with the Tshal pa Kanjur and Bhutanese versions of TZ, a major
omission and some other errors, which are avoided in the lemmata of the Tenjur version of TZComm.
Secondly, some of the marginal notes in tiny handwriting also show signs of having been copied along with
the main text from an earlier original, as we can tell from the type of mistakes that occur in their numbered
lists of topics. These offer reasonably strong evidence that the annotations have been copied from a previous
version of the manuscript, and not simply composed afresh by the scribe.3 The comments seem to develop a
numbered list of meanings or points (don). However, they begin with the third and fourth points in Chapter
One, and continue with the sixth point in Chapter Four, the seventh point in Chapter Five, and the eighth
point in Chapter Seven. The omission of the first, second and fifth points is most likely explained by the
hypothesis that the copyist omitted them: it is a harder task to copy numerous small comments in their
entirety than to copy a continuous main text, and it is easy to see how some might pass unnoticed. It is less
likely that the individual composing the comments would have omitted the missing points. Furthermore,
there are instances of repeated annotations (f. 1v and 4r), which again would suggest a copyist's dittography
rather than a fresh composition. There is also a case (f.3r) where the copyist has carelessly missed out a line,
perhaps rather less likely if a writer were composing and paying attention to the content of the main text.
Finally, an annotation at the opening of Chapter Eleven seems to indicate that the annotator is working with
the original chapter numbering, not the re-numbering given in the Dunhuang manuscript, which has lost
Chapters Thirty-one and Thirty-two. Thus, the note refers to the chapters on the rituals continuing up to
Chapter Thirty-nine, whereas in the re-numbered Dunhuang manuscript, they only go up to Chapter Thirtyseven.4 Nonetheless, although it is clear that the annotations must have been copied from an earlier
exemplar, there is also evidence that the original annotator was working after some textual corruptions had
been incorporated into the transmission. In the mantra list in Chapter 9, the set of annotations gives the
Tibetan equivalent names for the deities whose Sanskrit names are incorporated into the mantras. In the face
of an omission of four expected mantras, the annotator appended their names in an inappropriate position
(see p.53 below).
The manuscript mostly conforms to classical Tibetan spellings, although as we might expect, it also gives
many archaic and non-standard spellings. Even in these cases, its spellings are generally consistent and there
is no problem deciphering them. Archaisms include much use of the attached letter ya (ya btags), such as
mye for me, myed for med, myig for mig, and so forth, and some use of the reversed gi gu. The extra final
letter da (da drag) occurs but is rare (e.g. gyurd, 83v.6). There is frequent (but not exclusive) separation of
the genitive particle, 'i, given with a tsheg in front of it, such as in pa 'i, rather than pa'i, spu 'i, rather than
spu'i. The a chung final letter may be found in a few words where it is not in standard Tibetan (since it is
redundant): e.g. brgya' (31r.4, 48v.6, 49r.2) and 'byung ba' (45v.6). It may also be included where a word
3

Note that there is also some evidence that the annotations to IOL Tib J 331.III have similarly been copied from a previous
manuscript (see Cantwell and Mayer (2008: 94 note 24, 96 note 35). By the time IOL Tib J 321 was copied, the inclusion of a
set of annotations elucidating the tantric text, copied from one manuscript to the next, was a well-established procedure.
Note that there are are six occasions in the manuscript when the scribe has run out of space for an annotation and extends it
vertically up or down the right-hand side of the page. At first sight, this might suggest a fresh composition, since a copyist
could more easily estimate the space needed and adjust the positioning accordingly. However, generally throughout the
manuscript, where possible and where not interfering with a previous note, each annotation begins at the appropriate place
beneath (or less often, above) the words to which it relates, regardless of the amount of space it needs. This is done in each of
these six cases, and the lack of space is an inevitable consequence.

34

Textual Analysis

with a final a chung is marked with a final sa, e.g., sems dpa's rather than sems dpas (40r.1). The words,
mtha' and dga' are invariably given with the final a chung subscribed. Other words similarly given with the
final a chung subscribed are na bza' (31r.2, 4), bka' (34r.2 and 5, 38v.5, 53v.1 and 5, 70r.6, 82v.4), mkha'
(35r.2, 36v.4), 'ga' na (45r.5), gza' (52v.3, although here, the syllable thus avoids the string-hole circle), and
mud dga' (65r.3). However, other words do not have the a chung subscribed, for example, brgya' (31r.4,
although the size of the letter below the line would not make it easy to fit), btsa' (54r.4), and dpa' in sems
dpa' and dpa' bo.
The particle, ba, is frequently given for pa. In this case, it is often possible that pa is intended and this
apparent confusion is a writing style in which the two letters are virtually identical. Occasionally, a prefixed
or main letter ba also resembles pa (22v.3 brtan, 22v.4 'bar). Alternatively, it may be a spelling convention:
in texts produced in A mdo, the same convention can be seen today (for instance, throughout Mag gsar
2003). At various places, we have las stsogs pa as an alternative to la stsogs pa, which also occurs, where in
standard Tibetan we would find, la sogs pa. We find yan cad and man cad for yan chad and man chad
respectively (e.g. 31r.4-5, 71v.2), but this is a little inconsistent. Both yan cad and yan chad are given on
folio 81v.3. There is a similar usage of ca for other words which are conventionally spelt with cha. Thus we
find rin cen (in Chapters Two, Five, Six, and elsewhere), rgya cen (in Chapters Ten and Sixteen), mtshon ca
(or tshon ca) and dung cen (in Chapters Eleven and Twelve), mthu cen and dpal cen (in Chapter Eleven), but
again, this is not entirely consistent (dpal chen also occurs, as does chen po), tshigs cen and stabs cen (in
Chapter Sixteen), stobs cen (in Chapters Sixteen and Seventeen), gtum cen (in Chapter Twenty-two), and
dbang cen (in Chapters Twenty-four and Twenty-eight).
There are many instances of g.yos for g.yon, given in the expression g.yas g.yos (for g.yas g.yon, in
Chapters Eleven, Twelve and Sixteen). Where the word, g.yon, occurs alone, it is consistently given with a
final letter, na, but there is a consistent use of the spelling g.yos in g.yas g.yos. This would seem to be an
archaic spelling for this phrase; it also occurs in another Dunhuang text, IOL Tib J 557 (recto side, second
text, lines 1 and 14; verso side, line 11).
In Chapters Seven, Twelve and Thirteen, we find the archaic ral gyi for ral gri (see Tshig mdzod chen mo
Vol.3: 2667). There are a number of instances of long spyod for longs spyod, a spelling which of course
persists into modern times. In various places, we find nam ka for nam mkha', on several occasions in
Chapter Six, in Chapters Eleven, Twelve, and the final unlabelled chapter, corresponding to Chapters Fortyone and Forty-two of the root text. In Chapter One, rim (for stage/stages) is not invariably but fairly
consistently spelt rims. This also occurs in Chapters Six, Seven and Twelve. A number of times (in
Chapters Three and Seven, especially in the annotations), we find rigs/rigs pa for rig/rig pa, but rig/rig pa
also occurs. Similarly, Chapter Eight gives the spelling, rigs 'dzin, and Chapter Nine gives, rigs sngags. In
various places (Chapters One, Two, and the final unlabelled chapter), we find klung, most probably for
klong. This also occurs in Chapter Thirteen (53r.4), but klung, meaning, stream, would also work here. A
few times, the word, buddha, is transliterated as, 'bu ta (1r.1, 32r.4, 33ar.1). Twice in Chapter Six, and also
in Chapters Seven and Nine, we find sgreg mo for sgeg mo. The spelling, phub shud, for pu shud/phu shud,
is given on two occasions (54v.3, 56v.4; see bTsan lha ngag dbang tshul khrims 1997: 483).
Other non-standard spellings (which might also be spelling errors) are mkhyud, given for 'khyud, and
dbung rather than dbus, both in Chapter Six. In Chapter Five, there are two instances of brgya cu, apparently
for brgyad cu (interestingly, Chapter Two of the South Central edition of the Phur bu Mya ngan las 'das pa'i
rgyud chen po has a similar instance of brgya bcu, Cantwell and Mayer 2007: 127). The spelling, 'phral ba,
is given for dpral ba (55r.2), and gzhu nu is given for gzhon nu (82v.4). Finally, in the pagination in the
margins, the twenties are given as nyar rather than the more usual nyer.

c) Examination and Assessment of the Differing Boundaries of TZ in Different Editions

As with its Stric counterpart, Tantric scriptural transmission in Tibet comprises two parallel but
interdependent dimensions: the transmission of root scriptures, and the transmission of commentaries upon
the root scriptures. To understand the complexity of these two dimensions of transmission in the case of the
Thabs zhags, we must consider that while the root tantra TZ in all likelihood did originally exist as a standalone text, this stand alone version seems in some important cultural regions to have been displaced quite
early on by the commentary TZComm, in which a complete version of TZ was embedded. We now know
from stemmatic analysis that the stand alone version of TZ did in fact leave its traces within various editions
in the west, east and southern margins of the Tibetan cultural regions. However, by the time the Tshal pa
Kanjur and the ancestor of the Bhutanese NGB editions were being compiled, this stand alone version seems
to have been lost to view in several of Tibet's most influential cultural centres. We can infer this because we
have established that the redactors of the version of TZ shared by the Kanjur and the sDe dge NGB, as well
as the version shared by all the Bhutanese NGB manuscripts, felt compelled in each case to attempt their
own independent effort at re-extracting TZ from its commentary, TZComm. Presumably, they only did this
because they had no viable stand-alone version of TZ to refer to. On these two separate occasions of reextraction, the respective redactors arrived at somewhat different conclusions about what constituted
commentarial text and what constituted root text, with the result that different versions of what constituted
TZ became simultaneously current in different regions and editions. Here, we will focus on some of these
historically most influential canonical editions of TZ, the ones that were and still are used by most Tibetan
readers, and reflect on the different decisions their editors made about the text's boundaries.
The three currently most widely used versions of TZ (that is, the Tshal pa Kanjur/sDe dge NGB, the
Bhutanese NGB, and the South Central NGB) sometimes differ from each other radically. At first glance,
one might imagine the variation to result from an open recension, like those of the Kanjur texts where
different Tibetan translations, sometimes also deriving from differing Sanskrit sources, had descended
separately and then interacted within Tibet. If the Dunhuang finds had never been recovered, or if the local
Kanjurs were not consulted, this is possibly the kind of conclusion philologists might have come to for TZ.
But on closer inspection, the evidence actually indicates otherwise. The differences between the two extant
versions of TZComm can largely be explained by transmissional error causing loss or corruption of text in
one or the other version, while the striking variations between the different recensions of TZ itself appear to
have been generated from confusion as to which words of TZComm were lemmata, and which words were
commentarial. We believe the South Central NGB grouping together with the three local Kanjurs of
Bathang, Hemis, and Tawang, preserve the original boundaries of an archetypal stand alone root text TZ.
However, the historically much more influential Tshal pa Kanjur/sDe dge NGB and Bhutanese NGB editions
of TZ do not descend directly from an original stand alone root text TZ: they descend via TZComm, and on
these two different occasions, quite different and mutually varying decisions were taken about which
passages were TZ's lemmata and which were commentarial. So even though stemmatic textual analysis can
rarely produce absolutely incontestible proof, and in this case we may never be able absolutely irrefutably to
exclude all other possibilities such as an open tradition,1 nevertheless it is quite undeniable that a very major
factor in the historical transmission and variations in TZ over the last ten centuries has been its fuzzy
boundaries with its own commentary, as evidenced within its most widely used editions.

There are several possible scenarios, although only one likely one: a unitary root text might have begun life independently, and
later become expanded and contracted through editorial reference to the commentary; alternatively, a unitary root text might
first have appeared in Tibet embedded in its commentary, and variation generated as scribes tried to separate the lemmata from
the commentary; or a further alternative would be that there might have been an open tradition in which several differing
original root texts interacted over time with each other and with the commentary. We believe only the first of these
possibilities is at all likely in this case.

36

Textual Analysis

There are parts of TZComm where all the lemmata are made perfectly clear by the structure or wording of
the text. For example, most chapters commence with a clearly framed four line verse from TZ, or quite
often, passages of TZ embedded in TZComm may be signaled by wording such as, zhes gsungs te (it is said).
However, there are large sections where the lemmata are not explicitly marked off in such ways. The fact
that TZComm does not always clearly delineate the boundaries of TZ means that we should not be surprised
that editors attempting to extract TZ from TZComm faced a significant challenge. Were it not for stemmatic
analysis, it would be hard now to assess the relative merits of the differing versions of the textual boundaries
reflected in the different extant recensions of TZ. But we can have confidence that it was the Tshal pa
Kanjur and the Bhutanese NGB which failed correctly to preserve the original textual boundaries since, as
we shall show below (see p.50-54), they are part of a sub-branch of the transmission (the descendants of
hypearchetype c on the stemma codicum, p.43). In contrast, the South Central Tibetan NGB and each of the
local Kanjurs directly descend separately from the archetype, and their agreement indicates that they
preserve the original boundaries.
Highlighting of the Root text in the Dunhuang Manuscript
The Dunhuang scribe(s) clearly understood the importance of distinguishing root text from commentary.
In the Dunhuang manuscript of TZComm, lemmata are often highlighted with the semi-transparent wash
sometimes found in Indian and Tibetan manuscripts that is similar in principle to modern highlighting ink.
Unfortunately, it seems not to have been applied completely or consistently. The exact original extent of the
highlighting is uncertain. In some cases, the wash may originally have been applied, but later vanished.
There are also other cases where it has seeped through the page, so that some text on the reverse side
mistakenly appears to have been highlighted, although this is mostly a problem only for modern readers of
the digital images, since this feature is clearer when reading the actual document. In any case, the
inconsistency of the highlighting leaves major ambiguities. Sometimes, text that would clearly appear to be
part of the root text, and is so indicated by the phrasing of the text, is not highlighted. The dissolution of the
wash is not likely to explain all such examples of the absence of highlighting. For instance, there are no
disagreements between the different versions of the root text in Chapter 18, the root verses in this case
indicated clearly also by TZComm's wording, yet less than half of the root text is highlighted, in some cases
merely the opening words of a verse, in others the closing line or a line in the middle. The highlighted and
unhighlighted lines of the root text are beside each other on the same pages (62r-v), and it seems unlikely
that the wash dissolved neatly on some lines, but was preserved intact on others. A similar point can be
made about the contrast between the highlighted and unhighlighted lines of root text in Chapter 24 (67v).
It seems quite possible that for some areas where the commentary merely reproduces root text, the
manuscript's scribe may be satisfied with highlighting the opening words or otherwise leaving long passages
of root text unhighlighted. In fact, the chapter titles are also unhighlighted, although these are also given in
the root text editions. It is clear that highlighting of the chapter titles would seem redundant since it is
obvious that they are root text. In cases of long passages, it might similarly be assumed that the whole
passage is root text, and we need caution not to interpret absence of highlighting as necessarily indicating
that the text is not considered part of the root tantra. But even with this kind of example there is some
inconsistency: Chapter 9 on mantras is almost entirely root text, and highlighted throughout, yet the
following Chapter 10 on mudrs, which is also virtually all root text, is not highlighted at all, not even its
opening lines.2

Chapter 16 on the mudrs for the wrathful deities similarly lacks any highlighting, and the following Chapter 17 is somewhat
similar. Only the opening words are not root text, but there is no highlighting except for a few strings of words, and these
make no sense as root text apart from the text in which they are embedded. Chapter 14 which consists of the mantras for the
wrathful deities, begins like Chapter 9 by highlighting all the mantras, but then the highlighting strangely stops in the middle
of the mantra for the seventh wrathful deity (some apparent highlighting which appears on the digital image of the final line of
58v is merely the seeping through of highlighting from the other side of the folio). Following the mantras for the ten wrathful

The Differing Boundaries of the Root Text in Different Editions

37

More rarely, text which is not found in any of our extant root texts is highlighted. Very occasionally, the
manuscript highlights words or phrases that it clearly does not recognise as root text for instance, a few
citations of other texts are highlighted. Thus, in Chapter 3 (12r.6), half a verse citation from the Glang po
(clarified in the manuscript's annotations as indicating the Glang po rab 'bog lung) is highlighted in exactly
the same way as the root text highlighting. In this case, it is quite clear that the highlighting is not indicating
the TZ root text, and similarly, in Chapter 8 (30v.5-31r.2), one highlighted verse is attributed to the sGron ma
brtsegs pa and a following highlighted verse attributed to the rDo rje sems dpa'i le'u.3 However, the first
line of this first verse attributed to the sGron ma brtsegs pa is also given in Chapter 6 (17r.6), where it is
both highlighted and indicated as root text by the wording, zhes gsungs te, yet no mention is made of the
sGron ma brtsegs pa. A reader, then, might mistakenly assume that this line is intended as a line from the
TZ root text, although none of the editions of TZ in fact make an error here. In Chapter 19 (63v.3-4), two
tshig rkang which could certainly represent root text are highlighted, yet they are not included in any of the
editions of TZ. In various parts of the manuscript, a few words are apparently inappropriately highlighted, or
at least are not accepted as root text in any of the TZ editions (e.g. phur pa lnga btab la, Chapter 25, 69r.6).
Often, however, these are merely words added either at the beginning or the end of a highlighted root text
verse (e.g. Chapter 5, 14v.6, Chapter 26, 70r.3). Overall, the highlighting of text which does not seem to be
recognised as root text by any TZ edition is of little significance in comparison with the reverse, that is, the
significant sections of root text or text accepted as root text by the witnesses descending directly from the
archetypal TZ text which are not highlighted by the Dunhuang manuscript. The main point in drawing
attention to the highlighting of text which is apparently not root text is the confusion which it might
potentially generate.
A further source of possible confusion is in the case where TZComm repeats the words of the root text in
the course of its elucidation of the text. Generally, the manuscript does not highlight on the second and
subsequent references to the root text words, but occasionally it does highlight them. Mostly, however, there
is no problem since it is quite clear the root text is not repeating itself.
We cannot know how far other ancient manuscripts of TZComm might have applied this kind of
highlighting to draw attention to root text lemmata, and there is no evidence for it in the modern Tenjur
versions. However, it does enable us to see how the scribe of the Dunhuang manuscript understood the
boundaries of the root and commentary, at least for large sections of the text. And it is striking that the
highlighted sections neither correspond neatly to the root text as given in our edition, representing the most
likely archetype of TZ, nor to the root text as understood either by the Tshal pa Kanjur or the Bhutanese
traditions. Notwithstanding the possibilities that some of the highlighting may have worn off, or that the
scribe might have deliberately left long passages of root text unhighlighted (see above), there are some parts
where the Dunhuang manuscript's lack of highlighting seems to show that the manuscript's scribe did not
identify additions found in the Tshal pa Kanjur or Bhutanese versions as words of the root text. Conversely,
there are a few occasions where the manuscript's highlighting agrees with the Tshal pa Kanjur or the
Bhutanese NGB root texts, in opposition to our proposed edition (and the South Central and local Kanjur
versions).
Variations in the identification of Root text by the different editions
Let us first say that there are sections of the text where there is no confusion about the boundaries of the
root text. For example, there is agreement throughout Chapters 2, 9, 12, 21-23, 27-28, 30, 33, and 35-40, and

deities, there are a few words of commentary, and then a further mantra, the second half of which is highlighted, as are the last
mantras given in Chapter 14. After the opening highlighted words in the second section of the final chapter (which in fact
corresponds to the opening of a separate Chapter 42 in our textual edition restoring the chapter structure of TZ's archetype), the
first two verses are unhighlighted although recognised by all editions as root text, but then from the third verse, the
highlighting begins again (see folios 82v-83r).
This may be intended to indicate a separate text, or the Vajrasattva chapter within the sGron ma brtsegs pa.

38

Textual Analysis

apart from the chapter titles which are consistently unhighlighted, the root text of these chapters is
highlighted in full in the Dunhuang manuscript. In other sections (such as Chapters 13, 15-18 and 20), the
only anomaly is that the Dunhuang manuscript fails to highlight the root text in full. There is also no
disagreement over the extent of the root text in Chapters 31 and 32, chapters which are omitted in the
Dunhuang manuscript. Elsewhere, there are also sections where there is little difference between the
different versions, but put together, the disagreements are considerable in extent.
In Chapter 1, there are two minor disagreements. Some words embedded in TZComm, unmarked as root
text in the Dunhuang manuscript, are included only in the Bhutanese NGB edition of TZ. The words are not
necessary for the root text; like the rest of the passage of commentary in which they are embedded, they
clarify TZ's statement that the two truths are the cause and effect.4 Further on in the chapter, a phrase which
opens a verse of TZ, sems dpa' chen po nyon cig / (Listen, Great Being...), is omitted in the Tshal pa Kanjur
version alone. At this point, none of the root text is highlighted in the Dunhuang manuscript, so the
Dunhuang scribe's assessment here is unclear.
In Chapter 6, the Bhutanese version accepts rather more than the other TZ versions, agreeing with the
Dunhuang manuscript's highlighting of these extra words, which consist of one phrase and a series of
mantras, separated by other words of commentary. Despite the highlighting in the Dunhuang manuscript, the
Dunhuang scribe's assessment here is not altogether certain. In the case of the initial phrase, the highlighting
might merely have been drawing attention to the words, and the wording of the commentary does not clearly
signal either the phrase or the mantras as root text. Moreover, the Dunhuang manuscript only shares the first
words of the initial phrase (tshogs kyi dkyil 'khor); the following words are found only in the Bhutanese
versions of TZ. Yet it is quite likely that the Dunhuang version does assume the additions which are present
in the commentary to be root text. It does not generally highlight mantra syllables elsewhere.
The Bhutanese version of TZ also accepts as root text more of the series of mantras found in TZComm in
Chapter 7, but there (apart from a single syllable) the Dunhuang manuscript does not highlight the mantras
concerned. Moreover, in TZComm, the mantras are split into individual syllables, and separated by a good
deal of extra text. There is also a similar case of a few mantra syllables in Chapter 8, included in only the
Bhutanese version of TZ. Most but not all of the syllables are highlighted in the Dunhuang manuscript.
There is a further disagreement over the inclusion of other mantra syllables in Chapter 7. The five root
text verses following the opening verse, are each followed by a line of five mantra syllables. These verses
and five syllable lines are clearly highlighted in the manuscript, but the mantra syllables are omitted in the
Tshal pa Kanjur version of TZ. Since the verses are invoking the presence of the sets of each of the five
buddhas and his accompanying bodhisattvas arising at the different body parts in turn, it would certainly
seem appropriate to include the seed syllables in the root text, but the Tshal pa Kanjur editors seem to have
assumed them to be extra commentarial mantras.
In Chapters 7, 8 and 10, the Tshal pa Kanjur includes one line which is clearly commentary, clarifying
that Vajrasattva is being addressed,5 inserted within each chapter's root text opening verse. This is a rare
case of a disagreement about which, even without the benefit of stemmatic analysis, it would be clear that the
commentarial gloss has intruded into the root text in the Tshal pa Kanjur.
4

TZComm in translation gives (with the words included in the Bhutanese root text in italics): Since all dharmas arise without
characteristics, the ultimate truth is (their) cause. Since they appear like an illusion, (this) relative truth is to be known as the
result. (chos thams cad mtshan ma myed pa las byung bas/ don dam pa ni [QtGt pa'i] rgyu'o/ /sgyu ma tsam du snang bas
[QtGt insert /] kun rdzob ni 'bras bur rig par bya'o/ Dunhuang manuscript 4r.3-4; Gt 247-248). The Bhutanese texts give: don
dam pa ni rgyu'o / / kun rdzob ni 'bras bu yin par bya'o / (G51v.3-4).
zhes rdo rje sems dpa' la thos nas [Ms bos te] bshad do in Chapter 7; ces [Ms zhes] rdo rje sems dpa' la [Qk insert /; Nk omit
la] bos nas [Ms te] bshad do, in Chapter 8; ces rdo rje sems dpa' la bos nas bshad do, with the same variants in the Dunhuang
manuscript, in Chapter 10.

The Differing Boundaries of the Root Text in Different Editions

39

In Chapter 8, there would appear to be a case where the Dunhuang manuscript lines up with the Tshal pa
Kanjur against the other traditions. There are two verses, highlighted in the Dunhuang manuscript (apart
from the final two tshig rkang), which are found only in the Tshal pa Kanjur versions of TZ. In TZComm,
they are not altogether clearly marked off as root text by the wording (see TZ edition Chapter 8 below), but
the Tshal pa Kanjur editors and the Dunhuang scribe seem to have taken them to be root text.
In Chapter 11, we have a major disagreement over the boundaries of the root text. Only the opening two
verses6 and the mantras of this chapter are highlighted in the Dunhuang manuscript, yet the whole text is
within the Tshal pa Kanjur version of TZ, with no commentary omitted at all. Much of the chapter is
included in the other versions also. The long unhighlighted section at the beginning of the chapter is
accepted by all TZ versions, and these sections come across as the style of root text.7 However, in a lengthy
passage in the middle of the chapter, the Dunhuang manuscript's decision to highlight only the mantras is in
agreement with the South Central Tibetan, the Bathang and Tawang local Kanjur editions of TZ (the passage
is unfortunately lost through folio loss in the Hemis local Kanjur), which include only exactly the same
mantras as those highlighted by the Dunhuang manuscript. The Bhutanese version is essentially the same,
but adds a list of eight zoomorphic goddesses found further down in TZComm. At this point, the Dunhuang
manuscript's highlighting of the mantras suggests that it correctly identified the root text in the mantras
alone, especially since the manuscript does not generally single out mantras for highlighting.
We know from stemmatic analysis that the South Central NGB and local Kanjur versions remain faithful
to an ancient and correct tradition as represented in this case in the Dunhuang manuscript through its
highlighting of the mantras, while the Tshal pa Kanjur version has mistakenly allowed a good deal of
commentary to intrude into the root text. Yet there are powerful apparent justifications for the Tshal pa
Kanjur reading. The text at this point appears seamlessly to continue the reported conversation begun at the
beginning of the chapter, between Vajrasattva and Vajrapi (see note 7). In other words, it takes the literary
form of root text, the actual speech acts of the Buddhas, rather than the commentarial utterances of a human
voice. Thus the question remained for the Tshal pa Kanjur editors, does the passage count as commentary,
or is it intended as an integral part of the root? The text here describes and comments on the effectiveness of
the female deities who are listed in the first part of the chapter, and whose mantras are now given. The
phrasing of the additional text not given in the South Central, Bathang, Tawang and Bhutanese versions
could from appearances alone perfectly well be either root text or commentarial meditative instruction. It is
almost impossible to adjudicate between the different readings without stemmatic analysis.
Similarly, if judged purely on the basis of the semantic content, the passage which closes the chapter
could equally constitute root text or commentary. This final section is in fact included in full within the
South Central and local Kanjur traditions, but the Bhutanese version gives only the mantras.
The final mantras of Chapter 14, which are appropriately highlighted in the Dunhuang manuscript, are
omitted in the Tshal pa Kanjur version. This omission may even be a simple transmissional error, since they
are indicated in the chapter title.8 Nonetheless, it is possible that the error is not a scribal lapse, but a
deliberate decision by the Tshal pa Kanjur editors that these additional mantras for the ritual actions should
not be part of the root text.

6
7

In fact, not quite all - the final line of the second verse is unhighlighted.
This chapter consists of reported conversation between Vajrasattva as the expounder of the tantra, and Vajrapi as his
interlocutor. Elsewhere in TZ, however, it is Vajrasattva who is the interlocutor, and Vairocana the expounder: might this
section then be an interpolation from another Tantra? Yet if so, it seems that the archetype had already incorporated it.
"this is the fourteenth chapter, on the mantras for the wrathful deities and for ritual action combined together" (khro bo bcu'i
rtsa ba'i sngags dang las kyi sngags ril 'dus pa'i le'u ste bcu bzhi pa'o).

40

Textual Analysis

Following a passage included by all TZ versions in Chapter 19 (which is nonetheless only partially
highlighted in the Dunhuang manuscript), there are two mantras which make up the final words of
exclamation closing the passage. The second is highlighted in the Dunhuang manuscript. Both mantras are
included in the South Central, local Kanjurs and Bhutanese versions, but neither are given in the Tshal pa
Kanjur version.
Chapters 17, 24, 29, and 34 each begin a series of chapters on one of the four types of rituals, starting
with the destructive rituals (drag), then the captivating (dbang), increasing (rgyas pa), and pacifying (zhi ba)
rituals respectively. In each case, there is a verse which follows these chapters' opening root verses. These
second verses (or the third verse in the case of Chapter 17) parallel each other, for example, each of them
introduces an appropriate vajra, respectively the hatred vajra (zhe sdang rdo rje), the passion vajra (chags
pa'i rdo rje), the pride vajra (nga rgyal rdo rje) and the delusion vajra (gti mug rdo rje). The Dunhuang
manuscript partially highlights the verses in Chapters 24 and 29, and wholly highlights the verse in Chapter
34 (it omits the verse in Chapter 17, presumably in error). The Bhutanese version omits all these verses. It
seems clear that this is a deliberate editorial policy rather than an accidental omission, since the Bhutanese
version is consistent in its treatment of all four parallel verses. As in other cases of disagreement, it would
be hard to assess the relative merits of including or excluding the verses without the benefit of stemmatic
analysis. The root text is not rendered incoherent without them. Moreover, TZComm marks these additional
verses with the words, zhes bstod de, rather than, zhes gsungs te, which is more usual in its marking of root
text, so in this sense, appears to treat them a little differently.
The Dunhuang manuscript does not highlight any of the text within the penultimate Chapter 41, which
some versions count as the first part of the final chapter,9 not even the opening (although there is in any case
no clear opening verse marked off by the wording). The Bhutanese NGB version appears largely to accept
the apparent assessment of the unhighlighted Dunhuang manuscript, jumping from the chapter's opening
words to the content of Chapter 42 (which in some versions is not presented as a separate chapter; see note
9). Yet the Tshal pa Kanjur version includes every word of this section of the commentarial text, while only
the South Central NGB and three local Kanjurs correctly show a few lines from it. In this instance, the
whole section reads like the commentarial voice, glossing the meanings of the maala features described
earlier in the text, suggesting a good case for the Bhutanese decision in contrast to the Tshal pa Kanjur's
assessment. However, the specific lines given in the South Central NGB and the local Kanjurs list some of
the features without the glosses, and these make good sense as root text and are actually correct. Their
chapter title for this section, The Maala of the Great Captivator (dbang chen bsdus pa'i dkyil 'khor gyi
le'u), is appropriate for these lines, and survives also in the Tenjur version of TZComm.
Conclusion
Looking at the variations as a whole, it is not entirely clear that the different decisions can be neatly
categorised as consistent policies applied throughout the Tshal pa Kanjur or Bhutanese editions. For
instance, an examination of the sections of Chapters 8, 11 and 41 included in full in the Tshal pa Kanjur
alone, might suggest a policy of the Tshal pa Kanjur editors of inclusion in instances of doubt, and certainly,
the Tshal pa Kanjur version of the root text is considerably longer than the others. Yet it is the Tshal pa
Kanjur version which omits mantras in Chapters 7, 14 and 19, not to mention its failure to include various
short passages or phrases. The Bhutanese version seems even less consistent, in some instances including a
good deal of extra material (such as the mantra syllables in Chapters 6-8), while elsewhere omitting verses
(in Chapters 24, 29 and 34, as well as virtually the whole of Chapter 41).
We have seen that ancestors of the nowadays most prestigious editions of TZ, including the whole group
of Kanjurs of the Tshal pa line of descent along with the sDe dge xylograph NGB, and also the group of
9

The Dunhuang manuscript, Tshal pa Kanjur and Bhutanese NGB versions lose the title and thus the marking of the chapter
break at the end of Chapter 41. See the chapter endings given in Chapter 41 of our TZ and TZComm editions.

The Differing Boundaries of the Root Text in Different Editions

41

Bhutanese NGB manuscripts, have attempted to extract the root text from the commentary. They have failed
to identify the root text's boundaries in various places, displaying differing and particular decisions. Thus,
despite the persistence of the South Central NGB manuscripts with their continued preservation of the root
text's original boundaries, the Tibetan tradition as a whole has been quite undecided about TZ's boundaries,
perhaps for the last thousand years. There is no easy or obvious way of adjudicating most of these variations
without stemmatics. Although certain readings might appear preferable to the unaided learned eye, some of
their alternatives might nevertheless still appear defensible. Excluding the version which seems
unproblematically to descend from the original root text, at least four different decisions about TZ's
boundaries survive in the extant literature, and we have no idea how many others might have existed in the
past.10
A further difficulty facing Tibetan editors in accurately separating TZ's root text from its commentary is
that the broad or general literary conventions differentiating Buddhist root tantras from their commentaries
need not always be rigorously observed in all cases. As in Chapter 11, a commentarial text passage can
frequently look like root text, and vice versa: indeed, there are some occasions where a quite considerable
intertextuality can exist between commentaries and root tantras, as we have remarked in our previous works
on the Phur pa literature (Mayer 2004:134-136; Cantwell and Mayer 2008:148-149). A consequence of such
stylistic ambiguity means that root text lemmata can sometimes only be differentiated from their
commentary where some explicit literary device is consistently employed. The usual such device is to mark
out lemmata with such words as, zhes gsungs te, and in addition highlighting can sometimes be used, but in
this case neither were used with absolute consistency throughout the text. Thus on most occasions where the
Tshal pa Kanjur and/or Bhutanese versions diverge from TZ's original boundaries, in the absence of
absolutely consistent marking of lemmata in TZComm, the differences become extremely difficult to resolve
without recourse to complex stemmatic analysis. In fact, there are only a few occasions where resolution is
unambiguously possible without stemmatics, such as the example of the single line of commentary
explaining that Vajrasattva is being addressed, which intrudes into the Tshal pa Kanjur's Chapters 7, 8 and 10
(see above p.39). If the inadequacies in the Dunhuang manuscript's highlighting or application of verbal
markers such as zhes gsungs te is anything to go by, it is quite possible that the Tshal pa Kanjur and
Bhutanese NGB editors may not have been much helped by the layout of the TZComm witnesses they were
working from.
Perhaps there is little option but to allow, as do Tibetan lamas with some reluctance in actual social
practice, that the text of TZ has over the centuries taken several differing forms, most of them equally
defensible on purely semantic grounds. Despite the probable unitary origins of the text in a single archetype,
undoubtedly, the boundaries between the root text and the commentary have proven fuzzy for a thousand
years or more. If we accept an early interpretation of TZComm especially in its Dunhuang witness as
suggesting that TZ was originally associated with Padmasambhava (see below p.91-98), whom tradition
considered a fully enlightened being in human form who resembles the later rNying ma treasure revealers by
being able to produce both scripture and commentary alike, then it looks harder still to fix the fuzzy
boundaries without the benefit of stemmatics. The true text of TZ has remained singularly difficult to
separate from its commentary to this day and probably would have always remained so, without the benefit
of modern stemmatic analysis.
There is a further implication of this clear case of permeable boundaries between an NGB root text and its
commentary. Since so few NGB texts still have extant commentaries, it is not possible to assess empirically
10

In addition to the decisions represented by the Tshal pa Kanjur and Bhutanese groupings of root texts, we can add TZComm's
implicit marking of lemmata by its wording in some parts of the text, together with its apparent failure to mark off the lemmata
in other parts of the text (making its assessment of the boundaries uncertain in these areas of the text). Finally, we have the
added layer of the Dunhuang manuscript's version of the commentary with its highlighting, apparently clarifying TZComm's
assessment, but in fact, still leaving some uncertainty, and apparently underestimating the extent of the root text.

42

Textual Analysis

whether such restoration of root text editions through extraction from commentaries might once have been
more widespread. Accretion of commentarial text is a very well known phenomena in texts of all kinds, and
it may have happened in other NGB tantras as well. The Tshal pa Kanjur version of TZ suggests that its
editors at some junctures exercised caution by including more rather than less. Should it be the case that
other NGB texts became expanded in a similar fashion, then this would imply that the mere presence of some
Tibetan features within an NGB tantra need not in all cases be considered sufficient grounds to prove that its
main provenance was more Tibetan than Indian.

d) The Stemma of the Root Text

Diagram of the proposed Stemma of The Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi
don bsdus pa (TZ)
Note that the solid lines of descent indicate merely the direction of transmission, often through many
generations of copying, and the length of these lines has no significance. Thus, the South Central
transmission is likely to have descended through far more copyings than the Dunhuang manuscript. The
arrows represent tentative possible directions of limited contamination, from a branch other than b.

a origo

b (a
commentarial
transmission,
incorporating the
root text)

The Tenjur
edition
(Qt, Gt, Nt)

The South
Central rNying
ma'i rgyud 'bum
(T, R, K)

The Bathang
Kanjur
(Bth)

The Hemis
Kanjur
(He)

The Dunhuang
Manuscript (Ms)

The Tshal pa
Kanjur
(bKa' 'gyur)
(J, Qk, Nk, U,
Dk, D, Hk, V)

The Bhutanese
rNying ma'i
rgyud 'bum
Manuscripts (M,
G, G-a, Gr)

The Tawang
Kanjur from
O rgyan Gling
Temple (Ogl)

Textual Analysis

Discussion of the proposed Stemma

Stemmatics is a notoriously elusive art, sometimes yielding extraordinary clarity, at other times mere
uncertainty or worse. Its fruitful application demands a frank acceptance that the full range of historical
evidence is almost never available to us, and a consequent restraint of judgement that refuses to stray beyond
what is actually knowable. Such a focus on what can be said with confidence must then be complemented by
a realistic weighing up of further probabilities and possibilities. In this presentation of our stemmatic
reasoning, we therefore progress from the more certain conclusions supported by seemingly strong evidence,
to a number of less certain possibilities and speculations where the evidence remains less conclusive.
Possibly due to the importance of TZ as one of the key Mahyoga tantras, or perhaps because of the
protective and clarifying effect of its commentary, we do not see the kinds of significant recensional
variation we found in editing the NGB's Myang ngan las 'das pa'i rgyud chen po, nor the latter's long
obscure passages where contrasting attempts to solve a textual crux are most in evidence (Cantwell and
Mayer 2007: 22, 79-85). In TZ, the variants are mostly the obvious result of scribal lapses, non-standard (or
archaic) spellings, or variation created by its transmissional relation to TZComm (as described in section c
above), where the root text has in two of its recensions both gained and lost different passages of text in the
process of its extraction from the commentary.
Our procedure in creating this edition has been first to identify the main natural groupings and branches
within the textual transmission, and from that basis attempt a restoration of the archetype. There are a large
number of textual witnesses, and several of them are from a stemmatic viewpoint largely repetitive or
descriptive, so that the twenty-one witnesses of the text we have had to work with can for present purposes
most conveniently be taken to represent only eight distinct redactions. As explained above, we have therefore
not presented a collation of every witness throughout the entire edition. Rather, we present representative or
illustrative samples from each of the witnesses, plus a full collation throughout of the principal or most
suitable witnesses from each of the main groupings. Wherever significant variation occurs, the in most cases
confirmatory testimony of additional witnesses is brought to bear, even if repetitive.
To justify this policy to our readers, and to explicate the foundation of our stemma, we must therefore
first establish that a large number of the available witnesses are indeed merely descriptive or repetitive, and
that we are in fact justified in assessing TZ for present purposes to survive in only eight stemmatically
significant redactions. Hence we begin this discussion of the stemma by showing that the component
members of the four textual groupings of the Tshal pa Kanjur, the Bhutanese NGB, the Tenjur, and the South
Central NGB, have little else to offer than repeated presentations of their same respective testimonies.
The Eight Tshal pa Kanjur texts as multiple witnesses to a single redaction
Of the two major branches of the Kanjur, only the Tshal pa included a rNying rgyud section, complete
with a version of TZ. The Tshal pa rNying rgyud section enjoyed a number of re-editions over the centuries,
and we have collated sections from the versions of TZ deriving from eight of them, including in this case the
sDe dge NGB and the mixed Kanjurs of sNar thang, sDe dge, lHa sa, and Urga (see sections a and c above).
Apart from a small number of minor transmissional variations and a few trivial corrections, these eight
versions deriving from the Tshal pa rNying rgyud section have otherwise all proven identical, and can
therefore more usefully feature in our stemma as a single redaction, rather than as eight separate witnesses.
From the point of view of internal textual evidence, their descent from a common ancestor not shared by any
other extant version of TZ is most graphically demonstrated in their entirely unique decisions on the
boundaries of the text. The only plausible explanation for the intrusion of extensive passages of commentary
into this edition, and the omission of a few much shorter passages (as described in section c above), is that an

The Stemma of the Root Text

45

ancestor compiled TZ by extracting lemmata from a version of TZComm, and made mistakes in doing so.1
Throughout the text, smaller separative features in the form of distinct punctuation and other minor variants
also demarcate the group. Unlike its redactional boundary decisions, few if any of these transmissional
variants constitute clear indicative errors, but the sharing of differing particles and slight differences of
wording (such as te instead of nas; sngon chad instead of sngon yang; bzhin gnas instead of las ni or du
gnas, in Chapter 11) help demonstrate a distinct family affiliation. More serious apparent lapses, such as the
omission of sems dpa' chen po nyon cig/ in Chapter 1, on reflection are most likely explained by a deliberate
although mistaken boundary decision, rather than by a transmissional error. Similarly (again in Chapter
1), all the Tshal pa Kanjur versions place the homage in a non-standard position, following their initial title
but before the regular title: this decision is most likely explained by TZComm's positioning of its homage
after the TZComm title but before its exposition of the root text title. The Tshal pa Kanjur texts also seem
most obviously to have made some attempt to edit their mantras, frequently giving spellings in accordance
with the "correct" standard transcriptions for the Sanskrit terms (see TZ edition, Chapters 9 and 14), where
other editions do not so consistently do so. They also conjecture back-translations of two Tibetan names into
Sanskrit in their peaceful mantra series, giving dhvaja for rgyal mtshan, and mui for khu tshur, where the
other editions give ketu and sandhi respectively. Two of the best known subgroupings of the Tshal pa
Kanjur are the Peking and 'Phying ba stag rtse sub-lines, which we represent here through the Peking and
'Jang sa tham/Lithang Kanjurs. Within TZ however, the Narthang and Lhasa Kanjurs might also form a subgrouping, of which the Ulan Bator manuscript seems to be part (unfortunately, we do not have images of the
complete Ulan Bator manuscript so we cannot be certain if this apparent affiliation continues for the whole
text). In Chapter 1, these three editions share an error of byas for bcas, as well as other very minor variants,
and in Chapter 6, they share an error of sku gsung thugs in place of sku gsum zhes. However, the level of
variation between all the Tshal pa Kanjur versions is very minor, and they can for present purposes best be
taken as multiple witnesses to a single edition of TZ.
The Four Bhutanese NGB texts as multiple witnesses to a single redaction
All four Bhutanese NGB manuscript collections so far made available to scholars belong to the distinctive
Bhutanese NGB recension in 46 volumes, with its particular selection of texts and doxographical
arrangement. In the NGB texts we have edited previously, the Bhutanese versions have tended to differ from
one another only very slightly in their readings, and this is once again proven to be the case with TZ. Just as
with the Tshal pa Kanjur texts, the Bhutanese NGB texts are unequivocally demarcated as a distinctive
grouping separate from all other versions of TZ by their entirely unique boundary decisions. As with the
Tshal pa Kanjur texts, these mistaken decisions can most plausibly be explained as a worthy but imperfect
attempt on a single occasion to extract the root text TZ from its commentary TZComm. There are in addition
several other separative readings in the Bhutanese branch of the transmission: The Bhutanese texts alone
give a Sanskrit title at the beginning of the text, surely a recensional variation. Their homage gives
Vairocana rather than Vajrasattva as object, but this is most likely a transmissional scribal error, picking up
from the text below, which mentions Vairocana using the precisely the same wording. Also in Chapter 1, the
Bhutanese texts alone share a dittography, of 'di'i don 'di dang phyi nang gi chos thams cad 'dra ste / bden pa
gnyis su 'dus te / bden pa gnyis la tshul 'di ltar rig par bya ste /. In the final chapter, only the Bhutanese
versions give a unique chapter title, ngo mtshar bstod pa'i le'u, 'on the wondrous praises'. This makes good
sense and seems most apt, but it is the only example of this title, which is perhaps a recensional amendment
or insertion.2 At many chapter endings, the Bhutanese texts alone give skabs, bstan pa'i skabs or skabs bstan
1

As we argue above, in many of the boundary variations, it is not possible to judge the correctness of the different versions on
the basis of the content alone. Here, we simply note that these distinctive decisions demonstrate the family groupings. Our
reasoning for concluding that the Tshal pa Kanjur and Bhutanese NGB versions of the root text did not predate the shorter
version is because they both form part of a sub-branch descended from hypearchetype c, as explained below.
The Dunhuang manuscript and the Tshal pa Kanjur version omit any chapter title, so it is quite likely (given the shared
indicative errors between these versions and the Bhutanese manuscripts, as described below) that the editors of the Bhutanese
tradition were faced with an omission they attempted to correct.

46

Textual Analysis

pa rather than le'u (and this wording occurs in no other edition anywhere in the text). The Bhutanese texts
are further separated from the other TZ versions by the numerous small transmissional variants they hold in
common. Yet the Bhutanese NGB manuscripts have very little variation between them, most of which is
very minor, such as one manuscript inserting an additional shad. It is clear that dGra med rtse and the
original text of sGang steng-a (before its corrections, see section a, p.29 above) have a number of minor
orthographical errors avoided by mTshams brag, sGang steng-b and by the corrections in sGang steng-a, but
these differences are very small and not of any great stemmatic significance.3 Hence it serves present
purposes best if we take the Bhutanese texts as multiple witnesses to a single edition of TZ.
The Three South Central Tibetan NGB texts as multiple witnesses to a single redaction
The four NGB manuscripts of gTing skyes, Rig 'dzin, Kathmandu and Nubri originate from
geographically neighbouring regions, and also share a doxographical organisation and a very largely (but not
quite completely) identical choice of tantras. In the NGB texts we have edited previously, the readings of
these four manuscript collections have generally been very close indeed, including numerous indicative
errors demonstrating descent from a common ancestor not shared by any other NGB editions.4 The three
witnesses of TZ available to us from the South Central Tibetan grouping demonstrate a similar pattern of
close affiliation, by only rarely or in trivial manner disagreeing with each other, and by sharing readings of
all kinds, including several indicative errors. Such indicative errors include in Chapter 1, the omission of
chen po, the omission of don, and mistakenly reading mya ngan zhes for me zhes bya; in Chapter 8, reading
ldan pa yis for sbyor ba yis; in Chapter 9, reading o badzra pad ma for o dharma badzra; in Chapter 10
dittographically inserting the spurious words gung 'dzub 'byung ba'i phyag rgya yin/, erroneously omitting
the phrase ngos bzhi nas ni rtse mo gzugs/, and reading rdo rje sems ma'i in place of rdo rje 'dzin ma'i; in
Chapter 11 omitting the phrase bcom ldan 'das da ltar dang, and through eyeskip reading only rdo rje for
rdo rje 'khor lo la/ /rdo rje; in Chapter 12 erroneously omitting the phrase de yi bka' nyan pho nya ni/; in
Chapter 21 erroneously omitting the phrase /mud ga 'i chan dang sbyar ba dag/; and in Chapter 41
erroneously omitting the word nang. These clear indicative errors are found in no other versions of TZ, thus
showing the three South Central Tibetan texts of TZ to descend from a common ancestor not shared with any
other versions. In addition, the three South Central Tibetan texts agree closely in a much greater number of
smaller transmissional variants, yet show rather few and only minor and non-indicative differences from one
another. For these reasons, it serves present purposes best if we consider the three South Central texts as
multiple witnesses to a single edition of TZ, rather than as three separate editions.
The Three Tenjur texts as multiple witnesses to a single redaction
A version of TZComm is contained within the three seventeenth century Tenjur editions of Peking, dGa'
ldan (or Golden) and sNar thang. It is immediately apparent that these three versions descend from a
common ancestor because, as we have mentioned above, they alone share exactly the same major losses of
text at several junctures, notably from the middle of Chapter Six until the last lines of Chapter Ten; and from
the middle of Chapter Thirteen until the end of Chapter Seventeen, amounting to over thirty percent of
TZComm, but nearly forty percent of its lemmata representing TZ. In addition, they share numerous other
indicative errors, such as the omission of ba dan dang and of gsum in Chapter 1, a reading of 'od kyi instead
of dkyil 'khor, of tshogs chen 'khor lo instead of 'khor lo tshogs chen, and of rang bzhin instead of rngam
zhing in Chapter 5, or in Chapter 21 erroneously reading, lcags dag dang ni rdo rje dag/, instead of lcags
phye dug dang rdo phye rnams/. They also share numerous smaller transmissional errors not found in any
other version, such as in Chapter 1, pas for pa, the omission of ces, and dpyod for spyod. The differences
between the three texts are by contrast rare and extremely small. For these reasons, it serves present purposes
3

These small errors could have been corrected through conjecture, so they are not in themselves a satisfactory basis to establish
the directions of copying.
In some cases but not others, texts of these four manuscript collections can be seen to break down further into the two very
slightly differentiated sub-branches of gTing skyes and Rig 'dzin on the one hand, and Nubri and Kathmandu on the other
hand; however, we cannot tell if the two sub-branches are represented in the case of TZ, because the Nubri witness is missing.

The Stemma of the Root Text

47

best if we consider the three Tenjur texts as multiple witnesses to a single edition of TZComm, rather than as
three separate editions.
If the internal relations within these four textual groupings have not yet been precisely discerned, their
testimonies as collectives are nevertheless valuable, offering close approximations of their respective
hypearchetypal ancestors. In this way, the number of versions we must account for in the stemma is reduced
from twenty-one to eight.
The descendants of hypearchetype b: a distinctive commentarially-derived branch of the TZ tradition
In section c above we introduced the differing boundaries of TZ in its different versions. In our analysis of
indicative errors below (see p.50-54), we will further demonstrate that TZ's Tshal pa Kanjur and Bhutanese
witnesses must descend from a common ancestor c, which, given its position on the stemma codicum, can
only have had a commentarial derivation. Since the Tenjur and Dunhuang witnesses are still presented in the
form of commentarial lemmata, we must conclude that no fewer than four of the eight TZ editions and fifteen
of the twenty-one TZ witnesses available to us today actually descend from a commentarial source rather
than from a root text. Hence one might reasonably want to see in more detail the evidence that stemmatic
analysis can expose of a major branch of the TZ tradition which is commentarially derived. We believe such
evidence is clearly visible, and that it shows the surviving witnesses of this major commentarial branch to all
descend from the single source that we designate as our putative hypearchetype b. This hypothesis is
supported by internal textual evidence, in the form of errors and other distinctive variants, shared by all the
commentarially derived witnesses of TZ (Tenjur, Dunhuang ms, Tshal pa Kanjur and Bhutanese NGB),
against all other versions of TZ (South Central Tibetan NGB, and the three local Kanjurs of Hemis, Tawang
O rgyan gling, and Bathang). Unfortunately, the irretrievable loss of almost forty percent of the lemmata in
the Tenjur edition of TZComm and a smaller proportion in the Dunhuang ms means that the surviving useful
evidence is significantly less than it otherwise might have been: for where the Tenjur readings are missing,
logic cannot decide if the shared errors of the remaining commentarially derived texts represent the
inheritance of b, or of its descendant in the shape of the further hypearchetype that we will discuss shortly, c.
Nevertheless, the evidence from where the Tenjur's readings have been preserved is still sufficient to confirm
a distinctive commentarially derived branch of the TZ tradition. Here are some of the main examples:5
dge dang mi dge: Ms rgyu dang 'bras bu; JQkNkUDkDHkVMGGr rgyu dang 'bras bu'i; QtGtNt dgu dang
'bras bu; He dge dang myi dge'i (Chapter 1)
Here, the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese texts share an incoherent and indicative
error. The words, rgyu 'bras occur on the next line, and have most probably been picked up in error.
There is no doubt that the South Central, Hemis, Bathang and Tawang version is the correct and coherent
reading, supported also by the commentarial explanation (see the TZ edition of Chapter 1).
sangs rgyas tshe 'di: MsJQkNkUDkDHkVQtGtNtMGGr tshe 'dis sangs rgyas; K sangs rgyas 'di; He sangs
rgyas 'di tshe; Ogl sangs rgyas tshe 'dis (Chapter 1)
The contrasting readings here demonstrate the separate affiliations, but neither reading is clearly
incorrect. The Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese version would seem preferable, yet this
need not imply that the South Central, Hemis, Bathang and Tawang texts share an error. It may be that
the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese reading was introduced in their shared ancestor
because it represented a minor stylistic improvement.

Note that we have collated all available versions of the text for the examples listed in this chapter, apart from the Tshal pa
Kanjur texts, where we have systematically collected three witnesses, J, Qk and D (or Dk where D is missing), and more in
some cases, such as the examples from Chapter 1. Where they agree, the variants are most likely to occur also in the other
witnesses of the grouping. Note also that we draw attention to the specific reading under discussion by underlining the
relevant letters or words.

48

Textual Analysis

rigs sngags: MsJQkDQtGtNtMGGr omit; TRK rigs (Chapter 11)

In this case, rigs sngags is not indispensable for the coherence of the passage, but it would seem
preferable.
'byin: MsJQkDGtGGr 'dren (Chapter 11)
khrag skyug: Ms khrag du skyugs; Qk khrag skyugs; MG khrag tu skyug; QtGtNtGr khrag tu skyugs;
khrag skyug: Ms khrag du skyug; JDQtGtNtMGGr khrag tu skyug; Qk khrag tu kyug (Chapter 11)
This minor variant, the addition of tu/du, occurs twice in Chapter 11. It is not a serious error, although
the text is clearer without it. On the first occasion, the Tshal pa Kanjur texts do not give tu, perhaps
having emended the version they inherited.
sing ha mu kha dang/ bya gri mu kha sri ga dang/ shwa na: Ms bhya gra mu kha dang/ sing ha mu kha dag
dang [ldIn?] / /snri kha; JD bya gh mu kh dang/ singha; Qk bya gri mu kh dang/ sing ha; QtGtNt bya
tri mu ka dang/ sing ha mu ka dang/ sri la mu ka; MG bya gri mu ka dang/ sing ha; Gr bya khri mu ka
dang/ singha; TRK sing ha mu kha dang/ bya khri mu kha sri ka dang/ /sho na; Bth sing ha mu kha dang/
bya gri mu gi sra ga dang : shwa na; Ogl seng ha mu kha dang/ bya gri mu ka sri ga dang/ shwa na
(Chapter 11).
Here, the points to note are the contrasting order in this list of goddesses, and the omission of shwa na in
the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese grouping. Of course, there are variations in such
lists, but the South Central and local Kanjur version (the Hemis manuscript is missing at this point),
giving an order of Sihamukh, Vyghrmukh, glamukh, and vnamukh, is entirely plausible as
the originally intended list and, more importantly, is consistent with the discussion given below in the
same chapter of TZComm. It is also the ordering found in some other influential rNying ma traditions, for
example, it corresponds to the order of mantras for these goddesses given in Chapter 16 of the root rGyud
gSang ba'i snying po (gsang ba'i snying po de kha na nyid nges pa; in the mTshams brag NGB edition,
Volume Wa: 207), and also to the list in Karma Gling pa's Bar do thos grol, where the four goddesses are
placed at the four cardinal directions, just as in TZComm [Dorje 2005: 368-369; 397-399].
ha ha he he: MsJQkDMGGr ha ha de nas; QtGtNt ta ha de nas; TRK ha ha he he/; Bth he he (Chapter 11)
Either variant could fit, but clearly he he and de nas represent contrasting readings.
nas: MsJQkDGtGGr lhar (Chapter 11)
kho'i gzugs: Ms khro bo 'i gzugs; JDQtGtNtMGGr khro bo'i gzugs; Qk bro ba'i gzugs (Chapter 11)
The shared reading of the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese group, the wrathful one's
form, could fit, although in this context of drawing the form or effigy of the evil being or spirit, the South
Central and local Kanjur version of his effigy/body would seem more appropriate.
tsa kra bru h: Ms tsag kra [h?] [it appears that bh was possibly corrected to h]; JQkDMGGr
tsakra h; QtGtNt tsag kra bru (Chapter 11)
The South Central and local Kanjurs (Hemis is still missing at this point) give two final syllables. The
Tenjur has lost one of them while the Tshal pa Kanjur and Bhutanese texts has lost the other. The
Dunhuang manuscript may have inherited a reading of bhm, which it corrected to h.
de nas: MsJQkDQtGtNtMGGr de ltar (Chapter 11)
This is a minor variant, but it illustrates the affiliations.
ha ha zhes sang sang phyung: MsQtGtNt ha ha sang sang sgra phyung; JQkD ha ha sa sa sgra
phyung; TRK ha ha ha zhes sang sang phyung; Bth ha ha ha zhes sang phyung (Chapter 11)

The Stemma of the Root Text

49

The loss of zhes and the addition of sgra are further minor variants, but again, they illustrate the groups.
In this case, the Bhutanese edition is not represented, since it omits a number of lines here, incorrectly
assuming them to be commentary.
de yi bka' nyan pho nya ni/: Ms de 'i pho nya bka' nyan ni/; JDMGGr de yi pho nya bka' nyan ni/; Qk de
yi pho nya bka' brnyan ni/; QtGtNt de'i pho nya bka' nyan ni/; TRK omit; BthHeOgl de'i bka' nyan pho
nya ni/ (Chapter 12)
Here, the South Central version omits the line, but the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese
texts agree on a reading which has the words in reverse order from that in three local Kanjurs. Both
versions make good sense, but there is a clear contrast between the groups.
gsol bar bya: MsJQkDQtGtNtMG dbul bar bya; Gr 'bul bar bya; TRK gsol bar byos (Chapter 17)
Either variant fits well; again, the difference simply demonstrates the groupings.
/phra men ma rnams mngags nas ni/ /bkug nas bsregs na mod la 'chi/: Ms /'phra men mngags te bkug nas
ni / /thab du bsregs na mod la 'chi /; JQkD /'phra [D phra] men mngags [Qk mdags] te bkug nas ni/ /thab
kun bsregs na mod la 'chi/; QtGtNt /'phra men mngags te bkug [Gt bkugs] nas ni/ /thab khung bsregs na
mod [Gt mong] la 'chi/; MGGr /'phra men sngags te bkug nas ni/ /thab [Gr thabs] su bsregs na mod la
'chi/; TRK /'phra men ma rnams mngags nas ni/ /bkug nas bsregs na mod la 'chi/; Bth /phra men ma
rnams mang gi ni/ /bkugs nas bsreg na mchod la mchi/; He /phra myen ma rnams mngags nas ni/ /bkug
nas bsregs na mod la 'chi/ (Chapter 19)
Again, we have two distinctive versions, corresponding to the South Central and local Kanjurs versus the
Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese group. Both versions fit appropriately, although the
addition of thab du/su in the Dunhuang and Bhutanese version would seem helpful.
kun la mnga': Ms yul la dbang; JQkDQtGtNt yud la dbang; He kun la mnga' (He inserts the mnga' as a
correction in small writing); Bth kun la (Chapter 24)
At this point, the Bhutanese manuscripts omit the verse, presumably, since their ancestor understood it as
part of the commentary and not the root text. However, even though the Bhutanese witness is absent, it is
otherwise clear that we have a contrast between branch b and the other editions. Again, either reading
could fit. There is not an exactly parallel expression in the parallel verses in other chapters (Chapters 17,
29 and 34), although yud la or yud tsam occurs in those verses, so perhaps the Tenjur and Tshal pa Kanjur
reading (which the Dunhuang manuscript most probably corrupted) is more clearly in line with the
parallel verses.
mig dang: Ms de myig; JQkD de yi mig; QtGtNt de'i mig; Gr de'i mig dang; MG de yi mig dang; He
myig dang (Chapter 24)
This is a minor variant of little significance to the meaning, but the difference between the two versions is
clear.
gtor bas: MsJQkDQtGtNtMGGr mthong bas; Bth gtor ma (Chapter 27)
In this case, the South Central and local Kanjur version would seem more appropriate, and more in line
with the parallel chapters on the other ritual activities, in which gtor bas and gtong bas are found. It is
possible that mthong bas was an error for gtong bas.
las bstan pa'i: JQkDQtGtNtMGGr omit; He las kyi (Chapter 31)
This chapter is omitted completely in the Dunhuang manuscript, but the agreement between the group is
otherwise clear. These words are not necessary, and either version is equally acceptable in this context.
The Tenjur, Tshal pa Kanjur and Bhutanese version is in fact more closely in line with two of the parallel
chapters on the other ritual activities (Chapters 20 and 36), while the South Central and local Kanjur
version would seem more closely in line with the other parallel chapter (Chapter 26).

50

Textual Analysis

gtor ma bshad pa'i: JQkDQtGtNtMGGr gtor ma'i (Chapter 32)

Again, we are missing the Dunhuang witness for this chapter, but otherwise, the affiliated texts all omit
bshad pa, where the text reads perfectly well either way. In this case, the Tenjur, Tshal pa Kanjur and
Bhutanese version is in line with the parallel chapters on the other ritual activities (Chapters 21, 27 and
37).
gyur: MsJQkDQtGtNt grub; TRK 'gyur; He gyurd (Chapter 34)
As with the Chapter 24 variant listed above, we are missing the Bhutanese witness for this verse, since
their ancestor seems to have taken it as part of the commentary and not the root text. As in the case of the
Chapter 24 variant, there is not exactly parallel wording in the parallel verses in other chapters (Chapters
17, 24 and 29), although the South Central and local Kanjur variant would seem more similar to Chapters
17 and 24 (neither variant resembles the wording of Chapter 29).
dngul phye: MsJQkD dngul dang; QtGtNtMGGr dngul lam; T rdul phye; RKBthHeOgl rngul phye
(Chapter 37)
Here, dngul phye is the correct reading (the archetype almost certainly gave, rngul phye, rngul a common
spelling for dngul in old sources), while the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese versions
appear to be corruptions, with the Dunhuang and Tshal pa Kanjur reading possibly generated in
independent attempts to correct the obviously inappropriate Tenjur/Bhutanese reading.
gcig la gcig phyi nang du gzigs pa/: MsJQkNkDQtGtNt phyi nang du gcig la gcig gzigs pa; TRK gcig la/
gcig phyi nang du gzigs pa/; Ogl gcig la gcig phyir nang du gzigs pa/ (Chapter 41)
Here, we are missing a Bhutanese reading, since its version includes only the opening lines of Chapter 41.
Nonetheless, here again we clearly see a contrast between the two versions, although either is perfectly
acceptable in this context.
mthar phyin pa'i le'u: MsJQkNkD omit; QtGtNt rtog pa'i rgyal po don bsdus pa'i le'u; MGGr ngo mtshar
bstod pa'i le'u (Chapter 42)
Here, the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese editions do not share one reading, but all
lose the chapter title witnessed in the South Central and local Kanjur editions. The Tenjur seems to have
a scribal confusion with the main text title, while the Bhutanese line has a recensional variant on the title,
perhaps introduced on the basis of the chapter's actual subject matter.
le'u khri drug stong ba las/: MsJQkNkDQtGtNtMGGr omit; Bth le'u khri stong ba las; He le'u khri drug
stong las//; Ogl le'u khri drug stong pa las/ (Chapter 42)
These words are added into the text title in the concluding colophon. They do not occur as part of the title
anywhere else in the text, and it is possible that they are a shared error of the South Central and local
Kanjur group. However, if an ancestor of the South Central and local Kanjur group introduced this
addition, it must have been the result of a recensional editorial decision. This would seem rather less
likely than the hypothesis that the words existed in the archetype, and were lost from the Tenjur,
Dunhuang, Tshal pa Kanjur and Bhutanese line, either as a result of an editorial decision that they are
inconsistent with the text title elsewhere in the text, or as a result of scribal error, a scribe automatically
reverting to the regular title. It is quite possible that they were within the archetype of the text, alluding to
the ideal "original" version of the tantra in a sixteen thousand chapter text, from which the current text
was claimed to be extracted.
A distinctive sub-branch within the commentarial line, descending from the putative hypearchetype c
Exceptionally clear evidence exists for a further sub-grouping of texts within the commentarial line
descending from b. This sub-grouping comprises the Dunhuang manuscript together with the Tshal pa
Kanjur and Bhutanese NGB editions. Although, as we have mentioned above, each of the three members of
this grouping have their own highly distinctive single readings, they also share with one another a number of

The Stemma of the Root Text

51

clear indicative errors that can only be explained by a common ancestor that is not shared with any other of
our versions of TZ. All three can therefore be shown to descend in common from an error-carrying ancestor
that we identify as hypearchetype c. Unfortunately, some indicative errors occur in parts of the text where the
other extant member of the commentarial line, the Tenjur, has been lost, so that their testimony is
inconclusive: in such cases where the Tenjur is missing, we cannot know if the shared errors represent the
heritage of c or of b. However, in other cases the Tenjur text is available, and there we can see that it did not
reproduce some of these errors, so that the hypearchetype c can be clearly established.
The most significant such indicative error is a long passage of text which the Dunhuang, Tshal pa and
Bhutanese versions omit at the end of Chapter 10 (see Chapter 10 of TZ edition below). The Tenjur also
omits most of it, but only for the entirely unrelated reason that its falls within the Tenjur's major loss of text
from Chapter 6 until the final section of Chapter 10. When the Tenjur picks up the text near the end of
Chapter 10, it gives a few lines from the passage in question, showing that it did not share the ancestor of the
Dunhuang, Tshal pa Kanjur and Bhutanese editions which lost the entire passage.6

Chapter 10 is on the mudrs of the deities of the peaceful maala, who are listed in full in TZComm's Chapter 7, while their
mantras have just been given in a clear order in Chapter 9. The lost passage in the final section of Chapter 10 continues
exactly in sequence following on from the mudrs of the deities given in the chapter so far. It first gives the mudrs for the
male peaceful deities, although it appears to omit the four males in Vairocana's circle. Where we would expect them to occur,
the mudrs for the female deities commence, at first in no obvious logical sequence, which might suggest that we might have
some textual corruption at this point. On the other hand, the original tantra may not have had such a orderly schema as that
outlined in TZComm's Chapter 7, which in any case does not entirely correspond clearly either to the mantra or to the mudr
list (see Appendix). However, following what seem most likely to represent the mudrs of most of the principal female
deities, the text begins with the females of the retinues in a clear logical order corresponding to their layout in the maala, as
described in TZComm's Chapter 7, and given in precisely the same order as the sequence of mantras in TZ's Chapter 9. The
chapter stops abruptly in the versions which share the loss of the passage, with the second female member of Ratnasambhava's
retinue (rdo rje 'phreng ba), omitting the others of Ratnasambhava's group, as well as those of Amitbha's, Amoghasiddhi's
and Vairocana's groups. The closing passage is found in full in the South Central Tibetan version, and also in the Tawang and
Bathang local Kanjurs; unfortunately, we are missing the Hemis folios at this point, although the missing folios are of
approximately the right length to suggest that they included the passage. The passage begins at the correct place, with the
third female member of Ratnasambhava's retinue (rdo rje me tog ma), and continues as we would expect, apart from a puzzle
in the final lines of the text. There is one just conceivable but extremely unlikely alternative explanation to the hypothesis
that the South Central Tibetan, Tawang and Bathang versions preserve a passage once witnessed in the earlier archetype of all
the current versions. In this scenario, the editors of an ancestor of the South Central Tibetan, Bathang, Tawang and Tenjur
versions might have inherited the already corrupted version, noticed the omission of a large number of mudrs, and attempted
to restore the text by writing the extra lines. It would seem safe to discount this unlikely possibility, especially given the
distances separating these editions, as well as other evidence suggesting, for example, that the Bathang Kanjur has texts which
represent a tradition of extremely early readings (see above, p.27-28). Furthermore, Tibetan editors of such scriptural
collections seem rarely to re-write or add large sections of text, even where they find significant corruptions. There are other
incoherencies in the TZ text including the apparent loss of order in middle of the Chapter 10 mudrs mentioned above
which none of our editions has attempted to resolve. Moreover, if the final passage were a deliberate construction, one would
have expected it to supply all the missing mudrs, taking care that they correspond to the correct deities' names. However, the
final few appear to muddle Vairocana's male and female retinue, giving a single set comprising the male names with female
particles, and at the end, we additionally seem to be missing one of the principal female deities. If Vairocana's male retinue
are intended for the final names, it would surely make more sense for an editor involved in substantial re-writing, to insert
them in their logical place earlier in the chapter. Perhaps one slightly more conceivable possibility is that rather than a rewriting job, the editors found another tantra with the same set of deities and inserted the missing list from its list of mudrs,
thereby explaining the slightly odd presentation of the mudrs for the final deities. This hypothesis would gain support if such
a text were identified, especially if it has female versions of the list of Vairocana's male retinue as found in TZ. However,
unless such a text were to be identified, it would seem virtually certain that the South Central Tibetan, Tawang and Bathang
texts have preserved text which genuinely goes back to an earlier version before the loss of the passage in the other editions.
When added to the further evidence of shared errors in the Dunhuang, Tshal pa Kanjur and Bhutanese editions (whether or not
all those errors were shared with the Tenjur), we can feel confident that the most straightforward explanation for the loss of the
passage a scribal corruption within the Dunhuang, Tshal pa Kanjur and Bhutanese branch is the correct one.

52

Textual Analysis

Other indicative errors reinforce the picture of the Dunhuang, Tshal pa Kanjur and Bhutanese texts
forming a sub-branch that descends from hypearchetype c:
Further variants shared by the Dunhuang, Tshal pa Kanjur and Bhutanese texts against the Tenjur and
other editions:
MsJQkNkDkDHkVMGGr insert dpal (Chapter 1)
It is logically possible that the three versions might independently have introduced this adjective, but on
balance, more probable that the amendment was made on a single occasion, and inherited by all three
versions.
o badzra sma sha ni h dza/: Ms ba dzra kar ma h dz/; JD o badzra karma h dza; Qk o
badzra karma h dza; M o badzra karm h dza/; GGr o badzra karma h dza/; QtGtNt o
badzra sma sha ni h dza dza/; T o badzra smra sha ni h dza/; Bth o badzra [ba?] sha ni h
dza/; Ogl o badzra ba sha ni h dza/ (Chapter 11)
Here, the reading, karma, shared only by the Dunhuang, Tshal pa Kanjur and Bhutanese texts, seems to
be in error. The appropriate goddess here is man (see TZ edition Chapter 11).
'khu ba: MsJQkDMG 'khu byed; Gr 'khu 'byed (Chapter 11)
This is a minor variant and has no impact on the meaning, but it is more likely to have been generated
once and inherited by all three versions.
/'phags pa thabs kyi zhags pa pad ma 'phreng las/ dbang chen bsdus pa'i dkyil 'khor gyi le'u ste bzhi bcu
gcig pa'o// (Chapter 41)
Here, only the Dunhuang, Tshal pa Kanjur and Bhutanese texts lose this chapter title, and thus also the
chapter break between the final two chapters.
rdo rje 'du 'phro mdzad: MsQkNkJQkDMGGr rdo rje 'du 'phro khro (Chapter 42)
The reading, khro, is a rather distinctive error, and not very likely to have been generated independently,
although khro does occur two lines above, and has presumably been picked up from there.
A further example is suggestive of an error most probably found in the common ancestor of the group c,
but unfortunately at this point the Bhutanese NGB manuscripts fail to include a number of lines, assuming
them to be commentary rather than root text, so that we lack their testimony:
e ma'o chos e ma'o chos/: QtGtNtHe e ma'o chos/; MsJQkD omit; Ogl e ma'o chos/ e ma'o chos (Chapter
11)
This is an expression of jubilation which opens a verse citation from the Guhyasamja Tantra's Chapter
5. The first line of the verse then repeats the phrase, e ma'o chos. The omission shared by the Dunhuang
and Tshal pa Kanjur texts may have been due to an eyeskip to the first line of the verse, or might have
been a deliberate editorial decision, assuming the repetition to be redundant and dittographical. It is also
possible that c might have inherited a reading that had already lost one e ma'o chos (as in the Tenjur and
the Hemis version), thus giving the words the appearance of an unmetrical and unnecessary addition (see
TZ edition Chapter 11).
Variants shared by the Dunhuang, Tshal pa Kanjur and Bhutanese texts where the Tenjur reading is
missing:
As mentioned above, there are further examples of separative errors and other variants within the
commentarially derived lineages where, unfortunately, we are missing the Tenjur version of the text, so that
we cannot be sure whether these variants illustrate simply the legacy of c as preserved in the Dunhuang,
Tshal pa Kanjur and Bhutanese NGB versions, or the broader legacy of b, including also the Tenjur.
Nonetheless, it is worth listing them as further illustrations of the distinctiveness of this branch of the
tradition:

The Stemma of the Root Text

53

gtso bo lnga dang gtso mo lnga'i byin rlabs las/ /sems dpa' bcu drug dang/ /sems ma bcu drug gi rtogs pa:
MsJQkNkDMGGr omit (Chapter 7)
These words form part of the chapter title, which does still remain perfectly coherent without them.
Nevertheless, the omission is quite likely to have been a shared error, especially since the opening of the
commentary to Chapter 7 appears to reiterate these details of the chapter's topic. The reading at the very
least demonstrates the separate descent of this branch of the transmission.
bsgrubs: MsJQkNkDMG bsgrags; Gr sgrags; TBth bsgrub (Chapter 8)
In this case, the Dunhuang, Tshal pa Kanjur and Bhutanese reading is more appropriate in a context of the
consecrations of speech. However, it is possible that the earlier reading was in fact bsgrubs/bsgrub, and
emended in hypearchetype b or c.
o sti ra badzra o/ o dzwa la badzri o/ o sang ha ra badzra o/ o tsad tsa la badzra o/
MsJQkDMGGr omit (Chapter 9)
These four mantras represent the four females of Vairocana's entourage, appropriately positioned in the
sequence of mantras for the peaceful deities, following the mantras for the five principal females of each
buddha family, and immediately preceding the females of Akobhya's group (see Appendix). These
mantras appear to have become entirely lost either within b, or within c the loss of the Tenjur testimony
means that we can never know which. It appears the annotator of the Dunhuang manuscript, who
endeavoured to append deities' names conveniently beneath their mantras, became aware that the mantras
for Vairocana's females were lacking. Thus, in the absence of their proper mantras, he arbitrarily and
inappropriately appended their names beneath the four quite unrelated mantras at the chapter's end.
o badzra su ki tra: Ms ba dzra rad na ra ti tra; JD o badzra ratna ra ti tr; Qk o badzra radna
ra ti tra; M o badzra ratna ra ti tra; GGr o badzra ratna ra ti tra; Bth omits here, but gives o
badzra su ki dri in a less appropriate position above. (Chapter 9)
Here, the female deity whose mantra is expected is rDo rje bde ba ma, Vajra Bliss, and hence the South
Central and local Kanjur reading presumably represents the Sanskrit word, sukh. It may be that this
mantra was lost in an ancestor of the Dunhuang, Tshal pa Kanjur and Bhutanese NGB texts (either b or
c), and that the mantra element ratna rati was then introduced in an effort to make good the loss.
sgom: MsJQkDkMGGr sgrol; Ogl bsgom (Chapter 10)
Although sgrol is not incoherent at this point, sgom is clearer, and sgrol most likely picked up in error
from dgrol at the end of the line.
mgo 'gas: Ms 'gro 'o; JQkDMG 'gro ma; TRK mgo mags (Chapter 13)
The South Central reading is clearly a corruption of the local Kanjurs' mgo 'gas, which is confirmed in
TZComm's explanation of the verse, while 'gro would appear to be a shared error stemming from either
hypearchetype b or c.
grdza grdza h pha/: Ms ga rdza ga rdza ga rdza h phat /; JQkDMGGr gardza gardza gardza h
pha/; TRK gardza gardza h/; Bth ka dza h phat/; Ogl g rdza h pha/ (Chapter 14)
Here, it is uncertain whether the archetype had a double, single, or triple grdza/gardza. The South
Central, Hemis, Bathang and Tawang versions are divided between favouring a single or double, but the
Dunhuang, Tshal pa Kanjur and Bhutanese texts all favour a triple, which would be more consistent with
the majority of the following mantras in the sequence. It is impossible to be certain whether the mantra
was shortened or lengthened, but it is more likely that the archetype had a double or single mantra which
was lengthened in an ancestor of the Dunhuang, Tshal pa Kanjur and Bhutanese texts, either b or c, to
make it consistent with the mantras below.
dpa' bos: Ms rgyal bas; JDMGGr rgyal pos; Qk rgyal po'i (Chapter 16)

54

Textual Analysis

Again, the difference demonstrates the separate group affiliations; the South Central, Hemis, Bathang and
Tawang texts are here all agreed on the reading, dpa' bos. Either reading could fit equally well.
yang dag: MsJQkDMGGr yang du; Bth yang drag (Chapter 16)
In this case, yang du would seem to be a shared error, inherited either from b or from c.
dbus su: MsJQkDMGGr dkyil du; Bth dbusu (Chapter 17)
Either variant could work, and has essentially the same meaning, but once again, this example illustrates
the separate descent of this branch.
dngul lam dkar po'i shing rnams la/: MsJDMGGr phur pa kha dog dkar po la/; Qk unclear; QtGtNt omit
this tshig rkang; Bth rngul la dkar po'i shing rnas la/; He rngul lam kar po'i shing rnams la/ (Chapter
36)
Here, the Dunhuang, Tshal pa Kanjur and Bhutanese texts agree on a quite different reading. This is not
as appropriate in this context as the South Central, Hemis, Bathang and Tawang version, which is
consistent with the parallel chapters on the phur pas for the other three rituals, and also with the
commentary to Chapter 36.
The question of further affiliations within the descendants of c: is there an hypearchetype d?
The existence of a line of descent from hypearchetype c is clear, but what is less clear-cut is whether or
not we can reliably discern any further sub-lineages separating its heirs. For purposes of stemmatic
reconstruction, this is not a vital issue, since it has little impact on our understanding of the readings of b or
a. Nevertheless, it is worth examining the evidence, to achieve a clearer overall picture of the TZ tradition.
Especially near the start of Chapter 1, the Tshal pa Kanjur and Bhutanese NGB traditions do seem to share
separative errors, readings which appear on balance far more likely to represent a uniquely shared
inheritance than to have been independently or coincidentally generated. This would suggest that the
extraction of lemmata to create the Tshal pa and Bhutanese versions of TZ were made from very closely
related or even the same version of TZComm, which in turn implies the possible existence of an
hypearchetype d. Were we to have included it on our stemma diagram, this hypearchetype d would resemble
the Dunhuang ms in descending from c, and would in turn be the shared ancestor of the Tshal pa Kanjur and
Bhutanese NGB editions. Yet although the relevant shared variants do seem clearly indicative near the start
of Chapter 1, elsewhere in the text they on the whole seem less clearly so. For that reason, and bearing in
mind the probably great mass of invisible evidence from lost editions which will never be available to us, we
do not yet see the existing evidence as sufficiently decisive unequivocally to represent d as a separate
hypearchetype on our stemma diagram. After all, there is small chance the shared variants near the start of
Chapter 1 could be the result of contamination, while a great many of the other shared variants in the other
chapters are of the types that could simply be coincidental. Hence we conceive of d as a merely possible
hypearchetype that might acount for the shared readings between the Tshal pa Kanjur and Bhutanese NGB
edition, but not as an indubitable conclusion. Here is a selection of the readings in question:
Readings shared by the Tshal pa Kanjur and Bhutanese groups
bla re dang/ gdugs dang / rgyal mtshan dang/ ba dan dang/: JQkNkUDkDHkVMGGr give the order (with
minor variants): bla bre dang/ rgyal mtshan dang/ ba dan dang/ gdugs dang / (Chapter 1)
Here perhaps the Tshal pa Kanjur and Bhutanese reading could be said to conform to a slightly more
logical sequence the rgyal mtshan would often be mentioned first of the adornments but it is not very
likely that they would independently have re-ordered the list.
phyi nang gi mchod pa brgyad kyis mchod de/: JQkNkUDkDHkVMGGr nang gi mchod pa bzhi dang /
bdug spos la sogs pa'i phyi'i (MGGr omit phyi'i) mchod pa bzhis mchod de/ (Chapter 1)
Here, either reading could make good sense; the Tshal pa Kanjur and Bhutanese texts simply give a more
elaborate version. This shared variant perhaps originated with the intrusion of a marginal note into the

The Stemma of the Root Text

55

main text of a common ancestor. A marginal note of broadly similar import can be found in the Dunhuang
manuscript (see our edition of TZComm).
'dun: JQkNkUDkDHkVMGGr 'dud (Chapter 1)
Again, either reading could make good sense; the Tshal pa Kanjur and Bhutanese is quite appropriate, but
was perhaps the result of a psychological scribal error, linking the word to the phyag 'tshal following.
gsol ba btab pa: JQkNkUDkDHkVMGGr gsol ba (Chapter 1)
bskur bas: JD bsgyur na; Qk sgyur na; Nk bsgyur nas; MGGr sgyur bas (Chapter 6)
bya gri mu kha sri ga dang/ shwa na: Ms sing ha mu kha dag dang [ldIn?] / /snri kha; JDMGr singha;
QkG sing ha; QtGtNt sing ha mu ka dang/ sri la mu ka; TRK bya khri mu kha sri ka dang/ /sho na; Bth
bya gri mu gi sra ga dang : shwa na; Ogl bya gri mu ka sri ga dang/ shwa na (Chapter 11)
Here, the Tshal pa Kanjur and Bhutanese texts share the omission of one of the goddesses (sri ga =
glamukh), as well as sharing a reading with the Tenjur and Dunhuang manuscript, which was
discussed above.
las ni: JQkD bzhin gnas; MGGr du gnas (Chapter 11)
Here, las ni might possibly have been amended independently in the Tshal pa Kanjur and Bhutanese
texts, but the sharing of gnas suggests that their readings most likely have stemmed from a common
ancestor.
g.yas pa'i the bo yis: Ms g.yas kyI the bo yis; JQkD g.yas g.yon mthe bo yis (but Qk unclear); MGGr
g.yon pa'i mthe bong gis; Bth g.yas pa'i tho ba yis (Chapter 16)
Here, g.yas was most probably in the archetype; the variants can also make good sense. The Tshal pa
Kanjur and Bhutanese texts share g.yon, but the Tshal pa Kanjur also incudes g.yas. Possibly, a common
ancestor introduced g.yon and then the Bhutanese tradition dropped g.yas. There is no Tenjur for this
section.
dkyil du gnan: JQkD mthil du gnon; MG mthil du gnan; Gr mthil du mnan; TRK dkyil du mnan; Bth
dkyil tu gcan (Chapter 16)
The amendment to mthil makes little difference in meaning in this context, but dkyil and mthil are quite
different words, and it is not very likely that mthil would have been introduced separately in the two
groups. Again, there is no Tenjur for this section.
la sogs: MsHe la stsogs; JQkDMGGr sogs pas (Chapter 35)
sbyor ba'i: Ms 'byor pa'i; J spyod pa las kyi; QkD sbyong ba las kyi; MG gtor ma las kyi; Gr gtor ma'i las
kyi; Bth omits (Chapter 38)
The shared element here is the addition of las kyi. Given the context, this could have been independently
added, so it is not in itself conclusive evidence, but it is perhaps more likely that the amendment was
made once.
There are also a number of small variants in common between the Tshal pa Kanjur and Bhutanese texts,
such as agreements in verb forms, particles and punctuation, which add to the weight of evidence for the
Tshal pa Kanjur and Bhutanese traditions to descend from one ancestor not shared with the other branches of
the transmission. If so, it is most likely that they both descend from a closely related or even the same
version of TZComm, but quite independently made their choices of lemmata, since their versions of TZ
display quite different selections of commentarial text which have mistakenly been included, or of root text
which have mistakenly been omitted. However, many of their small shared variants could in fact represent
minor corrections of grammar or modernised spellings, which could have been independently generated.

56

Textual Analysis

Alternatively, in some cases, they might represent hypearchetype c, where its readings have been amended in
the Dunhuang manuscript.
Perhaps to complicate further the hypothesis of a shared ancestor d of the Tshal pa Kanjur and Bhutanese
NGB, there is also some evidence which on the contrary might seem to suggest alternative affiliations
between one of these groups and the Dunhuang manuscript. None of these shared readings are indicative
(we list examples below), but they add to an uncertain overall picture in which it is difficult to reach a firm
conclusion on relations within the sub-branch.
Readings shared by the Dunhuang and Tshal pa Kanjur texts
There are a number of minor variant readings shared by the Dunhuang manuscript and the Tshal pa
Kanjur, errors or amendments which do not appear in the Bhutanese texts:
thob par 'gyur ba'i: Ms thob pa 'di 'i; JQkNkUDkDHkV thob pa'i; TK thob par 'gyur ba; R: thob par 'gyur
bar (Chapter 1)
Here, there is a shared omission of par 'gyur. This could have been dropped independently, however, and
the sentence still makes sense without it.
o badzra kar ma ha: Ms kar ma ba dzra ha; JD o karma badzra ; Qk o karma badzra a; MGGr
o badzra karma a; TRK o badzra ka ra ma ha (Chapter 9)
This example relates to the order: the Dunhuang manuscript and the Tshal pa Kanjur give karma badzra
rather than badzra karma. The transposition was quite likely shared by the group including the
Bhutanese transmission (there is no Tenjur text at this stage), but the ancestor of the Bhutanese texts may
have corrected it in this case, perhaps even unreflectively, since badzra generally comes first in mantras,
and does so in most of the mantras listed in the chapter. Moreover, in the series of five mantras of which
this is the fifth, there is agreement across all editions on the word order only in the first mantra
(Vairocana's mantra), where badzra comes first. The next four mantras are for the four male deities of
Vairocana's retinue, and the c tradition gives badzra second for them all, apart from the Bhutanese texts
in the case of this final mantra of the group. It would seem more likely then, that the Bhutanese would
have inherited the reading of c and emended it. Even if the inherited group reading had put badzra first, it
is quite possible that the Dunhuang and Tshal pa Kanjur texts might have independently emended this
reading to conform to the three mantras for the other members of the retinue in their version of Chapter 9.
It would therefore be rash to draw the conclusion that this case indicates that the Dunhuang and Tshal pa
Kanjur texts might have a shared ancestor which the Bhutanese texts do not.
ni: MsJQkD la; QtGtNt kyang (Chapter 11)
Here, the Tenjur gives a different reading. Again, c may have introduced, la, which the ancestor of the
Bhutanese texts may have emended, even unreflectively.
mgon po dpa' bos yang dag bshad: Ms dpa' bo chen pos yang du gsungs; JQkD mgon po dpa' bos yang du
bshad; MGGr mgon po dpa' bos yang dag bshad; TRK mgon po dpa'os yang dag bshad; Bth mgon po
dbang pos yang bshad (Chapter 16)
Here, the Bhutanese texts correctly give yang dag and avoid the error of the Dunhuang manuscript and
the Tshal pa Kanjur (there is no Tenjur text at this stage). Yet in the verse above, the Dunhuang
manuscript, the Tshal pa Kanjur and the Bhutanese texts all give yang du, so perhaps their common
ancestor made the error, but in this instance, the Bhutanese line corrected it, possibly even unreflectively.
bsregs: MsJQkD bsres; MG bsreg; Gr sreg (Chapter 19)
Both readings here could make sense, and bsres if anything seems a little more appropriate. However,
bsregs was clearly in the archetype (the Tenjur gives it too), perhaps even in error: it could easily have
been picked up given the context and the word, bsregs, in the line above. It may be that c introduced
bsres, which the Bhutanese line emended to bsreg, possibly even unreflectively, given the context.

The Stemma of the Root Text

57

dmar po'i 'bras chan: Ms dmar po 'i 'bras can; JQkD dmar po'i 'bras can; M 'bras chen dmar po; GGr 'bras
chan dmar po; TRK dmar po'i sran ma; Bth dmar po 'bras chen; Ogl dmar po'i 'bras 'chan (Chapter 27)
Here, the Tshal pa Kanjur and Dunhuang manuscript's reading of 'bras can is clearly mistaken and could
easily have been emended in the Bhutanese line.
Readings shared by the Dunhuang and Bhutanese NGB texts
There are also a few shared readings or errors of the Dunhuang manuscript and the Bhutanese NGB texts,
which seem to be avoided by the Tshal pa Kanjur:
sbyar: MsMGGr gzhar (Chapter 10)
Either reading would fit here (there is no Tenjur text at this stage). Independent emendation to gzhar
would seem unlikely, although earlier in the chapter, the Bhutanese texts alone give gzhar where other
editions all give sbyar. It is possible that the Tshal pa Kanjur inherited the reading, gzhar, but emended
it, even unreflectively, to the more frequently used word, sbyar.
phyag rgya yin par 'dren pas bshad: Ms phyag rgya yin zhes dpa' bos gsungs; MGGr phyag rgya yin par
dpa' bos bshad; Bth phyag rgya yin bar 'dren bar bshad (Chapter 16)
dPa' bos and dren pas can both fit appropriately, the mudrs explained by the warrior bodhisattvas, or by
the buddha guides/leaders. It might seem unlikely that the Dunhuang manuscript and the Bhutanese texts
would have introduced dpa' bos separately, and at first sight, it is unclear how the Tshal pa Kanjur could
have recovered from this reading to reintroduce, dren pas. However, the line, phyag rgya yin par dpa'
bos bshad, occurs higher up in the chapter, and not far above this point, there are two instances of dpa'
bo/dpa' bos... bshad, so it is conceivable that the Dunhuang manuscript and the Bhutanese texts generated
the line independently. 'Dren pas bshad does not occur elsewhere in the chapter. (Note that there is no
Tenjur for this section.)
/zhe sdang rdo rje 'khor bcas pas/ /'jig rten khams ni ma lus kun/ /khros pas yud la brlag mdzad de/ /brlag
par ma gyur gang yang med/ (Chapter 17)
This verse of four tshig rkang has been omitted from the Dunhuang and the Bhutanese manuscripts. It is
included by the Tshal pa Kanjur and other editions (including the Tenjur). Although this might appear to
be a shared omission, in fact it is the first of a series of four verses which the Bhutanese texts omit in
Chapters 17, 24, 29, and 34. This therefore was a deliberate omission, and the accidental omission in the
Dunhuang manuscript is most likely coincidental.
kun kyang: MsMGGr kyang ni (Chapter 38)
Either reading could fit. It might seem a little unlikely although not by any means impossible for it to
have been introduced independently, or for the Tshal pa Kanjur to emend it back again, although the
meaning is similar, and both are commonly used in these kinds of verses.
Evidence to the contrary: assessing possible alternative relationships between the descendants of b
When we consider relationships between the Tenjur, the Dunhuang manuscript, the Bhutanese and the
Tshal pa Kanjur texts, there are a number of variant readings which do not fit neatly with the pattern of
affiliations outlined above. On balance, only a very small proportion of these variants seem likely to indicate
possible contamination affecting our neat pattern of stemmatic affiliations, and fortunately, they have little
impact on our reconstructions of the readings of b or a. Nevertheless, they are worth examining to increase
our overall understanding of the TZ tradition.
Readings shared by the Tenjur and Bhutanese texts
There are a few errors and variants shared by the Tenjur and the Bhutanese manuscripts. In Chapter 32,
which the Dunhuang manuscript unfortunately omits, one rather striking variant might indicate that the Tshal
pa Kanjur has avoided a questionable reading shared by all other heirs of b:

58

Textual Analysis

mkhas pas rgyas par bsgom zhing sbyin/: Qk mkhas pas rgyas pa bsgom zhing sbyin/; QtGtNt gtang [Qt
gtan] na thams cad rgyas par 'gyur/; MGGr btab pas thams cad rgyas par 'gyur/; TRK mkhas pas rgyas pa
bsgoms zhing phyin/ (Chapter 32)
There would seem to be three possible explanations for the clear contrast between the South Central, local
Kanjurs, and the Tshal pa Kanjur texts on the one hand, and the Tenjur and Bhutanese texts on the other
hand. The series of short chapters on the four rituals rely on stock phases, which are at least in part
predictable (dbang du 'gyur, rgyas par 'gyur, zhi bar 'gyur etc.; compare the line in Chapter 26: btab na
nges par de dbang 'gyur). It is even possible that the Tenjur and Bhutanese readings could have been
generated separately through loss of attention and the picking up of similar phrases elsewhere on the
page. Secondly, it may be that an inherited reading from b, perhaps of gtang, was corrupted to btab in c.
Perhaps, then, this reading might have been rejected by Tshal pa Kanjur editors, who could have picked
up a form of words from Chapter 37 below. The third explanation would suggest contamination: perhaps,
when copying the text inherited from b, the Tshal pa Kanjur scribe might have had some corrupted lines,
and he might have had access to another version of the text from which the line here was inserted.
The other three shared variants between the Tenjur and the Bhutanese manuscripts, where the Dunhuang
reading is also present, would seem less likely to suggest contamination:
he ru dpal: QtGtNtMGGr he ru ka (Chapter 11)
This emendation to the full word, heruka, could easily have been independently generated, even
unreflectively.
pad mas brgyan: J pad mos brgyan; Qk pad mos rgyan; D padmos brgyan; QtGtNt padma rgyas; MGGr
padmos rgyas; T pad ma brgyan; RK padma brgyan; Ogl padmas brgyan (Chapter 35)
The reading, brgyan is clearer and is more consistent with previous chapters. However, rgyas could have
been introduced independently, since a blooming lotus is a common symbol in this kind of tantric
literature.
dngul phye: MsJQkD dngul dang; QtGtNtMGGr dngul lam; T rdul phye; RKBthHeOgl rngul phye
(Chapter 37)
Here, lam would seem rather less appropriate than phye or dang, but it might have been accidentally
picked up from exactly such wording in Chapter 36 of the commentary, immediately above. While dang
could work, it is less appropriate than phye, but it may have been an ad hoc correction of the
inappropriate lam made independently by the Dunhuang ms and the Tshal pa Kanjur.
Readings shared by the Dunhuang and Tenjur texts
There are a number of questionable readings shared between the two commentaries, which are not also
shared by the Tshal pa Kanjur and the Bhutanese NGB manuscripts. Some such are almost certainly simply
commentarial glosses added to the root text, which have been identified corrected as such by the editors of
the Tshal pa Kanjur and Bhutanese editions, and thus of no significance for our root text stemma. For
instance, there is an addition of the word dam pa or dam pa'i in Chapter 4, the addition of yud tsam, linked
with the words, skad cig, in Chapter 11, and an additional word, zhing, also in Chapter 11. A number of
other shared readings might be taken to indicate a shared descent, or alternatively to add to the evidence for a
hypearchetype d as the common ancestor of the Bhutanese NGB and Tshal pa Kanjur editions. However,
with the posible exception of the readings in Chapter 26, they are not really weighty enough to justify a
revision of the proposed stemmatic affiliations:
dang/: Ms omit; QtGtNt /
chen po: MsQtGtNt omit (Chapter 2)

The Stemma of the Root Text

59

These two minor omissions in Chapter 2, occurring on consecutive lines, might possibly have been
generated independently. The omissions cause the loss of metre, and the word, chen po, is clearly an
important and meaningful addition to the verse here, describing the five poisons which are to be guarded
as the samaya. The Tenjur retains the shad in the line above, but also loses the word, dang. The
Dunhuang manuscript loses the seven syllable verse form, corrupting the next lines further to produce a
further conflated line of eleven syllables. The Tenjur keeps the four line verse, but its verse is unmetrical.
It is also possible that all b descendants lost chen po, but then a common ancestor of the Tshal pa Kanjur
and the Bhutanese manuscripts (the unproven d) emended the passage, having noticed the unmetrical
lines, and being familiar with the teaching associated with this phrase.
la'ang: MsQtGtNt la; TRKBthHe la yang (Chapter 6)
It is most probable that la yang was in the archetype, producing an unmetrical eight syllable line, which
might have been emended independently in the commentaries to the metrical reading, la.
spyod yul: MsQtGtNt spyod lam (Chapter 6)
In this context, these different meanings have little impact on the sense of the verse: one's entire field of
experience, or everyday activities, become the maala. Both words are very commonly used, and it is
quite possible that this variant could have been generated independently. Alternatively, we could
speculate that a shared ancestor of the Tshal pa Kanjur and the Bhutanese manuscripts, the unproven d,
emended a reading of spyod lam inherited from b.
'khor bcas: MsQtGtNt 'khor rnams (Chapter 11)
Again, this is a small variant which could even have been independently introduced.
gtun khung: Ms gtum cen; QtGtNt gtun chen; TRK brtun khung (Chapter 22)
Here, the word chen/cen is clearly in error. It is understandable in the Dunhuang manuscript's reading
with gtum, but gtum cen/chen does not make good sense here. It is very possible that the Tshal pa Kanjur
and the Bhutanese manuscripts (or their shared ancestor d, if we accept that unproven entity) inherited the
error, chen, and corrected it to the obvious reading of gtun khung.
lus ngag sems: MsQtGtNt lus dang sems (Chapter 24)
It is possible that the variant reading, dang, has been generated independently in the two commentaries, or
that it might have been corrected back to ngag (lus ngag sems is a commonly used phrase, and clearly fits
well here) in the hypothetical d.
kha dog dmar po'i shing/ /mgo bo: MsQtGtNt kha dog dmar po yi/ /shing la (Chapter 26)
The loss of mgo bo seems unlikely to have happened separately in the two versions. At first sight, it also
seems unlikely that a later scribe could inherit this error and yet restore the correct reading. However, the
other three chapters on ritual phur pas all begin their second lines with mgo bo. This series of chapters on
the four types of rituals are short and it is quite possible that a scribe might look up or down the page to
see the other chapters, or simply be aware that mgo bo was to be expected. Hence it is possible that a
hypothetical d might inherit this error, yet nevertheless fortuitously restore the archetypal reading; but this
would have to occur despite the fact that this chapter's phrasing is very slightly different from that of the
three other chapters on phur pas, none of which include the word kha dog. An alternative explanation for
this and the following variant within this short single verse chapter is contamination: perhaps a
hypothetical d might have lost the folio(s) for Chapter 26, and sought out another version to put in place.
It could then have inserted a version of Chapter 26 which did not descend from b.
phur pa'i las: MsQtGtNt phur pa; Bth phur bu'i las (Chapter 26)
This variant might easily have been independently generated: las is not necessary here, and it could easily
have been dropped in error. Moreover, the commentaries' reading is more consistent with the other

60

Textual Analysis

chapters on the ritual phur pas of different types. As noted above, a scribe might notice the other chapters
on phur pas, so the omission of las might even have been a considered decision.
la: MsQtGtNt te; MGGr nas (Chapter 36)
This is a minor variant, and here, the commentaries' reading of bsgoms te is consistent with the chapters
on the other three types of phur pas, Chapters 20, 26 and 31, and it would also seem preferable. The most
likely archetypal reading, la, might easily have been independently altered by the two commentaries, as it
seems also to have been in the Bhutanese line.
Thus, the above variants are often amenable to explanation as coincidentally shared readings, or the
possible restoration by emendation of the archetypal reading in d, the hypothetical common ancestor of the
Tshal pa Kanjur and Bhutanese editions. With so much of the transmissional history of TZ unavailable to us,
we might never know the causes of these problematic readings, but we can be thankful that their stemmatic
location means that they have little impact on the reconstruction of the archetypal readings.
The four editions descended from a root text: the Hemis, Bathang, Tawang and South Central NGB
manuscripts.
If the fifteen witnesses that constitute the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese versions of
TZ descend from a commentarial ancestor, it is equally clear that the remaining six witnesses descend from
an ancestor that was a root text. These six, namely the three South Central Tibetan NGB manuscripts and the
local Kanjurs of Hemis, Bathang, and Tawang, all preserve TZ's correct original boundaries, with no signs of
the telltale boundary confusions that mark the Tshal pa and Bhutanese editions. Three questions must be
asked of these four editions descended from a root text: (i) is there sufficient evidence to posit a common
hypearchetype not shared with b, from which all four are descended? (ii) might any one of them be the direct
ancestor of any other of their number? (iii) do at least some of them share a separate ancestor and thus form a
sub-branch distinct from the others? If the answer to merely question (i) is positive, notwithstanding any
other considerations, we will not be able to use stemmatic methods to recover any disputed archetypal
readings. The stemma would have two branches, descending from b on one side and from a hypearchetypal
ancestor of the South Central, Hemis, Bathang, and Tawang editions on the other. Where the two
hypearchetypes differed, there would be no way logically to infer which of the two readings was present in
the archetype. On the other hand, if the answer to all three of these questions is negative, then all four
editions must descend separately from the archetype, and we will be well placed to add their independent
testimonies to those of hypearchetype b to recover archetypal readings by majority decision. If the answer to
(i) is negative while the answers to either (ii) or (iii) are positive, then stemmatic recovery of archetypal
readings will remain possible, but its operation will be weakened because the range of independent
testimonies will be reduced. The internal evidence culled from their available variant readings, supplemented
to a smaller degree by the external evidence offered by their various bibliographic histories, together offer a
reasonably high degree of probability that these four editions do all descend separately from the archetype a.
Moreover, we believe this degree of probability is, very fortunately, high enough to justify the application of
stemmatic logic in the reconstruction of archetypal readings. Nevertheless, even a high degree of probability
is not comparable to the effective certainty offered, for example, by the evidence presented above for the
existence of hypearchetype c. It follows that in this case, the evidence needs to be weighed particularly
carefully.
(i) Other than the archetype a, is there any further ancestor shared by all four editions of TZ that
descend from a root text?
The nub of the question is whether some among the examples given above to demonstrate the descent
from hypearchetype b of the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese branch (or of the group
without the Tenjur, where the Tenjur has omissions), might be so ambivalent as to offer instead evidence of
an affiliation of the South Central and local Kanjur texts. If the South Central group of texts and the local
Kanjur manuscripts each descend separately from the archetype of all the available extant editions, then we

The Stemma of the Root Text

61

can confidently restore the archetypal reading where they all agree, or where at least three witnesses agree.
If, on the other hand, they share a common ancestor not shared with the Tenjur, Dunhuang, Tshal pa Kanjur
and Bhutanese group, and thus necessarily later than the archetype, then their agreed readings will merely
restore the hypearchetype of their line of transmission and we will have a bifid stemma that will stymie
further stemmatic analysis. The main method of resolving this question is to examine the variant readings
and errors within these four editions descended from a root text. If their shared variants are indicative,
agreeing significantly enough in error, we must deduce their common descent from a single error-bearing
ancestor, a mere hypearchetype. The degrees to which any shared errors are coincidental or otherwise
stemmatically insignificant, and to which their various errors are predominantly unshared single readings
unique to each edition, determine the degree of probability that they each descend independently from the
archetype.
In the evidence given above for hypearchetype b, out of the twenty-five examples where the Tenjur
reading is present, there are no clear-cut cases of the four editions descended from a root text unequivocally
agreeing in error. At worst, one could perhaps argue that the readings of the descendants of b could be
considered slightly preferable in five cases, namely in the second example from Chapter 1 above, the
example from Chapter 19, the first example from Chapter 24, and the examples from Chapters 31 and 32.
However, in all of these five cases, the emendation of the text could just as easily have taken place in b, as it
sought slightly to improve or tidy up an inherited reading that was serviceable enough but deemed slightly
inelegant, unclear or unsymmetrical. The situation is entirely similar in those sections where the Tenjur
reading is not present: the readings of the descendants of b would seem slightly preferable in only one or two
out of eleven cases (in the example from Chapter 8, and perhaps also the example from Chapter 14).
However, once again, the emendation of the text could just as easily have taken place in b or c, slightly
tweaking the inherited transmission to achieve a marginal improvement.
There are also many instances of very minor variants, including non-standard or incorrect spellings,
particles, and so on, which are shared by the South Central and local Kanjur texts, or by the majority of
them. In Chapter 36, they agree on a reading of mgo 'am (except for Bathang, which has probably corrupted
this further, and gives, mgo la), where b's descendants give the more coherent reading of mgo bo, consistent
with the parallel text in Chapters 20, 26 and 31. In Chapter 39, the South Central and local Kanjur texts give
phyag rgya las in place of the descendants of b's reading of phyag rgya dang, where dang would seem more
appropriate (phyag rgya las/ de las). The most likely explanation for such variants, however, is that they
were present in the archetype, and corrected in later texts. A few variants are quite likely to be explained by
the fact that b's is a commentarially-derived line, and commentary will occasionally add in an extra word to
elaborate on the original root text. Some such words seem to have persisted into both the Tshal pa Kanjur
and Bhutanese root texts. In Chapter 40, for instance, the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese
texts give the more clear reading of, las bsgrub/sgrub/grub pa'i phan yon, while the South Central and local
Kanjur texts merely give, bsgrub/bsgrubs/sgrub pa'i phan yon, without the additional las, which was most
probably originally a commentarial gloss. There are no other shared errors of the four editions descended
from a root text. Thus, although a number of the South Central and local Kanjur readings might at first
glance seem suggestive of a common hypearchetypal ancestor, on closer analysis, none of this evidence is
very strong.
(ii) Might any one of them be the direct ancestor of any other of their number?
All four have their own unique errors and are unlikely simply to represent the ma dpe or exemplar from
which any of the others were copied. This is especially clear in the case of Bathang, which has many scribal
errors throughout, including omissions of lines and passages (Chapter 28, for example, is omitted, perhaps
through an eyeskip error from one chapter ending to the next), dittographies and repetitions, and corruptions
of many words on virtually every page. It also rather consistently gives certain non-standard (or incorrect)
spellings, such as omission of the final letter sa in some words (thug instead of thugs, zang instead of zangs).

62

Textual Analysis

The South Central transmission has less apparently careless errors or non-standard spellings, yet its
probable transmission over many generations of copyings has left its mark in a number of indicative errors
(see p.46 above for some examples). We also know from their comparatively later dates that none of the
South Central manuscripts could be the direct ancestor of the somewhat earlier Hemis, Bathang or Tawang
manuscripts.
In contrast, the Hemis manuscript (probably dating from the 1630's) and the Tawang O rgyan gling
manuscript (dating from 1699) have relatively few errors. The Hemis version has a dittography (of sangs
rgyas bcom ldan 'das rnams kyi mkhyen pa/) in Chapter 1, and another (of las kyi sngags dang/) in Chapter
14. In Chapter 21, it has an omission it would be difficult for any later scribe to repair through conjecture (of
chan dang sbyar ba dag/ /sbyar bas gtor na de brlag 'gyur/ /'phags pa thabs kyi zhags pa), and a further
error (phur pa'i rather than gtor ma'i). Otherwise, the Hemis manuscript has only a small number of minor
variants. It has many archaisms, such as the use of the ya btags (myed rather than med), some but not all of
which are shared with the Dunhuang manuscript. The Tawang O rgyan gling manuscript is perhaps the most
remarkable of all: made by a master scribe from gTsang at the behest of the De srid Sangs rgyas rgya mtsho,
it has very few errors whatsoever, apart from the minor errors or non-standard spellings and so forth which it
shares with the other local Kanjur and South Central texts (as mentioned above). In fact, the Tawang
manuscript is impressively close to the archetype we have sought to restore. Its biggest errors are in Chapter
10, where it identifies a mudr as that of rdo rje sems pa rather than rdo rje legs pa, and reads byas pa for
btags pa. Elsewhere in Chapter 10, Tawang is also the only text correctly to give rdo rje dbyangs ma'i in
place of the rdo rje dbyings ma'i found in all other extant versions of TZ: surely a mistake in the archetype,
quite possibly corrected by the master scribe and orthographer who made the Tawang edition, but
nevertheless additional evidence that Tawang was unlikely to be the direct ancestor of any of the other
editions. The Tawang manuscript repeats a mantra in error in Chapter 9 (o badzra ba sha hri), but the
dittography was recognised and the repeated mantra clearly deleted. Elsewhere, Tawang repetitively inserts
zhe sdang (Chapter 2), reads phan phan for phan (Chapter 4), 'og ma for 'od (Chapter 10), sbus for pus
(Chapter 10), rgya for phyag rgya (Chapter 10), sbyir for cir (Chapter 11), dpa's for dpas (an archaism,
Chapter 11), omits bar in a non-metrical sentence (Chapter 11), reads gyur pa for ma gyur (Chapter 24), and
in addition has a small number of minor variants and errors. Having so few errors is a rather impressive
achievement in some fourteen folios of text.
(iii) Are there any sub-branches within these four editions?
If separative errors are found to unite two or more of the four versions that descend from a root text rather
than a commentary, that is, the South Central NGB and the three local Kanjurs, it would indicate their
common descent from an hypearchetype and reduce the range of independent testimony available for
stemmatic reconstruction. Fortunately, there is no compelling evidence of shared indicative errors between
any two or any three of these editions.
(a) Readings shared by the Bathang and South Central manuscripts: Inevitably considering their high
ratios of errors and the length of TZ, there are a number of instances where the Bathang and South Central
manuscripts appear as though they might share errors. However, none of the numerous major indicative
errors found in either of these two editions are mutually shared, and on further analysis, those few and trivial
errors that are shared look much more likely to be coincidental, or minor anomalies in the archetype that
have been emended independently. Here are their most striking shared readings:
kun tu bzang mo'i: Ms kun du bzang po 'i; JQkD kun du bzang mo'i; TRKBth kun tu bzang po'i (Chapter
10)
We do not have either the Hemis or Tenjur witnesses here. A female deity is to be expected; it may be
that the archetype gave po'i in error. The Dunhuang manuscript also gives a male; perhaps it inherited its
reading, and this was corrected independently in Tawang and in the ancestor to the Tshal pa Kanjur and
Bhutanese texts. Again, alternatively, this might be a coincidental error in the Bathang and South Central
texts.

The Stemma of the Root Text

63

rtswa yi: Ms rtswa la 'I; JQkDMGGr rtsa yi; GtNt rtswa'i; TRKBth rtsa ba'i; QtOgl rtsa'i (Chapter 11)
Here, we have a non-standard or incorrect spelling of rtsa for rtswa in many of the versions, as well as
other permutations. In a reading where so many scribes have evidently not given the correct spelling, the
addition of ba in Bathang and the South Central editions could be coincidental: any unmindful scribe
might reflexively follow the element rtsa with a ba. However one cannot entirely rule out a logical
possibility, however small, that this might represent a shared inheritance in error of South Central and
Bathang (although again, we have no Hemis reading at this point to offer greater perspective). The
reading could quite possibly have been transmited from the archaic archetype and independently taken
out in Tawang and hypearchetype b.
thib: MsJQkDQtGtNtMGGrOgl thim (Chapter 11)
Here again, the Hemis manuscript folios are missing. The correct reading is most probably, thib, with the
sense of the deities gathering together within oneself. However, we need to consider this case, because
although we have here accepted the South Central and Bathang reading, the alternative reading, given by
the Tenjur, Dunhuang, Tshal pa Kanjur and Bhutanese group, along with Tawang, remains possible as the
archetypal reading. In this case, thib would represent a corruption shared by Bathang and the South
Central texts. Nonetheless, it seems more likely that Tawang introduced the more common thim, and that
perhaps hypearchetype b independently did likewise.
/h he he pha h/: Ms /h / ha / he / phat / h /; JQkDQtGtNtMGGrOgl h ha he pha h/;
TRK las rnams h he he pha h/ (K's second he not clear); Bth h he he pha h/ (Chapter 11)
Again, the Hemis reading is missing; the question here is whether the text should read, ha he (as given in
Tawang and b) or he he, a reading which might have been introduced by an ancestor of the South Central
and Bathang texts, but might well have been the archetypal reading.
bsgul: JDTRBth bskul; QkK unclear; MG bsgol; Gr sgrol (Chapter 16)
In this case, bsgul is given in Hemis and Tawang. The reading, bskul, is surely incorrect, but it could
either have been introduced independently in the South Central and Bathang texts, especially given the
high frequency of minor errors in Bathang and the South Central's reasonable frequency of such errors. It
might similarly have been coincidentally introduced into the Tshal pa Kanjur texts also. Alternatively, it
could have been the archetypal reading (perhaps simply a non-standard spelling), corrected in the other
texts. Note that there is no Tenjur for this section.
las sbrus: Ms las sbus (but sbrus is given clearly in the commentary below, 65r.3); TK la sgrus (but R
gives la sbrus); Bth las sgrus; He la sbrus (Chapter 21)
Again, a coincidental error seems likely: sbrus and sgrus are easily confused visually, as well as being
homophones. Alternatively, it is possible that the archetype may have given las sgrus, but the correct
reading would be easily conjectured and emended.
In some of the above examples, where the Hemis reading is missing, we must also consider if rather than
a South Central/Bathang affiliation, we are witnessing a situation where the Tawang manuscript is lining up
with b. A further example might also suggest such an affiliation:
rang bzhin: MsOgl rang byung; JQkNkDQtGtNtMGGr rang 'byung (Chapter 5)
The text here is discussing the way the naturally existent maala absorbs and emanates (rang bzhin
dkyil 'khor 'du 'phro'i tshul), and the reading of rang bzhin fits also with the gloss which follows in
TZComm (Ms14v.5). The reading, rang byung, could also fit, however, even if a little less appropriately.
TZComm below speaks also of, the absorption and radiation of self-arisen light (rang byung 'od kyi 'du
'phro, Ms13v.4) and it is even conceivable, if unlikely, that rang byung might have been in the archetype.
It is much more likely that it was introduced as an error, although quite possibly, it may have been
independently introduced in the Tawang manuscript, and in the ancestor of the Tenjur, Dunhuang, Tshal
pa Kanjur and Bhutanese texts.

64

Textual Analysis

(b) Readings shared by the Bathang and Tawang manuscripts

The Bathang and Tawang manuscripts also share a number of readings, which again could in theory
represent a shared inheritance of error, but are much more likely to be coincidences in error, or minor errors
and archaisms in the archetype that have persisted here but been normalised elsewhere:
o badzra rad na sam bha ba tra: Ms ba dzra rad na sam bha ba tra: JD o badzra ratna sa bha
ba tr; Qk o badzra na rad na sam bha ba tra; MGGr o badzra ratna swa bha wa tra; TR o
badzra rad na sam (R sa ma) b tra; K o badzra rad na [...]; Bth o rad na sa ba bha tra/; Ogl o
rad na sam ba b tra (Chapter 9)
Here, there is an omission of badzra in Bathang and Tawang. Mantras have proven notoriously difficult
for scribes to copy accurately: in several of the NGB tantra transmissions we have looked at previously,
truly correct representations of the Sanskrit are only finally achieved with the late 18th century
preparations of the sDe dge xylograph, while as we know from Dunhuang, the earlier representation of
mantras was often particularly haphazard. The trajectory of rNying ma mantra representations seems in
many cases to have been one of gradual editorial improvement from the earlier editions to the later ones,
rather than of gradual scribally generated decline from a pristine original. The possibilities of a
coincidental error or of an archaic error transmitted from the archetype (and later corrected in some
editions) are therefore too strong to accept this as evidence of a shared hypearchetype uniting Bathang
with Tawang. Note also that we are missing the Hemis (and Tenjur) witnesses here.
o badzra shwa ra dha tu o: Ms ba dzra shwa ra dha du ; JD ho badzre shwa ra dh tu o; Qk
ho badzra shwa ra dha tu o; MGGr o badzra sha ra da du o; T o badzra shi ra da hu o; R o
badzra shwi ra da hu o; K o badzra shwa ra da h o; Bth o badzra shwa ra ta du; Ogl o badzra
shwa ra da tu (Chapter 9)
The issue here is Bathang and Tawang's omission of the final o. The same comments as above apply
here. Again, the need for o (in a list of mantras all ending in the syllable, o) would be easily
conjectured, and might even have been added unreflectively. There is still no Hemis (or Tenjur) witness
at this point.
o badzra sma sha ni h dza/: Ms ba dzra kar ma h dz/; JD o badzra karma h dza; Qk o
badzra karma h dza; M o badzra karm h dza/; GGr o badzra karma h dza/; QtGtNt o
badzra sma sha ni h dza dza/; T o badzra smra sha ni h dza/; Bth o badzra [ba?] sha ni h
dza/; Ogl o badzra ba sha ni h dza/ (Chapter 11)
Once more, the comments made above apply. Here, Bathang and Tawang both give ba sha ni where the
correct reading is sma sha ni ( ma n), but again this could be easily conjectured by the South Central
and Tenjur lines (Hemis is still missing at this point; RK do not share T's spelling error). Non-standard
and apparently inaccurate mantras are often found in early rNying ma tantric materials, so ba sha ni is
plausible as the reading in the archetype.
brlags: Ms plags; QkOgl blags; QtGtNt brlag; TRKHe rlags; Bth blag (Chapter 11)
Bathang and Tawang's reading of blag/blags would seem to be a different, less appropriate verb. Again, a
highly possible explanation is coincidence: the loss of an attached letter ra is a very minor change, and it
might also have been lost coincidentally in the Dunhuang text, as well as in one of the Kanjur texts. But
blag/blags may simply have been a non-standard spelling for brlags, which Hemis and the South Central
manuscripts emended.
sems dpa' chen po: BthOgl sems dpa' (Chapter 12)
Here, Bathang has rdo rje sems dpa' nyon cig/, while Tawang has sems dpa' nyon cig/. The correct
reading, given in all other editions, is probably: sems dpa' chen po nyon cig/. It is not at all clear that this
indicates a shared hypearchetypal ancestor of Bathang and Tawang.

The Stemma of the Root Text

65

grdza grdza h pha/: Ms ga rdza ga rdza ga rdza h phat /; JQkDMG gardza gardza gardza h
pha/; Gr gardza/ gardza/ gardza/ h pha/; TRK gardza gardza h/; Bth ka dza h phat/; Ogl g rdza
h pha/ (Chapter 14)
Once more, the comments made above about the representation of mantras apply. Here, Bathang and
Tawang share a single, as opposed to a double or triple mantra syllable, but it appears in a section on the
ten mantras of the Khro bo bcu in which there is confusion both within and between editions on whether
there should be single, double or triple mantras elements. (There is no Tenjur for this section.)
None of the above examples are convincing indications of affiliations between these manuscripts, and
moreover several of them are grouped within the chapters where the Hemis manuscript is missing. If the
readings in these cases were found also in Hemis, then they would no longer be readings shared only by
Bathang and Tawang, and we would deem it most likely that they have descended from the archetype.
Reading shared by the Hemis and Tawang manuscripts
Finally, there is a shared reading between Hemis and Tawang, which is entirely typical of archaic
Sanskrit representations, and so likely to have been in the archetype:
o badzra kro dha h ka rdza h/: Ms / ba dzra kro dha h ka ra h /; JQkD o badzra kro dha
h k ra h/; MGGr o badzra kro dha h ka ra h/; TRK o badzra kro ta h/ ka ra h/; Bth
o badzra kr dha h; He badzra kro dha h/ kardza h/; Ogl o badzra kr dha h ka rdza
h (Chapter 14)
We would expect the mantra for the wrathful deity, Hmkara, but both Hemis and Tawang give kardza/ka
rdza rather than ka ra. The reading may have been shared by Bathang, but Bathang lost the deity's name
altogether. It is quite likely that the ancestors of the South Central group and of the other branch of the
transmission emended the mantra. (Again, there is no Tenjur for this section.)
The paucity of clear indicative errors shared by the descendants of b with any of the other texts:
The basic shape of the stemma, which includes deriving the Tenjur, Dunhuang, Bhutanese and Tshal pa
Kanjur texts from a single hypearchetype b, would seem quite clear. There are only a few errors shared by
the descendants of b with any of the other texts, mostly explicable as coincidental errors or chance
emendations:
(i) There are a few readings shared by all descendants of b except the Bhutanese NGB texts, who instead
agree with the root-text derived editions. However, these few readings do not constitute evidence of any real
weight. In Chapter 1, the Tshal pa Kanjur and Tenjur texts share an omission of the word, rnams, at a place
where the Dunhuang manuscript omits a longer phrase (rendering its earlier reading uncertain). Here, the
Bhutanese texts apparently avoid an error descending from b, but the addition of rnams could have been
easily conjectured. There are two similar cases where the Dunhuang manuscript reading is present and
shares the b reading, but which the Bhutanese texts do not carry:
gnod: MsJQkDQtGtNt gnon (Chapter 11)
Either reading is fine, and in fact, gnon may be considered slightly more appropriate. It is possible that
the ancestor of the Bhutanese texts may have emended gnon, possibly even accidentally.
thams cad: MsJQkDQtGtNt rnams; Bth thad (Chapter 11)
Again, the Bhutanese texts avoid the reading of the rest of the branch, perhaps by unreflective
emendation.
(ii) As mentioned above, the Bhutanese and South Central texts share a final colophon not given elsewhere,
but nevertheless, this does not indicate an overall textual relationship. These two regional NGB traditions are
close geographical neighbours, and the shared final colophon is most likely due to a limited geographically
determined contamination, for example, the transmission from one area to the other of a title list or dkar

66

Textual Analysis

chag which included colophons. This deduction is supported by the fact that apart from this colophon, they
share only two minor variants throughout the whole text of TZ: both give dang sor instead of dang por in
Chapter 10 (which would have been easy to generate separately, either through the similarity of sor and por
in some scripts, or through accidental copying of sor above), and both give yin rather than zhes in Chapter
16, but this is in the context of the repetition of a phrase ending above with yin. The Dunhuang manuscript
has a crossed out yin, showing that its copyist had automatically written this, before realising the error.
(iii) Similarly, there are a few very minor variants shared by the Tshal pa Kanjur and South Central lines, but
none of any real stemmatic weight. The seemingly most significant is that both omit the mantra for rDo rje
bkrag gsal ma in Chapter 9 (o badzra ra ga ra ti hri), but this could be a coincidental omission. Both give
la rather than la'ang/la yang in Chapter 13, but they might separately have inherited la yang, which they
shortened for metrical reasons (la yang makes an eight syllable line).
(iv) Chapter 17 has an instance of a reading of dbyings, apparently shared by the Dunhuang, Bhutanese and
South Central manuscripts, where the correct reading is dbyangs: a small mistake easy to make and equally
easy to correct. Thus, dbyings could have been introduced separately and coincidentally into the South
Central NGB edition and into hypearchetype c, and then corrected in the Tshal pa Kanjur edition.
(v) In Chapter 11, the word, yang, is added in the Tenjur, the Dunhuang manuscript, and in the South Central
branch. The additional yang possibly improves the passage, and it might have been separately inserted in the
South Central and the other branch of the transmission. In the commentarial branch, it may have been added
as a commentarial elaboration, correctly identified as such by both the Tshal pa Kanjur and Bhutanese
transmission.
Summary: the South Central Tibetan NGB and local Kanjur editions
It is apparent from the analysis above that the evidence of a common hypearchetypal ancestor of the
South Central Tibetan and local Kanjur texts, or of any affiliations among their four lines of descent, is
decidedly weak; as is the evidence for affiliations between any of these editions and any descendants of b.
Even before we were able to collate the three local Kanjur witnesses, consideration of the Dunhuang
manuscript, Tshal pa Kanjur and Bhutanese texts' shared indicative errors had led us to conclude that the
South Central Tibetan tradition, which avoids them, must have descended separately from the archetype. We
could also sense that the TZ archetype was quite likely to have had similar boundaries to those still preserved
in the South Central Tibetan tradition. However, at that stage, we could not establish what those boundaries
were, since we had no way of distinguishing between the good readings inherited by the South Central
Tibetan tradition and its unique corruptions. It was only after the very belated acquisition of the three local
Kanjur witnesses in the final months of our work on this text that we have been able to restore the TZ
archetype and its correct boundaries with reasonable confidence. The errors and amendments within the
South Central texts can now be identified with certainty, where they differ from the three local Kanjur texts.
Existing evidence is not sufficient irrevocably to eliminate even the smallest possibility that the local
Kanjurs and the South Central Tibetan manuscripts might descend from a single shared hypearchetype, so
that the stemma would be bipartite or bifid, with these editions forming one branch, and the descendants of b
forming the other. However, as we have seen, there are no clear indicative errors shared between the local
Kanjurs and the South Central edition, so that probability remains small. Moreover, since all three of the
local Kanjurs, as also the South Central texts, come from geographical regions very distant from one another,
their separate descents from the archetype might seem all the more likely.
Conclusion: the Stemma with five branches
The diagram of the stemma above (see p. 43) illustrates the most likely overall picture of the relationship
between the extant texts, although it is also possible that there might be a further hypearchetype, d, a shared

The Stemma of the Root Text

67

ancestor of the Tshal pa Kanjur and Bhutanese NGB. In addition, there might have been some limited
contamination: there are a few instances of readings in either the Tshal pa Kanjur or the Bhutanese NGB
which might indicate that an ancestor might have consulted a different version of the text. It is possible, for
instance, that an ancestor of both of them (the hypothetical d) might perhaps have restored a lost Chapter 26
from a different ma dpe or exemplar, thus avoiding the errors of the Tenjur and Dunhuang manuscript for
Chapter 26 (but not elsewhere in the text). At a separate stage, an ancestor of the Tshal pa Kanjur editions
might have dealt with some corrupted lines in Chapter 32 by consulting an alternative ma dpe which avoided
the reading transmitted in the Tenjur and Bhutanese texts. Yet as we have noted above, it is also conceivable
in both cases that the archetypal readings might have been restored without such contamination.
Such caveats notwithstanding, we believe we can in many cases successfully apply stemmatic reasoning
to establish archetypal readings. For where a majority of the five branches deriving from the archetype agree,
other things being equal, there is a good probability that they will represent an archetypal reading. However,
it also follows that where the five branches are equally divided in their readings, we cannot be certain of the
archetypal reading. Fortunately, in the case of TZ, such disagreement generally occurs only over very minor
variants; and wherever the archetypal reading is uncertain in this way, we draw attention to the possible
alternatives in the apparatus by giving the reading in italics.

e) A Summary of The Commentary on A Noble Noose of Methods, the Lotus Garland Synopsis ('Phags
pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa)
Chapter 1 The Two Truths
TZComm begins with a twofold classification of the Mahyna spiritual path, as consisting of the Vehicle
of Characteristics, involving bodhisattva conduct, spanning a period of three aeons, and the Vajra Vehicle,
involving the Secret Mantra, and bringing realisation in this lifetime. In this Vajra Vehicle, all dharmas are
subsumed under the three maalas and the total purification of all objects and domains of experience brings
the enlightened qualities. The text title is glossed in terms of the features of the Vajra Vehicle: Noble since it
indicates accomplishment in one lifetime, and the Buddha's own practice; Noose since it represents the nonabandonment of sasra, and Methods since evil beings are liberated through its great compassionate action.
In the image of the Lotus Garland, Lotus symbolises wisdom and Garland symbolises methods. Synopsis
indicates that it represents a summary of all the key scriptural sources. The root text begins with Vajrasattva
offering to Vairocana and Vairocana teaching in response the sameness of all outer and inner dharmas,
summed up in the two truths. The commentary elaborates that ultimate truth is the primary cause without
essential nature, so that all dharmas arise without characteristics, while relative truth is the result, since they
appear like an illusion, as a creative display. With this understanding, everything becomes pure, and all
activities become the Two Accumulations. However much one may be immersed in all phenomena,
understanding the ultimate cause prevents the two kinds of obscuration, so this is the Accumulation of
Primordial Wisdom, while the immersion itself constitutes the Accumulation of Merit. The root text states
that by means of the vajra of even awareness, the ten bodhisattva levels are unified in evenness. The
commentary explains how each of the bodhisttva perfections are to be practised with this view. Since one
does not fall into habitual tendencies while immersed in phenomena, one is endowed with the fruits of the
Perfection of Giving. Since one has pure awareness of the karma of defilements as none other than complete
purity, this is the Perfection of Ethics. Having patience towards dharmas and not becoming defiled even
during conflicts, is the Perfection of Patience. Being aware of the four aspects of the conduct in everyday
life as the Accumulations of Merit and Wisdom, is the Perfection of Effortless Diligence. The Perfection of
Absorption is maintaining awareness of the five senses as the five family buddhas and the five sense objects
as their five consorts, and that there is no movement from the dharmadhtu even when the mind is agitated,
so that through a Vajra-like Samdhi, everything is entered into without obstruction. Awareness that the
dharmas of sasra in which one is immersed, and those of nirva are not different, is the Perfection of
Wisdom. Being aware of all defiled dharmas as the utterly pure dharmas, the purposes of self and others and
liberation are achieved. This is the Perfection of Skilful Methods. Awareness that oneself and the
enlightened are not different, and that all fields of experience and activity are utterly pure, is the Perfection
of Aspiration. Since being aware that sasra and defiled action are nothing other than nirva and
complete purification, the suppression of sasra and defilement is the Perfection of the Buddha powers.
Having realised that the person and dharmas have no self, so being aware that the habitual tendencies do not
even have their own essence, and not moving from such a primordial wisdom awareness at any time is the
Perfection of Primordial Wisdom. The fruits of these are that the ten bodhisattva levels can be perfected
instantaneously in this very life.
The results of virtuous and sinful actions are not nonexistent, but just as a light appearing within
darkness, so dark thoughts dissipate naturally when the light has come. By awareness of all phenomena in
their sameness nature, they are accomplished as buddha. There are four aspects of sameness: 1. the
sameness of the outer world's five objects and the Tathgatas; 2. The sameness of the inner mental
consciousnesses and the five primordial wisdoms; 3. The sameness on the relative level of the senses and
their objects, and their mode as male and female deities; 4. Their ultimate sameness in their birthless and
deathless mode. Without these awarenesses, non-virtuous actions make you fall into lower realms. Without
awareness, the buddha fields may be seen as the hells, but with awareness, the hells are seen as buddha
fields. If one is aware of the sameness of all dharmas, then virtuous and sinful dharmas are the same, and

A Summary of The Commentary

69

arise as the accumulations of merit and wisdom. If one engages in everything, in sameness, then in this very
lifetime, the state of omniscience is achieved.
The pleasures of the five senses are the natural qualities of the five kinds of offerings. Desire is the
Buddha Body of Supreme Enjoyment. Hatred, rage and wrath are the power of great compassion. If
conjoined with emptiness, without self, sentient beings' distorted behaviours can be joined with the truth of
non-duality. Since emanations are sent forth, hatred is the Emanation Body. Since all dharmas are
encompassed within complete non-conceptuality, delusion is the Dharmakya. Therefore, by not abandoning
the three poisons, they become the three Buddha Bodies; this is the short path.
Even if one were to work only on abandoning defilements, there are one hundred and ninety-eight of
them. Having abandoned these, one would achieve the fruit of the Arhat level. Then, having abandoned the
primordial defilement of what is to be known, through countless aeons, one will become a buddha. In this
way, like the rvakas, one may see the defiled and the pure dharmas separately. Yet if one is aware of the
utterly defiled phenomena as without self nature, they are completely pure, and there is no abandonment.
The level of omniscience is attained in a single moment. There is no need to tarry for a future buddha there
will be accomplishment in this life.
Chapter 2 The Samayas
The root text says: "Delusion, hatred, pride, desire and jealousy: are the code of commitments in which
there is nothing to guard against; they are the samayas which are vajra." Delusion is the samaya of
Vairocana: it lacks any abandonment of ignorance, and lacks any accomplishment of awareness, for they are
both of one taste in the dharmat. Hatred is the samaya of Akobhya, because it tames all beings, even
though not moving from the dharmadhtu. Pride is the samaya of Ratnasambhava, because it is enduringly
fixed in being aware of the truth of the sameness of dharmas. Desire is the samaya of Amitbha, because it
is determined to bring everything under its power. Jealousy is the samaya of Amoghasiddhi, jealous in
resenting the sending out of virtue and sin, when the sameness of dharmas is not realised. In short, if one
enters into the sameness of dharmas, and becomes endowed with self-aware primordial wisdom, there is
nothing to guard or not to guard. But if this is not accomplished, and one lacks the primordial wisdom
sameness in the mind, but does not guard samaya, then one arrives in the deepest of the hells.
The major samayas are the three aspects of purity, the five samayas, the five sacred foods, the samayas of
the absence of virtue and sin within phenomena, and of the lack of purity and impurity in foods. If these
samayas are unimpaired, then even if one engages in the five evils or poisons, one is not tainted by their
faults. This is the body vajra. Even if one says various things with the speech, one is untainted, and this is
the speech vajra. Being aware in this way is the mind vajra.
Chapter 3 Empowerment
The root text states: "these empowerments are obtained through the expressive power of one's own innate
awareness." Through the three maalas, when one is aware of the sameness of all dharmas, this is known
as obtaining empowerment through one's own natural qualities.
Means and wisdom are the great empowerments, so the root text instructs that vajra and bell are to be
held with awareness, and thus, all the dharmas of nirva and sasra are mastered. The male head of the
family is skilful means, the sambhogakya, while the female head is wisdom, the dharmakya. The
bodhicitta arisen from their consecration is the nirmakya. This is the empowerment of the three buddha
bodies. The vajra, the sign of means, is to engage in all phenomena. The bell, the sign of wisdom, is
awareness of all phenomena as illusory. Hence, all the phenomena of sasra and nirva, are engaged in
and brought under control. The Glang po is cited: "Since the buddha is the aware nature of mind, do not
seek the buddha anywhere else."

70

Textual Analysis

Chapter 4 Offering
It is crucial to understand "the supreme sacred offering". Because all enlightened and ordinary beings
partake of the dharmadhtu sameness, the yogin who is aware that they are not separate and not
differentiated, constantly engages in the five desirable sensual enjoyments. Through gratifying oneself, all
the noble ones are delighted; and by consecrating sasra in this way, all sentient beings also become
liberated and blissful. Thus should the yogin perform offerings at all times.
In performing the burnt offerings rite, with the samdhi that the noble ones and oneself are inseparable,
through the five seed syllables within one's own body, transformed into the celestial palace, the noble ones
are clearly visualised seated in the middle. The mouth is the homa fire offering pit, the tongue is the chief
deity, along with his emanations, and the hands are the homa's vajra ladles. Since one is endowed with the
mind of a noble one, one will be unmistaken as the almighty deity, so in this way, the food offering, adorned
by embellishments, should be made.
Chapter 5 The Samdhi maala
To explain the opening of the three maalas, the root text says: "Transcending the [eleventh
bodhisattva] level of universal light, becoming the Lotus-Eyed [Buddha], by means of the great consecration
of awareness, the naturally existent maala (rang bzhin dkyil 'khor) absorbs and emanates". The
dharmakya abides all-pervasively. Then by the consecration of effortless awareness, it is like self-arisen
light, radiating and re-absorbing. Arisen from the dharmakya consecration and absorbed into it again, the
maala abides with its own natural expression without any characteristics. Thus, the root text continues:
"Endowed with the [thirteenth bodhisattva level of] the assembled wheel of syllables, the major and minor
[buddha] marks all emanate from it. It is endowed with those whose forms are ravenous and fierce, and with
the swift males and females as messengers".
The five buddhas arise from the five syllables, and the five female buddhas arise from the five primordial
wisdom objects, and so too arise the sixteen male and sixteen female bodhisattvas, their total number of
thirty-two related to the major buddha marks. Each of the sixteen bodhisattvas has a diadem of the five
families, and these five times sixteen amount to a total of eighty, which are connected to the minor buddha
marks. The male and female wrathful deities are fierce towards the extremes of eternalism and nihilism.
The messengers go everywhere without going, indicating they are no different from the dharmadhtu. Since
they are all self-arisen, with no exertion, this is called the naturally existent maala (rang bzhin kyi dkyil
'khor).
The root text describes this as, "the vital seed vajra (thig le rdo rje)". Here, vital seed is the dharmadhtu.
Vajra is the mode of appearing as the male and female deities. The deep meaning is that it is what is known
as the vajradhtu maala. This samdhi maala brings realisation of the absence of self within the
person. The yogin, without needing effort, realises whatever arises as the expressions of these male and
female deities.
Chapter 6 The Representational maala
The root text asserts that the five elements are the tathgatas, and that moreover, each of the five are
within each one. The commentary elaborates that space is Vairocana, earth is Akobhya, water (or fire) is
Ratnasambhava, fire (or water) is Amitbha, while air is Amoghasiddhi. Also, earth on its own has hardness,
cohesion, heat, and mobility, and each of the five are latent within space. The other elements can be
understood similarly. Body is the form realm, speech is the desire realm, and mind is the formless realm.
Body is the sambhogakya, mind is the dharmakya, and speech is the nirmakya.
The root text says: "the five wisdoms and the five classes of beings are indistinguishable... by the mantra
and mudr empowerment, one's entire field of experience becomes the maala". Since the Tathgata and
oneself are not separate, there is no need to invite the Buddha from elsewhere. All speech is mantra, so there

A Summary of The Commentary

71

is no special reciting of essence mantras. Moving the body is the ritual mudr, so there is no need to uphold
mudrs. The mantra and mudrs are empowered by the creative power of the pure primordial wisdom
awareness of this great self identity.
The five faculties are the five male consorts, while their five objects are the five female consorts. The
body is Vairocana. The eye is Akobhya. The ear is Amitbha. The nose is Ratnasambhava. The tongue is
Amoghasiddhi. All forms are Vajralsy (Vajra Charm); all sounds are Vajragt (Vajra Melody); all smells
are Vajraml (Vajra Garland); all tastes are Vajranty (Vajra Dance); all touchables are Samantabhadr.
The term for the maala, the Assembled Centre and Circle (tshogs kyi dkyil 'khor) is glossed as follows.
Everything is gathered together and assembled through the consecration of oneself, so it is an assembly. It is
called, centre since everything emanates from one's own mind, and since all the primordial wisdoms are
emanated from the pure dharmat. With pure awareness, dharmas and mind become the same, so everything
is said to be centred in the mind. The reason it is called, circle, is because primordial wisdom, without centre
or circumference, is within everything, actively pervading and perfecting it.
Applying the technique of recognising the complete purity of the various aspects of the mind and
consciousness, they become the five primordial wisdoms and the five buddha families. [The Tenjur version
of this section is quite differently worded from the Dunhuang manuscript version, and seems rather more
coherent (see TZComm edition p.254-255), but following this passage, it then omits all the next text, up until
the final section of Chapter Ten.]
"Bodhicitta is the supreme siddhi." Without bodhicitta, even if one attains the divine superknowledges,
there will be no siddhis of the Buddha, nor understanding of the dharmadhtu nature. Doubts are the
dharmadhtu, so doubts are not cleared away. Since dharmas are not other than mental confusion, confusion
itself is unborn, unmoving in the expanse of incomprehensible non-conceptuality.
The means unite indivisibly with the wisdom, and the seed syllables are surrounding. Offerings should be
made to the lama. Uttering the appropriate mantras, and focusing on the syllables at the tip of the vajra and at
the centre of the lotus, the lotus petals are parted. Moving with the triple h, inconceivable light emanates,
and offerings are made to the Noble Ones. The bodhicitta flows down and is unified in the private place.
The bliss of the dharmat satisfies all sentient beings, thus bringing their benefit. The outer offering
goddesses are the various sensations, while the four inner goddesses relate to the inner union. When the
bodhicitta is produced, one meditates in absorption in the unborn, and on the mahmudr samdhi. If the
mind wanders, one meditates on the absorption and emanation of tiny vajras. Uniting in desire with the
mahmudr, the bodhicitta's natural qualities are meditated on in the immanent reality of great bliss.
Meditating within the samdhi of unborn bodhicitta is the dharmakya. The great bliss engendered through
the bodhicitta's natural qualities, is the buddha body of merit. The absorption and emanation of tiny vajras is
the nirmakya. Spontaneously unifying the three buddha bodies, together with the goddesses, one
becomes aware in a naturally non-conceptual manner.
Chapter 7 The characteristics of the deities and the stages of the Maala's Array, the consecrations of
the five principal males and the five principal females, the sixteen male and the sixteen female
bodhisattvas
The root text explains that the buddha bodies of the maala deities, in all their seemingly diverse
appearances, are entirely like jewels, so that all enlightened purposes and deeds will be achieved by
meditating on them. The commentary elaborates on the specifics of these deities of the peaceful maala.
They consist of twenty-five male and twenty-five female deities, divided into five groups of five. Each
group is headed by one of the five Buddhas (Vairocana, Akobhya, Ratnasambhava, Amitbha,
Amoghasiddhi) and their consorts. These deities are embodied in the outer world and its beings,
differentiated by family. The tathgata family is golden, the vajra family is white, the jewel family is blue,

72

Textual Analysis

the lotus family is red, and the karma family is green. They are "like jewels", since they are clear and blaze
with light, with stunning colours and shapes. Their mudrs are the wheel for the tathgata family, the vajra
for the vajra family, the jewel for the jewel family, the lotus for the lotus family, and the sword for the karma
family.
If one meditates on means and wisdom combined, the three kinds of siddhis will be attained. The objects
of the twenty-five primordial wisdoms are identical in their dharmadhtu character, yet arise separately.
Depending on one another, the means and wisdoms combine, and are endowed with bodhicitta. Consecrated
by primordial wisdom, the countless forms become one, while simultaneously, through awareness, becoming
countless different buddha bodies. Their suchness clear and radiating, they resemble jewels.
The deities are placed on the body parts with mantra syllables, and sensual experiences are transformed in
this way. Thus, Vairocana and his consort are uniting upon the head, and the four other pairs of their group
are on the eyes, ears, nose and tongue. Awareness of the emptiness nature of the seed syllable, there is
seeing in the mode of non-seeing, hearing in the mode of non-hearing, and so on, and in this manner, seeing
is universally activated, and so too the other sensual perceptions. Associations of the seed syllables and their
relationship to each other are given.
The vajra family deities are on the fingers of the right hand, with Akobhya in union with his consort
upon the middle finger. The right hand of the noble person is forceful in everything, so it is the right hand
which is used, since the male and female deities of the vajra family vanquish all opposing mras and all
siddhis are received.
The left hand of the noble person assembles every need and enables one to hold to the sacred. Thus, on
the middle finger is Ratnasambhava and his consort, with the other deities of the jewel family on the other
fingers. The primordial wisdom of sameness is non-abiding, so for this reason it burns up abiding in views
in terms of self.
Upon the middle toe of the left foot are Amitbha and his consort, and the other deities of the lotus family
are on the other toes. The left foot of the noble person in a leaping posture is held stably. Likewise, the male
and female deities of the padma family desire in the mode of desirelessness, and since this is made stable, it
is the left foot.
Amoghasiddhi and his consort are upon the middle toe of the right foot, and the other pairs of deities of
his family are on the other right toes. The right foot of the noble person in a leaping posture causes one to
arrive at the other shore, and likewise these deities bring accomplishments.
Chapter 8 The accomplishment of siddhi
The yogins and yogins realise the ultimate meaning of the three maalas in their inseparable cause and
effect. The yogins are meditated on as the twenty-five in the circle of the five buddhas, and the yogins are
meditated on as the twenty-five in the circle of the five objects. Thus, the yogins generate joy. The smooth
jewelled lotus circles are offered with bright beaming smiles, so that the joyful plunging vajras magnetically
invoke clouds of bodhicitta. When the bodhicitta flows, it is offered to the deities of the maala with the
exclamation, "a la la la ho!"
Many emanations are produced, and ritual activities of the four types are performed. With the mantras
and mudrs of the buddha families, the buddha family deities' natural qualities are generated in the medicinal
elixir, and creative seeds are produced in one's heart and in all the sense faculties. Drinking them in, the
supreme siddhi is obtained. Ultimately, consecrated by the conjunction of awareness and the spatial field's
nature, the arising of awareness which unmistakably realises the ultimate meaning is the bodhicitta and this
is Vajrasattva. Through this, subject and object dualism is transcended and the thirteenth (bodhisattva) level

A Summary of The Commentary

73

is attained, the supreme siddhi. The sGron ma brtsegs pa is cited as saying: "The bodhicitta becomes the
supreme siddhi, called, essential self-arisen awareness."
The root text continues: "The yogin, possessing the female vajra holders, bestows empowerment onto the
garments and offering foods, the substances of accomplishment, and in consuming (them), (they) are
transformed into the supreme siddhi". The offerings are made with the appropriate mantras and mudrs. As
they are consumed, the siddhis of vajra longevity are obtained. This means that the non-conceptual dharmat
is the buddha body. This essential nature is dressed in various appearances, and it appears within awareness.
Within the thusness of awareness, materiality is vanquished and consumed.
Then the yogin possessing the female vajra holders, consecrates the sounds of the songs and words as the
essence of buddha speech, and they are transformed into the supreme siddhi. The songs are heard in the ten
directions of the world, and mras are subdued through splendour. This means that the spatial field of the
inseparable male and female deities pervades everywhere and the awareness nature of the spatial field is the
sound which is heard.
Then the yogins, meditating on the male and female deities, emanate out all the movements of dance and
display in various light rays, and once again gather them in, meditating on the deities and their assembly
blazing and turning into light, so that supreme siddhi is attained. In other words, when all the male and
female deities are conjoining at one moment, even the individual families are indefinable. At once, the vajra
also transforms into a precious jewel. Visualising the transformations blazing in amassed light, the
realisation is the supreme siddhi.
Chapter 9 Mantras
The mantras are given for all the peaceful deities described in Chapter 7. The mantras come forth from
the buddha body, speech and mind vajras. First comes the mantra for Vairocana, then his accompanying
four male deities, then Akobhya, with his four male bodhisattvas, and so on. Then there are the five female
buddhas, followed by the list of female bodhisattva goddesses, starting with the four in the first group,
continuing with those in the the second, third, fourth and fifth groups, and ending with four concluding
mantras. Since the list is self-explanatory, the commentary adds little, although the Dunhuang manuscript's
annotations attach the names of the deities to their mantras.
Chapter 10 Mudrs
The root text makes the point that when one is actually the male and female deities, the body possessing
them at the five limbs (as described in Chapter 7 above), every movement is mudr. Meditating on the hands
as means and wisdom, the various hand mudrs are made. A description is given of all the mudrs for the
peaceful maala, and as in Chapter 9, the commentary makes no elaboration. Following the opening
mudrs, each mudr is given a name which identifies the deity whose mudr it is. The list begins with the
mudrs for the five buddhas, followed by those for the male bodhisattvas (apart from Vairocana's group), and
then continues with the female deities in a slightly less obvious order. The first four appear to be for four of
the principal females, and with some anomolies, the female bodhisattvas are given starting with the second
group. The Dunhuang, Tshal pa Kanjur and Bhutanese texts have a shared omission at the end of this
chapter (and the Tenjur only picks up the final lines following its omission of text starting in Chapter 6), but
the South Central NGB and local Kanjur manuscripts (apart from Hemis, which is missing these folios)
complete the list appropriately, although a few anomolies in the list as a whole remain (see the table of
peaceful deities in the Appendix, and especially note 3).
Chapter 11 Cutting off Vidy Mantras
Throughout the vast reaches of the ten directions of the world, the incomparable five male and female
deities' consecrations for accomplishment are infinite. The male and female bodhisattvas are now joined by
male and female wrathful deities with great powers of sorcery, who protect the world. The great heart vows

74

Textual Analysis

of yore are expressed: since all and every world without exception is unified with the Conqueror's field, one
aspires to unify with the mahmudr. Those possessing the sorcery of the great spiritual warriors are
requested to consecrate one's noblest intentions for accomplishment through the consecration of great
compassion. Prostrations are made to the buddha emanations unified with the three realms, and a samdhi
called, "wrathful display" is entered for the purpose of dealing with the harm caused by evil mantras. It is
explained that the tathgatas' body, speech and mind vajras already discussed will be effective.
Then the Victorious One entered into a samdhi called, "vajra weapon", and from his body, speech and
mind vajras, secret mantras were emitted, weapons fell and Vajrapi collapsed with his retinue. The
Victorious One revived him and explained the rituals to accomplish this secret mantra in future times.
On a sun maala, the Great Glorious Blood-drinker is to be invited. He has a smokey coloured body,
dark russet brown hair, three heads and six arms, and his heads are red, smokey coloured and black. In the
three right hands, he wields a large corpse staff, a human skull-cup filled with blood, which he stirs with a
vajra. In the three left hands, he holds a noose of corpse intestines, a huge garua hawk, and from the belly
of a thin corpse, rips out and eats the internal organs. On the crown of his head, Vairocana himself is seated.
Akobhya is on his right shoulder, Ratnasabhava on the left; Amoghasiddhi is on the right foot, and
Amitbha on the left.1
A vajra wheel is sent forth from the syllable bhru, and the vidy mantras of the swift females are
invoked to summon the evil spirits. An effigy is drawn, and singing vajra invocation songs, the messengers
are commissioned. The evil vidy mantras embodied in the effigy are put into a triangular homa container,
fire is ignited, and the evil perpetrators are called up by waving the scarf. Their vidy mantras are burnt.
The carnivorous spirits come for the feast. There is a further ritual attack; singing and brandishing the vajra,
the deities are dissolved into oneself. Then magical hybrid deities are summoned from the different
directions, their mantras are given, gtor mas are offered and they are entrusted with action. A variation on
the visualisation of the deity is given for destructive action. Then further wrathful goddesses are visualised
in the different directions of the outer circle, Sihamukh and so forth (the lion-, tiger-, fox-, and dog-headed
etc.) Five seed syllables are then to be recited. With h, the principal deity is generated; with ha, the four
inner goddesses are generated; with he, the four outer goddesses; with pha, all enjoined to be wrathful; while
with the final h, all blaze with yellow primordial wisdom light.
The chapter ends with a description of the tshogs feast offering rite. Prostrations are made with a request
for the evil ones who destroy samaya to be destroyed. With the practices of ritual union and liberation in the
great tshogs feast, the majestic splendour of the wrathful deities of buddha body, speech and mind will expel
all evil spirits. So majestic songs are sung at the feast and everyone becomes joyous. Having bound the
vajra lotus mudr at the forehead, further mantras are recited. The circle of the retinue attending the tshogs
feast praise the principal deity with song. The principal deity is delighted; he replies in a song which
bestows the samaya. Goddesses strike the drums and sing melodies.
Chapter 12 The Maala of the Wrathful Male Deities
[Note that the Dunhuang manuscript contains a passage near the end of its Chapter 12 which in fact was
misplaced, actually belonging in Chapter 13.]
The chapter gives a description of the wrathful male maala deities, with glosses throughout on the
imagery's significance. The root text begins: "In the middle of an expanse of radiating and re-absorbing
light rays, which blaze like the fire (at the end) of the aeon, the great wrathful deities abide, positioned in a
posture of stretching (the right legs out) and bending (the left inwards)." In order to tame by intimidating
and destroying the mras, obstacles, and so on, all the worldly realms blazed up in a great fireball, and light
1

Note that this placement fits with the schema for visualisation of the five buddhas at the parts of the body in Chapter 7 above.

A Summary of The Commentary

75

rays radiated from it. The wrathful deities were abiding in the middle, scowling with glaring eyes. The
burning fire of light rays are the five primordial wisdoms entered without obstruction, and they swirl within
the nature of the spatial field, embracing it within a non-dual samdhi, transcending the extremes. The
stretching of the legs indicates wisdom, while the bending of the legs indicates means, together uniting in
non-duality with the nature of the spatial field.
The root text says: "The central Great Blood-drinker is displayed in a form with nine heads and eighteen
arms; vajra-filled wings raised up (at the shoulders), and also eight legs." The heruka lord has a white
central head; the right is red and the left green. The next row's central head is yellow, the right light blue,
and the left light red. The final row's central head is black, the right reddish maroon, and the left dark green.
His hair is dark russet coloured, whirling upwards. He has a necklace of dried skulls, and a garland of fresh
wet heads as a shoulder belt. His body colouring is black, but sometimes he assumes other colours and body
forms. In the first two right and left hands, he stirs with a vajra and drinks from a skull-cup of blood. The
two hands beneath these pull out and consume the heart of a fresh human corpse, while the next two wield a
noose made from the innards of a human corpse. The hands beneath grasp a child's corpse and swallow it
whole and the next hands hold sun and moon maalas to his heart. The following pair of hands brandish
weapons which transform into different types, and the two beneath these hold up magical hybrid deities. The
two hands below hold maalas of wind and fire, scattering and burning up all worlds, while the next two
hold a water maala and mountains, tossing about and suppressing the worldly realms. The two hands
beneath hold a chest of jewels and a wish-fulfilling casket, creating armour protecting the yogins and
bringing them the attainments of siddhis.
His form (with nine heads?) represents the nine successive abidings in equilibrium, also not different
from the dharmat. The arms are also the eighteen dhtus; awareness of them the eighteen emptinesses.
Through eighteen primordial wisdoms, substantiality is destroyed, and this expresses means. Nine eyes look
no-where other than the dharmat; eighteen eyes look to the eighteen emptinesses, and so these express
wisdom.
Thus, means and wisdom are unified, so the eye, ear, nose, tongue, body and mind are seen as the
eighteen emptinesses; and eighteen times six makes one hundred and eight wrathful male deities. The objects
perceived by the wrathful male deities are the forms seen, and the sounds, smells, tastes, touchables and
mental dharmas, and their eighteen times six emptinesses, so the wrathful females also number one hundred
and eight.
Similarly, all dharmas are the male and female wrathful deities. The wings filled with vajras are raised
up, covering the worlds, and this means that primordial wisdom emanates from the dharmadhtu and
pervades everywhere. The first two right and left legs trample on male and female mahklas, the next two
trample on male and female mahdevas, the next pair on ganeas, and the final two on vinayakas. This
trampling of the eight worldly gods indicates the vow to mount upon the eight complete liberations, to
meditate on the four dhynas and the four equilibriums, and to realise the dharmat.
The root text continues: "His maala is of wrathful deities, with three heads, six arms, wings, and six
legs, upon vajra rock, with the stance of worldly protectors paying homage." The ten wrathful deities (khro
bo bcu) in the retinue of the r heruka are listed and described. Their names and some of their features are
consistent with other sources, and the list conforms to the usual sequence, apart from the deity of the above
direction, who is listed last instead of first. Although a standard set, there are some variations in the
appearance of the ten. In this case, they have similar colour schemes to the usual group given in most
Vajraklaya sources (such as Chapter 20 of the Myang ngan las 'das pa'i rgyud chen po in the NGB, see
Cantwell and Mayer 2007: 207-215), and also their first two hands similarly stir skull-cups of blood with
vajras. The specific implements given in their middle two hands are also generally the same or similar to
those given in Vajraklaya sources, while the final pair of hands are different since the Vajraklaya set simply

76

Textual Analysis

roll phur pas in their lower hands. Yet some of the extra implements found in the list here are present in the
Myang ngan las 'das pa'i rgyud chen po list. Then the names of the ten male attendant magical hybrid
deities ('phra/phra men) are given. They correspond exactly to the list of male magical hybrid deities to the
right of each wrathful deity in most Vajraklaya sources, each having the expected animal head.2 Here,
unlike the Vajraklaya sources, they wield corpses in their right hands, which they are about to eat, and in
their left hands, they carry weapons which can transform into anything needed.
All these deities have arisen from primordial wisdom as though as from the mind and the dharmas arisen
from the mind, and from primordial wisdom consecrations.
Chapter 13 The Glorious Vajra Rkass
The chapter concerns the wrathful female maala deities, who are said to have the same number of arms
and legs as their male counterparts, and they hold bells in their first right hands. The principal female is 'Jigs
byed chen mo (Bhairav), further called Ral pa gcig ma (Ekaja) in the commentary, who, however, is not
described. The set of ten female wrathful deities (khro mo bcu) have exactly the same names as those found
in the usual lists, such as in the Vajraklaya texts (see the discussion of the male set appearing in Chapter 12),
but the final four appear to be given out of order (see Appendix note 13). At this point, the Dunhuang
manuscript seems to have an omission, but the text given in the Tenjur version of the commentary is itself
perhaps a little garbled, and not all appropriate here. This text adds further goddesses and twelve inner
messengers, who are then referred to in the passage below in the Dunhuang text also. It also adds description
which is not altogether coherent. Then the list of female magical hybrid deities ('phra/phra men ma) is
given, in exactly the standard order, except that the lizard-headed, associated with the above direction, is in
most sources given at the top of the list, but here, in accordance with the order of the directions in Chapter
12, is given at the end of the list.
The root text then specifies that the student should cast a jewel or flower into this maala, the usual
procedure during empowerment, through which the appropriate deity for the student to focus on is
ascertained. The text adds that if the student then cultivates the deity wherever it lands, accomplishment will
follow. The student is shown the samaya of engaging in all the virtuous and non-virtuous dharmas.
Engaging in virtue is to be a rvaka, while engaging in non-virtue is to be a common sentient being. Here,
however, all dharmas are understood relatively as the male and female wrathful deities, and ultimately they
are the same, all equally unborn and unceasing. Then with (the master?) brandishing the vajra, (the student?)
accepts (the samaya?) and all the deities are beckoned and offerings made. At this point, the deities are
absorbed into one's body. The Lord Heruka pervades the whole body, Hkara is absorbed into the brain
centre, Vijaya into the forehead, Hayagrva into the region between the mouth and the chest, and so on, such
that all the ten wrathful deities are absorbed into different body parts. The principal deity or the chosen deity
becomes the whole body, and the messengers are absorbed into all the pores of the body hairs. The wrathful
female deities are then similarly absorbed. Through this primordial wisdom consecration, the spatial field
and the primordial wisdom should be unified in sameness within the essential dharmadhtu.
Glosses on the female magical hybrid deities are then given, in terms of their names or the animal
represented by their specific head, connected with features of the meditations transforming ordinary
experience. As a group, they are said to be known as magical hybrid deities ('phra/phra men ma), since their
2

The list of ten males and ten females are common in the Vajraklaya literature, including liturgies widely used in
contemporary times, such as the bDud 'joms gNam lcags spu gri cycle (see bDud 'joms Rin po che Volume Tha: 99-103), and
the Sa skya Phur chen (17v-21v). In the early Vajraklaya commentary known as the 'Bum nag, the imagery of the specific
animal heads is glossed in terms of metaphorical associations between the features of the animal and the specific wrathful
deity they accompany. For instance, the vivid variegated colouring (bkra ba) of the Dharma eye of the eastern wrathful deity,
Vijaya, is connected with the tiger-headed magical hybrid deity because of the tiger's stripes (bDud 'joms bka' ma edition,
Vol.Tha: 340.5-6; Boord 2002: 188).

A Summary of The Commentary

77

universal love is the spatial field consecration and with the primordial wisdom consecration, (they) are
consuming all material things. They are accomplished in the ordinary body, the spatial field as the body, and
the primordial wisdom as the limbs, and then they emanate from out of the pores of the skin. The root text
says that they go everywhere more swiftly than snapping the fingers, and the commentary explains that these
swift females, consecrated by the dharmadhtu, go everywhere without travelling since the spatial field
pervades everywhere. Further analogies follow, relating to the concealing female, the inciting female and the
killer female.
If the samaya should be broken, the female deities will cause the heads of the transgressors to burst. The
commentary explains that when the appropriate actions have not been accomplished, this transgresses against
the dharmadhtu and the spatial field's primordial wisdom unified in sameness. Thus, the transgressor will
come to circle around under the power of perverted thoughts. So this is said to be the samaya bursting the
head.
Chapter 14 The Mantras for the Wrathful Male Deities and for ritual action combined together
The mantra for casting the flower is given along with the invitation mantra (presumably, for the casting
and invitation as described in Chapter 13 above). Then the mantra for the principal deity (o badzra kro dha
shri he ru ka h/ a a a h pha/)3, is followed by a version of the standard mantras for the ten wrathful
male deities. Finally, mantras enjoining the four ritual actions, and the seed syllables of the five classes of
beings complete the chapter. As in Chapter 9, the commentary adds little, although again the Dunhuang
manuscript's annotations attach the names of the deities to their mantras.
Chapter 15 The Mantras for the Vajra Rkass
This short chapter gives the mantra for the female rkas (the archetype probably had o badzra kro dha
h, which some editions emend to o badzra kro dhe h) and the awareness mantras for the female
magical hybrid deities. There are two mantras to commission them; then syllables for dissolution, and for
the completion of ritual actions. The usual set of mantras for the ten wrathful female deities is not given.
Chapter 16 Mudrs
The opening point is similar to that made at the beginning of Chapter 10, in discussing the mudrs for the
peaceful deities. Here, it is explained that having understood the sameness of everything, all bodily
movements become mudrs. While Chapter 10 dealt with the hand gestures for the peaceful buddha family
deities, this chapter describes the full bodily postures needed to embody the various wrathful deities. First,
the positions to assume for the principal heruka deity are given, making the body like a vajra shape, and
becoming a single-spoked, a nine-spoked and an eleven-spoked vajra. The postures for each of the ten
wrathful male deities are then outlined, and these are followed by those for the magical hybrid deities as a
group. Once again, virtually all the text in this section is from TZ, and little elaboration is added by the
commentary, although the Dunhuang manuscript's annotations add a few clarifications, including the
appropriate mantras to accompany the movements of the wrathful deities (corresponding to their mantras as
given in Chapter 14).
Chapter 17 Ritual Actions
Invitations, offerings, and praises are made to the central Blood-drinker and the ten wrathful deities. The
chapter ends with a praise (omitted in the Dunhuang manuscript), giving the words for a praise of the hatred
vajra (zhe sdang rdo rje) and retinue, through whom all worldly realms are obliterated.

It is not entirely clear if this second phrase, a a a h pha/, is also intended as part of the principal male deity mantra, or if it
is a separate mantra.

78

Textual Analysis

Chapter 18 The Destructive Ritual

The series of chapters on the four rituals begins the sequence with the destructive ritual action (drag po'i
las), describing it in terms of the fierce actions of the vajra magical hybrid deities ('phra-men), seizing and
offering the evil spirits to the wrathful deities as food. The commentary supplies an inner interpretation, that
through the ritual, the viewing of sentient beings as substantial is consumed by its own true nature. The
active sense faculties are removed since in the context of the six primordial wisdom emptiness consecrations,
material things seen as solid wholes contradict their essential genuine nature. Thus, their qualities are made
to degenerate or they are naturally brought into sameness.
The magical hybrid deities are meditated upon as hindering the actions of and inciting divisions between
the objects of the rite, since material and one-sided accomplishment contradicts emptiness and sameness,
creates divisions within the tantric group and prevents ultimate accomplishment.
Chapter 19 The Destructive Homa
For the destructive fire ritual, the triangular hearth is prepared with three black phur pas and iron wire or
black rope. Thorny wood is used for the fuel, and the substances include poison, blood, black mustard,
realgar powder and iron filings. For the ritual, the magical hybrid deities are dispatched to summon the
objects who will burn up in the fire. The flesh mixed with the vegetarian burnt offerings is offered to all the
deities. Vajra songs are sung, and the offerings of the flesh, blood and bones of the elemental spirits are
consumed with delight by the deities. In conclusion, the commentary explains that the primordial wisdom
fire of buddha body, speech and mind burns up factors which are not conducive. With the striking arrows of
the primordial wisdom consecration, the deities delight in these factors, which naturally become conducive
to buddha body, speech and mind.
Chapter 20 The Phur pa for Destruction
The phur pa needed for destruction is described, made out of iron or black wood and featuring a head,
knot and three-sided blade. The commentary explains that Heruka and Ral pa gcig ma (Ekaja) are
meditated upon above the knot, and the male and female wrathful deities around the sides in appropriate
order. The rite of striking with it is said to transfix and take the life of the object, even if the object is a god.
Through the non-dual primordial wisdom emptiness consecration, the male and female wrathful deities strike
with the single phur pa of the nature of mind, and nothing is not permeated by this.
Chapter 21 The Destructive gTor ma
The recipe for making the destructive gtor ma is given, including iron filings, poison and gravel, mixed
with blood and mashed beans or black grains. In offering the gtor ma, a mere sight of it causes the vomiting
of blood. The commentary explains that this means that by meditating in thusness, with primordial wisdom
emptiness pervading everywhere, there will naturally be ruin and rot.
Chapter 22 The Destructive Ritual Union
The destructive ritual of union is described in terms of the male wrathful deity beating his hammer within
the female wrathful deity's mortar, thus beating the rite's object. The commentary explains that the
dharmadhtu nature is the mortar, with the spatial field of pure awareness the beating hammer. A concluding
remark on destructive rituals gives the instruction that black clothes should be worn for these practices.
Chapter 23 Conferring (transference of) Abode
The teaching on conferring the consciousness of the objects of the destructive ritual is described. The
doors of ordinary rebirth closed, the consciousness is taken up with the syllable h and elevated with the
syllable pha to the wombs of the female deity in union with the male deity in the place of Akaniha, thus
effecting liberation. The commentary adds meditations on each of the ordinary destinies to close the doors
in each case, involving recognition of the true empty nature of the emotional afflication associated with each
realm. For instance, the cause for sentient beings falling into hell is to experience hatred as though it has

A Summary of The Commentary

79

substantial reality. Recognising that hatred in fact has no real nature, the mind of hatred is not born, and this
closes the doors of hell. The same reasoning is applied to the other emotional afflictions and birth in the
other realms. The final meditation is of activating the male and female union of awareness within the
dharmadhtu spatial field, one's own belly as Akaniha, producing bodhicitta through which one's own
consciousness becomes accomplished.
Chapter 24 The Captivation Ritual
The chapters on captivating rituals are opened with an invitation to the male and female wrathful deities to
alight upon the eight spokes of the semi-circular maala, and then the making of offerings of songs, dance
and red items. The root text says that the vajra of passion, along with the retinue, captivates the objects of the
rite with passionate attraction, and as a result, all the worldly realms are entirely captivated.
The five types of garments are worn and the magical hybrid deities are sent out, summoning the objects of
the rite, and then they are gathered, along with their qualities of success and brilliance, by the Great Wrathful
deity, using the method of passionate attraction. They are thus captivated and will then do whatever is
desired. The commentary's interpretation is that the emptiness dharmadhtu pervades all, so outer and inner
things are indistinguishable. With the primordial wisdom of passionate awareness, one is aware that the outer
world is not different from oneself, so the outer is naturally attracted, revolving around oneself.
Since everything is oneself, it can be seen as, "the great self", with everything as one's own emanation.
The senses are transformed, and everything appears beautiful. With the awareness of this inseparability of
self and others, even the thoughtless become joyful and excited.
Chapter 25 The Captivation Homa
For the captivating fire ritual, the semi-circular hearth is prepared with five red phur pas and red boundary
rope. Fragrant smelling wood is used for the fuel, and the substances include copper filings, molasses, red
mustard and a drawing of the hearts of the rite's object, painted with shellac. In making the burnt offerings,
the vajra magical hybrid deities are meditated upon, burning up the thoughts of those who are unhelpful and
unfriendly towards you, and they will then obey you, and will start to think helpfully. The interpretation is
that the view regarding the self and all discursive thoughts arisen from it are burnt up in the primordial
wisdom emptiness fire. In the state of sameness, their great self-nature and single taste are said to cohere
together. Discursive thoughts are the great cause for sasra, so they are termed, unhelpful, and primordial
wisdom awareness without self, is termed, helpful thoughts, freed from sasra. Likewise, if there is
awareness that self and other are not different, others will also cohere together naturally, and come to think
helpfully.
Chapter 26 The Phur pa for Captivation
The phur pa needed for captivation is described, made out of copper or red-coloured wood and featuring
a head, knot and a semi-circular blade. The wrathful deities are visualised around the phur pa, and the rite of
striking with it captivates and brings everyone under control. The commentary explains that the pure
awareness's primordial wisdom phur pa will strike universally, and since there is no separation between
oneself and everything else, everything is naturally captivated.
Chapter 27 The Captivating gTor ma
The recipe for making the captivating gtor ma is given, including copper filings, molasses, mashed red
rice and other red grains. In offering the gtor ma, a mere sight of it causes the objects to become obedient.
If it is eaten, they will obey any commands whatsoever and become joyful too. The commentary explains
that since the dharmadhtu has no characteristic marks, it is all-pervasive. There is nothing to set apart the
great self-identity, and it will naturally captivate everything.

80

Textual Analysis

Chapter 28 The Captivating Ritual Union

The captivating ritual of union is described in terms of union of the maalas of the male and female
wrathful deities, which means that all will be captivated joyfully. The commentary explains that since the
dharmadhtu has no characteristic marks, it is all-pervasive, so the primoridial wisdom awareness unites with
it in non-duality. The unerring pure awareness of this non-dual state is bodhicitta, and it will naturally
captivate blissfully.
Chapter 29 The Increasing Ritual
The increasing rituals begin with an invitation to the deities to alight upon the eight spokes of the square
maala, and then the making of offerings of songs and yellow items. The root text says that the pride vajra,
along with the retinue, instantaneously increase the great successes and brilliance of the worldly realms.
The five types of garments are worn and the male and female wrathful deities, along with the magical
hybrid deities, are meditated upon as increasing the qualities of success and brilliance. If this meditation is
done, it will be actualised in a great self-identification. The commentary's interpretation is that the emptiness
dharmadhtu without characteristic marks pervades all, so the whole universe is indistinguishable from
oneself. Due to the primoridial wisdom awareness of this, all meritorious deeds, actions and so forth, are
oneself, or belong to oneself. If one is meditating on this, then it is like the sun arising in the skies, hot due
to the natural qualities of the sun, and also, since the heat itself is the sun, or the sun is light, so the light is
the same as the sun. All is a single identity, or the qualities belong to themselves.
Chapter 30 The Increasing Homa
For the increasing fire ritual, the square hearth is prepared with four yellow phur pas and yellow boundary
rope. The offered substances are the five types of jewels, the five medicines, the five grains, boiled rice pulp
and melted butter. In making the burnt offerings, the meditation is that all one's successes increase. The
interpretation is that with the dharmadhtu mirror-like primordial wisdom, the five male and the five female
principal deities arise as buddha body; the sixteen male and sixteen female bodhisattvas arise like the
buddha's major marks, and the five-fold diadems of each deity arise as the minor marks. The primordial
wisdom burns up material substance and it increases through its own natural qualities.
Chapter 31 The Phur pa for Increasing4
The phur pa needed for increasing is described, made out of gold or yellow-coloured wood and featuring a
head, knot and a four-sided blade. The deities are visualised above the knot, and the rite of striking with it
brings accomplishment and increases successes and brilliance. The commentary explains that the
sambhogakya is the samdhi phur pa, which increases successes and brilliance naturally.
Chapter 32 The Increasing gTor ma
The recipe for making the increasing gtor ma is given, including gold dust, mashed yellow rice grains,
milk and honey. In sending forth the gtor ma, successes and brilliance will increase and spread. The
commentary explains that since the dharmadhtu has no characteristic marks and is all-pervasive, with
nothing to set apart this great self-identity, it will naturally bring increase, and this is known as giving and
sending out primordial wisdom awareness.
Chapter 33 (numbered 31 in the Dunhuang Manuscript) The Increasing Ritual Union
The increasing ritual of union is described in terms of union of the maalas of the male and female
deities, and meditating on emanating bodhicitta. The commentary explains that if you meditate on taking
pleasure and increasing enjoyment, it will be accomplished accordingly. With union in the samdhi maala,
4

Note that Chapters 31 and 32 are missing from the Dunhuang manuscript, so the summary of these chapters relies on TZ and
the Tenjur version of the Commentary.

A Summary of The Commentary

81

pure awareness is the bodhicitta, so great enjoyment naturally increases.

Chapter 34 (numbered 32 in the Dunhuang Manuscript) The Pacifying Ritual
The pacifying rituals begin with an invitation to the deities to alight upon the eight spokes of the circular
maala, and then the offering of melodies and white offering gifts. The root text says that the delusion
vajra, along with the retinue, instantaneously pacify the vast extent of the worldly realms.
The five white garments are worn and the deities, along with their retinues, are meditated upon as
pacifying all disturbances, harm and evil. If this meditation is done, pacification will be instantaneous. The
commentary's interpretation is that the dharmadhtu emptiness samdhi of primordial wisdom pacification
will pacify everything by pervading all.
Chapter 35 (numbered 33 in the Dunhuang Manuscript) The Pacifying Homa
For the pacifying fire ritual, the circular hearth is prepared with six white phur pas. Silver filings and
white substances are offered. In making the burnt offerings, the meditation is that all disturbances are
pacified. The commentary's interpretation is that the dharmadhtu samdhi pacifies all discursive thoughts
through its own natural qualities, and the deities will also be venerated by this.
Chapter 36 (numbered 34 in the Dunhuang Manuscript) The Phur pa for Pacifying
The phur pa needed for pacifying is described, made out of silver or white-coloured wood and featuring a
head, knot and a round blade. The deities are visualised around the knot, and the rite of striking with it while
meditating on the deities pacifying would even pacify a god. The commentary explains that the elemental
nature's faultless essential pure awareness is the samdhi phur pa, so this pacifies everything naturally.
Chapter 37 (numbered 35 in the Dunhuang Manuscript) The Pacifying gTor ma
The recipe for making the pacifying gtor ma is given, including silver and mashed white rice grains etc.,
mixed with milk. In sending forth the gtor ma, even gods will be pacified. The commentary explains that
since the essential nature of the dharmat of all dharmas is to pacify, so primordial wisdom awareness
accordingly pacifies through pervading all material things.
Chapter 38 (numbered 36 in the Dunhuang Manuscript) The Pacifying Ritual Union
The pacifying ritual of union is described in terms of the accomplishment of the male and female deities,
generating majestic power and consecration, and pacifying everything. The commentary explains that the
dharmadhtu is in itself pacification, resembling the male and female deities in combining the dharmadhtu
and primordial wisdom awareness, and naturally pacifying. Thus, everything is pervaded with its
consecration, so everything becomes peaceful.
Chapter 39 (numbered 37 in the Dunhuang Manuscript) Summing up [all] the Ritual[s]
The root text says that the mudr which has no characteristic marks and the substantial mudrs arisen out
of it are explained by the warrior bodhisattva as unspecifiable, so the wise will be accomplished in doing
whatever they enjoy. The commentary adds that everything is the mudr without characteristic marks and
everything has arisen from the dharmakya. So if there is awareness of the flawless dharmakya, and the
yogin engages in any meditations, mudrs and marks of whatever type he likes, and commences the four
ritual actions, there must be accomplishment.
Chapter 40 (numbered 38 in the Dunhuang Manuscript) The Benefits of the Accomplishment
The root text says: "Through destructive [ritual], all are tamed; through captivation, the supreme union
is also accomplished. By increasing, [this accomplishment] is increased in the form of majestic splendour.
Through pacification, Supreme Bliss is also accomplished."

82

Textual Analysis

Through destructive performance, all gods and demons will be tamed. By captivating rituals, supreme
enjoyment will be experienced. Increasing rituals will stimulate successes and brilliance, while pacifying
rituals will bring attainment of supreme bliss. This is the great accumulation of merits, and the awareness
that there is nothing with characteristic marks is the great accumulation of primordial wisdom. Thus, the
four types of rituals should be continually engaged in, bringing this ultimate mahmudr into awareness.
Even if one deity is accomplished as explained above, the countless resultant siddhis will be inconceivable.
Chapter 41/42 (final unnumbered chapter in the Dunhuang Manuscript, which together with the Tshal pa
Kanjur and Bhutanese tradition, loses the title to Chapter 41 and thus the chapter division)
Chapter 41 (= Final Chapter Part 1 in the Dunhuang Manuscript), The Maala of the Great
Captivator)
The characteristics of the maala are that it is the maala of the Great Captivator (dbang chen bsdus
pa), and within it, the circular base indicates that the dharmadhtu is unfabricated. Abiding within the
circular base, made as a wheel with angled spokes, signifies that within the unfabricated dharmadhtu, the
consecrated deities proceed from its very nature. Made with four huge corners, which bring the entire
inanimate and animate universe under control, signifies increasing great bliss. From the centre, close to the
spatial field, the female deities are looking outwards, signifying seeing, the bodhisattva's approach to the
spatial field. The five male bodhisattvas looking inwards, signify viewing the approach to the spatial field.
Further glosses are given to the directions in which the other deities are looking, some looking inwards and
some outwards. The male and female wrathful deities as a whole look outwards, signifying the vanquishing
and taming of outer material substances and adversaries.
The spatial field resembles the maala of the sun. Primordial wisdom knowledge is clearly luminous
like the sun's light; its consecration is like the sun's warmth and heat. Its vanquishing of outer material
substances and so on is like the sun drying out and burning up objects.
The female deities signify pure awareness of the dharmadhtu, within the spatial field, while the male
deities spontaneously vanquish material substance and adversaries, and by their unification, the male and
female wrathful ones activate the four ritual actions and protect the world with compassion. The protection
of the world is demonstrated by the worldly gods and ngas, gathered into the retinue.
The eight auspicious symbols signify captivating everyone. They show the auspiciousness of remaining
unaffected by the bonds of sasra, even while engaging with it, and of remaining unaffected by the
rvaka's bias towards peace, even while engaging with nirva. Thus, this is the auspiciousness of not
being overawed by anything. The glorious endless knot is the auspiciousness of being endowed with
primordial wisdom knowledge. The wheel represents captivating all, and the auspiciousness of turning the
wheel of the dharma. The jewel is the auspiciousness of being endowed with inexhaustible treasure. The
lotus is the auspiciousness of not being in sasra's thrall. The parasol is the auspiciousness of being
untormented by the afflictions. The flask is the auspiciousness of being endowed with the supreme taste of
elixir. The conch is the auspiciousness of proclaiming the sounds which intimidate all adversaries. The fish
is the auspiciousness of satisfying the stream of all worlds and sentient beings of the five classes, with great
compassion.
The four continents consecrated by the five families demonstrate to sentient beings of the five classes that
they have the natural qualities of the tathgatas. This is also the affirmation that their benefit is
accomplished through compassion.
In this, the male and female principal deities, each with a fivefold retinue, do nothing whatsoever. The
twenty male and female bodhisattvas treat everyone as an only child, and through their love, together with
the male and female wrathful deities, they see all outer and inner material things like a hunter seeing game

A Summary of The Commentary

83

animals. By the force of such aversion, everything is unified in the spatial field itself. In this union, even
while engaging in the dharmas of sasra, there is great enjoyment since one is unaffected by the bonds of
the emotional afflictions, so this is the accomplishment of desire. Through the shining of the natural
condition of the spatial field of the dharmat, there is jealousy. Since the dharmat is unfabricated, this is
delusion. The dharmat is victorious over the three realms, so this is pride.
The consecrations of the five retinues of each of the five buddhas are located at the five parts of the body
(see Chapter 7 summary above). The male and female wrathful deities and their retinues share the
consecrations of their specific buddha families. Deeds involving solidifying are the work of the vajra family.
Those involving heating relate to the jewel family; moistening is connected with the lotus family, and
moving with the action family. The hallmark of the tathgata family is that these deeds are universally
accomplished from complete equilibrium.
The absence of inherent existence in solidifying, heating and the other deeds listed above constitutes the
consecrations of the five female buddhas and their retinues. The ability to perform action is the sign of the
male bodhisattvas, while the absence of inherent existence in this ability is the female bodhisattvas.
Likewise, pure awareness is the male and female wrathful deities. Although everything is absorbing and
radiating, in all outer and inner dharmas, these features are similarly to be seen.
Chapter 42 (= Final Chapter Part 2 in the Dunhuang Manuscript), with various titles, The Conclusion
or Ultimate Perfection in the South Central and local Kanjur root text versions; Praising the Wondrous
in the Bhutanese version; Synopsis of the King of Reflections in the Tenjur version; untitled in the
Dunhuang and Tshal pa Kanjur versions)
Vairocana addresses the Victorious One with a series of praises of the features and qualities of the
wrathful deities. For instance, their dark russet brown locks of hair coil upwards, to reach the summit of
existence, while their scowling eyes are taming perverted views. Baring their vajra fangs causes the root of
birth and death to be cut, and their body, speech and mind consecrations appear in their various implements.
Their loving rays of light flash like lightning and awesome voices roar like thunder. Moreover, the
consecration of great compassion activates the consecration within oneself, and these deities' attributes are
the primordial wisdom which transcends concepts. Their wrath has the natural condition of sameness and
generates enjoyments, in its vajra absorption and radiation. These verses of praise end the root text, with a
final visualisation of the maala being absorbed within the heart. The commentary adds a slightly cryptic
passage, perhaps first saying that the emergence from out of sameness means that when a noble being in pure
awareness has produced sounds articulated in speech, these are called, tantra, and the vajra wheel is thus
turned in Akaniha. Relying on this, the Buddha/Protector's body, speech and mind secrets are yogically
accomplished, so this (the speech, tantra or the Buddha's three secrets) is considered flawless. There is a
final verse praising Padma rgyal po, who unravels the great secret pith teachings from the expanse.

f) The "Citations" or Attributions of the Teachings in TZComm to other Tantras

The apparent citations of various tantras in TZComm (sometimes with names expanded further in the
Dunhuang manuscript's annotations) are of interest because they add significantly to our knowledge of which
Tantric text titles were current and considered authoritative in the post-Imperial period. They possibly might
also indicate a possible terminus ante quem for a number of rNying ma Mahyoga scriptural titles (but as we
show below this issue is complicated by various anomalies). Some of the titles are cited more than once, so
that in total, there are at least forty separate citations which are attributed to at least twenty-six different
named titles, depending on how one counts: some of the citations might be using different names for the
same text, while others might refer to collections rather than texts.1
The dPal mchog dang po is cited in Chapter 1, the Glang po or Glang po che in Chapters 3 and 25
(clarified in the annotations to Chapter 3 as the Glang po rab 'bog), the gSang ba'i rgyud in Chapter 6, and
Chapter 7's annotations seem to refer to a text called the Glog sgra bas brtsegs pa. There is a reference to a
text called the Glog gru in Chapter 8, and one to the Sa rba 'bu ta 'i don phyi ma, and also in Chapter 8, one
verse from the sGron ma brtsegs pa is cited, and a following verse from the rDo rje sems dpa'i le'u. Two
citations of the gCig las spros pa'i tan tra occur in Chapter 12; the Ri bo brtsegs pa is cited in Chapters 13,
19, 24 and 33; the dPal 'phreng dkar po in Chapters 13, 29 and 39 (which the Dunhuang manuscript labels
37); the rGyal po'i rtog pa in Chapter 16; the Kar ma ma le is cited three times in Chapter 18 and possibly
again in Chapters 22 and 34 (labelled 32 in the Dunhuang manuscript), where a text called the Las kyi
'phreng ba (Thabs kyi 'phreng ba in the Tenjur's Chapter 22) is referred to. The 'Gu hya (Gu hya ti la ka in
the Tenjur) and Ki la ya bcu gnyis kyi tan tra are cited in Chapter 20; the sGyu 'phrul dra ba is cited in
Chapters 21, 32 (omitted in the Dunhuang manuscript), 37 and 39 (labelled 35 and 37 respectively in the
Dunhuang manuscript); the Tan tra snying rje rol pa is cited twice in Chapter 23; the dBang c[h]en bsdus
pa'i tan tra is cited in Chapters 24 and 28; the Phur pa bcu gnyis kyi rgyud phyi ma is cited in Chapter 26.
The Zhags pa'i rgyud phyi ma gnyis (given as Thabs kyi zhags pa'i rgyud phyi ma gnyis in the Tenjur) are
referenced in Chapter 27; the Thabs kyi zhags pa itself is cited in Chapters 30 and 39 (labelled 37 in the
Dunhuang manuscript). Of course, the entire text of the root Thabs kyi zhags pa is cited and commented on
throughout the manuscript. However, at this juncture, and quite independently of the ongoing commentarial
progress through the root text, TZ's own title is mentioned in support of a particular view that is being
expounded in the commentarial glosses. The Tan tra Glang po chur 'jug is cited in Chapter 35 (labelled 33 in
the Dunhuang manuscript);2 the rTse gcig bsdus pa is cited in Chapter 36 (labelled 34); the rDo rje bkod pa
is cited in Chapter 38 (labelled 36). There are possibly a number of references in Chapters 8, 23, 31
(which is, however, omitted in the Dunhuang text), and two references in Chapter 40, labelled 38 to a
tantra called, Du ma, although this most probably simply means, many tantras. The first reference in Chapter
40 reads, ma ha yo ga'i tan tra du ma, which probably means, many mahyoga tantras, but may alternatively
mean, the mahyoga tantra (called), the Du ma. There is a further reference to the Tan tra sde lnga (the
Tenjur version of the commentary gives this as, Tan tra du ma sde lnga) in Chapter 39 (labelled 37 in the
Dunhuang manuscript), but this is quite likely to be a classificatory term referring to the three specified
mentioned tantras along with others, suggesting that many other texts of different classes of tantras share the
same teaching.3
1

There is also the question of whether or not to count the references in the Tenjur version of TZComm in chapters omitted in the
Dunhuang text. Of course, these chapters were certainly once present and omitted by scribal error from the Dunhuang version,
but we cannot use them in quite the same way as evidence for the post-Imperial period.
Klong chen pa, in the sNgags kyi spyi don tshangs dbyangs 'brug sgra, classifies the Glang po chur 'jug as the speech aspect
of buddha body, while the Glang po rab 'bog (as mentioned above, also referred to in TZComm simply as the Glang po,
assuming the identification given in the annotations to be correct) is the body aspect of buddha body (Dorje 1988: 33-34).
Certainty is illusive: it is possible that Tan tra sde lnga is intended to refer to a specific title rather than a category. For
instance, the NGB (Rig 'dzin edition Volume Cha: 35r-41r; mTshams brag Volume Tha: 527-543) has a text entitled, sPros pa
gcod pa sde lnga'i rgyud.

The "Citations"

85

Some of these titles are not easily identifiable from the existing titles of the rNying ma canons; we know
of no Glog gru, for instance. Nonetheless, most of the names are familiar from the traditional literature. For
example, although we have not found an extant rTse gcig bsdus pa, dPa' bo gtsug lag 'phreng ba (1504-1566)
includes this title in his list of the Eighteen Mahyoga Tantras (Dorje 1988: 34). Likewise, he gives a Du ma
'phro pa, possibly to be equated with the Du ma cited in TZComm, if this is not simply a general statement
about many texts containing the teachings referred to. Similarly, the title dPal 'phreng dkar po is found in
Klong chen pa's list, even though it has not been found so far amongst the extant NGB texts (Dorje 1988: 3334). Identification of the Sa rba 'bu ta 'i don phyi ma is not certain: many texts have Sarva buddha in their
title, including some NGB texts, such as the Sangs rgyas thams cad dang mnyam par sbyor ba series (four
texts are given in Volume Ma of the Rig 'dzin edition).
Most of the other cited titles are highly suggestive of famous Mahyoga tantras still extant and available
in the transmitted rNying ma canon: the dPal mchog dang po, the Glang po rab 'bog, the Ri bo brtsegs pa,
the Kar ma ma le, the K la ya bcu gnyis, the sNying rje rol pa and the Thabs kyi zhags pa. It is significant
that all of these occur in the NGB's core section for Mahyoga, known as the Eighteen Tantras of Mahyoga.
Other titles found here, such as sGyu 'phrul dra ba, are highly suggestive of other tantras of this same
grouping. We should also mention that other Dunhuang texts mention yet more titles from the Eighteen
Tantra section: a Zla gsang thig le is listed in PT 849, and a part of it said to survive as PT 281; a Sangs
rgyas mnyam sbyor is listed in IOL Tib J 454 and in PT 849 this may possibly also be indicated by
TZComm's Chapter 8, as noted above; a Ri bo brtsegs pa and dPal mchog dang po are cited in IOL Tib J 454,
as well as in TZComm; the gSang ba'i snying po is said to be cited in PT 332b and is listed in PT 849; and the
gSang ba 'dus pa survives as IOL Tib J 438.4 One of the mentioned titles in TZComm, the sGron ma brtsegs
pa, may possibly indicate the bDud rtsi bcud thig (or bcud bsdus) sgron ma brtsegs pa'i rgyud, which is
classified in the Man ngag sde spyi ti cycle of rDzogs chen tantras in the NGB. Perhaps the 'Gu hya/Gu hya
ti la ka might indicate the Thig le gsang ba'i brda' rgyud or the Thig le gsang ba de kho na nyid nges pa'i
rgyud, both of which are in the Man ngag sde gsang ba cycle of rDzogs chen tantras in the NGB.
Unfortunately, however, the situation is not without complications. It is true that some of the Dunhuang
Mahyoga textual survivals, such as TZ, TZComm and the gSang ba 'dus pa, are demonstrably more or less
the same texts as those extant today under the same titles. Yet this is by no means so clearly the case with
what appear to be presented as citations in TZComm. Although we have not been able to complete a
comprehensive search, we have attempted to search for them all, and we have not found any of the citations
in the extant versions of the titles it cites.5 We have also noticed other discrepancies: for example, a Phur pa
bcu gnyis phyi ma is mentioned, yet none of the three different extant Phur pa bcu gnyis texts has a phyi ma,
and similarly, TZ itself also has no phyi ma, although two are apparently referred to in Chapter 27.
In fact, the long series of citations at the ends of chapters in TZComm, when looked at collectively, seem
to form a single coherent thematic literary progression, and these ostensible quotations do not seem
necessarily to represent separate short passages taken from many different texts. We even find repetitions of
the same turns of phrase in a number of subsequent passages, as exactly the same points are applied to one
ritual activity after another. Thus, for instance, the point is made in Chapter 27's ostensible citation that the
dharmadhtu has no characteristic marks, so it is all pervading, and its own natural qualities will bring
captivation. The point is reiterated in Chapter 28 with some further elaboration and a slightly different
wording, emphasising that pure awareness will bring captivation through its natural qualities. Then again,
4

In collaboration with Sam van Schaik, we are making a database of such citations; these references are noted in this work-inprogress. Our thanks to him for the locations in IOL Tib J 454.
rNying ma tantras can vary considerably between different editions. The search cannot be considered complete until the
various editions have all been carefully examined. Such tasks will become possible when searchable electronic editions of the
texts are available. We have had to rely on our own browsing of one or two versions of each text and quite possibly missed
some citations.

86

Textual Analysis

Chapter 29's ostensible citation repeats the message of the dharmadhtu as lacking characteristic marks, this
time being said to pervade outer and inner material things, so primordial wisdom awareness brings
identification with positive qualities, like the sun's natural quality of heat. Chapter 30 speaks of the
dharmadhtu's primordial wisdom bringing increase through its own natural qualities, and the following
chapters continue in the same vein, following the chapters on increasing with pacifying, for instance, through
the natural qualities of awareness mentioned in Chapter 35 (numbered 33 in the Dunhuang manuscript).
Thus, we find virtually the same phrasing is repeated across different chapters, with some development of
slightly different nuances relating to the specific rituals under discussion, but on each occasion, the teaching
is attributed to quite different texts. The chances of so many different tantras repeating these same phrases in
a way that so exactly conforms to the message of this commentarial text would seem rather low. Most
tellingly, the commentator cites TZ itself in this series of citations, at the end of Chapter 30 yet his entire
work is in fact a word by word commentary on TZ and the words he attributes to the Thabs zhags in the
chapter-ending citation do not occur anywhere in the actual text, neither as the commentator presents it, nor
even in the accidentally omitted verses of root text! Hence we might suspect that these are not citations at all
in the usually accepted sense: perhaps, the commentator himself simply composed these passages, or else
adopted them all from a single existent text, with the dual intentions that each one sums up the
Commentary's gloss on the chapter in question, whilst simultaneously affirming these interpretations to
conform with the teachings of other famous scriptures.
In most cases, the references follow a commentarial gloss, using such wording as, zhes... las 'byung ngo
([what this] says is given in...), or, las bshad de/do (explained in.... ). It is probable that such phrasing is not
intended to indicate direct quotations from the scriptures it mentions, but is rather suggesting a similarity in
these sources and the teachings which it is seeking to propagate. Only a few of the references mark off the
citations clearly, giving the name of the scripture first, followed by the passage, which is then closed by
wording such as, zhes gsungs so.6 The passage in Chapter 30 (Dunhuang Ms 73v) which refers to the Thabs
zhags itself does appear to be referring back to the discussion in the Commentary's Chapter 5, relating to the
natural arising of the five male and female buddhas as the buddha body and the retinues as the major and
minor buddha marks. It seems the commentator is suggesting that the teaching given earlier in the text
implies the kind of increase now discussed in this chapter on the increasing burnt offerings ritual.

Such clear marking is found only for the dPal mchog dang po in Chapter 1, for the Glang po in Chapter 3, the sGron ma
brtsegs pa and rDo rje sems dpa'i le'u in Chapter 8, and for the Ri bo brtsegs pa in Chapter 13. However, even if these cases,
we have not identified the citations given.

PADMASAMBHAVA, THE THABS ZHAGS AND ITS COMMENTARY

In this section, we explore the question of what Padmasambhava may or may not have had to do with TZ
and TZComm. This question is raised by a four line verse of homage at the close of TZComm, as well as the
mention of Padmasambhava three times in the set of marginal annotations in small writing which occur
throughout the Dunhuang manuscript. Other scholars, following Kenneth Eastman's short discussion of this
text in the 1980's (Eastman 1983: 49-52), have assumed they represent Padmasambhava as author of
TZComm,1 but as we show below, the references are not unambiguous. For textual reasons which we
explain above (see Textual Analysis, Section b, p.32-33 above), it seems clear that these annotations do not
merely represent the reflections of a single reader of the manuscript. Rather, it is probable that they had been
written on the exemplar of the text which was copied by the Dunhuang scribe, and that they had thus been
copied along with the main text at least once. We might therefore perhaps consider them to reflect an
ancient teaching lineage on TZ, even if these teachings are no longer replicated in any other extant source.
Here we are concerned with only one characteristic of the content of these comments on the text, that is,
three somewhat enigmatic references to Padmasambhava, one right at the beginning of the text when
explaining the title, and two near the end. Unfortunately for us, the annotator makes no clear statement about
Padmasambhava's role in the text, rather as though he assumes the reader already knows this, and the point
of the annotations is simply to draw attention to some aspect of that role. Put together, the three comments
would seem to indicate TZComm, or perhaps TZ, as the work of Padmasambhava or at the least, as
representing his teachings, or as having some association with him.
Padmasambhava's Pith Instructions on the Garland of Views
As well as our analysis of TZ and TZComm, we made a comparison between the doctrines, language and
terminology of these compositionsand in particular TZCommwith another genuinely early work
attributed to Padmasambhava, the Man ngag lta phreng (Pith Instructions on the Garland of Views,
henceforth MTph), to see whether any obvious links between the two approaches can be discerned. In brief,
we can say simply that there are some similarities, but the style and content of these different kinds of texts
do not provide us with close parallels. MTph aims to provide a comprehensive classification of views, with
special attention given to the tantric path, while TZ and TZComm give a more limited classification of the
variety of paths, focusing primarily on the single Vajrayna perspective they represent. Moreover, TZComm
says little about stages through which the practitioner should progress, since it is principally concerned with
expressing and elaborating on its vision of the ultimate sameness of the dharmas of sasra and nirva, and
the total purity of the defiled dharmas as the enlightened body, speech and mind of the tantric deities.
Specifically, MTph's initial categorisation of erroneous views is not parallelled in TZComm, although
ordinary beings (sems can phal ba) are mentioned in TZComm's Chapter 13. Both texts start with a twofold
classification of the path into the Vehicle of Characteristics (mtshan nyid kyi theg pa) and the Vajra Vehicle
(rdo rje'i theg pa), but the way they describe these is rather different. In MTph, the Vehicle of
Characteristics is subdivided into the ravaka, pratyekabuddha, and bodhisattva paths, while for TZComm,
1

Eastman himself expresses some caution, finally concluding, "It appears... that we have one of the few surviving works of
Padmasambhava" (1983: 50, my emphasis). Jacob Dalton (2004: 763 note 17), states rather more positively that, "in the
interlinear notes to the Dunhuang versions of the Thabs kyi zhags pa pad mo'i 'phreng ba commentary (ITJ321), the
commentary is attributed to Padmasambhava". An article from Sam van Schaik (2008: 47) also states that, "the Dunhuang Ms
IOL Tib J 321 contains a colophon which states that Padmasambhava was the author of the commentary". However, van
Schaik reassesses the evidence in his blog (dated June 2007 but presumably written after the article), where he no longer refers
to a colophon and writes, "Finally, just in case I have given the impression that Padmasambhava actually wrote this manuscipt,
let me be clear that he didn't". However, it seems that he simply means that the manuscript is no autograph copy, since he
continues to speak of , "the attribution of this text to Padmasambhava", and interprets one of the annotations in this way (see
p.96, note 16 below).

88

Padmasambhava, the Thabs zhags and its Commentary

the twofold division is simply that of Mahyna in contrast to Vajrayna (Dunhuang ms.1r.2-4), so that in
TZComm, the Vehicle of Characteristics relates to the bodhisattva path. This includes, for example, the ten
pramits, regarding which TZComm mentions the associated long time scales (gradual accomplishment)
and difficulties (various ascetic practices). Pratyekabuddhas are only briefly mentioned in the Dunhuang
manuscript's marginal notes, and nowhere in the main body of TZComm. rvakas occur in TZComm's
Chapters 1, 13,2 and 41, where TZComm on each occasion contrasts the tantric perspective with that of the
Hearers.3 MTph's characterisation of the bodhisattva path does not differ materially from TZComm (both
mention the ten pramits), but rather than detailing its inadequacies, MTph's emphasis is instead on the
bodhisattva perspective itself, and how it understands the lack of any inherent nature (rang bzhin) ultimately,
while particular characteristic marks identified at the relative level are to be seen as illusory.
MTph divides the Vajra Vehicle into different tantra classes; TZComm does not. TZComm's Vajra Vehicle
the Ultra Great Vehicle of the Great Vehicle (theg pa chen po['i] yang theg pa chen po) is similar to
MTph's inner yoga method tantra (rnal 'byor nang pa thabs kyi rgyud). But TZComm does not have MTph's
division into the techniques of Generation, Completion (or Perfection), and Great Completion/Perfection
(bskyed pa'i tshul, rdzogs pa'i tshul, rdzogs chen po'i tshul). None of these three terms (nor the standard
categories used in the tradition of the Generation and Completion Stages (bskyed rim, rdzogs rim) are
mentioned in TZComm, even if the teachings they represent may be implicitly present.
For MTph, the maala is gradually established by the three samdhis, using the Generation technique. In
TZComm, the three samdhis are not discussed.4 In general in TZComm, there does not seem to be a sense of
the maala building up gradually it appears to be complete from the outset.
MTph's Completion technique is classified into ultimate and relative. The ultimate corresponds in part to
TZComm's fourth type of sameness, as given in Chapter 1: ultimate sameness in the birthless and deathless
mode (don dam par skye 'gag med pa'i [Gt par] tshul du mnyam pa'o). This follows a listing of three other
types, relating to 1. the outer world's five objects and Tathgatas; 2. the inner mental consciousness(es) and
five primordial wisdoms; 3. the senses and (their) objects on the relative (level) and (their) mode as gods and
goddesses. This fourth kind of sameness is also referred to in Chapter 13 (given in error within Chapter 12
in the Dunhuang manuscript), the sameness of all dharmas, ultimately equally unborn and unceasing. MTph
mentions its ultimate Completion technique in relation to the male and female deities, ultimately unborn and
unceasing (rdzogs pa'i tshul ni don dam par skye 'gag med pa'i lha dang lha mo). It adds that the nonconceptual ultimate middle way, is to be unmoving from the dharmadhtu. TZComm includes a brief mention
of the ultimate middle way (don dbu ma) once in Chapter 7, in relation to a meditation on seed syllables.
MTph's relative Completion technique is to meditate clearly on the Noble One's Rpakya ('phags pa'i
gzugs kyi sku). This term is not used in TZComm. Then MTph speaks of being accomplished through
meditating in sameness without any adulteration (mnyam la ma 'dres par bsgom pas 'grub), which does
sound reminiscent of TZComm's approach, but the words, without any adulteration (ma 'dres par), are not
used in TZComm.
MTph's Great Completion/Perfection technique (rdzogs chen po'i tshul) seems in keeping with
TZComm's ethos throughout. The second verse of the root text TZ's Chapter 6 is similar to MTph's
expression here, stating that all worldly and transcendent dharmas are indistinguishable, primordially the
2
3

The relevant passage is mistakenly placed within Chapter 12 in the Dunhuang manuscript.
Here, TZComm explains that the Vajrayna is not gradual and not biased towards the peace of nirva, and that its ultimate
and relative truths are summed up in the doctrine of sameness, which is quite unlike the rvaka view. rvakas are virtuous
and unlike ordinary beings, but the view of the sameness of all dharmas is superior.
There is one reference to the three samdhis in the annotations to the Dunhuang manuscript (Ms 66r: (it is) the three samdhis
which activate the union of the male and female deities), but it is not altogether clear whether or not the reference refers to the
three as usually explained, and in any case, they do not occur in the main text of TZComm.

Pith Instructions on the Garland of Views

89

maala of body, speech and mind. MTph supplies a tantra citation not in the root text TZ but similar in
approach equating the vajra skandha limbs with the five buddhas. TZComm's Chapter 41 (Ms 81v-82r)
outlines a series of associations, the Buddha Akobhya as the right hand or arm and the right sides of the
ribs, and so forth. TZComm's Chapter 6 expresses a similar view, with associations made between the
elements and the wisdom deities. However, TZComm gives the elements as the male deities, space as
Vairocana and so on, while MTph associates the elements with the five consorts. But then TZComm adds
that the five faculties are the five male consorts, while the five objects are the five female consorts. Overall,
then, the sentiment of this section in MTph is exactly the same as TZComm, but the specific associations
vary. Similarly, TZComm would be in entire agreement with MTph's statement that the three planes of
existence are totally pure, although the exact point is not parallelled in TZComm.
Apart from four inner goddesses mentioned Chapter 6, TZComm does not group the deities into sets of
fours in the way that MTph does. Unlike MTph, Samantabhadra and Samantabhadr are not a prominent pair
in TZComm. Kun tu bzang mo occurs in the list within TZComm Chapter 5 as the first of the female consorts,
and in Chapter 6 as one of the female deities (all touch is Samantabhadr), and also in the list of mudrs in
TZ Chapter 10, but neither kun tu bzang po nor kun tu bzang mo are given explicitly in the list of deities of
the peaceful maala of TZComm's Chapter 7, nor in the mantras in TZ's Chapter 9. In any case, the few
references to kun tu bzang mo do not amount to MTph's assertion of Samantabhadra as the nature of the
mental consciousness, bodhicitta, and Samantabhadr as the nature of conditioned and unconditioned objects.
There is a similar focus on finding Enlightenment by looking into one's own mind. TZComm's Chapter 3,
with reference to the Glang po, states: Since the buddha is the aware nature of mind, Do not seek the buddha
anywhere else.5 TZComm, however, has nothing quite like MTph's citation that all dharmas are in the mind,
mind dwells in space, and space no-where. There is also no close parallel in TZComm to MTph's citation that
all dharmas are empty in their essential nature, that they are primordially totally pure, entirely clear light,
nirva in nature, promordially the completely perfected buddhahood. But clearly, this is entirely in line
with TZComm's teaching.
MTph devotes a lengthy section to a presentation of the methods involved in the Great
Completion/Perfection. Here, there is an emphasis on yid ches (conviction or trust). The word does not even
occur in TZComm. Apart from a couple of references to paying respect and making offerings to the lama
(one of which is only found in the Dunhuang manuscript's annotations), there is little emphasis on devotion
or faith even empowerment is emphasised as taking place through one's own pure awareness. Of course,
MTph is stressing the gaining of conviction through one's own experience, but there is the suggestion that it
is through conviction or faithful believing in the depths of one's mind, that the realisation comes.
MTph gives a classification of four aspects of understanding (rtogs pa): 1. understanding of the single
cause (rgyu gcig par rtogs pa); 2. understanding the technque using the seed syllables (yig 'bru'i tshul rtogs
pa); 3. understanding through consecration (byin gyis rlob kyis rtogs pa); 4. direct realisation (mngon sum
par rtogs pa). TZComm does not use this classification, but its content could be seen as consistent with it.
For example, the single cause of all phenomena could be said to be expressed in TZComm's Chapter 1, that
the cause is the ultimate truth, all dharmas arising without characteristics, while the relative truth of them
appearing like an illusion is the result. For the second aspect, TZComm's Chapter 7 gives associations of
mantra syllables in a similar although not identical way to that of MTph, and much more extensively than the
MTph's short passage. There is nothing in TZComm like the imagery of dyeing the cloth which MTph uses to
illustrate the tantric transformation through consecration, but the concept of consecration (byin rlabs) recurs
throughout TZComm, and the focus on both the real empty nature and the wisdom manifestation. Direct
realisation is clearly expressed throughout TZComm and is fundamental to its vision.
5

rig pa'i sems nyid sangs rgyas te/ sangs rgyas gzhan du ma tshol cig. By comparison, in MTph we find: 'dus byas dang 'dus
ma byas pa'i chos thams cad rang gi sems las gud na med... rang sems so sor rtogs pa ni/ sangs rgyas byang chub de nyid do/

90

Padmasambhava, the Thabs zhags and its Commentary

In MTph, ultimate perfection in these four aspects takes place through the Three Characteristic marks
(mtshan nyid gsum).6 TZComm does not refer to these characteristic marks. They are mentioned briefly by
the Dunhuang manuscript's annotator in Chapter 1.7 The annotator's language is clearly in line with MTph's
list, which speaks similarly of a progression through the set of three. However, there is nothing in the root
text TZ or TZComm resembling MTph's elaboration using a further set of 1. connections/union ('brel ba); 2.
requirements (dgos pa); and 3. the ultimate requirements of the requirements (dgos pa'i yang dgos pa).
Nonetheless, although not expressed in these terms, TZComm's perspective is entirely in line with MTph's
discussion of these features. In TZComm's Chapter 1, the Vajrayna is described in terms of sameness, and
the result of buddhahood mentioned as free from accepting and rejecting; MTph defines requirements,
through the characteristic mark of engagement, in a similar manner (thams cad ye nas sangs rgyas pa'i
mnyam pa chen po la blang dor med par spyod pa ni 'jugs pa'i mtshan nyid do). MTph speaks of the ultimate
requirements in terms of the actualisation of the endless embellished wheel of enlightened body, speech and
mind (sku gsung thugs mi zad pa rgyan [brgyan] gyi 'khor lo mngon sum gyur pa). TZComm does not use
these terms, but clearly, the entire text and commentary is developing such a vision.
MTph divides the path into striving in the fourfold practices of Approach, Close Approach,
Accomplishment and Great Accomplishment. This classic yogic progression is not referred to in TZComm.
There is little mention of striving in TZComm, so for instance, in Chapter 1, there is a reference to perfecting
effortless diligence ('bad pa med pa'i brtson 'grus). The content of these MTph practices, however, is not
dissimilar from the Thabs zhags in approach, although again, the specifics are different. Approach is defined
in terms of applying the antidote so that dharmas are realised as uncontrived, primordially in the nature of
buddhahood. Here, we witness wording which does not occur in TZ or TZComm (gnyen pos bcos su med
par..) but clearly, the practice is in keeping with TZComm. Close Approach relates to realising oneself as the
deity primordially, a theme which runs throughout TZComm's maala descriptions. Accomplishment and
Great Accomplishment are connected with the consort practice, and are a little reminiscent of imagery given
in Chapters 6-8 of TZComm, although again, the specifics vary.
MTph follows the outline of these stages with a section on entering the maala of the Great Perfection.
Here, there is an elaboration of the stages of the empowerment ritual which interprets their significance in
terms of the appropriate meditative understandings associated with them. Although all the meanings are
inner glosses and not very different from the ethos of the Thabs zhags texts, in TZComm, even the
empowerment rite remains unelaborated. The emphasis in Chapter 3 is on empowerment through one's own
natural awareness, while the discussion on entering the maala in Chapter 13 stresses understanding unborn
and unceasing sameness. MTph describes the final stage of the Great Perfection as the Wheel of Syllables at
the stage of the Great Accumulation (yi ge 'khor lo tshogs chen gyi sa). This is a widely used term for the
culmination of the inner Vajrayna path, often described as the thirteenth bhmi. The Wheel of Syllables is
referred to in Chapter 5 of the root text TZ,8 in the context of the natural expression of the samdhi maala.
6

These categories remain very much a part of contemporary Mahyoga exegesis: the late Dudjom Rinpoche, for example,
analysed them in his survey of rNying ma doctrine, the bsTan pa'i rnam gzhag, citing verbatim from MTph: rtogs pa rnam pa
bzhi'i tshul rig pa ni shes pa'i mtshan nyid (Awareness (through) the methods of the four aspects of understanding is the
characteristic mark of knowledge); yang nas yang du goms par byed pa ni 'jug pa'i mtshan nyid (repeated familiarity with this
is the characteristic mark of engagement); goms pa'i mthus mngon du gyur ba ni 'bras bu'i mtshan nyid (and direct experience
through the inherent power of this familiarisation is the characteristic mark of the result). See Karmay's edition of MTph
(Karmay 1988: 167); also Dudjom Jikdrel Yeshe Dorje 1991, Vol 1: 265; Vol 2: 111; and bDud 'joms Rin po che 1979-1985,
Volume Kha: 310.
The annotations in TZComm's Chapter 1 (1r.5) present them as: "When (one) understands through the Characteristic mark of
Knowledge, by the inherent power of becoming familiarised with the Characteristic mark of Engagement, the Characteristic
mark of the Fruit is accomplished as Buddha Body, Speech and Mind." (shes pa'i mtshan nyid gyis rtogs na 'jug pa'i mtshan
nyid gyis goms pa'i mthus 'bras bu 'i mtshan nyid sku gsung thugs su 'grub bo).
It is also described as representing the Great Accumulation (tshogs chen) in most of the versions, but the Dunhuang
manuscript and the Bhutanese NGB version have it rather that it is endowed with the assembly (or accumulations, tshogs can).

Pith Instructions on the Garland of Views

91

In this case, the term is shared but the content is a little different in that MTph is speaking of the fruit
accomplished when the great siddhi is attained, while the elaboration in TZComm is discussing the natural
presence of full enlightenment, and the arising of the buddhas from the seed syllables.
MTph has a section pointing out that only mentally superior persons can understand primordial
enlightenment, and thus, others should practise more gradual paths, and the teaching is taught as the "ultra
secret vehicle" (yang gsang ba'i theg pa, an expression not occurring in TZComm) so as not to disturb such
ordinary people. TZComm, on the other hand, does not concern itself with the problems of those who do not
yet understand its profound teachings, and mainly uses the word secret (gsang) in the term, secret mantra
(gsang sngags), without explaining its rationale. The implication at the end of the final chapter, where the
buddha body, speech and mind secrets (gsang ba) are discussed, is that they are secret since inaccessible or
hidden to most.
MTph in conclusion discusses the different types of ascetic conduct and practices of those who accept the
different views it has discussed. The only parallel in TZComm is the contrast it draws in Chapter 13, using
different phrasing from MTph, between the approach of rvakas, practising virtue and avoiding non-virtue,
with the Vajrayna sameness teaching taught in TZComm. The ethical dilemmas mentioned by MTph, in
which bodhisattva compassion may need in some cases to take precedence over the basic precepts, are not
discussed in TZComm.
MTph cites the dam tshig chen po'i mdo which draws on the classic Buddhist imagery of the lotus to
illustrate that those established in the Buddha vehicle remain pure in morality while engaging in the
afflictions, like a lotus in muddy water. A similar point is made in different words in TZComm's Chapter 1 in
discussing the text's title, and the sentiment is also present in TZComm's brief treatment of the lotus in the list
of the auspicious symbols given in Chapter 41. However, there does not need to be any direct connection
between the two teachings. The lotus is a ubiquitous Buddhist symbol and this interpretation of it simply
expresses a general Vajrayna perspective.
Thus, we can conclude that although there may be a number of parallels to TZComm in the teachings
given in MTph, these are not especially striking. Clearly, they are different kinds of texts: a root tantra and
the commentary elaborating on its specific ethos on the one hand, and a scholarly outline of progressively
more advanced types of perspectives within the Buddhist teachings on the other hand. If Padmasambhava's
teachings are represented in both MTph and TZComm, he did not leave any obvious clues in the content and
styles of the texts which would confirm this.

The Evidence from the Commentary and its Dunhuang witness

Apart from the annotations on the Dunhuang manuscript, there is only one firm indication that the Thabs
zhags transmission owes anything to Padmasambhava, and this is in the homage at the end of the
commentary, which we discuss below. As most tantric scriptures, the root text, TZ, is in the anonymous
voice of the buddhas. TZ lacks any translator's colophons, apart from the Bhutanese and South Central NGB
editions, which share a final colophon identifying Vimala and gNyags Jnakumra as translators. This
colophon is not shared by the local Kanjurs and it seems most unlikely that the identification descends from
the archetype. Colophons may be detached and added to texts and to the catalogues of textual collections,
and the relative geographical proximity between Bhutan and the South Central regions mean that it is quite
likely that the attribution might have been more recently shared between the two textual traditions. In any
case, there is no mention of Padmasambhava. Apart from what appears to be a scribal signature given at the

92

Padmasambhava, the Thabs zhags and its Commentary

very end of the Dunhuang manuscript in the small writing of the annotations,1 neither the Dunhuang nor the
Tenjur versions of the commentary have any obvious authorial colophon, ending simply with the statement
that the text ends here, using the usual formula with the text title followed by rdzogs so.
However, the Tenjur versions are placed in a section of texts associated with names from the early rNying
ma pa historical accounts, including Padmasambhava. In particular, in all three of our Tenjur editions,
TZComm immediately follows a commentary on the Dur khrod khu byug rol pa, which is attributed to
mTsho skyes rdo rje, most likely to be identified with Padmasambhava.2
More specific evidence of an association within TZComm itself is a four line verse of homage
immediately preceding the final colophon. In the Dunhuang manuscript, it is separated by the page layout
from the rest of the text, and includes a new opening yig mgo, as though it might have been appended to the
text. No such setting is witnessed in any of the Tenjur versions, where the verse is explicitly included within
TZComm's final chapter, since its colophon refers to the chapter ending as well as the text ending. Whether
an integral part of the final chapter, or an addendum, the verse pays homage to, "he who has attained the
supreme siddhi, of great wonder, Padma rGyal po [The Lotus King] (who) is not worldly; (he who) unravels
from the expanse the tathgata's great secret pith instructions".3 There are two questions here: is this verse
paying homage to Padmasambhava, and if so, what is it indicating about him in relation to the text?
We are now in the position to answer the first question with a clear positive. At first sight, it may not
seem certain that this verse refers to Padmasambhava at all, even though the annotator seems to take it as
doing so. Taken on its own, there are only two hints. The name, Padma rGyal po [Lotus King], came to
refer to one form of Padmasambhava in the Guru Padma rnam thar literature, although we have not seen it
used in other Dunhuang manuscripts. Secondly, the first line of this eulogy consists of a few words which at
a later date are found, in a different order, in the famous supplication and prayer to Padmasambhava, the
Seven Syllable Supplication (tshig bdun gsol 'debs), but the coincidence of these few words in a different
order can hardly claim any significance.4

"Written by Bo'u-ko of Kam-cu (Ganzhou?)" (kam cu pa bo'u ko gis bris). The fact that the note is given in small writing
would seem to imply that it is not to be considered part of the main text itself. It is probable that this apparently Chinese name
is that of the scribe of this manuscript copy, but it is also possible, since the name is given in the small writing of the
annotations, that it represents the author of the annotations, which as mentioned above, have been copied from the exemplar.
Nyang ral, for instance, uses this name in his hagiography: see below p.94. This commentary is found also in the bKa' ma shin
tu rgyas pa (Volume 62: 403-756), and the expanded version of the bDud 'joms bKa' ma (Volume Thu 70: 249-594). Its
colophon (reproduced in the Tenjur and bKa' ma versions) refers also to Lotus Skull-Garlanded (padma'i thod phreng),
another name for Padmasambhava, as overseeing the translation and editing of the text. In full, it reads: "Here ends the
extensive commentary on all the tathgatas' body, speech and mind secrets, the Charnel Ground Cuckoo's Display, called the
Elixir's Vital Seed, composed by the (Vajra) Master, Lake-Born Vajra (mtsho skyes rdo rje). In the presence of the Indian
Scholar Lotus Skull-Garlanded, the chief editor, the Translator Jnakumara, edited, translated and codified it" (de bzhin
gshegs pa thams cad kyi sku dang gsung dang thugs gsang ba dur khrod khu byug rol pa'i rgya cher bshad pa/ bdud rtsi thigs
pa zhes bya ba slob dpon mtsho skyes rdo rjes mdzad pa rdzogs so/ /rgya gar gyi mkhan po padma'i thod phreng gi zhal snga
nas/ zhu chen gyi lo ts ba dzny na ku ma ras zhus nas bsgyur te gtan la phab pa'o).
dngos grub mchog brnyes ya mtshan chen po 'i [Tenjur: yis]/ /'jig rten ngam gyur [bsTan 'gyur: ma 'gyur] pad ma [Tenjur:
padma'i] rgyal po yis [Tenjur: las]/ /de bzhin gshegs pa'i man ngag gsang chen rnams/ /klung [Tenjur: klong] nas bkrol mdzad
de la phyag 'tshal lo//
The words are terms often used in association: "supreme siddhis" (dngos grub mchog) everywhere imply enlightenment; they
are often said to be discovered (rnyed pa), and describing them as wondrous (ya mtshan) need not be more than a coincidental
use of the same words. However, it is interesting that another old source also has a slightly similar categorisation of
Padmasambhava. A surviving manuscript copy of the ninth century 'Phang thang ma catalogue of translations, which begins
with a list of Indian masters, apparently representing the captions of illustrations on the earlier scroll from which the scribe
copied, describes him as: u rgyan gyi pa i ta grub pa brnyes pa'i pad ma 'byung gnas (dKar chag 'phang thang ma/ sgra
sbyor bam po gnyis pa 2003: f.1v.4; see also below, note 13).

The Evidence from the Commentary and its Dunhuang witness

93

Following up the hint that the name, Padma rGyal po, may connect this verse with the Guru rnam thar
hagiographies, and the phrasing of the verse as so enticingly evocative of Padmasambhava's later mystique,
we looked for similarities within the Zangs gling ma, the early gter ma hagiography of Guru Padma revealed
by Nyang ral Nyi ma 'od zer (1124-1192). And we found a closely parallel verse of praise to the great guru:
Final Verse of the Commentary to the 'Phags pa
thabs kyi zhags pa padma 'phreng gi don bsdus pa
(Dunhuang manuscript IOL Tib J 321 [Ms], f.84r;
bsTan 'gyur: Golden [Gt] rgyud 'grel Bu, 78-321,
Peking [Qt] rgyud 'grel Bu, 129b, sNar thang [Nt]
rgyud Bu 228)

[Ms ]

Nyang ral, Nyi ma 'od zer Slob dpon padma 'byung

gnas kyi skyes rabs chos 'byung nor bu'i phreng ba
zhes bya ba, rnam thar zangs gling ma (based
primarily on the Kathmandu National Archives
manuscript in dbu med (IMG_1670+1671, reel
E2703/10, f.16r.5-16v.1).5

[Ms ] [Ms ]
[GtQtNt

;=-]k

[Ms ]

(I) prostrate to he who has attained the supreme (I) prostrate to and praise the (buddha) body who
siddhi, of great wonder,
has attained the supreme siddhi, of great wonder,
Padma('i) rGyal po [The Lotus King] (who) is not the body of incomparable realisation, Padma rGyal
worldly;
po [The Lotus King];
(he who) unravels from the expanse

you (who) unravel from the expanse

the tathgata's great secret pith instructions.

the tathgata's great secret pith instructions.

It is clear that the verse is simply a variant on the phrasing, the first words of the second line a clear
improvement on TZComm's slightly obscure 'jig rten ma 'gyur. Nyang ral follows up the praise with a
further statement that the Guru then received two names: Padma Thod 'phreng, since he was wearing a
garland of skulls, and Padma rgyal po, since he had been made a king's son.6 Thus, the Zangs gling ma uses
the name, Padma rgyal po, for the Guru at exactly this point. It is worth adding that the name is not used
repeatedly in the Zangs gling ma: the Guru is mostly referred to as Padma 'byung gnas, or simply by the title,
5

Lewis Doney of the School of Oriental and African Studies, University of London, has worked on critically editing Nyang ral's
Guru Padma hagiography. He argues convincingly that the earliest and historically most influential recension is that
represented by two manuscripts in the National Archives in Kathmandu and two manuscripts from Bhutan, which he classifies
as ZL3. The version of ZL3 used here is Lewis Doney's discovery in the Kathmandu National Archives. We have emended
rtog in line 2 to rtogs, found in all the other witnesses of ZL3. The Rin chen gter mdzod chen mo version (Paro: Ngodrup and
Sherab Drimay, Kyichu Monastery, 1976, Volume Ka: 25), which has more recently become the most widely used version,
incorporates later material. It gives a variant second line (rtogs ba bla med mchog tu gyur pa yis/) for this verse.
zhes bstod nas/ thod pa'i 'phrengs pa sku la gsol bas/ padma thod 'phrengs du btags/ rgyal po'i sras mdzad pas/ padma rgyal
por btags/ (Kathmandu National Archives, IMG_1671, reel E2703/10, f.16rv.2-3).

94

Padmasambhava, the Thabs zhags and its Commentary

Slob dpon (master, guru). But as in the Guru's later hagiographies, many of the incidents related in the text,
and especially in the first part of the text focusing on the Guru's training and activities before the period in
Tibet, are followed by a note that at that time he was called by such-and-such a name. In the earlier
description of King Indrabodhi's discovery of the lotus-born emanation and his installation as the king's son,
the name he is given is, rGyal po mTsho skyes rdo rje (King Lake-born Vajra; p.9).7
In the fourteenth century Padma bKa' thang shel brag ma of O rgyan gling pa, we find a fully developed
system of the eight principal names of the Guru (gu ru mtshan brgyad, or sku gcig mtshan brgyad, Padma
bKa' thang Chapter 19: 88-9), which include Padma rgyal po,8 and the name is also used during the narrative
of his enthronement as a prince following his birth upon a lotus.9
It would seem that there is little doubt that the Padma('i) rgyal po who is praised at the end of TZComm is
none other than the Great Guru Padmasambhava, eulogised in the same manner as that used in the slightly
later hagiographical literature.10 It is worth adding that the first line of the praise occurs also just above in
Nyang ral's hagiography (IMG_1670, reel E2703/10, f.15v.4), again within a four line verse in the context of
the King of U rgyan's astonished wonderment at the sight of the Guru in union with his consort arising from
a lotus.
Assuming this identification of Padma('i) rgyal po in TZComm is correct,11 the significance of the homage
depends on whether it is seen as the final part of the text of TZComm, or whether it is an addendum. If it is
an addendum, it might be praising the author of TZComm; but if it is part of the main text, as seems likely by
its positioning in the Tenjur version, and the fact that even in the Dunhuang version it occurs before the
concluding text title, then it is the authorial voice of TZComm itself that is praising Padma rGyal po. The
context of the preceding lines of TZComm would seem to increase that likelihood that this eulogy is in fact
an integral part of the final chapter of TZComm. TZ's final chapter or section consists of a series of praises to
the maala deities, and their consecrations of oneself, and ends with a line cited by Klong chen pa in his
Phyogs bcu'i mun sel commentary on the rGyud gSang ba'i snying po root Mahyoga tantra, to the effect that
the maala becomes invisible through absorbing into the deity's own heart. The commentary adds further
lines which would seem to eulogise TZ as enlightened speech called "tantra", flowing from the
Buddha/Protector's turning of the vajra wheel, flawless and bringing realisation. The final verse praising
Padma rGyal po would seem to fit perfectly well with the tone of this passage, which is also written in the
same seven syllable meter and four line verse form.
7

10

11

The name mTsho skyes rdo rje (Lake-born Vajra), continues to be used in the narrative, both in praises (Rin chen gter mdzod
chen mo Volume Ka: p.22, 24), and in the Guru's self-appellation (p.48). It became one of the principal names for the Guru as
a meditational deity, used in many cycles of ritual meditative practices.
There is a set of homages to the group in Chapter 19; the one for Padma rgyal po reads: thabs mkhas rgya chen padma'i lho
phyogs su: ye shes klong chen padma rgyal por sprul: rin chen mkha' 'gro ma tshogs 'khor gyis bskor: cir snang rig pa gang la
gang 'dul bzhugs: padma rgyal po'i sku la phyag 'tshal bstod: (1985 edition, p.145). Again, in Chapter 41, there is a shorter
set of praises to the eight. The one for Padma rgyal po reads: khams gsum srid gsum dbang du sdud: padma rgyal po'i sku la
bstod: (1985 edition, p.281)
For instance, Chapter 18 begins with King Indrabodhi bestowing the name upon him (de la khye'u bzhugs rgyal por mnga'
gsol nas: mtshan yang padma rgyal por zhes byar btags: (1985 edition, p.145). The name, mTsho skyes rdo rje, which the
Zangs gling ma gives at the time of the "birth", is also used in this story at the site of the emanation. Then, Chapter 20 refers
to the Guru again with the name, Padma rgyal po, beginning with the words: de nas rgyal sras pad ma rgyal po yis: (1985
edition, p.147), and below, it speaks of him as the stainless supreme (buddha) body, Padma rgyal po, the second buddha who
is the lamp of the world (padma rgyal po'i sku mchog dri ma med: sangs rgyas gnyis pa 'jig rten sgron ma khyod: (1985
edition, p.148).
This is not the only textual parallel we have found between Dunhuang manuscript sources and Nyang ral's revelations. A
more substantial parallel in found within his bKa' brgyad bde gshegs 'dus pa, which we will be discussing in a forthcoming
article.
It is just possible that the verse was picked up by Nyang ral and given a different connotation from that in TZComm, but this
would seem a little unlikely, and clearly the annotator understands the text to refer to the great guru.

The Evidence from the Commentary and its Dunhuang witness

95

Perhaps, then, TZComm is implying that Padmasambhava is associated with the creative production of
enlightening tantric text, which has just been extolled in the preceding lines. Another possibility is that the
praise is focusing not on the composition of the text as such, but rather on the enlightening potential of its
teachings, teachings which are quintessentially those of the Great Guru. This might seem to make all the
more sense when we consider the context for the parallel verse in Nyang ral's hagiography. Here, following
the earlier praise mentioned above, the king pays obeisance and the Guru teaches him that one's own mind is
the unborn dharmakya. Immediately realising it as such, the king and his retinue attain the acceptance of
the unborn nature of dharmas. The amazed king then praises the Guru in the words of our verse. It may not
be entirely arbitrary that the context for the close of TZComm is some words of conclusion that realisation is
accomplished through the teaching, and this follows the final teaching of the root tantra on dissolving the
maala into the heart. Of course, this is not an exact parallel, but perhaps similar enough that we can reflect
that the verse is not simply a praise of the Guru's qualities, but a celebration of amazement at the ultimate
tantric teachings he embodies.
So far, we have considered only the main text of TZComm. If we then turn to the annotations, two
Padmasambhava references are given in this last section of the text. The final reference is placed directly
under the first line of the homage, and reads: The master ntigarba, having examined and (found it/him)
flawless, is praising Sambhava.12 [See IOL Tib J 321, folio 84r, image 85_83v-84r on the accompanying cd]
ntigarbha is known in early sources as a contemporary and colleague of Padmasambhava, often associated
with the Yogatantras. He is mentioned as the consecrator of the Imperial temple at bSam yas at the start of a
manuscript copy of the ninth century 'Phang thang ma register of translations authorised by the Tibetan
state.13 ntigarbha continues to play an important role alongside Padmasambhava in later rNying ma
literature as one of the so-called Eight Vidydharas of India, whom the rNying ma pa revere as important
founders of their Mahyoga tradition, and several of whom figure quite visibly in the Dunhuang literature.14
Although the annotator seems to assume we know, it is not clear exactly what he is implying about
ntigarbha's role in TZComm. He may be indicating that ntigarbha is the author of TZComm which
includes this final verse. Alternatively, he may be suggesting that ntigarbha's words of praise from
elsewhere are inserted here, or that ntigarbha has appended a homage to the text of TZComm. This third
possibility is rendered a little unlikely considering that the homage is placed within the text rather than after
the colophon, and we have noted that the Tenjur version even more clearly integrates the verse. It is hard to
judge between the other two possibilities. Since this is the first time the annotator has mentioned
ntigarbha while writing notes on some eighty-five folios of text, and since ntigarbha is also not named
in an authorial colophon, it seems unlikely the annotator is assuming we will take ntigarbha as TZComm's
author. But even if he is suggesting merely that some lines composed by ntigarbha are being drawn upon
12
13

14

slobs dpon shan ting gar bas brtags nas ma nor nas/ sam ba bha la stod pa 'o/ (f.84r.5)
See dKar chag 'phang thang ma/ sgra sbyor bam po gnyis pa (2003: Plate 2, f.1v.6-7; p.2: rgya gar gyi slob dpon bsam yas kyi
rab gnas mkhan sha ting gar bha). The surviving manuscript copy is doubtless a good deal later. It is hard to say whether
we can trust that ntigarbha was mentioned in the original ninth century document. Yet certainly, in listing the great Indian
masters at the start of the manuscript, the copyist mentions that the large paper scroll which he was copying contained
captioned drawings of these figures (de rnams ni shog dril chen po'i stod na sku gzugs bris pa rnams kyi kha yig yin/, f.1v.78), so we can surmise that his list reproduces the earlier captions. Elsewhere, the great fourteenth century canonical scholar Bu
ston mentioned ntigarbha as co-translator of the Sarva-durgati-pariodhana-tejo-rjasya tathgatasya arhato samyaksabuddhasya kalpa-nma, and from the lHan kar ma (AKA lDan kar ma), we know that this was amongst the early
translations of ntigarbha's time (Adelheid Herrmann-Pfandt 2008: 177, item 323). Martin 2006 (entry on ntigarbha) notes
that Bu ston's Yogatantra history (135.1, 140.1) lists his works and mentions his Yogatantra explanations, 'Bru 'grel rgan po.
As far as we are currently aware, without having made an exhaustive search, at least five out of this list of eight masters have
turned up so far in various Dunhuang tantric texts: Maju rmitra in IOL Tib J 331.1 and in IOL Tib J 1774 , Prabhahasti
seems perhaps to be referred to (as "pra be se") in PT 44 (Cantwell and Mayer 2008: 60), Padmasambhava in several (PT 44,
PT 307, IOL Tib J 321), ntigarbha in IOL Tib J 321; Vimalamitra (bye ma la mu tra, f.1) in IOL Tib J 644 and IOL Tib J
688 (on rosaries); Maju rmitra, Hkara (and Buddhagupta) in IOL Tib J 1774 (slob pon nI 'bu ta kub ta dang/ shI rI man
'ju dang/ hung ka ra).

96

Padmasambhava, the Thabs zhags and its Commentary

and inserted here, we have seen that the placement of the verse appears to fit appropriately as the text's own
final flourish.
What then is the annotator saying about this praise? The implied object of ntigarbha's examination is a
little ambiguous. He seems to say that ntigarbha finds either Padmasambhava's teaching flawless, or
perhaps Padmasambhava himself [See IOL Tib J 321, folio 84r, image 85_83v-84r on the accompanying cd].
Quite likely, the word ma nor (flawless) is picking up on the line in the text above. The previous verse
concludes that due to speech flowing forth when the great Protector turned the vajra wheel, the enlightened
body, speech and mind secrets have been accomplished and realisation attained, so this is explained as
simply flawless.15 Thus, the word flawless perhaps applies to the tantra's teachings (referred to in the homage
as: the tathgata's great secret pith instructions [de bzhin gshegs pa'i man ngag]), which are to be associated
with Sambhava. An alternative is that it is Sambhava himself who ntigarbha examines and finds flawless.
In fact, the realisation indicated as flawless in the line above could also be intended to refer to the
accomplished Sambhava, even if this is certainly not explicit in the text. In either case, it is clear that the
annotator takes it as given and not worthy even of explicit comment that the text owes its inspiration to
Sambhava, whether as the representative of the teachings expounded, whether as the composer of TZComm,
or even as the producer of TZ.
The second annotation referring to Padmasambhava, which occurs at the end of the folio above,
reinforces the impression that the annotator is drawing our attention to Padmasambhava's association with
the teaching presented in the root text and commentary [See IOL Tib J 321, folio 83v, image 85_83v-84r on
the accompanying cd]. TZ ends with the maala dissolution in the heart. This follows a concluding praise
of the deities of wrath with the natural condition of sameness, generating enjoyments and emanating and
reabsorbing. The final section of TZComm then begins with the commentarial words, "Since [it] has
emanated forth from out of sameness/evenness, (the) emergence means that..." (mnyam las phros te byung
ba'i don, 83v.6-84r.1). Beneath the first part of this explanation, the annotation adds: "this demonstrates
[that it] is not created by Padmasambhava idiosyncratically", or possibly, "Padmasambhava demonstrates
[that this] is not an idiosyncratic creation" (pad ma sam ba bhas rang gz[or?] byas pa ma yin bar ston,
83v.6). It is most likely that the comment is glossing the phrase in TZComm with which it is directly linked
(mnyam las phros te...).16 Quite what the note is suggesting about Padmasambhava's role in TZ's and/or
TZComm's teaching, however, is not entirely clear, again, presumably because the writer assumes that we
already know this, and he is focusing on commenting on the implications of that role or the content of that
teaching. At the least, however, it would suggest an association between Padmasambhava and the teaching
presented in the text, possibly TZComm, but again, perhaps TZ. If the annotation means, "this demonstrates
[that it] is not created by Padmasambhava idiosyncratically", then the positioning of the comment at the end
of the root text, several lines before the end of TZComm, would imply that it can only be TZ with which
Padmasambhava is linked, and which he has in some sense created or produced. The phrase, emanated forth
from out of sameness/evenness, seems to refer not only to the root text's preceding discussion of the deities,
but also connects with the next words of the commentarial elaboration, which eulogises the emergence of the
noble being's speech as "tantra". The implication of the annotation might therefore be that the phrase,
emanated forth from out of sameness/evenness, shows that the emergence of the speech of the tantra (referred
to on the next line) is not idiosyncratically produced by Padma, but produced through such a natural
emanation. In this case, the annotation suggests that Padmasambhava may have produced or inspired TZ.
15

16

mgon po bdag nyid chen po yis/ /rdo rje 'khor lo bskor ba na/ /ljags kyi dbang po bkram las [Gt pas] gsungs/ /de las brten te
mgon po'i [Gt yis]/ /sku gsung thugs kyi gsang ba rnams/ /rnal 'byor sgrub pas [Gt bsgrub pa] rtogs bya'i phyir/ /ma nor tsam
du bshad pa yin/ (Dunhuang Ms f.84r).
Our understanding here is a little at variance to Sam van Schaik's (2007) interpretation that the words rang gzor byas pa ma
yin bar refer to Padmasambhava's composition of TZComm as without fabrications of his own. However, van Schaik's overall
point, that Padmasambhava seems to be indicated as the architect of the teaching given in the text here, would indeed seem to
be implied.

The Evidence from the Commentary and its Dunhuang witness

97

On the other hand, if the annotation means, "Padmasambhava demonstrates [that this] is not an
idiosyncratic creation", then it might be drawing attention to the teachings on the natural process of the
maala emanating and reabsorbing, teachings which are to be seen as Padmasambhava's. In this case, the
annotation might be interpreting the words of commentary here (emanated forth from out of
sameness/evenness), implying that it is Padmasambhava's teaching that the process of the maala emanation
is not idiosyncratic. Alternatively, the annotation might not be interpreting the commentarial words, but
instead enlarging upon and adding to their interpretation of TZ s teaching. Padmasambhava's demonstration
would then be embodied in the words of TZ above, concerning the wrath with the natural condition of
sameness, producing vajra absorption and emanation.
Whichever specific connotation the annotation has, we need to add the caveat that it is perhaps not
intended for us to take the reference to Padmasambhava's demonstration or to his non-idiosyncratic creation
as representing his composition of the actual words of the text, whether of TZ or TZComm. The comment
may simply be linking Padmasambhava to the teaching (or the production of the teaching) presented here,
but it is certainly quite possible that the assumption of this annotation is that Padmasambhava is the main
architect of the text, most probably of TZ, but possibly of TZComm, or even of the entire text of root verses
and commentary.
The first reference to Padmasambhava in the manuscript comments on the words in the text title, pad ma
'phreng gi don bsdus pa'i 'grel pa', [see IOL Tib J 321, folio 1r, image 01_1r on the accompanying cd]
saying that the Buddha has summarized (the meanings) and Sambhava produced (them). This cryptic
comment might suggest that the Buddha condensed the meanings in TZ ('bu tas bsdus), while
Padmasambhava expanded them into TZComm (sam ba bhas byas [where byas = 'grel byas, cf 'grel pa byed
pa]). Alternatively, it might once again suggest that Sambhava is to be associated with TZ itself, that is, that
Sambhava has produced the Buddha's synopsis. Had the annotator wished to indicate that Sambhava
produced TZComm, he might have positioned this final part of the comment beneath the word, 'grel pa'
(commentary), rather than appending it to the comment connected with the word, bsdus pa (summarized). It
may even be that the word padma in TZ's text title may carry a further implication as well as the general
symbolic connotations of the word, lotus, in Buddhist literature. That is, as well as a "Lotus Garland
Synopsis", the title could be read as, "Padma's Garland Synopsis".17
Even if the comment is intended to imply that Sambhava was responsible for TZComm rather than TZ, the
wording by which Sambhava is linked to the Buddha's composition would suggest a close association with
the production of Buddha Word. And the imagery of the Buddha condensing the meaning, which is then
unravelled by Sambhava, might suggest some resonance with the kind of process which became central to
the gter ma tradition, by which the mystic revealer decodes and expands on the condensed kin script of
the initial revelation.
Clearly, questions still exist about these references to Padmasambhava. Do they suggest that
Padmasambhava was a great commentator or that he inspired the teachings given in TZComm, or are they
hinting even more that he was a revealer of tantras? We are unable fully to resolve these issues. One
conclusion from this examination of Padmasambhava's connection with TZ and TZComm is that we do not
appear to be dealing with a Padmasambhava represented merely as an ordinary human tantric scholar,
perhaps editing a text or composing its commentary. Rather, the final verse of TZComm itself gives every
impression of Padmasambhava mythologised as a fully realised and wondrous tantric master, eulogised in
17

TZComm's gloss on the title restricts itself to associating lotus with wisdom and garland with means, lotus further implying
that habitual tendencies do not develop despite engagement in all phenomena (pad ma [Qt padma] ni shes rab kyi rtags te [Nt
to]/ chos thams cad la spyod [QtGtNt dpyod] kyang [QtGtNt insert: /] bag chags su mi 'gyur ba'o/, Dunhuang ms. 2r).
However, it is not impossible that the additional connotation relating to Padmasambhava may have seemed obvious, just as the
fact that a lotus is a flower is not worthy of any remark.

98

Padmasambhava, the Thabs zhags and its Commentary

the same terms we find used of him in Nyang ral's hagiography, linked to the profound vajrayna teachings
of TZ, an enlightened figure perhaps implied to be at the basis of the textual tradition. The additional three
annotations relating to him in the Dunhuang manuscript only reinforce this impression. At the start of the
text, Sambhava is made responsible for producing the Buddha's condensed language. In the final chapter, he
is either said to produce the speech of the tantra naturally, or to demonstrate TZ's teachings on natural
emanation, while at the end, ntigarbha is said to confirm and celebrate Sambhava's realisation.

INTRODUCTION TO THE EDITIONS OF THE 'PHAGS PA THABS KYI ZHAGS PA PAD MA 'PHRENG GI DON
BSDUS PA AND ITS COMMENTARY
Our editorial policy pursues two objectives. In the first instance, we attempt, as far as possible, to restore
archetypal readings, that is, the readings of the most proximate common ancestor of all extant versions, freed
from the scribal corruptions and editorial interventions accumulated over the centuries.1 Assuming our
stemmatic analysis is correct, these are easily identified for TZ where a majority of the editions descending
directly from the archetype support a specific reading. Note that the hypearchetype of the Tenjur, Dunhuang,
Bhutanese, and Tshal pa Kanjur versions of TZ, which we have designated b on our stemma codicum (see
p.43 above), also counts as one voice in identifying the archetype. Uncertainties remain where the five
branches do not give a clear picture, say, where b has a single reading that is unlikely to be archetypal, but
the four others are evenly divided. In such cases which with TZ generally concern only very minor
variants we have had to make relatively arbitrary decisions, based on whichever reading might appear
preferable, or slightly more likely to have been the earliest.
To render our edition adequately accessible to a modern readership, we have not presented the TZ
archetype's archaic and non-standard orthography in the main body of its text, especially where we are quite
certain of the underlying intended word; nevertheless, all such archaisms and non-standard spellings are
carefully recorded in the apparatus in italics. By contrast, in those cases where an archaism or non-standard
spelling is difficult to analyse or remains ambiguous, we retain it within the main body of the edition.
It is perhaps instructive to recall that our proposed archetype of TZ must logically precede the Dunhuang
manuscript, which might have been copied as early as the late tenth century (although we cannot be certain
of this date). Nevertheless we cannot claim that our archetype represents an original text, even if logic
cannot entirely rule out that unlikely possibility. Much the same can be said of TZComm, although in that
case, no archetype can be restored. It follows that in both TZ and TZComm, the earliest or best recoverable
texts available to us will in all likelihood already incorporate previous scribal errors, which we cannot undo
with confidence. Moreover, given that tantric literature typically integrates and recycles sacred words from
the past, it is even possible that the original texts of TZ and TZComm may themselves have incorporated
prior passages of text. Such passages might have preserved their own irregularities, such as non-standard
lists of mudrs or deities, apparent inconsistencies, disordered materials (see Cantwell and Mayer 2007: 41,
81-82), or simply scribal corruptions. We cannot then expect that our archetype will be entirely free from
apparent error, nor that it will necessarily appear to be entirely consistent and coherent. Nonetheless, we
should be able to eliminate a good deal of further scribal accretion that has entered the tradition over more
than a thousand years.
It will be readily apparent that the extent to which we can achieve our first objective differs considerably
in the case of TZ and in that of TZComm. Our analysis above (see p.44ff) supports the possibility of
fruitfully restoring the archetype of TZ because five versions of TZ can be isolated that apparently descend
from the archetype independently, and this enables one to infer many archetypal readings. We have no such
luxury in the case of TZComm, where we have only two versions of the text, both of them afflicted by
omissions and scribal corruptions. The best we can do in this case is to restore at least the overall shape of
the text restoring chapters or passages lost in one branch of the transmission, and correcting misplacements
1

The idea of an archetype should never be confused with such conceptions as 'the original text' or 'the author's original
composition'. An archetype is merely the lowest common ancestor of all extant editions as represented on a stemma codicum
or stemma diagram, which is the furthest limit to which stemmatic logic can penetrate. An archetype might sometimes
coincide with 'the original text', but it certainly need not do so. In truth, 'the original text' or 'the author's original composition'
often proves to be a much less clearly defined entity than it might first appear, but even where such an entity is known to have
existed, it frequently remains quite unattainable to textual criticism.

100

Introduction to the Editions

of material but we cannot achieve a stemmatic reconstruction of the archetype, except where both versions
agree. Where the two versions differ in coherent readings, or where they both appear corrupt in different
ways, we cannot be certain which reading is earliest.
A second objective has been to acknowledge and showcase TZ's varied lines of transmission. As we point
out above (see p.20-25), Tibetan Buddhism reveres above all the living lineages of scriptural transmission
actually passed down by its lamas. Even where modern scholarship can reconstruct archetypes, an
adequately nuanced study needs also to take account of the differing versions of a text that have actually
served as its living representations in real historical contexts. For this reason, we have drawn attention to the
alternative readings of the major branches of the TZ transmission, even where they may not reflect those of
the archetype.
The Presentation of the Critical Editions
It is worth commenting here on our presentation of the two editions. For TZ, we have used Tibetan script,
both for the main text and the variants given in the apparatus, with supplementary notes and comments in
English. We hope that it will be reasonably accessible and clear for Tibetan readers, even for those without
English language skills. We have, however, retained transliteration in Roman script for the edition of
TZComm. The reason for this is that our text of TZComm primarily represents a diplomatic edition of the
Dunhuang manuscript, supplemented by readings from the other versions, and it is not yet straightforward to
represent all the idiosyncrasies of a Dunhuang document in the modern Tibetan fonts used in word
processing. Furthermore, thanks to the British Library, images of the original Dunhuang document are
included on the CD accompanying this volume, and are also readily available on the International Dunhuang
Project website hosted by the British Library, at <http://idp.bl.uk/database/oo_loader.a4d?pm=IOL Tib J
321>. We have therefore used well-established conventions for the transcription of the Dunhuang
manuscripts (see p.ix-x).
The Presentation of the Critical Edition of TZ
We have presented the readings which we believe most probably represent those of the archetype as the
main body of our edition, and where uncertainty prevails, we draw attention to the alternative possibilities by
italicising them in the apparatus. The exception is where the archetype appears to have contained a nonstandard or archaic orthography, or a trivial spelling error: as explained above, in such cases, if we are
certain of the intended reading, we do not restore the irregular or archaic forms in the main body of the
edition, but simply call attention to them in the apparatus by the use of italics. Given the number of versions
collated, we have not included all very minor variants of punctuation or of orthography; an exact and
detailed transcription of the Dunhuang manuscript is given in our TZComm edition.
Once again, it is worth reminding the reader that we do not in all instances have absolute proof of our
stemma, and most notably on one critical point: we cannot prove quite incontrovertibly that the three local
Kanjurs and the South Central editions all descend separately from the archetype, even though we can show
that this is most probably the case. If it transpires that they do not, and that instead they all descend from a
single hypearchetype, then wherever we have prioritised their shared readings over equally (or more)
coherent readings from b, our editorial choices will lose their logical underpinnings. However, this
eventuality is guarded against in the presentation of the edition: we consistently draw attention to all
meaningful variants by presenting them in the apparatus in bold face, so that wherever such a risk exists, it
will be readily apparent which meaningful variants are shared by the descendants of b.
Assuming our stemma to be correct, the bolding of alternative meaningful variants also has a further
purpose, which addresses our second objective outlined above (p.100). We recognise the historical fact that
despite their textual differences, not one but several of TZ's varied versions have played a significant rle in
TZ's promulgation in Tibetan society. Whatever meaningful readings might be found anywhere within the

Presentation of the Critical Editions

101

different versions of TZ, and however divergent from the archetype, they are nevertheless quite likely to have
played a valued rle in embodying the text to different people at different times and in different places. This
point is especially pertinent to the descendants of b. The celebrated Tshal pa edition of the Kanjur became
the basis for all printed editions of TZ, including the sDe dge NGB,2 while the extensive Bhutanese NGB is
arguably the most widely distributed edition in use today. Apparently the heirs of a transmission
incorporating the scribal errors of both b and of c, as well as introducing more errors of their own, and even
spuriously incorporating considerable passages of commentarial text (see p.35ff.), these two editions have
nevertheless become the most influential in the later historical transmission of TZ. Although their respective
intrusions of commentarial text were surely not in the archetype of TZ, we recognise that they have been so
long established and in such important historical transmissions, that it now seems inappropriate to relegate
such considerable addendums to mere footnotes. Thus we present them within the main body of the text, but
in the form of indented paragraphs, to make it clear that they should not be considered part of the restored
edition of the archetype. Our presentations of such passages therefore take as their base texts the Tshal pa or
Bhutanese editions as appropriate, but where these are at fault, we select readings from one or another of the
TZComm witnesses on an eclectic basis.
In total, we have consulted twenty-one witnesses of TZ. As we have shown, for stemmatic purposes,
these can be reduced to eight versions, comprising four single witnesses and four textual groupings (see p.26
and the diagram p.43 above). It would have been impractical exhaustively to represent all twenty-one
witnesses in full in the edition, and of very little benefit, given that many of them are descriptive or repetitive
and can thus bring nothing new to the investigation. A considerable quantity of the initial collation was
eventually removed from the earlier drafts to produce the more manageable critical apparatus presented here.
We have therefore represented every witness in full only for the first chapter, for purely illustrative purposes,
and if this has resulted in an excessively overburdened apparatus in Chapter One, we can only crave our
readers' patience. Throughout our edition, we have represented at least one witness for each of the eight
versions (see p.26-30 above). Wherever we represent a consequential scribal corruption, all witnesses that
could in any way prove relevant have been consulted, even though we do not necessarily represent this in our
apparatus. Thus, for example, we consulted all the available witnesses in the case of the major omission in
sub-branch c at the end of Chapter 10. Moreover, wherever we represent a more minor but still noteworthy
scribal lapse in one member of a textual grouping, all its sibling texts have been scrutinised to ascertain
whether the lapse was found throughout that textual grouping, or merely within the one member. Again, such
additional scrutiny is not normally recorded in our apparatus, except where it produced some unexpected
data.
To facilitate consultation of the original documents, we give location lists for all the collated editions at
the opening of each chapter and for each chapter title. We give further location lists following each occasion
when the TZComm witnesses have included commentarial additions between the root text verses. If the new
location list follows closely after a previous list, then only the TZComm edition page references are given.
The Presentation of the Edition of TZComm
Since we have been unable to restore an archetype of the text of TZComm, and since the Dunhuang
manuscript is the most complete extant edition as well as being such a valuable early historical witness, we
have where possible adopted it as the base text of our TZComm edition. As discussed above (see p.32-33),
the Dunhuang manuscript contains copious annotations in small writing, which themselves appear to
represent a tradition of interpretation, having apparently been reproduced from a previous copying of the
text. We present these in small italic print, placed in an approximately appropriate position, either above or
below the line, corresponding as much as possible to their placement in the original document. We also use
our own grey highlighting to indicate the lines and words that were highlighted in yellowish wash in the
2

In TZ's case, the sDe dge NGB simply reused the woodblocks from the sDe dge Kanjur.

102

Introduction to the Editions

Dunhuang manuscript, and these give some indication of the scribe's understanding of the extent of TZ's
lemmata (see the discussion above, p.36-37). As we have seen, however, the highlighting in the manuscript
does not always correspond exactly to the boundaries of TZ as proposed by our reconstruction of the
archetype. Hence, for ease of comparison, we present the words of TZ as given in our reconstruction, in bold
font.
In the apparatus, we present the variant readings of the Golden Tenjur version of the text, and in some
instances, also the Peking Tenjur and sNar thang Tenjur readings. While the Golden Tenjur version has been
fully collated throughout,3 we only fully collated representative samples of the Peking and sNar thang
Tenjurs, but sufficient safely to reconfirm their very close agreement with the Golden Tenjur as already
demonstrated in the edition of TZ. We also consulted one or both of these further Tenjur editions in all cases
where the Golden Tenjur differed significantly from the Dunhuang manuscript, or where the Golden Tenjur
appeared to have scribal errors or variants of any interest. Furthermore, we have taken account of readings
found in the Tshal pa Kanjur and Bhutanese versions of TZ, since as explained above, they must have
descended from a commentarial source, and they still incorporate lengthy passages from TZComm within
their redactions of the root text. Their readings are recorded when they diverge significantly from the two
commentary versions, or where they add a further perspective of interest. Of course, these readings are
found more fully in the TZ edition, as noted above.
While our edition is in most parts a diplomatic transcription of the Dunhuang manuscript with the
Tenjur's readings merely presented in the apparatus, it is also something more than that, and has an important
eclectic aspect. One of our main objectives has been to restore a complete text of TZComm, since, as we
discuss above, such a complete text has evidently been lost to the Tibetan traditions for many centuries. We
have therefore restored the sections lost to the Dunhuang manuscript through scribal omission in the first
instance by recourse to the Tenjur edition, and so, for these sections, it is the Tenjur's readings which we
present in the main body of the text. More rarely, where the Dunhuang reading is missing yet the Tenjur's is
faulty, our eclecticism extends as far as the Tshal pa Kanjur and Bhutanese versions of TZ, if they present the
passage of text in question and can offer a preferable reading (an example of this can be found near the
beginning of Chapter 3). We have moreover moved into Chapter 13 the section belonging to it, which had
been misplaced in Chapter 12 in the Dunhuang manuscript (see above, p.74). We also direct the reader to our
TZ edition for the final section of Chapter 10, a section which is omitted in full in the Dunhuang manuscript,
and in part in the Tenjur. Our edition therefore maintains a basically diplomatic disposition, but has recourse
to eclecticism where necessary. Finally, for the convenience of the reader, we have called attention to all
variants of any interest within the apparatus by the use of bold face. In this way, we have been able to
restore a complete version of TZComm, even though some textual puzzles and probable corruptions still
remain.

So as not to overburden the apparatus, we have not, however, included all extremely minor variants, such as Gt's standard
rendering of 'i following a letter without an intervening tsheg, where the Dunhuang manuscript generally gives a tsheg; thus,
the Dunhuang manuscript's pa 'i will be written, pa'i, in the Tenjur texts.

CRITICAL EDITION OF THE ROOT TEXT OF

THE 'PHAGS PA THABS KYI ZHAGS PA PAD MA1 'PHRENG GI DON BSDUS PA
[Note that within the apparatus (but not necessarily in the accompanying notes), a bolded reading indicates a
meaningful alternative reading; an italicised reading indicates a reading which may have been in the
archetype, but which we are not re-introducing, since it represents an archaic or non-standard spelling, or a
minor error, or it is less metrical (such as pa'i versus pa yi), or since the manuscripts most likely to be
preserving early readings are evenly divided, so the earliest reading is uncertain).]
Chapter 1
[Note that the commentaries commence 1r; 101a.1; 243; 176.5. The editions which
insert an initial title commence 294v.6; 299b.7; 816.5; 597.1; 597.1;
2
286r.1 ; 472v.7; 358v.1; 180r.5; 320r: title page]
[ 1v.4] [294v.6] [299b.8] [ 816.6] [597.1] [597.1] [286r.1] [472v.7]
[358v.1] [ 101a.6] [ 244.6] [ 177.2] [123.6] [51r.3] [46r.4] [395.4]

[180r.6] [320v.1] [204r.7] [31r.7] [367v.3]

4 5 678 910

1 See note 6 below. As explained there, there is variation and inconsistency in the different editions over the choice of

or / in the text title. Since the archetype most probably favoured / in most of the chapters and in the final

colophon, we are giving this here. | 2 and are made from the same blocks, generally with the sole difference that the

recto sides have different pagination (although there are a few rare spelling corrections in ). | 3
insert an initial title:

[ ] [ + ]

; ; ] ( also place

the homage here; see note 11 below); include a title for TZ within the TZComm cover title:

] ;

include a title for TZ within the main TZComm

title: [ + ]

] |

] ;

omit | 5 insert

includes the root text title, see note 3 above) | 6

(although the shad does not occur in the TZComm title which

; ; . Note that there is not only

variation between editions, but inconsistency in the choice of

consistently gives

insert a Sanskrit title:

or

in the text title: the Dunhuang manuscript alone

, while the Tshal pa Kanjur alone consistently gives ; the Tenjur versions (apart from

) give

in TZComm's cover title, but / elsewhere. The Hemis Kanjur gives here and in this first chapter title, yet

gives everywhere else. Similarly, the Tawang Kanjur gives in Chapter 1, yet / elsewhere, and the Bathang
Kanjur favours throughout the text, but chooses in Chapter 1. The two South Central versions which include a cover

title give

, but all three have / for the regular text title here. At the chapter endings, they more frequently give

, but also occurs. The likelihood is that there was an inconsistency in the archetype, most probably with /

in Chapter 1, and / in all the subsequent chapters. | 7 :

| 8 : (in where cited, f.1v, but note that

note 3 above);

is given in the TZComm title at the beginning of the text, f.1r, see

| 9 :

omit; | 10 : omits

104

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

11
2v.1] [ 101b.4] [ 245.6] [ 177.7]

i[

14

12

[31v] 13

[124] 15 16 17 18 1920

[ 246] [204v]

[ 178] 2122

23 [ 817] 24

2526

27 28

[180v] 29 30

i Throughout the edition, we give a new set of sigla page references each time the TZComm editions have included commentary

following the previous root text passage. Where, as in this case, the previous sigla location list is just above, we give only the
location list for the TZComm texts. We also give a location list at the beginning of each chapter, and for the title at the end of each
chapter.
11

: give the homage in this position (

);

give the homage following the TZComm title, before the root text lemmata begin, with no clear

indication that it is treated as root text (and not highlighted in ); give the

homage following the initial title, before the regular title (see note 3 above); give an alternative homage:

[ ]

It is likely that 's version and

positioning is the original one, and this positioning (shared also with the Bhutanese alternative homage) conforms to the standard
practice of positioning the homage following the title. The Bhutanese alternative homage is likely to have derived from a scribal
confusion, picking up the words from the opening lines of the text below. The Tshal pa Kanjur versions presumably followed
TZComm in giving the homage before the main root text title (although TZComm does position it following the initial commentary
title which of course includes the TZ title). | 12 : ; , with a following letter

deleted | 13 : omit | 14
insert

16

:
;

15

| 17

| 19 :

21 insert

unclear

| 20 :

18

| 22 : omit | 23 :

26 : omit

give the order [with variants noted above]:

27

| 24 : omit | 25 :

omit; omit

| 29

28

; | 30 :

omit

Chapter 1

105

31 3233

34

[46v] 35 36

37 38 39 [321r] 40 41 42
43

44
[ 3r.1] [295r.1] [300a.2] [ 817.2] [597.3] [597.3] [286r.3] [473r.3]

[358v.4] [ 101b.7] [ 246.3] [ 178.2] [124.3] [51r.7] [46v.1] [395.7]

[180v.1] [321r.1] [31v.3] [204v.2] [367v.6]

4546 47 4849 50 51
52

[396] [51v] 53 54 ii

5556 57

58

59 60 61

ii The words from sangs rgyas in the line above, up to here, are cited by Rong zom chos kyi bzang po in an untitled Thor bu, using

exactly the words of our edition, but lacking the shad break following spyod yul (1976 Rong zom bka' 'bum: 375).
31 :

; omit |

32 :

33

[ omit ] |

omit | 35 : ;

| 37 :
| 40

| 41

; |

48

: omit;

53

52

| 54 : omit | 55 :

56 :

49

| 51 inserts (dittographically)

| 42

| 43 :

; omits | 50

; omit | 39

| 36

| 44 : omit | 45 insert | 46 :

omits | 47

; ;

| 38

34

| 58

57 : ;

| 59

| 60 :

; omit | 61 : [] , prefixed is probably intended, but it appears

to have been an amendment by the original scribe of a previously written letter, perhaps ;

106

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

[ 3r.4] [ 102a.1] [ 246.5] [ 178.4]

6263 64 65

66 6768 69

70 71

[ 3v.1] [ 102a.3] [ 247.1] [ 178.5]

72 [368r] 73 74
7576

77 78 79 80

8182

[ 4r.4] [ 102a.8] [ 248.1] [ 179.2]

8384

85 86

87 8889

90 iii

iii This verse is cited in full in Rong zom chos kyi bzang po in an untitled Thor bu, using exactly the words of our edition, apart

from giving nas in place of na in the second line (1976 Rong zom bka' 'bum: 408).
62 :

63 : omit | 64 :

omit; gives

72

| 67 :

65

as a correction, inserted in small writing below | 66

; omits | 68 :

; ;

| 69 :

| 77

81

| 78

:
|

| 79

82 insert

; ;

| 70 : omit | 71

73 : | 74 : omit {but gives it} | 75

: omits |

| 76 :

80 :

(Note that at

first sight, the first part of this insertion appears to be a repetition, but these words are repeated in TZComm for the purpose of
explaining them, mingled in with further comment. The final part, from , is also in the commentarial text and not a
repeat of any root text words. Thus, here, the Bhutanese versions include commentary from TZComm in the root text, unlike the
Tshal pa Kanjur, South Central, or local Kanjur versions.) | 83 : ; adds
below | 84 : ;

| 85

as a correction, in small writing

; the scribe of has effaced about 6 syllables following here. | 86 : omits | 87 :

; | 88

: | 89 : omits | 90 :

Chapter 1

107

[ 5r.2] [ 102b.6] [ 249.1] [ 179.7]

[359r]
9192

93

9495

[ 180]

96

[ 6v.2] [ 103b.1] [ 251.1] [ 181.2]

97 98

99

[ 818] [32r]

100

101102

[47r] [321v] 103

[ 7r.1] [ 103b.4] [ 251.4] [ 181.4]

104 105

106107 108

109110

91 :

; ; ; ; |

; | 93
insert

98

| 94 :

92 :

| 95 : | 96 : | 97

omit (this would seem likely to

be an issue over the boundaries of the root text rather than an accidental scribal omission) | 99 : ;

100

102 :

; ; ; |

(Here,

101 :

103 :

104

alone have a really coherent meaning, also clearly shared in its

essentials by . This reading is also supported by the elaboration which follows in TZComm (

7r). It is most likely that preserve the archetypal

reading, which also implies that the Tenjur, Dunhuang manuscript, Tshal pa Kanjur, and Bhutanese NGB editions [the descendants
of b on our stemma codicum, p.43 above] share an indicative error here. Note that the Tenjur avoids the indicative error in
Chapter 10, although it has its own loss of text). | 105 : ; | 106 : ;

107

omits due to a tear from the page | 108

| 109

: | 110 :

:
;

108

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

111 112
[ 8v.6] [ 104b.3] [ 254.1] [ 183.2]

113 114 [ 9r]

115

116 117 iv

[52r] [598] 118 119120

121122 [181r] 123

[ 10r.6]

[295r.6]

[300a.8]

[ 818.3]

[598.1]

[598.1]

[286v.1]

[473v.3] [359r.4] [ 105a.5] [ 255.6] [ 184.3] [125.4] [52r.1] [47r.2]

[396.6] [181r.1] [321v.2] [204v.8] [32r.3] [368r.6]

124 125 126 127128 129 130

131 132 133

iv Chapter 13 of Klong chen pa's Phyogs bcu mun sel commentary on the rGyud gSang ba'i snying po cites this verse of the Thabs

kyi zhags pa. It gives two of the lines fairly closely, but the first and the final line of the citation are neither in the root text nor
commentary in any version we currently have, although they are clearly in line with the sentiment of our text. The citation reads:
thabs zhags las kyang / /gsang sngags tshul chen shin tu rmad byung ni/ /lam gyi nang nas nye lam chen po ste/ /thabs kyi nang
nas ma nor dam pa nyid/ shes rab nang nas shes rab khyad par can/ /zhes so/ (bdud 'joms bka' ma Volume La: 445-446).
111

: ; | 112 : ; | 113 : ; ;

114 :

: |

115

116 : | 117

| 119

: unclear;

118 :

: uncertain due to damage including a tear from the page | 120 :

slightly preferable, which perhaps explains its introduction) | 121

; (either 's order or the alternative order can fit, and the alternative version may seem

127

| 123

| 124

:
;

{but has };

[] (indistinct, damaged); | 133

| 125 :

: omit | 131

| 122 :

126

: ;

: ; omits, torn from page.

:
128

|
| 130

129

| 132 :

Chapter 2
i [
10v.1]

[295r.7]

[300b.1]

[ 818.3]

[598.1]

[286v.1]

[359r.4]

[598.2]

[286v.2]

[359r.6]

[ 105a.6] [ 256.1] [ 184.4] [52r.1] [47r.3] [396.7] [181r.1] [321v.3]

[204v.9] [32r.4] [368r.6]

123

78 9
1011
1213

[ 11v.2]

14 15

[295r.8]

[300b.2]

[ 818.4]

[ 105b.6] [ 257.4] [ 185.3] [52r.3] [47r.4] [397.1] [181r.2] [321v.4]

[205r.1] [32r.6] [368r.8]

16 17 1819 20

i Following our full collation of Chapter 1, from this point we systematically collate the following witnesses for the rest of the text:

the Dunhuang manuscript; three witnesses of the Tshal pa transmission, the Lithang or 'Jang sa tham Kanjur, the Peking Kanjur,
and the culturally important sDe dge NGB; the Golden Tenjur; two witnesses of the Bhutanese NGB transmission, sGang steng b
and dGra med rtse; two witnesses of the South Central NGB transmission, gTing skyes and Kathmandu; along with the local
Kanjurs of Bathang, Hemis, and Tawang Orgyan ling. We also continue to collate the other two Tenjur witnesses (Peking and
Narthang) for the next chapters until the Tenjur's major omission of text starting in Chapter 6. Furthermore, we collate the
Narthang Kanjur up to the end of Chapter 6, and the two Mongolian witnesses (Urga and Ulan Bator) in Chapters 2 and 3 (we
have only a few folios of Ulan Bator, so it was worth collating the first three chapters in full; the collation shows some
possible relationship with the Narthang/Lhasa sub-group, but not Urga). We continue to add interesting or stemmatically
noteworthy variants found in the other versions at various points of the edition.

| 2 :

omit | 3 : omit | 4

(inserted small beneath, as a correction) | 5

: omits | 6 : omit;

[] (indistinct, damaged);

| 8 :

| 9

| 7 :

omit | 10 : | 11 : | 12 : omits | 13 : | 14 :

15

18

16

: ( in subscribed);

; (a letter presumably sa has been deleted) |

: omit

17

19

( subscribed)

| 20

110

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

2122 23 24 25

21

24

omit

: | 25

22

23 insert

Chapter 3
[ 11v.3]

[295r.8]

[300b.2]

[ 818.5]

[598.2]

[286v.2]

[359r.6]

[286v.3]

[359r.7]

[ 105b.7] [ 257.5] [ 185.4] [52r.3] [47r.4] [397.2] [181r.3] [321v.5]

[205r.1] [32r.6] [368r.8]

1 [368v] 23 4 5
6

[295v] 7

8 9 [ 258]
10

111213

[ 12v.1]

[295v.1]

[300b.3]

[ 818.6]

[598.3]

[ 106a.5] [ 258.6] [ 186.2] [52r.4] [47r.5] [397.3] [181r.4] [321v.6]

[205r.2] [32r.7] [368v.2]

14 1516 17 18
1920 21 i

i Collation of the Urga and Ulan Bator Kanjurs ends here.

1

: ( corrected to beneath the line) |

insert

abbreviated sa);

: |

|6

omit | 5

(though a small sign following the gyi probably represents an inserted

; |

| 4 : ;

2 :

10

| 7 :

: ; |

| 12 :

| 13

11

: omits. (At first sight, appears to give a version of these two tshig rkang, but on closer inspection, and

comparing the text with the Tenjur version, it is clear that 's slightly elaborated reading in fact represents the following lines in
the commentary, which gloss the root text verse. Thus, we can tell that omitted the main root text lines here through accidental

haplography.) | 14 :

15

: ;

| 16 :

omit | 18

; |

17 :

omit | 19 : | 20 : ; | 21

Chapter 4
[ 12v.2] [295v.2] [300b.4] [ 818.7] [286v.4] [ 106a.5] [ 258.6]
[ 186.2] [52r.4] [47r.6] [397.3] [322r.1] [205r.2] [32r.8] [368v.2]

1 [ 259] 2
4 5 6

7 8

9 [ 819] [32v]

10

1112

[ 13r.5]

[300b.5]

[295v.3]

[ 819.1]

[286v.4]

[ 106b.4]

[ 260.2]

[ 187.1] [52r.5] [47r.7] [397.4] [322r.2] [205r.3] [32v.1] [368v.4]

13 14 1516 1718
19 20 2122

(although highlighted in the Dunhuang manuscript, this addition would

seem to be a commentarial gloss and not part of the root text.) | 2 :

omits (but do not omit); ;

);

| 8 : | 9 :

( unclear);

| 5 :

| 10

| 12

| 3 : | 4 :
| 6 :

( agrees)

: unclear; | 11

(but

: omit (the omission is clearly a scribal error; the verse might have been lost through eyeskip, on

.) Note that it is not only the lines of root text, but also some lines of commentary, found in

have been lost here. | 13 :

| 14

| 16

: omit | 15

, which

: ;

| 17

: omits | 18 : omit | 19 : | 20

insert ( do not) | 21 : | 22 :

Chapter 5
[ 13v.1] [295v.3] [300b.5] [ 819.1] [286v.5] [ 106b.5] [ 260.2]
[ 187.1] [52r.6] [47r.7] [397.5] [322r.3] [205r.4] [32v.1] [368v.4]

1 2
3 4 i

5 6 [47v]
7 89

[ 13v.6] [ 106b.8] [ 260.5] [ 187.4]

10 1112 13

14 15

16 17
1819

[ 14r]

20 [ 107a]

[ 14v.6] [ 107a.7] [ 262.1] [ 188.2]

21

22 23

24

i These two lines are cited in gNubs chen Sangs rgyas ye shes' bSam gtan mig sgron, Chapter 6 on Mahyoga: zhags pa las/ kun

tu 'od kyi sa las 'das/ /padma can du gyur ba na/ /zhes pa dang / (gNubs chen Sangs rgyas ye shes: 289). Klong chen pa's Phyogs
bcu mun sel, moreover, cites exactly the same lines with the same wording as gNubs (bdud 'joms bka' ma Volume La: 255).
1

: ; | 2 : ; | 3 :

; ;
;

| 7

| 5

| 6

| 11 :

probably of

| 13

| 9 :

( agrees) | 12 :

| 19 :

[damaged]

; ;

| 20

23

: |

14 :

| 10

15 :

( agrees); [?] (unclear)

| 18 :

; ;

| 21 :

24 :

; ; is damaged, but a is visible,

|
;

17

| 16 :

(note that 's reading fits with the gloss in TZComm below, 14v.5) | 8

[ ] ( unclear);

: ;

22 :

114

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

25

26

[ 15r.4]

[295v.5]

27 [ 15r] 28

[300b.7]

[ 819.4]

[286v.6]

[ 107b.2]

[ 188.5] [52v.1] [47v.2] [397.7] [322r.5] [205r.6] [32v.4] [368v.7]

[ 262.4]

29 30 3132

3334 35 36 37 38

25 : | 26 : | 27

30

31

| 28 : illegible (a gi gu is visible above the line) | 29

: ;

omit ( also omits)

followed by a deleted letter, perhaps originally intended as the zha of the following word, zhags pa

; ;

| 35 :

| 33

| 36 :

| 32 :

omit;

| 34 :

| 37 insert

38

Chapter 6
[ 15r.6]

[295v.6]

[300b7]

[ 819.4]

[286v.7]

[ 107b.2]

[ 188.5] [52v.2] [47v.3] [397.7] [322r.6] [205r.5] [32v.5] [368v.8]

[398] 2 34 5

7 [ 15v] 8

9 [322v] 10 11

[ 262.5]

[287r]

[369r]
12 13

[ 16r.1] [ 107b.7] [ 263.3] [ 189.2]

[301a] 14

15 16 17

18 19

20 21

22

_____________________________________________________________
Note that the Bhutanese edition of TZ includes further text here, which is for the most part found in
TZComm. We are including it here to draw attention to this additional text preserved in the Bhutanese
redaction of TZ, and it seems preferable to present it here rather than relegate it to footnotes. However, it
seems clear that the additions were not part of the archetypal text.
[ 16v.6] [52v.4] [47v.5]

[Comment: Here, only tshogs kyi dkyil 'khor is shared with the commentary: the rest of these two tshig
rkang do not appear to be witnessed in any other text, and they do not appear to be very appropriately
1 insert
omit | 3

[but

: |

15

| 22

20

| 19

| 11 : omit | 12 :

17

14 :

: ; ;

| 8 :

13 :

16 :

| 18 :

appears to read

( shares this reading); |

| 6 : |

| 9 : | 10

] |

: omits | 4 insert ; has a gap, resembling a space between tshig rkang, but

gives no shad. | 5 :

do not have

21

; ;

116

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

placed in connecting them with tshogs kyi dkyil 'khor. Nonetheless, the theme of the yab-yum union
does occur further down in TZComm, in introducing discussion of the mantras given below.23 The
words, tshogs kyi dkyil 'khor, are marked off by the wording of TZComm, as though they were root
text, and the Dunhuang manuscript furthermore highlights them. However, on closer examination, it is
not clear that TZComm is necessarily taking them as root text. Rather, it is drawing attention to them,
since in the context of the chapter, in explaining the principle of everything being integrated into the
maala, it is discussing the meaning of the components of the Tibetan term, as an illustration of the
point.]
[ 18r.2] [52v.4] [47v.5]

24

25

26 27

[Comment: The mantras occur in TZComm (or at least the Dunhuang version of it: the Tenjur has lost
the text at this point), where attention is drawn to them as seed syllables for the meditation on the
practice of union. They are highlighted in the Dunhuang manuscript, probably indicating that it takes
them as root text (it does not generally highlight mantra syllables elsewhere). However, the wording
of TZComm does not clearly signal them as root text.]
_____________________________________________________________
Note that the Tenjur version of TZComm stops running parallel to the text of the Dunhuang ms. in Chapter 6
[at 108a.6, 264.5, 189.7, corresponding to 16v.6], clearly as a result of an accidental
omission (the chapter titles for Chapters 6, 7, 8 and 9 are all missing, and so too are all the root text lemmata
for these sections). The Tenjur texts include just over a side of alternative material, most of which
corresponds to an alternative version of Chapter 6 material, abruptly moving in the last lines (of sixteen tshig
rkang) to the final section of Chapter 10, including the Chapter 10 title.
______________________________________________________________
[ 19v.4] [295v.8] [301a.1] [287r1] [52v.4] [47v.6] [398.2] [322v.2] [205r.7]

[32v.7] [369r.2]

28 2930 31

32 3334 35 i

i Collation of the Narthang Kanjur ends here.

23 The Dunhuang manuscript has the following words:
24

| 25

[further commentary...]
( also gives

:
| 28

| 26

... (

: [further commentary...]

: omit | 29

) | 30 : ;

32 : | 33 : omit; | 34 :

; ;

| 31

18r.1).

| 27 :

; ; | 35

Chapter 7
[ 20r.1] [296r.1] [301a.2] [287r.2] [ omits] [52v.5] [47v.6] [398.3]

[322v.3] [205r.7] [32v.7] [369r.2]

1 2

45
6 78

9
[ 21r.5] 1011

12

13 14 i

1516

[48r]

[ 21v.3] 17

i The Hemis Kanjur completes its folio 32v with

its folio 39.

1

; its next folios are missing and then it picks up the text in Chapter 11 with

omit; [...] (the folio is damaged, but the length of missing letters indicates that it had

either nyon cig, or words of a smilar length.) | 2 inserts (but only highlighting zhes) ;
insert

(Note that only highlights zhes, not the following tshig rkang,

indicating that it considers this line part of the commentary. An opening without this line would seem more appropriate to root
text, and is also consistent with the openings of previous chapters, but the Tshal pa Kanjur tradition's inclusion of it suggests again
some confusion over the boundaries of the root text and its commentary.) | 3 : | 4 : | 5 :

6 :

| 10 :

8 :

| 11 The Hemis Kanjur here has a tear out of the page, so the next words are

missing, until lnga gzhag below, which complete the folio. | 12 : | 13

| 15

| 16 : | 17

: |

14

: ;

(it is possible that intends a ha

a a a, but there does seem to be the sharp bend to the letter which distinguishes ta and da in the Dunhuang manuscript);
omit;

(Here, the Tshal pa Kanjur tradition fails to include the syllables as root text. They are

highlighted in the Dunhuang manuscript, and it would seem appropriate for the main text to include these and the further series of
five syllables below, which are also highlighted in the Dunhuang manuscript. On each occasion, they are connected with the verses
describing each group of five deities arising at the different body parts.)

118

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

_____________________________________________________________________
Note that the Bhutanese manuscripts include further mantra syllables here which are not found in the other
TZ versions. They occur in the Dunhuang manuscript commentary (the Tenjur TZComm version has lost the
text of Chapter 7), but they are only given individually, interspersed between commentarial text. There is
nothing in the commentary's wording to indicate that they are considered root text lemmata and, apart from
one instance of the syllable m ( 22r.5), the Dunhuang manuscript does not highlight them. As in
Chapter 6, we are including these extra mantras here in the main part of our edition, but demarcated by
indentation. Since they represent the popular Bhutanese recension, their relegation to the apparatus would
seem inappropriate. However, they are most unlikely to have been in the TZ archetype.
[ 21v.4-22v.6] [52v.7] [48r.1]

18
____________________________________________________________________
[ 23r.5] [296r.2] [301a.3] [287r.3] [52v.7] [48r.1] [398.4] [322v.4] [205v.1]

[369r.4]

19 20

2122 [53r] 23
24
25

[ 23v.5] 26

_____________________________________________________________________
Here, the Bhutanese manuscripts again include further mantra syllables which are unlikely to have been in
the original root text. See comments on the text just above, found at 52v.7 and 48r.1 (corresponding to
interspersed text, 21v.4-22v.6). These comments apply equally here.
[ 23v.6-24v.4] [53r.1] [48r.2]

27
18

(separated within the text)

; (identical to , but with shads between each syllable, and in place of m at the 8th syllable)

21 :

| 22 :

| 23

| 24

19

: |

20 :

| 25 :

| 26

omit (Again, we have mantra syllables which the Tshal pa Kanjur tradition alone does not recognise as root text, and again

they are highlighted in the Dunhuang manuscript, and would seem appropriate inclusions see note 17 above.) | 27

(separated within the text)

(25 syllables to 's 23, through an additional a ma before ma sha ti)

119

Chapter 7

_____________________________________________________________________
[ 24v.5] [296r.3] [301a.4] [287r.3] [53r.2] [48r.2] [398.5] [322v.5] [205v.1]

[369r.5]

28

29

30 31

32

[ 25r.5] 33

___________________________________________________________________
Here again, the Bhutanese manuscripts include further mantra syllables which are unlikely to have been in
the original root text. See comments p.118 above, on the text found at 52v.7 and 48r.1 (corresponding
to interspersed text, 21v.4-22v.6).
[ 25r.6-26r.2] [53r.3] [48r.3]

34
___________________________________________________________________
[ 26r.5] [296r.3] [301a.4] [287r.4] [53r.3] [48r.4] [398.6] [322v.5] [205v.2]
[369r.6]

35

36 [323r] 37

38
39

[ 26v.3] 40

28 :

29

: |

: ; ; |

30 :
32

; (, however, gives ); |

: |

33

31

: omit

(Again, we have the same issue of the extent of the root text, with the Tshal pa Kanjur tradition consistently omitting these series
of mantra syllables see note 17 above.) | 34
(separated within the text)

37

: ;

| 38

| 35 :

| 36 :

| 39 : ; | 40

omit (Again, the Tshal pa Kanjur tradition does not recognise these mantra syllables as root text see note 17 above.)

120

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

___________________________________________________________________
Here again, the Bhutanese manuscripts include further mantra syllables which are unlikely to have been in
the original root text. See comments p.118 above, on the text found at 52v.7 and 48r.1 (corresponding
to interspersed text, 21v.4-22v.6).
[ 26v.3-27v.2] [53r.4] [48r.5]

41
____________________________________________________________________
[ 27v.4] [296r.4] [301a.5] [287r.4] [53r.4] [48r.5] [398.7] [323r.1] [205v.2]
[369r.7]

42

43 44

45
46

[ 28r.4] 47

___________________________________________________________________
Here again, the Bhutanese manuscripts include further mantra syllables which are unlikely to have been in
the original root text. See comments p.118 above, on the text found at 52v.7 and 48r.1 (corresponding
to interspersed text, 21v.4-22v.6).
[ 28r.5-29r.1] [53r.5] [48r.6]

48
____________________________________________________________________
[ 29r.2] [296r.5] [301a.6] [287r.5] [53r.6] [48r.6] [398.7] [323r.2] [205v.3]
[369r.7]

49 [399] 50
:

41

| 45

(separated

within

| 42 :

| 46 :

| 47

the

| 43

text)

: |

44

omit;

(Again, the Tshal pa Kanjur tradition does not recognise these mantra syllables as root text see note 17 above.)

48

; ;

| 49

(separated within the text)

: ; ;

| 50 :

121

Chapter 7

51

52 53 54 55 56 57
58

[369v] 59 60 61 62

51

omit; has double tsheg | 52

| 54

: ; ;

| 58 : omit | 59 : | 60

omit | 53

| 55 insert

| 56 inserts

| 57

omit (The chapter title is perfectly coherent without these words, but it is quite

likely that the omission is an indicative error shared by the descendants of hypearchetype c [as shown in our stemma codicum, p.43
above], namely, the Dunhuang manuscript, the Tshal pa Kanjur and the Bhutanese editions, or possibly, hypearchetype b [but this
cannot be ascertained since the Tenjur is missing]. The opening of TZComm's Chapter 7 describes the content of the chapter as
follows:
;

; | 62 :

19v.6) | 61 :

Chapter 8
[ 29r.4] [296r.5] [301a.6] [287r.5] [ omits] [53r.7] [48r.7] [399.2]
[323r.3] [205v.4] [ omits] [369v.1]

2 3 [48v]
4

6
_____________________________________________________________________
Note that the Tshal pa Kanjur versions of TZ add two verses here. At first sight, these verses appear to be
clearly marked off by the wording of the commentary, and apart from the final two tshig rkang of the second
verse, they are highlighted in the Dunhuang manuscript (the Tenjur has lost the text of Chapter 8), suggesting
that it takes them as root text. TZComm marks off each verse with the words, "zhes brjod de" (the same
wording used in Chapter 6 for one of the mantras accepted only by the Bhutanese versions as root text, and
for a mantra in Chapter 19 which is not accepted by the Tshal pa Kanjur tradition) rather than, "zhes gsungs
te", the usual words which mark the root text lemmata. Although they read as appropriate root text, the root
text reads perfectly coherently without them. It seems clear that the Tshal pa Kanjur editors made an error in
including these verses in the root text. We include them here as representing the Tshal pa Kanjur tradition.
[ 29v.3] [296r.6] [301a.7] [287r.6]

7 8 9

1011

12 13

14

1 inserts but without highlighting

insert [ insert

[ inserts ] ( 's lack of highlighting indicates that it considers it part of the

commentary. As noted in the similar extra line at the beginning of Chapter 7, an opening without this line would seem more
appropriate to root text, and is also consistent with the openings of previous chapters, but the Tshal pa Kanjur tradition's inclusion
of it suggests again some confusion over the boundaries of the root text and its commentary.) | 2 :

( also gives ) | 3 : | 4 :
;

|6

| 5 :
| 7

; ;

; ; ; ;

: |

: | 9 : |

10

: |

11 :

| 12 : ; | 13 : ; | 14

123

Chapter 8

[ 29v.6] 1516

17 [301b]

18

[ 30r] 1920

21

_____________________________________________________________________
Note that the following mantra syllables are found only in the Bhutanese TZ manuscripts. They are found in
the Dunhuang TZComm manuscript, and the first four are highlighted but like the series of additional mantra
syllables included in the Bhutanese version of Chapter 7, they are only given individually, interspersed
between commentarial text. There is nothing in the commentary's wording to indicate that they are
considered root text lemmata. We include them here as representing the Bhutanese tradition.
[ 30r.2-30v.3] [53v.1] [48v.1]

22
_____________________________________________________________________
[ 31r.2] [296r.8] [301b.1] [287r.7] [53v.1] [48v.1] [399.3] [323r.4] [205v.5]
[369v.2]

[287v] 23 24

25

26

27 28

29

[ 31v.4] 30 31

3233 [296v]
34 35

3637

15 :
21

16 :

22

17 :

30

| 34 :

| 35

28

(separated within the text)

; ;
27

18 :

| 24

| 31

19 :

| 25

23

20 :

| 26 :

29

| 32 : | 33 :

; |

36 :

; ; ; (here, bsgrags would seem a more appropriate reading, and it

is conceivable that bsgrub/bsgrubs might represent a shared error of , but it is a possible reading and given the

124

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

[ 32r.5] 38 39

4041

42 [323v] 43

44

[ 32v.5] [296v.2] [301b.3] [287v.2] [53v.3] [48v.3] [399.5] [323v.1] [205v.7]

[369v.5]

45 4647

4849 50 51

weight of evidence supporting the stemma discussed in the introductory chapters, it is more plausible that an ancestor of the other
editions [hypearchetype c, or possibly, b, on our stemma codicum, p.43 above] introduced bsgrags.)
38

(here
;

44

( also gives );

| 40 : ; | 41 :

| 47 :

| 45

43

: omit | 46

: ;

| 42 :

: ; ;

; ; ; |

; | 50 : ; | 51

37

; ; ; ;

would seem to be in error with bskyed, but read bskyod.)

| 39

48

: omit |

49 :

Chapter 9
[ 33r.1] [296v.2] [301b.4] [287v.2] [ omits] [53v.4] [48v.4] [399.5]

[323v.1] [205v.7] [ omits] [369v.6]

1234 5
i
ii

i The peaceful deities of the maala, listed also in TZComm's Chapter 7, and below in TZ's Chapter 10 on mudrs, correspond to

a considerable extent with the deities of the Vajradhtu mahmaala of the Sarvatathgatatattvasagraha (henceforth STTS).
Here we note their Sanskrit equivalent names in the STTS where appropriate, and possible Sanskrit equivalents for the Tibetan
transcription of the name when there is no parallel deity in the STTS. This first and central deity is Vairocana. Besides TZComm's
Chapter 7 and TZ's Chapter 10, the Tibetan names are also noted in the marginal annotations given in the Dunhuang manuscript
version of TZComm's Chapter 9 list of mantras (see our TZComm edition, and the Appendix).
ii There are two puzzles in the Thabs zhags' tradition of the first set of four in Vairocana's circle. First, they are presented as
Vairocana's male group here [the first group of five including Vairocana, out of the twenty-five male deities, listed first, without
feminine particles]. They are explicitly explained as male, uniting with their female consorts, in the description in TZComm
(Dunhuang ms. 21v, 22v). Yet, the Tibetan version of the names given below in the TZ Chapter 10 mudr list give them with
female particles, and these feminine names would fit with the set in STTS. The second puzzle is inconsistency in the word order
of each name, that is, whether vajra/rdo rje comes first or after the other element of the name. It appears that apart from
Dharmavajra below, it is most likely that the archetype gave vajra first in these Chapter 9 mantras, as well as in the Chapter 10
mudr list. It may be that this order is correct for the Thabs zhags tradition, even though it is opposite to the STTS tradition, and it
moreover means that the same names are repeated, one in each of the other four buddhas' male groups. However, it may
alternatively mean that the archetype was in error, perhaps having banalised the less expected forms of sa twa badzra etc. into the
more familiar forms of badzra sa twa etc. In either case, it seems likely that an ancestor of the Dunhuang manuscript and the
Tshal pa Kanjur and Bhutanese NGB texts attempted to 'correct' these mantras. We cannot be certain; given the inconsistencies
across the chapters, there is a possibility that the archetype had the correct mantras which were then separately corrupted in the
local Kanjurs and South Central NGB traditions. Yet we must consider the fact that mantras are more often subjected to editorial
attention than other passages of text, as Tibetan editors engage in the constant struggle to achieve 'correct' Sanskrit renderings as
they go along. Moreover, it would have been more straightforward for the editors of the commentarial transmission to crosscheck the list in TZComm's Chapter 7. The Dunhuang manuscript is the only version we have for TZComm's Chapter 7; it gives
rdo rje following the other element of the name in three out of four cases, and is inconsistent in the case of rdo rje sems dpa'/sems
dpa' rdo rje. Note that the Dunhuang manuscript and the Tshal pa Kanjur and Bhutanese NGB traditions of this Chapter 9 share
the readings consistent with the STTS order of the names in three out of four cases, and in the fourth case, only the Bhutanese
amongst these groups give the apparently less 'correct' form.
1 :
6

; |

2 inserts

3 inserts

4 :

5 :

; ; ; [...];
;

Sattvavajr/Vajrapramit.) | 8

illegible;

|7

(STTS:

: ; ; ;

(STTS: Ratnavajr/Ratnapramit.)

126

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

10

11

12
13

14

15
16

17
18

19

20

(STTS: Dharmavajr/Dharmapramit.) | 10

Vajrasattva.)

11

: ;

(STTS:

: (dittography follows, but deletion is indicated by a faint line);

; (STTS: Vajrarga.) |

; ; ;

: ; ; ; (STTS:

13

(STTS: Vajrarja.) | 14

; (STTS: Karmavajr/Karmapramit.) |

Akobhya.) | 12

(STTS: Vajrasdhu.) | 16

15

: ; ;

: : ;

[...];

; (STTS: Ratnasambhava. The omission of badzra in may have been in the archetype, and

could have been corrected in the other editions.) | 17

: (inserted as a correction above the line);

; ; omit (STTS: Vajraratna.) |

18

: ;

; ; ; (STTS: Vajratejas. TZ seems to intend Vajrasrya;

the Tibetan name given in TZComm Chapter 7 is rdo rje gzi brjid, but in the TZ Chapter 10 mudr list [see p.136 below], the
Tibetan name is given as rdo rje nyi ma.) | 19

: ; ; ;

unclear [ ?]; (STTS: Vajraketu. The deity indicated here is Vajra Victory Banner; in Tibetan, rdo rje

rgyal mtshan. It would seem that the Tshal pa Kanjur tradition has attempted to construct a "correct" translation [just as it gives
mui for Vajra Fist rdo rje khu tshur below, where the earlier reading is likely to have been, san dhi].) | 20 :

; ; ; (STTS: Vajrahsa.)

127

Chapter 9

21

22

23

[400] 24

25

[ 33av.1] [206r]

[ 33r.6] [370r]

27

28

29

30

21

26

[?] ;

unclear

; ; ; (STTS: Loke vararja = Amitbha.)

:

22
23

omit;

(STTS: Vajratka.) | 24

(STTS: Vajradharma.)
;

: ; ; ;

(STTS: Vajrahetu. Vajracakra in the TZ mantra [and the name, rdo

rje 'khor lo, given in TZComm Chapter 7, Dunhuang ms. 26v, and in TZ Chapter 10 p.137 below] might seem surprising for this
deity, but Vajrahetu's associated samdhi is Sahacittotpdita-dharmacakra-pravartin.) | 25 : :

(STTS: Vajrabha. Note that

repetition is clearly deleted by small triangles of dots above the syllables.) | 26

(incorrectly placed at the beginning of the following folio);

repeats this mantra, but the

; (the mo is inserted as a correction between a and ka); [

];

28
29
30

(STTS: Amoghasiddhi.) |

27

(STTS: Vajrakarma.)

(STTS: Vajraraka.)

(STTS: Vajrayaka.)

: ; ; ; ;

(STTS: Vajrasandhi. The deity indicated here is Vajra Fist; in Tibetan, rdo rje khu tshur. It would seem that the Tshal pa Kanjur
tradition has attempted to construct a "correct" translation [just as it gives variants on dhwadza for Vajra Victory Banner rdo rje
rgyal mtshan above, where an earlier reading is likely to have been, ke tu], but it is likely that our other readings reflect the STTS
Vajrasandhi. The word, sandhi, may seem unrelated to khu tshur, but the associated samdhi of this deity is that of sarvatathgata
mui, the fist of all the buddhas.)

128

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

[ 33av.1]31 32

33

34

[49r] 35

36

37

38

39

40 41

31 Here, the Dunhuang manuscript ( 33av.1) gives,

, given in the appropriate position above in all other

versions. | 32

final

: ; ; ;

(possibly,

's omission of the

might indicate an error in the archetype? The intended name here is not certain, and the name is not given in STTS;

possibly, Vajre varadhtu is intended, a reversed variant of the more usual Vajradhtv var, but this would not seem to fit with
either of the two Tibetan names given in TZComm for Vairocana's consort, kun tu bzang mo and thams cad bdag nyid ma.
However, the Tibetan name of thams cad bdag nyid ma might correspond to Vi ve var.) | 33

given in STTS.) | 35

(Buddhalocan; not given in STTS.) | 34

: ;

(Mmak; not

; ;

[ ];

(Paravsin; not

: ; ; ;

given in STTS.) | 36

; ; (Samayatr; not given in STTS.) | 37 :

; (Sthiravajr; this female deity is not given in STTS, but the Tibetan name given in TZComm

Chapter 7 is brtan ma rdo rje.) | 38

: omit ( also omits); ***** (this female deity is

not given in STTS; the name should be equivalent to 'bar ma rdo rje, given in TZComm Chapter 7, thus suggesting, Jvlavajr or
Jvalavajr.) | 39

: [ ] ; omits (this female deity is not given in STTS; the

name should be equivalent to sdud ma rdo rje, given in TZComm Chapter 7, thus suggesting, Saharaavajr.)
40

(this female deity is not given in STTS; the name should be

equivalent to bskyod ma rdo rje, given in TZComm Chapter 7. The intended Sanskrit name here is very uncertain, but possibly,
Cacalavajr or Sacalavajr?) | 41

omit (This omission of the four females of Vairocana's circle, as given in TZComm Chapter 7, appears to be an error shared by the
Tshal pa Kanjur, Bhutanese and Dunhuang versions. The annotator of the Dunhuang manuscript seems to have noticed the
absence, and inappropriately labelled the final four mantras in Chapter 9 as corresponding to these four female deities.)

129

Chapter 9

4243
44

45

46
47

4849

[324] 50
51

52

53

42

: ; ;

(STTS: Vajralsy.) | 43 inserts

| 44

: ; ; ;

(STTS: Vajradhp.) |

45

; ; ;

(STTS: Vajrku a.) | 46

[then inserted small apparently as a correction in the left margin, vertically, its positioning indicated by a cross:]

(here, the letter de also has a zhabs kyu which has been deleted);
; ; [

] ; ;

(this female deity is not given in STTS; the name should be equivalent to rdo rje sgril ma, or even just possibly rdo rje

sgrol ma, but the mantra transcription would not seem to correspond to any likely Sanskrit related to either of these names.)
47

Vajraml.) | 48

(STTS:

: ; ; ; ;

; omits (STTS: Vajrapup.) |

49 insert

; inserts (here, seems only to have

made an error of positioning: this mantra occurs below in ; appears inappropriately to give it twice. It is possible,

since both and give it in this position, that it was present here in the archetype, perhaps repeated as in , but

since it is hard to judge whether , or preserve the archetypal version in relation to this mantra, we are accepting 's
more appropriate positioning.) | 50

(STTS: Vajrap a.) | 51

omits;

: ; ; ;

; omits here ( gives it above; see note 49. It is quite likely here that the ancestor of the Tshal pa

Kanjur, Dunhuang manuscript and Bhutanese editions lost this mantra, and then constructed an appropriate mantra to fill the slot,
perhaps intending Vajraratnarati. This female deity is not given in STTS; the name should be equivalent to rdo rje bde ba ma, and
it is most likely that it should be Vajrasukh.) | 52

: ; ;

; ; ; ; (STTS: Vajragt.) | 53 :
;

(STTS: Vajrlok.)

( ) ;

130

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

54

55

56

iii

57

58

59

60

61

62

63

iii It is not clear what the final four mantras represent. They do not appear to relate to any of the twenty-five deities of the

maala (although the annotations in the Dunhuang manuscript seem incorrectly to associate them with the females of Vairocana's
circle). They all give the name, Vajramui. It is possible that they might be meant to represent four protective figures or
gatekeepers in the four directions on the outer periphery of the maala. Alternatively, since the Vajra Fist is a well-known mudr,
it is also possible that the mantras relate to a ritual meditation using the mudr, to close the chapter.
54

: : ; ; ; ;

(STTS: Vajrasphoa.) |

55

: ; omit;

(this mantra, for rdo rje gsal bkra ma, is expected here; it is witnessed by the Bathang edition, Tawang, the Bhutanese edition

and the Dunhuang manuscript. This female deity is not given in STTS and the intended transcribed name is uncertain, possibly
Vajrargarati or Vajrarokarati?) | 56 : (here, the letter ha also has an attached ra and a gi gu, which
have been deleted);

; ; ; ; ;

(STTS: Vajranty.) |

57

which have been deleted);

58

: (here, the letter ha also has an attached ra and a gi gu,

; ;

; omits (STTS: Vajragandh.)

: [ (/)] (there is a clear 'greng bu on dha, but its deletion may be intended; moreover, the letter

ha also has an attached ra and a gi gu, which have been deleted);

; omits; (this female deity is not given in STTS; the name should be equivalent to rdo rje dril

bu, given in Chapter 10 below, or rdo rje dril 'khrol ma, given in TZComm Chapter 7, thus suggesting, Vajragha.)
59

(STTS: Vajrve a; here, the transcription of

Vajrarati does not seem to fit with the STTS, nor its equivalent Tibetan name, which should be rdo rje 'bebs pa ma.)
60

; omit | 62

; ; ; ;
; ; ;

| 63

61

131

Chapter 9

[ 33av.6] [296v.6] [301b.8] [287v.6] [54r.3] [49r.3] [400.4] [324r.2] [206r.3]

[370r.5]

64 65 6667 68
[302a] 69 70

64

: omit |

65

: |

66

: ; ; |

; ; | 68 : omit | 69 :

67 :

; | 70 :

Chapter 10
[ 34r.1] [296v.7] [302a.1] [287v.6] [ omits] [54r.4] [49r.3] [400.5]

[324r.2] [206r.4] [ omits] [370r.5]

23

4 5

6 7
[ 34r.4] 8 9

[601] i 10

11 12

13 [297r]

14 15

16

171819
2021 22

i At this point, folio 287 ends; in the Bodleian print, folio 287 is represented twice, while folio 288 is missing. We have

therefore collated this folio in .

1 inserts without highlighting

insert

(As noted in the similar extra lines at the beginning of Chapters 7 and 8, an opening without this line would seem

more appropriate to root text, and is also consistent with the openings of previous chapters, but the Tshal pa Kanjur tradition's
inclusion of it suggests again some confusion over the boundaries of the root text and its commentary. In this case, does not

highlight any of this chapter, so it is not possible to judge the highlighter's assessment.) | 2 : | 3 :

; | 4 : | 5 : ; | 6 : ;

; (but gives ); |

insert [ ; ] [ inserts ] (Again, this would appear to be a line of commentary

intruding into the root text in the Tshal pa Kanjur versions.) | 9 : | 10 : ;

; |

16 :

11 :

14 :

| 17 :

12

: ; ; |

15

13 :

: [?] (final nga and following tsheg small and uncertain)

; |

18 :

; ; |

; ; ; ; | 20 :

19 :

| 21 :

( gives ); unclear; | 22 : ; ; ; unclear

133

Chapter 10

2324 ii

25 [ 34v]
27 28

29

30 31

26

32 [370v]

[401] 33 34

35 [54v] 3637

38 3940
41

4243

[49v] 44 45

4647
48

49 [324v]
ii Note that Klong chen pa cites this verse in his Phyogs bcu mun sel, in the context of discussing the mudr representing the union

of the five family male and female deities, with their seed syllables visualised at the tips of the fingers of the left and right hands,
which are brought together. The whole passage reads: spyi'i phyag rgya ni g.yas kyi mdzub mo'i o bsam pa de dang g.yon gyi
gung mo'i m bsam pa de gnyis thad kar rtse sbyar te lhag ma g.yas g.yon gyi h dang l/ sw dang m/ dang py/ h
dang t ste yi ge 'bru brgyad so sor bsnol te rtse mo sprad pa ni rigs lnga'i rgyal ba kun dang ngo bo mnyam pa nyid du sbyor
ba'i dngos grub bde ba chen po tshe 'di nyid la sbyin par byed pa'i phyag rgya'o/ /de skad du'ang thabs zhags las/ sor mo nang du
bsnol te bkan/ /gung mdzub mthe bong che chung rnams/ /so sor rtse sprad phyir bskyed de/ /srin lag gshibs bsgreng dam tshig
rgya/ /zhes so/ /'di la sangs rgyas gsang ba ni rdo rje bsdam pa las gung mo gnyis gshibs te brkyangs pa thugs ye shes rtse gcig
pa'i phyag rgya ste rigs lnga spyi'i dam tshig gi rgyar 'dod do/ /kha cig rdo rje thal mo las g.yas pa'i sdig mdzub dang g.yon ma'i
gung mo gnyis rtse sprad de lhag ma rnams so sor sbyar bar 'dod de/ de'ang rnam pa ni thabs zhags ltar lhongs shig / (bdud 'joms
bka' ma Volume La: 279-280).
23

24

; ;

; unclear | 34

| 26

| 28 : | 29

: ; ;

| 27

| 32 :

unclear | 47 :

45

; [-] (unclear)

: ; ;

| 48

| 35 : | 36 : | 37 : | 38

| 41 :

| 30 :

| 42 :

| 40 :

; ; ; ; ; ; |

; ; unclear |

| 31

| 39

; ;
:

[?] (as correction below the line);

| 25 :

; ;

33

; ; unclear; omits |

| 43

44 :

46 : ;

: ; ; ; |

49 :

134

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

50 51

52

535455

5657 58
59 60

61 [206v]

62

6364 65

66

6768

69 [ 35r] 70

7172 7374

75

76

7778

79

80

81

82

8384 85

50 :

; ; |

; | 52 :

60

62

| 58 insert

; ;

| 65

77

70 :

| 73

83

; ;

; ; |

71 :

| 63 :

; |

| 74

: [?]

84

; ; ;

66

: unclear, though the

| 68 : ;

69 :

; ;

| 75 : | 76 :

| 81

; ; ;

72 :

| 79 :

56

(dittography, shared also by ) | 59 :

; ( gives );

55 :

: ;

| 78 :

: ; |

61

superscribed ra and the gi gu are visible | 67

: ;

| 64 :

| 53 : ; ; ; | 54 : ;

57 :

51

| 80

: ;

: |

82 :
85

135

Chapter 10

86 8788

[302b] 89 90
91 92

93 94

9596

97 98

99

100101102
103

104

105

[55r] 106107108

109

[ 35v] 110

111 112

113 114115 [297v]

[371r] 116 117 118 [325r]

[402] 119

120121 122 [602] 123124

86 : ;

appropriate reading, gcu.) |

89

87

(perhaps bcu was in the archetype and emended in , and to the more

| 90 :

| 97

103 :

| 94 :

100 :

| 91

101 :

; ; |

| 92

88

: (here, the prefixed letter ba has a

; ;

| 95 :

96 :

; | 98 : ; | 99 : ;

'greng bu which has been deleted);

| 93 :

| 109

104 :

; ; ; |

105 :

; |

102 :

106 :

| 107 : ; ; | 108 :

; omit ( also omits) | 111

110

: ; |

112 :

; |

113 :

; | 114 : [ ?] (attached ya not visible) | 115 : ; ;

; |

116 :

| 117 :

| 118

(but gives );

: ; |

| 121 : | 122

119

: omits |

120

: [-] | 123 : ;

; ; | 124 : [(/)];

136

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

125 126 127

128129130131

132 [50r] 133

134 135

136

137 138 139

140

141

142143 144

145

146147 148

149 150151
152153

154155

156 157158
159 160161

163

125

: omits |

126

162

164

: |

127 :

128 :

; |

129 :

; ; ; | 130 : omits | 131 :

132 :

| 137

;
|

| 134 :

| 138

| 149

152

144

| 135

| 150

| 142

; | 148

; ;

| 136 :

| 139 : | 140 :

| 146 : | 147 :

: unclear, [ (/ )] ;

| 145 :

133

| 141

| 143 :

, corrected from

: ;

: ; ; |

; ; ; |

151 :

153 :

| 154 : | 155 : ; | 156 : omits; | 157 :

;

| 160

| 158 :

: ;

(dittography) | 161 : [?] unclear;

| 163 : | 164

| 159 : [?] (in both cases unclear);

162

137

Chapter 10

[ 36r] 165 166

167

168

169170171172
173

174 175

176

177 178

179 180 181

182183 [207r]
184

185 186

187 188

189190191 192
193194

196

195

[55v] 197 198

199200
165 :

; |

201

166 inserts

(this would not seem to improve the description, and

moreover, where other mudrs in the sequence are described in 4 line verses, we already have 5 lines here, so this would take it to
6. It could possibly indicate a longer omission, but this seems unlikely, since rdo rje rnon po follows rdo rje chos in the list spelt
out in full in the commentary to Chapter 7 above. This line could have been picked up accidentally from the text below: it is the
first line of the verse for rdo rje smra ba's mudr.) | 167 : ; | 168 : ;

; ; |

169 : ;

; omits |

170 :

; ; ; | 171 : omits | 172 : [?] | 173 :

[(/)] (an indistinct mark might represent a very effaced 'greng bu)

| 174 :

185

183

| 197 : ; | 198

; | 201

| 186 :

[-] (unclear) | 189 :

| 176

| 187

[?]

| 191 :

: ;

[?] | 180 :

; ;

| 188

: ;

| 194 :

| 196 :

: ;

177

| 182 :

184

| 190 :

| 193 :

| 192 :

195

inserted beneath line) | 179

| 181

: ;

; (

; ;

| 178 :

| 175 :

; (inserted beneath line)

| 199 : ; | 200 :

138

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

202203

204

205

206 207 [325v] [303a] 208

209 [ 36v] 210

211

212213

[371v] 214
215

216

217 [50v] 218

219 220

[403] 221 222

223
224

225 226227 228

[298r] 229 230

231232

202

[-]

205 :

208

204

203

| 206

| 222 :

| 223

229 :

:
;

unclear;

210 :

| 216 :

| 220 :

226

230

| 231 :

: |

227 :

| 213 :

; ; |

; ; ;

[ ?]; [?]; ;

| 215 :

; |

[ unclear] ;

: ;

; (sic) | 225 :

| 212 :

214 : ;

| 219 :

| 207

| 209 :

211 :

| 218 :

| 217

| 221 :

| 224 :

228 :

;
232 :

139

Chapter 10

[289r] iii 233 234

235236

237

238

239 240 241

242243244245

246

247248249 [ 37r]

250

251 252
253

254

255 256257
258

iii The Bodleian version of , which has folio 288 missing, rejoins here with folio 289r.
233 :

| 237

234 :

235

: ; |

236 :

; ( also shares the reading, ;

agree with rdo rje 'dzin ma. The confusion here is most probably related to the fact that the text seems to lose its logical

(and systematic) sequencing of the deities at this point, and it is hard to assess the preferable reading. We might expect the male
members of Vairocana's circle, starting with Vajrasattva or Sattvavajra (see Chapter 9 note ii above), to be listed from here, but
they appear to have been omitted, while the first few of the following female deities do not seem to be in a clear order.
Nonetheless, rdo rje 'dzin ma does occur as an alternative name for rin chen spyan while rdo rje 'dzin pa does not. It would
seem that the female deities here perhaps correspond with four of the principal females, if we take rdo rje 'dzin ma as representing
rin chen spyan, while the females of the retinues are listed afterwards.) | 238 : | 239 : | 240 :

[?]; ; ; ; | 241 : omit | 242

: omits (eyeskip, from ) | 243 : ; ; ; | 244 :

; ; |

245 :

; ; ; |

248 :

| 249 :

| 250 :

deleted) | 252 : | 253

; ; ;

256 :

246 :

| 251

247 :

; ( also shares this reading);

[ ?] (the preceding ca appears to be

; ; |

254 :

255 :

257 : | 258

(The text has at this point lost its clear

sequencing. However, it would seem more likely here that we should have rdo rje sgrol ma, perhaps to be equated with the
principal female of Amoghasiddhi's circle, dam tshig sgrol ma. The positioning of rdo rje sgril ma makes better sense below,
where she is given in appropriate order by the Dunhuang manuscript version once the text recovers an obviously logical
sequencing. 's rdo rje sgron ma, not shared with , could represent rdo rje me sgron ma in Amitbha's circle, but

she appears to be given in the appropriate order below, in the text preserved only by , under the name, rdo rje mar
me. It is more likely that rdo rje sgron ma here is a corruption of the archetypal reading, rdo rje sgrol ma.)

140

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

259 260

261

262263

264 265

266

[56r] 267 268

269270

271272
273

[326r] 274
275276277
259 :

260 :

| 262 :

| 266

| 263 :

261

| 264

; ; ;

: |

265 :

: ; ; (Here, it is

hard to be sure what the archetype read. The Bathang manuscript and the South Central tradition including share the reading,

along with the Dunhuang manuscript, although

, as well as the Tshal pa Kanjur and Bhutanese versions, give

. Here, we would expect a female deity; it may be that the archetype read in error. Neither kun tu bzang po nor

kun tu bzang mo are in our list of deities of the peaceful maala given in Chapter 7 of TZComm, nor obviously in the mantra list
in TZ Chapter 9. However, kun tu bzang mo does feature in TZComm Chapters 5 and 6. In Chapter 5, she is specified as the
principal female, implying that she is Vairocana's consort. In Chapter 6, she is said to embody all touchables, and is thus again
linked with Vairocana [body]. In that list of the five buddhas and goddesses associated with sensual data, each of the other four
goddesses are the second females of the appropriate buddhas' circles. This would suggest kun tu bzang mo taking the place of
brtan ma rdo rje, given in Chapters 7 and 9 as Vairocana's second female. But it is more likely here that she is acting as
Vairocana's consort [principal female], just as she is presented in Chapter 5. The name for Vairocana's consort in Chapter 7 is
thams cad bdag nyid ma, and she is not mentioned at all in this mudr list [unless she is intended to be represented by the mudr
of union, which is given following Vairocana's mudr]. Although there are other apparent discrepancies in the list [such as the
apparent omission of the consort of Amitbha, gos dkar ma], it is probable that here, kun tu bzang mo is intended as the consort of
Vairocana, all the more likely since her mudr immediately follows those which seem to be for the principal females three of the
other buddha families.) | 267 : | 268 : ; ;

; ; |

269

: ; (Here, compare the line

which occurs below in the passage omitted in the descendants of hypearchetype c [as given in our stemma codicum, p.43 above],
namely, the Dunhuang manuscript, and the Tshal pa Kanjur and Bhutanese witnesses: ldan pa g.yas su byol pa las.) | 270 :

(note that give byol in the later verse which has parallel phrasing to this here, although the Tenjur reading is

again byo.) | 271 : ; ; | 272 :

; |

274

| 273 : ;

: [ ?] ; ; ; ;

; (Here, 's rkong rkong quite likely indicates, kong kong: the left (hand) is cupped. None

of the other variants seem entirely coherent.) | 275 : | 276 : ; | 277 :

141

Chapter 10

278 279280

281

282

283 284

285 286 287

288

[ 37v] 289290

291

[207v] 292

293

294

295296297

298299300 301

278 :

; |

285 :

287

279

: ; |

| 282 : | 283 :

280 : [?] | 281

| 284

: omit ( also omits)

; ; ; ; |

288

286

: ;

(There

are metrical issues here; apart from , all versions attempt to add the feminine particle, but this renders them unmetrical unless
the phrase phyag rgya yin is modified. However, in this sequence in the chapter, phyag rgya yin would seem to be expected, and
we find it in [as well as ], although not in , which gives, . It may be that the archetype

was unmetrical here.) | 289 : ; ; ; ; | 290 :

291

: |

292 :

; |

293 :

294

: ;

; ; (As noted above, rdo rje sgril ma is expected at this point in the sequence; it seems
most likely that the archetype was mistaken here, giving a na ro instead of a gi gu. rDo rje sgrol ma is not so likely here, and
probably occurs above [see note 258], while rdo rje sgron ma occurs as rdo rje mar me in the passage below preserved only by
. There is no rdo rje sgra ma in the list of deities given in Chapter 7 of the commentary or with the mantras in

Chapter 9. However, since only the Dunhuang manuscript gives the expected sgril ma, it is most likely that sgrol ma was given in
an earlier ancestor, and may have been corrected in the Dunhuang ms. However, we need to exercise some caution, given that this
Chapter 10 list does not follow an apparently logical sequence throughout, seems to omit some deities, and seems partially
corrupted. We also need to be cautious in following TZComm's Chapter 7 list, since it may have been to some extent tidying up a
less tidy picture of the maala as presented in the root text.

Moreover, TZComm itself seems to have some internal

inconsistencies, such as giving the name, kun tu bzang mo, for Vairocana's consort in Chapter 5, and the name, thams cad bdag
nyid ma, in Chapter 7.) | 295 : | 296 : inserts as a correction between dang and mtheu | 297 :

; ; ; ; |

298 :

| 301 : [?] (corrected, but the intended reading is not clear)

299 :

; |

300 :

142

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

302 [51r] [372r] 303

304

[Here we include the passage of text which concludes Chapter 10 in the South Central NGB tradition
(), and in the local Kanjurs of Bathang and Tawang (Hemis is missing these folios). It is omitted
from the other TZ versions (including also 826.5; 603.5; 477v.5; 362v.5;
134.2) and the Dunhuang manuscript, but there is little doubt that it was in the archetype (see the
discussion above, p.51), so we are restoring it here. We can infer that it must have been present in
hypearchetype b (as given in our stemma codicum, p.43 above), since the Tenjur tradition
( ) preserves the final lines of the chapter (apart from an obvious eyeskip, which
results in the loss of a few tshig rkang, see below). The Tenjur redaction has its own unrelated long lacuna
from Chapter 6 up until this final section of Chapter 10, including some of the passage, but not all of it. The
absence of the entire passage in the Dunhuang, Tshal pa Kanjur and Bhutanese witnesses must be seen as an
indicative error peculiar to them as descendants of hypearchetype c. Thus, we are include the passage in full,
collating all our available editions for it as far as they survive.]
[403.7] [184r.6] [326r.3] [207v.1] [372r.1]

305

306 307

[404] 308

309
310311

312
313

314

[184v] 315316

317

318319
302

(but

gives

) |

[ (/)] (gi gu not obviously present but other vowels on this line are unclear) | 304

| 305

| 306

| 307 :

303 :

| 308 :

| 309

| 310 : ; | 311 : | 312 : | 313 insert , restoring

the meter | 314 : | 315 : ; | 316 : | 317 : | 318

(Here,

's reading seems corrupt, while give the expected

reading, corresponding to the deities listed in Chapter 7, and the order of their mantras given in Chapter 9.) | 319 :

(restoring the meter);

143

Chapter 10

320

321
322
323

324 325

326

[326v]
327
328329

330

331

332333

334 335
336337

339

320

340

338

: |

321 :

322

: |

323 : ;

( all give , but we would expect here. Only

gives the expected

and it seems most likely that was an error in the archetype, corrected in . All versions give an appropriate mantra for rdo

rje dbyangs len ma in Chapter 9, and her positioning would seem most logical here, fitting with standard sets of these goddesses.
In this case, we cannot be entirely confident of the archetypal reading, since we lack much of the desired evidence: Hemis has
some folios missing at this point, the Tenjur has its lengthy lacuna, and as described above, the descendants of hypearchetype c
[the Dunhuang, Tshal pa Kanjur and Bhutanese NGB witnesses] have suffered a scribal mishap, losing the passage. However,
assuming that dbyings ma was in the archetype, although it is most likely mistaken, it is nonetheless worth noting that there are
also further slight variants below on some names as spelt in Chapter 7 of TZComm, even though these other variants do carry the
same meaning as the standard names. 's reading here possibly reflects a strand within the Thabs zhags textual
tradition. In Chapter 7 of TZComm, as represented by the Dunhuang manuscript (which is unfortunately our only witness for this
section of TZComm), her name is at one point ( 26v.1) given as rdo rje dbyings len ma, but it is possible that deletion of the gi gu
is intended. The name, rdo rje dbyings, is also given in the Dunhuang manuscript's marginal annotations of Chapter 5, 14r.3.)
324

: |

325

: |

329 :

| 334 :

332 :
339

340

326 :

330

; |

; ;

| 335

| 336 :

331

327

: |

| 333 :

328 :

unclear;

(restoring the meter);

| 337 :

338 :

(but there is possibly a tiny correction of an inserted nga, too small to make out in the reproduction)

144

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

341342

343

344 345

346

347

348

349350

351

352

353 354

355356357

358

359

[ 108b.4] [ 265.5] [ 190.5] iv

360 361 362 [372v]

363 364 365

366

[405] 367
368

iv here rejoin the text, following their long omission of text starting in the middle of Chapter 6.
341

: ; |

342 :

(restoring the meter); |

343 :

| 344 : | 345 : | 346 : | 347 : | 348 : | 349 :

350 :

351 :

(restoring the meter) |

354 : ; | 355 :
359 :

| 363

360 :

361

: |

353 :

| 356 : | 357 : | 358 :

| 364

(note that this reading is also given in a parallel line above);

367 :

352

| 368

| 366

; |
| 365

362

145

Chapter 10

369

370371372

373 [208r] [ 266]

374 375

376377 [185r]

378 [327r]

379

380381
382 383

384 385

386 387

388 389

390

v This is the end of the passage omitted in the descendants of hypearchetype c (as given in our stemma codicum, p.43 above).
369 :

370 :

371 :

372 :

; |

373

(The intended gender here is uncertain. We would expect to be continuing with

the female members of Vairocana's circle, but the names from here correspond to his male circle, omitted earlier in the text. It is
possible that it is the females which are intended, but simply labelled by their male consorts' names, rather as in the old-fashioned
English tradition of referring to a married woman by her husband's full name, preceded by the title, Mrs. Note that

accept feminine gender in the following instances, but give a masculine particle to the previous goddess listed

above [rdo rje 'bebs ma], so it would seem that are more likely simply to be in error rather than making a

considered judgement or preserving an earlier reading here.) | 374 : | 375 : | 376 :

377 : omits | 378 :

382 :

(restoring the meter) |

379

( omit the next lines, see below.) | 380 :

| 383 :

intended | 385

| 384 inserts

| 381 :

, but it is partially rubbed out and deletion appears to be

omit (eyeskip from lag gnyis sor mo, bearing in mind also that this verse largely repeats the phrasing of the

previous verse) | 386 : ; ; | 387 :

; ; | 388 : | 389 :

; | 390 :

146

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

[ 37v.3] [298r.6] [303a.8] [289r.6] [ 108b.7] [ 266.2] [ 190.7] [56r.4]

[51r.1] [405.3] [185r.2] [327r.2] [208r.2] [372v.4]

391 392393 394

395396397 398

391

omit

393 :

392

| 396 :

| 397 insert | 398

| 394 : omit | 395 :

Chapter 11
[ 37v.4] [298r.6] [303b.1] [289r.6] [ 266.2] [56r.5] [51r.2] [405.4]

[327r.2] [208r.3] [ omits] [372v.5]

12 3

4 5 6

89

10

1213

14

15

11

161718

1920

[ 38r] 21 22

23

[289v] 2425

26 27 [56v]

28 29 [298v] 30 31

32 33

34

35

36 37 38

1 : omits | 2 insert

: omit; has a

vertically ordered double tsheg, presumably indicating a shad, as elsewhere in this manuscript | 5 : | 6 :

: omit |

10

: ;

15 :

: ;

12

| 16 :

agrees with

: ; |

omit | 22

| 26

13 :

, although not its positioning;

: ;

| 37

| 27 :

11

is most likely a

: ; |

(sic) | 34

| 20 :

| 28

9 :

| 23 : ;

:
| 38

14 :

30 : | 31 : | 32 : | 33 :
36

's inappropriate reading of

| 17 : | 18 :

corruption of , which is given by ) | 19

| 25 :

21

| 24 :
29 :

| 35

; ;

148

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

39 40

41 42 [406]

43 [373r] 4445 46

47 48 [ 267] 49

50 [327v]
51
52 53 54

55

565758 [ 38v] 59

616263

6465

66

67 68

60

69 70
71

7273

[51v] 74 [208v] 757677 787980

81

82 83 84

85
39

40

| 43

46

49 :

omits | 56

| 52

: ;

| 50 :

41 :

: omit |
| 53

42 :

: ; |

: ;

| 44

| 51

57 :

47

: ; |

45 :

48 :

: [but ];

| 54

: ;

omit | 55 :

; (It is hard to

say whether the archetypal reading was or here; supports , slightly increasing the chance of 's reading

being correct) | 58 : omit | 59 : | 60 : ;

| 61

omit | 65 inserts

: omit | 62 : ; | 63 : | 64

| 66

: omits | 67

: omit | 68 :

(but ) | 69 : omit | 70
72

: ;

; unclear

: omit | 71 : omit

| 73 : omit | 74 : | 75

| 76 :

| 77 :

| 78

: ;
:

79 : ; ; | 80 : omit; gives a double tsheg | 81 inserts | 82 :

; | 83 : ;

| 84 : | 85

could have been generated as a kind of dittography, but it can be meaningful)

insert (this

149

Chapter 11

86

87

88 8990

91

92

93

94

9596

9798 99

100

101 102 103104105106 107 108 [304a]

109

[ 39r] 110 111112

113 [57r]

118

114 115 116

119

120121

122 123 124

[ 268] 125 126

86

: ;

; |
:

(though a very faint shad seems to be inserted as a correction in ) | 88

89

omit | 94

117

87

: omit

| 90 : omit | 91 : omit; has double tsheg | 92 :

| 97

; omits | 95
:

omit

; |

93

; |

; | 98 : ; | 99

96 :

omit ( do not omit; the omission is probably an eyeskip error made independently by and . It is also possible that

in creating its version of the root text from the commentary, assumed these words to be a commentarial filler, and omitted

them deliberately.) | 100 : omits | 101 insert | 102 : ;

; ; |

omits;

103

106 :

indistinct) | 107

110 : ;
omit; has double tsheg

114

113

| 115

:
:

; ; ; |

also omit) | 122

; | 125
;

120

: |

; inserts

118

: omits |

| 123

| 111 :

121 omits

| 112 :

117 :

119 :

(eyeskip;

: ;

105 :

109 :

| 116 : ;

; ; [];

| 108

omit;

; ; |

omit | 104 : omits;

| 124 :

| 126 :

150

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

127128129 130 [328r] 131 [373v] 132 133

[290r] 134 [407] 135136

137

138
139

140

141142 143 144 145 [299r] 146 147148

149 150 151152

153 154

[ 39v] 155156 157

158159
162

163

: omit |

128 inserts

; |

132

129 inserts

151

130 insert

| 142 :

| 145

omit

deleted);

| 143

150

omits;

| 158 :

| 147

133 : omits | 134

152

139

| 155

unclear,

141

: omits;

| 148 : omits

| 144

: ;

: omit

(here, the letter pa also has a na ro which has been

|

160 :

; |

; ;

| 159 :

| 157 :

: ;

: ; ;

| 146

131 :

: ;

: ; ;

| 153 : ; omits | 154

omit | 162

:
|

| 138

156 :

insert

149

perhaps:

| 135 : omits | 136 : ; | 137

; ;

140

160 161

127

163

161 :

:
;

151

Chapter 11

164 [52r]

165 166 167

168 169

170 171

172

173174
175

176

177

178 179180 181 182

183

184 185

186 187

188 [57v] 189 190

191192

193194 195 [209r]

196

[ 40r] 197 198 199 [ 269] 200 201

202

164 insert

insert

(this is most probably an intrusion of

commentary into the root text in both the Bhutanese and Tshal pa Kanjur editions; these words are not in any way marked off as
commentary in TZComm, so confusion would have been understandable. None of this passage is highlighted in the Dunhuang
manuscript, so it is not possible to judge the highlighter's assessment.) | 165 : ;

; ; |

; | 169

| 172

; |

166

: omits |

: ; omits | 170

; |

: omit | 171 :

174 : omit | 175 :

| 179

] | 184 :

189 :

| 190

198

201

: ;

199

| 191

| 197

186

| 187 : omits | 188 :

( unclear);

; ; ;

| 185 : ; unclear;

192 : omit | 193 : omit | 194 :

196

| 176 : | 177

: omit | 180 : | 181

omit | 182 : omits | 183 insert, perhaps by eyeskip on

168 :

| 173 : ; ; ;

( initial as correction); ;
| 178 :

167 :

| 195 inserts

: omit
200

; ;

| 202

152

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

203204205 206 207 [304b]

[328v] 208

209 210

211212 213 214 215216

217

218219

220 [374r]

221

222

223 [408] 224 225

226
228

227

229

230
231

232

[290v] 233 234

235 236237
238 239

240 [ 40v]

203

: omit |

204 :

inserts | 206 : | 207 : omit | 208 insert | 209

;
omit; inserts

222

omits ( also omit) | 213 insert

| 215

: omits | 216 :

: ; omits | 223

(dittography at new line);

:

236

| 228

| 231 :

234

| 229

| 232

| 218

211 :

| 220

| 224 : | 225

; ;

| 237 :

(dittography at new line); | 230 :

: ;

214

; ;

240 : ;

| 221

| 226

: ; ;

| 217 :

| 219 : ;

205

: omits | 210 :

; ; ; ; |

omit | 212

227

| 238 :

235

| 233

239

153

Chapter 11

241 242243

244 245

246247 248

249 250 [52v]

252 [299v]

251

253 254

255 256 257 258

259 260

261

262

263 264

265266

267

268269270
271272

273 274
241
242

give );

| 243 :

; ;

251

| 253
link

| 245

| 244 :

| 246 :

| 249 :

247 :

| 250

: ;

omits | 255

248 :

; omit (

agree on placing it in the new verse line) | 256 : ;

; ; ;

| 262

(note that links the

; ;

; ;

| 260

| 261

; ;

| 267

from rdo rje; also, but only after reversing rdo rje and 'khor lo.)
| 270 :

| 263

; ;

| 268

| 271 :

; ;

269

: ;

; appear to have had an eyeskip here

; ;

; ; ; ; | 273 : | 274

; ;

; ( agrees with

; ; ; ; | 265

omit | 266 :

to the next tshig rkang) | 259

's decision to

| 257 :

, who omit it altogether,

; ; ; ;
| 264

| 252 : [(/)];

to the previous tshig rkang would seem more logical, but the other versions, apart from
| 258

; ;

| 254 :

; ( both

272 :

: ; ;

154

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

275

276 [58r] 277278

279 280

281282283 284 285 [ 270]

286 287

288289 290

291

292

[ 41r] 293294

295

296 [329r] 297 298

275

; ;

(note the order reversal here and below, in which

contrast with all the other versions; it is not entirely

certain which is more "correct", since lists of these female deities can vary. However, all versions of a discussion in TZComm [see
p.163-164 below] agree with 's order here, and the order has some significance since the four goddesses

[Sihamukh, Vyghrmukh, glamukh, vnamukh] are associated with each of the cardinal directions in turn. This section
of text is witnessed not only in TZComm editions, but also in the Tshal pa Kanjur TZ texts, and the names [although not the full
text] are included in the Bhutanese NGB texts, again, in 's order. Moreover, 's version here is consistent

with some rNying ma traditions. For instance, Chapter 16 of the rGyud gSang ba'i snying po (mTshams brag NGB edition,
Volume Wa: 207) supports 's ordering in its mantra list for these four goddesses.) | 276 :

[ ?] ; ( agrees); ;

; ; ( share the

omission of sri ga; and together with , they also have the order reversal noted in the note above, and omit shwa na.)
277 :

278 :

; ; |

; |
;

281

: omits |

(Here,

279

: omits |

282 :

280 :

283 :

's reading seems clearer; an archaic meaning of ltod pa

according to bTsan lha ngag dbang tshul khrims (262): rtsod pa'i ming ste.) | 284 : omits | 285 :

; ( agrees) |
;

| 287

286

: ; ;

[..] ( 's yud tsam here may simply represent a commentarial addition) | 288 :

| 289 :

| 290

; ; ; ;
;

292

| 293 :

; ;

; ; |

295

: |

296 :

| 291 :

; |

| 297 : | 298 : ;

294 :

155

Chapter 11

299 300301

302 303

304305
306307

308309
310

311 312
313

[209v]

314

315 [305a] 316 [374v] 317

318 319

320321

322323 324

325 [409] 326

327328

329

330 331

332
334

333

335

299 : ; | 300

304

:
; | 301 : | 302 :

: |

305 :

| 308 :

306 :

| 309 :

| 303 :

| 310 :

307 :

(rtswa, 'straw', would seem more appropriate and was probably

intended, but perhaps rtsa'i or rtsa ba'i was in the archetype) | 311 : | 312 : ;

; |

313 : ; ;

314

[?] [it

appears that was possibly corrected to ]; ; (only have the two

final syllables, but it seems the was uncertain which to choose, while the other versions make different decisions) | 315 :

316 :

| 318 :

| 321 :

| 322

| 333

| 325

| 319

317

| 320

: ;

| 326 : | 327 :

; ; |

; |

| 323 : | 324

| 328 :

| 331

329 :

330

| 332 : ; [(/)?];

| 334 : | 335 :

156

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

336

337

[ 41v] 338339

340

341 342

343 344345

346

347 [291r] 348

[53r] 349

350351
352 353

354

355

[ 271]

[58v] 356357358

359

360361

362

363364

365366 367

336

368

: ;

; | 337 :

| 339 :

| 340

| 338

(either of these

readings could work; khrag might seem a little more appropriate, since we have already just above had the same line with sha, but
the repetition could well be correct. Either reading might represent a scribal error of repetition of the word given just above.
Here, has a reading shared with , but it could have come to this reading by amending khrag. Alternatively, both

the Tenjur and the ancestor(s) of the Tshal pa Kanjur and Bhutanese texts may have introduced khrag separately.) | 341 :

; |

342 inserts

(this addition, which is shared by

, would seem appropriate, and it is even just possible

that it might represent an earlier verse lost in the other editions, although it is not found in any other text, including . The

text remains coherent without it, although the two items to be consumed already mentioned, along with the two additional items
here, might possibly constitute four of the five articles mentioned just above [rnam lnga'i chas].) | 343 : | 344 :

345 :

| 348 :

| 353

; ; |
| 349

| 350

346 :

347

| 351 :

| 354 :

355 :

; ; ; | 356 : | 357 : | 358 : | 359

omits | 360 :

| 363

| 364 :

; | 367 : ; | 368

| 361 :

| 365

| 352 :

| 362 :

: ; |

: ; unclear

366 :

157

Chapter 11

[300r] 369

370

371 372

[329v] 373374 375

376

377 [ 42r] 378

_______________________________________________________________

, suggesting a break in the text; this may simply indicate the

[ (but not , or ) insert
commencement of the next section of mantras. From here up to 45r.3, the Tshal pa Kanjur versions
include the entire commentarial passage, and not merely the mantras. It seems that the editors were unable to
distinguish the root text from the commentary in this section. Perhaps this was all the more difficult since the
first part of Chapter 11's root text gives extensive detailed description. There might not appear to be any
obvious distinction between that material and the commentarial descriptions of the ritual actions and the
deities whose mantras are given next in the root text. The Dunhuang manuscript generally highlights the root
text, and here highlights only the mantras, suggesting that the scribe understood the boundaries of the root
text, but the Dunhuang manuscript also failed to highlight the extensive opening to this chapter, which all the
root text editions accept as root text. As in the other cases where the Tshal pa Kanjur or Bhutanese editions
have added commentarial passages, we are presenting the Tshal pa Kanjur additions (in an indented
paragraph format), not because they might be part of the archetypal root text but simply since their inclusion
has become part of the transmitted text in this branch of the tradition.]
[ 42r.1] [300r.1] [305a.5] [291r.3] [ 271.3]

379380 381 382

383

369 :

384

385 386

; |

370 : | 371 : | 372 : ; ;

373 dittographically inserts

387

374 :

375 :

(the archetypal reading is uncertain, but it is quite likely that it was thib, with the sense of the deities

gathering together within oneself.

may have introduced the more commonplace reading thim independently from

hypearchetype b [as given in our stemma codicum, p.43 above].) | 376 : ; | 377 :

; omits |

| 381

378 :

| 379 : | 380 : ; ;

(Note that here

| 382

| 383

: ;

give the standard order with the right hand mentioned before the

left, while the other versions give a non-standard order. It could be that the non-standard order is in error; or that the standard
order in was a hypercorrection, perhaps even a simple scribal error, since the standard order would be expected.
Although it is hard to judge, it is more likely that the non-standard order is correct: the same non-standard order is given in the
case of Be ta li below, and in that case, all versions agree.) | 384
385

| 386

| 387

omits;

: omits

158

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

388389 390
391
_______________________________________________________________
[ 42r.2] [300r.2] [305a.5] [291r.4] [ 271.4] [58v.3] [53r.3] [409.5]

[329v.1] [209v.5] [374v.6]

392
_______________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 42r.2] [300r.2] [305a.5] [291r.4] [ 271.4]

393394 395

397 398 399

396

400 401 402

_______________________________________________________________
[Here, the main root text resumes:]

403 404
_______________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 42r.4] [300r.3] [305a.6] [291r.5] [ 271.5]
405

406 407

408

409410 411 412413

388

| 390

(see note 383 above) |

:

| 391

389 :

; ;

omit | 392

; ; ;

[ ] ; ; unclear; ; (Variants are to

be expected in rNying ma pa literature. This first of the eight mamos is generally equated with Gaur, although Klong chen pa
[bdud 'joms bka' ma Volume La: 539] notes that most of the [rNying ma] root texts give the name as, ko'u r. See our TZComm
edition, Chapter 11, notes v and vi.) | 393 insert | 394 : ; | 395
insert | 396 : | 397 : omits | 398 :
400

: |

omit | 404

401 inserts

inserts

399 : omits

; |

403

: (it appears that ru written in error has been deleted in part,

it has been amended into the following lu);

| 405 insert

| 408 : ; | 409

insert | 412

402

| 406 insert

: omit | 413 : omits

| 407

| 410 : ; unclear | 411

159

Chapter 11

_______________________________________________________________
[Here, the main root text resumes:]

414
_______________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 42r.5] [300r.3] [305a.7] [291r.5] [ 271.6]

415416 417 418 419

421

422

420

423 [ 42v] [ 272]

_______________________________________________________________
[Here, the main root text resumes (though omits):]

424
_______________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 42v.1] [300r.4] [305a.8] [291r.6] [ 272.1]

4 2 5 4 2 6 4 2 7 4 2 8 4 2 9
430

431 432 433434 [305b] 435

436 437

414

:
;

insert

423

| 418 :

: ; ;

; ;

421

(The second of the eight mamos is equated

with Caur; see our TZComm edition, Chapter 11, note vii.) | 415 :
| 417 insert

| 419 inserts

omits

| 424

422

give

| 420

here) | 416

: ;

; ; omits (According to rNying

ma lists, such as that here and in the rGyud gSang ba'i snying po, the third of the eight mamos is Pramoh; see our TZComm
edition, Chapter 11, note viii.) | 425 : | 426 insert | 427 : | 428 :

| 432 :

| 429

insert | 430 :

omit | 437

| 433 insert

| 431 : ;

| 434 :

| 435

| 436 :

160

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

_______________________________________________________________
[Here, the main root text resumes (though omits):]

438
_______________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 42v.3] [300r.5] [305b.1] [291r.7] [ 272.2]

439 440

[291v]

441 442 443 444

445 446447 448 449450 451452

_______________________________________________________________
[Here, the main root text resumes:]

453
_______________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 42v.5] [300r.7] [305b.2] [291v.1] [ 272.4]

454 455 456

459 460 [ 43r]

465 466

438

;
;

457

458

461 462 463

464

| 446

450 :

omits | 453

| 442 insert

| 447 :

| 443 :

;
| 451

| 444 :

| 448

(note that here, gives

(indistinct,

| 462

| 465 : ; ;

: ; ;

; omits | 463
| 466

may be intended) | 452 :

(The fifth of the eight mamos is Pukkas; see

but agrees with ) | 457 insert

omits | 459 insert | 460

| 445 :

: omits | 449 insert

our TZComm edition, Chapter 11, note x.) | 454 : omit | 455 insert

: ; ; ;

omits (The fourth of the eight mamos is Vetl; see our TZComm edition,

Chapter 11, note ix.) | 439 inserts , and the following 7 syllables are written small | 440
441 :

456 :

| 458

| 461 : ;

464

161

Chapter 11

_______________________________________________________________
[Here, the main root text resumes:]

467
_______________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 43r.2] [300r.8] [305b.3] [291v.2] [ 272.5]

468

469

470 471 472473 [300v]

475

474

_______________________________________________________________
[Here, the main root text resumes:]

476
_______________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 43r.3] [300v.1] [305b.5] [291v.3] [ 272.6]

477 478

467

(The sixth of the eight mamos is Ghasmar; see our TZComm edition, Chapter 11, note xi.) | 468 insert

470 :

mantra variants also below) | 472 :

474 inserts

;
[?]

| 475 :

471

; ; ;

| 476

469 :

: ; (see

473 :

: ; ;

and agree);

(The expected seventh of the eight mamos is man; see our TZComm edition,

Chapter 11, note xii. Note that the reading, karma, appears to be an error shared by the Dunhuang manuscript, the Tshal pa Kanjur
and Bhutanese NGB, namely the descendants of hypearchetype c [as given in our stemma codicum, p.43 above]. 's reading,
ba sha ni, may have been in the archetype; perhaps 's and 's ancestors corrected it. 's ba

appears to be a correction from a still visible tsan, which has been rubbed out. Presumably, the scribe began writing the next name
on the list, and then noticed the error {in fact, Cal is given as seventh in Karma Gling pa's Bar do thos grol list [Dorje 2005:
369, 398], and in fifth position in the rGyud gSang ba'i snying po's Chapter 16}.) | 477 insert | 478 's
omission begins here (omission shared by and ); this seems to be a case of eyeskip to the next occurrence of

162

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

479 480 481

482483 484

486

485

_______________________________________________________________
[Here, the main root text resumes (though omits):]

487

_______________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage; while the Bhutanese version includes the list of
eight animal and bird headed deities in this passage, but not the other text (beginning further below, from
58v.4 and 53r.4):]
[ 43r.5] [300v.2] [305b.6] [291v.4] [ omits the first line, beginning
272.6]

488

[ 273] 490 491 492

489

493 494 495 [ 43v] 496

497 498

499 500501 502 [306a] 503

504

506 507 508 509

505

510

511

512 513 514

479 insert
483 :

480

| 484

481 : omit | 482 insert

| 485

: ; ;

| 487

omits;

| 486

; ;

unclear;

(The final eighth mamo is Cal; see

our TZComm edition, Chapter 11, note xiii.) | 488 insert | 489 rejoins here (along with and ),
after omitting

492 insert | 493 : | 494

from here to the end of the line);

omit | 497

500

498

| 510 :

514 inserts

| 509

| 508 :

| 495 :

| 501 insert | 502

inserts

| 491 :

| 505 inserts

(leaving one fifth of the line blank,

(dittography at the turn of the folio) |

499

496 :

: omit

| 503 inserts | 504

| 506 :

| 507

: [ (/ )] ( probably intended); ;

| 511 :

( agrees)

| 490 inserts

512 :

513

163

Chapter 11

515 516 517

521

518 519 520

522 523 524 525 526 527 528

529

530

531

532533534 535 536 537

539

538

543544

540 541

545

542

546 547 [292r] [ 44r] 548

549 550

551

552 553 554 555 556

557

558559 560 561 562 [ 274] 563

564 565

515 :

516

517 insert | 518 :

| 521 : | 522 : | 523 :

; ;

| 525

519 insert

520

| 524 : ;

(a small line above the

na apparently deletes it, but it is not entirely certain); ; | 526 insert

527 : | 528
532

529 : omits | 530 : ;

| 533 : | 534 : ;

omits | 537 : omits | 538

531 :

| 535 : | 536

539

insert | 540
: | 541 : | 542 : ; | 543 : ;

544 :

| 545 :

| 546 :

| 547 : ;

; ; (58v.4 and 53r.4. As noted above, include the animal

and bird headed deities given in this discussion, which are not given in the South Central/Bathang/Tawang texts [the Hemis folios
are missing]. The first of these eight phra men ma (Skt. pc) is Sihamukh; see our TZComm edition, Chapter 11, note xiv.)
548 inserts

(zhing kha na thogs pa is not entirely clear; is probably making a recensional intervention here, which

clarifies the sense). | 549 : | 550 : ; ; | 551 :

552

: ; ;

; (The second of these eight

phra men ma is Vyghrmukh. This is the second item on the list included by 58v.4 and 53r.5.) | 553 : ;

554 inserts

omit | 559 :

555

: ngan (perhaps intended) |

| 560

556 : omits | 557 :

: ; ;

558

; ;

(The third of these eight phra men ma is glamukh. This is the third item on the list included by 58v.4 and

164

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

566

567

568 569570 571572

573574 575 576577 578

579 580 581 582583

584

585 586

587 588 589

[ 44v] [301r] 590

592

591

593594

595 596

597 598 599 600

601 602 603

53r.5.) | 561

605

; ;
| 567

604

562 inserts

inserts

563 :

(but in both cases, may have been intended) |

: ; ;

565

| 570 : omits | 571

| 573 : | 574 :

564 :
566

; ; (The fourth of these eight phra

men ma is vnamukh. This is the fourth item on the list included by 58v.4 and 53r.5.)

569 :

568

| 575 : ; ;

omits

572 :

(The fifth of these eight phra men ma is Gdhramukh. This is the fifth item on the list included by 58v.4 and

53r.5.) | 576 insert | 577 :

| 580

| 582 insert

| 578 : | 579 :

; ; ( may be intended);

| 583 : omit | 584 :

| 581 inserts

| 585

: ;

; (The sixth of these eight phra men ma is Kakamukh. This is the sixth item on

the list included by 58v.4 and 53r.5.) | 586 : omits | 587 : ; | 588

: ;
omits | 592 :

595 :
599

| 589 :

; ; indistinct |
| 596

: ;

593

| 590 :
: |

594 :

| 597 insert

| 591 :

| 598 insert

(The seventh of these eight phra men ma is

Kkamukh. This is the seventh item on the list included by 58v.4 and 53r.5.) | 600 insert

(this appears to be misplaced see below and the placement is not shared by , although omitted altogether in )
601

| 604

602

: omits | 605

; ;

603

: ;

omit (We would expect this line here, rather

than above, where the Tshal pa Kanjur texts give it; most probably, is correct here.)

165

Chapter 11

606 607

608 609 610

[306b]

611 612 613 614 615

616

617

618619 620 621

622623624625626 627

629630

631

628

632 633 634

635 636 [ 275]

_______________________________________________________________
[ 44v.6] [301r.4] [306b.3] [292r.6] [ 275.1] [58v.5] [53r.5] [409.6]

[329v.2] [209v.6] [374v.7]

637

_______________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 45r.1] [301r.4] [306b.3] [292r.6] [ 275.1]

638

639 640

641642643
644645
;

606 inserts

inserts

607

: ; ;

(The final and eighth phra men ma is Ulkamukh. This is the eighth and last item on the list included by 58v.4 and 53r.5.)
608 inserts

618 :

| 614

623 :

609

: ;

| 619 :

610 :

| 615

| 620 :

: ;
| 621

611 :

| 616 :

: ;

| 624 : omits | 625 :

: ;

636 insert

| 641

inserts

| 633 :

| 637

('s second not clear);

| 642 inserts

| 643 :

| 644 :

| 622 :

| 626 :

| 631

;
| 638 inserts

omits | 613

| 617 inserts

: ;

| 634

omits | 627 : | 628 inserts | 629 : | 630 :

632

612

inserts

| 635 :

| 639 :

| 645 inserts

| 640 :

166

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

646

647

648

649 650 651 652

653

_______________________________________________________________
[From here, the commentary runs parallel to the root text as witnessed in the South Central, Bathang, and
Tawang Orgyan Ling editions. However, the Bhutanese versions omit much of the text, which relates to the
tshogs ritual, apart from the opening section, the mantras and the closing section of the chapter (see below).
Their mistaken judgement was understandable, given that the root text's mantra list above does not include
the elaborations of the retinue and rituals found in the commentary.]
[ 45r.3] [301r.5] [306b.4] [292r.7] [ 275.2] [58v.5] [53r.5] [409.6]

[329v.2] [209v.6] [374v.7]

654 [292v]
655

656657

[39r] i

658 659

660 661662 [375r]

663 664

665 666667

668 669 670 671672 [410] 673

i Here, the Hemis Kanjur picks up the text with its folio 39r. Folios 33-38 are missing, and its folio 32 ended in the first section of

Chapter 7.
646

650 :

661

647

| 655

: |

663 :

; omits | 651 :

omits | 658

unclear

659

; ;

| 656

648

| 652 insert

| 653 :

: ;

[-] (unclear) |

662

; ;

660 :

| 654 :

: ; ; |

| 664 :

649

| 657 :

unclear

; ;

( share blag/blags [ 's reading is suggestive of blags also]; this is a different, and not quite so appropriate verb

here, but since the sound is the same, it may have been a non-standard spelling, possibly going back to the archetype, though not
witnessed in .) | 665
tshig rkang from here to

: ; |

666

: ; |

667 58v.6 and 53r.6 omit 9

below, and after that mantra, omit a further 5 tshig rkang. This is most probably a

deliberate decision on the root text boundaries, rather than a scribal omission. | 668 : ; ;

669

; ;

; ; |
| 671

670

: ;

; omits | 672 inserts

| 673 :

167

Chapter 11

674675 676677

678679

[ 45v] 680

682 683 684 [210r] 685686

681

687 688

689690

691

692 693694

695696697 698 699

700701
702

703

704 705

[301v] [330r] 706 [307a] [ 276]

708709

674

: omit |

676 :

675 :

707

; ( and are divided over gzi byin/gzi brjid,

and so too are the descendants of branch b [as given in our stemma codicum, p.43 above], that is, the Tenjur and the descendants
of sub-branch c. Either reading can fit appropriately here.) | 677 : ; ; | 678
omit | 679 : | 680

omit | 682

omit

688

689

: ;

| 694 :

| 693

690

| 692

| 695

| 684

| 686 :

: ; [ ?] ;

[(/)];

: ;

; omit | 685 :

691

| 683

| 681

| 687

; ;

; ;

; |

696 :

; (here, it is hard to be certain which reading was in the archetype; both

and give mchong, but and clearly read mchod. Given the general accuracy of and ,
mchong is most likely. Either meaning could fit.) | 697 : ; | 698 : (here, the

archetype seems more likely to be byung, although phyung would seem rather preferable, but for this reason, it could have been
introduced unreflectively.) | 699 : omit | 700 : ; | 701 :

706

;
shads

702

[...] (unclear);

703

marks a break in the text, with two double

and a wide space ( now omit a further 5 tshig rkang.) | 704 :

705 : omit
707

| 708 : ; ; ; | 709 : omit

168

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

710711712

713 714715716

717

718

719 720

721

722

723 [ 46r] 724 725

727 728 729 730

726

731732733
734735

736737

738
[53v]

710

711 :

(the additional expression of exultation found in

| 713 :

(but unclear);
;

| 712 :

; |

| 714 :

| 715 : | 716 : | 717 : omit;

( and

agree);

was almost certainly present in the original archetype [as well as in the Guhyasamja parallel, see below note 725], but

perhaps edited out of hypearchetype c, the common ancestor of the Dunhuang, Tshal pa Kanjur and Bhutanese NGB, as not
adding anything of significance and likely to be in error. Perhaps this followed an initial shortening to the reading given [probably
coincidentally] in both the Tenjur and Hemis Kanjur, at which stage the addition had lost the meter and appeared dittographical.
rejoin after this line.) | 718 :

; ; |

| 724

; ; |

721 :

; |

722

719

: omit |

720 :

; |

723

| 725 A version of this song is found also within verses of homage in the

Guhyasamja's Chapter 5. See for example IOL Tib J 438, folio 14a (page 21), line 1-2, where we read:

726 A further omission by begins here, which again is most

probably a deliberate decision about the extent of the root text. | 727 : ; ;
insert

728

: |

729

730 : omit | 731

: omit |

; ; ;

| 732 : ( gives ) | 733

| 735 : omit | 736

| 737 : omit | 738

734

(a la la ho is the usual exclamation associated with enjoyment of the tshogs feast, so it would be easy for a triple la to

be amended unreflectively simply to a double la. See also note 747 below.)

169

Chapter 11

739

740 [59r]

741 742

743 744

745 746 [39v]

747

748 749

750

752

751

753

754755756757
758

759

760

761

[ 46r.5] [301v.3] [307a.3] [292v.6] [ 276.4] [59r.2] [53v.1] [410.7]

[330r.4] [210r.5] [39v.2] [375r.8]

[375v] 762763 764

739 :

744

748 :

; unclear;

740

: ;

;
750

| 745

747

; ; |

| 742 : omit | 743

| 746

;
;

741

| 749 : omit

| 751 Once again, omit these next non-mantra words, again most probably reflecting a deliberate

decision about the extent of the root text. | 752 : ; ; ; | 753

inserts double tsheg, indicating a break | 754

: omit |

: omits |

755 :

757 : omit | 758

; ;

omit;

;
;

| 760

[ ]

: ;

| 761 :

; omit; | 762

756

: ;

| 759

; ; |

; ;

; ; omits;

: ; ;

; ; ; | 764 : omit

| 763 :

170

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

765766 767 768 769

765 : | 766 : | 767 : ;

| 768

| 769 : omits

Chapter 12
[ 46r.6] [301v.4] [307a.4] [292v.6] [ 276.5] [59r.2] [53v.2] [410.7]
[330r.5] [210r.5] [39v.3] [375v.1]

1 2
[ 46v.1] [ 276.5]

3 45 6

[411] 78 9

10

11 12 1314

[ 47r.5] [ 277.6]

15 16

171819

20 21 [293r] 22
[ 278]

23 [ 47v]

24

[ 49v.4] [ 280.6]

25

26

27 [330v] 2829

: with prefixed da and suffixed a inserted as small letters; |

3 inserts | 4

30

12

: omit

11 :

| 5 : | 6 : ; ;

| 8 : |

: ; |

13

: omits |

10 :

| 7 :

14 : [(/)] (a correction has been

made; may be intended) | 15 : | 16 : | 17 : | 18 :

; |

19 :

; ;

20 : omits, but seems to attempt a correction by

(unclearly) inserting | 21 : ; ; ; ; | 22 :
illegible);
27

| 24 :

; | 23
| 25 :

| 28 : has this added as a correction below the line;

; | 30

| 26

(partially

| 29 :

172

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

[ 281] 3132
[ 53v.1]33 [ 284.5]

34

3536

37

3839

[ 53v.6] [301v.7] [307a.6] [293r.2] [ 285.3] [59r.6] [53v.5] [411.3]

[330v.1] [210r.8] [39v.5] [375v.4]

40 4142

4344 45 [ 54r] 46 47

31 :

; ( also gives ); |

seem a rather clearer reading);

32 :

('dul would

| 33 After a passage including two further root text verses displaced from Chapter 13 (the

passage is in Chapter 13 in the Tenjurs, while the two root verses are in Chapter 13 in all the TZ editions), the Dunhuang
manuscript here resumes running in parallel to the Tenjurs. | 34

| 35 :

| 36 : ; | 37

40

omit ( also omits);

: partially illegible | 38 : | 39 :

omit | 41

( also gives ); ; [ (/ )];

: ;

| 42 : ; ; | 43

: omit | 44 : | 45 :

(possibly, the introduction of thams cad here was a commentarial gloss,

which the Tshal pa Kanjur and Bhutanese editions did not remove, although if so, it seems the Dunhuang manuscript edited it out,
or omitted it through scribal error.) | 46 : ; | 47 :

Chapter 13
[ 54r.1] [301v.7] [307a.7] [293r.2] [ 285.3] [59r.6] [53v.6] [411.4]

[330v.2] [210r.8] [39v.7] [375v.5]

[210v] 1
2345

6 78

[ gives this text within Chapter 12, 52v.4] [ 287.3]

10

11 [59v] 12 13

14 15 16

17 18 1920

[ gives this text within Chapter 12, 53r.3] [ 287.6]

[302r] 21

22

23

[ 288] [54r] 24 [307b] [40r] 25

27
26

: ; ( also gives ) |

| 8

: illegible | 6

: omit | 9

|7

3 :

11

: |

| 25

19

20

13 :

| 18 :

| 22 :

: [-] (unclear);
| 27

| 10 :

; |

: ; ; ;

; ( also gives );

4 :

12 :

; ;

| 23 : ; | 24

: illegible | 21 :

: ; ;

; ; |

; ;

| 14 : | 15 : | 16 : | 17 :

28

| 26

| 28

174

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

[ 54v.6] [ 288.3]

29

3031 32

34

33

[ 55r] 35 36 37
[ 57r.6] [ 291.4]

38 39 4041 [376r]

42 43

44

[ 57v] 45

46

[ 57v.4] [ 292.1]

47 4849 50

51
52 53

54

_____________________________________________________________
Note that the Tenjur version of TZComm stops running parallel to the Dunhuang manuscript in Chapter 13 [at
118a.6, 292.2, 208.3, corresponding to 57v.6], clearly as a result of an accidental
omission (the chapter titles for Chapters 13, 14, 15 and 16 are all missing, and so too are all the root text
lemmata for these sections). The Tenjur texts re-join the Dunhuang manuscript near the end of Chapter 17,
corresponding to the Dunhuang manuscript 61v.1.
______________________________________________________________

29 :

; ;

31 :

| 32

; ; | 35
[

?] (unclear);

42 :

| 44

| 36 : | 37 :

| 40

| 53 : | 54 :

| 49 :

: illegible

34 :

unclear | 38

: [ ?] (unclear) | 41 :

; ;

30

| 33 :

; ; ;

; |

| 43 :

: | 45 : ; | 46 :

52

| 39

| 47 :

| 50 :

48 :

| 51 : omits

175

Chapter 13

[ 57v.6] [302r.3] [307b.2] [293r.6] [ omits] [59v.4] [54r.3] [411.7]

[330v.6] [210v.4] [40r.4] [376r.2]

55 [412] 56 [ 58r]

58 59

61

[331r] 60

62

57

636465

[ 58r.5] [302r.4] [307b.3] [293r.6] [ omits] [59v.5] [54r.4] [412.1]

[331r.1] [210v.4] [40r.5] [376r.3]

66 67 6869

70 71 7273 74

55

| 57 :

omits | 60 :

56

| 58 :

; ;

; ; ; |

59 :

| 61 : ; | 62 : ;

; ; |

63 :

64 :

; |

65

: ; ; (118r.5 shares the reading, mgo mags, which is surely a

corruption of 's mgo 'gas; this occurs in TZComm's explanation of the verse, and would seem to make better sense than the

other readings.) | 66 : omits | 67 : | 68 : ; (sic); ;

72

69 :

; ; |

| 73 : ; | 74

70

: omit |

71 inserts

Chapter 14
[ 58r.6] [302r.4] [307b.4] [293r.6] [ omits] [59v.6] [54r.5] [412.2]
[331r.2] [210v.5] [40r.5] [376r.4]

[ 58v]

[ 58v.1]

[ 58v.2]

|2

| 4

| 3

; ; ;
;

; ;

; ;

| 6

; [ ]; ; ;

; ; (The principal wrathful male deity is r heruka [TZComm Chapter

12, Dunhuang ms.49v.5], whose Tibetan equivalent is given as, khrag 'thung chen po [TZ Chapter 12 and Chapter 16.)
7

: ; ;

(The

principal deity is surrounded by the well-known set of ten Wrathful deities [khro bo bcu], beginning with Hkara, given last
rather than first on the list in TZComm Chapter 12. He is referred to by his Sanskrit name in TZComm Chapter 12 and Chapter 13,
shortened to h in the TZ mudr list in Chapter 16, where he is placed as usual at the top of the list.) | 9

; ;

(here, and in the following mantras, there appears to be inconsistency both within and between editions on

whether we have double, single or triple mantra syllables. Here, it may be that possibly the archetype had a double, or perhaps
even a single grdza/gardza, and in the following example, a double ha na, while the extra gardza/ha na might have been

177

Chapter 14

10

11 [293v] [60r]

12

13

15

14

16

17

18

introduced to create more consistency with the mantras below, most of which seem to be triple in and the other editions,
although not in .) | 10 : ; ; ;

(The mantra of Vijaya, whose

Tibetan name is given as rnam par rgyal ba in TZComm Chapter 12 and Chapter 13, and in the TZ mudr list in Chapter 16.)

11

the o of the next mantra);

(pha is written very small, squeezed in between h and

12

(The mantra of

Nladaa, whose Tibetan name is given as be con sngon po can in TZComm Chapter 12 and Chapter 13, and as dbyig sngon can
in the TZ mudr list in Chapter 16.) | 13

14

(The mantra of Yamntaka, whose transliterated Sanskrit name is used in TZComm Chapter 13. His

Tibetan name is given as gshin rje gshed po in TZComm Chapter 12 and as gshin rje 'joms pa in the TZ mudr list in Chapter 16,
and also in the annotations of the Dunhuang version of TZComm Chapter 14.) | 15

(pha is written very small, squeezed in between h and the o of the next mantra);

| 16

; ;
;

; ; (The mantra of rycala, whose Tibetan name is given as

mi g.yo mgon po in TZComm Chapters 12 and 13, and in the TZ mudr list in Chapter 16.) | 17

18

(in

an initial

, inserted dittographically, has been deleted);

; ; ;

(The mantra

of Hayagrva, who is simply called by his transliterated Sanskrit name in TZComm Chapters 12 and 13, shortened to 'gri ba in the
TZ mudr list in Chapter 16. The annotations of the Dunhuang version of TZComm Chapter 14 give his Tibetan name as dpal rta
mgrin.)

178

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

19

20

[54v]

21

[40v] 22

23

24

[ 59r]
26

25

27

_______________________________________________________________
Note that the Tshal pa Kanjur tradition omits the final mantras of the chapter, but this is clearly in error, since
the mantras (or at least the ritual action mantras) are indicated in the chapter title which all editions (apart
from the Tenjur, continuing its omission of several chapters) share.
_______________________________________________________________

19

| 20

; ;

; ; ; illegible; ;
(The mantra of Aparjita, whose Tibetan name is given as gZhan gyis mi thub pa in TZComm

Chapters 12 and 13, shortened to gZhan gyis mi thub in the TZ mudr list in Chapter 16.) | 21

22

:
;

(The mantra of Amtakualin,

whose Tibetan name is given as bdud rtsi 'khyil pa in TZComm Chapters 12 and 13, and in the TZ mudr list in Chapter 16.)
23

;
;

| 24

; ; (The mantra of Trailokyavijaya,

whose Tibetan name is given khams gsum rnam par rgyal ba in TZComm Chapters 12 and 13, and as khams gsum rgyal ba in the
TZ mudr list in Chapter 16.) | 25

| 26

illegible;

; (The mantra of Mahbala, whose Tibetan name is given stobs po che in TZComm Chapters 12 and 13, and

as stobs chen in the TZ mudr list in Chapter 16.) | 27

; illegible;

179

Chapter 14

[ 59r.2] [ omit] [ omits] [60r.3] [54v.1] [412.5] [331r.5] [210v.8]

[40v.1] [376r.8]

28
[ 59r.2]

30

[ 59r.3]

29

31

[ 59r.3] [302r.7] [307b.7] [293v.2] [ omits] [60r.3] [54v.2] [412.6]

[331r.6] [211r.1] [40v.2] [376r.8]

32 3334 35
3637

28

[ (/)]

's shan ta might be intended to read, shan ti);

omit;

( may intend the same readings here;

(TZComm explains that

this set of four mantras are for enjoining the four ritual actions: las bzhi bskul ba 'o, 59r.2. It is not altogether clear what all the

intended underlying words are, but they could be: 1. sumbha, perhaps for destructive action? 2. pui, for increasing; 3. nti,

for pacifying; 4. vaa or vaya, for captivating.) | 29

( adds that the name of the object should be inserted

into the mantra, although it highlights, thus recognising as root text, only the mantra syllables, reading in full,
omit;

| 30 :

);

( adds that the name of the object should be

inserted into the mantras, although again, it highlights [thus recognising as root text] only the mantra syllables, reading in full,

);

omit;

omit;

; ; ; ;

deleted; gives and gives ); ;

(The Dunhuang commentary,

| 31

appears to be

59r.3, explains its five syllables as designating the five paths for the mind-stream

[lam rgyud lnga brtag pa yin no], i.e. the five realms of rebirth (in which the asuras are not separated out from the gods).
give six [ give five, but 's list would seem to include the asuras, and perhaps omit the animal realm in

error]. The list of six as generally given [e.g. in the dKon mchog spyi 'dus, sngon 'gro section, p.25], is, a su n tri pre du,
respectively for gods, asuras, humans, animals, pretas and hell-beings.) | 32 : omits | 33 :

; ; |

34 :

; ; |

35 : omit

36 : ; ; | 37 : omit

180
38

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

[376v] 3940 41

38 inserts (dittographically)

: illegible | 41

39

40

Chapter 15
[ 59r.5] [302r.8] [307b.8] [293v.3] [ omits] [60r.4] [54v.3] [412.6]

[331r.6] [211r.1] [40v.3] [376v.1]

[331v] 2
3

[302v] 4

[308a] 5

[ 59r.6] 7
[ 59v.1] 8
[ 59v.1] [302v.1] [308a.1] [293v.3] [ omits] [60r.5] [54v.3] [412.7]

[331v.1] [211r.2] [40v.4] [376v.2]

9 1011 12

13 [413] 141516 17

omit;

5 :

illegible | 7

; ; (it is possible here that the

archetype's apparent reading of might have derived from an earlier, )

; ; ; | 6 :

: ;

by a shad break); ; ;

; ;

| 8 :

(not separated from the following three hs

| 9 : omits |

| 11 : ; ;

10

| 12 :

omit | 13 : ; | 14 : | 15 insert | 16 : ; | 17 :

Chapter 16
[ 59v.2] [302v.1] [308a.1] [293v.4] [ omits] [60r.6] [54v.4] [413.1]

[331v.2] [211r.2] [40v.5] [376v.2]

12 3

45

6
7 8

[ 59v.5] [ omits]

910

11 1213

1415 16

1718

[60v] 1920

[ 60r] 21 22

23 24

25 26
2728

29 3031

32 33

: omits | 2 : ; |

unclear;

: ; | 4 : ;

( agrees); | 5 : ; ; illegible | 6

: ; ; |

7 : omits | 8 :

14 :

| 16

10 :

11

: ; |

| 15 inserts

12 : ;

; |

13 :

: ;

| 17

9 :

18 : ; ; | 19 : ; omit | 20 : | 21 :

22

; ;

| 25 :

| 26 :

| 23 :

| 27 :

| 28 :

;
| 24

: ;

| 29 inserts

| 31 : no gap between the tshig rkang; small double tsheg inserted | 32 :

; ; ; ;

30

| 33 :

183

Chapter 16

34

35

36 37 [41r]

38 39

40 41 [55r]
42

43 44

45 46 47

48 49

50

51

[332r] 52 53

54

55 56
57

[294r] 58 59

[ 60v] 60 6162 63

64 65
34 :

35 :

; ; |

; ;

| 37

36

| 38

: |

| 40 : ; | 41 : ;
( also gives );
;

| 44

) | 49 :

);
;

| 46

| 47 :

| 56 :

| 54

but dbyig is an alternative/archaic word.) | 58 :

| 48 :

43 :

(but

( also gives

| 51

unclear);

| 55 :

; ; ;

(dbyug sngon is the more commonly used name,

| 57

| 64

; ;

| 63 :

| 45

; (but give

| 52 :

; ;

;
| 60

; ; |

(pa inserted as a small letter between rnams and ni) | 50

; unclear;

; ; ; ;

( agrees);
| 53

| 42 :

39 :

| 61

59

: ;

| 62 :

: ; ; ;

; (most probably, sku was an incorrect or non-standard spelling in the archetype) | 65 :

184

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

[377r] 66
68

67

69 70 71
72 73

75

74

76 77

78 79 [414] 80

[308b] 8182

[211v] 83
8485

8687 88

[61r] 8990

[303r] 91 92

93 94

959697
98 99
66 :

67

: ; |

68 :

; |

69

: ; ;

| 70 : ; ( also gives ); ; ('s reading

would seem preferable, and consistent with above, note 62, but / could also work) | 71 :

( also gives ) | 72 :

| 73

; ;

| 78 :

; ;
;

| 80 :

| 82 :

| 83

| 76 :

; ; ;

| 74

| 79

| 75

| 77

: ;
: ;

; ;

| 81

; ; ; ;

: ; ; |

84 :

| 85 : | 86 : ; (in an additional pa, inserted dittographically after g.yon pa, has been
deleted) | 87 : ; | 88 : ; ;

89 :

| 90 :

| 91

| 92

; (an

effaced superscribed letter appears above the ku); ; ; ; (as above, most

probably, skur was an incorrect or non-standard spelling in the archetype) | 93 : ; | 94 :

; ; ; |

below line, as a correction);

95

| 98

: |

96 :

| 99

( all give , which also seems more appropriate)

97 :

(zhes inserted

185

Chapter 16

100

[ 61r] 101 102

103104
105

106107
108109
110 111112

113

[41v] 114 [332v]

[55v] 115116117
119

120 121

122

123 124

118

125 126 127

[ 61r.5] [303r.4] [308b.5] [294r.5] [ omits] [61r.4] [55v.1] [414.5]

[332v.1] [211v.4] [41v.2] [377r.7]

128 129130 131132133

100
102

101

unclear;

| 103

line, as a correction) | 105

134

135

| 104 :

[-] (illegible)

; (zhes inserted below

: ;

; [

] ; |

106

: |

107 :

; | 108 : | 109 : ; ;

| 110

: illegible |
115 :

111

: |

112 :

; ;

113

114

| 116 : | 117 :

; ; (mchong would seem rather more appropriate and is given in but mchod could work here,

so given the apparent agreement between , we are accepting mchod, although it may have been an error in the
archetype.) | 118 : ; | 119 :

126

| 121

: | 122
124

: omits | 127 insert

| 128 :

; ; |

120

: illegible | 123

| 129

: ;
125

: ; ;

| 130 : ; ; | 131 insert | 132 : | 133 : omit

134 insert | 135

Chapter 17
[ 61r.6] [303r.4] [308b.6] [294r.6] [ omits] [61r.5] [55v.2] [414.5]
[332v.2] [211v.4] [41v.2] [377r.8]

12 3
45

67 [ 61v] 8
9 10

111213

14 [377v]

15

[ 61v.1] [ 292.2 rejoins] i

16 17

18 19 20

21 22

232425
26

i At this point, the brief passage of Chapter 17 text at the end of the Tenjur's conflated Chapters 13 and 17 rejoins.
1

: |

; | 5 :

2 :

; |

4 :

| 6 : ; unclear [ ?] | 7 : ; | 8

(the sa appears to have been deleted) | 9 :

| 10 :

| 11 :

; | 12 : | 13 : (in , an original seems to have been corrected to

by a small line crossing through the zhabs kyu.) | 14 : | 15 : ;

; | 16 :

| 17 :

| 18

| 19

; ;

(note in the presence under ga of a crossed out letter probably 'a indicating a long a vowel, or else ha);

| 20 : | 21

: omits | 22

(though is somewhat unclear) | 23 :

; ;

; [...] | 24 :

[...]; ( also give ); (dbyangs kyis, "made praises with vajra melodies",
would seem the most appropriate reading, and is given by both and [as well as ]; Bathang simply

gives kyi rather than kyis. dByings is shared by . 's reading, dbyings kyi, could be
meaningful: "made praises of the vajradhtu". It is possible that dbyings could have been independently introduced in error in the
South Central and the Dunhuang/Bhutan/ Tshal pa Kanjur line, but then corrected in the Tshal pa Kanjur transmission.) | 25 :
; | 26 : ;

187

Chapter 17

[ omits] [303r.6] [308b.8] [294r.7] [ 292.3] [ omit]27 [415.1] [332v.3]

[211v.5] [41v.5] [377v.2]

29

28

30 [294v]

31

32 33

34

35 36

[ 61v.3] [303r.6] [309a.1] [294v.1] [ 292.4] [61r.7] [55v.4] [415.1]

[332v.4] [211v.7] [41v.6] [377v.2]

[61v] 37 3839 40
4142 43

27 also omits

| 28 This verse of four tshig rkang has been omitted from the Dunhuang manuscript, and also from the

Bhutanese editions. Since the Bhutanese omission is doubtless a deliberate policy relating to the boundaries of the root text (see
p.40 above), the omission is almost certainly coincidental. See also the parallel verses in Chapters 24, 29 and 34 below. | 29 :

30 :

31 :

| 35 :

| 36 :

| 39 :

| 42 insert | 43

| 37

: omits | 38

; ;

32 :

33

34

: ;

| 40 : omit | 41

Chapter 18
[ 61v.4] [303r.7] [309a.1] [294v.1] [ 292.5] [61v.1] [55v.5] [415.2]
[332v.4] [211v.7] [41v.6] [377v.3]

1 2 3

45 6

7 8 9
10

11

[ 62r.4] [ 293.3]

12 13 14

151617

1819

20

21

[ 62v.3] [ 293.6]

222324 [212r]25

1 :

2 :

| 6 :

; |

3 :

4 :

5 :

; | 7 : ;

; ( also gives ); (here, the intended word might be as given

in the Dunhuang version, and the archetype's might be an archaic spelling for this. However, although would suggest

a more coherent reading, some sense can be made of / and even of the South Central and Bhutanese reading of . See

our TZComm edition, p.315.) | 8 : ; | 9 : | 10 : | 11 inserts

(deletion of bya appears to be indicated by a line through the letter, and the following 'gyur is written in small letters, probably as
an inserted correction) | 12

17 :

| 18 :

| 21

| 23 :

| 14

| 19 inserts

this is misplaced from another collection.

| 24 :

13

| 15

| 16 :

; [(/)];

| 20 :

| 22 :

| 25 gives a second folio 211 before 212, but

189

Chapter 18

26 27 28

2930

[303v]31 32

33

34 35 [ 294] 36

[ 63r.3] [303v.1] [309a.3] [294v.3] [ 294.3] [61v.4] [55v.7] [415.4]

[332v.6] [212r.1] [42r.1] [377v.5]

37 38 3940 [333r] 41 42
43 44454647 48

26 :

28

; ; |

27

; ;

initial ba of bcad is written small, inserted above the line as a correction);

; |

30 :

(the

| 29 :

31 The verso of this folio (303v) is handwritten, as presumably the

block for this page was lost or damaged. It holds 9 lines, instead of the usual 8. | 32 : ;

33

: |

34 :

as an addition above the line | 36 :

omits;

losing

| 38 :

; ;

35 : inserted in

37

: ;

| 41 : omit | 42 torn,

: ; |

| 46 insert | 47 : | 48

; |

; ; |

(the sa appears to have been deleted) | 39

| 40 :
| 43

44 :

45

Chapter 19
[ 63r.5] [303v.1] [309a.4] [294v.3] [ 294.4] [61v.4] [56r.1] [415.5]
[333r.1] [212r.1] [42r.1] [377v.6]

1 23 4
56
7

9 10 11

[ 63v.4] [ 295.2]

12 13

14 1516 17

1819 20

21

222324

25 26

27 2829

30 31

32 [378r] 33

35

: |

5 :

34

: |

; ;

| 10 :

3 :

; |

| 6 : | 7 : | 8 :

; ; ; ;

; |

13

| 9 :

| 11 : | 12 :

[ inserts

(dittography)]; ; ; ;

14

; ; ;

; ; |

16 :

; |

; ( agrees); |

omits | 22 : | 23 : | 24 : | 25

| 28 :

| 29

; ; | 15 :

20

17 :

18 :

: ( agrees) |

; ;

21

19 :

: ;

| 26 : | 27

: in small writing at the upper margin, preceded by a small mark that is

matched at the insertion point lower down on the page | 30

: | 31 : ; omits; | 32 :

33

: | 34

: ; ;

| 35 :

191

Chapter 19

[ 64r] 3637
[ 64r.4] [303v.4] [309a.6] [294v.5] [ 295.6] [61v.7] [56r.3] [415.7]
[333r.3] [212r.3] [42r.4] [378r.1]

38 3940 41 42 43 [62r] 44 [416] 45

36

; ; |

37

: omit

(almost certainly, this reflects a mistaken decision about the root text boundaries) | 38 : omit | 39 :

; ; | 40 : ; | 41

insert ; has a faint , presumably inserted retrospectively | 42 :

; ; ;

| 44 insert | 45

| 43

Chapter 20
[ 64r.5] [303v.4] [309a.6] [294v.5] [ 295.6] [62r.1] [56r.4] [416.1]
[333r.3] [212r.4] [42r.5] [378r.2]

1 2 [ 296]

45

67
8 9

[ 64v.6] [303v.5] [309a.7] [294v.6] [ 296.5] [62r.2] [56r.5] [416.1]

[333r.4] [212r.4] [42r.5] [378r.2]

10 11 1213

14 15 1617 [ 65r] 18

1 :

; |

: omits; | 4 :

5 inserts | 6 : ; | 7 : ; | 8

| 10

omit | 15

| 18

; |

omit | 11

| 9 :

: ; | 12 : ;

13 :

: ;

; ; |

| 16

: ;

14

| 17 insert

Chapter 21
[ 65r.1] [303v.6] [309a.7] [294v.6] [ 296.5] [62r.2] [56r.5] [416.2]
[333r.4] [212r.5] [42r.6] [378r.3]

123

45

6 7 8 910

11 12

[ 65r.6] [303v.6] [309a.8] [294v.7] [ 297.2] [62r.3] [56r.6] [416.2]

[333r.5] [212r.5] [42r.7] [378r.4]

13 1415

16 17 181920 [ 65v] 21 22

1 :

| 2 : | 3 :

(but gives

); ; | 5 :

| 4

(but is given clearly in its commentarial text below, 65r.3);

; ; | 6 : ;

; ; ; (However mud ga is spelt, we cannot find a definite identification, although it is quite

likely that the reading of mud ga derives from mung ga; Roerich [1986 Volume 7: 79] gives mung ga'i lo ma as the leaf of a pea,
so mung ga could be a kind of pea, which would fit the context (especially since the commentary below gives sran mud dga').
This could well be derived from an Indian word for mung beans (Hindi moong). If mud ga/ mud dga' does indeed derive from
mung ga, this might be one piece of evidence which would suggest the likelihood of an ancestor of all the root texts having been
copied from an dbu can source and introducing the error, since da/nga confusions are more likely in dbu can script.) | 7 omits
from here until below, in the title. | 8 : ;

: omit ( shares this omission) | 11 :

| 13

again here, with

17

; | 9 : |

: omit | 14

: ;

) | 15 : ; ;

: ;

21 :

| 18

: ;

10

| 12 : ;

| 16

; ( starts

: omit

| 19 insert

20 :

; inserted as a correction below the line, in black ink rather than the coloured ink of the main

chapter numbering the initial reading was simply | 22 :

Chapter 22
[ 65v.1] [303v.7] [309b.1] [294v.7] [ 297.3] [62r.4] [56r.7] [416.3]
[333r.5] [212r.6] [42r.7] [378r.5]

1 2

3 45

67 8

910 [295r]

[ 66r.1] [303v.8] [309b.1] [295r.1] [ 298.1] [62r.5] [56r.7] [416.4]

[333r.6] [212r.6] [42r.8] [378r.5]

11 [56v] 12 1314

15 161718 19

1 :

; |

;
;

illegible | 13

| 15

: |

| 6

: ;

; | 3 : ;

5 :

; ;

; ; ; ;

| 7 : ;

; ; ; |

: ; ;
: omit | 16

11

omit | 12

;
| 19

10 :

| 14 : ;

: ;

| 17 insert | 18 : ;

| 9 : |

Chapter 23
[ 66r.2] [303v.8] [309b.2] [295r.1] [ 298.2] [62r.5] [56v.1] [416.4]
[333r.6] [212r.6] [42v.1] [378r.6]

1 2 [333v] 34

5 6 7

8 9

10

11

12

13
14

1617

15

[ 67v.1] [303v.9] [309b.3] [295r.2] [ 299.6] [62r.7] [56v.2] [416.6]

[333v.2] [212r.8] [42v.2] [378r.8]

18 [304r] 19 2021 22
23 2425 26

5 :

|6

; ;

(abbreviated sa written under the postscript ba) | 4 : omits

; ; | 7 :

| 8 :

| 9

: ; ;

( agrees);

( agrees); ; ; (note that only and

give the presumably correct verb, rngubs/rngub [see our TZComm edition, Chapter 23, note 8]; the archetypal reading is
uncertain.) | 10 : ; | 11 : ;

; ; | 12 :

15 :

omit | 23

16

: omits | 18
;

; [abbrev] | 26

| 13 : unclear [?] | 14 :

: omit | 19 :

| 21 :

| 20

17

: ;

; |

; |

22 :

24 insert

25 :

Chapter 24
[ 67v.2] [304r.1] [309b.4] [295r.3] [ 300.1] [62v.1] [56v.3] [416.6]
[333v.2] [212r.8] [42v.3] [378r.8]

1 2

[378v]

4 [212v] 5

67 8

[ 67v.4] [304r.2] [309b.4] [295r.3] [ 300.2] [ omit]10 [416.7] [333v.3]

[212v.1] [42v.4] [378v.1]

11 12

13 [417]
14

15 1617

[ 67v.6] [304r.2] [309b.5] [295r.4] [ 300.3] [62v.1] [56v.4] [417.1]

[333v.3] [212v.1] [42v.5] [378v.2]

18 19

1 :

; | 2 : ; | 3

11

12

| 14

| 17

;
|

13

: ;

inserts the as a correction in small writing);

| 5 :

| 7 : ; omits | 8 : ; ; | 9

: | 4 :

| 15 :

| 10 also omits

; (

| 16

: ;

(and also ) omit (Although the Tshal pa Kanjur, Bathang, Hemis, Tawang and South Central versions

accept this verse as root text, it seems clear that this omission in the Bhutanese version is a deliberate editorial policy rather than
accidental. The Bhutanese version also omits parallel verses in the chapters opening the treatment of the drag po, rgyas pa and zhi
ba ritual actions (see Chapter 17, Chapter 29, and Chapter 34). The Dunhuang manuscript highlights the first line of the verse, but
the verse ends in the Dunhuang commentary with the words, ces bstod de [ gives ces ston to], rather than zhes gsungs te,

which closes the chapter's other root verses.) | 18 : ; | 19 :

; ; (perhaps an abbreviation is intended, but a tsheg separates the

final nga of mang and the initial ga of gnas).

197

Chapter 24

20 21

22 [ 68r] 23

24

25 26

[ 68v.3] [ 301.3]

2728 29

3031

32 3334

35

36 37

[ 69r2] [304r.4] [309b.6] [295r.5] [ 301.6] [62v.3] [56v.5] [417.2]

[333v.5] [212v.3] [42v.6] [378v.3]

38 3940

4142 43 4445 46

20 :

| 21 : ; ( agrees); | 22 :

; omits;
| 26

| 27

);

| 23 : unclear; | 24

[...] (unclear);

| 28

: omits | 29

(but

has

| 25 :

, and

also gives

(perhaps an abbreviation is intended, but a tsheg separates the final nga of mang and the

initial ga of gnasu). | 30 : ; ; ; ; ;

31 :

35 :
38

43 :

32 :

; ;

omit;

| 39

; ;
| 44

| 36

33 :

: ;

| 41

; |

| 37 :

: omit | 42 :

| 45 :

| 46

34 :

| 40 :

Chapter 25
[ 69v.4] [304r.4] [309b.7] [295r.5] [ 302.1] [62v.4] [56v.6] [417.3]
[333v.5] [212v.3] [42v.7] [378v.4]

23 4

8 9

[ 70r.1] [304r.5] [309b.8] [295r.6] [ 303.1] [62v.5] [56v.7] [417.4]

[333v.6] [212v.4] [42v.8] [378v.5]

10 1112 13
14 151617 18

1 inserts

(most probably, visually picking up from the start of Chapter 26) |

; ; |

: ;

: | 7 : ; | 8

here) | 9 :

; [?]

| 10

| 12 :

: | 3 :

; |
;

: omit | 11

5 :

; [--] (the text is torn

: ;

; ; |

13 :

omit | 14 : ; | 15 : (but and both give ); ;

; | 16

insert | 17 :

| 18 :

Chapter 26
[ 70r.3] [304r.5] [309b.8] [295r.6] [ 303.1] [62v.5] [56v.7] [417.4]

[333v.6] [212v.4] [43r.1] [378v.6]

1 2

345 [57r]

7 89 [310a] [334r]

10

11

[ 70v.5] [304r.6] [310a.1] [295r.7] [ 303.6] [62v.6] [57r.1] [417.5]

[334r.1] [212v.5] [43r.2] [378v.6]

12 13 14 [ 304] 1516 17
18 192021 22

: |

: |

3 :

(mgo bo at the

beginning of line 2 is consistent with the three other chapters on the phur pas of the different ritual actions [although they do not
repeat it in line 3]. None of the other chapters give kha dog; they all simply give the colours, and the words, shing rnams la
complete the first line. It seems that TZComm wanted also to give la here, and la is also found in the commentarial gloss.)
5 : | 6 inserts

(insertion shared by

and ; it seems most likely

here that had a simple dittography, moving from mgo la at the beginning of line 3 back to rgya mdud in

line 2.) | 7 : ; | 8 : ; (note that

's reading, shared by and , is more consistent with both the destructive phur pa verse in Chapter 20, and

also the increasing and pacifying phur pa chapters, Chapter 31 and Chapter 36. Moreover, the gloss which follows in the
commentary would seem to favour this reading, although it is not shared by the Dunhuang TZComm version of the root verse. It is
possible that it could have been introduced by the Tenjur editors.) | 9 : ; ; | 10 :

15

11 :

: ;

12

| 17 : omit | 18

omit | 13 :

| 16 :

14

; ;

; (see notes 4 and 8 above.

Here again, either reading could work, but in this case, the commentaries' shared reading is consistent with the chapters on the
phur pas for the other of the 3 ritual actions, apart from the South Central, Hemis, Tawang and Bathang version of Chapter 31 on
the rgyas pa'i phur pa.) | 19 : | 20 insert | 21 : | 22 :

Chapter 27
[ 70v.6] [304r.6] [310a.1] [295r.7] [ 304.1] [62v.7] [57r.2] [417.6]
[334r.1] [212v.5] [43r.3] [378v.7]

12

3 4

5 6 [ 71r]
7 [295v] [63r]

[ 71r.5] [304r.7] [310a.2] [295v.1] [ 304.5] [63r.1] [57r.3] [417.7]

[334r.2] [212v.6] [43r.3] [378v.8]

9 1011 12
1314 15

: ; ; |

2 :

( gives );

; (here, TZComm's elaboration below supports 'bras chan, which is clearly the earlier and

a more appropriate reading than 's 'bras can.) | 4 : | 5 :

; |

: ; |

(consultation of the chapters on gtor mas Chapters 21, 32, 37 is not entirely conclusive, but given the comparisons, a
reading of gtong bas or gtor bas might be expected. However, the reading mthong bas could make good sense, implying that
perceiving the gtor ma is considered to effect the ritual action.) | 8

| 9

11 :
14 :

| 15

: omit | 10

: ;

; ;

| 12 : omit | 13 insert

Chapter 28
[ 71r.6] [304r.7] [310a.3] [295v.1] [ 304.5] [63r.1] [57r.3] [417.7]

[334r.2] [ omits] [43r.4] [379r.1]

[ 71v]
1 2

[418] 3
4

[ 71v.6] [304r.8] [310a.3] [295v.2] [ 305.3] [63r.2] [57r.4] [418.1]

[334r.3] [ omits] [43r.5] [379r.1]

5 67

8 910 [ 72r] 11 12

1 :

; | 2 : ; ;

|3

: ; ;

; ;

omit | 9

insert | 11 :

| 12

| 4 : ;

: omit | 6

| 7 :

: ;

| 5

| 8

| 10

Chapter 29
[ 72r.1] [304v.1] [310a.4] [295v.2] [ 305.4] [63r.3] [57r.4] [418.1]

[334r.3] [212v.6] [43r.5] [379r.2]

1 2

3 4 5
67 8

9 10

11

[ 72r.3] [304v.1] [310a.4] [295v.3] [ 305.5] [ omit]12 [418.2] [334r.4]

[212v.7] [43r.6] [379r.3]

13

14 15 16

17 18

1920

2122

[ 72r.5] [304v.2] [310a.5] [295v.3] [ 305.6] [63r.4] [57r.5] [418.3]

[334r.5] [212v.8] [43r.7] [379r.4]

23 24

25 26 [ 306]
1 :

; ; has erased a few letters here in correction. |

(abbreviated sa written below the ma postscript) | 4

13

: ;

| 10

| 6

| 20 : omits | 21

| 18

| 11 :

3 :

| 7 :

| 19

| 12 also omits

| 14 has erased a few letters here in correction | 15 :

| 17 :

| 9

: omits | 5 :

2 :

16

: ;
|

22

(and also

) omit (As in the case of the parallel verses in the chapters introducing the other three ritual types, drag po, dbang, and zhi

ba [see Chapters 17, 24 above, and Chapter 34 below], it seems clear that the omission in the Bhutanese version is a deliberate
editorial policy rather than accidental. In this case, the Dunhuang manuscript highlights the first and the last line of the verse,
seemingly indicating them as root text, but as in the other chapters, the verse ends in TZComm with the words, zhes bstod de [this
time a reading given in both and ], rather than zhes gsungs te, which generally closes the root verses.) | 23 :

; omits | 24 : ;

| 26 :

; | 25 :

203

Chapter 29

27 28

3031 32

29

[ 73r.1] [304v.3] [310a.6] [295v.4] [ 306.5] [63r.5] [57r.6] [418.4]

[334r.5] [212v.8] [43r.8] [379r.5]

33 3435

36 [43v] 373839 40

27 : | 28 : ; ; | 29

: | 30 : ; ( as

inserted correction in small writing) | 31 : | 32 : (an erased ro is just about visible after

'gyuro) | 33 : omit | 34 : ; ; | 35 :

; |

( also omit);

36

omit | 37

omit

( include it against , with the descendants of b [on the stemma codicum p.43

above] divided) | 38 insert | 39 :

| 40 :

Chapter 30
[ 73r.2] [304v.3] [310a.6] [295v.4] [ 306.5] [63r.5] [57r.7] [418.4]
[334r.6] [213r.1] [43v.1] [379r.5]

12 3
4

[ 73v.4] [304v.4] [310a.7] [295v.5] [ 307.5] [63r.7] [57r.7] [418.5]

[334r.6] [213r.1] [43v.2] [379r.6]

8 [57v] [334v] 910 11

121314 15

: |

2 :

| 8

| 10 :

: |

: omit | 9

| 15

| 7

: ;

; ;

; [...] (unclear); ; ;

| 11 : omit | 12

| 13 insert

5 :

| 14 :

Chapter 31
[ omits] [304v.4] [310a.8] [295v.5] [ 307.5] [63r.7] [57v.1] [418.6]

[334v.1] [213r.2] [43v.2] [379r.7]

1
2 34

56

[63v]
[ omits] [304v.5] [310a.8] [295v.6] [ 308.2] [63v.1] [57v.2] [418.6]
[334v.1] [213r.2] [43v.3] [379r.7]

8 9 [310b] 1011 12
13 141516

: ; | 2 : ; | 3 : partially illegible

4 :

5 :

; |

(abbreviated sa written beneath the postscript ba) | 10

| 11 :

| 14

6 :

omit;

: ;

; ;

7 :

omits

| 12 : omit | 13

(note the variant readings in Chapter 26 on the dbang

gi phur pa, where all the root text versions include the word, las, but both versions of TZComm omit it; and all add bstan pa'i,
apart from the Tshal pa Kanjur versions which give bshad pa'i.) | 15 insert | 16 :

; ;

Chapter 32
[ omits] [304v.5] [310b.1] [295v.6] [ 308.3] [63v.1] [57v.2] [418.7]
[334v.2] [213r.3] [43v.4] [379r.8]

1 2 3

4 56
[379v]

7 8

[419]

[ omits] [304v.6] [310b.2] [295v.7] [ 308.6] [63v.2] [57v.3] [419.1]

[334v.3] [213r.3] [43v.5] [379v.1]

1011

12 131415

1 :

| 2 : | 3 : | 4 :

| 5 : | 6 : (here, seem repetitive, and moreover, this

reading [also shared with ] would not seem likely, given that the chapters on gtor mas for the other ritual actions Chapters

21, 27, 37 include appropriate liquids with which to mix the dry ingredients.) | 7

;
;

(small prefixed a chung inserted as a correction) | 9

: |

(The reading of the South Central, Bathang, Hemis, Tawang and Tshal pa Kanjur editions is consistent with the

corresponding wording for the pacifying gtor ma in Chapter 37, while the Tenjur and Bhutanese versions are slightly more in line
with the corresponding wording for the destructive and captivating gtor mas in Chapters 21 and 27.) | 10 :

; ;

omit | 13

; |

15 : ;

11 :

; ; |

12

| 14 insert

Chapter 33
[ 73v.6] [304v.7] [310b.2] [295v.7] [ 309.1] [63v.3] [57v.4] [419.2]
[334v.3] [213r.4] [43v.6] [379v.2]

1
2

4 [ 74r] [296r] 5

6 7

[ 74r.4] [304v.7] [310b.3] [296r.1] [ 309.3] [63v.4] [57v.5] [419.2]

[334v.4] [213r.5] [43v.7] [379v.2]

8 910 11
12 1314 15

1 :

| 2 : ;

5 :

; ;
8

10 :

13

omits | 9

| 6 : ; ; | 7 :

; ;

14 insert

| 3 : ; | 4 :

| 11 : omit | 12

15

Chapter 34
[ 74r.5] [304v.8] [310b.3] [296r.1] [ 309.4] [63v.4] [57v.5] [419.3]
[334v.4] [213r.5] [43v.7] [379v.3]

3
4 5

[ 74v] [44r] 7 [305r]

89

[ 74v.2] [305r.1] [310b.4] [296r.2] [ 309.6] [ omit]10 [419.4] [334v.5]

[213r.5] [44r.1] [379v.4]

11

12 13

14 15
16

[ 74v.4] [305r.1] [310b.4] [296r.2] [ 310.1] [63v.5] [57v.6] [419.4]

[334v.6] [213r.6] [44r.2] [379v.5]

1718

19

20

: ; ; |

; |

; ;

[...]

| 3 : | 4

: ; ; | 5 :

: ; unclear |

(illegible);

: |

| 9 has the entire tshig rkang (

) above the line, as a

correction. | 10 also omits | 11 (and also ) omit the next verse of four tshig rkang:

(As in the case of the parallel verses in the chapters

introducing the drag po'i las see Chapter 17 above dbang gi las see Chapter 24 above and the rgyas pa ritual see

Chapter 29 above it seems clear that the omission in the Bhutanese version is a deliberate editorial policy rather than accidental.
In this case, the Dunhuang manuscript highlights the whole verse, seemingly indicating it as root text, and it fails to highlight the
third verse, which all versions agree on as root text. But once again, as in the previous chapters, the verse ends in TZComm with
the words, zhes bstod de, rather than zhes gsungs te, which closes the chapter's other root verses.) | 12 :

13

; |

; ; |

17

14

: |

: |

18 :

15 :

; unclear |

19

16 :

: |

20

209

Chapter 34

21 22 23
25 26

24

[ 75r.2] [305r.2] [310b.5] [296r.3] [ 310.3] [63v.6] [57v.7] [419.5]

[334v.6] [213r.7] [44r.3] [379v.6]

27 28 29 [335r] 30 31

32 3334 35

21 :

; ; ; | 22 : | 23 : | 24 inserts

an accidental repeat of the phrase above. |

;

| 27

: omits | 28

25 :

: ;

| 30 :

omit | 32 : | 33

: omits; | 34

26

| 29

: ;

: ;

; |

insert | 35

31 :

: ;

Chapter 35
[ 75r.4] [305r.3] [310b.6] [296r.3] [ 310.4] [63v.7] [58r.1] [419.6]

[335r.1] [213r.7] [44r.4] [379v.6]

12 3 4

8 9 [64r]
[ 75v.2] [305r.3] [310b.6] [296r.4] [ 311.1] [64r.1] [58r.1] [419.6]

[335r.1] [213r.8] [44r.4] [379v.7]

10 1112 13

1415 16

: ; | 2 : | 3 : |

(almost

certainly, las was in the archetype here; las may also have been in the archetype for two of the comparable chapters on the other
ritual activities, although this is much less certain, since are evenly divided, supporting las in the case

of Chapter 19 [with giving la], and supporting las in Chapter 30 [with giving la]. In Chapter 25, however,
only gives las. agrees with in all four cases, giving la in one instance, and las in three. Given the inconsistency

across the chapters and between editions, and the greater appropriateness of la, we are assuming a minor error in the archetype in
this instance.) | 5 : ; ; ; ; ;

[ ] ; |

omits | 11

16

6 :

| 9

: ;

| 13 : omit | 14

: |

| 12 :

: ; ;

| 10

; ;

| 15 insert

( inserted small beneath line as a correction);

Chapter 36
[ 75v.4] [305r.4] [310b.7] [296r.4] [ 311.1] [64r.1] [58r.2] [419.7]

[335r.2] [213r.8] [44r.5] [379v.7]

23 4

56

[420] 7
[ 76r.2] [305r.4] [310b.7] [296r.5] [ 311.4] [64r.3] [58r.3] [420.1]

[335r.2] [213r.8] [44r.6] [379v.8]

8 [213v] 910
11 12 [380r] 13 14

unclear; omits this tshig rkang

(omission shared by and ); ; (It is clear from a

comparison with the chapters on the phur pas for the other three ritual activities Chapters 20, 26, 31 as well as TZComm,
which specifies silver or white wood in the commentarial line beneath the verse, that we require a line here as given in the South
Central, Bathang, Tawang and Hemis versions. Although the line shared by the Dunhuang manuscript, the Tshal pa Kanjur and
the Bhutanese version makes good sense, it has most likely been introduced by hypearchetype c, the ancestor of this branch of the
transmission, or perhaps hypearchetype b, their common ancestor with [as shown on the stemma codicum
p.43 above] , but this is cannot be ascertained since omit the line altogether.)

; (here, 's reading, inconsistent with Chapters 20, 26 and 31, and less coherent than mgo bo, would

appear to represent an error in the archetype, corrupted further by , and corrected in hypearchetype b.) | 3 :

| 4 :

| 5 :

; ; ;

(bsgoms te occurs in Chapter 20 [destructive phurpa], in Chapter 26 [captivating phurpa], and in Chapter 31 [increasing

phurpa], and would seem preferable; la was almost certainly in the archetype, and can fit, but it was probably emended by the two
commentaries, and by .) | 6 : | 7 : ; | 8 : omits;

| 11

: omit | 12

| 10 :

: omits | 13 : ; omits | 14

: ;

Chapter 37
[ 76r.3] [305r.4] [310b.7] [296r.5] [ 311.4] [64r.3] [58r.3] [420.1]

[335r.3] [213v.1] [44r.7] [380r.1]

1 2 3

4 5

6 7

[ 76v.1] [305r.5] [310b.8] [296r.6] [ 312.1] [64r.4] [58r.4] [420.2]

[335r.4] [213v.1] [44r.8] [380r.2]

9 1011 12
1314 15

(here, we would expect

dngul phye, to be consistent with the previous chapters on the gtor mas for the other ritual actions, Chapters 21, 27, 32. The ra
head letter in place of the da prefix is common in older Tibetan sources. It is found here in ; and in the Bathang
and Hemis versions of Chapters 35 and 36 [the South Central, Tawang and Tshal pa Kanjur versions of Chapter 35 apparently
corrupted it to rdul], and it was almost certainly present in the archetype here. has a nga/da confusion on the first syllable,
but clearly intends this reading. 's lam would seem rather less appropriate than phye or dang, possibly picked up

from the wording opening the previous chapter on the pacifying phur pa. 's dngul dang could work but is
less appropriate than dngul phye, and may have been a correction of the inappropriate dngul lam.) | 2 :

; [?] (

| 4 : ; unclear; ;

13

| 12 : omit

| 9

: omits | 14 insert | 15

| 6 :

| 11 : ;

; ;

| 8

| 5 :

| 7 :

is given but it appears to have been amended to ) | 3 : unclear;

omits | 10

: ; ;

Chapter 38
[ 76v.3] [305r.6] [311a.1] [296r.6] [ 312.2] [64r.4] [58r.5] [420.2]

[335r.4] [213v.2] [44v.1] [380r.2]

4 5 6
7

[ 77r.2] [305r.6] [311a.1] [296r.7] [ 312.5] [64.5] [58r.5] [420.3]

[335r.5] [213v.2] [44v.1] [380r.3]

8 9 1011 12
13 1415 16

1 :

2 :

| 7

13 :

; |

| 11

| 5

| 8

: omits | 9

; ; ;

| 10

: ; ;

12 : omit

: ; ; ; [

] ; omits | 15 insert
;

: ;

| 6 :

14

16

: ;

Chapter 39
[ 77r.3] [305r.7] [311a.2] [296r.7] [ 312.6] [64r.6] [58r.6] [420.4]

[335r.5] [213v.3] [44v.2] [380r.4]

1 2

345

6 [ 313]

7 8 9

[ 77v.3] [305r.7] [311a.2] [296v.1] [ 313.4] [64r.7] [58r.7] [420.5]

[335r.6] [213v.3] [44v.3] [380r.5]

10 1112 13
14 15 16

1 :

; ; ; ;

7
11

| 8 :

; |

; ;

: ; ;

16

: |

| 12 :

| 13 : omits | 14 :

| 9

| 10

4 :

omits

| 15 insert

: ; (though partially illegible);

Chapter 40
[ 77v.4] [305r.8] [311a.3] [296v.1] [ 313.5] [64r.7] [58r.7] [420.5]

[335r.6] [213v.4] [44v.4] [380r.5]

1 2 [64v]

[58v] 3 4
5

[335v] 78
[ 78v.1] [305r.8] [311a.4] [296v.2] [ 314.4] [64v.1] [58v.1] [420.6]
[335v.1] [213v.5] [44v.5] [380r.6]

9 1011 12
13

14 [305v] 15 1617

: |
5

7 :

: | 3 : ; [...] ( illegible) |

| 8 : | 9

omit | 13

| 11 :

: omits | 10

: ;

: ;

; |

; ; ;

12 :

; (the addition of las may have been a commentarial gloss which persisted into the Tshal pa Kanjur and

Bhutanese root texts) | 14 : ; | 15 : ; omits ( unreadable, but the shad is given also in
) | 16

: | 17 : illegible

Chapter 41
[The Dunhuang manuscript does not highlight any of the root text in Chapter 41, and the wording of
TZComm does not clearly indicate the root text. The Bhutanese version includes only the opening two lines.
The Tshal pa Kanjur version, however, treats the whole commentarial chapter as root text. As with previous
chapters where one or the other of these two versions includes commentarial additions, we are presenting the
Tshal pa Kanjur additions (in an indented paragraph format), not because they might be part of the archetypal
root text but simply since their inclusion has become part of the transmitted text in this branch of the
tradition.]
[ 78v.2] [305v.1] [311a.4] [296v.2] [ 314.4] [64v.2] [58v.2] [420.7]
[335v.1] [213v.5] [44v.5] [380r.7]

1 23

4 5 6
[ (and similarly, ) move from here directly to the start of Chapter 42 (/de skad ces), which
st
count as the 41 and final chapter.]

7
____________________________________________________________________
[The Tshal pa Kanjur adds the next commentarial words:]
[ 78v.3] [305v.1] [311a.5] [296v.2] [ 314.5]

8 9
____________________________________________________________________
[Here, the main root text resumes:]

1011

12

13 14

: illegible | 4

15

: omit | 5 :

| 6 :

; unclear; ;

] (the opening de la and closing ni

here are clearly commentarial additions); [ ] ;

(, however, gives );

( is subscribed as a correction);

10

: ;

; ;

| 13 : | 14

| 11 :

| 15

| 12

| 8 :

| 9

; ;

217

Chapter 41

____________________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 78v.4] [305v.1] [311a.5] [296v.3] [ 314.6]

161718 19 20 212223

24
25

26

27 28 29

30 [ 315] 31

34 [ 79r] 35

32 33

____________________________________________________________________
[Here, the main root text resumes:]

36 37 [421]

38

____________________________________________________________________
[The Tshal pa Kanjur adds the next commentarial words:]
[ 79r.1] [305v.3] [311a.7] [296v.4] [ 315.2]

3940 4142 43

44

____________________________________________________________________
[Here, the main root text resumes:]

4546 4748
16

: | 17 : | 18 :

| 22 :

| 19 : | 20 insert | 21

; unclear | 23 : omit | 24 while the shad in is

not clearly visible, a space is left, indicating that a shad break is intended | 25 : omit (while the shad in is not

clearly visible, a space is left, indicating that a shad break is intended) | 26 : omit | 27 : | 28 :

29

30 :

31 :

32 :

insert | 34 : ; unclear | 35 : | 36 :

39

43

| 37 :

: ;

| 40 :

| 44 inserts

| 45

: illegible

; ;

| 38 :

| 41 :

but it appears that a tsheg is added after na, and the final sa deleted) | 46 :

48

33

| 42 insert

; ; (nas is given,

; omits |

47 inserts

218

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

____________________________________________________________________
[The Tshal pa Kanjur adds the next commentarial words:]
[ 79r.2] [305v.3] [311a.7] [296v.4] [ 315.2]

4950 51

5253 54

____________________________________________________________________
[Here, the main root text resumes:]

5556
____________________________________________________________________
[The Tshal pa Kanjur adds the next commentarial words:]
[ 79r.3] [305v.4] [311a.8] [296v.5] [ 315.3]

5758 59
____________________________________________________________________
[Here, the main root text resumes:]

60

____________________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 79r.4] [305v.4] [311a.8] [296v.5] [ 315.3]

6162 63 [311b] 64 65 66 67
68 69

____________________________________________________________________
[Here, the main root text resumes:]
[380v]
49

52

56

70

[-]

: ;

| 59

(syllable unclear) |

: illegible | 57

62 :

50

| 53 : | 54

51

: ;

| 55 : omits

: ; | 58 : ;

60

: illegible |

| 63 inserts | 64 :

67 insert | 68 insert | 69 : | 70

61

| 65 inserts

: omit ( also omits)

| 66 insert

219

Chapter 41

____________________________________________________________________
[The Tshal pa Kanjur adds the next commentarial passage:]
[ 79r.6] [305v.5] [311b.1] [296v.6] [ 315.4]

71 72 73 74

79v] 7778

7576

79

8081 82 83
85

86

84

____________________________________________________________________
[Here, the main root text resumes:]

87 [ 316] 88 89
____________________________________________________________________
[The Tshal pa Kanjur adds the next commentarial words:]
[ 79v.4] [305v.7] [311b.3] [297r.1] [ 316.1]

90

____________________________________________________________________
[Here, the main root text resumes:]

91 92 93
____________________________________________________________________
[The chapter content ends here in , proceeding to the chapter title, but the Tshal pa Kanjur
adds the following commentarial passage:]
[ 79v.4] [305v.7] [311b.4] [297r.1] [ 316.1]

9495 96 97
71

| 80 : (pa appears to have been erased; gives ) | 81 : ;

75 :

82 inserts
86 inserts

; |

| 76 : omits | 77 insert
| 83 :

| 87

| 94

72 inserts

| 88 :

98

73 :

| 78 : omits | 79

74 :

| 84 inserts

| 89

omits

| 85 inserts

; ;

| 90 inserts

91 :

92

: ; |

93 :

insert | 95 : | 96 insert | 97 : | 98 inserts

220

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

99 100

101 102
103 104 105106

107 [ 80r] 108 109 110111 112

[306r] 113114115 116 117

118 119

121 122

120

123

124 125 126

127 128 129 130

131 132133

134 135

136 137138 139 140

[ 80v] [312a] 141

[ 317]
143

142

144 145

146147 148

149 150 151

99 :
104

100 :

101 insert
105

102 :

| 112 :

more appropriate here.) | 114 inserts

| 118

| 113 :

| 115 :

122 inserts

| 123 insert

| 127 :

| 124 :

| 128 insert

; ;

| 129

140 :

| 142 : omits | 143 :

: ;

| 141 insert

| 146 inserts

| 137 :

| 147 inserts

| 151 :

| 148

117 :

| 120

| 125 insert

| 126

131 : | 132 inserts

: ;

(sems ma would seem

121 :

| 130 :

| 135

: ; ;

133 inserts | 134

103 :

: ;

| 116 :

: | 119 : ;

: omits |

(tsha ba seems more appropriate here)

106 : | 107 : | 108 : | 109 inserts | 110

| 111 :

| 149 inserts

| 138 inserts
| 144 :

| 145

| 136

| 139
| 150

221

Chapter 41

152 153

154 155 156 157

158159

160 161 162

163164 165

166 167 168

169 170

171

[ 81r] 172173 174 175

176

177 178 179

181 182

183184

185

[297v] 180

186 187 188 189

190191

192193 194 [306v]

195 [ 318] 196

197198 199
152

161

| 157

153

:
:

( 's reading with the addition of

| 158

: omits | 162

| 166 inserts

[ ; ]

: ;

| 167

Vol.2, 1648) | 169 inserts

154 inserts

155 :

| 159 inserts

would seem most appropriate here.) | 165

: ;

| 160 inserts

| 163 inserts

| 168

156

| 164 :

: ;

: ;

(Bod rgya tshig mdzod chen mo gives: spa gong ba: [rnying] dpa' zhum pa,

170 :

171

(Here we

have three different readings, equally plausible, although perhaps the imagery drawn on in 's use of the word, rgyun (stream)

may seem rather more appropriate for the symbol of the fish. The Tshal pa Kanjur reading contrasts most with the other two
versions, not even sharing the /thugs rjes element, and omitting the final bkra shis so, which would be expected.) | 172 :

| 173 : | 174

177 inserts

182

unclear;

|
|

178 insert

183 :

| 186

179

| 175 :
:

184 :

180

| 176 : ;

; |

181 :

185 :

: omit (Here, the Tshal pa Kanjur's omission appears to be in

error.) | 187 : | 188 inserts | 189 : | 190 insert | 191 inserts | 192

insert

(Here, 's reading would seem clearer and more appropriate.) | 193 inserts

| 195 insert

| 196

| 197 :

| 198 inserts

| 194

| 199 :

222

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

200201

202 203 204 [ 81v] 205206

207208

209

210

211 212 213214 215216 217218

219

220

221

222

229 [312b]
223 224 225226 227228
230 231 232233 234

235 236 237238

239 240

241 242243 244 245 246 [ 82r] 247

248249 250
251

200 : | 201 inserts | 202

206 :

| 209 :

216 insert

| 213

| 217

207 :

| 210 :

: omits;

| 203 : | 204

unclear | 208 :

| 211 :

insert | 205 :

| 212

: ;

| 214 inserts | 215

| 218 :

omits

(Here, 's

reading would seem more appropriate. The head would fit with the body associations given in Chapter 7 of the commentary for
Vairocana and his retinue.)
221

232

219

unclear

omits | 240

| 236 :

| 227

| 234

| 233 insert

| 237

| 241

| 243 inserts

220

| 244

: ;

| 228 :

| 238 : omit | 239

| 245

231

| 229

| 235 :

| 242

: ;

| 222 : | 223 : | 224 : | 225

| 226 insert
230

omit (Here, the Tshal pa Kanjur's omission appears to be in error.) | 246

: ; (g.yas

| 248

is surely correct, or at least fits with the associations given in Chapter 7 of TZComm, as well as the line above.) | 247 :
inserts | 249 : | 250

| 251 :

223

Chapter 41

252 253254255

256 257 258

259 260 [ 319]

261 262

263 264 265

266 267 268269

270 271

272 273 274 275 276

277 278 279 280

281 [298r] 282 [ 82v] 283 284 285

286 287 288 289 290

291 292 293

294 295 296 [307r] 297

298299

300 301 302

303 304

_____________________________________________________________
Note that the Dunhuang manuscript and the Tshal pa Kanjur version recognise no section break, and continue
seamlessly into the text of the next chapter (/de skad ces), while the Bhutanese version rejoins the text with
the new chapter beginning.
______________________________________________________________
252 : | 253

: omits | 254

258 : | 259 inserts | 260

omits | 264

276

267 :

insert

288

268

| 272 :

| 279 :

| 283 :

: ;

| 271 :

inserts | 255 : | 256 : omits | 257 inserts

| 280

261 :

265

262 inserts | 263 :

: ;

: |

269 : omit | 270

266 :

| 273 insert | 274 inserts

; inserts | 275 :

281 :

| 277 :

| 284 : | 285

: |

| 278 :

289 :

290

; ; ;

282

| 286 :

287

291

inserts | 292 insert | 293 : | 294 : | 295 : | 296 inserts

297 :

298 :

na ro is surely accidental) | 302 : | 303

299 :

300

: omits |

301

(but the

: | 304 : ;

224

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

[ 319.4] [421.2] [335v.3] [213v.7] [44v.8] [380v.1]

305 306 307308 [45r]

309

305

311 312313314

310

: | 306 : | 307 :

310

314 inserts

: illegible |

311

; (though unclear) | 308 : | 309 :

; |

312 :

; |

313 insert

Chapter 42
[ 82v.4] [307r.1] [312b.7] [298r.2] [ 319.5] [64v.2] [58v.2] [421.2]
[335v.3] [213v.8] [45r.1] [380v.2]

2 3

4 5 6 7
89

1011 12

131415 [214r]

[ 320] 16 17
18 19

[ 83r] 20 21

22

23 24 2526

27

28

29 30 31
32

33

34 35

3637
1 inserts

);

| 2 : ; | 3 :
: omit; unclear | 5

| 4

; omit | 7

;
;

| 12 :

| 22 :

30

| 26 :

13 :

| 20 : [(/ ?)];

| 6 :

| 23 :

| 27

| 19

is given also in );

| 36 : ; ;

| 34 :

| 35

| 29 inserts

; ;

| 33

15 :

| 25

: ;
31

) | 11 :

| 21 :

| 32 :

; ; omits;

| 24 :

| 28

14 :

| 18 insert

| 10

( partially illegible); ( also gives

: ;

| 17 :

: ;

: omit | 9 :

; ; |

16 inserts

| 8

; (but give

( gives , but

| 37 :

226

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

38 [313a] 39 40 4142

43 44

45

47 [336r] 48
46
49 50 51

52 53

54

55 56 5758

59

60

[ 83v] 61 [65r] 62 [422] 6364

65 66

67 68

[59r] 70

69

71 72 73

74

7576

[381r] 77
38 inserts

41 :

39

| 42 :

| 43 :

40

| 44

: ;

; ; (but gives ); ;

('gyur, shared by

, might have been present in the archetype, and corrected

independently in the branches leading to , but it is hard to be sure, given that the more appropriate

reading, 'gyu, is given in ). | 45 : ; | 46 :

47

54

;
|

| 51

;
58

| 55

48 :

; |

| 52

: ;

: ; ;

49 inserts

50

: ;

| 53 :

| 56 :

| 57

seems to give a verse of six rather than four tshig rkang; quite possibly, the

extra lines were intended as a commentarial elaboration, which mistook for the main text and then dropped the original verse

lines) | 59 : | 60 : | 61 inserts | 62
: | 63 :

| 64 : | 65 : ; ;

| 68

74

| 66 :

| 69 :

| 70

| 67 :

| 76 : | 77 :

| 71 : omit | 72 :

[?] (yang seems to have been amended to 'ang); |

75

; |

73 :

227

Chapter 42

[45v]

78 79 80


81 82

84

83

[ 321] 85 86 i

87 88

8990 919293 94 ii
_______________________________________________________________
Note that the Dunhuang manuscript and the Tshal pa Kanjur version omit any chapter title. Besides the
disagreement over the chapter numbering due to the omission of the Chapter 41 title in the Dunhuang, Tshal
pa Kanjur and Bhutanese versions, there are three suggested titles for the final chapter. Here, we are
privileging the title shared between the South Central, Bathang, Hemis and Tawang editions, which was most
likely present in the archetype, and we are giving the numbering which they share with the Tenjur, which was
surely the earlier numbering. Note, however, that the Bhutanese title would seem very apt, given the chapter
content. The Tenjur title would seem rather less likely, since it appears to pick up rather on the text title,
suggesting a scribal confusion. However, each version including the omission of a final chapter title for
what could be considered a postscript rather than a separate chapter could be seen as having its own
rationale and tradition.
_________________________________________________________________
[ 321.4] [65r.4] [59r.3] [422.3] [336r.5] [214r.6] [45v.2] [381r.3]
95

9697 9899

_______________________________________________________________

i Note that Klong chen pa cites the three lines from /mnyam nyid... up to here in his Phyogs bcu mun sel; the citation is precisely as

given in our edition, except for agreeing with the variant, khro bas, rather than khro bos (bdud 'joms bka' ma Volume La: 488489).
ii This line is cited exactly in Chapter 22 of Klong chen pa's Phyogs bcu'i mun sel, (thabs kyi zhags pa las/ nyid kyi thugs kar zhugs
nas mi snang bar gyur to/ /zhes dkyil 'khor bsdu bas mi snang bar gsungs shing/, bDud 'joms bKa' ma Volume La: 618-619). See
also G. Dorje 1988: 1295- 6.
78 inserts | 79

82

84

; |

96

| 88

94 :

83 :

95 insert

; ; ; ;

86

| 89

: omit |

92 :

; |

| 85

91 :

98
omit

: |

87 :

: | 80 : | 81 : ; ; ; ;

; ; |

| 99

90 :

93 :

| 97 insert

228

Critical Edition of the Root Text of the 'Phags pa Thabs kyi zhags pa padma 'phreng gi don bsdus pa

Note that there is also some disagreement over the main text title. Moreover, the South Central, the Bathang,
Hemis and Tawang versions, along with the Dunhuang manuscript and the Bhutanese NGB, are all agreed
that rtog pa'i rgyal po['i don] bsdus pa should be part of the main text title (and note that the Tenjur also
includes this phrase above, perhaps given in error in the chapter title),100 although rtog pa'i rgyal po is not
included in the title given at the start of the text.
_________________________________________________________________
[ 84v.4] [313a.5] [298r.7] [ 321.5] [65r.4] [59r.3] [422.4] [336r.4] [214r.6]

[45v.3] [381r.3]

101 102

103104105106 107108 109 110

111112

100 It is just conceivable that this phrase might once have been part of the chapter title, and preserved as such only in the Tenjur

version, but given the agreement between the South Central, the Bathang, Hemis and Tawang versions, as well as the Dunhuang
manuscript and the Bhutanese NGB, this would seem most unlikely. | 101 : | 102 :
omit;

(in tantric lore, there is often

reference to a text as an excerpt from a vast work no longer extant. Here, the specific number in this reference to the tantra being
extracted from a sixteen thousand chapter text may have some connection with the symbolism of the tradition of a sixteen
thousand line Myjla Tantra, mentioned in the Majurnmasagti, which is supposed to be extracted from it. See Davidson
1982: 2, 44, and see also David Gray 2009: note 23, citing Claudio Cicuzza 2001, The Laghutantrak by Vajrapi: A Critical
Edition of the Sanskrit Text, Serie Orientale Roma 86, Roma, Istituto Italiano Per l'Africa e l'Oriente: 49, 123, 126, 127).
103

: ; ;

illegible | 105 :

110

| 107

omit;

| 108 :

omit;

104 :

| 109

; ;

| 106 :

omit

(here,

are of course simply giving the TZComm title addition [although omitting the word, rgyud]; we include the full readings

here to enable a better comparison between the versions.) | 111 : ; | 112 The
Bhutanese and the South Central versions give a final colophon. 65r.5

59r.4

422.5

336r.4.

THE

EDITION OF THE COMMENTARY,

'PHAGS PA THABS KYI ZHAGS PA PAD MA 'PHRENG GI DON BSDUS PA'I 'GREL PA

[Note that bolded text in the edition indicates the root text as understood by our TZ critical edition;
highlighted text indicates text which has been highlighted in the Dunhuang manuscript. Pagination and line
numbers in the Dunhuang manuscript are given in square brackets; this also applies to the pagination of other
versions, preceded by their sigla. Text in small italics indicates marginal annotations given in small writing in
the Dunhuang manuscript. Within the apparatus, a bolded reading indicates a meaningful alternative
reading.]
Chapter 1
[Ms1r.1] [Qt101a.1] [Gt244.1] [Nt176.5]
1

$ / /'phags pa thabs kyi zhags pa zhes bya ba/2

/pad ma3 'phreng gi don bsdus pa'i 'grel pa'4/

[beneath the line] rdo rje sems dpa'I dngos grub thob par bya ba dang/ bsam ba la bgegs myi 'jug cing bsam ba mthar phyin par bya
ba'i don/ 'bu tas bsdus sam ba bha byas

/dpal rdo rje sems dpa' [2] la phyag 'tshal lo/5

[beneath line 2] mos pa dang rtogs pa 'o/

/theg pa chen po la yang6 rnam pa gnyis te/

mtshan nyid kyi theg pa dang/
rdo rje 'i theg pa 'o/
/7mtshan nyid kyi [3] theg pa ni/
/mdo sde 8 la stsogs9 pa'i ***** tshul du10/
[beneath the line] mdo sde 'i gzhung gis

byang chub sems dpa'i spyod ***** pa spyad pas/11

bskal pa 12 gsum kyis13 [4] 14 gnon pa ste/
[beneath the line, continuing onto line 4, the continuity indicated by the crosses at the end of line 3 and beginning of line 4] mos pa
spyod pa'i sa nas sa dang por + +por bskal pa gcig dang po nas brgyad du gcig brgyad nas kun du 'od du gcig

/yun ring po nas pha rol du15 phyin pa bcu la stsogs16 pa/
dka' ba sna tshogs 17 spyad18 pas 'grub19 ces 'byung/
[beneath 'grub ces 'byung] sangs rgyas 'grub ces mdo sde las 'byung20

/rdo rje 'i theg [5] pa ni gsang sngags kyi sgo ra21 nas 'jug pa ste/
/rdo rje lta bu rnam pa gsum gyi mthu rnyed nas/
tshe 'di nyid22 la 'grub ces 'byung 23 [Nt177] ste/
[beneath the line, continuing at the bottom of the sheet, in a further row; the continuity is indicated by the crosses at the end of the
row underneath line 5 and at the beginning of the second row for the annotation] $// shes pa'i mtshan nyid gyis rtogs na 'jug pa'i
mtshan nyid gyis + + goms pa'i mthus 'bras bu 'i mtshan nyid sku gsung thugs su 'grub bo

1 Qt,

Gt (243) and Nt insert a title: /thabs kyi zhags pa pad mo'i [Gt: ma'i] 'phreng ba'i 'grel pa bzhugs [Gt: bzhugso]// | 2 zhes bya
ba/: QtGtNt omit | 3 pad ma: QtGtNt padmo | 4 pa': QtGtNt pa | 5 Note that in our TZ edition, this homage follows the title, while
the Tshal pa Kanjur version of TZ gives it following their initial title but before the main text title, perhaps seeking to reproduce
the order found in TZComm from which it was extracted. In TZComm, the commentarial discussion of the root text title is below,
and the homage is not repeated in the lemmata beneath it. However, there is no indication that the Dunhuang manuscript is
treating the homage here as part of the root text at all. | 6 yang: QtGtNt don | 7 QtGtNt insert: de la | 8 QtGtNt insert: dag | 9 stsogs:
QtGtNt sogs | 10 du: QtGtNt gyis | 11 /: QtGtNt omit | 12 QtGtNt insert: grangs med pa | 13 kyis: QtGtNt gyis | 14 QtGtNt insert: sa
(Nt as a correction); this is most probably an omission in the Dunhuang manuscript. | 15 du: QtGtNt tu | 16 stsogs: QtGtNt sogs
17 QtGtNt insert: pa | 18 spyad: QtGtNt dpyad | 19 'grub: QtGtNt grub | 20 'byung: partially smudged out, but not in doubt | 21 ra:
there is a small line above indicating deletion; QtGtNt omit | 22 nyid: QtGtNt omit | 23 QtGtNt insert: ba

230

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

[above 1v line 1, starting from the first "chen po"] $ /chen po ni rgya chen po dang 'bras bu chen po dang po/ [nams?]24 dag rnam
gsum gyi tshul gyis sku gsung thugs la bya /

de ni theg [1v.1] pa chen po yang theg pa chen po25 zhes bya 'o/
[beneath the line, starting from the second "chen po"] ye shes chen po dang dmyigs pa chen po

/'di 'i26 mgo mjug du27 'jig rten dang 'jig rten las 'das pa'i chos thams cad dkyil 'khor [2] gsum du 'dus shing/

[beneath the line, starting from "dkyil 'khor"] rang bzhin dang [gzugs]28 29 [2] brnyan dang yid /

zhing dang spyod yul30 yongs su dag pas 31 yon tan gyi khyad 32 gong na myed33 de/

[beneath the line] dkyil 'khor gsum du 'dus pa nyid kyis spyod yul yongs su dag par gzhan las 'phags par sbyar/

/bcom ldan 'das kyis gsungs [3] pa'i34 nang na yang /35
[beneath the line] 'di ltar rtogs pa las

bkri ba'i don ma yin * gyi/

[beneath the line] dka' thub dang tshul 'khrims

nges pa'i don gtan la bab par bsta ***** n pa'i sde dang /36
[beneath the line] ma ha yo ga 'i rgyud

rmad du byung ba'i sde [4] las btus te bstan to/

[beneath the line, continuing onto line 4, the continuity indicated by crosses] sangs rgyas dang byang chub sems dpa' + [4] + thams
cad kyis shin du sngags

/ 37 'phag ***** s pa thabs kyi zhags pa38 pad ma39 'phreng **** gyi40 don bsdus pa /41
[beneath the line, starting under thabs] bsdus pa ni gsung rabs sde bcu gnyis dang tan tra sde bco brgyad las bsdus pa 'o

zhes bya ba 42 de la [5]

[beneath the line] gleng gzhi stan te

'phags pa ni/
nyan thos dag kyang 'jig rten las 43 'phags mod kyi/44
[beneath the line, under 'phags mod kyi] dka' thub dang tshul khrims kyis

skabs 'dir de la myi45 bya yi46/

[beneath the line] rgyud 'di skabs 'dir

[Gt245] theg pa chen po47 yang chen po

[beneath the line, continuing onto line 6, the continuity indicated by crosses] mdo sde kri ya u pa ya yo ga dang/ theg pa chen po +

[6] nyan thos dang yun ring po 'i48 grub49 pa las 'phags pa ste/
[beneath the line] + yin gyi 'di ni bdag las kyang bsgrub pa'i thabs kyis 'phags te ma ha yo ga 'I don la zhugs pa nI

/tshe 'di nyid las50 'grub51 pa dang /

[beneath the line] ma ha yo ga 'I don la zhugs pa ni i

sangs rgyas bcom ldan 'das [2r.1] $ / /chen po la bya 'o/

/thabs kyi zhags pa zhes bya ba ni /52
[beneath the line] nga 'dzin pas

thabs ni snying rje chen po ste/

mtshams myed53 pa lnga la stsogs54 pa [2] la rol cing/
[beneath the line] myi dge ba bcu dag bsdu [beneath the line, starting under la stsogs] so so ci ltar ['dul?] ba bzhin

i line 6 ma ha yo ga 'I don la zhugs pa ni this repeated annotation appears to be a scribal dittography, suggesting that the
annotations are a copy (see also f.4r, and the discussion above, p.32-33).
24 nams?: rnam pa or rnam intended? | 25 theg pa chen po yang theg pa chen po: QtGtNt theg pa chen po'i yang chen po | 26 'di 'i:
QtGtNt 'di | 27 du: QtGtNt tu/ | 28 gzugs: letter za unclearly written | 29 continuity between the comment beneath line 1 and at the
beginning of line 2 is here indicated by a kind of ornamental asterisk in both places, consisting of a triangle of three small circles
30 spyod yul: QtGtNt dpyod yul | 31 QtGtNt insert: / | 32 QtGtNt insert: par | 33 myed: QtGtNt med | 34 gsungs pa'i: QtGtNt gsungs
pa thams cad kyi | 35 /: QtGtNt omit | 36 gtan la bab par bstan pa'i sde dang /: QtGtNt gtan la phab pa dang | 37 QtGtNt insert: de
la | 38 QtGtNt insert: / | 39 pad ma: QtGtNt padmo | 40 gyi: QtGtNt gi (the Dunhuang manuscript is in error here: the correct
spelling is given in the title above, f.1r) | 41 /: QtGtNt omit | 42 QtGtNt insert: ni/ | 43 Gt insert: 'das | 44 /: QtGtNt omit | 45 myi:
QtGtNt mi | 46 bya yi: QtGtNt bya'o (this reading would seem clearer here) | 47 po: QtGtNt po'i | 48 po'i: QtGtNt po nas | 49 grub:
QtGtNt 'grub | 50 las: QtGtNt la | 51 'grub: QtGtNt grub | 52 /: QtGtNt omit | 53 myed: QtGtNt med | 54 stsogs: QtGtNt sogs

Chapter 1

231

phyin ci log la zhen pa'i [Qt101b] sems can rnams sgrol ba 'o55/
[beneath the line] rdo rje can du rtogs par bya ba'i phyir ram dngos po myed do
[beneath the line, starting under rnams sgrol and continuing under the next words] dngos po myed par rtogs par bya ba'i phyir shes
rab kyi ngo bo myed par chud pas 'khor bar myi [gtong?]56 ba 'o

/zhags pa57 ni myi58 gtong ba ste/

'khor ba nyid mnya pa'i [3] chos su mkhyen nas59/
'khor bar60 myi61 **** gtong bar sbyor ba la bya 'o62/
[beneath the line, under myi gtong] phyin ci log du [beneath the line, under bya 'o] chos nyid ['du(/ldu)?]

/'di ni sbyor ***** ba rnam gnyis zhes bya/ 'o63

[beneath the line] thabs dang zhags pa'i don

/che mchog [4] 64 gnyis shes65 kyang 'byung/

[beneath the line, continuing onto line 4, the continuity indicated by crosses] theg pa 'og ma + + dag la thabs de lta bu myed thabs
dang shes rab mkhyen bar can

pad ma66 'phreng67 **** zhes bya ba la/

pad ma68 ni shes rab kyi rta ***** gs te69/
chos thams cad la spyod70 [5] kyang71 bag chags su myi72 'gyur ba 'o73/
[beneath the line, under chags su myi] ye nas sangs rgyas pas

'phreng74 zhes bya ba75 ni76 thabs kyi rtags te/

[beneath the line, starting with thabs kyi, and continuing under the next tshig rkang] $//'di man cad gleng gzhi ston te 'phrin las rnam
bzhi las stsogs pa

sprul pa'i 'phreng77 ba rgyun myi chad78 pa la bya 'o/ /

[6] mdor79 na thabs dang shes rab zung du80 'brel pas/
'khor ba dang mya ngan las81 'das pa la myi82 gnas pa'i don to/
[beneath the line, linked to 'khor ba] 'jig rten pa [beneath the line, linked to 'das pa] nyan thos pa
[beneath the line, linked to myi gnas pa'i] 'di gnyis ka la

/don bsdus pa ni83 [2v.1] gsung rab bcu gnyis dang/

[beneath 2v.1] yi ge 'grel [pa (/ba)] rabs gcig las 'di la bam po bco brgyad yod pa las bsdus pa zhes 'byung ngo

tan tra sde bco brgyad la stsogs84 pa la85 bsdus pa 'o86/

/de nas bcom ldan 'das dpal kun rig87 rnam par [2] snang mdzad la/
[beneath the line] gong ma'I 'phro yin pas de nas zhe na'o / de 'i 'grel pa ni gsal bar byas pa 'o / gleng gzhi ston pa
[beneath line 2] ston pa phun sum tshogs pa [stsogs?] [-] stan pa

bcom ldan 'das dpal88 rdo rje sems dpas89 /90

[beneath the line, under rdo rje sems dpas] sdud pa phun sum tshogs pa

rig pa'i byin rlabs91 las byung ba'i mchod pa92 /

[beneath the line] ma 'ongs pa'i sems can slobs dpon dang bla ma la bkur sti cher byed du zhugs pa'i don no/

55 ba

'o: QtGtNt ba'o | 56 gtong: final nga unclearly written | 57 The wash over zhags pa in line 2 is probably intended but not
certain perhaps it has faded; similarly the second pad ma on line 4. | 58 myi: QtGtNt mi | 59 nas: QtGtNt pas | 60 bar: deletion of
the final ra indicated by a line above; Gt par; QtNt bar | 61 myi: QtGtNt mi | 62 bya 'o: QtGtNt bya'o | 63 bya/ 'o: 'o is written
slightly lighter, and a little smudged; it could have been inserted afterwards, or alternatively, the shad might have been added and
the 'o deleted; QtGtNt kyang bya | 64 QtGtNt insert: rnam | 65 shes: QtGt zhes | 66 pad ma: QtGtNt pad mo | 67 'phreng: QtGtNt
omit | 68 pad ma: Qt padma | 69 te: Nt to | 70 spyod: QtGtNt dpyod | 71 QtGtNt insert: / | 72 myi: QtGtNt mi | 73 ba 'o: QtGtNt ba'o
74 'phreng: QtGtNt phreng | 75 Our TZ edition gives zhes bya ba'i rgyud, but the versions of the root text extracted from the
commentary do not. The manuscript's highlighting possibly indicates agreement with its inclusion, but the placement does not
agree: it should follow don bsdus pa. Here, TZComm's wording might rather suggest that the zhes bya ba is intended merely as a
quotation mark for the word, 'phreng. Note that the Tenjur tradition also includes zhes bya ba following don bsdus pa, but that the
Dunhuang manuscript does not. | 76 QtGtNt insert: / | 77 'phreng: QtGtNt phreng | 78 myi chad: QtGtNt mi 'chad | 79 mdor: QtGtNt
dper | 80 du: QtGtNt omit | 81 las: QtGtNt omit | 82 myi: QtGtNt mi | 83 ni: QtGtNt zhes bya ba ni/ | 84 stsogs: QtGtNt sogs | 85 la:
QtGtNt las | 86 pa 'o: QtGtNt pa'o | 87 rig: QtGt rigs; Nt rig, with a following letter deleted | 88 dpal: QtGtNt omit (it is omitted in
our TZ edition, but found in the Tshal pa Kanjur and Bhutanese texts) | 89 dpas: QtGtNt dpa' | 90 /: QtGtNt omit (as do the Tshal pa
Kanjur texts) | 91 rlabs: QtGtNt brlabs | 92 pa: QtGtNt pas/

232

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

rgya yongs su ma chad [3] pa'i [Gt246] gzhal yas [Nt178] khang /
[beneath the line, continuing onto line 3, the continuity indicated by crosses] de bzhin gshegs pa la ni mchod pa'i sprin dang + +
rang bzhin gyis ldan bas gud nas mchod pa [g(/k)]yang myi dgos te

bla re 93 gdug **** s dang /

rgyal mtshan dang/
ba dan dang/94 *****
sgeg mo la stsogs95 pa phyi nang gyi96 [4] mchod pa brgyad kyis mchod de/97 *****
[beneath the line] 'khor phun su tshogs pa
[beneath the line, under phyi nang gyi] nang gyi bzhi dang phyi 'i bzhi
[beneath line 4] $// de bzhin gshegs pa dga' ba'i yul bzhi sku gsung thugs dang ldan pa

/lan gsum bskor98 ba byas te99/

[beneath the line] lan gsum gsol ba btab pa'i don

so so **** ma yin ba'i100 rig pa'i101 'dun pas102 [5] phyag 'tshal nas103 spyan sngar 'dug ste/
[beneath the line, continuing onto line 5, the continuity indicated by crosses] kun rig dang rdo rje sems dpa' ni sku gsung thugs kyi
don tshul te so so ma yin ba'i phyir te phyi ma 'i bsgrub pa po + +rnams kyi ched du /

'di skad ces104 gsol to/

zhes bya ba ni /105
bcom ldan 'das106 dpal rnam par snang mdzad [6] la107 rdo rje sems dpas don bsdu bar108
[beneath the line] gsol ba btab pa phun sum tshogs pas stan

gsol ba'i phyir /

rang gyi109 byin rlabs110 las byung ba'i mchod pa'i sprin chen pos mchod de/
[3r.1] $/ /'og nas 'byung ba111 skad gsol ba btab pa'i don to/
/bcom ldan 'das kyis112 dus gsum du gshegs pa 'i113 sangs rgyas 114 [2] kyis mkhyen pa'i115/
[beneath the line] gleng bsla[ng?] [beneath the line, starting from gsum] gsol ba btab pa dang phun sum tshogs pa bstan
[beneath line 2, mkhyen pa'i] thabs te yab
ii sangs rgyas bcom ldan 'das rnams116 kyi spyod yul117/
[beneath the line, linked to spyod yul] shes rab ste yu

nyan thos dang thun mong ma lags pa thabs gang gyis118 [3] rtogs pa 119 /
[beneath the line] mdo sde dang nyan thos rang gyas don 'di ltar ma rtogs pas
[beneath the line, linked to thabs gang gyis] tha[b]s dang shes rab skye b[o?] skal ba can la

tshe 'di nyid la thams **** cad mkhyen pa'i go 'phang thob pa 'di 'i120 [Qt102a]**** don 121[b]shad122
du gsol/
[beneath the line] skad cig ma nyid la

ii

See TZ edition, Chapter 1 note ii, on the citation of this passage by Rong zom chos kyi bzang po.

93 QtGtNt

insert: dang/ | 94 ba dan dang/: QtGtNt omit | 95 la stsogs: QtNt la sogs; Gt sogs | 96 gyi: QtGtNt gi | 97 Note that the
Tshal pa Kanjur and Bhutanese texts have a rather different version of this passage: gzhal yas khang dang/ bla bre dang/ rgyal
mtshan dang/ ba dan dang/ gdugs dang/ sgeg mo la sogs pa nang gi mchod pa bzhi dang / bdug spos la sogs pa'i phyi'i
mchod pa bzhis mchod de/ | 98 bskor: QtGtNt skor | 99 te: QtGtNt nas | 100 ma yin ba'i: QtGtNt omit | 101 pa'i: QtGtNt pa
102 'dun pas: QtGtNt 'dun pa'i (the Tshal pa Kanjur and Bhutanese texts give 'dud pas; see comment p.55 above) | 103 QtGtNt
insert: / | 104 ces: QtGtNt omit | 105 /: QtGtNt omit | 106 QtGtNt insert: / | 107 QtGtNt insert: / | 108 bsdu bar: QtGt bsdus par; Nt
bsdus bar | 109 rang gyi: QtGtNt rig pa'i | 110 byin rlabs: QtGtNt byin gyis brlabs | 111 'byung ba: QtGtNt byung ba'i 'di | 112 kyis:
QtGtNt omit | 113 pa 'i: QtGtNt pa'i | 114 QtGtNt insert: bcom ldan 'das (so too do the various TZ editions) | 115 pa'i: QtGtNt pa
116 rnams: Gt final sa subscribed, almost illegible | 117 spyod yul: QtGtNt dpyod yul | 118 gyis: QtGtNt gis | 119 QtGtNt insert: de
(the Tshal pa Kanjur and Bhutanese texts also give de) | 120 pa 'di 'i: QtGtNt par 'gyur ba'i | 121 QtGtNt insert: brlabs te
122 [b]shad: prefixed ba is probably intended, but it appears to have been an amendment by the original scribe of a previously
written letter, perhaps ma; QtGt gshad (Nt has bshad)

Chapter 1

233

zhes lan [4] gsum du gsol ba btab bo123 zhes **** bya ba'i tha tshig go /
[beneath the line, continuing onto line 5 (not the following line 4),iii the continuity indicated by crosses] rgya chen pos mgo smos pa
dang rgyud thams cad kyi +

/de skad ces124 gso **** l ba btab pa dang/

[beneath the line] don gsum pa iv gleng slang ba 'chang

cang myi125 gsung [5] bar 'dzum ba mdzad nas126 'di skad bka' stsal to127/
[beneath the line] +don ni zhags pa'i don 'dir 'dus pa 'o
[beneath the line, linked to bka' stsal to] gleng bslang ba ston pas gleng slang ba

zhes bya ba ni128 rdo rje sems dpa's129 don bsdu bar gsol ba gnang nas/
da ltar [6] dang ma 'ongs pa'i sems can thams cad la phan pa'i phyir/
don bsdu [Gt247] bar gsol ba 'di lags so130 snyam nas dgyes te131 'dzum ba132 [3v.1] mdzad nas /133
'og134 nas 'byung ba skad bka' stsal ces bya ba'i tha tshig go/
/sems dpa' chen po nyon cig /
[beneath the line] don bzhi pa ci ltar yongs su shes par bya ba bstan pa ni bden ba gnyis kyi le'u ston

don 'di dang135 phyi nang gyi136 [2] chos thams cad 'dra ste/
[beneath the line] tha[b]s137 kyi zhags pa 'di dang

bden ba138 gnyis su 'dus so /

bden ba139 gnyis la yang tshul 'di ltar rig par bya 'o140 /
zhes bya ba ni141 go [3] bar bya ba'i phyir /
sems dpa' chen **** po zhes142 rdo rje sems dpa' la bos pa **** 'o143 /
'di 'i144 don dang phyi nang gyi145 don146 thas [4] cad 'dra ste zhes bya ba ni /
thabs kyi *** zhags pa pad ma147 'phreng gyi148 don dang /
[beneath the line] 'di nyid bden pa gnyis dang thabs dang shes rab zung du 'brel pa dang spyir ro

phyi *** rol149 snod kyi 'jig rten dang/

bcud [5] kyi 'jig rten gyi chos thams cad 'dra 'o150 zhes bya ba'i don to/
/ci 'i151 phyir 152 zhe na/
bden ba gnyis su 'dus shes bya ba153 smos te/
/ [6] snod bcud kyi 'jig rten gyi chos dang /
thabs kyi zhags pa pad ma154 'phreng gyi don dang 'dra ste/
[beneath the line] thams cad 'dra 'o zhes bya ba'I don

/155 thams [Nt179] cad bden ba156 gnyis su 'dus [4r.1] $ / /shes bya ba'i157 tha tshig go/
/de la bden pa gnyis 158 gang /
iii

Note the annotation continues having omitted a line this could be possible further evidence of the annotations being a copy,
where the copyist carelessly missed out a line, perhaps a little less likely if a writer were composing and paying attention to the
content of the main text. On 5r, there is a similar omission of a line, the annotation continuing from lines 3 to 5, but in this case,
the second part of the annotation may simply be attempting to avoid confusion with a different annotation on line 4.
iv don gsum pa: the third point. This series of points is developed through the first few chapters, but the first and second points
appear to have been omitted in error by the copyist. The fourth point is found below, while a sixth, seventh, and eighth point are
found in Chapters 4, 5 and 7 respectively. The fifth point also appears to have been omitted. The omissions suggest that the
annotations are most likely to have been copied from a previous version of the commentary, along with the main text; see the
discussion p.32-33 above.
123 QtGtNt

insert: / | 124 ces: QtGtNt omit | 125 cang myi: QtGtNt ci yang mi | 126 QtGtNt insert: / | 127 to: QtGtNt pa | 128 QtGtNt
insert: / |
QtGtNt dpa' | 130 lags so: QtGtNt yin no | 131 QtGtNt insert: / | 132 ba: QtGtNt pa | 133 /: QtGtNt omit | 134 'og:
QtGtNt 'og ma | 135 There is a mark above the tsheg following dang, and dang is bracketted by a semi-circular line above. It
might possibly be intended to indicate deletion, but dang is given in all other editions. | 136 gyi: QtGtNt gi | 137 thas: presumably,
thabs intended | 138 ba: QtGtNt pa | 139 ba: QtGtNt pa | 140 bya 'o: QtGtNt bya'o | 141 QtGtNt insert: / | 142 QtGtNt insert: / | 143 pa
'o: QtGtNt pa'o | 144 'di 'i: QtGtNt 'di'i | 145 gyi: QtGtNt gi | 146 don: QtGtNt chos (MGGr, who include some of this passage, also
give chos, which is of course given in the root text citation above) | 147 pad ma: QtGtNt padmo | 148 gyi: QtGt gi | 149 phyi rol:
QtGtNt phyi | 150 'o: QtGtNt bo | 151 ci 'i: QtGtNt ci'i | 152 QtGtNt insert: 'dra | 153 bden ba gnyis su 'dus shes bya ba: QtGtNt zhes
bya ba | 154 pad ma: QtGtNt padma | 155 QtGtNt insert: chos | 156 ba: QtGtNt pa | 157 shes bya ba'i: QtGtNt pa'i | 158 QtGtNt insert:
ni
129 dpa's:

234

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

tshul ci ltar rtogs159 par bya zhe na /

bden pa gnyis la tshul 'di ltar [2] rig par bya zhes smos te /
bden ba160 gnyis ni don dam pa'i bden pa dang /
kun rdzob kyi bden pa 'o161 /
de lta yang nyan thos 162 ltar ma yin [3] gyi163/
[beneath the line] de dag kyang bden pa gnyis sdu 'dod mod kyi

chos thams cad mtshan ma myed164 *** pa las byung bas /

[beneath the line, starting from mtshan, and continuing under the next tshig rkang] mtshan nyid myed pa dang dbyings la lha dang
lha mo la stsogs pa sgyu ma tsam du lha dang lha mo las stsogs pa sgyu ma tsam du v

don dam pa ni165 rgyu 'o166 / ***

/sgyu ma tsam [Gt248] du snang bas167 kun rdzob [4] ni 'bras bur rig par bya 'o168/
zhes bya *****ba'i tha tshig go/169
/de nyid bstan pa'i phyir *****
[beneath the line] gong du sgyu ma'I don bshad pa'i phyir ro

/bden pa170 gnyis rgyu dang 'bras bu 'i171 [5] tshul /

de bzhin nyid la gnas bcas na172 /
/phyi nang [Qt102b] snod bcud rnam par dag /
/byed spyod173 thams cad tshogs gnyis 'gyur zhes /174 vi
[6] gsungs te /
/snod kyi 'jig rten lta bu /
chen po lnga175 gcig gyi176 nang na yang lnga lnga177 yod de/
/178de 'i chas gzhigs na/179
/ chen po [4v.1] lnga yang ma grub ste180 rang bzhin myed181 pa dang /
182 bcud kyi 'jig rten lta bu yang183 sems dang sems las byung184 ba'i chos/185
snga ma ni 'gags186 [2] phyi ma ni ma 'ongs 187/
da ltar gyi188 ni myi189 gnas pas rang bzhin myed de190/
rang bzhin myed191 pa ni192 don dam pa'i bden ba193 ste rgyu 'o194/
195de las so so 'i196 [3] rtsal sgyu ma tsam du yod pa ni/
kun *** rdzob kyi bden ba197 ste 'bras bu 'o198 /
chos de dag **** gi de bzhin nyid ni don dam pa ste/
[4] rgyu la myi199 gnas pa'i tshul gyis gna ***** s na/
[beneath the line, linked to rgyu la myi] shes rab [beneath the line, linked to gnas] thabs

phyi nang snod200 kyi 'jig rten 'bras **** bu 'i201 chos thams cad202 la ci203 ltar [5] spyod204 kyang /205
bsgrib206 pa rnam gnyis su myi207 'gyur te/
[beneath the line] nyon mongs pa dang shes bya 'I sgrib pa

de ltar chos thams cad la spyad208 kyang /

lha dang lha mo la stsogs pa sgyu ma tsam du: the repetition in this annotation seems likely to be a dittography rather than
intended, perhaps further evidence of the annotations being a copy (see also a similar example, f.1v above).
vi See TZ edition, Chapter 1 note iii, on the citation of this passage by Rong zom chos kyi bzang po.
v

159 rtogs:

QtGtNt rtog | 160 ba: QtGtNt pa | 161 pa 'o: QtGtNt pa'o | 162 QtGtNt insert: kyi | 163 gyi: QtGtNt gyis | 164 myed: QtGtNt
med |
ni: QtGtNt pa'i | 166 rgyu 'o: QtGtNt rgyu'o | 167 QtGtNt insert: / | 168 bya 'o: QtGtNt bya'o | 169 zhes bya ba'i tha tshig
go/: QtGtNt omit | 170 pa: QtGtNt omit | 171 bu 'i: QtGtNt bu'i | 172 na: QtGtNt nas | 173 byed spyod: QtGtNt byed dpyod | 174 'gyur
zhes /: QtGtNt 'gyur/ zhes | 175 lnga: Qt la | 176 gyi: QtGtNt gi | 177 lnga lnga: QtGtNt lnga | 178 / /: Gt / | 179 the shad appears to be
an amendment from an earlier written letter, probably a final sa. | 180 QtGtNt insert: / | 181 myed: QtGtNt med | 182 QtGtNt insert:
nang | 183 QtGtNt insert: / | 184 byung: QtGtNt 'byung | 185 /: Qt omits | 186 'gags: QtGtNt 'das/ | 187 QtGtNt insert: pa'o/ | 188 gyi:
QtGtNt omit | 189 myi: QtGtNt mi | 190 myed de: QtGtNt med do | 191 myed: QtGtNt med | 192 QtGtNt insert: / | 193 ba: QtGtNt pa
194 rgyu 'o: QtGtNt rgyu'o | 195 QtGtNt insert: / | 196 so so 'i: QtGtNt so so'i | 197 ba: QtGtNt pa | 198 'bras bu 'o: QtGtNt 'bras bu'o
199 myi: QtGtNt mi | 200 snod: QtGtNt gi snod bcud | 201 bu 'i: QtGtNt 'bu'i | 202 thams cad: QtGtNt omit | 203 ci: QtGtNt ji
204 spyod: QtGtNt dpyad | 205 /: QtGtNt omit | 206 bsgrib: QtGtNt sgrib | 207 myi: QtGtNt mi | 208 spyad: QtGtNt dpyad
165 pa

Chapter 1

235

mtshan nyid myed209 pas210 bag chags su myi211 [6] 'gyur ba ni /212
ye shes kyi tshogs so/
bdag dang 'phags pa thams cad mnyam bar bya ba'i213 phyir214 spyod pa ni /215
[beneath the line, starting from mnyam] mnyam ba rnam bzhI 'I don rtogs pas na
[beneath the line, linked to spyod pa ni] sd[o(/e)]216 gsu gyis

bsod nams kyi tshogs [5r.1] $/ /217zhes bya ba'i don to218/
[beneath the line, linked to don to] gal lam

/bsod nams dang ye shes kyi tshogs ni/ [Gt249]

[beneath the line, and continuing under the next tshig rkang] mdo sde lugs kyis ni pha rol tu phyin [pa bcu?] mthar phyin par spyad
nas rdzogs shing grub pa['i?] mtha' bsod nams dang ye shes kyi tshogs su 'gyur zhes so

pha rol du219 phyin pa bcu spyad220 pas rdzogs shes221 'byung na/
/ [2] de dag rdzogs sam zhe na/
rdzogs par bstan pa'i phyir/
[beneath the line, and continuing under the next tshig rkang] gsang sngags kyi theg pa la pha rol tu phyin pa dang tshogs [g?]nyis
rims gyis bgod pa yin cig car du

/pha rol phyin bcu rdzogs pa ste/

/sa bcu rim kyis222 bkod pa las223/
[beneath the line] nyan thos sam mdo sde 'i gzhung [k(/g)]yis

[3] mnyam nyid rig pa 'i224 rdo rje 'is225 / *** [Nt180]
[beneath the line] mnyam nyid rig pa'i skyes bu des ye shes skad cig ma se gol

/sa rnams mnyam ba226 nyid du sbyor/ ****

[beneath the line] gtogs pa tsam gyis pha rol du phyin pa bcu dus gcig du rdzogs par sbyor ro

/zhes gsungs te/

bag chags su [4] myi227 'gyur bas /
[beneath the line, continuing onto line 5 (not the following line 4), the continuity indicated by crosses] $/ sems dang po dang rig pa
can la +

chos thams cad **** la spyod228 pa ni /

[beneath the line] bems po dang ril po thams cad la myi dmyigs par spyod pa ni pha rol tu phyin pa 'o

sbyin pa'i pha rol du229 phyin **** pa 'i230 'bras bu dang bcas pa 'o231/
[beneath the line] des bsod nams dang ye shes kyI tshogs ni sangs rgyas kyi 'bras buo 'o

/ [5] nyon mongs pa las232 rnam par byang ba myi233 gzhan bar234 rig nas/
[beneath the line] + spyod pa nyid thams cad de nyid myi dmyigs pa ni pha rol du phyin pa 'o bsod nams dang ye shes kyi tshogs te
sangs rgyas kyi 'bras bu 'o

bdag dang 'phags pa thams cad so so ma yin tha myi235 dad par rtogs236 [6] pa ni/
[beneath the line, linked to rtogs] de myi 'dul ba n[o(/i)]

tshul khrims kyi pha rol du237 phyin pa 'o238/

/'di lta bu 'i239 chos la bzod pa 'am/
[beneath the line, linked to bzod pa 'am] myi skrag myi sngang bar
[above the line, bracketed and linked to: 'khrug par] zhe sdang gyis

'khrug240 par [Qt103a] 'gyur yang bsgrib241 par myi242 'gyur [5v.1] ba dang/
[beneath the line] khro bo 'i tshul du ston gyis kyang [beneath the line, linked to bsgrib par myi 'gyur] mnya ba nyid kyi don rig na
243

dpe byad 'og nas 'byung ba ltar rig pa ni/

209 myed: QtGtNt med | 210 QtGtNt insert: / | 211 myi: QtGtNt mi | 212 /: QtGtNt omit | 213 bar bya ba'i: QtGtNt pa'i | 214 QtGtNt
insert: / | 215 /: QtGtNt omit | 216 sd[o(/e)]: there would seem a good possibility that sgo is intended; the vowel more closely
resembles a na ro, a small stroke would turn the apparent da into ga, and this reading would fit best. | 217 QtGtNt insert: su 'gyur
218 to: QtGtNt no | 219 du: QtGtNt tu | 220 spyad: QtGtNt dpyad | 221 shes: QtGtNt zhes | 222 kyis: QtGtNt gyis | 223 las: QtGtNt la
224 pa 'i: QtGtNt pa'i | 225 'is: QtGtNt yis | 226 ba: QtGtNt pa | 227 myi: QtGtNt mi | 228 spyod: QtGtNt dpyod | 229 du: QtGtNt tu
230 pa 'i: QtGtNt pa'i | 231 pa 'o: QtGtNt pa'o | 232 las: QtGtNt omit | 233 myi: QtGtNt mi | 234 gzhan bar: QtGtNt zhan par | 235 myi:
QtGtNt mi | 236 rtogs: QtGtNt rtog | 237 du: QtGtNt tu | 238 pa 'o: QtGtNt pa'o | 239 bu 'i: QtGtNt 'bu'i | 240 'khrug: QtGtNt 'khrul
241 yang bsgrib: QtGtNt kyang sgrib | 242 myi: QtGtNt mi | 243 QtGtNt insert: mtshan dang

236

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

bzod pa'i pha rol du244 phyin pa 'o245/

[beneath the line] bzod pa la myi skrag myi sngang la

/spyod lam rnam bzhir gnyid log kyang /

[beneath the line] gnyid log pa ni bsod nams kyi tshogs nyid ngo bo nyid ngo bo myed pa ni ye shes kyi tshogs so

bsod nams [2] dang ye shes gyi246 tshogs su rig pa ni/

'bad pa myed247 pa'i brtson 'grus kyi pha rol du248 phyin pa 'o249/
[beneath the line, linked to 'bad pa myed pa'i, and continuing in a second row] gtod bsgrub myi dgos gyi ye nas yin pa 'o250
[beneath the line, starting from brtson 'grus, and continuing in a second row still under this line] brtson 'grus la mang251 rnam pa
gnyis ste 'bad pa dang bcas pa dang 'bad pa [myed?] pa 'o/ /'dI 'i skabs su 'bad pa myed pa dang sbyar ro252

/dbang po lnga ni rigs lnga/

[beneath the line, continuing onto line 3, the continuity indicated by crosses] dmyig253 la rdo rje 'I rigs rna ba la bad ma'I rigs sna la
rin po che ['i?] rigs +

yul lnga ni [3] yum lngar/254

[beneath the line, continuing under the next tshig rkang] + lce las kyi rigs lus la d[e?] bzhin gshegs pa'i rigs de 'i yul *** lnga yang
dbang po lnga dang 'dra

rig nas/
sems g.yo **** s kyang255 chos kyi256 dbyings las ma257 'gyur par258 **** rig cing/
rdo rje lta bu 'i259 ting nge 'dzin [4] gyis/
[beneath the line, continuing onto line 4, the continuity indicated by crosses] 'joms pa [da(/ra)?]ng myi tshugs pa dang myi 'gyur ba
dang + [beneath line 4] + 'dus ma bya pa 'o
260

thams cad la thogs pa mye **** d261 par 'jug pa ni/

bsam gtan gyi pha ***** rol du262 phyin pa 'o263 /
'khor ba'i chos [5] la spyod kyang /
mya ngan [Gt250] las 'das pa dang /
so so ma yin bar264 rig pa ni /
nyan thos las khyad bar265 che bas /
[beneath the line] nyan thos kyi spong len pas

shes pa'i nang na rab266 ce 'am267 phul [6] zhes bya ste /
shes rab kyi pha rol du268 phyin pa 'o269 /
/kun nas nyon mongs pa'i chos thams cad270 rnam par byang ba'i chos su rig cing/ /
[6r.1] $ / / 'phags pa thams cad 271 mnyes pas /
bdag dang gzhan gyi don du 'gyur zhing grol ba ni /
thabs kyi pha rol du272 phyin pa 'o273/
/ [2] bdag dang 'phags pa thams cad so so ma yin pa dang /
zhing dang spyod274 yul thams cad yongs su dag par rig pa ni /
smon lam kyi275 pha rol [3] du276 phyin pa 'o277/
/'khor ba dang kun nas **** nyon mongs pa las
[beneath the line] gnyi ga mnyam ba rnam bzhi nang [-]278

/ mya ngan las 'das pa **** dang/

rnam par byang ba gud na myed279 par rig [4] ste/
244 du:

QtGtNt tu | 245 pa 'o: QtGtNt pa'o | 246 gyi: QtGtNt kyi | 247 myed: QtGtNt med | 248 du: QtGtNt tu | 249 pa 'o: QtGtNt pa'o
annotation is marked off to the right by a long vertical jagged line, avoiding confusion with the next comment; "yin pa 'o"
is given beneath the first part of the annotation. | 251 la mang: la yang? | 252 again, this annotation is similarly marked off to the
right by a long vertical jagged line; the second part of the comment, from "[myed?] pa" is given beneath the first part of the
annotation. | 253 dmyig: or possibly, /myig. In any case, myig (eye) would seem to be intended. | 254 /: QtGtNt omit | 255 QtGtNt
insert: / | 256 QtGtNt omit: kyi | 257 ma: QtGtNt mi | 258 par: QtGtNt bar | 259 bu 'i: QtGtNt bu'i | 260 QtGtNt insert: chos | 261 myed:
QtGtNt med | 262 du: QtGtNt tu | 263 pa 'o: QtGtNt pa'o | 264 bar: Nt par | 265 bar: QtGt par | 266 rab: QtGtNt khyad par
267 QtGtNt che 'am/ | 268 du: QtGtNt tu | 269 pa 'o: QtGtNt pa'o | 270 QtGtNt insert: / | 271 QtGt insert: chad (but deletion possibly
indicated by dots in Gt; Nt does not insert it) | 272 du: QtGtNt tu | 273 pa 'o: QtGtNt pa'o | 274 spyod: QtGtNt dpyod | 275 kyi:
QtGtNt gyi | 276 du: QtGtNt tu | 277 pa 'o: QtGtNt pa'o | 278 nang [-]: possibly nas/, but given the similar annotation under line 4
below, nang / would seem more likely | 279 myed: QtGtNt med
250 this

Chapter 1

237

'khor ba dang kun nas nyon mong ***** s pa gnon pa ni280 stobs kyi pha rol du281 ***** phyin pa 'o282 /
[beneath the line, linked to nyon mongs] gnyi ga mnyam ba rnam bzhi 'i nang /

gang zag la bdag myed283 [5] par rtogs pas /

bag chags la stsogs pa 'i284 gnas myed285 cing /286
chos la bdag myed287 pas288 bag chags rang gyi ngo bor289 [Nt181] yang myed290 par rig cing291 [6] de ltar rig
pa 'i292 ye shes 293 dus thams cad du myi294 'gyur ba ni /
ye shes kyi pha rol du295 phyin pa 'o296/
/de 'i297 'bras bu ni298 sa dang po nas sa299 bcu 'i300 [6v.1] bar du ste /
[beneath the line] pha rol du phyin pa 'o
[beneath the line, curving round the end of the line on 6r.6 to fit] nyan thos dang mdo sde 'i lugs gis

nyan [Qt103b] thos kyi ltar rims kyis bkod301 pa ni ma yin gyi /
[beneath the line] $/ /de yang drang ba'I don du sa bcur bzhag go yang dag pa'i don la sor bzhag du myed do

stong pa nyid kyi ye shes rdo rjes302 mnyam pa nyid du sbyor bas /
[beneath the line, linked to ye shes rdo rjes] thabs dang shes rab [beneath the line, linked to sbyor bas] rtogs pas
sa bcu [Gt251] yang [2] tshe 'di nyid la skad cig gis rdzogs par byed do / vii

/ci 'i303 phyir zhe na /

sems dpa' chen po nyon cig /304
zhes bya ba ni/
kun nas nyon mongs [3] pa'i chos dang305 rnam par byang ba'i chos306 ***** mtshan nyid gcig du307 ston pa
la /308
[beneath line 2, linked to kun nas...] the tsom gsal ba'i phyir

the tsom309 *** za bar dogs nas310 'di skad gsungs [4] so/
/rdo rje lta bu 'i311 ye shes kyis/
[beneath the line] rdo rje dang 'dra/

**** /chos kun sgyu mar rig pa'i phyir / *****

[beneath the line, continuing under the next tshig rkang] gong du mtshan nyid myed par bstan pa'i [blo(/byo) zhan(/zhag) ba?]312
dag kun rdzob kyi rgyu 'bras kyang myed par dogs pa la gzhan dbang lta bu
[above the line, bracketed around bag chags] kun du rtogs lta bu

/dngos dang dngos myin313 bag chags [5] bsregs/

/yang na ye shes myi314 zhes bya/
/zhes gsungs te /
stong pa nyid kyi ye shes kyis /
[beneath the line] mtshan nyid myed par rig pa'i

/chos thams cad sgyu mar rig pa 'i315 [6] phyir/

[beneath the line] thabs dang shes rab

/dngos po rtog par lta ba dang /

vii

At this point, the systematic collation of Qt and Nt ends; their readings are only mentioned where it seemed useful to check
whether Gt's reading represents the Tenjur group, or in passages where a full collation was desirable, such as passages omitted in
the Dunhuang manuscript.
280 QtGtNt

insert: / | 281 du: QtGtNt tu | 282 pa 'o: QtGtNt pa'o | 283 myed: QtGtNt med | 284 stsogs pa 'i: QtGtNt sogs pa'i
QtGtNt med | 286 /: QtGtNt omit | 287 myed: QtGtNt med | 288 QtGtNt insert: / | 289 rang gyi: QtGtNt rang rang gi ngo
bo | 290 myed: QtGtNt med | 291 QtGtNt insert: / | 292 pa 'i: QtGtNt pa'i | 293 QtGtNt insert: su | 294 myi: QtGtNt omit | 295 du:
QtGtNt tu | 296 pa 'o: QtGtNt pa'o | 297 de 'i: QtGtNt de'i | 298 QtGtNt insert: / | 299 sa: QtGtNt omit | 300 bcu'i: QtNt bcu pa'i; Gt bcu
ba'i | 301 rims kyis bkod: QtGtNt rim gyis bgrod | 302 QtGtNt insert: / | 303 ci 'i: Gt de ci'i | 304 sems dpa' chen po nyon cig /: there
may once have been highlighting here, as there appears to be darkening around the letters, but this could simply be slight
smudging. The same applies to, "/rdo rje lta bu 'i ye shes kyis/" below. | 305 Gt inserts: / | 306 Gt inserts: / | 307 du: QtGtNt tu | 308 la
/: Gt las | 309 the tsom: Gt the tshom | 310 Gt inserts: / | 311 bu 'i: Gt bu'i | 312 blo bzhag intended? | 313 myin: Gt med | 314 myi:
QtGtNt me (the Dunhuang manuscript is surely in error here: our TZ edition agrees with the Tenjur reading, which is also found in
most of the versions, including the Tshal pa Kanjur and Bhutanese texts) | 315 pa 'i: Gt pa'i
285 myed:

238

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

stong pa chad par lta ba la stsogs pa316 las legs nyes 317 kyi bag chags thams cad318 bsregs
[7r.1] $ / / pa 'i319 phyir /
ye shes kyi mye320 zhes bya 'o/
/yang gsungs pa /
[above the line, continuing to the end of line 1] tshig rkang pa gnyis po 'di dpe yang yin don yang yin no *** rgyu 'bras myed du dogs
pa'i mgo gnon yin no mnya pa [continuing sideways, to the right of the page] nyid kyi dpe

/rgyu321 dang 'bras bu rnam smyin322 ni/323

[beneath the line, continuing to the end of line 1] gong du mtshan nyid myed par rtogs pa bstan pa dang rgyu 'bras myed par dogs pa
ste ma rtogs kyang rgyu 'bras myed do [+]

/myed myin324 rgyu 'bras 'thun pas 'grub325/

[2] de bzhin mnyam nyid ye shes kyis /
[beneath the line] + do zhes bya ba myi [r?]ung326 ste/

/ rig pas tshogs su gyur pas327 so /

/ zhes gsungs te/
las dge sdig gi 'bras bu ni myed pa ma yin [3] te /328
dper na mun pa'i nang na329 sgron ma330 byung na /
mun pa'i bsam bas /331
sgron ma 'ongs pas 332 'bros333 so snyam yang myi sems/334
sgron mas [4] kyang335 mun pa gnag pa bstsal336 lo ***** snyam yang myi337 sems mod kyi /
sgron ma ***** byung na338 mun pa chos nyid kyis bdang339 [5] dang 'dra bar340 [Gt252] las dge sdig gis /341
mtho **** ris dang342 ngan song bsgrub bo snyam yang myi sems343 ** /
mtho ris dang344 ngan song gis kyang345 dge sdig [6] bsgrub bo snyam myi346 sems mod kyi /
sems kyi khyad bar347 gyis de ltar grub bo /
de bzhin du chos thams cad mnyam pa348 nyid du rig [7] pas kyang /
sangs rgyas su sgrub349 bo snyam du myi350 sems mod351 kyi/
chos nyid gyis 'grub352 ces bya ba'i tha tshig go /
/de la [Nt182] yang mnya [7v.1] ba nyid la 353 rnam pa bzhi ste /
phyi354 yul chen po lnga 355 de bzhin [Qt104a] gshegs par356 mnyam pa dang /
[beneath the line, linked to mnyam] rang snang ba la bya

nang gyi357 sems rnam par shes pa ye shes lngar [2] mnyam pa dang /358
[beneath the line, linked to nang gyi] rtogs pa can la [beneath the line, linked to ye shes lngar] rtogs pa myed par

kun rdzob du359 dbang po dang yul /360

lha dang361 lha mo 'i362 tshul du mnyam pa363 dang /
don dam par skye 'gag myed pa'i364 tshul du mnya ba 'o365 /
[3] de lta bu 'i366 mnyam pa nyid kyi rig pa dang **** myi367 ldan na /
myi368 dge ba bcu 'i369 las kyi **** 'bras bu ni370 ngan song gsum 371 du ltung372/
[4] dege373 ba bcus ni 'dod khams kyi lha **** dang myir 'grub374 /
316 stsogs

pa: Gt sogs pa/ | 317 QtGtNt insert: gnyis | 318 QtGtNt omit: thams cad | 319 pa 'i: Gt pa'i | 320 mye: Gt me | 321 rgyu: Gt
dgu |
Gt smin | 323 See the TZ edition p.107 for this line; the South Central and local Kanjur texts have a more coherent
reading: dge dang mi dge rnam min ni/ | 324 myed myin: Gt med min | 325 'thun pas 'grub: Gt mthun bar grub | 326 it is not certain
that rung is intended, but if one compares rung given in the main text, 8v.1, this seems the most likely reading | 327 gyur pas: Gt
'gyur bas | 328 myed pa ma yin te /: Gt omits | 329 na: Gt du | 330 sgron ma: Gt sgron me | 331 /: Gt omits | 332 QtGtNt insert: mun
pa | 333 'bros: Gt bros | 334 yang myi sems: Gt du yang mi sems kyi | 335 Gt inserts: / | 336 gnag pa bstsal: QtGtNt bsal | 337 yang
myi: Gt du mi | 338 Gt inserts: / | 339 bdang: QtGtNt 'gro ba | 340 Gt inserts: / | 341 gis /: Gt kyang | 342 Gt inserts: / | 343 myi sems:
Gt omits | 344 dang: Gt omits | 345 Gt inserts: / | 346 myi: Gt mi | 347 bar: Gt par | 348 pa: Gt ba | 349 sgrub: Gt bsgrub | 350 myi: Gt
mi | 351 mod: Gt omits | 352 gyis 'grub: Gt kyis grub | 353 Gt inserts: yang | 354 phyi: Gt phyi'i | 355 Gt inserts: ste/ | 356 par: QtGtNt
pa lngar | 357 gyi: Gt gi | 358 rnam par shes pa ye shes lngar mnyam pa dang /: QtGtNt ye shes mnyam pa dang/ | 359 du: Gt tu
360 dbang po dang yul /: QtGtNt omit | 361 Gt inserts: / | 362 mo 'i: Gt mo'i | 363 pa: Gt ba | 364 myed pa'i: Gt med par | 365 ba 'o: Gt
pa'o | 366 bu 'i: QtGtNt: bu chos | 367 myi: Gt mi | 368 myi: Gt mi | 369 bcu 'i: Gt bcu'i | 370 Gt inserts: / | 371 QtGtNt insert: gyi sdug
bsngal | 372 ltung: QtNt lhung ngo; Gt ltung ngo | 373 dege: a small line seems to indicate the deletion of the first 'greng bu, thus
making dge. | 374 myir 'grub: Gt mir grub bo
322 smyin:

Chapter 1

239

/myi375 g.yo bas ni ***** gzugs dang gzugs myed376 pa'i [5] khams su skye ste377/
de dag ni rnam par smyin378 pa zad nas379 ngan song gsum380 du ltung381 ngo /
/de bas na dpal mchog dang po las /
[6] sems dpa' chen po /382
rang gi rig pa'i ye shes kyis /383
chos384 mnyam pa nyid du385 rtogs pa myed386 na /
sangs rgyas kyi sa la spyi bo nas
[8r.1] $ / / lag gyis387 drangs te bzhag na yang /
sems can dmyal ba'i gnas su [Gt253] mthong ngo /
rang rig dang ldan na /
dmyal ba'i [2] gnas kyang388 sangs rgyas kyi zhing yongs su dag par mthong ngo/
/zhes gsungs so /
dper na lha 'i389 bu dri ma myed390 pa'i 'od kyang/
ngan [3] song sbyong ba'i dkyil 'khor byas pas / **
/ ngan song las mtho ris su thar par zad *** kyi391 /
rang gyi rig pa'i392 ye shes myed393 [4] pas394 sangs rgyas su ma grub395 zhes ***** bshad do /
/ de bas na396 chos thams cad **** la mnyam pa nyid du rig na397 /
/ [5] dge sdig gyi398 chos thams cad 'dra ste /
bsod nams dang399 ye shes kyi400 tshogs su 'gyur ro401 zhes bya ba'i tha tshig go /
/ mnyam pa402 nyid [6] kyI don gsal bar rig pa myed403 pa rnams kyis ni /
/dge ba bcu la goms par bya ba'i404 rigs so /
/ mnyam ba nyid du rtogs pa rnas [8v.1] kyis kyang /
dge ba bcu405 gtsor len na/406
mnyam par myi407 rung ste/
nyan thos kyi sa pa tsa408 yin no/
/ yong ni nyan thos dang409 theg pa chen [2] po 'i chos gnyis ka /
sangs rgyas kyis410 snod kyi411 rim pa dang sbyar te/
sems412 can kyi413 don du 'gyur pas414 gsungs pa415 'dra bas/ /
[3] nyan thos kyi chos la yang416 lhag par **** sun byung417 zhing spang du myed do418 /
/rang gyi419 shes **** rab rtul zhing blo ma byang420 bas /
[4] [Nt183] chos mnyam pa nyid kyi don421 ma rtog **** s pa'i phyir422/
/chos mnyam pa la423 sun **** dbyung bar myi bya 'o424/
/rnam gcig [5] du425 na sems can dmyal ba'i chos kyang myi426 bkol zhes427 'byung na /
nyan thos lta ci428 smos te / viii

viii Compare the passage in gNubs chen Sangs rgyas ye shes' bSam gtan mig sgron Chapter 6 on Mahyoga, which includes an
apparent citation in the context of a discussion of Conduct which is conducive to harmony, somewhat re-arranging the presentation
here (gNubs chen Sangs rgyas ye shes: 264): /mthun 'jug 'tshams pa'i spyod pa ni/ ji ltar spyad ce na/ bsam pa la mi sgrib na 'gro
don gyi spyod pa gang yang mi dor bar spyad de/ zhags pa las/ rnam gcig tu sems can dmyal ba sogs pa'i spyod paang mi bkol/
zhes 'byung / nyan thos dang lta ci smos kyi thams cad [ma] la sun dbyung zhing lhag par spyod par mi bya'o/
375 myi:

Gt mi | 376 myed: Gt med | 377 skye ste: Gt skyes te | 378 smyin: Gt smin | 379 Gt inserts: / | 380 gsum: Gt omits | 381 ltung:
Gt lhung | 382 /: Gt omits | 383 /: Gt omits | 384 chos: QtGtNt sems can thams cad | 385 du: Gt omits | 386 myed: Gt med | 387 gyis:
Gt gis | 388 Gt inserts: / | 389 lha 'i: Gt lha'i | 390 myed: Gt med | 391 par zad kyi: Gt gyi | 392 rang gyi rig pa'i: Gt rang rig pa'i
393 myed: Gt med | 394 Gt inserts: / | 395 su ma grub: Gt ma grub bo/ | 396 de bas na: Gt de bas | 397 na: Gt nas | 398 gyi: Gt gi
399 Gt inserts: / | 400 kyi: subscribed in small writing | 401 ro: Gt omits | 402 pa: Gt ba | 403 myed: Gt med | 404 bya ba'i: Gt bya bar
405 Gt inserts: la | 406 na/: Gt pa ni/ de nyid | 407 myi: Gt mi | 408 pa tsa: it appears that a small line indicates deletion of tsa, but
this is not entirely certain; Gt omits | 409 yong ni nyan thos dang: Gt omits | 410 kyis: Gt kyi | 411 kyi: Gt omits | 412 sems: final sa
rather smudged | 413 kyi: Gt gyi | 414 pas: Gt bar | 415 pa: Gt par | 416 yang: Gt omits | 417 sun byung: Gt sun dbyung | 418 spang du
myed do: QtGt spang bar mi rung ngo | 419 gyi: Gt gi | 420 zhing blo ma byang: QtGtNt omit | 421 kyi don: Gt omits | 422 Gt
inserts: ro | 423 la: Gt nyid | 424 myi bya 'o: Gt mi bya'o | 425 rnam gcig du: Gt rnam pa gcig tu | 426 myi: Gt mi | 427 zhes: Gt ces
428 ci: subscribed in small writing

240

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

thams cad la429 mnyam bar spyod do430/

[6] tshe 'di [Gt254] nyid la thams cad mkhyen pa'i go 'phang thob bo zhes431 gsungs pa bstan pa'i phyir /
/lam kyi432 nang433na nye lam bde rdzogs
[beneath the line, continuing under 9r.1] yi ge [rang sa?] gcig la lam kyi nang na nye lam mdo thabs kyi nang da pa [she(/ste)] +

[9r.1] $ / /pa/
[beneath the line, continuing under the next tshig rkang] + bsgrubs na [chud(/khud)] kyang myi za la sgo 'i nang na yangs pa yin zhes
kyang 'grang ngo/434

thabs kyi nang na ma nor dam pa ste /

/bsgrubs na435 tshe 'dis sangs rgyas che ba yang /436 ix
/ blang dor myed pa'i437 sku gsum438 [2] rdo rje thob /
zhes439 gsungs te /
'dod pa'i yon tan rnam440 lnga ni /
mchod pa rnam lnga 'i441 rang bzhin pas442/
'dod chags ni long[s?]443 spyod [3] rdzogs pa'i sku 'o444/
/zhe sdang gtum khro *** thugs rje chen po 'i445 stobs /
stong pa bdag *** myed pa446 dang sbyar nas /
sems can log [4] par spyod447 pa dag gnyis448 su myed449 pa'i **** don la sbyor zhing/450
gnas451 stobs pa'i sbyor **** bas sprul pa 'gyed pa'i phyir452 zhe sdang [5] sprul pa'i sku 'o453/
gti mug454 ni chos thams cad rnam par myi455 rtog456 pa'i khongs su 'dus pas457 chos kyi sku 'o458/
/de459 bas na dug gsum [6] ma spangs par460 sku gsum du 'gyur bas461 nye lam mo/
/rnam gcig du462 nyan thos463 ltar bya464 na/
khams gsum kyi465 nyon mongs pa466 [9v.1] bsgom ba467 spang bar bya ba 'ba' shig468 kyang /
brgya dgu bcu rtsa brgyad yod do/469
de dag spangs nas470 dgra bcom ba'i471 'bras bu thob bo/ /
[2] de nas gdod472 shes bya 'i bsgrib473 pa spong ste/
bskal pa grangs myed474 pas/
475mnyam ba476 nyid rtogs nas sangs rgyas so/
477de ltar chos mnyam [3] pa nyid rtogs nas/
nyan thos kyi [Gt255] ltar **** kun nas nyon mongs pa dang478 rnam par byang **** ba'i chos so sor lta
zhing479 nyon mongs [4] pa ni spong /
rnam par byang ba ni480 thob par ***** 'dod pa lta bu ni ma481 yin gyi/
rims482 gyis **** spang du myed483 de/
kun nas nyon mongs [5] pa'i chos484 rang bzhin myed485 par rig na486 /
ix

Note that Chapter 13 of Klong chen pa's Phyogs bcu mun sel commentary on the rGyud gSang ba'i snying po cites this verse of
the Thabs kyi zhags pa (bdud 'joms bka' ma Volume La: 445-446; see our TZ edition, Chapter 1 note iv).
429 la:

Gt omits | 430 mnyam bar spyod do: Gt mnyam par dpyad de | 431 bo zhes: Gt ces | 432 kyi: Gt gyi | 433 gsungs pa bstan pa'i
phyir / /lam kyi nang: highlighting unclear and uncertain. | 434 a long downwards line, striking through ngo, is possibly intended
simply to mark off the comment as belonging to the above text. | 435 na: Gt nas | 436 for these two verse lines, highlighting may
once have been present, suggested by an apparent slight smudging appearance. | 437 myed pa'i: QtGtNt med pas (the Tshal pa
Kanjur and Bhutanese texts agree with the Dunhuang manuscript, giving med pa'i; but our TZ edition gives med par) | 438 gsum:
QtGtNt omit (but the TZ editions do not) | 439 / zhes: Gt ces | 440 rnam: Gt rnam pa | 441 rnam lnga 'i: Gt rnam pa lnga'i | 442 pas: Gt
bas | 443 final sa very unclear but appears to have been inserted as a small letter beneath the line; Gt longs. | 444 sku 'o: Gt sku'o
445 po 'i: Gt po'i | 446 myed pa: QtGtNt med pa'i don | 447 spyod: Gt phyogs | 448 gnyis: QtGtNt / h gis rnam par shes pa bsdus/
ho ye shes kyi rlung gis btags te/ /dmigs (this addition would seem to be in error.) | 449 myed: Gt med | 450 /: Gt omits | 451 gnas:
QtGtNt gang | 452 sbyor bas sprul pa 'gyed pa'i phyir: QtGtNt phyir/ | 453 sku 'o: Gt sku'o | 454 Gt inserts: gi | 455 myi: Gt mi
456 rnam par myi rtog: highlighting uncertain: may simply be smudging. | 457 Gt inserts: / | 458 sku 'o: Gt sku'o | 459 QtGtNt insert:
lta | 460 Gt inserts: / | 461 'gyur bas: Gt gyur pas | 462 rnam gcig du: Gt rnam pa gcig tu | 463 Gt inserts: kyi | 464 bya: Gt byas
465 kyi: Gt gyi | 466 pa: Gt omits | 467 ba: Gt pas | 468 shig: Gt zhig | 469 do/: Gt omits | 470 nas: Gt na | 471 ba'i: Gt pa'i | 472 gdod:
Gt sdod | 473 bya 'i bsgrib: Gt bya'i sgrib | 474 myed: Gt med | 475 QtGtNt insert: chos | 476 ba: Gt pa | 477 Gt inserts: / | 478 Gt
inserts: / | 479 Gt inserts: / | 480 ba ni: Gt ba'i chos | 481 ma: Gt omits | 482 rims: Gt rim | 483 myed: Gt med | 484 chos: Gt omits
485 myed: Gt med | 486 na: Gt nas

Chapter 1

241

rnam par byang ba yin bas487 spang du myed488 de/

dper489 nyi ma shar na490 yul khams brgyud491 de shar pa492 [6] lta bu493 [Nt184] ni ma yin gyi/
'dzam bu gling dus gcig du lham zhes shar pa494 dang 'dra bar /
chos mnyam pa nyid495 rig na496/
bsgrib pa rims [10r.1] $/ / kyis497 spang du myed498 de/
thams cad mkhyen pa'i go 'phang dus gcig du499 thob ces500 pa'i tha tshig go/ x
/bcom ldan 'das kyis [2] kyang501 sems can drang ba'i thabs bsam gyis myi502 khyab pa zhig gsungs te/
de dag gyi503 nang na504 dam pa ni505 'di yin no/
de bas na 'di las [3] 'byung ba bzhin bsgrubs na /
tshe 'di nyid **** la506 ye shes shar te/507
mnyam pa nyid kyi **** klung508 chen du gyur nas/
sangs rgyas [4] 509 phyi mar bsdad myi510 dgos te tshe 'di nyid **** la 'grub bo/
nyan thos kyi511 ltar dge ba **** ni rung /
myi512 dge ba ni myi rung bar sgros [5] 'dogs pa513 lta bu ni ma yin gyi/
ci514 spyad515 kyang bsod nams dang516 ye shes kyi tshogs su 'gyur ro517 zhes bya ba'i tha tshig go /
/ [6] 'phags pa thabs kyis zhags pa pad ma 'phreng las/518
bden pa gnyis bstan pa'i le'u ste519 dang po 'o520// : :521 //

x gNubs chen Sangs rgyas ye shes' bSam gtan mig sgron Chapter 6 on Mahyoga cites these lines in the context of a discussion of
natural purification of the defilements (gNubs chen Sangs rgyas ye shes: 271-272): zhags pa las/ sgrib pa rims kyis spangs du
med de thams cad mkhyen pa'i go 'phang la dus gcig tu sbyor ro/ zhes 'byung /
487 bas:

Gt pas/ | 488 myed: Gt med | 489 Gt inserts: na | 490 na: Gt nas | 491 brgyud: Gt rgyud | 492 pa: Gt ba | 493 lta bu: Gt omits
zhes shar pa: QtGtNt lhag gis 'char ba (QtGtNt is surely more appropriate here) | 495 Gt inserts: du | 496 na: Gt nas
497 bsgrib pa rims kyis: Gt sgrib pa rim gyis | 498 myed: Gt med | 499 du: Gt tu | 500 ces: Gt po zhes | 501 Gt inserts: / | 502 myi: Gt
mi | 503 gyi: Gt gi | 504 na: Gt nas | 505 ni: Gt omits | 506 la: Gt las | 507 shar te/: Gt omits | 508 klung: QtGtNt klong | 509 QtGtNt
insert: kyi sa | 510 myi: Gt mi | 511 kyi: Gt omits | 512 myi: Gt mi | 513 myi rung bar sgros 'dogs pa: Gt sgribs 'dos pa | 514 Gt inserts:
ltar | 515 spyad: Gt dpyad | 516 Gt inserts: / | 517 ro: Gt omits | 518 pad ma 'phreng las/: Gt pad 'phreng gi | 519 Gt inserts: / | 520 po 'o:
Gt po'o | 521 Here there is a repeated ornamental punctuation mark, consisting of two small vertically arranged circles, marking the
break between chapters. The same kind of punctuation mark occurs after each of the subsequent chapters, although for most
chapters, there is only a single rather than a double mark.
494 lham

Chapter 2
i [Qt105a.6]

[Gt255.6] [Nt184.4] ii

[Ms10v.1] //da ni sgo 'dir1 [Gt256] 'jug pa'i dam tshig2 bstan pa'i phyir/
/sems dpa' chen po dam tshig 'di dag3 ni/
/bdag dang gzhan la phan pa'i [2] phyir ngas bshad do/
/zhes gsungs te/
gti mug zhe sdang nga rgyal4 'dod chags phrag dog5 ni/
/bsrung ba6 myed7 pa'i sdom pa dam tshig [3] rdo rje can/ 8
/zhes gsungs te/ *****
iii /gti mug ni rnam par snang mdzad kyi dam **** tshig ste /
ci 'I phyir zhe na/ /
[4] ma rig pa9 spang du myed10 /
rig pa ni b*****sgrub du myed11 de/
chos nyid12 kyi dbyings su ***** ro gcig pa'i phyir ro/
iv /zhe sdang ni13 myi [5] 'khrug14 pa'i dam tshig ste/
chos kyi dbyings las kyang ma 'gyur par15 thams cad 'dul ba'i16 phyir ro/
v /nga rgyal nI rin cen17 'byung ldan gyI dam [6] tshig18 ste/
chos19 mnyam ba20 nyid du bden bar rig nas /
de la sran 'dzugs21 pa'i phyir ro22/
vi /'dod chags ni snang ba snang23 mtha'24 yas kyi da
[11r.1] $/ / tshigs25 ste/
thams cad dbang du byed pa'i tha tshig go/

Note that Chapter 1 of Klong chen pa's Phyogs bcu mun sel commentary on the rGyud gSang ba'i snying po cites the Thabs kyi
zhags pa on this chapter (bdud 'joms bka' ma Volume La: 63). Unless some section of the root text has been lost in the now extant
root text versions, it seems that it is in fact TZComm which is cited or rather paraphrased. Note the parallel sentences added into
the notes below, and see also Dorje 1988: 393. The passage is given in the context of a discussion of the primordial identity of
sentient beings and buddhas, mundane and enlightened phenomena.
ii As noted in Chapter 1 note vii above, apart from the sample of the first section of Chapter 1, and elsewhere where the Dunhuang
manuscript has an omission, Qt and Nt are not fully collated. They have been consulted and variants noted where appropriate to
confirm whether or not Gt's variants of significance are shared by the Tenjur tradition in general.
iii Compare the Phyogs bcu mun sel's apparent citation (bdud 'joms bka' ma Volume La: 63): 'di ltar gti mug ni rnam par mi rtog
pa'i khongs su 'dus pas/ spang blang med pa'i mtshan nyid ni sku de bzhin gshegs pa'i rigs so/
iv Compare the Phyogs bcu mun sel's apparent citation (bdud 'joms bka' ma Volume La: 63): /zhe sdang ni chos kyi dbyings las ma
g.yos bzhin du gdug pa 'dul ba'i mtshan nyid rdo rje'i rigs so/
v Compare the Phyogs bcu mun sel's apparent citation (bdud 'joms bka' ma Volume La: 63): /nga rgyal ni chos mnyam pa nyid la
sran tshugs pa'i mtshan nyid rin po che'i rigs so/
vi Compare the Phyogs bcu mun sel's apparent citation (bdud 'joms bka' ma Volume La: 63): 'dod chags ni chos thams cad la
mnga' brnyes pa'i mtshan nyid padma'i rigs so/
1 sgo

'dir: Gt sgor | 2 dam tshig: QtGtNt 'tsham tshim (the Dunhuang manuscript reading seems clearer here) | 3 'di dag: Gt omits
(the TZ versions do not omit it) | 4 Gt inserts: / (the TZ versions give: dang/) | 5 Here, the Dunhuang manuscript and the Tenjur
texts both omit chen po, which is given in the TZ versions. | 6 ba: Gt du | 7 myed: Gt med | 8 dam tshig rdo rje can/: QtGtNt can/
/de ni dam tshig rdo rje yin/ (Note that this reading is shared by all other editions; the Dunhuang manuscript appears to be in
error.) | 9 Gt inserts: ni | 10 myed: Gt med | 11 bsgrub du myed: Gt bsgrubs tu med | 12 nyid: Gt omits | 13 ni: subscribed in the same
handwriting | 14 myi 'khrug: Gt mi 'khrung | 15 par: ra subscribed; Gt bar | 16 'dul ba'i: QtGtNt du 'dus pa'i | 17 rin cen: Gt rin chen
18 dam tshig: Gt dam dam tshig (QtNt do not share this dittography) | 19 QtGtNt insert: thams cad | 20 ba: Gt pa | 21 sran 'dzugs: Gt
sran tshugs | 22 ro: Gt omits | 23 snang: Gt omits | 24 mtha': 'a subscribed | 25 da tshigs: Gt dam tshig

Chapter 2

243

vii /phrag

dog ni26 gdon myi27 za bar grub pa'i dam tshig ste/
chos mnya par [2] 28rtogs pas29 dge sdig [myed(/gyed)(/gyid)]30 pa'i 31phrag dog go
/mdor na chos32 mnyam ba nyid du chud de/
rang33 rig gI34 ye shes can du gyur na/35
bsrung ba [3] dang myi srung ba myed36 de/
[beneath line 3] bsrung ba yang myed de ye shes lnga mnyam ba nyid las ma gtogs par de ltar bsrung du myed
[above line 3, bracketted and linked to bsrung du myed...] bsrung ba yang myed do so sor btab pa yang ma yin

chos37 thams cad **** chos kyi klung38 du gyur pas39 grub pa yin **** pa'i phyir bsrung du myed40 kyi/
ma grub [4] cing sems la ye [Gt257] shes mnya pa nyid ***** myed41 bzhin du dam tshig ma srungs42 na/
**** dmyal ba'i yang gting du 'gyur gyi/ [5] te /43
mnyam pa'i ye shes ma44 thob pas45/
mnyam ba46 la skur pa47 btab pa yin no/
dam tshig chen po ni48 dag pa rnam gsum ste/ /
[beneath the line, linked to dam tshig]49 gi rtsa ba ni [beneath the line, linked to rnam gsum] sku gsung thugs

[6] de nyams su myong zhing50 lon bar51 bya ba'i gnya' nan52 du/
dam tshig lnga dang /
[above the line, bracketted and linked to lnga] kha

dam zas53 lnga dang /

[beneath the line] bdud rtsi lnga
[above the line, bracketted and linked to chos] ga

chos la dge sdig myed54 pa dang /

[beneath the line, linked to dge sdig] dogs pa'i gnyen po

zas la [11v.1] gtsang smye myed55 pa'i dam tsig mnyams56 na/
[beneath 11r.6, linked to zas la] rtog pa'I gnyen po

/gtugs pa57 lnga la58 spyod kyang/

/de 'I skyon kyis myi59 gos pa nI/
/60lus rdo rje zhes bya 'o61/ viii
/ngag gis [2] sna tshogs su smras kyang/
/nyes pa 'i skyon kyis myi62 gos pa'i phyir ngag rdo rje zhes bya63/
/de ltar rig pa'64 ni sems rdo rje zhes bya 'o//
// /'phags pa'65 [3] thabs kyi zhags pa'66 pad ma67 'phreng68 las/ ****
/dam tsig69 bstan pa'i le 'u ste70 gnyis pa 'o // :

vii

Compare the Phyogs bcu mun sel's apparent citation (bdud 'joms bka' ma Volume La: 63): phrag dog ni chos mnyam par ma
rtogs pa rnams la las dge sdig yod phyir 'gro ba'i don byed pa'i mtshan nyid las kyi rigs so/
viii Rong zom pa cites this passage in an untitled Thor bu, omitting a few words (bka' 'bum 1976: 397-398): /de ltar yang 'phags
pa thabs kyi zhags pa las/ dam tshig chen po ni dag pa rnam pa gsum ste/ de nyid nyams su len par bya ba'i gnya' nan du chos la
dge sdig med pa dang/ zas la gtsang rme med pa'i dam tshig ma nyams na/ gdug pa lnga la spyod pas kyang de'i skyon gyis mi gos
pa ni lus rdo rje zhes bya ba la sogs pa gsungs pa'o/
26 QtGtNt

insert thams cad | 27 myi: Gt mi | 28 QtGtNt insert: ma | 29 Gt inserts: / | 30 [gyed(/myed)(/gyid)]: QtGtNt yod
insert: phyir | 32 QtGtNt insert: thams cad | 33 rang: Gt nang (QtNt rang) | 34 gI: Gt pa'i | 35 /: Gt omits | 36 myi srung ba
myed: Gt mi bsrung ba med | 37 chos: QtGtNt omit | 38 klung: QtGtNt klong | 39 Gt inserts: / | 40 myed: Gt med | 41 mnya pa nyid
myed: QtGtNt nyams pa med | 42 srungs: Gt bsrungs | 43 gyi/ te/: Gt te | 44 ma: subscribed in the same writing, with a cross above
the line marking the position. | 45 thob pas: QtNt rtogs pas; Gt rtogs bas | 46 ba: Gt omit | 47 skur pa: Gt bskur pa | 48 Gt inserts: /
49 The intended positioning is not entirely clear: it may be that this is intended to link to the line below (dam zas lnga), and the
annotation, bdud rtsi lnga. | 50 Gt inserts: / | 51 lon bar: QtGtNt log par | 52 gnya' nan: QtGtNt gnya' non | 53 zas: Gt rdzas
54 myed: Gt med | 55 smye myed: Gt sme med | 56 dam tsig mnyams: QtGtNt dam tshig ma nyams | 57 gtugs pa: QtGtNt dug
58 la: QtGtNt omit | 59 kyis myi: Gt gyis mi | 60 / /: Gt omits | 61 bya 'o: Gt bya | 62 kyis myi: Gt gyis ma | 63 bya: Gt bya'o | 64 pa':
'a subscribed; Gt pa | 65 pa': 'a subscribed; Gt pa | 66 pa': 'a subscribed; Gt pa | 67 pad ma: Gt padmo | 68 'phreng: Gt phreng ba
69 dam tsig: Gt dam tshig | 70 Gt inserts: /
31 QtGtNt

Chapter 3
[Qt105b.7] [Gt257.5] [Nt185.4]
[Ms11v.3] //da ni dkyil [4] 'khor gsum bshad de/1
/sems dpa'2 ***** chen po dbang 'di dag ni /
[beneath the line, continuing under the next tshig rkang] don du dbang gyi mtshan nyid bshad pa pa 'o/ dbang la yang yo byed kyis
thob pa dang rigs pa thob pa gnyis la 'dir ni rigs pa'i rtsal gyis

rang gyI3 rig pa'I ***** rtsal gis4 thob bo/

/zhes gsungs [5] te/
dkyil 'khor rnam gsum gyI tshul gyis/
[beneath the line, linked to tshul gyis] 'og nas 'byung ba'i

chos thams cad mnyam pa nyid du rig pa na5/

dbang rang bzhin gyis thob6 bo zhes bya7 [6] ba'I8 don to/
/thabs dang shes rab dbang chen te/
[beneath the line] longs spyod rdzogs pa'i sku dang sprul pas sku gsum thob pa la bya dbang chen rigs kyis rtsal gyis thob bo/9
[above the line, bracketted and linked to rdo rje] thabs [above the line, bracketted and linked to dril bu] shes ra[b]

/rdo rje dril bu rigs10 pas 'dzin/

[beneath the line] kI la ltar rdo rje dang dril bu dngos su 'dzin pa zhes na 'dir ni dngos su

____________________________________________________________________
[Here, the Dunhuang manuscript appears to have an omission.11 Our version of the root text lines is based on
the Tshal pa Kanjur/Bhutanese TZ texts, noting the Tenjur variants, since in this case the Tenjur gives an
reading which does not occur in the other root texts and appears to be in error:
/mya ngan 'das dang 'khor ba yi/
/chos la 'dod bzhin12 [Gt258] dbang byed do13/
zhes gsungs te/]
____________________________________________________________________
/de yang ci lta bu [Qt106a] zhe na/
mya ngan las 'das [12r.1] $/ /pa dang /
'khor ba'i chos thams cad la ci14 ltar 'dod pa bzhin dbang byed15 do zhes16 gsungs te /
[beneath the line, linked to ci ltar] thabs dang shes rab gnyis su myed par rig na

rigs kyi gtso bo ni thabs te/

[2] rdzogs long spyod pa'i sku 'o17/
/gtso mo18 ni shes rab ste19 chos kyi sku 'o20/
/de dag gi byin rlabs21 las byung ba'i22 sems skye ba ni/
sprul pa'ii23 [3] sku 'o/
/de ltar spyod pa'i24 'phags ***** pa thams cad dang /
bdag du25 tha myi26 dad ***** par spyod27 pa ni28 sku gsum thob pa'i [4] dbang ngo /
/rdo rje ni thabs kyi rtags ***** te29 chos thams cad la spyod30 pa 'o/ *****
/dril bu ni shes rab kyi rtags te/
1 bshad

de/: Gt omits | 2 dpa': 'a subscribed | 3 gyI: Gt gi | 4 gis: Gt gyis | 5 na: Gt ni | 6 thob: Gt 'thob | 7 bya: Gt omits | 8 ba'I: Gt
pa'i |
than a usual shad, there is a long wavy vertical stroke here, separating this comment from the next. | 10 rigs: Gt rig
11 The most likely explanation of what at first sight appears to be a dittography in QtGtNt, is that first we have the root text
metrical verse lines, given here, and second (below), there follows a wordier prose version which is in fact commentary. Thus, the
Dunhuang manuscript, which gives only the commentarial lines, has in fact omitted the root text lines altogether through
haplography. | 12 chos la 'dod bzhin: QtGtNt 'dod pa'i chos la (a reading given only in the Tenjur texts) | 13 do: QtGtNt pa'o | 14 la
ci: Gt la/ ji | 15 byed: Gt byad | 16 do zhes: Gt ces | 17 rdzogs long spyod pa'i sku 'o: QtNt longs spyod rdzogs pa'i sku'o; Gt longs
spyod pa'i rdzogs pa'i sku'o. | 18 gtso mo: QtGtNt rigs kyi gtso bo | 19 Gt inserts: / | 20 sku 'o: Gt sku'o | 21 byin rlabs: Gt byin gyis
brlabs | 22 Gt inserts: byang chub kyi | 23 pa'ii: 'a subscribed; the double gi gu presumably in error; Gt pa'i | 24 de ltar spyod pa'i:
Gt de dpyod pa | 25 du: Gt tu | 26 myi: Gt mi | 27 spyod: Gt dpyod | 28 Gt inserts: / | 29 Gt inserts: / | 30 spyod: Gt dpyod
9 rather

Chapter 3

245

[5] chos thams cad31 sgyu mar rig pa'i phyir/

'khor ba dang mya ngan las 'das pa'i chos thams cad la/
ci32 ltar 'dod par dba[ng?]33 [6] sgyur zhing spyod34 do/
[beneath line 5] thabs dang shes rab gnyis su myed pa [---] + [beneath line 6] + spyad ky[i?] bag chags su myi 'gyur

/glang po las kyang /

[beneath the line] glang po rab 'bog i lung gyi khungs so

/ phyogs bcu 'i35 'jig rten gang nas kyang /

/sangs rgyas rnyed par yong myi36 'gyur/
[12v.1 /rig pa'i sems nyid sangs rgyas te/
/sangs rgyas gzhan du37 ma tshol cig /
zhes38 'byung ba lta bu 'o/
/'phags pa thabs [2] kyi zhags pa pad ma39 'phreng40 las dbang gyi41 le'u ste42 gsum pa 'o43// :

i Glang

po rab 'bog: one of the 18 Mahyoga tantras, as classified by the rNying ma pa.

31 Gt inserts: la | 32 ci: Gt ji | 33 dba[ng?]: the final nga is partially smudged out, but would seem to have been intended. There is
some damage to the paper at the bottom right-hand corner and this has affected the end of this line and the annotation below; Gt
dbang | 34 spyod: Gt dpyod | 35 bcu 'i: Gt bcu'i | 36 myi: Gt mi | 37 du: Gt tu | 38 zhes: Gt ces | 39 pad ma: Gt padmo | 40 'phreng: Gt
'phreng ba | 41 gyi: Gt gi | 42 Gt inserts: / | 43 pa 'o: Gt pa'o

Chapter 4
[Qt106a.5] [Gt258.6] [Nt186.2]
[Ms12v.2] //da ni mchod pa 'i cho ga1 bstan pa'i2 [3] phyir /
[beneath the line] don drug pai mchod pa'i mtshan nyid bstan pa 'o

/sems dpa' chen po mchod **** pa3 [Gt259] dam pa mchog4 ni 'di ltar rig par bya 'o/
____________________________________________________________________
[Here, the Dunhuang manuscript has an omission. We give the QtGtNt version of the commentarial words,
but rely on the TZ texts for the root text verse:
/nyan thos las khyad par cher 'phags pa'i phyir mchod pa'i mchog ces bya'o/ /
ci' phyir mchod pa dang mchod ces gsungs te/
bdag mchod5 kun mnyes thams cad 'drangs/
/kun la phan pa'i sman yin te/
/rnal 'byor bdag nyid chen po yis/
/thams cad bza' zhing brgyan par6 bya/]
____________________________________________________________________
**** zhes gsungs te/
'phags pa [4] thams cad dang /
stong khams kyi s*****ems can dang /
bdag du7 chos kyi dbyings **** mnyam pa nyid kyis /8
so so ma yin9 [5] tha myi10 dad par rig pa'i rnal 'byor pas/
11'dod pa'I yon tan rnam12 lnga rtag du spyad de13/
bdag tshim bas/
'phags pa thams cad kyang14 [6] mnyes par 'gyur ba15 dang /
/'di lta bu 'i byin rlabs16 kyis /17
khams gsum gyi sems can thams cad kyang sgrol zhing18 dga'19 bas / /
[13r.1] $/ / rnal 'byor pas ni20 dus thams cad du mchod pa dang21 gtor ma bya ba dang /
hom yang 'phags pa dang22 bdag so so ma yin pa'i ting nge 'dzin [2] gyis/23
lus kyi nang du24 yi ge 'bru lngas25 gzhal myed26 khang du gyur pa'i dbus na /
'phags pa rnams bzhugs par gsal bar dmyigs te27 [3] dper na28 kha ni hom khung /
[beneath line 3] he las stsogs pa'i rang bzhin he 'i dkyil du gru gsum byas la/

lce ni las **** gang byed pa'i29 gtso bo sprul pa30 dang bcas pa *** /
[beneath the line] yi ge hu las rdo rjer 'gyur

lag pa ni hom rdo rje31 gzar32 te / [Gt260]

[beneath the line] gsung las byung ba'i hom gzar kyi rang bzhin du byas te

i See Chapter 1 note iv above: the annotator provides a numbered list of points although some are missing; the third and fourth
occur in Chapter 1, the fifth is missing, but the seventh occurs below in Chapter 5.
1 cho

ga: QtGtNt mchog | 2 bstan pa'i: the sa head letter and the 'i are slightly affected by washing out, presumably by the (water?)
damage to the top right-hand corner of this sheet, which (see above) has affected the bottom right-hand corner of folio 12r. In this
case, the writing remains clearly legible. | 3 pa: Gt pa'i | 4 dam pa mchog: Gt dam pa'i mchog | 5 mchod: Gt omits (but QtNt do not
omit) | 6 bza' zhing brgyan par: QtGtNt brgyan cing bza' bar | 7 du: Gt tu | 8 kyis /: Gt kyi | 9 yin: na subscribed (end of line)
10 myi: Gt mi | 11 QtGtNt insert: bdag nyid zhal zas | 12 rnam: Gt omits | 13 rtag du spyad de: Gt rtag tu dpyad te | 14 kyang: Gt
omit | 15 'gyur ba: Gt gyur pa | 16 byin rlabs: Gt byin gyi brlabs | 17 /: Gt omits | 18 sgrol zhing: QtGtNt 'drangs shing | 19 dga': 'a
subscribed | 20 Gt inserts: / | 21 Gt inserts: / | 22 Gt inserts: / | 23 /: Gt omits | 24 Gt inserts: ni | 25 'bru lngas: QtGtNt bru las/
(generally, bhr is the syllable for the Palace, which would seem to support the Tenjur reading, although the five buddha seeds
might also make sense in this context) | 26 myed: Gt med | 27 dmyigs te: Gt dmigs te/ | 28 dper na: Gt omit | 29 byed pa'i: Gt bya
ba'i | 30 sprul pa: Gt omits | 31 rdo rje: QtGtNt omit | 32 QtGtNt insert: du dmigs

Chapter 4

247

[4] de ltar phun sum tshogs pa'i don ***** yin bas/
bdag nyid du33 khyad par du ***** 'phags pa'i sems dang ldan bas34 [5] bdag ma nor pa'i35 dbang phyug chen
po yin ba'i phyir/
zhal zas dang rgyan gyis brgyan par bya 'o/ /
zhes bya ba'i tha tshig go / /
[beneath the line and extending vertically up the side of the edge of the page] rang bzhin kyi dkyil 'khor dang yid kyi dkyi36 'khor
dang gzugs brnyan kyi dkyil 'khor ro ii

[6] 'phags pa thabs kyis zhags pa pad ma 'phreng37 las mchod pa'i le'u ste38 bzhI pa 'o// :

ii This appears to be the annotator's understanding of the three maalas mentioned in the main text above (in Chapter 1
[Ms1v.1-2], and at the beginning of Chapter 3 [Ms11v.3-4]; see also the annotation connected with Ms1v.1-2, where yid is given
third in the list). The three maalas are discussed in the text at greater length from Chapter 5, where we find rang bzhin kyi dkyil
'khor (rang byung kyi dkyil 'khor is given in the Dunhuang manuscript and commentarial transmission of the root text lemmata,
although rang bzhin kyi dkyil 'khor occurs in the commentarial text below, Ms14v.5) and rdo rje'i dbyings kyi dkyil 'khor
(Ms15r.2), with the chapter title referring to, ting nge 'dzin kyi dkyil 'khor. Then, gzugs brnyan gyi dkyil 'khor is the topic of
Chapter 6. In the annotations within Chapter 5, the impression is given that the first part of the chapter (up to Ms14v.5) concerns
rang bzhin kyi dkyil 'khor, while the final section concerns ting nge 'dzin kyi dkyil 'khor (annotation bracketted to Ms14v line 6: 'di
man cad...). However, it does not appear that the annotator intends this as a separate category to be equated with the yid kyi dkyil
'khor listed second here. An annotation at the beginning of Chapter 7 suggests that this chapter concerns, yid kyi dkyil 'khor.
However, the term, yid kyi dkyil 'khor, does not occur anywhere in the main text. The three are not clearly specified in the main
text, but it would seem more likely that TZComm's understanding is reflected in three terms used in its text, most probably, rang
bzhin kyi dkyil 'khor and ting nge 'dzin kyi dkyil 'khor, explained in Chapter 5, and gzugs brnyan gyi dkyil 'khor, explained in
Chapter 6.
33 du:

Gt omit | 34 Gt inserts: / | 35 pa'i: Gt ba'i | 36 dkyi: sic; dkyil presumably intended | 37 'phags pa thabs kyis zhags pa pad ma
'phreng: Gt 'phags pa padmo phreng gi don | 38 Gt inserts: / (QtNt do not)

Chapter 5
[Qt106b.5] [Gt260.2] [Nt187.1]
[Ms13r.6] //da ni dkyil 'khor [13v.1] gsum phye ste bshad par bya'o/
[beneath 13r.6 and extending vertically up the side of the edge of the page] don bdun pa i dkyil 'khor kyi mtshan nyid bstan pa 'o

/kun du1 'od kyi sa las 'das nas ni/2

[beneath the line] kun du 'od kyi sa 'i 'og na yod pa 'am bzhag nas pad ma'i spyan kyis steng na yod pa ma yin te kun du 'od nyid la
myi gnas pa la 'das zhes bya 'o
/pad mo'I3 spyan du gyur pa na4/ ii

/rig pa'i byin rlabs5 chen po [2] las/

/rang byung dkyil 'khor6 'du 'phro 'i7 tshul/
zhes8 gsungs te/
chos kyi sku ni kun du9 khyab par gnas so/
[beneath the line] kun du 'od ces bya ci 'i phyir kun du 'od ces bya zhe na sa 'i phyir kun du 'od ces bya 'o

/de las 'das par [3] gnas10 pa ni/

[beneath line 3] kun du 'od do

ma chags pa'i spyan ***** gyis chos kyi sku la bltas na/
[beneath the line, linked to spyan] mye long lta bu 'i ye shes kyis

ci11 lta bu **** 'i tshul du yang ma mthong ba12 de ltar [4] mthong ba'i tshe/
[beneath the line] kun du 'od nyid kyang [beneath the line, linked to ma mthong] mtshan pa'i rnam par dngos po dang mtshan +
[beneath line 4] + mar 'dzin cing chags pa myed par

'bad pa myed13 par rig **** pa'i byin rlabs14 las/

rang byung15 'od kyi ***** 'du 'phro ltar 'gyur te16/
dang po yang [5] chos kyi sku 'i byin rlabs17 las byung18 /
slar yang der 'dus te/
mtshan myed19 pa la20 rang bzhin gyis gnas so/
[beneath the line] thig le chen po yi ge myed pa dang mmthun21 thig le

de ltar byung22 ba gang zhe na23 [6] bstan pa'i/ phyir24

/yi ge 'khor lo 'i tshogs can te25/
[beneath the line] stong pa chen po yi ge lnga

/mtshan dang dpe byed26 kun las 'phros27/

[beneath the line] mtshan bcu drug phyed phyed thig le can

See Chapter 1 note iv above regarding the list of points (don) found in the annotations.
gNubs chen Sangs rgyas ye shes' bSam gtan mig sgron Chapter 6 on Mahyoga cites these two opening lines, and moreover,
Klong chen pa (Phyogs bcu mun sel, in the bdud 'joms bka' ma Volume La: 255) cites exactly the same lines with the same
wording as gNubs (see our TZ edition of Chapter 5 note i).

ii

1 du:

Gt tu | 2 nas ni: Gt omits | 3 pad mo'I: the 'a is subscribed and its gi gu is above mo; Gt padma'i | 4 pa na: Gt pas ni | 5 byin
rlabs: Gt byin brlabs | 6 rang byung dkyil 'khor: QtGtNt rang 'byung 'od kyi (but note that the other versions of TZ agree with the
Dunhuang manuscript on dkyil 'khor here, except that the Tshal pa Kanjur texts shorten it to dkyil. Also, although the Tshal pa
Kanjur and Bhutanese texts also give rang 'byung, and Tawang agrees with rang byung, the South Central, Hemis and Bathang
texts give rang bzhin. See the discussion, p.63 above.) | 7 'phro 'i: Gt 'phro'i | 8 zhes: Gt ces | 9 du: Gt tu | 10 gnas: QtGtNt khyab
11 ci: Gt ji | 12 Gt inserts: / | 13 myed: Gt med | 14 byin rlabs: Gt byin brlabs | 15 byung: QtGtNt omit | 16 'gyur te: Gt byung ngo
17 byin rlabs: Gt byin gyis brlabs | 18 byung: Gt 'byung | 19 mtshan myed: Gt mtshan ma med | 20 la: Gt las | 21 mmthun: there
appears to be a crossed out head letter ma or sa above the prefixed ma; possibly, it was written unclearly and the scribe decided to
write it more clearly below. | 22 byung: Gt 'byung | 23 zhe na: Gt bcas na/ | 24 / phyir: phyir is written small, apparently inserted;
most probably, the shad is intended to follow the insertion; Gt phyir/. | 25 'khor lo 'i tshogs can te: QtGtNt tshogs chen 'khor lo ste
(but note that the other versions of TZ agree with the Dunhuang manuscript on the order here, although only the Bhutanese texts
agree with can rather than chen.) | 26 byed: Gt byad | 27 'phros: Gt 'phro

Chapter 5

249

/rngam zhing gtum28 pa'i gzugs can [14r.1] $ / /dang /

mgyogs pa mgyogs ma 'pho29 nyar ldan/
zhes gsungs te /
yig30 'bru lnga lta bus31 mtshon pa las/ [Gt261]
[above the line, bracketted and linked to mdzad] ka

ye shes lnga las [2] gtso bo rnam par snang mdzad dang /
[beneath line 2] [cho?]s kyi dbyings kyi ye shes
[above the line, bracketted and linked to 'khrug] kha

myi32 'khrug pa dang /

[beneath the line] mye long lta bu 'i ye shes
[above the line, bracketted and linked to cen] ga

rin cen33 dpal dang

[beneath the line] mnyam ba nyid kyi ye shes
[above the line, bracketted and linked to ba] nga

/snang ba mtha'34 yas dang /

[beneath the line] so sor rtog pa'i ye shes

gdon myi35 za bar grub pa 'o/

[beneath the line] bya ba nan tan kyi ye shes

[3] ye shes lnga 'i yul lnga las36/ *****

gtso mo37 kun du38 bzang mo 39/
yangs pa'i40 spyan dang /
[beneath the line] sgreg mo iii

ma ma ki41 dang /
[beneath the line] rdo rje dbyings iv

na bza'42 dkar dang /

[beneath the line] 'phreng ba

dam [4] tshig bsgrol ma 'o/

[beneath the line] gar byed ma

/de 'i sbyor pa'I43 ***** byin rlabs44 las byung ba ni45 /
[beneath the line] yong ni 'bu ta la mtshan bzang po sum bcu rtsa gnyis dang dpe byed bzang po brgyad cu mnga' ste de la 'dul ba'i
dbang gyi lha dang lha mo 'i sku grangs myed kyis ma smos te tshul 'di ltar sbyar ba yin no

sems dpa' **** bcu drug dang /

sems ma bcu drug46 [5] phrag47 gnyis na48 sum bcu49 rtsa gnyis te/
iii

sgreg mo = sgeg mo. See the list of goddesses in the maala specified below in Chapter 7. The four specified in the
annotations here correspond to the second of the four females in each of the buddha family lists given in Chapter 7. They are also
listed in Chapter 6 explicitly as goddesses of the senses. However, the presentation of these annotations highlights a slight
discrepancy within the textual tradition. The second female associated with Vairocana's family brtan ma rdo rje is not here
linked with kun du/tu bzang mo as one might expect. Rather than an omission in the annotations, the most likely reason is that in
the Chapter 6 list of the sense goddesses associated with the five buddhas (see below f.16v), the principal female of the group, kun
du/tu bzang mo, is given rather than the second female. Thus, if the annotator had been following the Chapter 6 list, there would
be no further goddess to add to the main text here.
iv here, rdo rje dbyings seems to be equivalent to, or a minor error for, rdo rje dbyangs (or rdo rje dbyangs len ma). See Chapter 6
below, f.16v.1. The annotator (or copyist) also appears to have confused the placement of rdo rje dbyings/dbyangs and rdo rje
'phreng ba. rDo rje dbyings/dbyangs should be connected to Paravsin (Amitbha's family), and rdo rje 'phreng ba should be
placed with Mmak (Ratnasambhava's family).
28 rngam

zhing gtum: QtGtNt rang bzhin gdum (but note that the other versions of TZ agree with the Dunhuang manuscript here,
although the Bhutanese texts give rngam pa'i gtum.) | 29 ma 'pho: Gt ma'i pho | 30 yig: Gt yi ge | 31 lta bus: Gt bos | 32 myi: Gt mi
33 rin cen: Gt rin chen | 34 mtha': 'a subscribed | 35 myi: Gt mi | 36 las: Gt nas | 37 gtso mo: Gt gtso bo | 38 du: Gt tu | 39 Gt inserts:
dang | 40 pa'i: Gt ma'i | 41 ma ma ki: Gt m ma k | 42 bza': 'a subscribed | 43 sbyor pa'I: 'a subscribed; Gt omit | 44 byin rlabs: Gt
byin gyis brlabs pa | 45 Gt insert: rtsa gnyis dang sbyar | 46 Gt insert: ste/ | 47 phrag: Gt bcu drug | 48 na: Gt ni | 49 sum bcu: Gt
sum cu

250

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

mtshan bzang po50 sum bcu51 rtsa gnyis dang sbyar/

[beneath the line] thun mong dang thun mong ma yin pa 'o

sems dpa' bcu drug la/

dbu rgyan lta bu52 [6] rigs lnga lnga zhig 'byung ba ni/
spyir brgya53 cu ste dpe byed54 bzang po brgya55 cu dang sbyar56/
sems ma la la yang57 dbu brgyan58 yod mod59 kyi/
[beneath the line] sems ma la dbu rgyan myed [dpa?] zhe na

[14v.1] gnyis kyi byin brlabs60 las 'byung bas/

/dpe byed61 kyi skabs su myi62 bshad/
rtsis las 'das pa'i nang du ni gtogs63 so/
[beneath the line] yab yum gnyis ka 'i sbyor ba'i byin rlabs las byung bar 'dra ba'i phyir rigs lnga ma 'byung 'ga'64 yang myed +

/rig pa'i [2] byin rlabs las byung ba'i sems dpa'65 dang /
[beneath the line] + te sems ma rtsis las 'das gtogs so/ bcu drug las sum cu rtsa gnyis brgya' bcur bgrang ba ni gtso gzhung ste
bgrang

sems ma66 ma gtogs pa'i mtshan dang dpe byed67 ni/

[beneath the line, linked to mtshan dang...] sem mo zangs kyi mdog 'dra ba dang dbu mthon thing g.yas su 'khyil ba ni

'khor los bsgyur68 ba'i rgyal po la69 [3] yang yod de /

de ni sangs rgyas ma yin **** no/
/khro bo dang khro mo rnams ni70 mtshan **** ma 'i spyod yul71 la72 khro zhing gtum [4] pa ste/73
[beneath the line] dngos su [na?] rig pa'i sems nyid

rtag chad kyi mtha'74 la khro **** gtum zhes bya ba'i don to/
/pho nya ni **** 'gro ba myed75 par kun du76 phyin pa ste77 [5] chos kyi dbyings [Gt262] dang tha myi78 dad
pa'i don to/
'di dag 'bad rtsal myed79 par rang 'byung bas/
rang bzhin kyi80 dkyil 'khor zhes81 bya 'o/ /
[above line 6, bracketted] $// 'di man cad ting nge 'dzin kyi dkyil 'khor

[6] de la de82 ni thig le rdo rje yin/

[beneath the line] gnyis su myed pa'i byang chub kyi sems yab yum gnyis su myed [pa?]'i byang chub kyi sems

/'di zhes83 gang zag bdag myed pas84 /

/rtog85 par bya phyir ngas bshad de/
/mkhas pas de la 'byor par86 [15r.1] $/ /gyis/
/zhes gsungs pa ni87/
de la thig le ni chos kyi dbyings so /
/rdo rje ni lha dang lha mo 'i tshul du yod pa la bya 'o88 /
[beneath the line] blo rts[e] gcig pa 'o [beneath the line, linked to yod pa...] $/ rang gyi lus nyid lha 'i mtshan nyid de

[2] don du89 na rdo rje'i90 dbyings kyi dkyil 'khor zhes bya ba'i91 tha92 tshig go /
[beneath the line] sems 'khrug pa 'am gsal ba

/ting nge 'dzin kyi93 dkyil 'khor 'di bshad pas /

50 bzang

po: Gt omit | 51 sum bcu: Gt sum cu | 52 lta bu: Gt lta bu'i | 53 brgya: Gt brgyad (clearly, the text intends eighty; brgya cu
may be an archaism) | 54 dpe byed: Gt dpe byad | 55 brgya: Gt brgyad | 56 Gt insert: ro | 57 yang: Gt omit | 58 dbu brgyan: Gt dbu
rgyan | 59 mod: Gt omit | 60 kyi byin brlabs: Gt gyi byin gyis brlabs | 61 dpe byed: Gt dpe byad | 62 su myi: Gt sa; Nt su (no
negative) | 63 ni gtogs: Gt rtogs | 64 'ga': final 'a subscribed | 65 dpa': Gt omit | 66 ma: Gt las | 67 dpe byed: Gt dpe byad | 68 los
bsgyur: Gt lo bskor | 69 la: Gt omit | 70 Gt inserts: / | 71 spyod yul: Gt dpyod yul | 72 Gt inserts: / | 73 ste/: Gt de | 74 mtha': 'a
subscribed | 75 myed: Gt med | 76 du: Gt tu | 77 Gt inserts: / | 78 myi: Gt mi | 79 rtsal myed: Gt rtsol med | 80 kyi: Gt gyi | 81 zhes: Gt
ces | 82 de: Gt 'di | 83 zhes: Gt nas (but note that the TZ versions agree with the Dunhuang manuscript here.) | 84 myed pas: Gt med
pa | 85 rtog: Gt rtogs | 86 'byor par: Gt sbyor bar (but note that the other versions of TZ agree with the Dunhuang manuscript here,
and also that the Tenjur agrees with the manuscript's reading at the end of the TZComm chapter below (Gt262.4, Nt188.4). | 87 pa
ni: Gt te | 88 bya 'o: Gt bya'o | 89 du: Gt tu | 90 rdo rje'i: Gt rdo rje | 91 bya ba'i: Gt bya'i | 92 tha: smudged but still legible | 93 ting nge
'dzin kyi: Gt ting 'dzin gyi

Chapter 5

251

gang zag la bdag [3] myed94 par rtogs par bya ba'i95 phyir bshad do **** zhes bya ba 'i96 don to /
[beneath line 2, linked to gang zag...] $// nyan thos kyi gang zag la bdag +
[beneath line 3] +myed pa ni spangs pas thob ces zer ba ni 'di 'i skabs su spyad la shes rab zin pa yin no/

/rnal 'byor pas 'ba****d myi97 dgos pas98 lha dang lha mo 'i tshul [4] du ci99 ltar 'byung ba100 rtogs pa 'am /
****
/'byor par gyis shi[g?] pa 'i tha tshig go/
/ 101 thabs kyi zhags pa pad ma102 'phreng 103 [5] las/
ting nge 'dzin104 gyi dkyil 'khor gyi le'u ste105 lnga pa 'o// :

94 myed:

Gt med | 95 par bya ba'i: Gt pa'i | 96 bya ba 'i: Gt pa'i | 97 myi: Gt mi | 98 Gt inserts: / | 99 ci: Gt ji | 100 ba: Gt bar | 101 Gt
inserts: 'phags pa | 102 pad ma: Gt padmo | 103 Gt inserts: ba | 104 ting nge 'dzin: Gt ting 'dzin | 105 Gt inserts: /

Chapter 6
[Qt107b.2] [Gt262.5] [Nt188.5]
[Ms15r.5] //da ni gzugs brnyan gyi dkyil 'khor bstan [6] pa'i phyir /
/sems dpa' chen po1 spyod yul thams cad ni2 'di ltar rig par bya 'o3/
/zhes gsungs te /
chen po lnga yang4 de [15v.1] bzhin gshegs /
/5gcig la lnga lngar rig par gyis/
/khams gsum bdag nyid yin pa ste/
[beneath the line] rigs lnga 'am bdag gi lus

/sku gsum zhes kyang de la bya6/

[2] zhes gsungs pa ni /7
nam ka8 rnam par snang mdzad/ [Gt263]
/sa myi9 'khrug pa/
i mye10 rin cen11 'byung ldan/
chu12 snang ba13 snang ba mtha'14 yas/
[3] rlung gdon myi15 za bar grub pa 'o16/
***** /sa gcig po17 la yang sra ba dang /
[beneath the line, linked to sra ba] rdo rje

'du ba **** dang18 dro ba dang /

[beneath the line, linked to 'du ba and dro ba respectively] pad ma rIn cen

g.yo ba19 ste/

[beneath the line] a mo gha

de rnams [4] gzhigs na/

nam ka 'i20 bag la21 zha **** bas22 lnga lnga yod de /
[beneath the line] bhe ro tsa na

de bzhin du lhag ***** ma rnams la yang23 de bzhin24 rig par [5] bya 'o/
[beneath the line] chu dang mye dang rlung dang nam ka dang lnga lngar rig par bya 'o

/de dag25 ni sems dpa' 26 dang sems ma'i mtshan nyid do/
[beneath the line] lnga lnga nyi shu lnga 'o

lus ni gzugs kyi khams27/

ngag ni 'dod pa'i khams/
[6] yid ni gzugs myed28 pa'i khams/
lus ni long spyod29 rdzogs pa'i sku30/
sems ni chos kyi sku 31/
i

Note that the Tenjur texts have the placing of fire and water reversed. There are variations in the associations between the
elements and the buddha families but the Tenjur order fits better with the order of the list of qualities below (sra ba, 'du ba, dro ba,
g.yo ba) and it corresponds to one established convention. For instance, Klong chen pa's Phyogs bcu mun sel commentary on the
rGyud gSang ba'i snying po links the five consorts with the elements as follows: sa sangs rgyas spyan/ chu m ma k/ me gos dkar
mo/ rlung dam tshig sgrol ma/ nam mkha' dbyings kyi dbang phyug mar gnas gyur, bdud 'joms bka' ma Volume La: 65. It is most
likely that the Dunhuang manuscript has the error here, although not altogether certain. Associations of qualities with the buddha
families in Chapter 41 (Dunhuang ms. 82r.2-4) would seem to favour the Dunhuang reading here.
1 chen

po: Gt omits (note that the TZ versions agree with the Dunhuang manuscript on chen po here) | 2 Gt inserts: / | 3 bya 'o: Gt
bya'o | 4 yang: Gt la | 5 / /: Gt pa | 6 bya: Gt bya'o | 7 zhes gsungs pa ni /: Gt omits | 8 nam ka: Gt nam mkha' | 9 myi: Gt mi | 10 mye:
QtGt chu | 11 rin cen: Gt rin chen | 12 chu: QtGt me | 13 snang ba: QtGt omit (clearly, this is a dittography in the Dunhuang
manuscript) | 14 mtha': 'a subscribed | 15 myi: Gt mi | 16 pa 'o: Gt bo (but Qt gives pa'o) | 17 po: Gt pu | 18 Gt inserts: / | 19 g.yo ba:
QtGt g.yo zhing bskyed pa | 20 nam ka 'i: Gt nam mkha' | 21 la: Gt ma | 22 zha bas: Gt zhi bas | 23 yang: Gt omits | 24 Gt inserts: du
25 dag: Gt omits | 26 Gt inserts: dpa' (dittography) | 27 khams: Gt khas | 28 myed: Gt med | 29 long spyod: Gt longs spyod | 30 Gt
inserts: yin | 31 Gt inserts: yin

Chapter 6

253

ngag ni sprul pa'i sku 'o /

[16r.1] $/ /yang gsungs pa /
/ye shes lnga dang32 'gro ba lnga /
/mnyam ba33 so sor dbyer myed34 de /
[beneath the line] rnam par shes pa brgyad

/sngags dang phyag rgya35 dbang bskur [2] bas /

/spyod lam36 kun du man da la37/
/38zhes gsungs te /
de ltar mnyam ba39 nyid du rtogs pa ni /
de bzhin gshegs pa dang40 [3] bdag so so ma yin bar rig pas/ **
/sangs rgyas gud na41 spyan drang42 myi43 dgos/ ***
/ngag smra 'o chog44 sngags yin bas/
[4] snying po lhag par brjod du myed45 /
lus **** skyod46 do chog47 phyag rgya [Qt108a] yin bas /
phyag **** rgyas bsrung myi48 dgos te/
sngags dang [5] phyag rgya49 ni50 bdag nyid chen po 'i/51
rig pa'i ye shes kyi rtsal chen pos52 dbang bskur ro /
/ dper na dbang po lnga ni yab lnga
/ yul [6] lnga ni yum lnga ste / [Gt264]
lus ni be ro tsa na /
myig ni ag sho bhya53 /
/rna ba ni a myi dha ba54/
sna ni rad na sam bha ba /
lce ni a mo ga55 [16v.1] sid dhI56 /
/ ii gzugs thams cad ni rdo rje sgreg57 mo /
sgra thams cad ni rdo rje dbyangs /
dri thams cad ni rdo rje 'phreng58 ba /
bro59 thams cad ni [2] rdo rje gar ro /
/reg thams cad ni kun du60 bzang mo 'o61 /
tshogs zhes bya62 ba ni63 bdag gyi byin rlabs64 las65 thams cad 'dus shing66 tshogs [3] pa'i phyir tshogs shes
bya67/
dkyil zhes68 *** bya ba gang zhe na /
thams cad bdag gi sems *** las sprul pa'i phyir /
ii This

group of five goddesses feature in the maala of deities given in Chapter 7 of TZComm, with their mantras and mudrs
given in Chapters 9 and 10 of TZ respectively. The first is the second female in Akobhya's retinue, the second is the second in
Amitbha's female circle, the third is the second in Ratnasambhava's female retinue and the fourth the second Amoghasiddhi's
female retinue. We would expect Vairocana's second female, brtan ma rdo rje, for the fifth goddess, rather than kun tu bzang mo,
who does not occur in the list given in Chapter 7 and 9, but who is given as the principal female (Vairocana's principal female) in
Chapter 5 (14r.3), while her mudr is given in the root text of Chapter 10 (see below, f.37r.3). The associations of each of the first
three goddesses make good sense here rdo rje sgeg mo is the goddess of beauty, rdo rje dbyangs of music and rdo rje 'phreng ba
of flower garlands. They also match up appropriately with the list of the five buddhas given just above, except that here, the
goddess for Vairocana (body/touch) occurs in the fifth rather than the first position.
32 lnga

dang: Gt dang ni (note that the TZ versions agree with the Dunhuang manuscript here) | 33 ba: Gt pas | 34 myed: Gt med
Gt rgyas | 36 spyod lam: QtGtNt agree with spyod lam, but the TZ versions give spyod yul here | 37 man da la: Gt ma 'da'
la | 38 / /: Gt omits | 39 ba: Gt pa | 40 Gt inserts: / | 41 na: Gt nas | 42 spyan drang: Gt dpyad | 43 myi: Gt mi | 44 smra 'o chog: Gt
smras so cog | 45 myed: Gt med | 46 skyod: Gt bskyod | 47 do chog: Gt do cog | 48 myi: Gt mi | 49 rgya: Gt rgyas | 50 Gt inserts: /
51 /: Gt omits | 52 pos: Gt por | 53 myig ni ag sho bhya: Gt mig ni akobhya | 54 a myi dha ba: Gt a mi de ba | 55 a mo ga: Gt a mo
kra (Qt a mo ka) | 56 sid dhI: Gt si ti | 57 sgreg: Gt sgeg | 58 'phreng: Gt phreng | 59 bro: QtGtNt ro (In terms of the context of the
sense objects, QtGtNt's reading makes the best sense; the Dunhuang manuscript reading does not fit the passage, even if it seems
to fit with the goddess's name!) | 60 du: Gt tu | 61 'o: Gt omit | 62 zhes bya: Gt bye | 63 Gt inserts: / | 64 gyi byin rlabs: Gt gi byin
brlabs | 65 las: Gt omits | 66 Gt inserts: / | 67 shes bya: Gt zhes bya'o | 68 zhes: Gt ces
35 rgya:

254

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

ye shes thas [4] cad kyang69 chos nyid dag pa las sprul ***** pas/
chos dang sems70 rig pas mnya ***** par gyur pa la71 dkyil te sems la [5] bya 'o /
'khor gang zhe na /
thams cad du dbung72 mtha'73 myed74 par ye shes kyis75 khyab cing rdzogs par spyod pa ni 'khor zhes bya76/
/
[6] de lta bu ni77 tshogs kyi dkyil 'khor zhes bya 'o /
78rnam gcig du79 sgo lnga 'i rnam par shes pa rang thad du ma 'phros te / iii
ye shes su bltas [17r.1] $ / / pa ni /
rnam par dag pa'i thabs su rig pas /
kun du lta ba'i ye shes bya 'o /
[beneath the line] ye shes 'di dag lta bur cha 'dzin

rigs ni las kyi rigs so /

yid ni [2] rnam par shes pa yul la rang thad du ma 'phros te /
ye shes su 'phros pas /
rnam par dag pa'i thabs su rig pas /
[beneath the line] ye shes 'di thugs kyi cha 'dzin

mye long [3] lta bu 'i ye shes te /

rigs ni rdo **** rje 'i rigs so/
/nyon mongs pa can *** gyi yid yul la rang 'thad du ma [4] 'phros te /
ye shes su 'phros pa nI **** gzung ba dang 'dzin pa spangs pas / *****
rnam par dag pa'i thabs su rig [5] na mnyam pa nyid kyi ye shes te /
rigs ni rin po che 'i rigs so/
iii QtGtNt

[Qt108a.6, Gt264.5, Nt189.7] stop parallelling the Dunhuang manuscript here, and following an alternative version of
the next section, omit the remainder of Chapter 6, all of Chapters 7, 8, 9, and most of Chapter 10, including the entire extent of
Chapter 10 as given in the Dunhuang manuscript, apart from its final title. It is clear that an accidental omission has taken place,
because the chapter titles for Chapters 6, 7, 8 and 9 are all missing. Instead of the text of these chapters, the Tenjur versions give a
relatively short piece of alternative text which runs from Gt's 264-266. Although this text does not have any visible break, it in
fact consists of two passages which run into each other, but they are quite clearly distinct in content. The first section continues
the topic of Chapter 6, carrying on the present discussion which concerns applying the technique of recognising the complete
purity of the different aspects of the mind and consciousness as the five primordial wisdoms. About half of the Tenjur version's
lines on this topic occur within this next section of the Dunhuang manuscript's Chapter 6, but in a different order (apart from the
Tenjur version's additional lines, the parallel lines occur in the Dunhuang manuscript in the order 17r.5-6; 17r.2-5; 16v.6-17r.1).
An interesting feature of the Tenjur version of this section is that the passage coheres well, and along with the extra material, it
appears to represent some improvement on the parallel text we find in the Dunhuang manuscript. For instance, the Dunhuang
manuscript gives only four families, activity, vajra, jewel and tathgata, and the primordial wisdom linked with the activity family
is certainly non-standard (kun du lta ba'i ye shes, 17r.1). QtGtNt give the five families in more usual sequence, beginning with the
tathgata family, and giving the more standard, bya ba nan tan gyi ye shes, Gt265, for the activity primordial wisdom. Moreover,
the Tenjur version's beginning follows on from the preceding passage a little more coherently, with a seemingly less abrupt change
in topic, moving from the previous discussion of the meaning of dkyil 'khor, to introduce the next material as showing how all the
maalas arise. Furthermore, some corruption in the Dunhuang manuscript version at this point would appear to be indicated by a
further rather abrupt change in topic following its version of this passage, which might suggest that it lost some material at this
point. It is not possible to be certain which of the extant version of the passage is closer to their common ancestor, but it is most
likely that the divergence between the two versions was unrelated to the Tenjur version's major lacuna, most probably the result of
the loss of a large number of folios. What is clear is that in the middle of a verse line in the Tenjur, there is a sudden leap from the
topic dealt with in Chapter 6, to Chapter 10's theme of the ways to exhibit the mudrs of the deities in the peaceful maala. The
sixteen verse lines given in the Tenjur version before the Chapter 10 title (Gt265-266) run parallel to the final section of TZ's
Chapter 10 (as given in the South Central, Tawang and Bathang texts). The omission of the passage at the end of Chapter 10 from
the descendants of hypearchetype c, namely, the Dunhuang manuscript, Tshal pa Kanjur and Bhutanese NGB, represents their
indicative error, which could not have existed in the Tenjur version of TZComm, prior to the appearance of its great lacuna from
here in Chapter 6 up to the final part of Chapter 10.
69 kyang:
76 bya:

Gt omit | 70 Gt insert: dang | 71 la: Gt las/ | 72 dbung: Gt dbus | 73 mtha': 'a subscribed | 74 myed: Gt med | 75 Gt inserts: /
Gt bya'o | 77 ni: Gt na | 78 Gt insert: yang | 79 du: QtGt tu/

Chapter 6

255

[above the line, bracketed above bsod rnams] lus [rgya?]

/bsod nams kyi gzhi dag pa las chos kyi dbyings [6] dag pa'i ye shes su gyur te /
rigs ni de bzhin gshegs pa'i rigs so /
___________________________________________________________________
(As noted above, QtGtNt give an alternative version of the above lines before the major omission of material
up until the closing lines of Chapter 10, and in some ways this Tenjur passage appears more coherently
structured than the version in the Dunhuang manuscript. For comparison, we include this material here.)
[Qt108a.6] [Gt264.5] [Nt189.7]
yang rnam gcig tu/ sems dang yid dang / rnam par shes pa zhes bya ste/ 'di gsum du kun 'dus pas/ dkyil 'khor
thams cad 'byung ba'i rgyu yin no/ [Nt190] /de la sems kyi rang bzhin kun gzhi'i rnam par shes pa la/ bag
chags sna tshogs pa bsags pas gnas gyur te/ bsod nams kyi zhing dag pa ni/ chos kyi dbyings rnam par dag
pa'i ye shes te/ de bzhin gshegs [Gt265] pa'i rigs so/ /kun gzhi'i rnam par shes pa phyi nang gi 'jig rten la/
rgyun mi 'chad par rtogs shing sna tshogs kyi rang bzhin du 'jug pa nyid/ rang thad du [du: Nt tu] 'phros te/ ye
shes su 'phros pas rnam [Qt108b] par dag pa'i thabs su rig cing / gnas su gyur pa ni me long lta bu'i ye shes
rdo rje'i rigs so/ /nyon mongs pa'i yid bdag tu lta ba gnas su gyur pa ni mnyam pa nyid kyi ye shes te/ rin po
che'i rigs so/ /sgo lnga'i rnam par shes pa rang thad du ma 'phros te/ ye shes su blta bas su/[Qt omit /] rnam
par dag pa'i thabs su rig cing gnas gyur pa ni/ /bya ba nan tan gyi ye shes te las kyi rigs so/ /yid kyi rnam par
shes pa gzung 'dzin gyi dang [QtNt ngang] du rtogs pa can/ yul la rang thad du ma 'phros pas/ rnam par dag
pa'i thabs su rig cing gnas gyur pa ni/ padma ltar thabs kyi sgo sna tshogs kyis/ so sor kun tu rtogs pa'i ye
shes te padma'i rigs so/ /ye shes su ltas pas ni/ rnam par dag pa'i thabs su rig pas/ kun tu lta ba'i
(Here, QtGtNt move abruptly to the final section of Chapter 10.)
__________________________________________________________________
[Ms17r.6] /byang chub sems ni dngos grub mchog / iv
[17v.1] zhes gsungs te /
byang chub kyi sems ma gtogs par /
lus la 'od 'byung ba dang /
lha 'i mngon bar shes lnga dang /
nam ka dang brag [2] dang ri la thogs pa myed par rgyug cing /
chu la myi 'bying bar 'gro ba ni /
'jig rten gyi lha klu mthu bo che dang /
bdud can dag la yod de [3] de dag gyi sangs rgyas kyi dngos grub du **** myi chud de /
chos kyi dbyings nyid myi shes *** pa mang rab ces 'chaddo/
/the tsom [4] ldang ngo mchog chos kyi dbyings yin **** bas /
[beneath the line, under chos kyi dbyings] chos nyid kyi dus na

the tsom [bsal?]80 du myed do / ****

/sems kyi 'khrul pa las chos [5] gzhan na myed de /
'khrul pa nyid ma skyes pa 'o /
ma skyes ma g.yo chen po pa rnams kyis ni /
/gang zhig g.yo su zhig rtog [6] ces smra ba dang /
[beneath the line] myi g.yo zhe 'am ste myi rtog

myi g.yo zhes zer ba nyid mtshan ma nyid la chags su zad kyi /

iv

Although there is no suggestion here that this is a citation from a different text, in Chapter 8 (see below, f.30v), this tshig rkang
begins a highlighted verse attributed to the sGron ma brtsegs pa.

80 the

ba prefix is written with thick ink and it has a zhabs kyu beneath. Most probably the ba prefix is a correction of a previously
written letter meant to be deleted, perhaps du.

256

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

nam [-]81 sangs rgyas kyi bar du /

myi rtogs pa'i dbyings [18r.1] $/ /su chud myi srid do /
thabs dang shes rab du ldan zhing gnyis su myed par sbyor ba'i tshe /
mtshan kyi thad kar yi ge 'i 'bru rnas [2] bkod pa ni /
g.yon du bzhag ste bla ma'i mchod pa bya 'o /
82/sa ma ya stwa/ sa ma ya ho / ra go ya myi83 / ra go ya ham84 /
[3] zhes brjod de /
pra be shas85 / pad *****ma'i 'dab ma phral la /
rdo rje 'i rtse mor **** 86 /
pad ma'i dkyil du h dmyigs te [4] rten87 cing 'brel pas hm gsum kyis ***** bskyod de /
bsam gyis myi khyab par ***** 'od 'phro zhing 'phags pa la mchod [5] do /
/mchod pa ni rdo rje me tog /
bdug pa mar mye /
byug pa 'byung ste /
ci 'i phyir zhe na /
gsang ba'i rgyud 'di 'i skabs su /
[6] yo ga phyi pa dang myi 'dra ste /
me tog ni dkyil 'khor du dor ba yin bas /
gsang ba'i dkyil 'khor du byang chub kyi sems babs pa dang [18v.1] sbyor ro /
bdug pa ni byang88 chub kyi sems 'phro ba dang sbyar ro /
/spros pa ni byang chub kyi sems mye mar du gyur pas /
sku 'i rang bzhin [---]89 [2] gyis 'od gsal zhing gcig la gcig myi sgrib par gsal ba'i rang bzhin dang sbyar /
/byug pa ni chos nyid bde ba la reg pa'i tshe tshim [3] ba ste /
de lta bu 'i bde ba la sems can *** thams cad dgod pas /
sems can gyi don *** byed pa dang sbyar ro /
yang rnam gcig [4] du na sgreg mo90 ni 'dod chags kyi yu*****l /
'phreng ba ni mkhyud pa /
glu ni dga'91 ***** bar gyur pa /
gar ni spyod pa /
du pe ni [5] pog ste yid du 'ong ba'i yul lo /
/pus pe ni bskyod92 pa 'o /
di pa ni dga'93 ba 'o/
dri chu ni reg pa 'o /
sgo bzhi stod smad nas [6] mkhyud pa dang /
gzung ba dang /
myi shor ba dang /
rdo rje dang /
pad mor ldan ba ste /
rims bzhin no /
/'di bzhin du sbyor ba ni nang ngo /
[19r.1] $/ / dngos su mchod pa ni phyi 'o/
81

Insertion written beneath the line as correction to the text; the position is marked clearly with a cross above the line, but the
letter to be inserted is not clear, possibly ma or sa, although neither of these would seem appropriate additions here. Possibly par
or ka might be intended, if rnam par, or nam ka (=nam mkha') is the intended word. | 82 These mantras are included in the
Bhutanese NGB TZ version; variants in the Bhutanese texts are given below.) | 83 ra go ya myi: GGr ra ga ya mi | 84 ra go ya ham:
GGr ra ga ya ha | 85 pra be shas: GGr o pra be sha | 86 : GGr | 87 highlight not very clear; a little uncertain.
88 highlight not very clear; a little uncertain; it may rather be discolouration of the paper. | 89 A word of two or three letters is
deleted, probably through deliberate smudging in the original; while it is mainly impossible to read, the ending seems to have a
zhabs kyu and a final ra. | 90 sgreg mo = sgeg mo (note that there is an instance above of rdo rje sgreg mo [Ms16v.1], where the
Tenjur texts give rdo rje sgeg mo, and other instances elsewhere in the text.) | 91 dga': 'a subscribed | 92 ba prefix tiny, inserted
below | 93 dga': 'a subscribed

Chapter 6

257

pha/
yang na phyag rgya mo dang /
myos ma mo dang /
tshe 'phel ma 'o /
'di bzhi ni nang gyi lha mo [2] bzhI 'o /
/la sye ni phyag rgya mo ste mkhyud cing sbyor ba 'o /
ma le ni myos ma mo ste bde bar gyur pa 'o/
/nir te ni zab mo ste [3] byung ba za ba 'o /
'gir te ni tshe 'phel *** ma ste /
zos pas sku gsungs thugs kyi **** bdag po tshe dang yon tan dpag du [4]
myed par 'grub bo /
/yang de ltar **** byang chub kyi sbyor ba mdzad pa na /
byang chu*****b sems byung ba'i tshe /
ma skyes [5] pa'i ting nge 'dzin du bsgom zhing /
phyag rgya chen po 'i ting nge 'dzin gsal bar bya 'o /
/de nas sems 'phros par gyur na /
rdo rje phra mo [6] 'i 'du 'phro sgom /
yang na byang chub kyi sems las phyag rgya chen po 'dod chags kyi sbyor ba mdzad de /
dang po 'dod chags ldang ba'i sems [19v.1] kyang myi skyed /
tha mar zhi bar yang myi bsam /
byang chub kyi sems rang bzhin gyis bde ba chen po de kho na nyid du bsgom mo / /
[2] byang chub kyi sems ma skyes pa'i ting nge 'dzin du bsgom ba ni chos kyi sku 'o/
byang chub kyi sems rang bzhin gyis bde ba chen por gyur [3] pa ni /
bsod nams kyi sku 'o / *****
/rdo rje phra mo 'i 'du 'phro ni sprul pa'i sku 'o /
/sku gsum lhun gyis grub pa'i [4] sbyor ba ste /
gong ma'i lha mo rnams ***** kyang rang bzhin gyis dmyigs su myed pa'I94 **** tshul du rig par bya 'o /
/ 'phags [5] pa thabs kyi zhags pa pad ma 'phreng las /
gzugs brnyan gyi dkyil 'khor gyi le'u ste drug pa 'o //
:
//

94 pa'I:

'a subscribed

Chapter 7
[Ms19v.6]
[above the line] i $// don brgyad pa ii yid kyi dkyil 'khor ston

da ni gtso bo lnga dang gtso mo lnga 'i byin rlabs /

[beneath the line] $/ /'phags pa nor bdun kyi 'grel pa [rg]yu1 'i nor dang 'bras bu 'i nor ro

sems dpa' bcu drug dang /

sems ma bcu drug gi rtog pa bshad par bya 'o /
[20r.1] $/ /sems dpa' chen po zhes /
rdo rje sems dpa' la bos te bshad do /
rgyal ba'i rgyal ba rnams kyi sku /
[beneath the line] rgyal ba ni nyon mongs pa spangs pas rgyal ba rgyal ba'i nyon mongs pa spyad pas rgyal ba 'o

/sna tshogs [2] cir yang snang mod kyi /

/kun du rin cen lta bu ste /
/don dang las yin de bsgoms 'grub2 /
/zhes gsungs te /
[3] gtso bo nyi shu rtsa lnga dang /
gtso mo nyi *** shu rtsa lnga las stsogs pa rnams kyi **** sku ni /
snod kyi 'jig rten dang /
[4] bcud kyi 'jig rten du bsdus pa /
kha **** dog dang dbyibs cir yang snang mod kyi / *****
mdo tsam du rigs dbye na /
[beneath the line, continuing under line 5 below as indicated] glog [sgra?] bas brtsegs pa las rig pa gcig las rig pa sna tshogs + + su
'gyur ro kha dog kyang de bzhin du sbyar ro

de bzhin [5] gshegs pa'i rigs ser /

/rdo rje 'i rigs ni dkar /
rin po che 'i rigs ni dksngo/
/pad ma'i rigs ni dmar /
las kyi [6] rigs ni ljang ku ste /
/gsal zhing 'od 'bar ba la stsogs pa /
khyad bar du kha dog bzang ba dang /
dby[i?]bs legs pa la /
[20v.1] rin cen lta bu zhes bya ste /
bsgrub pa'i skabs su de ltar bsgom pa'i don /
lha dang lha mo 'i rnams kyi kha dog rnam lnga 'i don /
rigs lngar [2] sbyar ba ni /
las kyi bye brag gis /
phyag rgya 'i dbyibs kyang tha dad par snang ste /
/de bzhin gshegs pa'i rigs la 'khor lo / /
[3] rdo rje 'I rigs la rdo rje /
/rin po che *** 'i rigs la rin po che /

i There is no clear bracketting indicating that this annotation should be linked to line 6 rather than line 5, but the context would
seem to make it more appropriate that the point should be linked to the new chapter, while the need for a different annotation
beneath the line would have limited the available space, so that this comment was put above.
ii See Chapter 1 note iv above regarding the list of points (don) found in the annotations.
1 rg

is not clearly written but seems clear from the context. | 2 de bsgoms 'grub: the Tshal pa Kanjur texts give bsgoms pas 'grub,
and this reading is supported by most of the other TZ texts.

259

Chapter 7

/pad ma'i rigs ** la pad ma /

/las kyi rigs la [4] ral gyi ste /
/thabs dang shes rab ***** 'brel par bsgoms na /
dngos grub ***** rnam gsum yang rims kyis thob [5] par 'gyur zhes bya ba'i don /
/yang gzhan du na /
ye shes lnga dang /[...]3
/ye shes lnga 'i yul lnga 'i byin brlabs dang bcas pa /
[6] spyir ye shes nyi shu rtsa lnga 'i yul chos kyi dbyings mtshan nyid gcig pa la /
so sor snang ba de ltar nyi shu rtsa lnga yod de /
skad cig gcig4 [21r.1] $/ / la /
rig pa'i ye shes nyi shu rtsa lnga ste /
[beneath the line] byang chub kyi sems

/thabs dang shes rab 'brel pa dang /

[beneath the line, respectively under thabs and shes rab] ye shes

yum

/byang chub kyi sems su bcas te /

[beneath the line] thabs dang shes rab 'brel pas 'byung ba

[2] dus gcig du 'byung bas sku gcig du yang bshad la /

[beneath the line] mu las 'phro mu las 'dus pa

ye shes kyi byin brlabs grangs myed pa zhig du 'gyur zhing rig pas /
[beneath the line, under rig pas] rig pa'I rtsal gis [rigs?]/

sku grangs pa myed [3] pa sna tshogs su 'gyur te /

[beneath the line] rig pa'i rtsal gis bye brag gis go ba

de nyid gsa*****l zhing 'phro ba ni rin cen dang 'dra ste / *****
[beneath the line, respectively under de nyid gsal and rin cen] 'od gis gsal ba

du ma[r?]

/de ni don nyid do /

[beneath the line] [don?] dang bya ba'i 'brel pa rnal 'byor pas rtogs pas /

don la gzhol ba [4] dang /

gong ma ltar ye shes grangs myed ***** pa zhig 'khrugs pas / **********
[beneath the line] ye shes skad cig la rig pa'i don so sor snang ba ni 'khrugs shes bya ba yang na 'dzings shes bya ste gcig gis gcig
rig pa

/ don ma nor bar brtags nas [5] 'dzin pa ni las yin te /

[beneath the line] rnal 'byor pas rtogs pas bya ba'I don no [beneath line 5] rigs pa las kyi phyag rgya las yin bya ba'i 'grel pa

dngos grub kyang de nyid du thob bo zhes tan tra du ma las bshad do /
/dbu myig rna ba sna lce la / /
[6] zhes bya ba lnga bzhag [s?]te /
/ gtso bo lnga dang gtso mo lnga /
/sbyor ba'i tshul du bsgom par bya /
/zhes gsungs te / /
[21v.1] dbu la rnam par snang mdzad dang /
thams cad bdag nyid ma /
myig la rdo rje sems dpa' dang /
brtan ma rdo rje /
sna la rin cen rdo rje dang 'bar ma [2] rdo rje /
/ rna ba la chos rdo rje dang /
sdud ma rdo rje /
lce la las kyi rdo rje dang /

3 the

ink is smudged here; some letters have possibly been deliberately deleted. | 4 apparent highlighting over words in the final
lines of this side of the page is almost certainly the result of highlighting on the reverse, which has seeped through, affecting this
side of the folio inadvertently.

260

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

bskyod ma rdo rje /

rnams sbyor ba'i tshul du bsgom / /
[beneath the line] snang mdzad yab yum

[3] de 'i byin rlabs ni thams cad yin te *****

/nye bar bstan pa ni 'di rnams yin no / *****
iii /a ha a a a /5
[beneath the line] sems dpa' bzhi yang yum dang rigs bzhi sbyar

/gzhan du [4] na chos kyi sku 'i phyag rgya e gru gsu ****** las 'das nas /
[beneath the line, under e] thabs [beneath the line, under gsu] sku gsung thugs

/rig pa'i phya******g rgya mu la 'phyo bas las /

[beneath the line, from mu] myi gnas pa'I tshul de shes rab

[5] rang byung kyi ye shes kyis mu don gang yang ma yin ba'i ngo bor rig pa dang /
[beneath the line, from gang yang] $/ /rnam par snang mdzad dang thams cad bdag nyid ma ste myi dmyigs pa'i don no

/ mthong ba myed pa'i tshul du mthob6 dang / /

[beneath the line, under mthong] shes rab [beneath the line, from myed pa'i] $/ /dbyig gyi rnam par shes pa dang m gnyis su myed
pa ni 'di bcu las stsogs pa 'o

[6] thos pa myed pa'i tshul du thos pa dang /

[beneath the line] rna ba'i rnam par shes pa myi dmyigs pa'I tshul du thos pa dang thabs

/ci yang myed pa'i [drI?]7 [tshuor]8 ba dang /

[beneath the line] sna 'I rnam par shes pa dbang po lnga 'am rnam par shes pa lnga

/ro thams cad dang bral ba'i tshul gyis myong [22r.1] $/ /ba dang/
[beneath the line, under bral] shes rab [beneath the line, under myong] thabs
[above the line] lus kyi rnam par shes pa

/reg bya thams cad dang bral ba'i tshul gis reg pa dang /
[above the line] {myig

/ mthong bar byed pa dang /

[beneath the line, and continuing under the next tshig rkang, and down the right-hand side of the page] mthong bar myed pa'I tshul
kyis mthong bar byed pa po zhes kun la sbyar ro 'di man cad ni la stsogs pa [dbang?] po dang rims bzhin du sbyar ro
[above the line] rna ba

nyan par byed pa dang /

[2] tshor bar byed pa dang /
/myong bar byed pa dang /
reg par byed pa rnams te /
[beneath the line] lus kyi dpang po9

/bcu po de dag la bya ba'i don /

/de la chos thas [3] cad m du 'dus pas /
[beneath the line] shes rab du

/mu kyi****don rig pa ni /

[beneath the line] thabs yin de

/chos thams cad****rig pa nyid la yang snang la / /

[4] rig pa'i byin rlabs gzhan la yang****snang bas /

iii

Compare the syllable series in this chapter with those of the rGyud gSang ba'i snying po's Chapter 4. The Tshal pa Kanjur omits
these mantra syllables, and the similar sets of mantra syllables below (presumably assuming them to be a commentarial addition).

ha a a a: it is possible that a ha a a a is intended (the Bhutanese and other root texts give a ha a ha a), but there
does seem to be the sharp bend to the letter which distinguishes ta and da in this manuscript. | 6 mthob: mthong ba intended?
7 [drI?] is an insertion, placed below the line, with a cross above to mark the position | 8 tshuor: unclear with some smudging
which continues onto the following, ba. It appears that a correction of tshul to tshor was most probably intended; it is possible
that deletion of the ba was also intended. | 9 dpang po: dbang po probably intended.
5 a

Chapter 7

261

rig pa ni ste rnam par******snang mdzad do /

[beneath snang mdzad] dbu la

/m nyid thas [5] cad kyi de bzhin nyid du gcig pas /

[beneath the line, extending down the side of the page, the first part affected by some smudging of the ink] m la yul la [bya?] ba
kun la khyab pa'i phyir

/thams cad bdag nyid ma 'o /

/de la nI chos thams cad gang yang ma yin ba'i ngo bor [6] rig pa 'o /
[beneath the line, continuing on line 6] mdo sde las de bzhin gshegs pa rtag tu [sk]ye ba myed pa'I chos zhes bya ba dang mtshungs
pa ste rtogs pa 'am ++ chos thams cad bde bar gshegs dang 'dra zhes bya ba dang mthud

/m ni gang yang ma yin ba'i ngo bor rig pa 'o /

[beneath rig pa 'o] dbyings las bzhag go

/de nyid kyang /

mtshan nyid myed par rig pas na / /
[beneath the line] dang m so sor 'dug la de gnyis mnyam ste dbyar myed pa 'o
[above the line] don du m rgyu ba myed pa la bya

[22v.1] myi gnas pa'i sgrar 'phyo zhes bya /

/sbyor ba ni m las10 brten nas 'byung11 /
[beneath the line] thams cad bdag nyid ma rnam par snang mdzad

/ gyis m ni /
[beneath the line] be ro tsa nas chos nyid12

myi rig pa'i tshul kyis [2] rig par sbyor ba 'o /
/m dang so so ma yin bar rig pa 'byung ba ni yi ge lnga pa 'o /
[beneath the line, from 'byung...] gnyis su myed pa'i byang chub kyi sems

/ de la m gyi don myi 'gyur ba'i yongs [3] su grub pa ni /

[beneath the line] be ro tsa na chos nyid rtogs kyang rung ma rtogs kyang rung [beneath line 3, grub] dmyig

brtan ma rdo rje 'o / *****

/m la gzhol zhing lta ba ni sems dpa' **** rdo rje 'o /
[beneath the line] mi 'gyur ba la [beneath the line, under lta ba] rig pa

/m gis mtha'13 gnyis14 [4] bsregs pa ni 'bar ba rdo rje 'o / *****
[beneath the line] chos nyid kyi ngo bo nyid kyis mtha'15 dag pa yin bas

/m gyI don dbu ma phyin ci ma log pa'I ***** yongs su grub pa ni

[beneath the line, under grub pa ni] rigs pas rtogs pa

/rIn cen [5] rdo rje / 'o

/m du don bsdus pa ni sdud pa rdo rje 'o /
[beneath the line] ngo bo myed pa [beneath the line, under rdo rje] sna la

/de ltar rig pa ni chos rdo rje'o/

/m nyid du myi gnas pa ni bskyod ma rdo rje [6] 'o /
[beneath the line] 'jig pa myed ba [beneath the line, under rdo rje] lce la la

/myi gnas pa'i rjes su bsgrub pa ni las rdo rje 'o /

[beneath the line, under bsgrub pa ni las] rtogs par byed pa rig pa nI

de ltar m dang du 'brel pa'i skad/s cig gcig la / /

[beneath the line, under m dang ] thabs dang shes rab

[23r.1] $/ /nyi shu rtsa lnga phrag gnyis kyi lha dang lha mo 'i 'og nas 'byung ba dang /
dus gcig du gnas so zhes tan tra du ma las 'byung ngo / /
[2] dbu dang dmyig la stsogs pa ni /
/yan lag gi dam pa 'am spyi yin bas te /

10 las:

it appears that a line above the line on the final sa is indicating its deletion | 11 'byung: prefixed a chung small, inserted
gi gu not clearly visible, but it is present, on the zhabs kyu of the mantra syllable above. | 13 mtha': 'a subscribed | 14 gnyis:
smudged, a little unclear | 15 mtha': 'a subscribed

12 nyid:

262

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

yan lag gzhan de la rten pa bzhin du /

rnam [3] par snang mdzad dang /
/thams cad ****** bdag nyid ma'i 'khor la stsogs pa yang ***** don kyi gtso bo yin bas /
[beneath the line] [sang?]16 'og nas 'byung ba'i 'khor kun kyi rgyu yin no [beneath kyi gtso bo] kun la khyab pa'i phyir

dam pa [4] yin la /

lha dang lha mo thams cad ****** de las 'byung zhing rten pas spyi yin te / *****
[beneath the line] 'og nas 'byung ba [beneath the line, from las 'byung] snang mdzad dang bdag nyid ma 'o

/de bzhin gshegs pa'i rigs ni [5] dbu yin no /

/lag pa g.yas kyi sor mo la /
/h zhes bya ba'i lnga bzhag ste /
/sems dpa' lnga dang sems ma lnga / /
[6] sbyor ba'i tshul du bsgom par bya /
/zhes gsungs te /
skyes bu 'i lag pa g.yas pa ni thams cad du b[tsa?]n te /
[beneath the line] nus pa 'am byed pa'I don [beneath the line, from du, continuing under next tshig rkang] nus pa 'am skyon sel cing
byed pa'i don

skyon sel te [23v.1] yon tan len par byed do /

/de bzhin du rdo rje 'I rigs kyi lha dang lha mo yang /
bdud la stsogs pa rgol pa thams cad 'joms shing [2] dngos grub thams cad len pa'i phyir lag pa g.yas pa yin no
/
[beneath the line] nyon mongs pa dang sgrib pa dang nad dang sdug bsngal kun

/gung mo la myi bskyod pa dang /

[beneath the line] phyag na nyi ma'i dkyil 'khor ['dzan]17 yang na sbyor ba'i phyag rgya

/yangs pa'i spyan / /

[beneath the line] phyag na spyan 'dzIn

[3] mdzub mo la rdo rje sems dpa' dang / ******

[beneath the line] phyag na rdo rje dang dril bu yang sbyor ba'i rgya

rdo rje sgreg mo /

[beneath the line] phyag gnyis khu tshur dkur brtan

/srin lag la rdo rje rgyal***po dang /

[beneath the line] phyag na lcags kyu yang na sbyor ba'I [phyag rgya]18

rdo rje bdug pa ma /

[beneath the line] phyag na pog [p?]or19 yang na sbyor ba'i rgya

/mthe'u [4] cung la rdo rje 'dod pa dang /

[beneath line 4] phyag na mda' gzhu yang sbyor ba'i phyag rgya

rdo rje kyo ga ***** ma dang /

[beneath the line] phyag na lcags kyu yang na sbyor ba'i rgya

mthe bo la rdo rje legs pa dang ***** /

[beneath the line] phyag na rdo rje yang na sbyor ba'i rgya

rdo rje sgril ma rnams sbyor ba'i tshul [5] du bsgom /

[beneath the line] phyag na 'phreng ba 'dzin yang na sbyor ba'i rgya

/de dag gyi byin rlabs ni 'di rnams so /

16

[sang?]: most probably, sa is intended | 17 'dzan: 'dzin most probably intended; it could be that the rather large tsa 'phru is
intended to be a gi gu. | 18 phyag rgya uncertain due to the string hole, but most probable in the context | 19 pog [p?]or: pog phor
presumably intended

Chapter 7

263

/ka kha ga gha nga /

[beneath the line] myI skyod pa
[starting beneath skyod pa, apparently in continuation or as a further comment, and continuing onto line 6, as indicated by the
crosses] sems dpa' bzhi dang sbyor $/ /de ltar rna par snang mdzad mu las gnas par rig pa ste nyid [dang?] dang mu
du sbyor ba las byang chub ++sems 'byung ba ni myi skyod pa la stsogs pa 'o

/gzhan du na mye long lta bu 'i ye shes [6] de /

[above the line, bracketed above a 'i..]20 shes rab yangs pa'i spyan [bracketed above ma ste] thabs ste myi skyod pa

/a 'i don la shar pa ni ma ste /

/g.yo ba myed pas myi bskyod21 pa 'o /
[beneath the line, under myi bskyod pa] thabs myi skyod pa

/a nyid ni dngos po nyid las [24r.1] $/ /'das pas yangs pa ste /

[beneath the line, continuing onto 24r, under line 1, as indicated by the crosses] $/ /rnam par snang mdzad ni dbu la gnas par myi
rtog ++ myi bskyod pas ni rnam par snang snang22 mdzad dang [$?]m gi don mdzad pa yid byed pa'i tshul du mthong ba ni myi
g.yo ba'i tshul du rtogs pas dbyings ni ngo bo myed pas [a?] 'o de ltar rig pa ni ma ste myi skyod pa 'o

/myi bskyod23 pas g.yo ba myed pa mthong ba'i myig 'dzin pa'i phyir spyan no / /
[2] sbyor ba ni a las rten nas ma 'byung /
[beneath the line] yangs pa'i spyan dbyings las reg pa ni ma myi skyod pa myi skyod pa'i ngo bo myed pa ni spyan no

/mas a 'i don rig pa ni sbyor ba 'o /

[beneath mas] ye shes [beneath don] dbyings

/a dang ma so so ma yin bar rig pa [3] 'byung ba ni yi ge lnga 'o /

[beneath the line, continuing onto line 3, as indicated by the crosses] 'brel pa'i don myi skyod pa dang yangs pa'i spyan gyis sbyor pa
las ++ byang chub kyi sems byung ba ni 'di rnams so/

/ a dang ma 'I don so sor bkra bar sems *** dpa' ni rdo rje sems dpa'o / /
[beneath rdo rje] rtogs pa

[4] bkra ba ni gzhol zhing 'dud pa yin ***** pas rdo rje sgreg mo 'o /
[beneath the line] yid la byed pa 'am myi gting ba'i don

/a dang ma ***** 'i don gyis dngos po 'jom pa'i [5] zhe sdang ni rdo rje rgyal po 'o /
[beneath rgyal po] dngos po 'o

/bkra ba'i don gyis rgyun tshim par byed cing /

/zhe sdang bskyed pa ni rdo rje bdug pa 'o / /
[beneath ni rdo rje bdug] myi gtong zhing 'dod pa 'o

[beneath pa 'o] srin lag go

[6] a dang ma 'I don la gzhol ba ni rdo rje 'dod pa 'o /

/'dod pa grub nas bkra bar gyur pa nyid ni rdo rje kyo ga ma 'o /
[beneath ga ma 'o] mthe bo 'o

[24v.1] a dang ma 'i don bkra bar gsal bar chud pa ni /

rdo rje legs pa 'o /
/bkra bar gsal bra24 snang bar gsal ni rdo rje sgril ma 'o/
[2] yang ma ni dngos po myed pa'i sa yin te /
/dngos po ni sel cing /
dngos po myed pa ni len pas /
/rdo rje rigs ni lag la g.yas [3] pa 'o /
/yang mye long lta bu 'i **** ye shes ni ma skyes pa yin te /
[beneath the line, beginning under yin te and continuing under the next tshig rkang] dngos po myed par dang rig pa myed pa'i tshul
yin te

/dngos po **** myed pa'i sa la lta ba'i phyir ro / /

[4] sa de ni zhi ba yin ba'i phyir shan ting ***** zhes smos te /
20 the

bracketing in front of this and the following annotation appear to mark them off from the above line, linking them to line 6
below. | 21 bskyod: ba prefix inserted, tiny, beneath the line | 22 snang snang: sic. | 23 bskyod: ba prefix inserted, tiny, beneath the
line | 24 bra: bar, with ra subscribed? Perhaps the ra was added; initially, there might have been a transposition of bar here and the
ba following (note the deletion of its final ra).

264

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/zhi ba ni tshim par ***** byed pas /

/byams pa chen po [5] zhes kyang bya 'o /
/lag pa g.yon gyi sor mo la /
/tra zhes bya ba lnga bzhag te /
/sems dpa' lnga dang sems ma lnga / /
[6] sbyor ba'i tshul du bsgom par bya /
/zhes gsungs te /
skyes bu 'i lag pa g.yon pa ni /
/dgos pa thams cad stsogs
[25r.1] $/ /shing dam du 'dzin par byed do /
/de bzhin du rin po che 'i rigs kyi lha dang lha mo yang /
sems dpa' dang sems ma 'i mtshan nyid [2] rig pa'i don stsogs pa 'am /
/dam du 'dzin par byed pa'i phyir lag pa g.yon pa yin no /
/gung mo la rin cen 'byung gnas dang /
[beneath the line] phyag na rin po che 'i 'khor lo 'dzin yang na sbyor ba'i phyag rgya

[3] rin cen spyan /

/mdzub mo la rdo rje *****rin cen dang rdo rje 'phreng ba /
[beneath the line] phyag na rin po che 'i myu gu 'dzIn yang na sbyor ba'i phyag rgya/25
[beneath the line, under rdo rje 'phreng ba] phyag na na26 rIn cen 'phreng ba'

/srin lag la *** rdo rje gzi brjid dang /

[beneath the line] phyag na nyi ma 'dzIn yang na sbyor ba'i phyag rgya

/rdo rje me tog [4] ma/

[beneath the line] phyag na me tog yang na sbyor ba'i phyag rgya27

/mthe'u cung la rdo rje rgyal mtsha*****n dang /

[beneath the line] phyag na rgyal mtshan yang na sbyor ba'i phyag rgya

/rdo rje zhags pa ma /

[beneath the line] phyag na zhags pa yang na sbyor ba'I phyag rgya/28

mthe bo ***** la rdo rje bzhad pa dang /

[beneath the line] phyag na '[tshe]ms29 kyi 'phreng ba 'dzIn yang na sbyor ba'i phyag rgya

rdo rje bde ba'i [5] rnams sbyor ba'i tshul du bsgom /
[beneath, continuing under line 5, as indicated by the crosses] phyag na bsil yab ++ yang na sbyor ba'i phyag rgya

/de 'i byin brlabs ni 'di rnams yin no /

/ ta tha da dha na /
/'di bzhin du mnyam ba [6] nyid kyi ye shes a 'i don gnas te /
[beneath, continuing under line 6, as indicated by the crosses] mu dang 'brel pa'i mtshan nyid myed par ++ mnyam bar rig pa
nyid kyi e a ni rin cen spyan no / dang m a dang e 'i don mnyam ba nyid du rig pa ni dang m gyi sbyor ba las byang
chub kyi sems byung ba yin te e ni rin cen 'byung ldan no/

/a ni mnyam ba nyid de /
rin cen zhes bya 'o30/
a las byung nas a nyid la lta zhing [25v.1] gnas pa ni e ste /
/e ni rin cen 'byung nas so /
/e la gnas pa
[/] a ni rin cen spyan te /
mnyam ba nyid du rig pa'i [2] ye shes 'dzin pa'i phyir /
/rdo rje 'dzin ma zhes kyang bshad pa ste /
[beneath the line] gung mo la
25 long

ornamental shad, marking this annotation off from the next | rIn na na: sic | 27 rgya subscribed, small and a little unclear,
but probable from the context | 28 long ornamental shad, marking this annotation off from the next | 29 '[tshe]ms: possibly, 'choms
or 'tshoms, but 'tshems would seem more likely to be intended | 30 'o appears to be deleted, but the line may possibly have been
accidental.

Chapter 7

265

[above the line, bracketed above a las] yum

/sbyor ba ni a las brten nas e 'byung /

[beneath a las] dby[a(/i)]ngs [beneath the line, under e 'byung] rIn cen 'byung nas

/es a 'i don [3] rig pa ni sbyor ba zhes bya 'o / *****
[beneath line 2] rtogs pa [beneath line 3] dbyings kyi don

/e dang a so so ma yin bar yang rig ***** pa 'byung ba ni yi ge lnga 'o / /

[4] e dang a 'i don rig pa ni rdo rje rin *****cen no /
[beneath the line] rIn cen 'byung ldan dang rin cen spyan kyi sbyor ba las byang chub kyi sems 'phro ba ni 'di rnams so

/rig pa la bkra ba ni rdo rje 'phre*****ng ba ma 'o /

[beneath the line] tshig tsam ma yin ba la dran ba la

[beneath ma 'o] 'dzub mo

/a 'i don ngang gis mnya [5] ba nyid du es31 gsal bar rig pa ni rdo rje gzi brjid do /
/a 'i don mnyam ba nyid du gsal ba ni /
rig pa la mdzes pa'i phyir rdo rje me tog ma 'o [/]
[beneath the line] srin lag la lo

[6] a dang e 'i don gyis mnyam ba nyid du rig pas /

/bdag du lta ba thams cad bcom ba ni /
/nga myed pa'i nga rgyal te rdo rje rgyal mtshan no /
[26r.1] $/ /a 'i don e dang don so so ma yin pa rig nas 'brel pa ni rdo rje 'gas pa ma 'o /
[beneath 'gas pa ma 'o] mthe'u cung la 'o

/a 'I don gsal zhing mdzes par /

es rig [2] pa nyid ni rdo rje bzhad pa 'o /
/a 'I don e la gsal zhing mdzes par byed pa ni /
rdo rje bde ba ma 'o /
[beneath the line] mthe bo la 'o

[yang e?]32 ni mye ste bdag du lta ba thas [3] cad sreg cing mnyam ba nyid dam du 'dzin****pas /
/rin po che 'i rigs ni lag pa g.yo**n pa 'o /
/ mnyam pa nyid kyi ye shes [4] ni myi gnas pa yin te /
/bdag ***** du lta ba'i gnas bsregs pa'i phyir ro/ *****
/de ni mngon spyod yin te sreg pa'i [5] phyir ro /
/de ni gnas phyi mo dang sprod par byed pas /
snying rje chen po zhes kyang bya 'o /
/ rkang pa g.yon gyi sor mo la /
[beneath the line] 'di 'i dam tshig gi 'dod chags chen po 'o

/hri zhes bya ba [6] lnga bzhag ste /

/sems dpa' lnga dang sems ma lnga /
/sbyor ba'i tshul du bsgom par bya /
/ zhes gsungs te /
gung mo la snang ba [26v.1] mtha'33 yas dang/
[beneath line 6, and up the right-hand edge of the page] phyag na zla ba'i dkyil 'khor dang pad ma'i 'phreng gis bskor ba yin yang
sbyor ba'i rgya

gos dkar ma /
[above the line, above rdo rje dbyings len ma] phyag na hong kung yang sbyor ba'i rgya

mdzub mo la rdo rje chos dang rdo rje dbyings34 len ma /

[beneath the line] phyag g.yas kyi dang po g.yon kyi kha 'byed pa yang na sbyor ba'i phyag rgya

31 final

sa inserted, subscribed | 32 the letters are all clear but the separating tsheg is apparently missing | 33 mtha': 'a subscribed
is most probably an error for dbyangs, given in the repeat of the name below (27r.2). It is possible that a tiny line is
indicating deletion of the gi gu, but this is uncertain, since the annotation is written at exactly this place. Nonetheless, the apparent
error might represent a tradition in the Thabs zhags transmission; rdo rje dbyings ma seems to have been in the archetype of TZ in
the Chapter 10 mudr list.

34 dbyings

266

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/srin lag la rdo rje rnon po dang /

[beneath the line] phyag na ral grI 'dzin pa yang na sbyor ba'i phyag rgya

rdo rje mye sgron ma / /

[beneath the line] phyag na mye mar sgron ma yang na sbyor ba'i phyag rgya

[2] mthe'u cung la rdo rje 'khor lo dang /

[beneath the line] phyag na 'khor lo 'dzin yang na sbyor ba'i phyag rgya

rdo rje l[c?]ags sgrog ma /

[beneath the line] phyag na lcags sgrog yang na sbyor ba'i phyag rgya/

mthe bo la rdo rje [smab?]35 dang /

[beneath the line] phyag na rdo rje lcag kyu [d'dzin?]36 yang na sbyor ba'i rgya/

rdo rje gsal bkra ma rnams sbyor ba'i tshul du sgom /

[beneath the line] phyag na rma bya ba'i mdongs 'dzIn yang na sbyor ba'I phyag rgya

[3] 'di 'i byin rlabs ni 'di rnams so /

/pa pha ba bha ma /
/gzhan du na ***** so sor rtog pa'i ye shes kyis / /
[4] a 'I don la bltas na /
mtha'37 dang dbus **** myed de /
kun du mtha'38 yas so / /
***** de ltar snang ba ni bam ste /
[beneath the line, continuing onto line 5, as indicated by the crosses] m dang o la 'brel pa las mmyam pa nyid la sbyor bas rig pa
ni ba ste ++m dang gyi sbyor ba la byang chub kyi sems yin te de bas na snang ba mtha'39 yas so

/bam ni [5] snang ba mtha'40 yas so /

/a'i don ngang gis dag pa ni gos dkar ma 'o /
[beneath the line, under dkar ma 'o] gung ma la 'o

/sbyor ba ni a la brten nas bam 'byung /

[beneath the line, under sbyor ba] shes rab [under bam 'byung] bltos pa

/bam kyIs [6] a 'I don mtha'41 myi dmyigs par rig pa ni sbyor ba zhes bya 'o /
[beneath line 5] rtogs pas dbyangs kyis [beneath line 6, under sbyor ba zhes bya] ye shes chos nyid la

/a dang ba so so ma yin bar rig par byung ba ni yi ge lnga 'o / /

[beneath the line] snang ba mtha'42 yas dang gos dkar mo sbyor ba'i byang chub kyi sems pa 'phros pa'i rnams so

[27r.1] $/ /a dang bam kyi don gyis dngos po bstsal te /

/rig pa 'byung ba ni rdo rje chos so /
/dngos po bstsal te ma chags par go ba'i sgra [2] nI rdo rje dbyangs len ma 'o /
[beneath len ma] 'dzub ma43 la 'o

/a dang bam gyi don gyis dngos po myed pa nyid du chod pa ni rdo rje rnon po 'o /
/a 'i dngos po myed pa nyid du [3] gsal ba ni /
/rdo rje mye sgron ma 'o /
[beneath sgron ma 'o] srin lag la 'o

/as kun la khyab pas thams cad rang bzhin **** gyis mya ngan las 'das par rig pa rdo rje [4] 'khor lo 'o /
a dang thams cad 'bre****l pa nyid ni rdo rje lcags sgrog go / *****
[beneath sgrog go] mthe'u cung la 'o

/a dang bam gyI don rig pa ni brjod [5] pa byed pa ste rdo rje smra ba 'o /
/a dang bam gyI don nyid rig pa la bkra zhing gsal te /
/brjod pa byed ni rdo rje gsal bkra ma 'o / /
[beneath bkra ma] mthe bo la 'o

[6] yang na bam ni chu ste 'dus pa 'am sdud par byed pa yin bas brtan pa'i phyir /
35 smab?

Clearly written but uncertain. The likely intended word is smra ba, which occurs on the next side (27r.5). The mantra
for rdo rje smra ba (= Vajrabha) is given in the root text of Chapter 9, and his mudr in Chapter 10. | 36 d'dzin: clearly written;
perhaps dang 'dzin intended, or simply a da prefix in error? | 37 mtha': 'a subscribed | 38 mtha': 'a subscribed | 39 mtha': 'a
subscribed | 40 mtha': 'a subscribed | 41 mtha': 'a subscribed | 42 mtha': 'a subscribed | 43 'dzub ma: 'dzub mo intended?

Chapter 7

267

/ pad ma'i rigs ni rkang pa g.yon pa yin no / /

[27v.1] so so rtog pa'I ye shes ni 'gags pa yin note /
kun du mtha'44 yas par gnas pa'i phyirro/
/de ni dbang po yin te sdud pa'i phyir ro / /
[2] a dang bam gyi don bde ba chen po la gnas pa'i phyir /
dga'45 ba chen po zhes kyang bya 'o /
/skyes bu 'i rkang pa g.yon pa ni /
/stabs su mchongs [3] pa'i tshe /
brtan nas phul bar byed do /
/de bzhin du pad mo 'i rigs kyi lha dang lha mo **** yang /
chags pa myed pa'i tshul gyis [4] chags te /
[beneath line 3] shes rab [beneath line 4] thabs

/brtan par byed pas ***** rkang pa g.yon pa 'o /

/rkang pa g.******yas kyi sor mo la /
/a zhes46 [5] bya ba lnga bzhag ste /
/sems dpa' lnga dang sems ma lnga /
/sbyor ba'i tshul du bsgom par bya /
/zhes gsungs te /
skyes bu 'i [6] rkang pa g.yas pa ni /
mchongs pa'i stabs kyis gnas pha rol slebs par byed do /
/de bzhin du las kyi rigs kyi lha dang lha mo yang
[28r.1] $/ /grub pa thams cad mthar phyin par byed pas rkang pa g.yas pa 'o /
/gung mo la [d]on47 yod pa grub pa dang /
dam tshig sgrol ma /
[2] 'dzub mo la rdo rje las dang /
[beneath the line] phyag na byi shwa badzra48 yang na sbyor ba'i rgya

/rdo rje gar byed ma /

[beneath the line] phyag gar byed pa yang na sbyor ba'i rgya

/srin lag la rdo rje bsrung ba dang //

[beneath the line] phyag na rdo rje 'i phrub49 yang na sbyor ba'i rgya

rdo rje50/
[beneath the line, continuing under the rest of the intended tshig rkang] phyag na byug pa yang na sbyor ba'i rgya

byug pa ma /
mthe'u cung la rdo rje gnod sbyin dang [3] rdo rje dril 'khrol ma /
[beneath line 2] phyag na m[ch(/tsh)]e ba 'dzin yang na sbyor ba'i rgya [beneath line 3] phyag na dril bu 'dzin yang na sbyor ba'i
rgya

mthe bo la rdo rje ***** khu tshur dang /

/rdo rje 'bebs pa ma rnams ***** sbyor ba'i tshul du bsgom /
de 'i byin [4] rlabs ni 'di rnams so /
/tsa tsha dza ***** dzha nya /

44 mtha':

'a subscribed | 45 dga': 'a subscribed | 46 The highlighting might possibly have washed out here, but the paper gives no
hint of this. It would seem unlikely that highlighting of the rest of the verse but not these two words was intended; more likely, it
was omitted in error. | 47 [d?]on: da not very clearly written, but obviously intended. | 48 byi shwa badzra: probably viva vajra is
intended (in fact badzra appears more like, bdzu, but it seems certain that a ra btags, not a zhabs kyu, is intended). | 49 phrub: it
seems most likely that phub is intended. It is also possible that the mark which appears as an attached ra might be intended as a
deletion sign, deleting the zhabs kyu, thus, phab, although this would seem less appropriate. | 50 rdo rje: written small, perhaps
inserted. The following shad was most likely written first, so although not actually deleted, it is most probably not intended to
remain.

268

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

gzhan du na bya ba bsgrub pa'i ****** ye shes kyis a 'i don la 'byor [5] pa ni don yod par grub pa 'o /
[beneath line 4, continuing onto line 5 as indicated by the crosses] dang mu du 'brel pa'i mtshan nyid myed par rig pa dang es
mnyam bar rig pa dang bam gis so so ma yin bar rig pa ni a te dang m ++ du sbyor ba las byang chub kyi sems byung ba yin
te don yod grub pa 'o

a 'i don nyid dam tshig kyang yin la /

de la 'byor pas /
chos thams cad grol ba'i rang bzhin yang yin bas /
[6] dam tshig sgrol ma 'o /
[beneath sgrol ma 'o] gung mo la 'o

/sbyor ba zhes bya ba ni /

a 'i don la brten nas yang51 'byung /
/ya 'i don gyis a myi dmyigs [28v.1] par rig pa ni sbyor ba zhes bya 'o /
/a dang ya so so ma yin bar rig pa 'byung ba ni yi ge lnga 'o /
[beneath the line] don yod grub dang dam tshig bsgrol ma gnyis su myed pa 'di rnams so

/a dang ya so so ma yin bar [2] rig pa ni rdo rje las so /

/so so ma yin bar snang ba ni /
lta ba'i yul bas rdo rje gar byed ma 'o /
/a dang ya 'i don gyis dngos po [3] thams cad ma zin ba ni rdo rje bsrung ba 'o /
/a 'i don ya la reg pa ni rdo rje byug pa ma 'o/
[beneath byug pa ma 'o] srin lag go

/a 'i don dngos po las 'das par [4] rig pa ni /

'phags pa 'am / ******
dngos po zos pas /
rdo rje gnod sbyin no / ******
/a 'i don ni 'phags pa 'am zos [5] par snang bar go ba'i sgra ni rdo rje dril 'khrol ma 'o /
[beneath 'khrol ma 'o] mthe'u cung la 'o

/a dang ya 'i don 'brel par rig pa ni rdo rje khu tshur ro /
/a dang ya 'i don 'brel [6] par snang ba ni /
/rdo rje 'bebs ma 'o /
[beneath 'bebs ma 'o] mthe bo la 'o

/yang ya ni rlung ste /

/a 'i don la phyin par byed pas /
las kyi rigs ni rkang pa g.yas pa 'o
[29r.1] $/ /bya ba bsgrub pa'i ye shes ni /
myi dmyigs pa yin te /
a dang ya 'i don nyid ma 'grub pa'i phyir ro /
/de ni rgyas pa yin te [2] thams cad bdag du 'dus pa'i phyir ro /
/dmyigs par bya ba myed pas btang snyoms chen po zhes kyang bya 'o / /
'phags pa thabs kyi [3] zhags pa pad ma 'phreng las / *****
/lha 'i mtshan nyid dang dkyil 'khor bkod **** pa'i rim pa52 bstan pa'i le'u ste [4] bdun pa 'o //

51 yang:

ya intended? | 52 Here, the Dunhuang manuscript, Tshal pa Kanjur and Bhutanese texts omit some words found in other
versions of the title: /gtso bo lnga dang gtso mo lnga'i byin rlabs las/ /sems dpa' bcu drug dang/ /sems ma bcu drug gi rtogs
pa (compare TZComm's opening of Chapter 7, which describes the content of the chapter in similar words, Ms19v.6).

Chapter 8
[Ms29r.4]
// da ni dngos groub bsgrub pa'i rtog pa ****** bshad par bya 'o /
/sems dpa' [5] chen po nyon cig /
/zhes rdo rje sems dpa' la bos te bshad to / i
/rnal 'byor rigs 'dzin mar ldan bas /
/lnga zhing lnga [6] dang sbyor ba yi /
/ tshom bu lnga 'i lha bsgoms na /
/ thams cad bdag nyid chen por 'gyur /
/ zhes gsungs te / /
[29v.1] rnal 'byor pa dang /
rnal 'byor ma /
dkyil 'khor gsum kyi don rgyu 'bras dbyer myed par rtogs pas /
/ rigs lnga 'i 'khor nyi shu [2] rtsa lngar rnal 'byor pa rnams kyis bsgom zhing /
/yul lnga 'i 'khor nyi shu rtsa lngar rnal 'byor ma rnams kyis bsgoms la / /
[3] rnal 'byor ma rnams kyis dga'1 ba bskyed ***** nas /
nye bar bskul bar bya ste /***
ii / *** rin cen pad ma'i dkyil 'jam ni / /
[4] 'dzum ba'i mdangs kyis 'bul lag******s kyis /
/dgyes mchog rdo rje gsor ****** nas kyang /2
/byang chub sems sprin [5] dbang du3 gsol /
/zhes brjod de /
/thugs kar sen mos btab la /
[beneath the line, and continuing under the next tshig rkang] thugs kar sen mos btab pa ni gnyis su myed pa'i sbyor ba'i dga'4 ba 'dis
sangs rgyas thas cad kyi thugs te zab mo yin ba'i don

/ phyag rgya bstan par bya 'o /

/ rnal 'byor pha rnams [6] kyis kyang dga'5 ba bskyed nas phyag rgya sbyin bar bya ste /
/ rdo rje gsor pa'i dga'6 mchog ni /
/ 'dzum zhing 'khyil pa'i gnas su gsor /
[30r.1] $/ /myi dmyigs 'du7 dang 'phro ba ni /
[beneath the line] longs spyod rdzogs pa'i sku la sprul pa'i sku

/rdo rje 'dzin mar ldan las 'byung /

/ zhes brjod de /
thugs kar sen mos btab ste / /
[2] snying kar h /
lte bar /
rked par o/
rkang pa gnyis la sv h /
/dkyil 'khor bcu 'i sbyor ba gsal bar bya zhing /

i This

unhighlighted line of commentary has intruded into the Tshal pa Kanjur version of TZ.
verse and the following verse marked, zhes brjod de, have intruded into the Tshal pa Kanjur version of TZ (and these verses
appear to be recognised as lemmata by the Dunhuang scribe's highlighting).
ii This

1 dga':

'a subscribed | 2 /dgyes mchog rdo rje gsor nas kyang/: for this line, the Tshal pa Kanjur tradition gives, /dgyes pa'i rdo rje
rab gsor nas/ | 3 dbang du: the Tshal pa Kanjur texts give dbab par | 4 dga': 'a subscribed | 5 dga': 'a subscribed | 6 dga': 'a
subscribed | 7 'du: the Tshal pa Kanjur texts give bde; either reading could fit (the Dunhuang reading would seem perhaps
preferable)

270

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/ byang [3] chub kyi sems bab na a la la la ****** ho zhes dkyil 'khor thams cad la dgyes **** par mchod
nas /
[beneath a la la la] gnyis su myed pa'i don la shin du dgyes par

sprul pa mang po bsgo [4] zhing las rnam bzhi yang brtsam mo /
[beneath line 3] ba spu thams cad nas sprul pa 'gyed pa 'o

/sman la rigs thams cad kyi sngags ***** dang phyag rgyas rdo rje sems dpa' 'am/
[5] rigs kyi lha 'i rang bzhin du mos pa bskyed nas /
/snying ka dang dbang po thams cad la yang thig le bgyi /
/khong du 'thungs na [6] dngos 'grub mchog thob bo /
/ don du na chos kyi dbyings ngang nyid du /
/ ye shes lnga 'i don so sor 'dzin pa dang /
/chos kyi dbyings [30v.1] kyi ngang nyid rig pa'i rig pa la so sor snang ba ni /
/lha dang lha mo nyi shu rtsa lnga phrag gnyis te /
/ de ltar dbyings nyid dang8 rig pa'i gnyis [2] su 'byor pa'i byin rlabs las /
don ma nor bar rtogs pa'i rig pa 'byung ba ni /
/ byang chub kyi sems te /
rdo rje sems dpa' yin /
[3] des gzungs 'dzin las 'das nas / *****
/gnyis su myed pa m gi sa la snyoms par ***** 'jug go /
[beneath the line] m gi dbyings la 'dus pa'i 'khor yin la byang *** chub kyi sems ni de [na?]9 gnas par gsungs so

/de 'i phyir rdo rje sems [4] dpa' ni /

rdo rje 'dzin kyi sa la gnas ***** te /
sa bcu gsum pa zhes kyang bya 'o ******* /
/ 'di ni dngos grub mchog yin te [5] sgron ma brtsegs pa las kyang /
10/ byang chub sems ni dngos grub mchog /
[beneath the line, and continuing below for the whole of line 6 as indicated by the crosses] byang chub kyi sems ma gtogs par lus 'od
'byung pa dang mngon bar shes pa lnga dang brag la thogs pa myed pa dang chu la bying ba myed pa la stsogs pa rdzu 'phrul ni 'jig
rten gyi ++lha dang klu mthu bo che dang bdud sdig can la yang yod de de dag gis ni sangs rgyas kyi dngos grub du gal chud de
chos nyid kyi mtshan nyid ma rtogs kyi myi shes pa mang rab ces 'chad do de bas na byang chub kyi sems ni dngos grub gi mchog
go/

/ de rig rang 'byung nyid ces bya /

/ 'od dang mngon she[r(/s)] [6] lnga las stsogs /
/ ldan yang gnod sbyin chos myi shes /
/ zhes gsungs pa /
rdo rje sems dpa'i le'u las kyang /
/ stong la [31r.1] $/ /'byor pas rtag du dam pas 'byung /
/ bde gshegs bde mdzad bde bar mdzad pa ste /
/ rig pa'i skyes bu rdo rje sems dpa' nyid / /
[2] de na11 byang chub sems su rig pas rig /
/zhes gsungs te /
lta bu ste /
rnal 'byor rigs 'dzin mar ldan bas /
/ na bza'12 dang ni zhal [3] zas rnams /
/ grub pa'i rdzas su dbang skur nas /
/ gsol na dngos grub mchog **** du 'gyur /
8 dang

subscribed small, a correction in the original hand | 9 na? subscribed, small, in the original hand | 10 This verse and the one
below are marked off by the words, "zhes gsungs pa"/"zhes gsungs te" following, and by the highlighting. In this case, however, it
seems that they are highlighted because they are apparently citations from other texts, not because they are part of TZ. Note also
that the verse accidentally repeated below (f.33v.1-2) is not highlighted there, although nor is the following root text verse
highlighted either. | 11 de na: note that the accidental repetition below (f.33v.2) gives de ni. | 12 bza': 'a subscribed

Chapter 8

271

/ zhes gsungs te /
[4] rnal 'byor pa rnams kyis na bza'13 dang *****
/zhal zas la so so 'i rigs kyis ****** sngags brgya' yan cad 'bum [5] man cad phyag rgya dang bcas par bzlas
shing /
/ lha dang lha mo 'i ngang nyid dam /
byin brlabs su mos pa bskyed nas /
/ rnal [6] 'byor pa rnams kyis hu rnams la gsol zhing bstabs /
/ hu rnams kyis kyang gsol zhing bstabs na /
cis kyang myi tshugs [31v.1] pa dang rdo rje tshe la stsogs pa'i dngos grub thob bo /
/ gzhan du na chos nyid myi dmyigs ni sku 'o /
ngang nyid sna tshogs su snang ba ni bgo ba [2] yin te /
sku nyid rig pa la snang ba ni mas stabs pa yin te /
/rig pa nyid sku la blta ba ni /
hus bstabs pa yin te /
[3] rig pa de nyid kyi nang du dngos po th[o(/i)]m14 zhing **** bcom ba zos pa [nayin?] te /
de bzhin 'byo*****r pas dngos grub mchog yin no /
[4] zhes du ma las 'byung /
/rnal 'byor ***** rigs 'dzin mar ldan bas /
/ glu dang ***** tshig gi sgra rnams ni / /
[5] gsung gi ngo bor byin brlabs te /
/ bsgrags na dngos grub mchog du 'gyur /
/ zhes gsungs te /
rnal 'byor pa rnams kyis [6] glu dang tshig la stsogs pas bsgrags te /
/ gsung gi yan lag gis phyogs bcu 'i 'jig rten du go nas /
bdud ni zil kyis [32r.1] $/ /mnan/
rigs 'dzin thams cad ni mos pa 'phel bar bsgoms na dngos grub mchog thob tho /
/ gzhan du na lha dang [2] lha mo 'i dbyings dang /
so so ma yin bas kun du khyab la /
/ dbyings nyid rig pas thos pa'i sgra dang /
rig pa nyid dbyings la nyan [3] pa'i sgra dste /
/sgra des rig pa tshim****ba'i phyir glu 'o /
/lha dang lha mo 'i m*****tshan nyid brjod pa ni tshig yin te /
[4] de lta bur rtogs pa ni dngos grub m*****chog yin no /
sa [rba?] 'bu ta 'i don phyi ***** ma las bshad do / /
[5] rnal 'byor rigs 'dzin mar ldan ba'm /
/bro dang gar gyis bskyod byas nas /
/ 'od kyi 'du 'phror byin brlabs na /
/ dngos grub [6] mchog kyang thob par 'gyur /
/zhes gsungs te /
rnal 'byor pa rnams kyis /
lha dang lha mor bsgoms la /
/ bro dang gar gyis bskyod [32v.1] pa thams cad 'od gzer sna tshogs su 'phros pa las /
/slar 'khyil te /
lha'I tshogs dang bcas par 'bar nas 'od du gyur par bsgos na /
[2] dngos grub mchog thob par 'gyur ro /
/gzhan du na lha dang lha mo thams cad dus gcig du 'byor pa las /
/rigs kyang ma nges te /
[3] dus gcig du rdo rje yang rin po cher 'gyur /

13 bza':

'a subscribed | 14 It is probable that an original o has been amended to an i.

272

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/ de bzhin du 'khrug cing 'gyur ba dang / *****

/ de dag myigs15 pa ni 'od phung du 'bar [4] ba yin te /
/ de ltar rtogs pa ni ****** dngos grub mchog yin no zhes / *****
/ glog gru las 'byung ngo / /
[5] thabs kyi zhags pa pad ma 'phreng las /
/ dngos grub bsgrub pa'i le'u ste brgyad pa 'o // :
______________________________________________________
[The final page of the chapter is repeated (with an omission) on the verso side of folio 33. This sheet might
have been originally intended as 31r. The page begins, like a recto side, with an opening yig mgo, and the
first part is an almost exact repeat of the section which is found on 31r.1-4. It then jumps to the chapter
ending. It is quite likely that the scribe noticed the error of the omission and re-started the page. Later, so as
not to waste the paper, perhaps the text of 33r was written on the other side. We are giving the transcription
in indented format, to indicate that, although not clearly marked as such, it represents text repeated in error.]
[33v.1] $/ /'byor pas rtag du dam pas 'byung /
/ bde gshegs bde mdzad bde bar mdzad pa ste /
/ rig pa'i skyes bu rdo rje sems dpa' nyid / /
[2] de ni byang chub sems su rig pas rig /16
/zhes gsungs lta bu ste /
/rnal 'byor rigs 'dzin mar ldan bas /
/ na bza'17 dang ni [3] zhal zas rnams /
/ grub pa'i rdzas **** su dbang bskur nas /
/ gsol na dngos **** grub mchog du 'gyur /
/ zhes [4] gsungs te /
rnal 'byor pa
[The final section (continuing without any break or shad, presumably in error, perhaps due to an eyeskip to
'byor pa las/, which are the words preceding the following text) is an exact parallel to the Chapter 8 end
above (32v.2-5):]
rigs kyang ****** ma nges te /
dus gcig du rdo rje yang ****** rin po cher gyur /
/ de bzhin du [5] 'khrug cing 'gyur ba dang
/ de dag dmyigs pa ni 'od phung du 'bar ba yin te /
/ de ltar rtogs pa ni dngos grub mchog yin no zhes [6] glog gru las 'byung ngo /
/thabs kyi zhags pa pad ma 'phreng las /
dngos grub bsgrub pa'i le'u ste brgyad pa 'o // : //
_______________________________________________________

15

myigs: sic, dmyigs is surely intended. dMyigs is given in the alternative final page of the chapter, given on the verso side of
folio 33 (see below). | 16 This time, this verse is not highlighted. | 17 bza': 'a subscribed

Chapter 9
[Ms32v.5]1
i //

da ni rigs [33r.1] $/ / da ni rigs2 sngags3 kyi kyi le'u bshad par bya ste /

[beneath the line] sngags dang phyag rgya 'chad de

/ de nas bcom ldan 'das de dag gyi sku dang gsung dang /
thugs rdo rje las 'di dag phyung ngo /
[2] ba dzra b[o(/e)]4 ro tsa na /
[beneath the line] rnam par snang mdzad

/ sa twa ba dzra5 h /
[beneath the line] sems dpa' rdo rje

/ rad na ba dzra tra /

[beneath the line] rIn cen rdo rje

/ dhar ma ba dzra hri /

[beneath the line] chos kyi rdo rje

/ kar ma ba dzra6 ha /
[beneath the line] las kyi rdo rje

/ ba dzra a ksho [3] bhya h /

[beneath the line] myI 'khrug pa

/ ba dzra sa twa h / ****

[beneath the line] rdo rje sems dpa'

/ ba dzra r dz h /
[beneath the line] rdo rje rgyal po

/ ba dzra ra ga h / ****
[beneath the line] rdo rje 'dod pa

/ ba dzra ra ga h /7
/ ba dzra [4] sa dhu h /
[beneath the line] rdo rje legs pa

/ ba dzra rad na ***** sam bha ba tra /

[beneath the line] rdo rje rin cen 'byung8 gnas

ba dzra rad na tra9

[beneath the line] rdo rje rin cen

i This

chapter presents the Dunhuang version's understanding of the root text's mantra list. For further comments on these mantras,
see our TZ edition p.125-130.
1 This

page has only five lines. | 2 Note the dittography. If we are correct in the suggestion that the page with the alternative ending
of Chapter 8 (finished at the end of the sixth line) was written first before its error of omission was discovered, then it would make
sense that we begin here with the beginning of the chapter. Later, after the scribe realised the mistake, the final part of Chapter 8
was re-copied with the omitted passage, so that 32v ended up finishing on the fifth line. Perhaps the new chapter was then opened
in error at the end of that line. | 3 rigs sngags: rig sngags is almost certainly intended (see p.34 above). | 4 Presumably, be is
intended, although the vowel more closely resembles a na ro (most of the TZ texts give bai). | 5 sa twa ba dzra: the Tshal pa Kanjur
and Bhutanese TZ texts agree with the order here and in the following mantra, but the other root texts do not. | 6 kar ma ba dzra:
the Tshal pa Kanjur TZ texts agree with the order here, but the Bhutanese and other root texts do not. In the previous mantra,
however, most of the root texts agree with the order given here. | 7 Repeated, and the repetition does not occur in any version of
TZ; there is a very faint line perhaps crossing through the middle letters, presumably indicating deletion of the whole mantra.
8 'byung inserted, superscribed. | 9 ba dzra rad na tra: this is inserted in the handwriting used for the annotations, but it
appears to be an omission from the main text (this is confirmed by its inclusion in all the TZ editions, apart from the South Central
tradition which omits it in error). It is written just above the line, with the final syllable underneath at the same level as the rest of
the text. It is thus squeezed in as much as possible between the gap from the previous mantra and that following. The annotation
is then written in the usual place under the line.

274

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/ ba dzra su rya tra / ****

[beneath the line] rdo rje nyi ma

/ ba dzra ke 'u re10 tra / /

[beneath the line] rdo rje rgyal mtshan

[5] ba dzra ha sa tra /

[beneath the line] rdo rje bzhad pa

/ ba dzra a myi dha ba hri /

[beneath the line] rdo rje snang ba mtha'11 yas

/ ba dzra dhar ma hri /12

[beneath the line] rdo rje chos

/ ba dzra trig sha hri /

[beneath the line] rdo rje rnon po

/ ba dzra tsag kra hri / /

[beneath the line] rdo rje 'khor lo

[6] ba dzra pa she hri

[beneath the line] [rdo rje smra ba?]13

14/

ba dzra kar ma ha /

[beneath the line] rdo rje las

/ ba dzra ra ksha ha /
[beneath the line] rdo rje srung ba

/ ba dzra ya ksha ha /
[beneath the line] rdo rje gnod sbyin

/ ba dzra san ti dhi15 ha /

[beneath the line] rdo rje khu tshur

[33av.1] ii ba dzra a mo gha sid di ha /16

[beneath the line] rdo rje don yod grub pa

/ ba dzra shwa ra dha du /

[beneath the line] thams cad bdag nyid ma

/ ba dzra 'bu ta lo tsa na h /

[beneath the line] yangs pa'i spyan

/ ba dzra ma ma ki tra /
[beneath the line] rIn cen spyan

/ ba dzra pan dha ra [2] ba si ni hri /

[beneath the line] gos dkar mo

/ ba dzra sa ma ya ta ra h /
[beneath the line] dam tshig sgrol ma

ii

Unnumbered folio following 33r. As mentioned above (p.32), we are referring to it as f.33a, and calling this the verso side
(neither side has pagination) since the other side gives the opening yig mgo. However, this side follows on from 33r, and
completes the chapter. The reverse "recto" side commences with the first three mantras given here, but then is not completed.
Possibly, the scribe realised there were a few spelling errors and decided to discard the side, re-copying and finishing the page on
the verso side.
10 ba

dzra ke 'u re: here, the Tshal pa Kanjur TZ texts give a different mantra, badzra dhwadza in J, but Vajraketu is probably the
intended deity (see TZ edition p.126). | 11 mtha': 'a subscribed | 12 The Tshal pa Kanjur texts omit this mantra, presumably in error
(the Bhutanese and other root texts include it, and we would expect Vajradharma's mantra at this point in the sequence, as the first
male in Amitbha's retinue; see Chapter 7 above, f.26v.1.) | 13 [rdo rje smra ba?]: written very faintly, but smra ba is reasonably
certain (and rdo rje most probable). | 14 here, we would expect, ba dzra a mo gha sid di ha, for the mantra of Amoghasiddhi to
start this group before the mantras for his retinue (listed in Chapter 7 above, f.28r): this mantra has been displaced to the end of the
group of his retinue, and given at the beginning of the next folio (see below). | 15 ba dzra san dhi: here, the Tshal pa Kanjur texts
give a different mantra, badzra mui, but Vajrasandhi is probably the intended deity (see TZ edition p.127). | 16 ba dzra a mo
gha sid di ha: this mantra is displaced; it should occur above, before ba dzra kar ma ha.

Chapter 9

275

_________________________________________________________________
[Note that we are missing the mantras for the four female deities of Vairocana's circle, which should come at
this point. It seems that they have been lost in the Dunhuang, Tshal pa Kanjur, and Bhutanese NGB
transmission. They might have been lost by hypearchetype c, or prior to that, by hypearchetype b (see our
stemma codicum, p.43 above), but this cannot be ascertained since the Tenjur has a major lacuna which
includes this chapter. The mantras are: /o sti ra badzra o/ /o dzwa la badzri o/ /o sang ha ra badzra
o/ /o tsad tsa la badzra o/ (see our TZ edition.)]
_________________________________________________________________
/ ba dzra la sye h /
[beneath the line] rdo rje sgreg mo

/ ba dzra du pi h/
[beneath the line] rdo rje bdug pa ma

/ ba dzra ang ghu sha h /

[beneath the line] rdo rje lcags kyu ma

/ ba dzra dwa ti [3] shwa r[a(/e)]17 h /

[beneath the line] rdo rje sgril ma

ba dzra dwa deu shwa re h18

[beneath the line, along the edge of the left hand side of the page] rdo rje sgreg mo

/ ba dzra ma le tra /

[beneath the line] rdo rje 'phreng ma

*** ba dzra pu spe tra /

[beneath the line] rdo rje me tog ma

/ ba dzra pa sha tra /

[beneath the line] rdo rje zhags pa ma

**** / ba dzra rad na ra ti19 tra / /

[beneath the line] rdo rje bde ba ma

[4] ba dzra 'gir ti hri /

[beneath the line] [-]20 rdo rje dbyangs len ma

/ ba dzra a **** lo ke hri /

[beneath the line] rdo rje mye mar ma

/ ba dzra spo ta hri /

[beneath the line] rdo rje lcags sgrog ma

/ **** ba dzra ra ga ra ti hri /21

[beneath the line] rdo rje gsal bkra ma

/ badzra nir [5] ti hria /

[beneath the line] rdo rje gar byed ma

/ ba dzra gan te hria /

[beneath the line] rdo rje byug pa ma

17 There

is some smudging above the letter, so it is unclear whether a vowel was originally present or not. | 18 This line is inserted
small as a correction on the left margin, vertically, and marked by a cross; its place in the text is indicated by a matching cross.
The line runs from top to bottom of the text. There is no highlighting. It seems simply to represent a correction for the previous
mantra not a new mantra (and this is confirmed by the single mantra with variants on this given in the root texts). The extra
label with the Tibetan name (rdo rje sgreg mo) seems inappropriate and is most probably in error for rdo rje sgril ma, as given
beneath the originally written mantra (see also the sequence given in Chapter 7 for Akobhya's retinue, f.24r-24v). | 19 ba dzra rad
na ra ti: here, the Tshal pa Kanjur and Bhutanese TZ texts have variants on this mantra, but the other root texts give a different
mantra, badzra su ki (see TZ edition p.129). | 20 one syllable here is smudged out, almost certainly deliberately deleted. | 21 The
Tshal pa Kanjur texts omit this mantra, presumably in error (the Bhutanese and other root texts include it, and we would expect it
at this point in the sequence, as for the final female in Amitbha's retinue; see Chapter 7 above, f.26v.)

276

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/ ba dzra gan dh[a(/e)]22 hria /

[beneath the line] rdo rje dril 'khol23 ma

/ ba dzra ra ti ha /
[beneath the line] rdo rje 'phebs ma
iii / ba dzra mu sti h /
[beneath the line] brtan ma rdo rje

/ ba dzra mu sti tra /

[beneath the line] 'bar ba rdo rje

/ ba dzra mu [6] sti hri /

[beneath the line] bsdud ma rdo rje

/ ba dzra mu sti ha /
[beneath the line] bskyod ma rdo rje

/ thabs kyi zhags pa pad ma 'phreng las sngags kyi le'u ste dgu pa 'o // : //
___________________________________________________
[The recto of unnumbered folio 33a contains the following text, a version of the opening three mantras given
on 33av. As with the repeated ending of Chapter 8, we are giving the transcription in indented format, to
indicate that, although not clearly marked as such, it represents text repeated in error.]
[33ar.1] $/ // ba dzra a mo k sid dhi ha /
/ ba dzra shwa ra dha ru /
/ ba dzra 'bu ta li tsa na h /
[As noted above (note ii), the probability is that the scribe began on this side, but for some reason (a few
errors?) discarded the side, re-copying the mantras and finishing the page on the verso side.]
____________________________________________________

iii Note

that the annotator takes the final four mantras to represent the four female deities of Vairocana's circle, as listed in Chapter
7 above (see f.21v, f.22v). However, these final mantras of the chapter do not bear any relation to their names. In our TZ edition
(see p.128 note 41 and p.130 note iii), it will be seen that the mantras for these deities do in fact occur in the expected position
following the group of the principal female deities, but that they have been lost in the Dunhuang, Tshal pa Kanjur, and Bhutanese
NGB transmission (see also the discussion in our TZComm edition above, at f.33av.2). Presumably, the annotator noticed their
omission and assumed the final four mantras to represent these goddesses.
22

dh[a(/e)]: there is a clear 'greng bu but it is smudged. Its deletion may be intended, but the smudging may be due to the
smudging deleting the vowel on the following letter. | 23 dril 'khol: dril 'khrol intended?

Chapter 10
[Ms34r.1]
$/ /da ni phyag rgya 'i le'u bshad par bya ste /
[beneath the line] sngags dang phyag rgya 'brel bar bshad de

/sems dpa' chen po nyon cig /

/zhes rdo rje sems dpa' la bos nas bshad do / /
[2] lha dang lha mo 'i ngang nyid du /
/ yan lag lnga la 'byor pas na /
/ bskyod pa thams cad phyag rgya yin /
/ de skad dpa' bos bka'1 stsal to /
[3] zhes gsungs te /
/gong du bshad pa'i tshul** du bsgoms na /
/lus kyi yan lag thams*** cad phyag rgya yin ba'i don / /
[4] de la phyag rgya 'i mtshan nyid ni // *****
//thabs dang shes rab ngang nyid du / *****
/lag gi 'du byed bsgoms nas ni //
[5] thal mo dam sbyar rtse bsnol gshibs /
/ rdo rje thal mor bka'2 stsal to /
/ sor mo nang du bsnol nas bkan /
/gung 'dzub mthe bo [6] mthe'u cung rnams /
/ so sor rtse sprad phyir bkyed de /
/srin lag gshibs bsgr[e?]ng3 dam tshig rgya / i
/rkang sor g.yas kyis g.yon pa [34v.1] gnan /
/lag mthil g.yas kyis g.yon pa gnan /
/mthe bo rtse sprad pad du blang4 /
/rnam par snang mdzad rgya cen yin / /
[2] rdo rje khu tshur bcangs pa la /
/g.yas kyi 'dzub mo g.yon gyis gzung /
/tshogs pas gci5 la gcig bzung ba' /
/ 'di ni sbyor [3] ba'i rgya mchog go /
/mthe bo gnyis dang mthe'u cung gnyis /
/ so sor rtse sprad zlum *** por ste /
/ lhag ma6 'od 'phro rting [4] phye ba /
/myi bskyod pa'i phyag rgya yin *****
/rting sprad gung mo rtse sbyar te / ******
/mdzub mo tshigs bcag gung la sbyar7/
[5] srin mtheb rtsa sprad mthe'u cung swa' /
i

Klong chen pa cites this verse in his Phyogs bcu mun sel (bdud 'joms bka' ma Volume La: 279-280; see our TZ edition, Chapter
10 note ii).
1 bka':

'a subscribed | 2 bka': 'a subscribed | 3 bsgr[e?]ng is a correction placed below the line; a previous word of a single syllable
above it has been deleted, and the 'greng bu is uncertain (it could simply be part of the deleted syllable), but is most likely here,
and agrees with the Tshal pa Kanjur, Bhutanese, and most of the other root texts. Compare IOL Tib J 331.III, 5r.4: "gung mo
gnyIs gshIbs te bsgreng ba 'I bar du phur pa gzung la", "(you) hold the phur pa between the two middle fingers [which are]
positioned upright". | 4 pad du blang: the Bhutanese texts read, 'phang du blang, here (most other versions agree with the
Dunhuang reading). | 5 gci: presumably, a scribal error for gcig given in the other versions of TZ. | 6 lhag ma: the Bhutanese texts
read, lha mo, here (most other versions agree with the Dunhuang reading). | 7 sbyar: the Bhutanese texts read, gzhar, here (other
versions agree with the Dunhuang reading).

278

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/rin cen 'byung gi phyag rgya yin /

/gung 'dzub srin lag rtse sprad de /
/mthe'u cung ngos la gzhag [6] par8 bya /
/mthe bo rtse sprad zlum por te /
/snang ba mtha'9 yas phyag rgya yin /
/thal mo sbyar nas sor mo dgrad /
/bskyod nas [35r.1] $/ /'khor lo ltar byas te /
/'glog gi10 'khor [ba?]r11 bya /
/ don yod grub pa'i phyag rgya yin /
/ phyag rgya'i rgyal po 'di dag ni /
[2] thugs kar 'chang12 zhing nyid du mos /
/'du 'phro sgrol13 zhing mkha' la dgrol /
/las dang thun mong nyid du bshad /
/rdo rje dril bu bzung [3] nas ni /
/thugs ka dang ni bkur ***** brtan nas /
/ gcu thabs zur kyis lta ***** ba ni /
/ rdo rje sems dpa'i rgya cen yin14 [4] lag pa bcsnol15 nas khu tshur bcang *****
/gung mo gnyis ni brkyang bar bya / *******
/'dzub mo lcags kyu ltar byas pa /
[5] rdo rje rgyal po 'i rgya cen yin /
/ lag pa khu tshur bcangs pa las/
/ g.yon gyi mtheb 'dzub rtse sprad de /
/ g.yas kyi gung mo [6] brkyang nas gdang /
/ rdo rje 'dod pa'i rgya cen yin /
/ lag pa gnyis kyi sor mo rnams /
/ so so 'i gung mo 'i tshigs dang por /
[35v.1] ngos bzhi nas bzhi ngosnas16 ni rtse mo gzugs /
/rdo rje legs pa'i rgya cen yin /
/rting sbyar gung mor rtse sprad de /
/mtheb mdzub srin [2] lag tshigs bcag nas /
/ drug car rtse sprad mthe'u cung s[b(/w)]as17 /
/rdo rje rin cen rgya cen yin /
/ mthe bo mthe'u cung rtse sprad nas / /
[3] lu gu rgyud du sbrel bar bya /
/ lhag ***** ma 'od ltar byas pa ni /
/ rdo rje nyi ma'i ***** rgya cen yin /
/ lag ngar g.yon pa [4] bsgreng nas ni /
/ rdo rje khu tshur bcangs ***** pa la /
/ g.yas pas gru mo gzung bar **** bya /
/ rdo rje rgyal mtshan rgya cen yin18 /
[5] g.yon kyi steng du g.yas bzhag nas /
/ so so 'i tshigs bcags rtse sprad de /
/ kha 'i thad kar gzung bar bya /
8 gzhag

par: the Tshal pa Kanjur texts read, sbyar bar, here (most other versions agree with the Dunhuang reading). | 9 mtha': 'a
subscribed | 10 The Dunhuang manuscript omits 'phreng (or phreng) bas, found in all the TZ editions. | 11 'khor [ba?]r: an original
'khor lor appears to have been corrected to 'khor bar, found in all the TZ texts apart from the Bhutanese, which give bskor bar.
12 'chang: most of the root texts give bcing. | 13 sgrol: the Tshal pa Kanjur and Bhutanese TZ texts agree with sgrol but most of the
other root texts give sgom. | 14 the expected shad is found in the other versions of the text. | 15 deletion of ca (presumably written
in error) indicated by a line above. | 16 the repetition of bzhi nas is presumably in error; it is not found in the other versions.
17 s[b(/w)]as: sbas is found in most of the other versions. | 18 yin abbreviated into a stack

Chapter 10

279

/ rdo rje bzhad pa'i rgya cen [6] yin /

/g.yon gyi sor mo 'i rtse rnams ni /
/[ma phrad?]19 rtsam du bsdus pa las20 /
/g.yas kyis 'byed pa ltar byas ste / /
[36r.1] $ / / rdo rje chos kyi rgya cen yin /
/g.yas kyi lag ngar bsgreng ba las /
/mthe bo mthe'u cung [sprad?]21 de/
/lhag ma gsum [2] yang bsgreng bar bya /
/g.yon pa'i khu tshur dkur brtan ba /
/rdo rje rnon po 'i rgya cen yin /
/lag pa rgyab sbyar22 rtse bsnol nas / /
[3] mthe bo gnyis kyang rtse sprad de / *****
/bskor ba lta bur byas pa ni /
/rdo rje 'khor **** lo 'i rgya cen yin /
/sor mo nang du [4] bsnol pa las /
/gung mo rtse sprad ***** mdzub mo gzhar23 te /
/ mthe bos gung ***** mo gnan par bya /
/ rdo rje smra ba'i [5] rgya cen yin /
/ lag pa'i khrig ma bsnol ba las /
/gung ma gnyis kyi tshogs24 dang por /
/ ngos bzhi nas ni sor mo brtan / /
[6] rdo rje las kyi phyag rgya yin /
/ lag pa gnyis kyi sor mo rnams /
/ phan tshun thags [pasu?]25 btags pa las /
/ mthe bo [36v.1] gnyis ni bsgreng bar bya /
/ rdo rje bsrung ba'i phyag rgya yin /
/ rdo rje khu tshur bcangs pa las /
/ mdzub mo gnyis ni bsgreng nas kyang / /
[2] kha 'gram gnyis su gzung bar bya /
/ rdo rje gnod sbyin rgya cen yin /
/ rdo rje khu tshur bcangs pa las /
/ bsnol te kha ni sbyar nas kyang /
[3] thugs kar dam du gzung bar bya / *****
/rdo rje khu tshur rgya cen yin /
/rigs kyi ***** phyag rgya bsgoms nas ni / /
[4] lag gi 'du byed 'di dag gis / *****
/'du 'phro mos pa bskyed nas kyang / *****
/bcing zhing mkha'26 la dgrol bar bya /
[5] mthe bo mthe'u cung nang du sprad /
gung mo rtse sbyar bzhi yang bsgreng /
/ rdo rje ltar byas27 rting sprad pa /
/ rdo rje 'dzin pa'i28 phyag rgya yin / /

19 [ma

phrad?]: ma phrad is presumably intended, but the ma and pha are positioned close together, and a tsheg after ma (perhaps
inserted small as a correction) is barely visible (ma phrad is found in most of the other versions). | 20 las: sa subscribed
21 [sprad?]: the other versions give rtse sprad here. | 22 rgyab sbyar: the Tshal pa Kanjur texts give rgya gram here (other versions
agree with the Dunhuang reading). | 23 gzhar: the Bhutanese TZ texts agree with gzhar but most of the other root texts give sbyar.
24 tshogs: most probably an error for tshigs given in the other texts. | 25 [pasu?]: possibly, a correction from an original pas into su
(the reading in the other texts) is intended, but there is no indication of deletion. | 26 mkha': 'a subscribed | 27 rdo rje ltar byas: rdo
rje 'dzin ltar is given in the other texts | 28 pa'i: the Tshal pa Kanjur texts agree with pa'i, but the other texts give ma'i, which
seems more appropriate (see TZ edition p.139).

280

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

[6] rting sbyar [mthebu?]29 mdzub rtse sbyar te /

/ mdzub mo 'i myig du mthe'u cung drang /
/ lhag ma rtse gsum ltar byas pa /
/ rdo rje spyan kyi rgya [37r.1] $/ / cen yin /
/ lag [cngar?]30 g.yon pa bsgreng nas kyang /
/ sor mo ud pa la ltar la31 /
/ g.yas kyi gru mor gzung bar bya //
[2] rdo rje sgril ma'i32 phyag rgya yin /
/ rgang pa33 gnyis kyi pus mo dang /
/ lag pa gnyis kyi khru mo rnams /
/ sprad nas lag pa gdangs [3] pa34 ni /
/ kun du bzang po 'i35 phyag ***** rgya yin /
/ rdo rje khu tshur dkur brtan ***** te /
/ glan pa g.yon du36 byo ba las /
[4] chags pa'i myig du lta ba ni / *****
/ rdo rje sgreg mo phyag rgya yin / *****
/ g.yon pa skong sko[r?]37 por byas te /
[5] g.yas pa gru mo bzung nas kyang /
/ bskong zhing sor mo 'phro ba ni /
/ rdo rje bdug pa'i phyag rgya yin /
/ lag pa gnyis kyi rgyab [6] sbyar nas /
/ rdo rje khu tshur bcangs pa las /
/ mthe'u cung sbrel nas mdzub mo bsgreng /
/ rdo rje lcags kyu 'i38 phyag rgya yin / /
[37v.1] mthe'u cung lu gu rgyud du sbrel /
/ sor mo rnams ni so sor bsgreng /
/ thal mo bkan te lta ba ni /
/ rdo rje sgril ma'i phyag rgya yin /
[2] srin lag dang ni mthe bo gnyis /
/ so sor rtse mor sprad pa las /
/'phral ba'i thad kar drangs pa ni /
/ rdo rje 'phreng ba'i phyag rgya [3] yin /
[The Dunhuang manuscript, together with the Tshal pa Kanjur and Bhutanese texts of TZ (the descendants of
c on our stemma codicum, p.43 above), here omit a passage of around two Tibetan text sides of lemmata,
which is found in full in the South Central, Tawang and Bathang texts, and the last few lines of which are
included in the Tenjur version of TZComm, where it begins again after its long lacuna, towards the end of
Chapter 10 [Gt265.5]. Thus, we can be confident that TZComm originally had this passage. Since the
passage has no additional commentary apart from the root text lemmata, we are not including it here to avoid
a long repetition: for this section of TZComm, then, see the TZ edition for Chapter 10, p.142-145.]

29

[mthebu?]: uncertain whether mthe'u or mtheb is intended, perhaps mtheb was corrected into mthe'u (mtheb/'theb is given in
most of the other texts). | 30 [cngar?]: deletion of ca is uncertain (a small faint line is perhaps above the letter), but probably
intended (ngar is given in the other texts). | 31 la: byas pa is given in the other texts | 32 rdo rje sgril ma'i: rdo rje sgrol ma'i is
given here in most of the other texts and seems more appropriate (see TZ edition p.139). | 33 rgang pa: presumably, rkang pa (as in
the other texts) is intended | 34 gdangs pa: the Tshal pa Kanjur texts agree with gdangs pa, but the other texts give gdengs pa (see
TZ edition). | 35 bzang po 'i: bzang mo'i is given here in most of the other texts (we would expect a female deity; see TZ edition
p.140). | 36 glan pa g.yon du: ldan pa g.yas su is given here in the other texts | 37 skong sko[r?]: various variants are given in the
other texts; most probably, kong kong is the correct reading (see TZ edition p.140). | 38 rdo rje lcags kyu 'i: there are several
slightly different versions of this tshig rkang; all agree on rdo rje lcags kyu, but the others all attempt to add a feminine particle.

Chapter 10

/39 thabs kyi zhags pa pad ***** ma 'phreng40 las /

//
phyag rgya 'i le'u ste41 bcu pa ***** 'o // :

39 Gt

inserts: 'phags pa | 40 pad ma 'phreng: Gt padmo 'phreng ba | 41 Gt inserts: /

281

Chapter 11
[Ms37v.4] [Gt266.2]
[above the line] $/ /'di man cad le'u sum cu rtsa dgu pas ni 'byor pa'i las bstan to i

da ni rig1 sngags gcad2 pa'i le 'u bshad **** de/

[beneath the line, continuing under the following lines 5 and 6, as indicated by the crosses] las drug ste 'byor pa'i las dang drag po 'i
las dang dbang gi las dang rgyas pa [da?]ng3 /zhi ba dang spyir btang ba 'o/ 'byor pas nI 'phrin las [gang?] bya na [ya?]ng kri
[ya?] ltar gcig nyams na yang dngos grub myi thob pa yin gyi skyon dang yon tan gnyis su myed par ++ rig na las gang byas kyang
skyon myed pa ni 'byor pa zhes bya 'o/ las gcig la yang lnga 'o/ de yang gang zhe na drag po 'i las dang drag po 'i ho dang drag po
'i phur pa dang drag po 'i gtor ma dang drag po 'i sbyor ba 'o/ 'phrin las dbang dang ++ rgyas pa dang zhi ba la'am/ 'dra bar lnga
lnga yod pas [sp?]yir bshad do

/de nas dkyil 'khor gyi lha dang4 ***** lha mo rnams la /
/5phyag rgya6 rdo rje [5] la stsogs7 pa byang chub sems dpa' brgyad dang /
'khor phal po ches8 dbyangs dang mthun bar9 bstod pa10/
/11
phyogs bcu12 'jig rten rab [6] 'byam na13 /
/rgyal ba gang bzhugs 'khor dang bcas /
/bla myed14 lha dang lha mo lnga'15/
/grub pa'i byin rlabs16 bsam myi17 khyab / /
[38r.1] $/ /18sems dpa'19 bcu drug rigs kyi gtso /
/sems ma bcu drug rigs kyi yum/
/mthu cen20 khro bo khro mo dang / /
[2] pho nya 'jig rten skyong bar21 ldan/
/ /22
sngon kyi23 thugs dam chen po ni /
/'jig rten ma lus thams cad kun /
/rgyal ba'i zhing du [3] sbyor bas na /
/phyag rgya chen por bda*****g sbyor cig /
/ /24
dpa' bo chen po mthu25 **** ldan rnams /
/ thugs rje chen po26 byin [4] rlabs27 kyis /
/ bdag gyi28 lhag pa'i29 bsa ***** ba30 rnams/
/grub pa31 byin gyis brlab **** du gsol /
/ /32
khams gsu [5] [Gt267] zhing du sbyor la phyag 'tshal lo /

i This

comment if it is suggesting that the various rituals it lists below are the subject of the next chapters up to and including
Chapter 39 would seem to indicate that the annotator is working with the correct numbering of the chapters, since discussion of
these rituals does go up to Chapter 39, numbered only 37 in the Dunhuang manuscript.
1 rig:

Gt rigs | 2 gcad: Gt gcod | 3 The original da seems clear, but a long vertical stroke goes right through it, perhaps as a deletion,
but perhaps it could have been merely an ink mark made in error. | 4 Gt inserts: / | 5 / /: Gt omit | 6 phyag rgya: QtGt phyag na
(there is no doubt that phyag na rdo rje is the correct reading, and is moreover found in all the TZ editions) | 7 stsogs: Gt sogs
8 ches: Gt che | 9 bar: Gt par | 10 pa: Gt nas | 11 /: Gt o | 12 bcu: Gt bcu'i | 13 'byam na: Gt 'byams ni | 14 myed: Gt med | 15 lha
dang lha mo lnga': Gt lha lnga lha mo yi (Gt's order is not supported by any of the TZ editions; the Tshal pa Kanjur and Bhutanese
texts lose the word lnga altogether, agreeing with dang and with Gt's reading of yi) | 16 byin rlabs: Gt byin brlabs | 17 myi: Gt mi
18 QtGt insert: o (the Dunhuang, Tshal pa Kanjur and Bhutanese TZ texts all omit o here, but the other root texts support QtGt's
inclusion) | 19 dpa': Gt pa | 20 cen: Gt chen | 21 bar: Gt dang | 22 /: Gt o | 23 kyi: Gt gyi | 24 /: Gt o | 25 mthu: Gt mthur
(none of the root texts support Gt here) | 26 po: Gt po'i | 27 byin rlabs: Gt byin brlabs | 28 gyi: Gt gi | 29 lhag pa'i: added in small
writing as subscribed correction, the positioning marked by a cross | 30 bsa ba: Gt bsam pa | 31 pa: Gt par | 32 /: Gt o

Chapter 11

283

/dmyigs myed33 chos kyi sku la phyag 'tshal lo /

/longs spyod rdzogs pa'i sku la phyag 'tshal lo /
[6] cir yang sprul pa'i sku la34 phyag 'tshal lo /
/zhes bstod nas nyid kyi thugs kar zhugs te /
/rol pa khro ba35 zhes bya ba'i ting nge [38v.1] 'dzin la snyoms par zhugs nas /
/khro bo 'i rtog pa bshad do /
/de nas bcom ldan 'das dpal36 rdo rje sems dpa' la /
[beneath the line] 'phags pa'i nor bdun gyi le'u 'o ii

/lag na [2] rdo rje37 'di skad ces gsol to /

/bcom ldan 'das da ltar dang38 ma 'ongs pa'i dus na /
gcer bu pa dang /
kun du39 rgyu la stsogs pas /40
so so 'i [3] rig sngags41 kyi las42 kyis /
43ma ha44 de ba **** la stsogs pa45 'jig rten gyi lha dang / *****
/gnod sbyin mthu bo che dang /
mthu mo [4] che dang /46
'phra men ma47 thams cad bsku*****l nas /
/sems can la48 srog gi ***** bar chad bgyid/49
'tshe bar bgyid/
[5] shin du sdug bsngal bar bgyid na /
/de las50 thabs gang gis bzlog par bgyi51/
/sbyor ba gcad52 par bgyi /
/bka'53 stsal pa /
[6] lag na rdo rje nyon cig /
/'di ni de bzhin gshegs pa 'i thams cad kyi54 sku dang gsung dang thugs rdo rje 'i55 bdag pos sngon
yang bshad de / /
[39r.1] $/ /lag na rdo rje khyod nyid56 kyis byas pa'i rig sngags57 dang /
srid pa zad pas58 byas pa'i59 gsang sngags thams cad kyang bzlog par byed /
/gnon [2] par byed /
sbyor ba gcod par byed60 na /
/ kun du61 rgyu la stsogs62 pa gzhan lta ci smos /
/ de [Gt268] nas bcom ldan63 rdo rje mtshon ca zhes bya ba'i64 [3] ting nge 'dzin la snyoms par zhugs
nas / ****
ii Note

also the annotation at the beginning of Chapter 7 (f.19v.6) concerning this set, dividing them into those of cause and those
of the fruit. The standard set consists of: faith, morality, giving, listening (to the teachings), decorum, self-respect, and wisdom
(dad pa, tshul khrims, gtong ba, thos pas rgyud grol, khrel yod, ngo tsha shes, shes rab). But it is not clear how they fit here, or
how they fitted with the text in Chapter 7 either.
33 myed:

Gt med | 34 sku la: Gt omit (none of the root texts support Gt here) | 35 khro ba: Gt khro bo (the Tshal pa Kanjur and
Bhutanese TZ texts all give khro bo or khro bo'i here, but most of the other root texts support the Dunhuang manuscript reading)
36 dpal: Gt omit (most of the root texts support the Dunhuang manuscript reading) | 37 rdo rje: Gt rdo rjes | 38 Gt inserts: / | 39 du:
Gt tu | 40 stsogs pas /: Gt sogs pas | 41 rig sngags: Gt rigs sngags | 42 las: Gt lus (a clear error in Gt) | 43 QtGt insert: lha | 44 ha: Gt
h | 45 stsogs pa: Gt sogs pa/ | 46 mthu mo che dang /: QtGt omit (other editions also omit this although the Bhutanese texts include
it; it seems to be a kind of dittography, yet the mo means that it is not a simple dittography and could fit appropriately. Since
neither the local Kanjurs, nor the South Central NGB, nor the Tshal pa Kanjur texts give it, it would appear to be an error of the
Dunhuang manuscript, perhaps coincidentally shared with the Bhutanese texts.) | 47 'phra men ma: Gt 'phra men | 48 la: Gt omit
49 /: Gt omit | 50 las: Gt omits (the other versions give la here) | 51 bgyi: Gt bgyid | 52 gcad: Gt bcad | 53 bka': 'a subscribed
54 thams cad kyi: Gt omit (none of the root texts support Gt here) | 55 thugs rdo rje 'i: Gt / thugs rje'i (none of the root texts support
Gt here) | 56 nyid: Gt omit | 57 rig sngags: Gt gsang sngags (all the root texts support Gt here) | 58 pas: Gt par | 59 byas pa'i: Gt bya
ba'i | 60 sbyor ba gcod par byed: QtGtNt omit: / (eyeskip) | 61 du: Gt tu | 62 stsogs: Gt sogs | 63 bcom ldan: Gt bcom ldan 'das kyis
(the Tshal pa Kanjur and Bhutanese TZ texts support Gt; the other root texts give 'das but not kyis) | 64 mtshon ca zhes bya ba'i: Gt
mtshon cha'i/ (only the Dunhuang manuscript gives zhes bya ba'i)

284

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/sku dang gsung dang thugs rdo rje las / *****

/ gsang sngags bzlog65 cing66 gnon pa dang/
[4] sbyor ba gcod pa zhes bya ba 'di gsungs ***** so /
/ dz la ba dzra h / *******
/man tra sum bha ni /
/ ba dzra [5] 'dz la h /
/ bhr / h bhru / tra bhru / hrI bhru /a bhru /67
/'di gsungs ma thag du68 /
phyogs bcu nas mtshon ca 'i69 char chen por70 [6] bab ste /
lag na rdo rje nyid kyang71 kha sbub du72 'gyel nas73 khrag du74 skyugs par gyur to/
/lha dang gnod sbyin thams cad kyang75 brgyal te /
[39v.1] lus dum por76 dum puor77 chad par gyur to /
/de nas bcom ldan 'das kyis78 de dag gi sbyor ba chad par rig nas79 slar bslang80 te/
/ ba dzra sa twa ra [2] dz dz81 /
/de nas bcom ldan 'das kyis82 lag na rdo rje la bka'83 stsal pa/
/lag na rdo rje khyod nyid kyang84 'di las 'da' bar ma byed cig / /
[3] khyod85 kyi 'khor thams cad kyang 'di las **** 'da' bar byed na /
/thams cad brlag86 ***** par 'gyur ro /
/ de bzhin du lag [4] na rdo rje khyod kyis /
lha dang gnod sbyin tha*****ms cad la yang sgo shig87 /
/de nas ***** lag na rdo rjes gsol ba'88/ /
[5] bcom ldan 'das 'di na /
srid pa89 gsum thams cad90 'dul bar bgyid pa /
/mthu bo che thams cad kyi91 nga rgyal gcog92 pa bdag93 lags94 [6] ci ltar bcom ldan 'das shag kya95
thub pas96 mthu stobs dang ldan ba'i slad du/
/bdag mying97 lag na rdo rje zhes btags na98/ /
[40r.1] $ / / 'dir rdo rje sems dpa's [Gt269] bcom ldan 'das kyi byin gyi99 brlabs kyis /
/sngags 'di smras pas /
/ bdag kyang khrag du100 skyug [2] cing 'gyel bar 'gyur101 na /
gzhan lta ci smos te/
/'di ni 'da' bar dka'102 ba 'o/
/de bzhin du lha dang103 gnod sbyin rnams la yang bsgo 'o104/
[3] de nas bcom ldan 'das kyis ma 'ongs **** pa'i dus na /
gsang sngags 'di bsgrub105 ***** pa'I cho ga bshad de106 /
ma las nyi ma'i [4] dkyil 'khor du /107
dpal cen108 khrag ***** thung109 spyan drang ngo /
/sku ni dud ka110 chen ***** po111 la /
[d(//)]bu112 skra kham nag ral pa
65 bzlog:

Gt bzlos (a clear error in Gt) | 66 cing: Gt shing | 67 / dz la ba dzra h / /man tra sum bha ni // ba dzra 'dz la h
/ / bhr / h bhru / tra bhru / hrI bhru /a bhru/: Gt o dzwa la badzra h/ mantra sum ba ni/ o dzwa dzwa la
h/ o bru/ h bru tra bru tra bru/ a bru/ (the repeated tra bru in Gt seems to be in error; most other verions
agree with hri/hr) | 68 du: Gt tu | 69 mtshon ca 'i: Gt mtshon cha'i | 70 por: Gt po | 71 kyang: Gt omits | 72 du: Gt tu | 73 Gt inserts:
/ | 74 du: Gt tu | 75 kyang: Gt omit | 76 dum por: por is inserted above; Gt omits (none of the root texts support Gt here) | 77 dum
puor: the na ro above is crossed through; Gt dum bur | 78 Gt inserts: / | 79 Gt inserts: / | 80 bslang: Gt langs | 81 ba dzra sa twa ra
dz dz: Gt o badzra swa da twa ra dza dzha | 82 Gt inserts: / | 83 bka': 'a subscribed | 84 Gt inserts: / | 85 khyod: Gt khyed
86 brlag: Gt rlag | 87 sgo shig: Gt bsgo zhig | 88 ba': Gt pa | 89 pa: Gt omit | 90 thams cad: Gt omit (most of the root texts support the
Dunhuang manuscript reading) | 91 kyi: Gt kyis | 92 gcog: Gt gcod | 93 bdag: Gt dag | 94 QtGt insert: te/ | 95 shag kya: Gt shkya
96 Gt inserts: / | 97 mying: Gt gi ming | 98 na: Gt nas | 99 gyi: insertion subscribed as a correction; Gt gyis | 100 du: Gt tu | 101 'gyur:
Gt gyur | 102 bka': 'a subscribed | 103 de bzhin du lha dang: QtGtNt omit (a clear error in QtGtNt) | 104 'o: it appears that a previous
go has been corrected to 'o. | 105 bsgrub: Gt grub | 106 de: Gt do | 107 /: Gt omits | 108 cen: Gt chen | 109 thung: Gt 'thung | 110 dud ka:
Gt dung ka | 111 po: Gt mo | 112 [d(//)]bu: Gt dbu

Chapter 11

285

[5] pa113 can /

/dbu gsum pa114 la phyag drug ste /
/dmar dang dud ka gnag pa115 yin/
/g.yas kyi gsum na thogs pa 'i116/ /
[6] sti ra 'i dbyug pa chen po dang /
/dung cen117 dmar gis gang118 ba la /
/rdo rjes dkrug cing gsol 119 bar brtag /
/ g.yon gyi gsum [40v.1] na thogs pa ni /
/ sti ra 'i rgyu 'i zhags120 pa can dang121 /
/khyung kyi122 khra chen thogs pa dang /
/sti ra phra mo 'i gsum123 pa nas / /
[2] glo snying nang grol124 'dren cing za /
/g[ts(/c)]ug du125 rnam par snang mdzad nyid /
/126 phrag pa g.yas la127 a ksho bhya /128 h/
/g.yon na rad na129 sa [3] ba bha /tr130/
/rkang pa g.yas na a mo gha****sid dhi131 a/
/g.yon na a myi dha ba132 ste133****hri134 /
/bhru135 las 'khor lo rdo rje rnams/136
[4] ni rig sngags137 brtag /
iii /ke'u ri138 dang ***** ni tse'u ri dang /
/pra mo ha dang be ta lI139/ *****
iv /gzhan yang bhya gra mu kha140 dang /
[5] sing ha mu kha141 dag142 dang [ldIn?]143 /
/ snri kha144 la stsogs145 mgyogs ma rnams/
/sti ra la ni 'bri tshugs146 su /
/lta zhing ltod147 pa'i tshul las [6] ni// [Gt270]
/ ha ha148 de nas yud tsam skad cig149 la/
/rig sngags150 nyid dang sbod dang151 gtong bcas te /
/ 'gug152 pa'i pho nyar brtag par bya / /
[41r.1] $/ /ras sam gro ba153 'am rtswa rnams la /
/khro bo 'i gzugs su bris pa 'am /
/yang na skud pa nag pos bcing /
/ha ha154 gad [2] mo chen po dang /

iii

Note that here the text briefly introduces the first members of two sets of goddesses, who are described further in the
commentary below, and in the case of the first set of eight, their mantras are also given in the root text (see below, Ms42r-44v).
iv Note that the order of these goddesses given here is not the same as that in the later discussion in the chapter (see Ms43v-44r),
and the order as given below in fact corresponds to that in our edition of TZ (see p.154 above).
113 pa

pa: dittography at new line | 114 pa: Gt po | 115 pa: Gt pa'i | 116 'i: Gt ni | 117 cen: Gt chen | 118 gis gang: Gt gyis bkang
Gt 'thung (all the root texts support Gt here) | 120 'i zhags: Gt ma thogs | 121 can dang: dang inserted below crossed out
can | 122 kyi: Gt gi | 123 gsum: QtGt gsus (this is surely an error in the Dunhuang manuscript) | 124 nang grol: Gt nang khrol
125 g[ts(/c)]ug du: Gt gtsug tu | 126 / /: Gt o/ | 127 la: Gt pa | 128 a ksho bhya /: Gt ako bhya | 129 rad na: Gt ratna | 130 sa ba
bha /tr: the attached a chung in the syllable tr is placed above the attached ra; Gt sam bha ba tra | 131 a mo gha sid dhi: Gt
a mo si ti | 132 a myi dha ba: Gt a mi de ba | 133 ste: Gt omit | 134 hri: Gt hr | 135 bhru: Gt bru | 136 'khor lo rdo rje rnams/:
QtGt rdo rje 'khor lo dang / /rdo rje rnams (most of the root texts support Gt here, although they give la/ rather than dang/)
137 rig sngags: Gt rigs sngags | 138 ke'u ri: Gt kai ri | 139 be ta lI: Gt pe ta la | 140 bhya gra mu kha: Gt bya tri mu ka. | 141 kha: Gt
ka | 142 dag: Gt omit | 143 ldin: very slight smudging on lda which might possibly indicate deletion, the gi gu is more above the na
than lda, so perhaps nI is intended; Gt omit (as do other versions of the text) | 144 snri kha: Gt sri la mu ka | 145 stsogs: Gt sogs
146 'bri tshugs: Gt 'bru tshugs | 147 ltod: Gt rtod (no other versions share the Dunhuang manuscript version some give bstod
yet ltod seems clearer here) | 148 ha ha: Gt ta ha | 149 yud tsam skad cig: Gt skad cig yud tsam | 150 rig sngags: Gt sngags (none of
the root texts support Gt's reading) | 151 dang: Gt omit | 152 'gug: Gt 'gugs | 153 gro ba: Gt 'gro ba | 154 ha ha: Gt ha ha'i
119 gsol:

286

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/ rdo rje bskul ba'i glu blangs te /

/ pho nya rnams la155 bngag156 par bya /
de nas rig sngags157 bzung nas kyang /
/rtswa la158 'I gzugs [3] la gnas pa dang /
/'khor lo ***** la stsogs159 mtshon ca 'is160/
/de dag ***** phye mar klag161 par bya /
/ badzra tsag kra [4] h162 /
/phye ma de dag btus163 nas ni164 / *****
/thod pa gzhu dang bcas pa 'i165/ ******
/hom khung gru gsum sgo can du /
[5] /ram166 zhes bya 'i167 mye sbar te168 /
/drang srong 'khor rnams g.yab mos 'bod169/
/de nas rig sngags170 bsreg par bya /
/ de ltar lha dang [6] bdag sbyar nas /
/rnam lnga 'i chas dang ldan bar171 bya /
/ ma mo172 chen mo la sha la spyon173 /
/ 'byung bo 'i sha ni zan bas zhim / /
[41v.1] 'byung po 'i khrag174 ni chang bas zhim /
/ ma mo chen mo sha175 la spyon176/ 177
/glu dang gar dang gang rgyangs gyis178/
/so sor mchod de bsgrub [2] par bya /
/zin pa'i sems can gang yang rung /
/de dag skud pa nag pos bcing /
/de yang dum por179 gcad par bya /
/de nas mye ni180 [3] snga ma bzhin /
/tshe gnas [Gt271] bya phyir **** 'bru lnga yis /
/gnas lnga dag du rnam181**** bkod de /
/zin pa'i sems can [4] bsrung182 bar bya /
/tsha zhing rtsub ****** pa'i shing dang ni /
/sha dang chang dang ****** yo byed183 rnams /
/gnag cing tsha [5] ba rnams sbyar te /
[beneath the line] yungs nag bca' sga [beneath line 5] lcags phye

/ lha chen rnams la mchod par bya /

/de nas las rnams grub pa dang /
/ha ha la stsogs184 gad mo dang /
[6] /glu dang rdo rje dbyug pa yis /
/lha rnams bdag185 la thim par bya /
/'di ni bdag nyid bsrung ba dang /
155 la:

Gt kyang | 156 bngag: Gt mngag | 157 rig sngags: Gt rigs sngags | 158 la: Gt omits | 159 stsogs: Gt sogs | 160 mtshon ca 'is: Gt
mtshon cha yis | 161 klag: Gt rlag | 162 h: slightly uncertain, it appears that bh was possibly corrected to h; Gt bru (the
Dunhuang, Tshal pa Kanjur and Bhutanese texts all give h; the South Central and local Kanjur texts give bru h) | 163 btus:
Gt gtubs | 164 ni: Gt kyang | 165 'i: Gt yis | 166 ram: Gt ra | 167 'i: Gt ba'i | 168 mye sbar te: Gt me spar nas | 169 'bod: Gt bod | 170 rig
sngags: Gt rigs sngags | 171 bar: Gt par | 172 ma mo: corrected from mo mo with a smudge | 173 spyon: Gt sbyon | 174 khrag:
smudged and erasure of an earlier gi gu seems to be intended. | 175 sha: Gt khrag (the Dunhuang manuscript reading is supported
by the South Central and local Kanjurs; Gt agrees with the Tshal pa Kanjur and Bhutanese texts) | 176 spyon: Gt sbyon | 177 QtGt
insert: /ma mo chen mo rkang la sbyon/ /'byung po'i rkang ni mar bas zhim/ /'byung po'i rus pa zhim pa ste/ /'di dag gsol te
mnyes par mdzod/ (This would seem an appropriate addition but it is confined to the Tenjur tradition. The text remains coherent
without it, although the two items to be consumed already mentioned, along with the two additional items here, might possibly
constitute four of the five articles mentioned just above [rnam lnga'i chas].) | 178 gyis: Gt kyis | 179 por: Gt bur | 180 mye ni: Gt mi
de | 181 du rnam: Gt tu rnams | 182 bsrung: Gt srung | 183 yo byed: Gt yo byad | 184 stsogs: Gt sogs | 185 bdag: Gt dag (an error in
Gt)

Chapter 11

287

/las rnams thas cad [42r.1] $ / / grub par byed/

phyogs na 'phra men ma ko'u re186 kha dog sngon mo /187 vi
phyag gnyis pa ste188/
phyag g.yon189 pa 190 tsan dan [2] dmar pos bkang ba thogs 191/
g.yas192 zhing chung ngu 'i193 dbyig194 pa thogs pa ste /
ba dzra ko'u ri dz dz h195/
zhes stong rtsa brgyad bzlas te bcol [3] na/
/sangs rgyas kyang 'gugs 196 na/ ***
/gzhan lta smos kyang ci dgos / *****
/che ge mo zhig 197 ces bcol la /
[4] gtor ma dang198 tsan dan dmar chen gyi199 a ***** rga la /
/200 ru rulu201 ru lu h bhyo /
/zhes brjod cing /
a rga lan gsum [5] sbreng ngo /
/ lho phyogs na202 tse'u ri kha dog ser mo /
phyag na mda' gzhu203 'gengs pa/
/ ba dzra tso'u ri h 'dz204/ vii
v /shar

[beneath the line, continuing onto line 6] gdon dang dgra 'i gzugs byas la mda' 'i/ bar bzhag gam gang la gdug pa'i + [line 6] phyogs
kha bltas te mda' 'phangs te phog cing gzir la la bdag du gi dbang du 'gyur ro

zhes stong rtsa205 brgyad bzlas te bcol [6] na /

phyag na rdo rje 'i mthu yang /206
rdul tsam yang myed par rku nus pa na /207
/ gzhan lta smos kyang ci dgos /

v Note

that here the text describes a set of eight goddesses, versions of which are familiar from Buddhist Sanskrit sources, such as
the Hevajra literature, where their names are typically given as: Gaur, Caur, Vettl, Ghasmar, Pukkas, abar, Cal and
omb. In rNying ma pa sources, there are variations on the list, but generally, abar and omb are replaced by, Pramoh and
ma n. A very similar version of the set to that found in the TZ is given in the mantras listed in Chapter 16 of the rGyud gSang
ba'i snying po de kha na nyid nges pa (mTshams brag NGB edition, Volume Wa: 207), the one difference being the placement of
Cal, here given at the end of the list of eight, but in the rGyud gSang ba'i snying po's Chapter 16, in position 5, at the head of
the second set of four. It will be seen that the first group of four are associated with the cardinal directions, while the second
group of four are associated with the intermediate directions (see below). Klong chen pa's Phyogs bcu mun sel commentary on the
rGyud gSang ba'i snying po (bdud 'joms bka' ma Volume La: 539) gives interesting glosses of the names of these eight goddesses,
who he refers to as the eight mamos, and later, the eight site mamos (gnas kyi ma mo brgyad, bdud 'joms bka' ma Volume La: 605).
See the notes on the individual goddesses below.
vi In discussing the first member of this group of goddesses in the Phyogs bcu mun sel, Klong chen pa first acknowledges the
name, Gaur, meaning White Lady, but then notes further that most of the former authoritative root texts rather give the name as,
ko'u r. Rather than a spelling error, he glosses this as having many meanings, the relevant one of which is mdzod ldan. This is
the Tibetan equivalent for the ancient Indian town of Kau mb, famous from the stories of the Buddha's life. The text reads: gau
r ni dkar mo'o/ /de sngon gyi gzhung phal che bar ko'u r zhes brjod pa'ang de nyid de/ mdzod ldan dang / mkhas pa dang /
padma la dga' ba la sogs pa du ma la 'jug kyang / mdzod ldan du bsgyur ro/, bdud 'joms bka' ma Volume La: 539.
vii This second of the eight mamos is equated with Caur, who is also given second in the rGyud gSang ba'i snying po's Chapter
16; see Klong chen pa's commentary, bdud 'joms bka' ma Volume La: 539.
186 ko'u

re: Gt ke'u ri ma/ / | 187 /: Gt omits | 188 pa ste: Gt te | 189 g.yon: QtGt g.yas (the Tshal pa Kanjur tradition, which includes
this commentarial passage, agrees with the Dunhuang manuscript on the ordering of the hands. The Tenjur may have
unreflectively used the standard order of giving the right first.) | 190 Gt insert: na ban dha (the Tshal pa Kanjur tradition agrees
with Gt) | 191 Gt insert: pa | 192 g.yas: QtGt g.yon | 193 zhing chung ngu 'i: Gt zhing tshung gi | 194 dbyig: Gt dbyug | 195 ba
dzra ko'u ri dz dz h: Gt o badzra ke ri dza dza h (see note vi above) | 196 Gt insert: nus | 197 Gt insert: khug cig
198 dang: Gt omit | 199 gyi: Gt gyis | 200 : Gt o | 201 rulu: it appears that ru written in error has been deleted; in part, it has been
amended into the following lu. | 202 na: Gt nas | 203 gzhu: Gt omit (the Tshal pa Kanjur texts support the Dunhuang manuscript
here) | 204 ba dzra tso'u ri h 'dz: Gt o badzra tse'u ri ha dza dza | 205 stong rtsa: Gt brgya rtsa (QtNt, and the Tshal pa
Kanjur tradition, agree with the Dunhuang manuscript here) | 206 /: Gt omits | 207 rdul tsam yang myed par rku nus pa na/: QtGtNt
'phrog nus na/ (the Tshal pa Kanjur tradition essentially agrees with the Dunhuang manuscript here; see p.159 above)

288

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/nub phyogs na pra mo ha dmar po /208

[42v.1] phyag na chu [Gt272] srin gyi rgyal mtshan thogs pa /
[beneath the line] spyi bor la bskor bar bsam

ba dzra pra mo ha h dza dz209 viii zhes stong210 rtsa brgyad bzlas te bcol na/
/phyag na rdo rje yang rang 211 gi212 mthu [2] gzhag213 nas dbang du 'ong na /

[beneath the line] lha dang myi yang bdag gi dbang du 'dus par bsgom

/gzhan lta cI smos214 /

/byang phyogs na be ta li215 kha dog nag po216/
phyag g.yon na zhing chung ngu217 'dzin za ba218 /
[3] phyag g.yas na rdo rje phyar/219
ba dzra be ta*****li h dz220 ix zhes stong rtsa brgyad bzlas te bcol na / ****
/'jig rten dang 'jig rten las 'das [4] pa'i dngos grub thams cad kyang grub221 na ******/
/'jig rten gyi bsgrub pa phran222 tshegs lta ****** ci smos /
/shar lho 'i223 mtshams na224 [5] pus ka si225 226 dmar ser /
/phyag gnyis zhing chung ngu 'i227 ba su ta 'dren cing za ba228/
ba dzra pus ka si h dz229 x/
zhes stong rtsa brgyad230 bzlas te bcol [6] na /
phyag na rdo rje yang 'ching231 zhing 'dren par nus na /
gzhan za bar bya ba232 lta ci smos /
/ lho nub kyi mtshams na /
kas ma li233
[43r.1] $/ /kha dog ljang ku234 /
phyag g.yon pa na235 ban da236 tsan dan dmar pos bkang ste237 rdo rjes dkrug cing 'thung ba /
ba dzra kas ma ri h dz238 xi/ /
[2] zhes stong rtsa brgyad bzlas te bcol na /
phyag na rdo rje yang myos par 'gyur na /
gzhan lta smos kyang ci dgos /
nub byang mtshams na239 badzra kar [3] ma240 kha dog ljang ku241/
zhing chung ngu242 lag pa ***** gnyis kyis bzung ste243 snying kha244 nas za ba /

viii

This third of the eight mamos is Pramoh, who is also given third in the rGyud gSang ba'i snying po's Chapter 16; see Klong
chen pa's commentary, bdud 'joms bka' ma Volume La: 539.
ix This fourth of the eight mamos is Vetl, who is also given fourth in the rGyud gSang ba'i snying po's Chapter 16; see Klong
chen pa's commentary, bdud 'joms bka' ma Volume La: 539.
x This fifth of the eight mamos is Pukkas, who is given sixth in the rGyud gSang ba'i snying po's Chapter 16, following Cal,
here given in position eight; see Klong chen pa's commentary, bdud 'joms bka' ma Volume La: 539.
xi This sixth of the eight mamos is Ghasmar, who is given seventh in the rGyud gSang ba'i snying po's Chapter 16; see Klong
chen pa's commentary, bdud 'joms bka' ma Volume La: 539.
po/: QtGt / kha dog dmar mo/ (the Tshal pa Kanjur texts support dmar mo/) | 209 ba dzra pra mo ha h dza dz: Gt
o badzra pra mo ha dza dza/ | 210 stong: QtGtNt brgya (the Tshal pa Kanjur texts agree with the Dunhuang manuscript here, and
in future occurrences in the passage below, all versions give, stong) | 211 QtGt insert: bzhin (the Tshal pa Kanjur texts agree with
the Dunhuang manuscript here) | 212 gi: Gt gyi | 213 gzhag: Gt gzhan (the Tshal pa Kanjur texts give bzhag) | 214 cI smos: cI is
added in small writing as a correction, below the line; Gt smos kyang ci dgos | 215 be ta li: Gt pe ta li | 216 nag po: Gt nag mo
217 zhing chung ngu: Gt zhing chung | 218 za ba: Gt cing za ba | 219 phyar: Gt 'phyar ba | 220 ba dzra be ta li h dz: Gt o
badzra pe ta li h dza dza/ | 221 grub: Gt 'grub | 222 phran: Gt 'phran | 223 lho 'i: Gt lho | 224 na: Gt nas | 225 pus ka si: Gt pu ka si
226 Gt inserts: kha dog (the Tshal pa Kanjur texts support Gt here) | 227 zhing chung ngu 'i: Gt zhing chung gi | 228 za ba: Gt brab
(indistinct, za ba may be intended; Qt gives za ba) | 229 ba dzra pus ka si h dz: Gt o badzra pu ka si h dza dza
230 stong rtsa brgyad: QtGtNt omit (the Tshal pa Kanjur texts support the Dunhuang manuscript here) | 231 'ching: QtGt 'chi (a
clear error in QtGt) | 232 za bar bya ba: QtGt omit (the Tshal pa Kanjur texts give bcing zhing za ba) | 233 kas ma li: QtGt kas ma
ri | 234 ljang ku: Gt ljang gu | 235 g.yon pa na: Gt g.yon | 236 ban da: Gt ban dha | 237 Gt inserts: / | 238 ba dzra kas ma ri h
dz: Gt o badzra kas ma ri h dza dza | 239 Gt inserts: / | 240 kar ma: QtGtNt sma sha ni | 241 ljang ku: Gt ljang gu | 242 zhing
chung ngu: QtGt zhing chung | 243 Gt inserts: / | 244 snying kha: Gt snying ga
208 dmar

Chapter 11

289

/ ba dzra kar ma h dz245 xii/

zhes stong rtsa [4] brgyad bzlas te bcol na /246
las gang ***** bya ba thams cad 'grub bo /
/byang shar ***** gyi mtshams na /
tsan ta li kha dog [5] ser skya phyag g.yas na /
zhing chung ngu 'i ban da 'dzin /
g.yon zhing nyid 'dzin247 /
ba dzra tsan da li h dz / xiii
zhes stong rtsa brgyad bzlas te [6] bcol na /248
za ma [Gt273] dang ma ning la249 yang rgyud 'phel bar byed250 na /
dbang po tshang zhing nyams pa myed251 pa252 /
tshe dang dbang thang 'phel bar bya ba [43v.1] bya ba253 lta ci smos /
/'phra men ma 'di rnams 'phrin las mdzad par bskul254 ba'i tshe/
a rga255 dang gtor ma dbul ba ni/256
gong du smos pa bzhin [2] 257 tshul bzhin 258 bgyi/
pa/259
khros na260 thams cad 261 phyag cha tril shu la262 'bar ba /
g.yon [rd?]e'u cung263 bsnams 264/
gdan mye265 smug gnag266 267/ /
[3] sku 'i kha dog nag mo268 la /
khrag dang ***** zhag dang thal ba'i thig les tshum269 bur byas ***** shing /
ba su tas rked par270 bcings271 /
[4] gser kyi272 pags pa gos su bgos273/ *****
/sro na ta an dz nas274 lus brgyan te / *****
gcer bu 'i tshogs smug gnag275 'bar ba276 [5] g.yon rde'u cung277 bsnams278 /
/gdan mye279 smug gnag280 /
sku 'i kha dog nag mo la281 'jig rten 282 khams zad cing bsreg283 par byas nas / /
[6] slar dpal chen po 'i sku la284 thim par bya 'o285/ / /286
/da ni287 phyi288 rim 289 bshad par bya ste /

xii This seventh of the eight mamos should be man, who is given eighth in the rGyud gSang ba'i snying po's Chapter 16; see
Klong chen pa's commentary, bdud 'joms bka' ma Volume La: 539.
xiii This final eighth mamo is Cal, who is given fifth in the rGyud gSang ba'i snying po's Chapter 16; see Klong chen pa's
commentary, bdud 'joms bka' ma Volume La: 539.
245

ba dzra kar ma h dz: QtGtNt o badzra sma sha ni h dza dza (this is an error shared by the Dunhuang manuscript,
the Tshal pa Kanjur, and Bhutanese NGB, which give karma.) | 246 QtGtNt omission begins here; this seems to be a case of
eyeskip to the next occurrence of: zhes stong rtsa brgyad bzlas te bcol na | 247 g.yon zhing nyid 'dzin: the Tshal pa Kanjur texts
give g.yon gyis zhing dbyug 'dzin pa | 248 QtGtNt omit: las gang bya ba thams cad 'grub bo / /byang shar gyi mtshams na / tsan ta
li kha dog ser skya phyag g.yas na / zhing chung ngu 'i ban da 'dzin / g.yon zhing nyid 'dzin / ba dzra tsan da li h dz / zhes
stong rtsa brgyad bzlas te bcol na | 249 la: Gt omits | 250 byed: Gt 'gyur | 251 myed: Gt med | 252 Gt inserts: lta ci smos | 253 bya ba
bya ba: dittography at the turn of the folio | 254 bskul: Gt skul | 255 a rga: Gt argas | 256 /: Gt omits | 257 Gt inserts: du | 258 Gt
inserts: du | 259 pa/: Gt omit | 260 khros na: Gt 'khros nas | 261 Gt inserts: kyi | 262 tril shu la: Gt tri shul | 263 cung: Gt chung
264 Gt inserts: pa | 265 mye: Gt me | 266 gnag: Gt nag | 267 QtGt insert: gi steng na bzhugs pa (the Tshal pa Kanjur tradition agrees
with the Dunhuang manuscript here) | 268 mo: Gt po | 269 tshum: Gt tshom | 270 par: Gt pa | 271 Gt inserts: pa | 272 kyi: Gt gyi
273 bgos: Gt gyon pa | 274 sro na ta an dz nas: a small line above the na apparently deletes it, but it is not entirely certain; Gt sro ta
a 'dzas (most probably, this indicates a transcription of Sanskrit, srotojana, a white collyrium applied to the eyes) | 275 gnag: Gt
nag | 276 Gt inserts: / | 277 cung: Gt chung | 278 Gt inserts: pa | 279 mye: Gt me | 280 gnag: Gt nag | 281 QtGt omit: / sku 'i kha dog
nag mo la (this seems to be an error in QtGt) | 282 Gt inserts: gyi | 283 bsreg: Gt sreg | 284 sku la: Gt skur | 285 bya 'o: Gt byas so
286 Here we find two punctuation marks, each of a small circle positioned in the middle of the line, and evenly placed between
separating shads. | 287 da ni: Gt de'i | 288 phyi: Gt phyir | 289 Gt inserts: par

290

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

shar phyogs su290 sing ha mu kha291 xiv/

kha dog ser mo [44r.1] $ / /zhing kha na thogs pa /
lag pa g.yas g.yos292 bsnol pa293 seng ge 'i mgo can /
/ lho phyogs na bhya gra mu kha294 /
kha dog dmar [2] mo295 /
lag pa gnyis bsnol te296 mdun ngan297 zhing gnas pa la lta ba298 stag gyi299 mgo can /
/nub phyogs na snri kha300 /
kha dog nag mo301 [3] lag pa gnyis kyis302 [Gt274] zhing chung ngu303 'dzin ***** cing lces ldag304 pa /
wa305 mgo can/ ****
/byang phyogs na shwa na mu kh306/ /
[4] kha dog mthing ka307 lag pa gnyis ***** kyis khong dral te308 lta ba /309
khyi310 mgo can *****/
shar lho 'i mtshams na311 kri ta 312 kha [5] dog dmar gnag /
zhing chung ngu 'i ba su ta kha na thogs pa /
lag pa g.yas pa313 na ban dha /
g.yon 314 gri thogs pa /
bya rgod mgo [6] can/
/lho nub 315 mtshams na316 kang ka 317/
kha dog dkar dmar318 /
phrag pa g.yon pa na319 zhing khel320 zhing rkang pa nas bzung ba / /
[44v.1] g.yas pa321 mgo 'dzin 322/
bya bkang ka 'i323 'go la324 mchu ring 325/
thor to gug326 ze ba dmar 327/
/nub byang gi mtshams na/
ka ka328 kha dog nag mo /
[2] mchus zhing thogs pa /
lag pa g.yon bas329 pad mo330 ban dha 'dzin 331/
bya rog mgo can332 /
byang shar kyi333 mtshams su u lu334 kha dog sna tshogs335 [3] lag pa g.yon336 pa dung cen337 khrag gis
gang338**** ba la339 'thung ba /
xiv

Here we begin description of a group of eight animal-headed goddesses beginning with Sihamukh (Tibetan, seng gdong ma),
whose mantras are given in the same order in Chapter 16 of the rGyud gSang ba'i snying po (gsang ba'i snying po de kha na nyid
nges pa; in the mTshams brag NGB edition, Volume Wa: 207). They are discussed in Klong chen pa's Phyogs bcu mun sel
commentary on the rGyud gSang ba'i snying po (bdud 'joms bka' ma Volume La: 539), as the eight phra men ma (Skt. pc).
290 su:

Gt nas | 291 sing ha mu kha: Qt singha mu ka; Gt sidha mu ka | 292 g.yos: Gt g.yon | 293 pa: Gt ba | 294 bhya gra mu kha: Gt
bya gri mu ka (= Vyghrmukh; Tibetan, stag gdong ma) | 295 dmar mo: final ra of dmar subscribed; Gt dmar po | 296 Gt inserts: /
297 ngan: Gt na (perhaps /na intended) | 298 Gt inserts: / | 299 gyi: Gt gi | 300 snri kha: Gt sri la mu ka (= glamukh; Tibetan, wa
gdong ma) | 301 nag mo: Gt nag po/ | 302 kyis: Gt na | 303 zhing chung ngu: Gt zhing chung | 304 ldag: Gt lngag (but ldag was
probably intended) | 305 wa: the letter is not a clearly written wa but is consistent with the wa given in pha wang on 54v.4; Gt wa'i
306 shwa na mu kh: Gt sho na mu ka (= vnamukh; Tibetan, spyang ki gdong ma) | 307 mthing ka: Gt mthing ga /zhing | 308 Gt
inserts: / | 309 /: Gt omits | 310 khyi: Gt khyi'i | 311 Gt inserts: / | 312 kri ta: Qt kri ta mu ka /; Gt kri ta ka kha (= Gdhramukh;
Tibetan, bya rgod gdong ma) | 313 Gt omits pa | 314 Gt inserts: na | 315 Gt inserts: kyi | 316 Gt inserts: / | 317 kang ka: Gt kang ka mu
ka (= Kakamukh; Tibetan, dur bya gdong ma) | 318 dkar dmar: QtGt dmar ba (the Tshal pa Kanjur tradition agrees with the
Dunhuang manuscript here) | 319 phrag pa g.yon pa na: Gt lag pa g.yas pa la (the Tshal pa Kanjur tradition agrees with the
Dunhuang manuscript here) | 320 khel: Gt 'gel | 321 g.yas pa: Gt g.yon pas (the Tshal pa Kanjur tradition agrees with the Dunhuang
manuscript here) | 322 Gt inserts: pa | 323 bkang ka 'i: Gt kang ka'i | 324 'go la: Gt mgo can | 325 Gt inserts: ba | 326 gug: Gt bkug pa
327 Gt inserts: ba | 328 ka ka: Gt khwa kha mu ka (= Kkamukh; Tibetan, bya rog gdong ma) | 329 g.yon bas: Gt g.yas pa (the
Tshal pa Kanjur tradition agrees with the Dunhuang manuscript here) | 330 pad mo: Gt padma | 331 Gt inserts: pa | 332 bya rog mgo
can: Gt omits (the Tshal pa Kanjur tradition also lacks this here, but gives it above. However, we would expect it here.) | 333 kyi:
Gt gyi | 334 u lu: Gt hu lu mu ka/ (= Ulkamukh; Tibetan, 'ug pa'i gdong ma) | 335 Gt inserts: / | 336 g.yon: Gt g.yas (the Tshal pa
Kanjur tradition agrees with the Dunhuang manuscript here) | 337 cen: Gt chen | 338 gang: Gt bkang | 339 la: Gt omits

Chapter 11

291

g.yas340 rdo rje lcags kyu ***** 'dzin /341

'ug pa'i mgo can / /
[4] nang rim 'ug pa'i 'khor gis342 bskor 343*****/
de 'i phyi rim bya rog gi 'khor gis344***** bskor /345
/de 'i 346 rim 347 kang ka348 mchu [5] rings kyi349 'khor kyis350 bskor 351/
/ de 'i phyi rim bya rgod kyi 'khor gyis bskor 352 /
de dag thams cad kyang zhing mang po gsar rnying spungs pa'i [6] steng du353 /
zhing la ltod354 pa'i phyir 'thab pa'i tshul ston pa she dag355 go /
/ de356 dag gi snying po ni [Gt275] 'di rnams so /
/h / ha / he / phat / h / 357
[45r.1] $/ /'di kun 'bru re re 358 yang lan bzhI bzhi brjod do /
/hung359 gis gtso bos360 bskyed /
/has nang gi lha mo bzhi361 bskyed362 /
he363 phyi 'i lha [2] mo bzhi364 bskyed /
/pha gyis365 thams cad khro bar366 bskul367 /
/h gyis368 thams cad ye shes kyi369 'od gser370 'bar bas371 phyogs bcur [3]
khyab par bgyi 'o /
/lha mo blo372***** ldan tshogs kyi bdag /
/he ru dpal373 la ***** phyag 'tshal lo/
/slobs dpon374 [4] la375 ni brnyas byed dang /
rnal 'byor bar ***** du gcod pa dang /
/dam tshig gnyan ***** po 'dral ba dang /
/rdo rje spun [5] la 'khu byed376 rnams/
/khyed377 kyis rdul du plags378 nas nI /
/ gnod sems ma lus gtan nas phyung /
/dpal chen sta na 'ga'379 na 'i [6] gnas su
/380tshogs chen po la gshegs pa'i tshe /
/sku gsung thugs kyi khro bo 'i gzi byin381 kyis /
/gdug pa thams cad gtan [45v.1] nas 'byin cing long spyod382 chen po 'i tshogs chen po la /
'gying383 zhing 'gying384 ba'i glu blangs te /
/thams cad dgyes par mdzad pa'i385 dbyangs [2] 'dis dgyes par mdzad do /
/rdo rje pad mo 'i386 phyag rgya 'phral bar387 bcings nas phyar te mchong zhing388 /
'khor thams cad gcer bur byung389 [3] nas/
skra bshig cing rgyug ste /390 *****
/sngags 'di gsungs so391 /
/o e 'i **** a ra li ho / hrng hrng hrng hrng / /392

340 g.yas:

Gt g.yon pa (the Tshal pa Kanjur tradition agrees with the Dunhuang manuscript here) | 341 'dzin /: Gt 'dzin pa | 342 gis:
Gt gyis |
inserts: ba | 344 gis: Gt gyis | 345 bskor /: Gt bskor ba | 346 Gt inserts: phyi | 347 Gt inserts: bya | 348 ka: Gt ka'i
349 Gt omits mchu rings kyi (the Tshal pa Kanjur tradition agrees with the Dunhuang manuscript here) | 350 kyis: Gt gyis | 351 Gt
inserts: ba | 352 Gt inserts: ba | 353 du: Gt na | 354 ltod: Gt stod | 355 she dag: Gt sha dag | 356 de: Gt 'di | 357 /h / ha / he / pha /
h / : Gt /h ha he pha h/ | 358 Gt inserts: la | 359 hung: Gt h | 360 bos: Gt bo | 361 bzhi: Gt omits (the Tshal pa Kanjur
tradition omits any enumeration here and below) | 362 bskyed: Gt bskyod | 363 he: Gt hes | 364 bzhi: Gt brgyad (the Tshal pa Kanjur
tradition omits) | 365 gyis: Gt kyis | 366 bar: Gt bor | 367 bskul: Gt skul | 368 gyis: Gt gis | 369 ye shes kyi: Gt omits (the Tshal pa
Kanjur texts support the Dunhuang manuscript here) | 370 'od gser: Gt 'od zer | 371 Gt inserts: / | 372 blo: Gt 'od (none of the root
texts give 'od) | 373 he ru dpal: Gt he ru ka (the Bhutanese texts also give he ru ka, but the other root texts support the Dunhuang
manuscript) | 374 slobs dpon: Gt slob dpon | 375 la: Gt lha | 376 byed: Gt ba | 377 khyed: Gt khyod | 378 plags: Gt brlag | 379 'ga':
final 'a subscribed | 380 /gnas su: Gt omits (none of the root texts support Gt here) | 381 gzi byin: Gt gzi brjid (the root texts are
divided here) | 382 long spyod: Gt / longs spyod | 383 'gying: Gt 'ging | 384 'gying: Gt 'ging | 385 pa'i: Gt pa yi/ | 386 pad mo 'i: Gt
padma'i | 387 'phral bar: Gt dpal 'bar | 388 mchong zhing: QtGtNt mchod cing (the root texts are divided here) | 389 byung: QtGtNt
phyung (the root texts are divided here) | 390 /: Gt omits | 391 gsungs so: Gt gsung ngo | 392 /o e 'i a ra li ho / hrng hrng hrng
hrng / : Gt /o e a ra li ho hri hri hri/ (BthHeOgl give hring hring hring hring for the second set of syllables)
343 Gt

292

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

[4] de nas tshogs su gshegs nas / ****

'khor thams cad la spyan bgrad nas393 ***** gzigs shing 'dzum [Gt276] ba'i394 brdas /395
[5] ha ha sang sang sgra phyung ste/
'khor thams cad kyis kyang396 mgrin gcig du/397
/dbyangs 'dis398 gtso bo la bstod do / /
[6] 399e ma 'o chos kyi don 'byung ba'400 /
/chos don dag pa bdag myed401 pa /
/ nam ka402 dang ni mnyam pa'i gnas/
/rdo rje rgyal la [46r.1] $/ / phyag 'tshal lo /
/zhes bstod nas403 gtso bo mnyes pas404 'khor thams cad la gzigs shing /
dam stsal pa'i405 dbyangs [2] 'di gsungs so /
/kha ga lhya/ sha ta ba ka / a la la ho/ ma ha pan tsa406 a mri ta407 /
kha ram yo gi ni /
kha ram408 ta ki ni /
[3] a la la ho /
bha ga bhan409/
kha kha kha/ ****
lha mo rnams rnga rdung410 zhing dbyangs len /
/rgyud gsum gyi glu tshig / /
[4] te ghem / te ghem / te ghem /411
'ghe****** tam / 'ghe tam / 'ghe tam /412
ta k+hem ****** ta k+hem / ta k+hem /413
k+hem k+hem [5] k+hem /414
/'phags pa thabs kyi zhags pa pad ma415 'phreng las /
rigs sngags gcod pa'i le'u ste416 bcu gcig pa 'o// : //

393 Gt

inserts: / | 394 ba'i: Gt pa'i | 395 brdas/: Gt rdas | 396 Gt inserts / | 397 du/: Gt tu | 398 'dis: Gt 'di | 399 QtGtNt insert: e ma'o
chos/ (the archetype most probably gave, e ma'o chos e ma'o chos/, which was perhaps shortened to the Tenjur version, then lost
completely in the Dunhuang, Tshal pa Kanjur and Bhutanese texts) | 400 ba': Gt ba | 401 myed: Gt med | 402 nam ka: Gt nam mkha'
403 bstod nas: Gt brjod pas (none of the root texts support Gt here) | 404 pas: Gt nas/ | 405 stsal pa'i: Gt gsal ba'i (none of the root
texts support Gt here) | 406 pan tsa: Gt panytsa | 407 a mri ta: Gt a m ta | 408 ram: Gt ra | 409 bhan: Gt bn | 410 rdung: Gt
brdung | 411 te ghem / te ghem / te ghem /: Gt ta kem ta kem ta kem/ | 412 'ghe tam / 'ghe tam / 'ghe tam/: Gt ge tem ge tem/
413 ta k+hem ta k+hem / ta k+hem / : Gt ta kem ta kem/ | 414 k+hem k+hem k+hem / : Gt kshem kshem kshem/ | 415 pad ma: Gt
padma | 416 Gt inserts /

Chapter 12
[Ms46r.6] [Gt276.5]
da ni rdo rje srin po rnams kyi dkyil 'khor bshad de /
/sems dpa' chen po nyon cig /1
zhes2 gsungs pa ni /
rdo rje srin po 'i dkyil [46v.1] 'khor bshad pa'i phyir /
/rdo rje srin po3 la bos pa 'o4/
/5 bskal pa'i6 mye7 ltar 'bar ba 'i8/
/'od zer 'du 'phro 'i klong dkyil na9/ /
[2] khro bo chen po rnams bzhugs pa /
/ brkyang bskum tshul du bzhugs10 pa [s(/l)]tie11/
zhes gsungs te /
bdud dang bgegs la stsogs [3] pa [Gt277] spa 'gong12 zhing brlhag13 par 'dul14 ba'i **** phyir/
'jig rten gyi khams thams cad ***** mye phung15 chen po 'bar zhing16 dug com kyi17 [4] sgra 'i18 nang du /
/'od gzer19 sna tshog*****s 'phro zhing 'khyil pa'i dkyil20 na / *****
/khro bo 'i gnas21 'byung ba'i rnams22 [5] zhabs brkyang bskum gyi tshul du /
/23khro gnyer sdang myig du24 gzigs shing bzhugs pa'i don 25/
/gzhan du na rtag chad kyi mtha'26 dang /
[beneath the line] ting nge 'dzin kyi ltar bya na bskal pa'i mye ltar 'bar ba ni

[6] mtshan ma 'I spyod yul dang /

lta ba'i gnas drug cu rtsa gnyis ni /
yang dag pa'i dbyings nyid la 'jug pa'i bar du gcod pas27 bgegs su [47r.1] $/ / gyur28 la / spyir myi rtog29 par
gyur pas /30
bdud de dag 'joms zhing31 sreg32 pa'i shes rab kyi mye33 la bya /
/ye shes lnga 'i 'od gzer34 [2] sna lngas35 thogspa36 myed37 par 'jug pa dang /
[beneath line 1] 'od gzer 'phro ba ni

dbyings nyid la 'khyil te /

/38gnyis su myed39 pa'i ting nge 'dzin du tham40 nas/
mtha'41 las 'das te / /42
[3] sku myi dmyigs pa ni43 dkyil lo / *****
/don gong ma ltar go ba ni sgra ste /
kun du44 khyab **** pa'i phyir sgra chen po zhes bya 'o/ /
[4] ye shes 45 tha dad par rig cing go ba ni /46 ****
sgra sna tshogs 47 yin zhes /48
gcig las49 ***** 'phros pa las 'byung ngo /50
1

/: Gt omits | 2 zhes: Gt ces | 3 srin po: QtGt sems dpa' (the Tenjur reading would seem more likely here; the Dunhuang
manuscript reading most probably derived from a repetition) | 4 pa 'o: Gt so | 5 QtGt insert: h (none of the root texts support
QtGt here) | 6 pa'i: Gt pa | 7 mye: Gt me | 8 'i: Gt yis | 9 na: Gt nas | 10 bzhugs: Gt gnas (the root texts support Gt here) | 11 [s(/l)]tie:
A correction has been made here, but the intention is not clear; it seems that the head letter la has been amended to sa; both a gi gu
and a 'greng bu are given, perhaps the intention was to correct the gi gu into the 'greng bu; Gt ste | 12 spa 'gong: Gt spa skong
13 brlhag: perhaps brlag intended, and possibly, the ha beneath was intended to indicate something else; Gt brlag | 14 'dul: Gt gdul
15 mye phung: Gt me dpung | 16 Gt inserts / | 17 dug com kyi: Gt tug com gyi (possibly, tug chum?) | 18 sgra 'i: Gt sgra | 19 'od gzer:
Gt 'od zer | 20 dkyil: Gt dbus | 21 gnas: Gt 'og nas (the Dunhuang manuscript is perhaps clearer here) | 22 ba'i rnams: Gt ba rnams/
23 / /: Gt omits | 24 sdang myig du: Gt sdang smig tu | 25 Gt inserts: to | 26 mtha': 'a subscribed | 27 Gt inserts / | 28 gyur: Gt 'gyur
29 myi rtog: Gt mi rtogs | 30 /: Gt omits | 31 zhing: Gt shing/ | 32 sreg: Gt skrag (the Dunhuang manuscript is perhaps clearer here)
33 mye: Gt me | 34 'od gzer: Gt 'od zer | 35 lngas: Gt lnga las/ | 36 thogspa: Gt thogs pa | 37 myed: Gt med | 38 / /: Gt omits | 39 myed:
Gt med | 40 tham: Gt thim (the Dunhuang manuscript reading if 'tham is intended is perhaps clearer here, although Gt's reading
could also fit) | 41 mtha': 'a subscribed | 42 / /: Gt omits | 43 myi dmyigs pa ni: Gt mi dmigs pa | 44 du: Gt tu | 45 Gt inserts: lnga
46 /: Gt omits | 47 Gt inserts: pa | 48 zhes /: Gt ces | 49 las: Gt la | 50 'phros pa las 'byung ngo /: Gt spros pa'i tan tra las 'byung ste/

294

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/brkyang ba [5] ni shes rab ste51 dbyings las dmyigs52 pa'i don to /
/ bskum ba53 ni thabs te /
dbyings nyid dang54 myi gnyis55 par chud pa'i don to /
/dbus kyi [6] khrag thung56 chen po57 ni /
/dbu dgu phyag kyang58 bco brgyad ste59/
/gshog pa rdo rje gtams pa gdengs60 /
/zhabs kyang [Gt278] brgyad du rnam par [47v.1] bstan /
/ zhes gsungs te /
khro bo 'i 'og61 gnas62 'byung ba'I63 nadbus na /
/64he ru ka gtso bor bzhugs te /
/ dbu dang po dbus dkar /
[2] g.yas dmar /
g.yon ljang ku65 /
bar ma dbus ser /
g.yas sngo skya66 /
g.yon dmar skya /
tha ma dbus nag po/
g.yas dmar smug /
[3] g.yon ljang gnag67 go /
dbu skra kham gna****g ral pa68 gyen du 'khyil ba /
/thod pa ska **** po 'i 'phreng bas brgyan pa 'o / /
[4] sku mdog ni69 nag po zhes kyang 'byung ste *****/
yong mdog dang sku ni70 cir71 yang 'gyur zhes72 'chad pa mang 73/
/thod pa rlon pa'i [5] 'phreng bse74 rel khar thogs pa /
/phyag 75 g.yas g.yos76 dang po gnyis na77/
/rdo rje 78 dung chen khrag gang79 gis gang80 ba dkrug cing gsol 81
[6] de 'I 'og ma gnyis na82 sti ra 83 chen po 'i84 rlon pa'i 85 snying 'byin cing gsol 86/
/de 'i 'og ma gnyis na87 sti ra chen po 'i nang zhags par bsdogs88 [48r.1] $/ / te/
lam rgyud lnga 'i sems can thams cad la 'debs te89 dbyings la sbyor zhing lta na byed /
/de'i 'og ma gnyis na/
sti ra chung ngu90 [2] chang par bcangs te91 ril myid mdzad 92/
/de 'I 'og ma gnyis na93 nyi ma dang94 zla ba'i dkyil 'khor95 thugs kar 'dzin /
/ de 'i 'og ma gnyis na [3] cir yang 'gyur ba'i mtshon ca thogs96/ ****
/de 'I 'og ma gnyis na /
'phra men ma97 mgyo***gs ma zhes bya ba dang /
gsod ma zhes [4] bya ba'i98 khyung dang 'ug pa lta bur sprul **** pa'I99 lta khra thogs nas /
[Gt279] dam las 'gal **** ba la100 gtong ngo zhe 'o101 /
de 'i 'og ma [5] gnyis na rlung dang mye 'i102 dkyil 'khor thogs te103/

51 Gt

inserts / | 52 las dmyigs: Gt la dmigs | 53 ba: Gt pa | 54 dang: Gt la | 55 myi gnyis: Gt mi dmigs (the Dunhuang manuscript
reading is perhaps clearer here, although Gt's reading could also fit) | 56 thung: Gt 'thung | 57 po: Gt bo | 58 kyang: Gt rgya (a clear
error in Gt) | 59 ste: Gt do | 60 pa gdengs: Gt pas rdongs (rdongs is clearly in error) | 61 'i 'og: correction below the line, marked by
cross above it | 62 gnas: Gt nas | 63 ba'i: 'a subscribed | 64 / /: Gt omits | 65 ljang ku: Gt ljang khu | 66 sngo skya: Gt sngo | 67 gnag:
Gt nag | 68 kham gnag ral pa: Gt ral pa kham nag | 69 ni: Gt omit | 70 yong mdog dang sku ni: Gt sku yang ni mdog dang sku
71 cir: Gt ci | 72 zhes: Gt ces | 73 Gt inserts: ngo | 74 bse: a tiny dot after ba may in fact be intended as a tsheg, so perhaps, ba se is
intended; Gt ba se (that is: 'phreng ba se rel khar) | 75 Gt inserts: bco brgyad la/ | 76 g.yos: Gt g.yon | 77 na: Gt ni | 78 Gt inserts:
dang | 79 gang: Gt omits (a clear error in the Dunhuang manuscript) | 80 gang: Gt bkang | 81 Gt inserts: ba/ | 82 Gt inserts / | 83 Gt
inserts: zhing | 84 'i: Gt omits | 85 Gt inserts: glo | 86 Gt inserts: ba | 87 Gt inserts: / | 88 zhags par bsdogs: Gt zhags pa bsngogs | 89 Gt
inserts / | 90 chung ngu: Gt chung ngun | 91 Gt inserts / | 92 myid mdzad: Gt mid mdzad pa | 93 Gt inserts / | 94 dang: Gt omits | 95 zla
ba'i dkyil 'khor: Gt zla ba | 96 mtshon ca thogs: Gt mtshon cha | 97 'phra men ma: Gt 'phra men | 98 ba'i: Gt ba | 99 'a subscribed
100 la: G omits | 101 ngo zhe'o: Gt zhes pa'o | 102 mye 'i: Gt me'i | 103 te: Gt pa ste

Chapter 12

295

'jig rten thams cad g'thor104 zhing bsreg105 go /

/ de 'i 'og ma gnyis na106 chu 'i dkyil 'khor [6] dang ri thogs te107/
/'jig rten gyi khams gnon cing g.yengs108 par mdzad do /
/de 'i 'og ma gnyis na109 rin po che 'i mdzod dang /110
yid bzhin [48v.1] gyi za ma tog thogs te111/
/rnal 'byor pa rnams la112 cis kyang myi113 tshugs pa'i go ca114 la stsogs115 pa dang /
/ dngos grub rnam gsum yang thob [2] par mdzad116 do /
/gzhan du na mthar gyis gnas pa'i snyoms par 'jug par117 dgu/
chos nyid dang tha myi118 dad la119 rims su yang120 snang ba 'o /
[3] phyag kyang khams bco brgyad /121
stong pa bco *** brgyad du rig pa de nyid /
/ye shes kyi dbco ** brgyad kyis122 dngos po 'jig pa'i phyir ro/
[4] phyag bco brgyad dang sbyar te /
de ni thabs ***** yin no /
/spyan dgu ni chos nyid las ***** gzhan du myi gzigs so / /
[5] /spyan bco brgyad ni123 stong124 pa bco brgyad la gzigs125 pas shes rab bo /
/de ltar thabs dang shes rab 'brel pas /
spyan dang snyan dang shangs dang126 [6] ljags dang /
sku dang127 thugs kyang stong pa bco brgyad du 'jigs128 pas /
/bco brgyad phrag drug ste129 khro bo brgya' rtsa brgyad do /
/khro bos [49r.1] $/ /gzhigs130 pa'i yul gzugs dang sgra dang dri dang ro dang131 reg bya dang /
chos kyang132 stong pa bco brgyad phrag drug ste /
/ de ni dmyigs133 pa'i [Gt280] yul [2] yin pas /
khro mo yang134 brgya' rtsa brgyad do /
/de lta bu she dag135 ste/
chos thams cad kyang136 khro bo dang khro mo yin par mkhas pas137 rig ste /138
[3] grangs myed do139 zhes gcig las 'phros **** pa140 las 'byung ngo / 141
/gshog pa ni rdo rje **** gtams142 pa ste143 'jig rten khebs144 pa [4] tsam du gdengs pa 'o /
/gzhan du na ***** chos kyi dbyings las145 ye shes 'phros ***** pa146 dang /
/147thams cad du khyab pa'i148 [5] 149 pa'I don to /
/zhabs brgyad kyis g.yas g.yos150 dang po151 gnyis ni /
ma ha ka la152 pho mo la153 brdzis pa 'o /
/de 'I 'og ma gnyis ni ma ha [6] de ba pho mo la brdzis pa 'o/154
/de 'i 'og ma gnyis ni tshogs kyi bdag po pho mo la155 brdzis pa 'o /
104 'thor:

Gt mthor (the Dunhuang manuscript is perhaps more likely here) | 105 bsreg: Gt sreg | 106 Gt inserts / | 107 te: Gt ste
Gt g.yeng | 109 Gt inserts / | 110 /: Gt omits | 111 thogs te: Gt ste | 112 la: Gt ya | 113 myi: Gt mi | 114 go ca: Gt gro cha
115 stsogs: Gt sogs | 116 deleted na ro above mdzad, some smudging of the syllable | 117 par: Gt pa | 118 myi: Gt mi | 119 Gt inserts /
120 yang: Gt omits | 121 phyag kyang khams bco brgyad /: Gt omits | 122 ye shes kyi dbco brgyad kyis: QtGt ye shes brgyad kyis
phyi nang gi (presumably, the Dunhuang manuscript's dbco is simply for bco; possibly, the prefixed da might once have been
intended as a shad, although that would not seem very appropriate here. QtGt's addition of phyi nang gi may be a useful addition,
but brgyad is most probably in error for bco brgyad) | 123 Gt inserts / | 124 stong: Gt stod | 125 gzigs: Gt dmigs | 126 spyan dang
snyan dang shangs dang: Gt spyan dang/ snyan dang/ shangs dang/ | 127 dang: Gt gsung (the Dunhuang manuscript seems more
likely here) | 128 'jigs: Gt gzigs (Gt would seem to be correct here) | 129 Gt inserts / | 130 gzhigs: Gt gzigs (Gt would seem to be
correct here) | 131 yul gzugs dang sgra dang dri dang ro dang: Gt yul gzugs dang/ sgra dang/ dri dang/ ro dang/ | 132 chos kyang: Gt
omits | 133 dmyigs: Gt dmigs | 134 khro mo yang: Gt khro bo | 135 she dag: sha dag intended?; Gt sha stag | 136 kyang: Gt omits
137 Gt inserts / | 138 rig ste/: Gt rigs te | 139 myed do: Gt med do/ | 140 'phros pa: Gt spros pa'i tan tra | 141 Gt inserts: mkha' lding
gi rgyal po'i tshul du rdo rje gshog pas gdangs pa'o/ /phyag mang po gdengs pa la yang bya zhes kyang bshad/ (this
addition could fit appropriately, although the Dunhuang manuscript version remains coherent without it) | 142 gtams: Gt stams
143 Gt inserts / | 144 khebs: Gt gdengs | 145 Gt inserts / | 146 'phros pa: Gt 'phro ba | 147 / /: Gt omits | 148 pa'i: Gt pas | 149 Gt inserts:
phyin (omission in the Dunhuang manuscript is most likely) | 150 g.yos: Gt: g.yon | 151 dang po: Gt omit | 152 ma h ka la: Gt ma
ha ka la | 153 la: Gt gnyis | 154 /de 'i 'og ma gnyis ni ma ha de ba pho mo la brdzis pa 'o/: QtGt omits (QtGt is clearly in error here)
155 la: Gt gnyis
108 g.yengs:

296

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/ de 'i 'og ma gnyis ni156 byi na ya ka [49v.1] pho mo la157 brdzis pa 'o /
/gzhan du na158 rnam par thar pa brgyad la zhon pa'i tha tshig ste /
bsam gtan bzhi dang /
snyoms par 'jug pa bzhi [2] dang /
bsgom159 zhing /
/160chos nyid ma rtogs pa ni /
'jig rten gyi161 lha 'i bsam gtan no /
/chos nyid rtogs pa ni rnam par thar pa brgyad [3] du 'gyur te /
/de 'i phyir 'jig rten gyi **** lha brgyad zhabs kyis brdzis pa ni 162/ ** / ***
/chos nyid163 rtogs164 pas rnam par thar pa [4] brgyad du 'gyur te / / la zhon zhes ***** bya ba'i don to /
/de 'i dkyil 'kho*****r khro bo ni /
/dbu gsum phyag [5] drug gshog pa can /
/ zhabs drug rdo rje165 brag steng166 na /
[Gt281] /'jig rten skyong ba167 'dul168 tshul lo /
/zhes gsungs te /
shrI169 he ru ka 'i [6] 'khor gyi khro bo170 /
/shar phyogs su khro bo rnam par rgyal ba sku mdog dkar /171
dbu gsum172 dkar 173/
g.yas sngo /174
g.yon ljang ser /
[50r.1] $/ /phyag g.yas g.yos175 dang po gnyis na /
/rdo rje 176 dung cen177 dmar gyis gang ba178 dkrug cing gsol 179/
/de 'i 'og ma gnyis na /
tri shu la180 dang [2] 'khor lo bsnams /
tha ma gnyis na181 sti ra 'i dbyug pa dang pha bang bsnams so182 /
/gshog pa 183 gong du bshad pa dang 'dra184 /
/rdo rje brag [3] gtams pa la brtan185 te /
'jig rten ***** skyong ba brgya' byin186 zhabs kyis brdzis te ***** bzhugs so /
/gzhan du na187 dbu [4] gsum ni188 rnam par thar pa189 gsum dang ***** sbyar ro /
/phyag drug ni ye shes drug 190 ***** yul drug la /191
myi dmyigs pa'i192 [5] tshul gyis myi193 'gyur bar brtan194 te /
brtan195 pa'i phyir 196 brag ces smos so /
/'jig rten skyong ba ni197 dge ba bcu 'i las kyis grub 198 la/ /
[6] de la bcibs199 pa ni200 dge ba bcu las 'das pa'i phyir201 mnyam ba202 nyid du203 spyod pa'i don to /

156 Gt

inserts / | 157 la: Gt gnyis | 158 na: Gt omit | 159 bsgom: Gt sgom | 160 / /: Gt omits | 161 gyi: Gt omits | 162 QtGt insert: 'jig
rten gyi lha'i bsam gtan chos nyid du mkhyen to (the Dunhuang manuscript seems coherent without this, although it could fit; it
is uncertain whether this is a repetition in error by QtGt, or an omission by the Dunhuang manuscript) | 163 chos nyid: Gt chos
164 rtogs: Gt rtags | 165 rdo rje inserted in small writing as a correction below the line, the positioning marked by a cross.
166 steng: Gt stengs | 167 ba: Gt bas | 168 'dul: Gt 'dud (note that the TZ editions agree with Gt here; the Dunhuang manuscript
seems to make rather better sense but it is probably in error) | 169 shrI: Gt shr | 170 QtGt insert: la re re la yang yan lag nyi shu
rtsa bdun/ phyag rgya lnga bcu rtsa bzhi dang ldan par bstan no (it is hard to know what to make of this addition; it appears
rather incoherent here.) | 171 /: Gt ba | 172 gsum: Gt dbus (we might expect gsum la dbus here; see below for the parallel
descriptions of the other wrathful male deities) | 173 Gt inserts: ba | 174 /: Gt omits | 175 g.yos: Gt g.yon | 176 Gt inserts: dang
177 cen: Gt chen | 178 gang ba: Gt bkang pa | 179 Gt inserts: ba | 180 tri shu la: Gt tri shul | 181 Gt inserts / | 182 so: Gt pa'o | 183 Gt
inserts: ni | 184 'dra: Gt mthun no | 185 brtan: Gt brten | 186 brgya' byin: Gt brgya byin | 187 du na: Gt omits | 188 Gt inserts / | 189 pa:
Gt pa'i sgo | 190 Gt inserts: dang sbyar ro/ /zhabs drug ni ye shes drug/ (there appears to an eyeskip at this point in the
Dunhuang manuscript; comment on the zhabs drug is needed, so Gt is surely correct.) | 191 /: Gt omits | 192 myi dmyigs: Gt mi
dmigs | 193 myi: Gt mi | 194 brtan: Gt bstan | 195 brtan: Gt bstan | 196 Gt inserts: rdo rje (Gt is almost certainly correct here)
197 Gt inserts / | 198 Gt inserts: pa | 199 bcibs: Gt bcings (the Dunhuang manuscript reading seems more appropriate here) | 200 Gt
inserts: / | 201 Gt inserts: / | 202 ba: Gt pa | 203 du: Gt la

Chapter 12

297

/de bzhin du khro bo dang204 khro mo 'i205 'og206 gnas207 'byung ba'i phyag [50v.1] mtshan rnams kyang tshul
de ltar sbyar/ ro/208
/shar lho 'i209 mtshams su be con sngon po can210/
/sku mdog sngo skya /
dbu gsum la [Gt282] dbus sngo /
[2] g.yas dkar211 g.yon dmar skya212/
/phyag g.yas g.yos213 dang po gnyis214 215/
rdo rje dung cen216 khrag gis gang ba217 dkrug cing gsol 218/
/ bar ma gnyis219 [3] dbyug pa rin po che dang mye dbal220 bsnams ****/
tha ma gnyis 221 sti ra 'i dbyug pa dang khwa bsna*****ms so222/
/zhabs223 drug mye lha la224 [4] brdzis pa 'o /
/gzhan ni gong ma dang ***** 'dra /
/lho phyogs su gshin rje gshed ***** po ste /
/sku mdog sngon po / /
[5] dbu gsum la dbus sngo225 /
/g.yas sngo skya /
g.yon ljang gnag /
phyag g.yas g.yos226 dang po gnyis 227 rdo rje dung cen228 khrag gis gang ba la229 [6] dkrug cing gsol 230/
/bar ma231 gnyis 232 dbyug pa myi233 mgo can dang dgra sta bsnams /
/tha ma gnyis sti ra 'i234 dbyug pa dang bya rog bsnams/
[51r.1] $/ /zhabs drug 'gshin rje235 la brdzis te bzhugs /
/lho nub mtshams na myi g.yo mgon po /236
sku mdog ljang sngon / /
[2] dbu gsum la /237
dbus ljang sngon /
g.yas sngon po/
g.yon smug gnag238/
phyag g.yas g.yos239 dang po gnyis 240/
rdo rje dung cen241 khrag242 gis [3] gang ba243 dkrug244 cing gsol 245/
bar ma gnyis 246 **** ral gyi247 dang 248 zhags bsnams /
tha ma gnyi*****s 249 sti ra 'i dbyug pa dang 'ug pa bsnams/
[4] zhabs drug bden bral la250 brdzis pa 'o/ *****
/nub phyogs su ha ya gri ba251 sku mdog ***** dmar 252/
dbu gsum la dbus dmar [5] g.yas dkar /
g.yon sngo253/
phyag drug la g.yas g.yos254 dang po gnyis 255/
[Gt283] rdo rje dung cen256 khrag gis gang ba257 dkrug cing gsol 258/ /
[6] bar ma gnyis 259 zhags pa dang spu gri bsnams 260/
/tha ma gnyis 261 sti ra 'i dbyug pa dang /
nam ka262 lding bsnams 263/
204 dang:

Gt omits | 205 'i: Gt omits | 206 'og inserted beneath the line, with a cross above the line to mark position. | 207 gnas: Gt
nas |
ro/: Gt sbyar ro/ | 209 'i: Gt omits | 210 can: Gt omits | 211 dkar: Gt kar/ | 212 dmar skya: Gt dmar | 213 g.yos: Gt
214
g.yon |
prefixed ga inserted in small writing beneath the line | 215 Gt inserts: na | 216 cen: Gt chen | 217 gang ba: Gt bkang nas
218 Gt inserts: ba | 219 Gt inserts: na | 220 mye dbal: Gt me rbal | 221 Gt inserts: na | 222 so: Gt pa | 223 zhabs: Gt zhag | 224 mye lha
la: Gt me lha | 225 sngo: Gt sngon po | 226 g.yos: Gt g.yon | 227 Gt inserts: na/ | 228 cen: Gt chen | 229 gang ba la: Gt bkang nas
230 Gt inserts: ba | 231 bar ma: Gt bar | 232 Gt inserts: na | 233 myi: Gt mi | 234 sti ra 'i: Gt na sta ra'i | 235 'gshin rje: Gt gshin rje
236 na myi g.yo mgon po /: Gt su mi g.yo mgon po | 237 /: Gt omits | 238 gnag: Gt nag | 239 g.yos: Gt g.yon | 240 Gt inserts: na
241 cen: Gt chen | 242 khrag: Gt khrug | 243 gang ba: Gt bkang nas | 244 dkrug: Gt dkrag | 245 Gt inserts: ba | 246 Gt inserts: na
247 gyi: Gt gri | 248 Gt inserts: dar | 249 Gt inserts: na | 250 la: Gt omits | 251 ha ya gri ba: Gt ha ya gr ba/ | 252 Gt inserts: po
253 sngo: Gt sngon po | 254 drug la g.yas g.yos: Gt omits | 255 Gt inserts: na | 256 cen: Gt chen | 257 gang ba: Gt bkang nas | 258 Gt
inserts: ba | 259 Gt inserts: na | 260 Gt inserts: pa | 261 Gt inserts: na | 262 nam ka: Gt nam mkha' | 263 Gt inserts: pa
208 sbyar/

298

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/zhabs264 drug klu la [51v.1] brdzis pa 'o/

/nub byang gi mtshams na /265
gzhan gis myi thub pa266 sku mdog dmar skya /
dbu gsum la dbus dmar skya /267
g.yas ser po268 [2] g.yon ljang ku269 /
phyag drug g.yas g.yos270 dang po gnyis 271 rdo rje dung cen272 khrag gis273 gang274 ba dkrug cing gsol 275/
/bar ma gnyis 276 ba dan277 dang [3] rngul yab278 bsnams 279/
tha ma gnyis 280 **** sti ra 'i dbyug pa dang ske ga281 bsnams 282/ ****
/ zhabs drug be rem ba283 la brdzis284 pa [4] 'o /
/byang phyogs su bdud rtsi 'khyil **** ba285 /
sku mdog ljang gnag286 /
dbu gsum ***** pa287 dbus ljang gnag288/
g.yas ser [5] po /
g.yon dud ka /
phyag drug g.yas g.yos289 gnyis 290 rdo rje dung cen291 khrag gis gang ba292 dkrug cing gsol 293/
/bar ma gnyis 294 dbyug tho295 [6] dang /
rdo rje296 rgya gram bsnams 297/
/tha ma gnyis 298 sti ra 'i dbyug pa dang phub shud299 bsnams /
/ zhabs drug gnod sbyin brdzis pa 'o / /
[52r.1] $/ /byang shar mtshams su khams gsum rnam par rgyal ba /
sku mdog ser po /
dbu gsum la /300
dbus ser 301/
g.yas dmar / /
[2] g.yon ljang ku302 /
phyag drug g.yas g.yos303 dang po gnyis 304/
rdo rje dung cen305 khrag gis gang306 ba dkrug cing gsol 307/
[Gt284] bar ma gnyis 308 gzhu309 dang [3] mda' bsnams /
/tha ma gnyis 310 ***** sti ra 'i dbyug pa dang srin bya bsnams / *****
/zhabs drug dbang ldan la311 brdzis pa [4] 'o /
/'og gyI312 phyogs su stobs ***** po che /313
sku mdog dud ka /
dbu gsum ***** la dbus dkar314 /
g.yas nag po /
[5] g.yon dmar po /
phyag drug g.yas g.yos315 dang po gnyis 316 rdo rje dung cen317 khrag gis gang318 ba dkrug cing gsol 319/
/bar ma gnyis zhags [6] pa dang tho ba320 bsnams 321/
264 zhabs:

Gt zhags | 265 /: Gt omits | 266 gzhan gis myi thub pa: Gt gzhan gyis mi thub pa'i | 267 dbu gsum la dbus dmar skya /: Gt
omits |
inserts: / | 269 ljang ku: Gt ljang gu | 270 drug g.yas g.yos: Gt omits | 271 Gt inserts: na | 272 cen: Gt chen | 273 khrag
gis: Gt omits | 274 gang: Gt bkang | 275 Gt inserts: ba | 276 Gt inserts: na | 277 ba dan: QtGt ban dha (the sources vary, but some give
ba dan as a characteristic implement of this deity, so the Dunhuang reading would seem more likely here, especially since the first
hands already hold a dung chen) | 278 rngul yab: presumably, for rnga yab, which would be expected for this deity (Gt shares the
reading, rngul yab) | 279 Gt inserts: pa | 280 Gt inserts: na | 281 ske ga: Gt skra ga (almost certainly, skye ka, archaic for magpie, is
the intended reading) | 282 Gt inserts: pa | 283 be rem ba: Gt be ra ma (it is most probable that Varua[s] is indicated here)
284 brdzis: Gt rdzis | 285 ba: Gt pa | 286 gnag: Gt nag | 287 pa: Gt la | 288 ljang gnag: Gt ljang | 289 g.yos: Gt g.yon dang po | 290 Gt
inserts: na/ | 291 cen: Gt chen | 292 gang ba: Gt bkang pa | 293 Gt inserts: ba | 294 Gt inserts: na | 295 dbyug tho: Gt dbyug to | 296 /
rdo rje: Gt omits | 297 Gt inserts: pa | 298 Gt inserts: na | 299 phub shud: Gt pu shud | 300 /: Gt omits | 301 Gt inserts: po | 302 ljang
ku: Gt ljang gu | 303 g.yos: Gt g.yon | 304 Gt inserts: na | 305 cen: Gt chen | 306 gang: Gt bkang | 307 Gt inserts: ba | 308 Gt inserts:
na | 309 gzhu: zhabs kyu indistinct | 310 Gt inserts: na | 311 la: Gt omits | 312 gyi: Gt gi | 313 /: Gt omits | 314 dkar: Gt dud ka (Gt's
reading here appropriately fits with the main body colour) | 315 g.yos: Gt g.yon | 316 Gt inserts: na/ | 317 cen: Gt chen | 318 gang: Gt
bkang | 319 Gt inserts: ba | 320 zhags pa dang tho ba: Gt tho ba dang zhags pa | 321 Gt inserts: pa
268 Gt

Chapter 12

299

/tha ma gnyis gtun shing dang gshol322 bsnams /

/zhabs drug phag chen la323 brdzis pa 'o /
/ steng gi [52v.1] phyogs su hung ka ra324 /
sku mdog sngo skya /
dbu gsum la dbus sngo skya /
/ g.yas ser po /
g.yon dmar po/
phyag drug g.yas g.yos325 [2] dang po gnyis /326
rdo rje dung cen327 khrag gis gang328 ba dkrug cing gsol 329/
/bar ma gnyis 330 mda' gzhu bsnams /
tha ma gnyis 331 tho ba dang chu gri bsnas [3] so332/
/zhabs drug gnyi333 zla dang334 gza'335 ***** skar la336 brdzis pa 'o /
/'di337 dag thams ***** cad kyang338 thod skam kyi339 dbu rgyan dang340 /
[4] thod rlon kyi341 'phreng ba can342 she dag343 go / *****
[The next passage of text given in the Dunhuang manuscript, from 52v.4 up to 53v.1, belongs to Chapter 13.
It is found in the Tenjur's Chapter 13, and moreover, the root text verses are found in Chapter 13 of the
different versions of TZ, not Chapter 12. The inclusion of the passage at this point in the Dunhuang
manuscript was presumably due to a folio misplacement at some stage of the text's transmission, so we have
moved this text to its correct placement in Chapter 13 of this edition, and we continue here with the text from
53v.]
[53v.1] [Gt284.5] /de 'i pho nya bka'344 nyan ni /
/rdo rje 'phra men gtum bag can /
/'jigs pa'i gzugs can mang po ste /
[2] /de dag cir yang grub345 par byed /
ces gsungs te /
'phra men ma stag 346 mgo can dang /
g.yag 347 mgo can 348/
sha ba'i mgo can 349 /
gzig 350 mgo [3] can 351/
[Gt285] bya ba'i352 mgo can 353 /
spyang ka 'i354 mgo can 355 /
seng ge 'i mgo can /356
dom gyi mgo can 357/ ***
/dred kyi mgo can 358 /
phag gyi359 mgo [4] can rnams te /
'di360 dag lag pa361 g.ya*****s pa na362 sti ra thogs te363 za zhing rgyug / ****
/ g.yon na364 cir yang 'gyur ba'i365 [5] tshon ca366 thogs nas /
/khro bo 'i bka'367 bzhin du368 las byed do369 /
/gzhan du na rdo rje phra mo370 ste/
sems dang sems las byung ba'i chos lta bur371 [6] ye shes dang ye shes lta bu 'i372 byin brlabs373 las byung ngo
/
322 gtun

shing dang gshol: Gt gshol dang gtun shing | 323 chen la: Gt omits | 324 hung ka ra: Gt h ka ra | 325 g.yos: Gt g.yon
Gt na | 327 cen: Gt chen | 328 gang: Gt bkang | 329 Gt inserts: ba | 330 Gt inserts: na | 331 Gt inserts: na | 332 so: Gt pa | 333 gnyi:
Gt nyi | 334 dang: Gt omits | 335 gza': 'a subscribed | 336 la: Gt omits | 337 'di: Gt de | 338 kyang: Gt omits | 339 kyi: Gt gyi | 340 dang:
Gt byas pa | 341 kyi: Gt gyi | 342 can: Gt omits | 343 she dag: Gt sha stag | 344 bka': 'a subscribed | 345 grub: Gt 'grub | 346 Gt inserts:
gi | 347 Gt inserts: gi | 348 Gt inserts: dang | 349 Gt inserts: dang | 350 Gt inserts: gi | 351 Gt inserts: dang | 352 bya ba'i: Gt byi la'i
(Gt's reading is presumably correct here) | 353 Gt inserts: dang | 354 spyang ka 'i: Gt spyang ki'i | 355 Gt inserts: dang | 356 seng ge 'i
mgo can/: QtGt omits (QtGt are clearly in error) | 357 Gt inserts: dang | 358 Gt inserts: dang | 359 gyi: Gt gi | 360 'di: Gt de | 361 lag
pa: Gt omits | 362 na: Gt ni | 363 te: Gt nas | 364 na: Gt pa ni | 365 ba'i: Gt pa'i | 366 tshon ca: Gt mtshon cha | 367 bka': 'a subscribed
368 du: Gt omits | 369 do: Gt pa'o | 370 phra mo: Gt 'phra men ma | 371 bur: Gt bu ste/ | 372 lta bu 'i: Gt kyi | 373 byin brlabs: Gt byin
rlabs
326 /:

300

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/374 thabs kyi zhags pa pad ma375 'phreng las /

/khro bo 'i dkyil 'khor 376 gyi377 [54r.1] $/ /le'u ste bcu gnyis pa 'o// :

374 Gt

inserts: 'phags pa | 375 pad ma: Gt padmo | 376 Gt inserts: thams cad (the Tshal pa Kanjur and Bhutanese TZ texts agree
with thams cad, but the other root texts do not include it) | 377 gyi: Gt kyi

Chapter 13
[Ms54r.1] [Gt285.3]
//da ni khro mo 'i1 dkyil 'khor bshad de /
/rdo rje srin mo chen mo ni /
[2] 'jigs byed chen mor snang ba ste / 2
/mthun te zhabs kyang de dang 'dra /
/zhes gsungs pa ni3/
/khro mo rnams kyi sku dang /4
phyag dang zhabs khro bo [3] dang 'dra /
/myi mthun bar5 phyag g.ya*****s dang po6 dril bu thogs so7/
/gzhan du ****** na dril ni8 shes rab/
rdo rje ni9 thabs so/
[4] de zung du 'brel pa las /
[prur?]10 mo dang11 ***** rdo las mye phyung12 nas btsa' la13 bsregs14 pa'I ***** tshul du /
/ pho nya 15dang pho nya mo [5] rnams16 byung nas17 ci yang brlag par byed ces bya ba'I18 don to/
/de la khro mo ni gtso mo ral pa19 gcig ma /
/shar phyogs su rdo rje bsnyems20 ma/
[6] de bzhin du rdo rje 21 dur khrod dbdag22 mo dang23/
rdo rje gtung24 khung ma dang /
rdo rje gtum mo dang /
rdo rje mda' snyegs25 ma dang /
[Gt286] rdo rje sgra 'byin ma dang /
[54v.1] rdo rje gsod ma dang /
rdo rje skul byed ma dang /
rdo rje rlung 'byin ma
_________________________________________________________________
[QtGtNt give a passage here, which is not present in the Dunhuang manuscript. There seems little doubt that
the Dunhuang manuscript has suffered from an omission. However, QtGtNt's passage does not seem fully
apt either, and seems to have been somewhat corrupted. The verb immediately following the Dunhuang
omission, snams (or bsnams in Gt), would suggest that the goddesses listed in the previous passage are
holding implements which have just been described, but no implements at all are mentioned in the Dunhuang
version of the text. The need for some text here is also supported by the following reference to eleven/twelve
sense-bases, which most probably relates to the twelve inner messengers listed in the text included by the
Tenjur. Moreover, in Chapter 18 (Dunhuang manuscript f.62r), there is a possible reference back to some of
the content of this list, when the text refers to sending out the Multiform Means emanations, such as the
Garua, the Raven, and (additionally in the Tenjur version,) the Vulture-headed (khyung dang khwa la stsogs
[+QtGtNt bya rgod mgo can la sogs] pa khra thabs su btang ste). However, the description given in
QtGtNt's passage does not seem entirely to fit with what follows, and it is not clear at the end of the passage
that we do have a list of the expected hand implements. Possibly, the passage is indicating rather that the
various further emanations are being held in the goddesses' hands, although this is not altogether clear.
Alternatively, perhaps some further text is missing. In any case, the passage as given cannot
1 mo

'i: Gt bo'i | 2 Gt inserts: phyag kyang srin po chen po ltar/ (this tshig rkang, a version of which is included in the root texts,
is certainly omitted from the Dunhuang text in error) | 3 pa ni: Gt pa'o | 4 /: Gt omits | 5 myi mthun bar: Gt mi mthun pa | 6 dang po:
Gt par | 7 so: Gt omits | 8 dril ni: Gt dril bu | 9 ni: Gt omits | 10 [prur?]: perhaps srur intended; Gt dur | 11 dang: Gt omits | 12 mye
phyung: Gt me byung | 13 btsa' la: Gt / brtsal | 14 bsregs: Gt bsreg | 15 Gt inserts: pho | 16 rnams: Gt omit | 17 Gt inserts: / | 18 bya
ba'I: Gt bya'i | 19 pa: Gt omits | 20 bsnyems: Gt snyems | 21 Gt inserts: sder mo/ rdo rje (there is no doubt that the Dunhuang
manuscript is in error in this omission.) | 22 dbdag: Gt bdag | 23 dang: Gt omits | 24 gtung: Gt gtun | 25 mda' snyegs: Gt bsnyems
(although mda' snyegs makes good sense, mda' snyems is the more usual name for this wrathful female.)

302

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

unproblematically be assumed to represent the original text entirely accurately. Unfortunately, we have no
other versions of the passage, so we simply present it here.]
[Qt116a.2; Gt 286.1; Nt204.2] dang / rdo rje brjod byed ma dang / rdo rje 'jigs byed ma dang / rdo rje [Nt
rdoe] bdud rtsi sbyor byed ma dang / rdo rje 'phrog ma rnams so/ /nang gi pho nya bcu gnyis ni rdo rje
rmongs byed ma dang [Gt omits dang]/ rdo rje yid 'phrog ma dang / rdo rje khrag 'jib ma dang / rdo rje che
'joms ma dang / rdo rje rmugs byed ma dang / rdo rje srin mo snying za ma dang / rdo rje skul byed ma dang
/ rdo rje 'bebs ma dpal bdag ma dang / rdo rje rkang 'jib ma dang / rdo rje gnas ma mgyogs byed ma dang /
rdo rje tshe [Qt che] sgrub ma dang / tshe 'dzin ma rnams so/ /'di rnams kun kyang rin po che dang / rdo rje
rgyu las byung ba'i gshog pa gdengs pa/ dbu dang sku la sogs pa/ thod rlon sbrul nag dang / zhing chen dang
/ chu srin gyi stod g.yogs dang / smad g.yogs dang / za byed kyi kha 'bar bas zhags pa kun gyi phyag na/
sbod gtong rdo rje nam mkha' lding dang / khra dang 'ug pa dang / bya rgod dang / byan lag dang / kang ka'i
lus la chu srin gyi mgo can/ gshog sbub dang spu [Nt sbu] gseb thams cad nas/ lcags kyi rgyun btabs pa 'dra
ba 'byung ba/ sder mong thams cad rdo rje lcags kyu /mchu rdo rje rtse gcig pa'i bzod dang / rdo rje so gri
dang / rdo rje kha gdengs dang / /rdo rje kar ma bud ka /rdo rje 'ching bar byed pa'i mchu dang / 'jigs par
byed pa'i gzugs can/ 'di lta bu phyag [Gt287] na
_________________________________________________________________
[54v.1]26 snams27 so /
/gzhan du na28 rnam par grol ba'i29 skye mched bcu gcig gis30 [2] yul gzhigs te31/
chos kyi dbyings lasu32 gyur pa la bya zhes /
rtse gcig bsdus pa las 'byung ngo /
/pho nya ni33 'phra men ma ste34 / /
[3] bya rgod mgo can dang /
bya rog mgo ***** can dang /35
'ug pa'i mgo can dang /
khwa 'I ***** mgo can dang /
phub shud36 mgo can [4] dang /
khra 'i mgo can dang /
pha wang37 ***** mgo can dang /
sre mo 'i38 mgo can dang / *****
/byi ba dang39 smrigs40 bu 'i mgo can [5] rnams so / i
/gzhan du na chos kyi dbyings kyi byin brlabs41 las 'byung zhes 'chad /
[The next passage of text was misplaced within Chapter 12 in the Dunhuang manuscript, from 52v.4 up to
53v.1. Since it clearly belongs here in Chapter 13, we are placing it here.]

It is worth noting that this list of the wrathful female emanations corresponds exactly to the standard order given for them in
many rNying ma tantras and practice traditions, apart from giving the Lizard-headed at the end rather than at the beginning of the
list. This alternative placement, however, fits with the order in which the male wrathful deities are listed above in Chapter 12,
where the male wrathful deity associated with the above direction is given at the end of the list, Ms52v.1, and so too his male
emanation, Ms53v.2.
26 Here

we resume the text in the Dunhuang manuscript. | 27 snams: Gt bsnams | 28 na: Gt omits | 29 ba'i: Gt ba yi/ | 30 bcu gcig gis:
QtGt bcu gnyis kyis (QtGt are possibly correct here, and the twelve would correspond to the set of twelve inner messengers listed
in the Tenjur's passage above. However, Chapter 18, Ms62r.2-3, gives rnam par grol ba'i skye mched bcu gcig, and there the
reading, bcu gcig, is shared by all copies of the Tenjur. It seems more likely that the correct intended number here is a
non-standard eleven, which the Tenjur emended to the standard list of twelve, perhaps unreflectively.) | 31 gzhigs te: Gt gzhig ste
32 lasu: Gt su | 33 ni: Gt omits | 34 ste: Gt rnams te | 35 Gt omits bya rog mgo can dang/ | 36 phub shud: Gt pu shud kyi | 37 pha
wang: Gt pha bang gi | 38 sre mo 'i: Gt sre mong gi | 39 byi ba dang: Gt bya'i mgo can dang/ (the standard list of these emanations
would support the Dunhuang manuscript's reading, and in any case, Gt's reading would seem too non-specific) | 40 smrigs: the
attached ra is faintly written which may possibly indicate an attempted deletion, but this is an alternative spelling, occuring also
below, 56r.5; Gt / smigs | 41 byin brlabs: Gt byin rlabs

Chapter 13

303

[52v.4] [Gt287.3] / de lta bu 'i42 dkyil 'khor du /

/rin ce******n43 me tog bor nas ni44 /
/gang la [5] bab pa'i lha de ni /
/gnyer na45 grub pa46 the tsom myed /47
/zhes gsungs te48 /
gong ma lta bu 'i dkyil 'khor du /
rin po che 'am49 [6] me tog bor te /
gang la bab pa'i50 lha de bsgrubs na 'grub bo51 zhes bya ba'i don to /
/ gzhan du na52 chos kyi dbyings khro bo dang53 khro mo
[53r.1] $/ /'i tshul du snang ba54 la /55
gnyis su myed56 pa'i ye shes kyis57 rig pa ni /
ye shes kyi58 rig pa'i59 me tog bor60 zhes bya ste / /
[2] khro bo dang61 khro mo thams cad dang62 bdag du63 tha myi64 dad pas grub65 pa yin no zhes /66
rtse gcig bsdus pa las 'byung ngo /
/ de nas slob ma67 [3] dam tshig bstan te /
/dge dang ***** myi dge 'I68 chos rnams la /
/lhag par ***** bsten 69 par myi bya 'o70 /
/ mnyam [Gt288] nyid [4] rig pa'i klung71 zhugs ste72 /
/ci la yang73 ***** thogs pa myed74 par spyod75 /
zhes bsgo 'o76 ****** zhes gsungs te /
/ lhag par [5] dge ba la gtsor spyod pa ni /
/nyan thos yin no77 /
/myi78 dge ba la gtsor spyod pa ni79 sems can phal pa80 yin no /
/theg pa chen [6] po ni81 chos thams cad kun rdzob du82 khro bo dang83 khro mor mnyam la /
/ don dam par skye 'gag myed84 par mnyam bas85 /
/ci la yang thogs [53v.1] pa myed86 par 'jug pa'i don to /
[Following the above passage of text which was misplaced in the Dunhuang manuscript, from 52v.4 up to
53v.1, we now return to the place where we left off within the text correctly labelled as Chapter 13.]
[54v.5] /de dag87 'khor dang bcas pa spyan drang ba ni / / 88

42 de lta bu 'i: Gt de'i (the root texts support the Dunhuang manuscript here) | 43 rin cen: Gt rin chen | 44 ni: Gt su | 45 gnyer na: Gt
nye bar (the root texts support the Dunhuang manuscript here) | 46 grub pa: Gt 'grub pa | 47 the tsom myed/: Gt the tshom med
48 zhes gsungs te: Gt ces gsungs pa yin no | 49 che 'am: Gt che'i | 50 gang la bab pa'i: Gt gsang ba'i (the Dunhuang manuscript's
reading would seem more appropriate here) | 51 na 'grub bo: Gt na/ 'grub po | 52 Gt inserts: / | 53 Gt inserts: / | 54 ba: Gt pa | 55 /: Gt
omits | 56 myed: Gt med | 57 kyis: Gt kyi | 58 kyi: Gt omits | 59 rig pa'i inserted as a correction, below the line, the position marked
by a cross above the line. | 60 bor: Gt bor ro | 61 dang: Gt omits | 62 dang: Gt omits | 63 du: Gt tu | 64 myi: Gt mi | 65 grub: Gt /
bsgrub | 66 /: Gt omits | 67 ma: Gt ma'i | 68 myi dge 'I: Gt mi dge'i | 69 bsten: Gt bstan (the root texts are divided here) | 70 myi bya
'o: Gt mi bya ste | 71 klung: Gt klong (either reading is possible: stream or expanse of awareness. Elsewhere in this manuscript,
there are instances where the Dunhuang text gives the spelling, klung, where klong is intended. This may be the case here: it is an
archaic spelling for klong. However, in this case, the root texts support the reading, klung. Thus, either an archaic spelling has
persisted into all branches apart from the Tenjur, or the reading klung is really intended here.) | 72 ste: Gt nas | 73 la yang: G la'ang
74 myed: Gt med | 75 par spyod: Gt pa spyod do | 76 'o: Gt bo | 77 nyan thos yin no: Gt nyan thos pa yin | 78 myi: Gt mi | 79 Gt
inserts: / | 80 pa: Gt ba | 81 po ni: Gt po'i | 82 du: Gt tu | 83 dang: Gt omits | 84 myed: Gt med | 85 bas: Gt pas | 86 myed: Gt med
87 dag: Gt nas | 88 Gt inserts: (also Qt116b.8 and Nt205.6): de yi pho nya bka' nyan ni/ rdo rje 'phra men gtum bag can (This
addition does not seem altogether appropriate here, but it could fit. These two tshig rkang occur in TZ Chapter 12, given in the
Dunhuang manuscript on folio 53v.1, immediately following the passage displaced from Chapter 13 into Chapter 12. It is not
entirely clear whether they are an accidental scribal repetition in Gt, or whether they are intended to be repeated in the
commentary here. Perhaps their occurrence at this point might be related to the Dunhuang manuscript displacement, or rather, its
return to the Chapter 12 material. It is most likely that a displaced folio explains how the Dunhuang manuscript jumped to the text
of Chapter 13 while within Chapter 12. However, the repetition of this line at precisely this point in Chapter 13 might explain
how the Dunhuang scibe [or previous copyist] returned back to the Chapter 12 material following this line.)

304

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

[6] /rdo rje dbyugs pas89 blangs nas su /

/g.yab mo lngas kyang g.yab par bya /
/ de nas dkyil 'khor zhugs90 par brtag /
/mchod pa chen [55r.1] $/ /por91 dbul bar bya /92 zhes gsungs te /
/de la bdag gi93 lus [la]94 bsdu95 na /
/96gtso bo 97 he ru ni98 lus la khyab par bsdu 'o/ /
[2] /hung ka ra ni99 glad100 pa'i dkyil du101 bsdu 'o/
/rnam par rgyal ba ni 'phral102 bar bsdu 'o/
/ha ya gri ba103 ni kha nas mthong ka 'i104 bar du bsdu 'o/ /
[3] khams gsum rnam par rgyal ba ni **** lag pa g.yas pa la bsdu 'o/ ****
/be con105 sngon po ni lag pa106 g.yon pa [4] la107 bsdu 'o/
/bdud rtsi 'khyil pa ni108 ***** snying kar109 bsdu 'o/
/ya man ta ka110 ni ***** lte bar bsdu 'o/
/gzhan gis [5] myi thub pa ni111 rkang pa g.yas pa la bsdu 'o/
/myi112 g.yo mgon po ni rkang pa g.yon par113 bsdu 'o/
/stobs po che ni114 gsang bar115 [Gt289] bsdu 'o/
[6] /de ltar 'dus pa las116/
dbus kyi gtso bo he ru ka 'am /
gang bsgom pa'i lha 'i skur117 bsgom mo/
/pho nya thams cad ni118 ba spu 'i bu ga thas [55v.1] cad du119 bsdu 'o/
/srin mo rnams kyang tshul de ltar bsdu 'o/
/gzhan du na sems 120 las byung ba'i chos gnas par gyur121 pa la bya ste/ /
[2] ye shes kyi byin brlabs122 las 'byung123 ba ni/
chos kyi dbyings nyid 124 la/
dbyings kyi ye shes kyis mnyam pa nyid du sbyor zhing /
/125myi dmyigs126 par [3] snyoms par 'jug pa'i tshe /
/sems ****** dpa' gnas gyur pas ni /
reg pa dang /127 *****
yid la byed pa dang128 ['(/b)]dun129 pa dang / /
[4] mos pa dang /
dran pa dang /
ting nge 'dzin dang / *****
'byed pa dang /
dad pa dang /
gnyid gnas ****** gyur pa'i bar du slong zhing bskul130 bar [5] byed pa ste/
chos kyi dbyings dang ye shes las byung nas /
slar dbyings su 'dzud pas /
/131sems pa 'gyur pa ni132 'phra men ma133 skul byed ma134 zhes bya ste/
89 dbyugs

pas: Gt dpyugs pa | 90 zhugs: Gt bzhugs (the Tshal pa Kanjur and Bhutanese texts support the Tenjur reading here, but
the other root texts agree with the Dunhuang manuscript) | 91 por: Gt po | 92 bar bya /: Gt bas | 93 gi: Gt gis | 94 la: it seems that an
earlier lnga was corrected (there would seem to be a very faint nga beneath the la which has been rubbed out; la would make better
sense here); Gt la | 95 bsdu: Gt sdu | 96 / /: Gt omits | 97 Gt inserts: shr | 98 he ru ni: Gt he ru ka | 99 hung ka ra ni: Gt h ka ra
100 glad: Gt klad | 101 dkyil du: Gt rgyas su | 102 'phral: Gt pral | 103 ha ya gri ba: Gt ya ha gr ba (but Qt gives ha ya gr ba)
104 mthong ka 'i: Gt thong ka'i | 105 be con: Gt be chon | 106 lag pa: Gt omits | 107 la: Gt omits | 108 ni: Gt omits | 109 snying kar: Gt
snying gar | 110 ya man ta ka: Gt ya manta ka | 111 gzhan gis myi thub pa ni: Gt gzhan gyis mi thub pa | 112 myi: Gt mi | 113 par: Gt
pa la | 114 ni: Gt omits | 115 bar: Gt ba la | 116 'dus pa las: Gt bsdus la | 117 lha 'i skur: Gt lhar | 118 ni: Gt omits | 119 du: Gt la | 120 Gt
inserts: dang sems | 121 gnas par gyur: Gt gnas gyur | 122 byin brlabs: Gt byin rlabs | 123 'byung: Gt byung | 124 Gt inserts: de | 125 /
/: Gt omits | 126 myi dmyigs: Gt mi dmigs | 127 /: Gt omits | 128 Gt inserts: / | 129 ['(/b)]dun: prefixed 'a is probable but the letter is
joined beneath in faint ink. It seems most likely that a ba prefix was first written, but amended to 'a with thick ink. | 130 bskul: Gt
skul | 131 / /: Gt omits | 132 sems pa 'gyur pa ni: Gt sems dpa' gnas 'gyur ba ni/ | 133 'phra men ma: Gt 'phra men | 134 skul byed
ma: Gt bskul ma

Chapter 13

305

[6] gzhan reg 135 la stsogs pa snyoms par 'gyur ba'i136 gnas /
/gzhan gis137 bsad pa'i god138 la 'bring139 ba'i phyir /
/bya rgod 140 mgo can zhes bya 'o /
[56r.1] $/ /rI bo brtsegs141 pa las ni /142
mgyogs pa'i phyir bya143 mgo can zhes144 smos so145/
/'dun par gyur pa ni /146
dbyings dang ye shes mnya [2] pa nyid du 'dod pas /
/'phra men ma147 za [ba] mo148 zhes bya ste149 /
/bya rog 150 mgo can no/ [Gt290]
/ yid la byed par gyur151 pa ni152/
dbyings dang ye shes mnya [3] ba nyid du 'dod153 pas/
'phra men ma 'drangs154 ***** ma zhes bya 155 ste /
/ chos kyi dbyings la156 ***** nyal zhing /157
ye shes kyi phyir 158 'brang [4] bas /
mtshan mo langs la nyin mo ****** nyal bas159 'ug pa'i mgo can no / *****
/mos par gyur pa ni dbyings dang ye [5] shes mnyam bar gyur pa 'dzin pas /
/'phra men ma 'dzin ma160 zhes161 bya ste /
khra162 mgo can no /
/163 dran par gyur pa ni /164
mnyam par gyur [6] pa nyid ma brjed pas165 'phra men ma bsgrims166 ma zhes bya ste pha bang167 mgo can
no/
/'byed par gyur pa ni /168
mnyam 169 nyid kyi nang du/170
rdul tsa
[56v.1] yang log par rtog pa myi171 gnas te /
dper 172 mar mkhol ma'i173 nang du chu myi174 gnas pa dang 'dra bar /
sel cing skems pa'i phyir /
/'phra men ma [2] gsod ma175 zhes bya 'o176 ste/
sre mo 'i177 mgo can no /
/ting nge 'dzin178 gyur pa ni/
mnyam pa nyid las gzhan du myi179 g.yo bas /
'phra men ma sbed [3] ma180 zhes bya ste /
khwa181 mgo can 182/ *****
/dad par183 gyur pa ni mnyam 184 nyid du 'gro bas ******
/'phra men ma mgyogs ma zhes bya [4] ste/
phub shud185 mgo can no/ ******

135 Gt

inserts: pa | 136 'gyur ba'i: Gt gyur pa'i | 137 gis: Gt gyis | 138 god: Gt gong | 139 'bring: Gt 'brang | 140 Gt inserts: kyi
Gt rtsegs | 142 /: Gt omits | 143 bya: Gt / bya rgod (surely, an omission in the Dunhuang manuscript) | 144 zhes: Gt ces
145 so: Gt pa'i phyir ro | 146 /: Gt omits | 147 ma: Gt omits | 148 za [ba] mo: there is a clear na ro on the ba, but also an ink line
probably intended as a deletion; za ba mo occurs below, f.56v.2; Gt za byed | 149 zhes bya ste: Gt ces bya'o | 150 Gt inserts: gi
151 par gyur: Gt pa gnas gyur | 152 ni: Gt omits | 153 'dod: Gt shes | 154 ma 'drangs: Gt nam 'grangs | 155 Gt inserts: ba | 156 la: Gt ni
157 /: Gt omits | 158 Gt inserts: mi | 159 Gt inserts: / | 160 ma: Gt omits | 161 zhes: Gt ces | 162 khra: Gt khra'i | 163 / /: Gt omits | 164 /:
Gt omits | 165 Gt inserts: / | 166 bsgrims: Gt bsgrin | 167 pha bang: Gt / pha wang gi | 168 /: Gt omits | 169 Gt inserts: pa | 170 /: Gt
omits | 171 myi: Gt mi | 172 Gt inserts: na | 173 mkhol ma'i: Gt khol ba'i | 174 myi: Gt mi | 175 gsod ma: Gt bsod ma | 176 'o: deletion
appears indicated by a tiny line above; Gt omits | 177 sre mo 'i: Gt sre mong gi | 178 ting nge 'dzin: Gt ting 'dzin gnas | 179 myi: Gt
mi | 180 sbed ma: Gt sbad ma (Gt's reading is plausible; bTsan lha nNag dbang tshul khrims 1997: 608, gives for sbad pa, gtong
ba'am skul ba'i don. However, given the gloss here, it would seem that the Dunhuang manuscript reading which could mean,
Concealed/Veiled/Concealing, might possibly be more apt. It may also be that sbad here is an alternative spelling for sbed. Note
that in the second and third verse of TZ's Chapter 18, there is agreement across most of the versions, on the verb for the phra men
mas' activity as, sbad byas te, where either meaning could fit well.) | 181 khwa: Gt khwa'i | 182 Gt inserts: no | 183 par: Gt pa | 184 Gt
inserts: ba | 185 phub shud: Gt pu shud kyi
141 brtsegs:

306

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/gnyid gyur pa ni /
mnyam pa nyid gsal ***** ba ste186/
mnyam pa dang myi 'thun ba'i187 [5] rnams myos par byed pas /
'phra men ma myos byed ces bya 'o188 ste /
/smrIgs bu 'i189 mgo can no/
/de rnams kyi las kyi gnas190 skabs [6] kyi bye brag la la191 ni [Gt291] 'phra men ma myo 'bros ma zhes bya
ste /
byI ba'i192 mgo can no /
ci 'i phyir 'phra men ma zhes193 bya zhe na/ /
[57r.1] $/ /dbyings dang ye shes kyi 194 byin brlabs pas195 /
/dbyings kyi byin rlabs196 ni de kun la byams /
/ ye shes kyi byin brlabs197 kyis ni dngos [2] po kun za ba'i phyir ro /
/de rnams 198 lus su grub pa ni199 dkyus nas smos pa nyid do /
/dbyings dang ye shes ni lus dang yan lag go / /
[3] de las 'phro zhing 'byung ba ni ba spu 'i200 ****** bu ga nas 'byung ba201 ste 'phra men no / ******
/gzhan yang 'phra men ma202/
khab203 [4] spu can dang /
ral gyi204 spu can dang205/ ******
/lcags kyi pags pa206 can dang / *****
zangs kyi khab207 spu can dang /
zangs [5] kyi pags pa can la stsogs208 pa/
bye ba 'bum phrag 209 brgya rtsa brgyad cig 'byung ba yang /210
'di rnams las 'phros so zhes /211
gcig las [6] 'phros pa'i tan tra las 'byung /
/se gol gtogs pa tsam gyis kyang /
/mgyogs ma212 phyogs bcur213 gar yang phyin / /
[57v.1] de la zhon te214 'gro byed pa /
/ bya rog mgo can la stsogs215 yin 216/
zhes gsungs te /
gang la bya ba'i skabs kyis /217
mgyogs [2] ma dang za ba mo la stsogs218 pa zhon te 'gro na /219
'grub po220 pa 'i don221/
gzhan du na mgyogs ma nyid chos kyi dbyings kyi222 byin brlabs pas223/
[3] dbyings kyi224 kun la khyab pas /
'gro ***** bar myed225 par [Gt292] phyin pa 'o /
/zhes226 pad mo227 ***** dpal 'phreng dkar po las bshad do228/
[4] sbed byed ma 'is229 sbas pa na/ / *****
mthu chen lha srin gang gis kyang /*****
186 ba

ste: Gt te | 187 myi 'thun ba'i: Gt mi mthun pa'i | 188 ces bya 'o: deletion of 'o appears indicated by a tiny line above; Gt ma
zhes bya | 189 smrIgs bu'i: Gt spyang ki'i (in the context of the other emanations listed here, the Dunhuang manuscript's reading
would seem more likely) | 190 gnas: Gt omits | 191 la la: Gt la | 192 byI ba'i: Gt bya'i (as noted on the list at f.54v.4 above, the
Dunhuang manuscript's reading seems more likely and Gt's reading would seem too non-specific) | 193 zhes: Gt omits | 194 Gt
inserts: dbyings kyi | 195 byin brlabs pas: Gt byin gyis brlabs pa ste | 196 byin rlabs: Gt byin gyi rlabs kyis | 197 byin brlabs: Gt byin
gyi rlabs | 198 Gt inserts: kyi | 199 ni: Gt omits | 200 ba spu 'i: Gt / pa spu'i | 201 ba: Gt omits | 202 'phra men ma: Gt 'phra man ma
203 khab: Gt kha | 204 ral gyi: Gt ral gri'i | 205 dang: Gt no | 206 pags pa: Gt spu | 207 khab: Gt kha | 208 stsogs: Gt sogs | 209 Gt
inserts: dang | 210 /: Gt omits | 211 zhes /: Gt / zhes | 212 ma: Gt mas | 213 bcur: Gt bcu | 214 te: Gt nas | 215 stsogs: Gt sogs | 216 Gt
inserts: pa kun (the Tshal pa Kanjur and Bhutanese texts give la sogs kun for la stsogs yin, but the other root texts support the
Dunhuang manuscript reading, giving la sogs yin) | 217 /: Gt omits | 218 stsogs: Gt sogs | 219 /: Gt omits | 220 po: Gt omits | 221 Gt
inserts: to | 222 kyi: Gt kyis | 223 byin brlabs pas: Gt byin gyis brlabs las | 224 kyi: Gt kyis | 225 'gro bar myed: Gt 'gro ba med | 226 /
/zhes: Gt zhes / | 227 pad mo: deletion seems to be indicated by a line through pad, and smudging of mo; Gt omits | 228 do: Gt de
229 ma 'is: Gt mas ni

Chapter 13

307

/de la mthong bar myi230 nus na / /

[5] gzhan gis231 mthong bar ga la nus/
/gzhan du na sbed ma ni232 chos kyi dbyings myi dmyigs pa'i233 byin brlabs234 yin bas /
dmyigs su [6] myed235 de / ii
/des sbas pa ni rang bzhin gis myi mthong ngo zhes /
rtse gcig bsdus pa las 'byung ngo /
/skul byed las ni gzhen 'des236 [58r.1] $/ /yin /
mgyogs pa'i las ni phyin par 'gro /
/ gsod ma'i las ni bsgrubs pa yin /
/so so ma bsgrubs nyid 'gro [2] 'o/
/zhes gsungs te /
rnal 'byor pas las gang bsgo ba ma bsgrubs na /
gang gis ma bsgrubs pa 'am /
'phra men nyid dam [3] gyis mgo 'gas she 'o / / ****
gzhan du na las ma bsgrubs dang /
/chos ***** kyi dbyings dang /
dbyings kyi ye shes [4] kyis mnyam bar sbyor ba las 'gal ***** bas /
/de dang 'gal na log par rtog ***** pa'i dbang gis 'khor bar 'gyur bas /
[5] dam gis mgo 'gas pa yin zhe [na?]'o237/
/thabs kyi zhags pa pad ma 'phreng las /
/sri/238 rdo rje srin mo 'i le'u ste bcu gsum pa [6] 'o // :

ii QtGtNt lose the text here, omitting the following chapters, re-joining the Dunhuang manuscript version near the end of Chapter
17, corresponding to the Dunhuang manuscript f.61v.1, but also including some stray text in their final lines of that chapter. The
Tenjur chapter numberings have not been adjusted, so the Chapter 17 title is correctly identified in the Tenjur texts.
230 myi:

Gt mi | 231 gis: Gt gyis | 232 sbed ma ni: Gt sbed byed ma ni/ | 233 myi dmyigs pa'i: Gt mi dmigs pa'i | 234 byin brlabs: Gt
byin rlabs | 235 dmyigs su myed: Gt dmigs su med | 236 gzhen 'des: presumably, this is for gzhen 'debs, which is the reading given
in the Tawang Kanjur. The South Central, Hemis Kanjur, Tshal pa Kanjur and Bhutanese texts, give gzhan 'debs. | 237 [na?]'o:
smudged letter, most probably na, deletion probably intended. | 238 only the Dunhuang manuscript includes sri/.

Chapter 14
[Ms58r.6]
//da nI rig sngags kyi le'u bshad par bya ste /
/ ba dzra kro dha /
sa ma ya h /
/ ba dzra kro dha /
[58v.1] sa ma ya phat /
ba dzra kro ta sa ma ya stwam /
/me tog dor ba'i sngags so /
/I he hyi /
bha ga bhan /
dza h ba ho /
ru lu ru lu ru lu [2] h /
spyan drang ba'i sngags so /
i / ba dzra shrI he ru ka h /
[beneath the line] gtso bo 'i sngags

h phat /
/ ba dzra kro dha h ka ra h /

[beneath the line] h ka ra 'o

ga rdza ga rdza ga rdza h phat /

[3] ba dzra kro dha bi dza ya h /
[beneath the line] rnam par rgyal ba'i

ha na ha na ha ***** na h phat /
/ ba dzra kro dha ni la tan ta1 ***** da ha da ha da ha h phat /
[beneath the line] dbyIg po sngon po

/ ba dzra [4] kro dha ya man ta h /

[beneath line 4] gshin rje 'joms pa

pa tsa pa tsa pa tsa ***** h phat /

/ ba dzra kro dha /
a rya ***** a tsa la h /
[beneath the line] myi g.yo ba

bhan dha bhan dha bhan dha [5] h phat /

/ ba dzra kro dha [pa?]ha2 ya gri ba h /
[beneath the line] dpal rta mgrin

hu lu hu lu hu lu h phat /
/ ba dzra kro dha /
/a pa tsi ta h /
[beneath the line] gzhan gyis myi thub pa

/ti shha I shtha I shtha h pha /

[6] ba dzra kro ta /
a mri ta kun da li h /
[beneath the line] bdud rtsi 'khyil pa 'o

tshIn dha tshIn dha tshIn dha h phat /

i See
1 the

our edition of TZ p.176-178, for notes on the names of the deities in these mantras.

root texts complete this mantra with, h/, which is undoubtedly omitted in error here. | 2 [pa?]ha: prefixed pa is in error here
(and not shared by the root texts).

Chapter 14

309

/ ba dzra kro dha tre log gya bi dza ya h /

[beneath the line] khams gsum rnam par rgyal ba'i

bhIn dha bhIn dha bhIn dha [59r.1] $/ / h phat /

ba dzra kro dha ma ha ba la h /
[beneath the line] stobs po che

spo ta spo ta spo ta h phat /

rtsa ba'i sngags dang /
las kyi sngags ril 'dus pa 'o /
[2] bskul ba ni /
sum bha /
pu stang /
shan ting /
swas ting /
las bzhi bskul ba 'o /
h mying phat mying /
h mying h /
h mying sw h /
[3] swa mying ha /
gzing pa'i sngags / *****
a nrI trI prI dhu /
lam rgyud lnga brtag pa ***** yin no /
/thabs kyi zhags pa [4] pad ma 'phreng las khro bo bcu dang ***** las kyi sngags ril 'dus pa'i le'u
***** ste bcu bzhI pa 'o // :

Chapter 15
[Ms59r.4]
// [5] da nI srin mo rnams kyi sngags ste /
'phra men ma rnams kyI rig pa bshad do /
ba dzra kro dha h /
ha la ha la ha la ha la h /
ma 'i /
[6] ba dzra1 h /
ru lu ru lu ru lu ru lu ru lu h /
ma ma ma ma ma2 /
bhyo bhyo bhyo bhyo bhyo h /
mngag pa /
'dz 'dz 'dz /
slar bsdu ba /
[59v.1] h h h /
yud tsam gyis las zIn bar bsgom /
/ thabs kyi zhags pa pad ma 'phreng las /
/rdo rje srin mo 'i sngags [2] kyi le'u ste bco lnga pa 'o // :

1 The

root texts (apart from the Bhutanese texts, which omit this mantra) insert kro dha or kro dhe here, and this has surely been
omitted in error in the Dunhuang manuscript. | 2 The root texts (apart from the Hemis Kanjur) insert h here.

Chapter 16
[Ms59v.2]
//da nI phyag rgya 'i le'u bshad par bya ste / i
/mnyam nyid du rtags1 byas na / /
[3] bsgul zhing bsgigs pa2 thams cad kyang / **
/phyag rgya yin zhes rgyal bas bshad /
/ **** mdo tsam du ni 'di ltar brtag / /
[4] zhes gsungs te /
dag pa rnam gsum ii ***** kyI tshul du mnyam pa nyid du rig na / *****
/lus bsgul zhing bsgigs pa [5] thams cad phyag rgya yin zhes /
rgyal po 'i rtog pa las bshad do /
[beneath the line] rnam par snang mdzad kyis bshad do

/lag pa thal sbyar gyen du bsgreng /

[beneath the line] he ru ka 'i phyag rgya

/rkang pa mthil sbyar thur du [6] gzugs /

/tshigs cen bzhi yang gcag par bya /
/ mgo dang mtshan yang snying po brtag /
/rdo rje rtse gcig rtse dgu dang / /
[beneath the line, and continuing under the top line of the next folio] rtse gcig ni dpal he ru ka 'i skur smon pa la bya ste rtse dgu ni/
+

[60r.1] $/ /bcu gcig du yang rnam brtag ste /

[beneath the line, and continuing onto line 2, as indicated by the crosses] + dpal he ru kar gnas pa las lus la khro bo brgyad bkod pa
la dgu 'o/ bcu gcig ni bdag gtso bor he ru kar bsgoms pa 'am yang na lus khro bo 'i bcu la bkod pa la bya 'o rnam gcig du phyag rgya
bcings par bya +

/'dI ni khrag 'thung chen po yi /

/phyag rgya yin bar dpa' bos bshad /
/ lag pa thal [2] sbyar gyen du bsgreng /
[beneath the line] + zhes sngar bshad do3

/rkang pa phral nas brkyang bskum zhing /

[beneath the line, beginning under brkyang] hu ka ra 'I phyag rgya ste 'di man cad [sa?]lha4 so sor byas na yang bzang ngo/ o
badzra g[h?]ro dha hu ka ra hu gar rdza gar rdza hu pha

/goms pa'i stobs5 kyis gar byed pa /

/'dI ni h gyI rgya cen yin /

Note that Chapter 10 is also described as on mudrs, but there is little overlap. This chapter focuses entirely on the bodily
postures for embodying the various wrathful deities, whereas Chapter 10 lists the hand gestures for the peaceful buddha family
deities.
ii dag pa rnam gsum: the three aspects of purity. These are referred to in the interlineal notes of Chapter 1 (above the line of
f.1v.1). There the annotator refers to them in the context of the methods of the three bringing enlightened body, speech and mind.
1 rtags:

the root texts all give rtogs. | 2 bsgigs pa may possibly derive from the verb, sgrig pa, and could make sense here (implying
re-arrangements), but it is more likely that our reading of bsgigs pa is in fact a corruption of the root editions' reading of
gsigs/bsigs (the Tshal pa Kanjur and Bhutanese texts give gsigs; the South Central NGB and local Kanjurs give bsigs), swinging,
which would seem most likely the archetypal and the preferable reading here. | 3 Despite the plus sign which generally indicates
continuity, this comment does not continue the thread of the comment on the immediately preceding line. It appears to refer
simply to the words given here, pointing out that they repeat a line given above, and thus mean the same as for their former
occurrence. | 4 a sa is written clearly as a prefix, but presumably in error. | 5 goms pa'i stobs: the root texts all give the more
appropriate, gom pa'i stabs.

312

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

[3] rkang phral mthil ni thur du gzugs / ****

[beneath the line] stobs po che 'i

/lag pa phral nas gar byed cing / *****

[beneath the line] o badzra grod dha ma ha pa la hu spo ta spo [dha?] hu pha/

/ stabs su mchongs pa byed pa ni / /

[4] 'dI ni stobs cen phyag rgya yin / *****
/rkang pa g.yon gyIs g.yas pa mkhyud ******
/lag pa gar thabs b[skor?]6 nas [5] kyang /
/ thal mo drag du brdab pa ni /
[beneath the line] badzra g[h?]ro dha bhI dz ya hu ha na ha na hu pha

/ rnam par rgyal ba'i phyag rgya yin /

/ lag pa g.yas7 kyI the bo yis /
/mthe'u cung sen mo [6] 'I dkyil du gnan /
/gdang8 pa'i thabs su bsgreng ba ni /
[beneath the line] o badzra ghrod dha ni la tan dha dha ha dha ha hu pha/

/ dbyug sngon can gyi phyag rgya yin /

/ lag pa g.yas kyI khu tshur gyis/
[60v.1] brdeg pa'i thabs su gdangs9 pa dang /
/g.yon pa dku la brtan pa ni /
/gshin rje 'joms pa'i phyag rgya yin /
[beneath the line] badzra g[h?]rod dha ya man dha ka hu pa tsa pa tsa pa tsa hu pha/

/ lag pa g.yas [2] kyI thal mo 'is /

/gdab pa'i thabs su gdangs10 pa dang /
/g.yon gyis 'dzIn pa ltar byas pa /
/myI g.yo sgrol11 gyI phyag rgya [3] yIn /
[beneath the line, beginning under the previous tshig rkang, but apparently connected to this line] badzra ghrod dha a rya a tsa la
hu bhan ***12 dha bhan dha bhan dha hu pha

/lag pa thal mo sbyar nas nI / ****

/gung mo rgyab du bsnol bar bya / ******
/mthe bo gnyis kyang phral te [bsgreng?]13 /
[4] spyI bor bzhag na grI ba'i rgya /
[beneath the line] badzra ghrod dha ha ya 'ghrI ba hu hu lu hu lu hu lu hu pha/

/rkang ****** pa g.yas pa thur du gzugs / *****

/lag pa g.yas pa gyen du bsgreng /
[5] g.yon kyi rka14 lag phyar ba ni /
gzhan gis myi thub phyag rgya yin /
[beneath the line] o badzra ghrod dha a pa ra dzI dha hu dhi shtha dhi shtha dhI shtha hu pha/

/lag pa g.yon pa15 dkur brtan te /

/rkang pa gshibs te [6] mchongs pa ni /
/bdud rtsi 'khyil pa'i phyag rgya yin zhes16 /
6 b[skor?]:

the ink is partially erased from the page, but it appears to read, bskor, the reading shared by the root text editions (see
TZ edition p.183). | 7 g.yas: this reading is supported by the South Central NGB and local Kanjurs, but the Tshal pa Kanjur give
g.yas g.yon, and the Bhutanese texts give g.yon. | 8 gdang: this reading is supported by the Tshal pa Kanjur texts, which give
gdangs, but the South Central NGB and local Kanjurs give gdengs, which seems rather more appropriate (the Bhutanese texts give
brdabs, but then give gdengs instead of the following bsgreng). | 9 gdangs: here, the root texts agree on gdengs. | 10 gdangs: here,
most of the root texts agree on gdengs (although the South Central NGB is divided between rdangs and rdengs). | 11 sgrol:
undoubtedly, this is a scribal error in the Dunhuang manuscript; the root texts give the more appropriate, mgon. | 12 an attached ra
(on the syllable, sgrol) in the line above prevents writing here. | 13 [bsgreng?]: the ink is partially erased from the page, but it
appears to read, bsgreng, the reading shared by the root texts (see TZ edition). | 14 rka: undoubtedly, a scribal error in the
Dunhuang manuscript for the root texts' reading, rkang. | 15 g.yon pa: the root texts give g.yas g.yon. | 16 zhes: inserted below line,
as a correction

Chapter 16

313

[beneath the line] badzra ghrod dha mri dha kun 'dha li hu tshin 'dha tshin dha tshIn dha hu pha/

/dpa' bo chen pos yang du17 bshad /

/lag pa g.yas g.yos bsnol ba dang /
[61r.1] $/ /rkang pa brgyang bskum byas pa ni /
/ khams gsum rgyal ba'I phyag rgya zhes /
[beneath the line, beginning under ni of the previous tshig rkang] badzra ghrod dha tre log gya bhI rtsa ya hu bhin dha bhin
dha bhin dha hu pha/

/18 dpa' bo chen pos yang du19 gsungs /

[beneath the line] rnam par snang mdzad kyis bshad

/ lag pa [2] g.yas g.yos sprugs pa dang /

[beneath line 1] 'phra men ma [k('d)]un

/ lus kyang drag du bsigs pa ni /

/ rdo rje 'phra men ma rnams kyi /
/ phyag rgya yin zhes dpa' bos [3] gsungs /
/ phyag rgya 'di dag bcings ****** nas ni /
[beneath the line] rnam par snang mdzad 'am rdo rje sems dpa' 'khor ba'i sdug bsngal las

/ stabs cen drag du mchongs byas ***** na /

khams gsum dag las mchongs [4] par 'gyur /
/ gzhan yang go 'phang ***** thams cad kun /
[beneath the line] yon tan phun sum tshogs pa'i [beneath the line, under kun] ma lus pa

/ 'grub 'gyur 'di ***** la the tsom myed /

[beneath the line, under the tsom myed] nges par 'grub

/ rig pa'i skyes [5] bu chen pos bshad do /

[beneath line 5] rnam par snang mdzad kyis bshad

/ 'phags pa thabs kyi zhags pa pad ma 'phreng las /

phyag rgya 'i le'u ste bcu drug pa 'o // :

17 yang
19 yang

du: the South Central NGB and local Kanjurs give the more appropriate, yang dag. | 18 the root texts insert: mgon po.
du: the South Central NGB, local Kanjurs and Bhutanese texts give yang dag.

Chapter 17
[Ms61r.5]
// [6] da ni las kyi le'u bshad par bya ste /
/ 'khor lo chen po rtsibs brgyad par /
/ dkyil du khrag thung spyan drangs te /
/ mdun du hung yang [61v.1] spyan drang ngo /1
/ rgyab du stobs cen spyan drangs nas /
/ 'khor lo rtsibs kyi steng daug2 du /
/bcu po rnams kyang rim bzhin no /
[beneath the line] khro bo gzhan dag bsnams

[Gt292.2] i /so so'i [2] gnas su spyan drangs te /

[beneath the line] phyogs mtshams su

/ dmar gyI3 a rga4 gsol nas ni5/

/mchod yon gzhan yang dbul bar bya /
[beneath the line] 6sha chen bzhi las stsogs te sman sna lnga drung du bzhag ste

/ rdo rje 'I7 dbyings8 kyis bstod [3] byas te /

/ de nas las rnams bsgrub ***** par bya / / ****
[The Dunhuang manuscript appears to have an accidental omission here of a further root text verse before the
chapter ending (this verse is also omitted in the Bhutanese version, but apparently by design, since it omits
the parallel verses in Chapters 24, 29, and 34). We give here the version in QtGtNt, which also supplies the
text given as an interlinear note beneath line 2 in the Dunhuang manuscript (see above). It is possible that
the Tenjur editors included a marginal note in error, but since the line could fit perfectly well into the main
text, it is also possible that the Dunhuang manuscript mistook a line of the main text for an annotation.
Qt118a.7; Gt292.3; Nt208.4
zhes gsungs te/
sha chen dang bzhi la sogs pa sman sna lnga dang sbyar te/
zhe sdang rdo rje 'khor bcas pas/
/'jig rten khams ni rab 'byams kun9/
/khros pas yud las10 rlag mdzad de/
/rlag par ma gyur gang yang med/ ces bstod do/]
[61v.3] /11 thabs kyi zhags ***** pa pad ma12 'phreng las /
las kyi le'u [4] ste bcu13 bdun pa 'o// :

i At
1

this point, the Tenjur version's conflated Chapter 13 and 17 rejoins the Dunhuang manuscript.

There appears to be highlighting here (and as indicated below) but it is not dark, and it is possible it may simply be
discolouration of the paper. Alternatively, it may have partially rubbed out, and perhaps more of these lines recognised as root
text in our TZ edition may once have been highlighted also? But if so, the evidence no longer remains on the other words.
2 daug: an original dug seems to have been corrected to dag by a small line crossing through the zhabs kyu. | 3 gyI: Gt chen (none
of the root texts support Gt here) | 4 a rga: Gt arga | 5 ni: Gt kyang | 6 Gt gives this interlinear note within its main text, after its
version of the following two lines (see below) | 7 rdo rje 'I: Gt rdo rje | 8 dbyings: Gt dbyangs (the root texts are divided here, but
dbyangs was probably in the archetype and also seems a more appropriate reading) | 9 rab 'byams kun: ma lus kun, in the South
Central NGB and local Kanjurs | 10 las: la, in the root texts | 11 Gt inserts: 'phags pa | 12 pad ma: Gt padmo | 13 Gt inserts: /

Chapter 18
[Ms61v.4] [Gt292.5]
//da ni drag po 'i las 1 bshad par bya ste / *****
/rdo rje 'phra men sbyor ba yis2/ /
[5] yud tsam gyis nI de3 bkug nas/
/bltod4 pa'i tshul du rab bstabs na5/
/lha yang rung ste brlag par bya /6 'gyur /
zhes gsungs te / /
[6] 'phra men ma stag 7 mgo can la stsogs8 pa i mgyogs pa9 rnams mngag ste10/
/gang la bya ba yud tsam gyis11 snying la bzung nas/
khro bo [62r.1] $/ //dang khro mo 'i zhal du [Gt293] bstabs te12/
[beneath the line] [chu?] klong gsol bar bsgoms [na?] bskabs par 'gyur ro [rlung?]13 la stsogs pa yang chad par bsgyur ro zhes/

ltod14 pa'i tshul du gsol bar bsgoms na15/

/lam rgyud16 lnga 'i17 sems can gang yang [2] rung ste18 mod la 'chi bar 'gyur ro/
/gzhan du na chos kyi dbyings dang/
stong pa nyid kyi ye shes lnga 'i byin rlabs19/
rnam par grol ba'i [3] skye mched bcu gcig gyis20/
lam rgyud **** lnga 'i sems can dngos por lta ba ni/
/**** stong pa nyid kyi21 dam tshig las 'gal [4] bas/
/de dag rang bzhin gis22 za ***** bar 'gyur zhes tan tra kar ma ma le las ***** bshad do/
/'ug pa'i mgo can23 [5] la stsogs24 pa25 /
/'phra men ma26 rnams sbad byas te27/
/myig28 dang rna ba sna dang lce/
/de sems 'phrog29 bsgoms nyams par 'gyur/30
[6] zhes gsungs te/
khyung dang khwa la stsogs pa31 khra thabs su32 btang ste/
/gang la bya ba la33 bya ba'i dbang po 'byin par bsgoms na//
[62v.1] 34 dbang po nyams nas myig35 long ba dang/
i Note the lists of the male and female emanations above, beginning with the male Tiger-headed: Ch. 12, f.53v.2-3, and Ch. 13,
f.54v.3- 4.
1 Gt

inserts: rnams | 2 yis: Gt yi | 3 de: Gt 'dir | 4 bltod: QtGtNt stod (The local Kanjur and Tshal pa Kanjur version, brtod/rtod,
which was probably in the archetype, would suggest, "using the method of staking/tethering (them)". The Tenjur, Bhutanese and
South Central NGB versions - bstod - would mean: "while praising", which could also fit. However, bltod, given in the Dunhuang
manuscript, is a further alternative, which might seem rather more apt, and brtod may even have been an archaic spelling for this.
Dan Martin 2007: 122, notes that bltod nas = skrag nas, citing Katsumi Mimaki's work on dBus pa blo gsal. The commentarial
text below gives ltod (although again, bstod in the Tenjur), for which bTsan lha ngag dbang tshul khrims [1997: 262] gives rtsod
pa, which might suggest a confrontational or aggressive manner. See also Chapter 11 of the TZ edition above, p.154, and further
elaboration in TZComm's Chapter 11, p.291 above.) | 5 na: Gt nas | 6 bya /: deletion of bya appears to be indicated by a line through
the letter, and the following 'gyur is written in small letters, probably as an inserted correction; Gt omits | 7 Gt inserts: gi | 8 stsogs:
Gt sogs | 9 pa: Gt ma | 10 mngag ste: Gt mngags te | 11 Gt inserts: / | 12 bstabs te: Gt bstab ste | 13 rlung: this is rather uncertain, but
on comparison with other letters, seems the most likely word here. Other possibilities are: drung/hung/ru/rung. | 14 ltod: QtGt
bstod (see note 4 above) | 15 na: Gt nas | 16 rgyud: Gt brgyud | 17 lnga 'i: Gt drug gi (lnga would seem more appropriate here,
however, given the imagery of the five primordial wisdoms, and the repetition of lnga 'i below). | 18 Gt inserts: / | 19 byin rlabs: Gt
byin brlabs | 20 gyis: Gt gis | 21 kyi: Gt omits | 22 gis: Gt gyis | 23 'ug pa'i mgo can: the female emanation associated with the
southern direction, listed on f.54v.3. | 24 stsogs: Gt sogs | 25 pa: Gt pa'i | 26 'phra men ma: Gt 'phra men | 27 byas te: Gt de | 28 myig:
Gt mig | 29 sems 'phrog: Gt bas mchog (Gt is clearly in error here) | 30 /: Gt omits. The highlighting on this verse is faint and
uncertain where it occurs. Perhaps originally the whole verse was highlighted, or perhaps previous highlighting has disappeared.
31 la stsogs pa: QtGtNt la sogs bya rgod mgo can la sogs pa | 32 khra thabs su: Gt khrag 'thung (the Dunhuang manuscript's
reading seems more appropriate here) | 33 bya ba la: Gt omits | 34 Gt inserts: de'i | 35 myig: Gt de'i mig

316

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

rna ba 'on pa dang/

sna zhom ba dang/
lce lkugs pa dang/
sems myos te glen par 'gyur ro36/ /
[2] gzhan du na/37
stong pa nyid kyi ye shes drug gyI38 byin brlabs39 kyis/
/ril por lta ba'i dngos po rnams/40
yang dag pa nyid las 'gal ba'i [3] phyir/
/rang bzhin 41 nyams par42 byed **** ces tan tra kar ma ma les43 'byung ngo/ ****
/'phra men rnams ni sbad byas te44/
[4] de 'i bsgrub pa bar gcad45 cing/
/ ***** gsang sde46 'khrug47 par bsgoms byas na/ ****
bsgrub48 pa nyams shing 'bye [Gt294] bar49 'gyur/
[5] zhes gsung pa te50/
/'phra men ma rnams kyis/
gang la bya ba de 'i51 las thams cad ma52 'grub par bgegs53 byed pa 54/
/gsang sde thas [6] cad 'bye ba'i55 dkrugs ma byed par bsgoms56 na/
/don du gnyer ba myi57 'grub cing /58
gsang sde gcig la59 gcig shar gnyer te 'bye60 bar [63r.1] $/ /'gyur ro/
/gzhan du 61 stong pa nyid kyI ye shes kyi byin rlabs62 kyis /
dngos po la gtsor63 bsgrub par64 byed pa dang/
tshig gi sgra la [2] gtsor byas te65 phyogs gcig du66 byed pa ni/
stong67 pa nyid dang mnyam pa nyid68 las 'gal ba'I phyir/
/don du bsgrub pa myi69 'grub cing/
gsang sde [3] yang 'bye bar70 'gyur ro zhes ma les71 **** las 'byung ngo / ***
/72thabs kyi zhags *** pa pad ma73 'phreng las/
drag po 'i [4] las kyI le'u ste74 bco brgyad pa 'o //
:

36 ro: Gt omits | 37 /: Gt omits | 38 gyI: Gt gi | 39 brlabs: Gt rlabs | 40 /: Gt omits | 41 Gt inserts: gyis | 42 nyams par: Gt mnyam bar
(the Dunhuang manuscript's reading seems more appropriate here) | 43 kar ma ma les: Gt karma ma le las | 44 byas te: Gt bya ste
45 bar gcad: Gt rab gcad | 46 sde: Gt ste | 47 'khrug: Gt 'khrugs | 48 bsgrub: Gt bsgrubs | 49 'bye bar: Gt 'byed par | 50 gsung pa te: Gt
gsungs te | 51 bya ba de'i: Gt bya ba'i | 52 ma: Gt mi | 53 bgegs: Gt gegs | 54 Gt inserts: dang | 55 'bye ba'i: Gt 'byed pa'i | 56 par
bsgoms: Gt pa sgoms | 57 myi: Gt mi | 58 /: Gt omits | 59 la: Gt omits | 60 'bye: Gt 'byo | 61 Gt inserts: na | 62 byin rlabs: Gt byin
brlabs | 63 gtsor: Gt gtso bor | 64 bsgrub par: Gt bsgrubs pa | 65 Gt inserts: / | 66 du/: Gt tu | 67 stong: the letters, sa and ta, are
smudged but the intention would seem clear. | 68 dang mnyam pa nyid: Gt omits | 69 myi: Gt mi | 70 'bye bar: Gt byed par | 71 ma
les: Gt kar ma ma le | 72 Gt inserts: 'phags pa | 73 pad ma: Gt padmo | 74 Gt inserts: /

Chapter 19
[Ms63r.5] [Gt294.4]
//da ni drag po 'i hom1 gyi las ***** bshad de /
/hom khung zur gsum [5] gdengs pa la /
/rdo rje mda' gzhu ldan byas te /
/ thod pa dag gis rnam par brgyan /
/spyan drangs2 bsregs na lha yang 'chi /
/zhes [6] gsungs te /
/thab khung zur gsum3 du byas la /
/grwa mda' gzhu rnams4 bkang nas /
khro bos 'phen bar bsgom /
tha mar5 phur pa [63v.1] nag po gsum btab la /
/ lcags thag gam /6
thag pa nag pos bskor te/
/khro bo dang khro mo dang 'phra men gong nas 'byung ba7 rnams spyan [2] drangs la8 /
/ [tsh]er9 ma'I shing las mye sbar10 nas /
/tsha [Gt295] ba'I11 mar dang lan tsa12 dang /
/dug dang myi13 khrag ra 'i khrag /
/ske tse14 ldong ros lcags phye [3] la15 /
/khro bo khro mo 'phra men gyi / ****
/gsang sngags rnams ni bzlas nas su / *****
/gang la bya ba'i gzugs dag ni / /
[4] mying16 rus bsregs na 'grub par 'gyur/ *****
/'phra men mngags te bkug17 nas ni / *****
/thab du18 bsregs na mod19 la 'chi /
[5] gsur dang bsres20 pa'i sha rnams kyis/
/ lha rnams thams cad mchod par bya /
/rdo rje glu ni blangs nas su /
/'byung po 'i sha ni shin du21 zhim/
[6] 'byung po 'I khrag ni shin du22 zhim/
/'byung po 'i rus pa zhim ba23 ste /
/gsol te thams cad dgyes par mdzod /
/a la la la24 ho /
[64r.1] $/ /ha ha ha h25 zhes brjod par26 bya 'o/
/gzhan du na/27

1 hom:

Gt ham | 2 drangs: Gt drang | 3 gsum: Gt omits | 4 rnams: QtGt rnam par (A singular would seem more likely: QtGt's
reading would mean, in the form of [a] bow and arrow, and this may refer to a bow shape which is made within the hearth.
Another possibility would be that the bow shape might be drawn around the hearth, connecting the triangle points, but this is less
likely since it would seem inconsistent with the parallel chapters below on the fire rituals for the other ritual types Chapters 25,
30 and 35 where the different symbols mentioned would seem to be depicted on or within the hearth.) | 5 tha mar: Gt tha ma | 6 /:
Gt omits | 7 gong nas 'byung ba: Gt ma | 8 la: Gt te | 9 [tsh]er: the paper here has rucked up partially obscuring the letter tsha, but
the intention would seem clear. | 10 mye sbar: Gt mer spar | 11 ba'I: Gt kha'i | 12 tsa: Gt tshva | 13 myi: Gt mi | 14 ske tse: Gt skye
tshe | 15 phye la: Gt phya las | 16 mying: Gt ming | 17 bkug: Gt bkugs | 18 du: Gt khung | 19 mod: Gt mong | 20 bsres: Gt bsregs
(either reading is possible; the Tshal pa Kanjur also gives bsres, but all other editions give bsregs.) | 21 du: Gt tu | 22 du: Gt tu
23 ba: Gt pa | 24 la la la: Gt la la (a la la ho is the usual refrain in the practice tradition; Gt's reading is shared by the Bhutanese
texts and the Hemis Kanjur. The Tshal pa Kanjur texts omit these syllables, but the South Central NGB and the other local
Kanjurs agree with the Dunhuang manuscript's reading, which is most likely to have been in the archetype.) | 25 Gt inserts: /
26 par: Gt pa | 27 /: Gt omits

318

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

sku gsung thugs kyi ye shes kyi myes28 /

sku gsung thugs dang myi 'thun ba'i29 [2] phyogs 30 bsregs31 pa dang /
ye shes kyI byin rlabs32 kyis mda' 'phangs pa ltar 'phro bas phog33 ste/
sku gsung thugs dang 'thun34 par gyur [3] pa ni /
khro bo dang khro mo 'i rang bzhin ***** gyis mnyes pa yin zhing /
de ltar go ba ***** ni 35 dbyangs blang shes /36
rI bo brtsegs [4] pa'I tan tra las 'byung ngo37 / *****
/thabs kyi zhags pa pad ma38 'phreng las *****
/ drag po 'i hom gyi le'u ste39 bcu [5] dgu 'o40 // :

28 mye:

Gt me | 29 myi 'thun ba'i: Gt mi mthun pa'i | 30 Gt inserts: thams cad | 31 bsregs: Gt bsreg | 32 byin rlabs: Gt byin gyis
brlabs | 33 'phro bas phog: Gt / 'phro bar phog phog | 34 'thun: Gt mthun | 35 Gt inserts: glu | 36 blang shes /: Gt blangs shing
37 ngo: Gt de | 38 pad ma: Gt padmo | 39 Gt inserts: / | 40 dgu 'o: Gt dgu pa'o

Chapter 20
[Ms64r.5] [Gt295.6]
//da nI drag po 'i 1 las bshad de/
/lcags sam nag po 'i shing rnams [Gt296] la/
/mgo bo rgya mdud2 [6] rtse zur gsum/
/khro bo khro mo 'khor3 bsgoms te/
/btab na4 lha yang brlag5 par 'gyur/
zhes6 gsungs te/
lcags dang7 shing kha dog [64v.1] nag po tsher ma can la/8
phur pa mgo bo rgya mdud/9
rtse zur gsum du bzhogs10 te/
rgya mdud kyi steng du he ru ka11 dang /
ral pa gcig ma/12
ngos [2] su go rims bzhin du /
khro bo dang khro mo rnams bsgoms te/
gang la bya ba'i gzugs dang /13
mying14 rus la btab na /
[beneath the line] [chu?] klung la btab na skams so ril15 btab na ra16 rnyil lo

lha yang gzer thabs dang [3] rIms kyI17 btab ste18 'chi bar 'gyur ro/*****
/ri rab dang gleng19 gbzhI20 'I ngos la stsogs21 ***** pa la bris te btab na /
/phur pa [4] gcig gis phyogs bcur btab par ***** 'gyur ro/
/gzhan du na gnyis su myed22 ***** pa'I ye shes stong pa nyid kyi23 byin [5] rlabs24 khro bo dang khro mo
yin la /
des ma khyab pa myed25 de /
sems kyi phur pa gcig btab pas /
phyogs bcur btab par 'gyur zhes [6] 'gu hya26 dang ki la ya bcu gnyis kyi tan tra27 las 'byung ngo /
thabs kyi zhags pa pad ma28 'phreng las /
/drag po 'i phur pa'i le'u ste29 [65r.1] $/ /nyi shu 'o30// :

1 Gt

inserts: phur pa'i (presumably, the Dunhuang manuscript omits this in error) | 2 rgya mdud: Gt brgya mdud | 3 'khor: Gt
rnams (the root texts support Gt here) | 4 na: Gt ste | 5 brlag: Gt rlag | 6 zhes: Gt ces | 7 dang: Gt sam | 8 kha dog nag po tsher ma
can la/ : Gt nag po kha dog can la | 9 phur pa mgo bo rgya mdud/: Gt phur pa/ rgya mdud | 10 bzhogs: Gt gzhogs | 11 he ru ka: QtGt
h ka ra (the Dunhuang manuscript reading would seem more appropriate here) | 12 /: Gt omits | 13 /: Gt omits | 14 mying: Gt
ming | 15 ril: ri la intended? | 16 ra: perhaps for, rwa; or ri might be intended. | 17 kyI: Gt kyis | 18 Gt inserts: / | 19 gleng: presumably,
gling intended; QtNt gling; Gt gleng | 20 gbzhI: it appears that gzhI has been amended to bzhI; Gt bzhi. | 21 stsogs: Gt sogs
22 myed: Gt med | 23 kyi: Gt kyis | 24 byin rlabs: Gt byin brlabs/ | 25 myed: Gt med | 26 'gu hya dang: QtGt gu hya ti la ka dang/
27 tan tra: Gt tantra | 28 pad ma: Gt padmo | 29 Gt inserts: / | 30 nyi shu 'o: Gt nyi shu pa'o

Chapter 21
[Ms65r.1] [Gt296.5]
//da ni1 drag po 'i gtor ma bshad de2/
lcags phye dug dang rdo phye rnams/3
/khrag las sbus4 [2] te bsres5 pa dang/
/mud ga 'i chan6 dang sbyar ba dag/
/sbyar bas gtor na de brlag7 'gyur/
/zhes gsungs te/
lcags phye8 dang 9 dug rnas [3] myI10 'am khyi [Gt297] 'am ra 'i khrag las11 ***** sbrus te/
sran mud dga'12 'am/
'bru nag ***** po 'i13 chan dang14 sbyar te gtor ma byas [4] la15/
gang la bya ba de16 bkug ste byin ***** na/
mthong bas khrag du17 skyug te/18
rul ***** nas 'chi bar 'gyur ro/ /
[5] gzhan du na stong pa nyid kyI ye shes kyis kun la khyab pa'i phyir19 de nyid bsgoms pas rang bzhin gyis
zhigs20 shing rul par21 'gyur ro22 [6] zhes sgyu 'phrul dra ba las 'byung ngo/
/thabs kyi zhags pa pad ma23 'phreng las /
drag po 'i gtor ma 'I le'u ste/
/nyi shu [65v.1] gcig go24/ // :

1 da

ni: Gt omits | 2 de: Gt do | 3 lcags phye dug dang rdo phye rnams/: QtGtNt lcags dag dang ni rdo rje dag/ (none of the root texts
support the Tenjur reading here, which also seems less appropriate) | 4 sbus: the expected word, sbrus, might be intended, but there
is no clear indication of the attached ra. The instance of sbrus in line 3 below is perfectly clear and quite different from the
syllable here, so sbus is likely simply to be a scribal error; Gt: sbrus | 5 bsres: Gt bsreg (only the Tenjur gives this unlikely
reading) | 6 mud ga 'i chan: Gt mu dga' can (on mud ga/mud dga', see the discussion in the TZ edition, p.193.) | 7 brlag: Gt rlag
8 lcags phye: Gt unclear | 9 QtGt insert: rdo rje (we appear to have omitted reference to rdo phye in the root verse above; QtGt's
rdo rje may perhaps be an error for it? Note that QtGt also give rdo rje in their citation of the TZ root verse above.) | 10 myi: Gt /
mi | 11 las: Gt la | 12 a chung subscribed. | 13 nag po'i: Gt nag po | 14 chan dang: Gt tsan dan dang (the Tenjur reading seems
inappropriate here) | 15 gtor ma byas la: Gt omits | 16 de: Gt omits | 17 du: Gt tu | 18 skyug te/: Gt skyugs | 19 Gt inserts: / | 20 gyis
zhigs: Gt omits | 21 rul par: Gt rul bar | 22 Gt inserts: / / | 23 pad ma: Gt padmo | 24 gcig go: Gt gcig pa'o

Chapter 22
[Ms65v.1] [Gt297.3]
//da nI drag por1 sbyor ba'I las bshad de/
/khro mo 'i2 dkyil 'khor gtum cen3 du/
/khro bo 'i [2] tho bas brdungs pa yi/
/rdo rje gtun 'phrugs4 bsgoms byas na5/
/lha yang rung ste brdungs6 par 'gyur/
zhes gsungs ste7/
khro bo8 rnams kyi dkyil9 [3] du/
/khro bo rnams kyis brdungs par10 bsgom*****s na/
/gang yang rung ste11 brdungs par 'gyur ro/*****
/gzhan du na khro mo lta bu chos kyi dbyings [4] nyId ni/
/'jIg rten thams cad la12 khyab ***** pa'i gtun no/
/khro bo lta bu 'i dbying*****s 13 rig pa ni/
/'jig rten thams [5] cad 14 brdungs pa'i15 tho ba zhes/
las16 kyi 'phreng ba'i tan tra las 'byung17 ngo/
/gzhan gis18 drag po 'i sbyor ba zhig byas na yang/
las 'di [6] rnams kyis byed pa nyid la slar log19 par 'gyur ro20/
/drag po 'i las byed [Gt298] pa'i tshe/
chas gos nag po rnam 21 lnga dang ldan bar bya 'o/
[66r.1] $/ /drag po 'i las rnams bshad zin to22/
/thabs kyi zhags pa pad ma23 'phreng las/
/drag por 'byor pa'i24 le'u ste25 nyi shu [2] gnyis po26// :

1 por:

Gt po'i | 2 khro mo 'i: Gt khro bo khro mo'i (only the Tenjur gives this less appropriate and unmetrical reading) | 3 gtum cen:
Gt gtun chen (but gtun khung, found in the root texts, is surely the correct reading) | 4 'phrugs: Gt phrugs | 5 na: Gt nas | 6 brdungs:
Gt rlag (none of the root texts support Gt here, and brdungs occurs also in the commentarial explanation below) | 7 ste: Gt te
8 khro bo: Gt khro mo (Gt's reading is surely more likely here, and fits with the quoted root text.) | 9 dkyil: Gt dkyil 'khor (Gt's
reading is surely more likely here, and fits with the quoted root text.) | 10 brdungs par: Gt brdung bar | 11 ste: Gt te | 12 la: Gt omits
13 dbyings: Gt dbyings nyid | 14 Gt inserts: la | 15 brdungs pa'i: Gt brdung ba'i | 16 las: QtGt thabs | 17 'byung: Gt byung | 18 gis: Gt
gyis | 19 log: Gt ldog | 20 'gyur ro: Gt 'gyuro | 21 Gt inserts: pa | 22 to: Gt no | 23 pad ma: Gt padmo'i | 24 drag por 'byor pa'i: Gt drag
po'i las kyi | 25 Gt inserts: / | 26 po: Gt pa'o

Chapter 23
[Ms66r.2] [Gt298.2]
//da ni gnas bstabs1 pa'i cho ga bstan t[e(/o)]2/
/drag po 'i las kyis3 bsgrubs pa'i4 rnams/ /
[3] 'khor ba'i5 lam lnga6 sgo bcad nas/ *****
/rnam shes h gyis7 bsdus brdub te8/ *****
/'og myin9 bla myed gnas mchog [4] na10/
/gtso bo lnga dang11 gtso mo lnga/ *****
[beneath the line, linked to gtso bo] ye shes [beneath gtso mo] thabs [beneath lnga] shes rab

/sems dpa' dang ni sems ma dang/ *****

/khro bo dang ni khro mo rnams/ /
[5] sbyor ba'i lhums su phat12 kyis 'phang/
[beneath the line] yab yum kyI sbyor ba mdzad pa'i ting nge 'dzin gsum

/zhes gsungs te/

drag po 'i las kyis bsgrubs pa'i13 rnams/
/lam rgyud lnga14 sngags [6] 'bru lngas sgo bcad15 par bsgoms na16/
/rnam par shes pa h gyis bsdus te17/
/bdag gi ltor brngubs18 nas/
'og myin19 gyi gnas na [66v.1] lha gong nas20 smos pa'i rnams sbyor pa'i21 mdzad pa'i lhums su22 phat23 kyis
'phangs te btang na/
/nges par 'grub ces 24 du ma [2] las 'byung ngo/
/tan tra snying rje rol pa las kyang 'byung ngo/
/gzhan du na chos kyi dbyings mnyam pa nyid la snyoms par zhugs [3] pa'I tshe/
/sems can dmyal *** bar ltung ba'i25 rgyu ni zhe sdang la/26
/dngos po yod par lta ba 27 yin te/
/zhe sdang la [4] rang bzhin myed28 par shes shing/ *****
/zhe sdang gi sems myi29 skye bas na/ *****
dmyal ba'i [Gt299] sgo bcad30 pa yin no/
1 bstabs:

Gt btab (gnas bstabs is from gnas su stob pa, which implies the conferring of a higher place/abode of rebirth. This is a
well-established feature of sgrol ba [forcible release] rites. Amongst Dunhuang sources, for instance, the term occurs in PT42
[pages r.25 line 2, and v.71 line 1, as numbered in the Mellon International Dunhuang Archive images; see also Meinert 2006:
118-119, 122, 124]. Gt's reading, gnas btab, might suggest the establishment of a higher place/abode of rebirth, thus expressing
the same idea. However, none of the versions, including Gt, support Gt's reading in the concluding title at the end of the chapter
[although Gt repeats gnas btab just above this title]). | 2 bstan t[e(/o)]: a 'greng bu has been written over a na ro or vice versa (the
na ro looks more like the original writing but it is hard to be certain); Gt bya ste | 3 kyis: Gt kyis kyis (but it appears the first may
be deleted by dots above the letter; Qt does not share the additional kyis) | 4 bsgrubs pa'i: Gt sgrol ba'i (none of the root texts
support Gt here) | 5 ba'i: Gt ba | 6 lnga: Gt lnga'i | 7 gyis: Gt gis | 8 bsdus brdub te: Gt bsdu rdungs nas (Here, the Dunhuang
manuscript reading of brdub is likely to be an error for brngub, which would seem more appropriate. It is also consistent with the
reading, brngubs, which occurs in the Dunhuang manuscript version of TZComm below the root text verse. The Tawang Kangyur
gives rngubs, and the Bhutanese texts also give rngubs or rngub, but the other root texts similarly give the main letter as da or ta.
This is possible evidence of an earlier scribal error due to an dbu can source creating a nga/da confusion. The South Central NGB
and Hemis Kanjur's various readings [bsdubs/bdubs/bsdus/brtugs] may represent further corruptions, but quite probably, the error
was in the archetype [and perhaps corrected in the Tawang and Bhutanese texts]. Gt's reading of brdung would seem even more
remote from an earlier clear reading.) | 9 myin: Gt min | 10 bla myed gnas mchog na: Gt gnas mchog bla na med (none of the root
texts support Gt here) | 11 dang: Gt 'am | 12 phat: Gt pha | 13 bsgrubs pa'i: Gt bsgral ba'i (Gt's reading is plausible, although as
noted above, Gt's reading of sgrol ba'i in the root text citation was not supported by any of the root text editions) | 14 Gt inserts: /
15 sgo bcad: Gt sgo cad | 16 na: Gt nas | 17 gyis bsdus te: Gt gis bsdus nas | 18 brngubs: Gt brdungs | 19 myin: Gt min | 20 nas: Gt na
21 rnams sbyor pa'i: 'i appears to be deleted by a line crossing it through; Gt rnal 'byor ba | 22 Gt inserts: / | 23 phat: Gt pha | 24 Gt
inserts: tan tra | 25 ltung ba'i: Gt lta ba'i | 26 /: Gt omits | 27 Gt inserts: ni | 28 myed: Gt med | 29 myi: Gt mi | 30 sgo bcad: Gt sgo
gcad

Chapter 23

323

[5] de bzhin du 31 'dod chags dang32 gti mug la rang bzhin myed33 par shes shing/
/de dag gi sems myi34 skye bas na/
yi dags dang35 byol song gi sgo [6] yang bcad pa yin no/
/lha dang myir36 skye ba'i rgyu ni/
dge ba bcu dang bsam gtan bzhi dang/
snyoms par 'jug pa bzhi bsgrub37 pa yin [67r.1] $/ /te38/
de dag la39 rang bzhin myed40 par rig cing41/
sems myi42 skye ba ni lha dang myi 'i43 sgo bcad44 pa yin no/
/bdag gi lto zhes [2] bya ba nI chos kyi dbyings te/
bla na myed45 pa'i phyir 'og myin gyi46 gnas shes47 bya ste/
/chos nyid kyi dbyings48 nyid rig pa'i tshe/
/lha dang [3] lha mo'i tshul du 'brel par49 rig pa ni **** sbyor ba mdzad pa ste50/
/de nyid byang chub kyi **** sems yin bas/
rang gi rnam par [4] shes pa de 'byor par51 bsgoms na/52 *****
nges par 'grub bo/
dper na rgyal po ra ***** ma nas53 bsad pa thams cad lhar [5] skye bar54 smon lam btab pas/
de bzhin du ra ma nas55 bsad pa lhar 'grub bo zhes/56
/drang srong mngon bar57 shes pa can gyis bshad pa ste/
[6] 'di ltar chos kyi dbyings la snyoms par sbyor ba lta ci smos58/
/zhes59 snying rje rol pa'i tan tra las 'byung ngo/
/gnas bstabs60 pa'i [67v.1] cho ga bshad zin to/
/thabs kyi zhags pa pad ma61 'phreng las/
gnas bstabs62 pa'i le'u ste63 nyi shu gsu mo64// :

31 Gt

inserts: yang | 32 Gt inserts: / | 33 myed: Gt med | 34 myi: Gt mi | 35 dang: Gt omits | 36 myir: Gt mir | 37 bsgrub: Gt bsgrubs
Gt no | 39 la: Gt omits | 40 myed: Gt med | 41 cing: Gt pa ni | 42 myi: Gt mi | 43 myi 'i: Gt mi | 44 sgo bcad: Gt sgo gcod
45 myed: Gt med | 46 myin gyi: Gt min | 47 shes: Gt zhes | 48 chos nyid kyi dbyings: Gt chos kyi dbyings | 49 tshul du 'brel par: Gt
tshul 'brel bar | 50 sbyor ba mdzad pa ste: Gt omits | 51 de 'byor par: Gt der 'byo bar | 52 /: Gt omits | 53 nas: Gt as | 54 Gt inserts:
'gyur bar | 55 ra ma nas: Gt sngar nas | 56 / /: Gt omits | 57 bar: Gt par | 58 smos: Gt smos te | 59 zhes: Gt omits | 60 gnas bstabs: Gt
gnas btab | 61 pad ma: Gt padmo | 62 bstabs: Gt bstab | 63 Gt inserts: / | 64 gsu mo: Gt gsum pa'o
38 te:

Chapter 24
[Ms67v.1] [Gt299.6]
// [2] da ni dbang gi las bshad de/ [Gt300]
/zla gam dkyil 'khor rtsibs brgyad par/
/khro bo khro mo1 spyan drangs te/
/mchod pa'i glu la stsogs pa'm2 mchod/ /
[3] de nas las rnams brtsam bar3 bya/ *****
/zhes gsungs te/
dbang gyi4 dkyil 'khor du *****
/lha go rims5 bzhin du spyan drangs te/
[4] de nas glu dang6 gar dang a rga las stso*****gs pa7 yo byed8 dmar9 pos10 mchod par bya 'o11*****/
chags pa'i rdo rje 'khor bcas pa/
[5] 'jig rten ni rab 'byam12 kun/
/chags pas yul13 la dbang mdzad de/
/dbang du ma gyur gcig kyang med/14
/ces bstod de15/
de nas chas [6] gos rnam lnga dang16 ldan bas las bya 'o/17
/rdo rje 'phra men mngags nas su/
/bkug nas khro bo chen po 'is18/
/chags pa'i tshul gis19 [68r.1] $/ /de lus las20/
/dpal dang gzi brjid bsdus na21 'grub/
zhes22 gsungs te/
rdo rje 'phra men23 rnams mngags ste24/
/gang dbang du [2] bya ba de bkug25 ste/
/de 'i26 lus dang sems dang ngag las27 dpal dang28 gzi brjid bsdu zhing/
/ci yang nyan par bsgoms na29 dbang du 'gyur te/
[3] bu la ma byams pa 'am30/
[beneath line 3, and continuing under line 4, as indicated by the crosses] $// pha rol du phyin pa zhig la byas na bdag gtso bo nyid
bsgoms ** pa ** la rang gi dbyings mar bsgrubs te sbyor ba byas pa ni 'grub [d?]es 'byung ba ni bud med dbang du bya ba'i don no
kha gcig las pha rol dbang du bya zhe na skabs su +

lus la **** grib ma 'khor ba dang 'dra bar/

/bdag la **** 'khor zhing rigs myi 'thun pa31 zhig na [4] yang32 lus ston cing/33
[beneath line 4, continuing from line 3] + shes rab ni skyes pa ste bdag nyid gtso bor bsgoms pa'i dbyings mar bsgoms te sbyar na
bub ces bshad do/ mdor na pho mor gang la yang bdag phyag rgya chen por 'Is34 'dzin pa/

/ci bsgo ba ***** rnams35 bran bzhin byed de/

ci sems pa ***** yang smra na/
gzhan lta ci smos [5] te 'grub bo36/

1 khro

mo: Gt khro mor | 2 stsogs pa'm: Gt sogs pas | 3 brtsam bar: Gt rtsam par | 4 gyi: Gt gi | 5 lha go rims: Gt lha rnams go rim
dang: Gt glu | 7 a rga las stsogs pa: Gt arga la sogs pa/ | 8 yo byed: Gt yo byad | 9 dmar: prefixed da subscribed | 10 pos: Gt po
11 mchod par bya 'o: Gt mchod pa bya'o | 12 ni rab 'byam: Gt khams ni rab 'byams (the Dunhuang manuscript is unmetrical here,
and clearly in error, as is clear also from the root text editions) | 13 yul: Gt yud (the Tshal pa Kanjur texts support Gt here, while
the Bhutanese texts omit the whole verse; the other root texts give, kun la mnga', but any of the readings could fit) | 14 gcig kyang
med/: Gt gang yang med (the root texts support Gt here) | 15 /ces bstod de: Gt ces ston to | 16 dang: Gt yang | 17 bas las bya 'o/: Gt
par bya'o/ de yang gzhung las/ | 18 'is: Gt yis | 19 gis: Gt gyis | 20 las: Gt la | 21 na: Gt nas | 22 zhes: Gt ces | 23 'phra men: Gt 'phra
men ma | 24 ste: Gt nas | 25 bkug: Gt bkug pa | 26 de 'i: Gt de'i | 27 sems dang ngag las: Gt ngag dang sems dang | 28 dang: Gt
dang/ | 29 na: Gt na/ | 30 pa 'am: Gt pa'am | 31 myi 'thun pa: Gt mi mthun pa | 32 yang: Gt yang/ | 33 cing/: Gt cing | 34 'Is: sa
subscribed, possibly sa is intended as a separate word here, so this would read, sa 'dzin pa, which could make good sense.
35 rnams: Gt omits | 36 bo: Gt pa'o
6 glu

Chapter 24

325

/gzhan du na stong pa nyid kyi chos kyi dbyings kyis/

/phyi nang gi dngos po thams cad la khyab pas/ [Gt301]
tha myi37 dad pa ni chags [6] par38 rig pa'i ye shes kyis/
phyogs bcu39 snod dang40 bcud kyi 'jig rten thams cad/
bdag dang tha myi41 dad par rig pas/
rang bzhin gis42 [68v.1] byams shing 'khor ba yin te/
/bran bzhin byed pa yang/43
bdag dang tha myi44 dad pas/
/thams cad bdag yin pa'i phyir/
/bdag nyid [2] chen po zhes kyang de 'i45 phyir bya/
/rnam gcig du na/46
thams cad bdag gi47 sprul pa yin no48 zhes49/
/dbang cen bsdus50 pa'i tan tra las 'byung ngo/
[3] 'phra men51 rnams ni52 mngags nas su53/ ***
/de myig54 rna ba sna dang lce/
/lus dang sem****s bsgyur rnam55 bsgoms na56/ /
[4] kun la mdzes shing dga'57 bar 'gyur/ *****
/zhes gsungs te/
rdo rje 'phra men ma ***** rnams mngag ste58/
de 'i dmyig59 dang [5] rna ba dang sna dang lce dang/60
lus dang sems61 bdag la dga'62 ba dang/
/bdag63 gis nyes pa byas na yang64 mdzes par snang bar65 dbang po66 bsgyur [6] bar67 bsgoms na/
/de bzhin du 'grub ces bya ba'i tha tshig go/
/gzhan du na/68
stong pa nyid kyI ye shes kyis kun la khyab pas69 [69r.1] $/ /bdag dang tha myi70 dad par rig pa nyid kyis71
dbang po la khyab par bsgyur72 ba yin te/
dper73 [p?]la74 bcol mkhan gyis myi75 rigs pa [2] zhig byas na yang/
kun dga'76 bar 'gyur te/77
rgod pa78 lta bu 'o79/
/zhes rI bo brtsegs pa las 'byung ngo/
/thabs kyi zhags pa pad ma80 'phreng [3] las/
:
/dbang gi las bstan pa ste81 **** nyi shu [Gt302] bzhi pa 'o82//

37 myi:

Gt mi | 38 ni chags par: Gt chags pa yin/ de ltar chags pa'i | 39 bcu: Gt bcu'i | 40 snod dang: Gt snod | 41 tha myi: Gt tha
Gt gyis | 43 yang/: Gt yang | 44 myi: Gt mi | 45 de 'i: Gt de'i | 46 du na: Gt tu | 47 gi: Gt gis | 48 no: Gt omits | 49 zhes: Gt zhes
bya | 50 cen bsdus: Gt chen sdus | 51 men: Gt men ma | 52 ni: Gt omits | 53 nas su: Gt so (none of the root texts support Gt here)
54 de myig: Gt de'i mig | 55 rnam: Gt rnams (the root texts support Gt here) | 56 na: Gt nas | 57 a chung subscribed | 58 mngag ste :
Gt mngags te | 59 de 'i dmyig: Gt de'i mig | 60 dang/: Gt dang | 61 sems: Gt sems/ | 62 a chung subscribed | 63 la dga' ba dang/ /bdag:
Gt omits (eyeskip) | 64 yang: Gt yang/ | 65 par snang bar: Gt pa dang/ | 66 po: Gt pos | 67 bar: Gt omits | 68 na/: Gt na | 69 pas: Gt
pas/ | 70 myi: Gt mi | 71 kyis: Gt kyis/ | 72 bsgyur: Gt 'gyur | 73 dper: Gt dper na | 74 [p]la: bla most likely intended but letter appears
to read sla; Gt bla | 75 myi: Gt mi | 76 a chung subscribed | 77 'gyur te/: Gt gyur te | 78 rgod pa: Gt rgong ba | 79 bu 'o: Gt bu'o
80 pad ma: Gt padmo | 81 pa ste: Gt pa'i le'u ste/ | 82 pa 'o: Gt pa'o
42 gis:

Chapter 25
[Ms69r.3] [Gt302.1]
//da ni dbang gi hom bshad de1/ /
[4] hom khung zla gam mdengs pa la2/ ******
/rdo rje 'i3 zhags pas brgyan par bya/ *****
/de nas lha rnams spyan drang4 te/ /
[5] myi phan5 rnam rtog bsregs na 'grub /
/zhes6 gsungs te/
hom khung zla gam la/
rdo rje zhags pas brgyan te/
/kha dog dmar po 'i7 [6] phur pa lnga btab la8 /
/mtha'9 thag pa dmar pos bskor te/
shim10 dri ma mnam pa la mye sbar11 nas/
zangs phye dang/12
bu ram dang [69v.1] yungs kar dmar po dang /
gang la bya ba'i snying po rgya skegs13 kyis bris te /
/bsreg cing rdo rje 'phra men mas bdag la myi14 phan zhing15 [2] myi 'jar bar16 sems pa'i rtog pa bsregs par
bsgoms na /
ci bsgo ba nyan cing phan bar17 sems par 'gyur ro /
/gzhan du 18 rnam par [3] rtog pa ni bdag du19 lta ba las 'byung ***** ste /
stong pa nyid kyi ye shes kyi myes20 ***** bdag du21 lta ba dang /
de las byung ba'i [4] rnam par rtog pa thams cad bsregs ***** nas /
mnyam pa nyid kyis bdag nyid ***** chen po dang22 ro gcig pa ni 'jar ba [5] zhes bya /
/bdag du23 24 dngos por zhen nas rnam par rtog pa ni /
/25'khor ba'i26 rgyu chen po yin pas /
myi27 phan pa zhes bya ste /
[6] bdag myed28 pa'i ye shes rig na /29
'khor ba dang bral bas phan bar sems pa zhes bya 30/
/de ltar rig na31 bdag duang32 gzhan du tha myi33 [70r.1] $ / /dad pas /
gzhan yang rang bzhin gyis 'jar34 zhing phan par35 sems par [Gt303] 'gyur ro /
/zhes glang po36 las bshad do/
/thabs kyi [2] zhags pa pad ma37 'phreng las/
:
dbang gyi38 hom gyi le'u ste39 nyi shu lnga pa 'o//

1 de:

Gt par bya ste | 2 mdengs pa la: Gt gdengs pa las | 3 rdo rje 'i: Gt rdo rje | 4 spyan drang: Gt spyan drangs | 5 myi phan: Gt mi
'pham (the root texts all support the Dunhuang manuscript reading, which makes clearer sense here, and is reinforced by the
commentarial elaboration below) | 6 zhes: Gt ces | 7 po 'i: Gt po | 8 lnga btab la: Gt btab ste | 9 mtha': 'a subscribed | 10 shim: QtGtNt
shing (QtGtNt's reading of shing would simply mean wood, and may be correct. However, if shim is correct, Roerich [Volume 9:
253] gives "Dalbergia Sissoo" [i.e. Indian rosewood] for shim sha pa, so it is possible that shim is short for this. Another
alternative is suggested by bTsan lha ngag dbang tshul khrims [1997: 939], who gives shim shim as equivalent to zhim pa, ie.
aromatic, a meaning which would reinforce the other adjectives. In this case, the implication that a type of wood is being referred
to would simply be implied.) | 11 mye sbar: Gt me spar | 12 /: Gt omits | 13 rgya skegs: Gt rgya skyags (= rgya skyegs, shellac)
14 myi: Gt mi | 15 Gt inserts: / | 16 myi 'jar bar: QtGtNt mi mdza' ba'i (QtGtNt's reading could fit here, but given the repeated
occurrence of 'jar below, the Dunhuang manuscript reading would seem preferable) | 17 bar: Gt par | 18 Gt inserts: na | 19 du: Gt tu
20 myes: Gt mes/ | 21 du: Gt tu | 22 Gt inserts: / | 23 du: Gt tu | 24 Gt inserts: lta ba'i | 25 / /: Gt omits | 26 'khor ba'i: Gt 'khor | 27 myi:
Gt mi | 28 myed: Gt med | 29 na /: Gt nas | 30 phan bar sems pa zhes bya: Gt sems zhes bya ste | 31 na: Gt nas | 32 duang: deletion of
the zhabs kyu appears to have been intended, the vertical stroke of the da extended further below. | 33 myi: Gt mi | 34 'jar: Gt 'dzar
(but QtNt give 'jar, and Gt may intend this) | 35 phan par: Gt omits | 36 glang po: Gt glang po che | 37 pad ma: Gt padmo | 38 gyi: Gt
gi | 39 Gt inserts: /

Chapter 26
[Ms70r.2] [Gt303.1]
//da ni dbang gi phur pa'i [3] las bshad de /
/zangs sam kha dog **** dmar po yi/
/shing la rgya mdud rtse zla **** gam/
1/mgo la khro bo 'khor2 [4] bsgoms te /
btab na nges par de ***** dbang 'gyur3/
/zhes gsungs te/ *****
zangs sam kha dog dmar po 'i shing [5] la
/phur pa mgo4 rgya mdud la/5
rtse zla gam du bzhogs la6
/rgya mdud kyi steng dang ngos7 khro bo dang khro mo rnams8 bsgoms te/ /
[6] gang du dbang du9 bya ba10 de 'i gzugs sam11 mying12 rus 13 la btab ste/
bka'14 bzhin15 byed par bsgoms na /
nges par dbang du 'gyur ro/
[70v.1] ri rab dang16 gling bzhi la stsogs17 pa la18 bris te btab na/
phur pa gcig btab pas19/
phyogs bcur btab par 'gyur te/
thams cad [2] dbang du 'gyur ro zhes bya ba'i don to/
/gzhan du na chos kyi dbyings kyis20 phyogs bcu thams cad du khyab la/
de ltar rig pa'i ye shes ni [3] phur pa yin te/
thams cad du btab ***** par 'gyur zhing /
/bdag dang tha myi21 dad ***** pas22 las thams cad rang bzhin gis23 [4] dbang du 'gyur ba yin no zhes/
*****
/24 phur pa bcu gnyis kyi rgyud phyi ma las ***** 'byung ngo/
/dbang gi phur pa bshad zin [5] to/
/thabs kyi zhags pa [Gt304] pad ma25 'phreng las/
:
dbang gi phur pa bstan pa'i le'u ste26 nyi shu drug go27//

1 Gt

inserts: /mgo la rgya mdud rtse zla gam/ (this seems to be a dittography) | 2 khro bo 'khor: QtGtNt khro bo khro mor (none of
the root texts support the Tenjur reading here, but see the TZ edition, p.199 note 8) | 3 de dbang 'gyur: Gt 'grub bo | 4 mgo: Gt mgo
bo | 5 /: Gt omits | 6 la: Gt pa | 7 ngos: Gt ngos su | 8 khro bo dang khro mo rnams: Gt khro bo khro mo | 9 gang du dbang du: Gt
gang dbang du | 10 ba: Gt ba'i | 11 Gt inserts: / | 12 mying: Gt ming | 13 Gt inserts: bris pa | 14 bka': 'a subscribed | 15 bzhin: Gt bzhin
du | 16 dang: Gt omits | 17 stsogs: Gt sogs | 18 la: Gt omits | 19 pas: Gt na | 20 Gt inserts: / | 21 myi: Gt mi | 22 pas: Gt par | 23 gis: Gt
gyis | 24 yin no zhes/ /: Gt yin no/ /zhes | 25 pad ma: Gt padmo | 26 Gt inserts: / | 27 drug go: Gt drug pa'o

Chapter 27
[Ms70v.5] [Gt304.1]
//da ni [6] dbang gi gtor ma bshad de /
/zangs phye dang ni bu ram dang /
/dmar po 'i 'bras can1 la stsogs2 pa /
/sbyor3 bas gtor ma gtang [71r.1] $/ /bar bya /
/mthong bas nges par de dbang 'gyur4/
zhes gsungs te /
zangs phye dang 'bras dmar po 'i chan la stsogs5 pa/
'bru [2] kha mdog6 dmar po 'i chan dang /
bru mar7 chu la stsogs8 pa la sbyar te /
rdo rje 'phra men gyis9 gang dbang du bya ba de byin na / 10
mthong bas de [3] dbang du 'gyur te /
zos pa lta ci smos11 **** ci bsgo ba yang nyan cing dga'12 bar13 'gyur ro / ****
/gzhan du na chos kyi dbyings mtshan [4] nyid myed14 pas thams cad du khyab ste/ *****
/ bdag nyid chen po dang tha myi dad15 pas / *****
rang bzhin gis16 dbang du 'gyur te /
[5] de ltar rig pa'i ye shes ni17 gtor ma btang ba yin zhes 18/
19zhags pa'i rgyud phyi ma gnyis las 'byung ngo /
/thabs kyi zhags pa pad ma20 'phreng las21 dbang gi gtor ma'i le'u ste22 nyi shu bdun no23 // :

1 can:

Gt chan | 2 stsogs: Gt sogs | 3 sbyor: Gt sbyar | 4 de dbang 'gyur: Gt dbang du 'gyur | 5 stsogs: Gt sogs | 6 kha mdog: Gt kha
dog |
mar: perhaps, 'bru mar intended; QtGt bu ram (the Tenjur reading would seem more consistent with the root text verse
above; bru mar might be an error for this) | 8 stsogs: Gt sogs | 9 Gt inserts: / | 10 de byin na /: QtGt bkug ste phyin nas | 11 Gt
inserts: / | 12 dga': 'a subscribed | 13 dga' bar: Gt dbang du | 14 myed: Gt med | 15 tha myi dad: Gt tha dad | 16 gis: Gt gyis | 17 Gt
inserts: / | 18 btang ba yin zhes: Gt gtang ba yin no zhes | 19 QtGt inserts: thabs kyi | 20 pad ma: Gt padmo | 21 Gt inserts: / | 22 Gt
inserts: / | 23 bdun no: Gt bdun pa'o
7 bru

Chapter 28
[Ms71r.6] [Gt304.5]
// da ni dbang du 'byor pa'i1 las bshad de/
/khro mo rnams [71v.1] kyI dkyil 'khor du /
/khro bo 'i dkyil 'khor bsgoms byas nas2 /
/ lha d.yang3 rung ste dbang du 'gyur /
/de ni bde ba chen po yin / /
[2] zhes gsungs te /
khro mo 'i dkyil 'khor du /
khro bo 'i dkyil [Gt305] 'khor gyi4 sbyor ba'i las kyis /
lha man cad5 gang yang rung ste /
thams [3] cad dga'6 zhing dbang du 'gyur bar bsgoms ***** na7 /
de bzhin du 'gyur ro8 /
/gzhan du na ***** chos kyi dbyings mtshan nyid myed9 pas [4] kun la khyab pa /
de ltar rig pa'i ***** ye shes ni10 gnyis su myed11 par 'byor te / *****
/ de ltar 12 nyid ma nor bar rig pa ni/
[5] byang chub kyi sems zhes bya ste /
/de ni bde ba bde ba13 chen po 'i dbang ngo /
/ kun du14 khyab par rig pa ni15 rang bzhin gyis dbang du 'gyur ro zhes16
[6] dbang cen17 bsdus pa'i tan tra las 'byung ngo /
/thabs kyi zhags pa pad ma18 'phreng las /
/dbang du19 'byor pa'i le'u ste /
[72r.1] $/ /nyi shu brgyad do20// :

1 dbang

du 'byor pa'i: Gt sbyor ba'i | 2 bsgoms byas nas: Gt sgoms byas na | 3 d.yang: presumably, error for yang; Gt yang | 4 gyi:
Gt omits | 5 man cad: Gt man chang | 6 dga': 'a subscribed | 7 na: Gt nas | 8 'gyur ro: Gt 'byor 'byor ro | 9 myed: Gt med | 10 ni: Gt
kyis ni/ | 11 myed: Gt med | 12 Gt inserts: 'byor cig pa | 13 bde ba bde ba: Gt bde ba | 14 du: Gt tu | 15 Gt inserts: / | 16 Gt inserts: /
17 cen: Gt chen | 18 pad ma: Gt padmo'i | 19 du: Gt omits | 20 do: Gt pa'o

Chapter 29
[Ms72r.1] [Gt305.4]
//da ni rgyas pa'i las bshad de /
/gru bzhi1 dkyil 'khor rtsibs brgyad par / /
[2] lha rnams 'khor bcas spyan drangs te /
/mchod pa glu la stsogs2 pas mchod /
/ de nas las rnams brtsam par bya /
/ zhes gsungs [3] te /
rgyas pa'i dkyil 'khor du /3 *****
lha 'khor dang bcas pa spyan drangs te /*****
/yo byed4 ser pos5 mchod6 la / /
[4] nga rgyal rdo rje 'khor dang bcas7 / *****
/ 'jig rten khams ni8 rab 'byam9 kun / *****
/ dpal dang gzi brjid chen po 'i10 rnams/
[5] yud tsam gis11 ni rgyas par mdzod12 /
/ces13 bstod de14 /
de nas chas gos rnam15 lnga dang ldan bar bya 'o/
/ khro bo khro mo rnams dang bcas16 / /
[6] rdo rje 'phra men 'khor bcas pa17 / [Gt306]
/ dpal dang gzi brjid 'phel bsgoms18 na /
/ bdag nyid chen por de19 'gyur ro /
/ zhes gsungs te /
[72v.1] lha rnams kyis bdag gi dpal dang gzi brjid 'phel bar bsgoms20 na /
/ nges par de bzhin du 'grub 'par21 'gyur ro22 zhes bya ba'i don to /
[2] gzhan du na chos kyi dbyings mtshan nyid23 myed24 pas /
phyi nang gi dngos po thams cad la khyab ste /
/ de ltar rig pa'i ye shes kyis25 [3] snod26 bcud kyi 'jig rten thams cad dang27 ***** bdag dang28 tha myi29 dad
pas30 phyogs bcu ***** thams cad duna31 /
/ bsod nams [4] 32 spyod pa las stsogs pa thams **** cad 33 bdag yin pa 'am34 /
bdag gi yin **** bar35 bsgoms na36 /
/ dper37 nyi ma'i [5] dkyil 'khor bar snang la shar pa38 /
/ nyi ma'i rang bzhin dro ba yin bas /
gzhan du dro ba yang39 nyi ma yin ba 'am /
yang na nyi ma'i [6] 'od yin bas /
nyi ma 'i40 yin pa dang 'dra bar /41
thams cad bdag gam /
bdag gir42 yin bar43 'gyur te44 /
de 'i phyir bdag nyid chen [73r.1] $/ /po yin zhes/45
dpal 'phreng dkar po las 'byung ngo /

1 bzhi:

Gt bzhi'i | 2 stsogs: Gt sogs | 3 /: Gt omits | 4 yo byed: Gt yo byad | 5 pos: Gt por | 6 mchod: Gt byas | 7 dang bcas: Gt bcas
pa | 8 ni: Gt 'di | 9 'byam: Gt 'byams | 10 po 'i: Gt po | 11 gis: Gt gyis | 12 rgyas par mdzod: Gt rgyas mdzad de | 13 ces: Gt zhes | 14 de:
Gt do | 15 rnam: Gt rnam pa | 16 rnams dang bcas: Gt 'khor bcas yang (most of the root texts give, 'khor bcas dang here) | 17 pa: Gt
pas | 18 bsgoms: Gt sgoms | 19 de: Gt der | 20 bsgoms: Gt sgoms | 21 'par: Gt par | 22 'gyur ro: Gt 'gyuro/ | 23 mtshan nyid: Gt nyid
24 myed: Gt med | 25 Gt inserts: / | 26 Gt inserts: dang | 27 Gt inserts: / | 28 dang: Gt tu | 29 myi: Gt mi | 30 Gt inserts: / | 31 duna: it
appears that an original du has been deleted by a small stroke; Gt nas | 32 Gt inserts: dang | 33 las stsogs pa thams cad: Gt thams
cad la sogs pas/ | 34 pa 'am: Gt pa'ang | 35 bar: Gt par | 36 na: Gt nas | 37 dper: Gt dper na | 38 shar pa: Gt shar ba ni | 39 dro ba yang:
Gt 'gro ba 'am | 40 nyi ma 'i: Gt nyi ma | 41 bar /: Gt bas | 42 gir: Gt gi | 43 bar: Gt par | 44 te: Gt ro | 45 zhes/: Gt ces

Chapter 29

/thabs kyi zhags pa pad ma46 'phreng las /

/47 rgyas pa'i las [2] kyi48 le'u ste49 nyi shu dgu 'o50// :

46 pad

ma: Gt padmo | 47 / /: Gt omits | 48 las kyi: Gt omits | 49 Gt inserts: / | 50 dgu 'o: Gt dgu pa'o

331

Chapter 30
[Ms73r.2] [Gt306.5]
// da ni rgyas pa'i hom bshad de1 /
/ hom khung gru bzhi gdengs pa la /
/ rdo rje 'khor [3] los2 brgyan par bya /
/ de nas lha ***** rnams spyan drangs te /
/ bsregs pas *** mchod cing dpal 'phel bsgom3 /
[4] zhes gsungs te /
hom khung gru bzhI ***** la /4
rdo rje 'i5 'khor [Gt307] los brgyan par ***** bri ste6 /
phur pa kha dog ser po [5] bzhi btab la /
mtha'7 thag pa ser pos bskor te /
de nas lha rnams spyan drangs la8 /
rin po che sna lnga 9/
sman sna lnga /10
'bru sna lnga [6] dang 11 'bras chan dang /
mar khu la stsogs12 pas bsrags13 shing mchod la14 /
bdag gi dpal la stsogs15 pa 'phel bar bsgoms na /
[73v.1] de bzhin du 'grub par 'gyur ro16 zhes bya ba'i don to /
/ gzhan du na chos kyi dbyings 17 mye18 long lta bu 'i ye shes dmyigs19 pa'i dus gcig na /
[2] gtso bo lnga dang gtso mo lnga ni20 sku bur 'byung la21 /
/sems dpa' bcu drug dang sems ma bcu drug ni22 mtshan lta bur 'byung23 la /
de lta bu 'i24 [3] dbu rgyan lnga lnga ni dpe byed25 du 'byung26 ste / ****
/ spyir ye shes kyis27 dngos po bsregs shing*****
/rang bzhin gyis rgyas28 par 'gyur la /
[4] nyan thos las khyad par che ba'i phyir 29 ***** dpal 'phel zhes /30
thabs kyi zhags ***** pa las 'byung ngo /
/thabs kyi [5] zhags pa pad ma31 'phreng las /
rgyas pa'i hom kyi32 le'u ste33 sum cu pa 'o// :

1 de:

Gt do | 2 los: Gt lo | 3 bsgom: Gt sgom | 4 /: Gt omits | 5 rdo rje 'i: Gt rdo rje | 6 bri ste: Gt bris te | 7 mtha': 'a subscribed. | 8 la:
Gt te | 9 Gt inserts: dang | 10 sman sna lnga / : Gt omits | 11 Gt inserts: / | 12 stsogs: Gt sogs | 13 bsrags: Gt bsregs | 14 la: Gt na
15 stsogs: Gt sogs | 16 Gt inserts: / / | 17 Gt inserts: la/ | 18 mye: Gt me | 19 dmyigs: Gt dmigs | 20 lnga ni: Gt lnga'i | 21 bur 'byung la:
Gt lta bur byung | 22 Gt inserts: / | 23 'byung: Gt byung | 24 de lta bu 'i: Gt de'i | 25 dpe byed: Gt dpe byad | 26 'byung: Gt byung
27 kyis: Gt kyi | 28 rgyas: QtGt brgyan (but the Dunhuang manuscript's reading would seem more appropriate) | 29 Gt inserts: ro/ /
30 zhes /: Gt zhes bya | 31 pad ma: Gt padmo | 32 kyi: Gt gyi | 33 Gt inserts: /

Chapter 31
[Here, the Dunhuang manuscript omits two entire chapters which are found in the Tenjur version, and the
root verses are also given in all editions of TZ. Chapter 31 as counted by the Dunhuang manuscript
corresponds to Chapter 33 in all other editions, and Dunhuang's chapter numbering henceforth runs two
behind all other versions until the end of the text. There is no doubt that these chapters did once exist in the
archetype; the structure of the work requires them here, and it is virtually inconceivable that this section of
the text would have been written without them. Here, we present the Tenjur version.]
[Ms omits] [Qt124a.4; Gt307.5; Nt219.1]
//da ni rgyas pa'i phur pa'i las bshad de/
gser ram ser po'i shing rnams la/
/mgo bo rgya mdud rtse zur bzhi/
/lha rnams bsgoms te gdab par bya/
/'di ni rgyas pa'i las la shis/
zhes gsungs te/
gser ram ser po'i shing rnams la/
/mgo bo rgya mdud zur bzhi par [Gt308] bzhogs te/
rgya mdud1 la lha rnams bkod de/
/gdab na gang la bya ba de2 la btab na 'grub cing/
dpal dang gzi brjid rgyas par 'gyur ro/
/gzhan du na chos kyi dbyings la longs spyod rdzogs pa 'byung/
/longs spyod rdzogs par 'byung ba nyid/
/ting nge 'dzin phur pa yin pas/
/rang bzhin gyi dpal dang gzi brjid rgyas pa yin no/
/zhes tan tra du ma las 'byung ngo/
/thabs kyi zhags pa padmo 'phreng las/
rgyas pa'i phur pa'i le'u ste/
sum cu gcig pa'o//

1 On

the basis of the structure of the parallel passages relating to the phur pa for the other ritual actions as given in the previous
chapters, it might be supposed that an earlier version might have read rather: gser ram ser po'i shing rnams la/ /phur pa mgo bo
rgya mdud/ rtse zur bzhi du bzhogs te/ rgya mdud... The addition of "kyi steng dang ngos" might also be anticipated from the
chapters above, but they do not occur in the later chapter on the pacifying phur pa. | 2 Again, an addition of "gzugs sam mying rus"
might be anticipated from the chapters above, but they do not occur in the later chapter on the pacifying phur pa.

Chapter 32
[Ms omits] [Qt124a.7; Gt308.3; Nt219.4]
//da ni rgyas pa'i gtor ma bshad de/
gser phye 'bras chan ser po rnams/
/'o ma sbrang rtsi gtor ma ni/
/rgyas par 'gyur ba'i1 gtor ma ste/
/gtang na thams cad rgyas par 'gyur/2
zhes gsungs [Qt124b] te/
gser phye dang 'bras ser po'i chan la sogs pa/
'o ma dang sbrang rtsir sbyar te/
/btang na dpal dang gzi brjid rgyas shing 'phel bar 'gyur ro/
/gzhan du na chos kyi dbyings mtshan nyid med pa thams cad la khyab ste/
bdag nyid chen po dang tha mi dad pa rang bzhin gyis thams cad rgyas par 'gyur te/
de ltar rig pa'i ye shes ni/
byin cing gtong ba yin ces/
sgyu 'phrul dra ba las 'byung ngo/
/'phags pa thabs kyi zhags pa padmo'i 'phreng las/
/rgyas pa'i gtor ma'i le'u ste/
sum cu gnyis pa'o//

1 'gyur

ba'i: QtNt 'gyur pa'i | 2 Note that there is a significant contrast between the version of the fourth verse line given in most of
the root texts, and this version, which is largely shared by the Bhutanese NGB (see the TZ edition above)

Chapter 33
[Ms73v.5] [Gt308.6]
// da ni rgyas par1 sbyor pa'i2 las [Gt309] [6] bshad de /
/ lha mo rnams kyi dkyil 'khor du /
/ lha 'i dkyil 'kho[r]3 rdzogs par bsgom /
/ byang chub sems ni 'phro4 [74r.1] $/ / bsgoms na /
/ de ni5 rgyas pa'i las 'grub bo /
/ zhes gsungs te /
lha mo rnams dang /6
lha rnams 'dzum ba7 mdzad cing 8 [2] bde ba chen pos rol cing /9
long spyod10 rgyas par bsgoms na11 /
de bzhin du 'grub bo /
/ gzhan du na /12
ting nge 'dzin gyi dkyil 'khor [3] na13 /
lha dang lha mo 'i tshul du sbyor la ***** de ltar rig pa ni14 byang chub kyi sems ***** 'phro bas15 /
/long spyod16 chen po [4] rgyas pa yin zhes /
ri bo brtsegs ***** pa las 'byung ngo /
/ rgyas pa'i las b*****shad zin to /
/17 thabs kyi zhags [5] pa pad ma18 'phreng las /19
rgyas par20 'byor pa'i21 le'u ste22 sum cu gcig go23 //
:

1 par:

Gt pa'i | 2 pa'i: Gt ba'i | 3 'kho[r]: the final ra has been smudged out, but the traces make it certain. | 4 'phro: Gt spro | 5 ni: Gt
yis | 6 /: Gt omits | 7 rnams 'dzum ba: Gt sbyor ba chen po (either alternative is possible; perhaps the Dunhuang manuscript
reading might seem slightly less repetitive and obvious.) | 8 Gt inserts: / | 9 /: Gt omits | 10 long spyod: Gt longs sbyong (longs
spyod is surely intended) | 11 na: Gt nas | 12 /: Gt omits | 13 na: Gt ni | 14 Gt inserts: / | 15 'phro bas: Gt spros pas | 16 long spyod: Gt
longs spyod | 17 Gt inserts: 'phags pa | 18 pad ma: Gt padmo | 19 /: Gt omits | 20 par: Gt pa'i | 21 'byor pa'i: Gt sbyor ba'i | 22 Gt
inserts: / | 23 gcig go: Gt gsum pa'o

Chapter 34
[Ms74r.5] [Gt309.4]
// da ni zhi ba'i las bshad de / /
[6] zlum po 'i dkyil 'khor rtsibs brgyad par /
/ lha rnams 'khor bcas spyan drangs te /
/ dkar po 'i yon gyis mchod par bya / /
[74v.1] de nas las rnams brtsam par bya1
zhes gsungs te /
dkyil 'khor zlum por2 byas te /
/ lha rnams spyan drangs nas3 /
yo byed4 dkar po dang dbyangs la stsogs5 pas [2] mchod pa6 byas nas /
zhi ba'i las brtsam zhes bya ba'i don to /
/ gtI mug rdo rje 'khor bcas pas /
/ 'jig rten khams ni [3] rab 'byam7 kun /
/ yud tsam nyid ***** la zhi mdzad de8 /
/ zhi bar grub la phyag ***** 'tshal lo /
/ zhes [Gt310] bstod de /
[4] chas dkar po rnams lnga9 dang ldan ***** par byas te10 las bya 'o /
/ lha rnams **** 'khor bcas thams cad kyi11 / /
[5] 'khrug pa thams cad zhi mdzad par /
/ bsgoms na thams cad yud tsam gyis /
/ zhi bar 'gyur12 te the tshom myed13 / /
[6] zhes14 gsungs te /
lha dang lha mo 'khor dang bcas15 'khrugs pa la stsogs16 te /
/gnod sdig thams cad zhi bar mdzad par17 [75r.1] $/ / bsgoms na /
de bzhin du 'gyur ro /18
/gzhan du na chos kyi dbyings stong pa nyid kyi 19 ting nge 'dzin gyis /
/thams cad 20 [2] khyab pas thams cad 21 zhi bar 'gyur ro /
/ zhes 22 las 23 'phreng ba las 'byung ngo /
/24 thabs kyi zhags pa pad ma25 'phreng26 las /
[3] zhI ba'i las kyi le'u ste27 sum cu gnyi*****s so28 // :

1 de

nas las rnams brtsam par bya: this is written above the line, as a correction | 2 por: Gt par | 3 nas: Gt te | 4 yo byed: Gt yo byad
Gt sogs | 6 pa: Gt par | 7 'byam: Gt 'byams | 8 de: Gt do | 9 dkar po rnams lnga: Gt gos rnam pa gsum | 10 te: Gt de | 11 kyi:
Gt kyis | 12 'gyur: Gt gyur | 13 myed: Gt med | 14 zhes: Gt ces | 15 dang bcas: Gt bcas pa | 16 stsogs: Gt sogs | 17 mdzad par: Gt
omits | 18 de bzhin du 'gyur ro /: Gt omits | 19 Gt inserts: ye shes zhi ba'i | 20 Gt inserts: du | 21 Gt inserts: du | 22 zhes: Gt gzhan
23 Gt inserts: kyi | 24 Gt inserts: 'phags pa | 25 pad ma: Gt padmo'i | 26 'phreng: Gt 'phreng ba | 27 Gt inserts: / | 28 gnyis so: Gt bzhi
ba'o

5 stsogs:

Chapter 35
[Ms75r.3] [Gt310.4]
// da nI zhI ba'i hom bshad de / /
[4] hom khung zlum po gdengs pa la1 / ****
/ rdo rje pad mas brgyan2 par bya3 /
/ phur ***** pa dkar po drug btab ste / /
[5] spyan drangs dngul phye la stsogs4 mchod /
/ zhes5 gsungs te /
hom khung gong ma lta bur /6
lha gong ma 'khor dang bcas pa spyan [6] drangs te /
/ yo byed7 dkar po8 bsreg cing mchod de /
'khrug pa la stsogs9 pa thams cad zhi bar bsgoms na /
/ de bzhin du [75v.1] 'grub bo /10
gzhan du na /11
chos nyid kyi ting nge 'dzin gyis /
rnam par rtog pa la stsogs pa12 'khrug pa thams cad rang bzhin gyis [2] zhi zhing /
lha rnams kyang mchod par 'gyur ro13 /
[Gt311] tan tra glang po chur 'jug14 las 'byung ngo /
/ 15 thabs kyi zhags pa pad ma16 'phreng las17 [3] zhI ba'I hom gyI le'u ste18 sum cu *****gsum mo19 //
:

1 la:

Gt las | 2 pad mas brgyan: Gt padma rgyas (the Bhutanese texts support Gt here, but the other root texts support the Dunhuang
manuscript, and brgyan is surely the correct reading) | 3 bya: Gt bya'o | 4 stsogs: Gt sogs | 5 zhes: Gt ces | 6 gong ma lta bur: Gt
zlum po gdengs pa las/ | 7 yo byed: Gt yo byad | 8 po: Gt pos | 9 stsogs: Gt sogs | 10 de bzhin du 'grub bo/: Gt omits | 11 /: Gt
omits | 12 stsogs pa: Gt sogs pa/ | 13 'gyur ro: Gt 'gyuro zhes | 14 'jug: Gt zhugs pa | 15 Gt inserts: 'phags pa | 16 pad ma: Gt padmo
17 Gt inserts: / | 18 Gt inserts: / | 19 sum cu gsum mo: Gt sum cu lnga ba'o

Chapter 36
[Ms75v.3] [Gt311.1]
// da ni zhi ba'i phur pa'i las bshad de /
[4] phur pa kha dog dkar po la /1 ****
/ mgo bo rgya mdud rtse zlum por2 / *****
/ lha rnams bsgoms te gdab par bya /
[5] 'dI ni zhi ba'i las la shis /
/ zhes gsungs te /
dngul3 lam shing kha dog dkar po la4 /
phur pa mgo bo rgya mdud rtse5 zlum por [6] bzhogs6 la /
rgya mdud la lha rnams bkod de /
zhi bar bsgoms nas7 btab na /
lha yang8 zhi bar 'gyur ro /
gzhan du na chos [76r.1] $/ /nyid ma nor par rig pa nyid ting nge 'dzIn gyi phur pa yin te /
/ thams cad rang bzhin gyis zhi bar 'gyur zhes /9
rtse gcig [2] bsdus pa las 'byung ngo /
/ 10 thabs kyi zhags pa pad ma11 'phreng12 las /
/ zhi ba'i phur pa'i le'u ste /
/sum cu [3] bzhI13 pa 'o // :

phur pa kha dog dkar po la /: QtGtNt omit this tshig rkang (the Tshal pa Kanjur and Bhutanese texts support the Dunhuang
manuscript reading here, but the other root texts all give: dngul lam dkar po'i shing rnams la/; see comment in the TZ edition,
p.211.) | 2 por: Gt po | 3 dngul: Gt rngul | 4 kha dog dkar po la: Gt mdog dkar po las | 5 mgo bo rgya mdud rtse: Gt rgya mdud
6 bzhogs: Gt gzhogs | 7 nas: Gt te | 8 yang: Gt gzhan | 9 /: Gt omits | 10 Gt inserts: 'phags pa | 11 pad ma: Gt padmo'i | 12 'phreng: Gt
'phreng ba | 13 bzhI: Gt drug

Chapter 37
[Ms76r.3] [Gt311.4]
//da ni zhi ba'i gtor ma bshad de / *****
/dngul dang1 'bras can2 dkar po rnams/
[4] 'o mas3 sbyar pa'i4 gtor ma ni / *****
/zhi bar gyur pa'i5 gtor ma ste / *****
/mkhas pas zhi bar bsgoms zhing [5] gtor /
/zhes6 gsungs te /
dngul dang 'bras chan dkar po7 la stsogs pa8 'o ma dang sbyar te /
/gtor na /9
lha man cad10 zhi11 bar [6] 'gyur ro /
/gzhan du na12 chos thams cad kyi chos nyid ngang nyid13 kyis /14
zhi ba yin pas15 de ltar rig pa'i ye shes [Gt312] kyis / /
[76v.1] dngos po thams cad las16 khyab pas zhi bar 'gyur ro /
zhes sgyu 'phrul dra ba las bshad do /
/ 17 thabs kyi zhags pa pad ma18 [2] 'phreng las /
zhI ba'i gtor ma'I le'u ste19 sum cu lnga 'o20// :

dang: Gt lam (Gt's reading, shared with the Bhutanese texts, is less appropriate here; the Dunhuang reading is shared with the
Tshal pa Kanjur texts, but the other root texts give dngul phye) | 2 can: Gt chan | 3 'o mas: Gt 'o mar (Gt's reading is shared with
the Tshal pa Kanjur and Bhutanese texts, but the Dunhuang reading is supported by the other root texts) | 4 pa'i: Gt ba'i | 5 gyur
pa'i: Gt 'gyur ba'i | 6 zhes: Gt ces | 7 'bras chan dkar po: Gt 'bru dkar po'i chan | 8 stsogs pa: Gt sogs pa/ | 9 /: Gt omits | 10 man cad:
Gt man chad | 11 zhi: QtGt 'chi (clearly, the Dunhuang manuscript reading is more appropriate here) | 12 na: Gt omits | 13 ngang
nyid: Gt omits | 14 kyis /: Gt kyi | 15 Gt inserts: / | 16 las: Gt la | 17 Gt inserts: 'phags pa | 18 pad ma: Gt padmo | 19 Gt inserts: /
20 sum cu lnga 'o: Gt sum cu bdun pa'o

Chapter 38
[Ms76v.2] [Gt312.2]
//da ni zhI bar1 'byor pa'i le'uas2 bshad de / /
[3] lha mo rnams kyi3 dkyil 'khor du / *****
/lha 'i grub pa bsgom par bya /
/byin ***** gyis4 thams cad zhI bsgoms na5 //
[4] /lha rnams kyang ni6 zhi bar 'gyur7 zhes ***** gsungs te /
lha dang lha mo 'i8 grub pa'I ***** byin rlabs kyis9 /
thams cad zhi [5] bar bsgoms na /
lha nyid kyang zhi ba yin na10 gzhan lta ci smos / 11
/gzhan du na chos kyi dbyings /
lha dang lha mo12 lta bu yang13 yong [6] gis14 zhI ba yin la /
de ltar rig pa'i ye shes lha 'i rang bzhin lta bu yang /15
rang bzhin gis16 zhi ba yin te /
de 'i byin rlabs17 kyis [77r.1] $/ /thams cad du khyab la18 /
rang bzhin du19 zhi ba yin bas /
thams cad 20 zhi bar 'gyur ro /
zhes rdo rje bkod pa'i tan tra las [2] 'byung ngo /
/21 thabs kyi zhags pa pad ma22 'phreng las /
/zhi bar23 'byor pa'i24 le'u ste /
sum cu drug pa [3] pa 'o25// :

bar: Gt ba'i | 2 'byor pa'i le'uas: it appears that le'u has been amended to las. The 'greng bu has been smudged and partially
erased; the zhabs kyu is not fully wiped out but also smudged and it seems almost certain that this is the intention; the sa is written
in thick ink over a probable original smudged a chung; Gt sbyor ba'i las | 3 kyi: Gt kyis | 4 byin gyis: Gt byin rlabs (none of the
root texts support Gt here) | 5 bsgoms na: Gt sgoms nas | 6 kyang ni: Gt kun kyang (the Bhutanese texts support the Dunhuang
manuscript here, but the other root texts support Gt) | 7 Gt inserts: /. In the Dunhuang manuscript, there is some smudging over
some of this verse, but no real evidence of the usual highlighting wash; it is possible that it had been applied but has disappeared.
8 'i: Gt omits | 9 byin rlabs kyis: Gt dkyil 'khor las | 10 Gt inserts: / | 11 Gt inserts: zhi bar 'gyur ro/ | 12 / lha dang lha mo: Gt kyi ye
shes lha mo'i rang bzhin (Gt's reading seems a little more appropriate here) | 13 Gt inserts: / | 14 yong gis: Gt yongs kyis | 15 /: Gt
omits | 16 gis: Gt gyis | 17 byin rlabs: Gt byin brlabs | 18 la: Gt pa yang | 19 du: Gt gyis | 20 Gt inserts: ni | 21 Gt inserts: 'phags pa
22 pad ma: Gt padmo'i | 23 bar: Gt ba | 24 'byor pa'i: Gt sbyor ba'i; the Tshal pa Kanjur and Bhutanese texts also insert las kyi here,
but the addition is not supported by the other root texts. | 25 sum cu drug pa pa 'o: Gt sum cu rtsa brgyad pa'o

Chapter 39
[Ms77r.3] [Gt312.6]
//da ni las sdud pa1 bshad de /
/mtshan **** nyid myed2 pa'i phyag rgya dang / /
[4] de las dngos byung3 phyag rgya rnams / ****
/nges pa myed4 par [Gt 313] dpa' bos bshad /
/mkhas pas ci dga'r5 bsgyur6 na [5] 'grub /
/zhes7 gsungs te /
gong8 du bshad pa thams cad mtshan nyid myed9 pa'i phyag rgya ste /
/chos kyi sku las10 dngos por snang ba dang /
mtshan mar bsgoms pa thams cad 'byung bas /
chos kyi skur 'dus te /
/chos kyi sku ma nor bar rig na /
rnal 'byor pas11 bsgom12 [77v.1] pa dang /
phyag rgya dang mtshan ma thams cad ci13 ltar bde ba dang /
ci rigs su sbyar zhing14 las rnams15 bzhI brtsams na /
myi16 'grub pa myed17 [2] do zhes18 rnam par snang mdzad19 nyid [k]yis20 bshad de21 /
tan tra thabs kyi zhags pa dang /
dpal 'phreng dkar po dang /
sgyu 'phrul dra ba la stsogs [3] te22 tan tra 23 sde lnga las bshad de24/ ****
/ 25 thabs kyi zhags pa pad ma26 'phreng 27 ***** las /
las bsdus28 pa'i le'u ste /
[4] sum cu bdun no29 // :

1 sdud

pa: Gt bsdu ba | 2 myed: Gt med | 3 byung: Gt 'byung | 4 myed: Gt med | 5 dga'r: 'a subscribed, perhaps intended as a
correction, although the final ra has not been deleted; Gt dga' | 6 bsgyur: Gt sgyur | 7 zhes: Gt ces | 8 gong: Gt gang | 9 myed: Gt
med | 10 las: Gt la | 11 pas: Gt bas | 12 bsgom: Gt sgoms | 13 ci: Gt ji | 14 Gt inserts: / | 15 rnams: Gt rnam | 16 myi: Gt mi | 17 myed:
Gt med | 18 Gt inserts: / | 19 rnam par snang mdzad: Gt bai ro tsa ni | 20 [k]yis: kyis is probably intended although it appears like
gyis; generally, in this manuscript, the ga is open at the top while here it is closed, so it is probable that the stroke apparently
making ka into ga was accidental; Gt kyis | 21 de: Gt do | 22 stsogs te: Gt sogs te/ | 23 Gt inserts: du ma | 24 de: Gt do | 25 Gt inserts:
phags pa | 26 pad ma: Gt padmo'i | 27 Gt inserts: ba | 28 bsdus: Gt btus | 29 sum cu bdun no: Gt sum cu dgu pa'o

Chapter 40
[Ms77v.4] [Gt313.5]
//da ni las bsgrub pa'i1 phan yon bsha*****d de /
/drag pos thams cad 'dul [5] bar 'gyur /
/dbang gis 'byor pa mchog kyang 'grub /
/rgyas pas2 gzi brjid rnam3 par 'phel /
/zhi bas bde ba4 mchog kyang [6] 'grub /
/ zhes5 bshad6 de7 /
drag po spyad pas ni /8
'di dang gzhan du /9
lha dang bdud la stsogs10 pa thams cad thul par11 'gyur ro /
[78r.1] $/ /dbang gyi12 las byas pas ni /13
'di dang gzhan du long spyod14 mchog dang / [Gt314]
'khor phun sum tshogs par 'gyur ro / /
[2] rgyas pa'i las spyad15 pas ni 'di dang gzhan du dpal dang gzi brjid 'phel bar 'gyur ro /
/ zhi ba'i las spyad pas ni /
'di dang gzhan [3] du dpal dang gzi brjid 'phel bar ***** 'gyur ro /
/zhi ba'i las spyad pas ***** ni / 16
'di dang gzhan du bde ba'i mchog [4] thob par 'gyur te17 /
/'di dag ni ***** bsod nams kyi tshogs chen po yin ***** te /
mtshan nyid myed18 pa las [5] myi gzhan bar19 rig pa ni /
/ye shes kyi tshogs chen po yin bas20 /
phyag rgya chen po 'i don rig par byas la /
las rnam bzhi rgyun [6] du21 brtsam par bya'o zhes /22
ma ha yo ga 'i tan tra du ma las bshad do /
/gong nas 'byung ba'i lha gcig23 bsgrubs24 na yang /
dngos grub [78v.1] grangs myed de /25
bsam gis myi26 khyab bo27 zhes 28 du ma las29 bshad do /
/ 30 thabs kyi zhags pa pad ma31 'phreng 32 las /
las bsgrub [2] pa'I33 phan yon gyi le'u ste/
sum cu brgyad pa 'o34 // :

bsgrub pa'i: Gt drag po'i (Gt seems inappropriate here, probably picked up in error from the first verse line) | 2 pas: Gt par
Gt rtsam (Gt's apparent reading may simply be due to a poor print. Qt clearly gives rnam) | 4 ba: Gt ba'i | 5 zhes: Gt ces
6 bshad: the sha letter has been smudged but remains clearly legible. | 7 de: Gt do | 8 /: Gt omits | 9 /: Gt omits | 10 stsogs: Gt sogs
11 par: Gt bar | 12 gyi: Gt gi | 13 /: Gt omits | 14 long spyod: Gt longs spyod | 15 spyad: Gt byas | 16 'di dang gzhan du dpal dang gzi
brjid 'phel bar 'gyur ro / / zhi ba'i las spyad pas ni /: apparent dittography; Gt omits | 17 te: Gt ro | 18 myed: Gt med | 19 myi
gzhan bar: Gt mi gnas par | 20 bas: Gt pas | 21 du: Gt tu | 22 zhes /: Gt / zhes | 23 gcig: Gt cig la | 24 bsgrubs: Gt bsgrub | 25 myed de
/: Gt med pa de | 26 gis myi: Gt gyis mi | 27 bo: Gt pa'o/ | 28 Gt inserts: kyang | 29 Gt inserts dittographically: du ma las (Qt does
not) | 30 Gt inserts: 'phags pa | 31 pad ma: Gt padmo | 32 Gt inserts: ba | 33 pa'I: Gt pa | 34 sum cu brgyad pa 'o: Gt bzhi bcu pa'o
3 rnam:

Chapter 41
[Ms78v.2] [Gt314.4]
//dkyil 'khor 'di 'i mtshan nyid ni /
dbang chen bsdus [3] pa'i dkyil 'khor ste1 /
de la dkyil **** 'khor gzhi zlum por2 bya ba ni /
/chos ***** kyi dbyings spros pa myed pa'i don to/
[4] 'khor lo 'i rtsibs zur can du ***** bya zhing /
gzhi zlum po 'i nang na gnas ***** pa ni /
chos kyi dbyings spros3 pa [5] myed4 bzhin du /5
byin rlabs6 kyis /
lha dang lha mo dang /
khro bo dang khro mo dang pho nya la stsogs7 pa ngang8 nyid kyis 'phro ba'i don to /
rnam [6] [Gt315] gcig du9 na /10
chos kyi dbyings rnam par rtog pa myed11 bzhin du /
thams cad mkhyen pa'i ye shes ci yang mkhyen pa /12
[beneath the line] sdo sde las rnam rtog myi mnga' cir yang sa ler mkhyen ces bya ba da[ng m?]thun

lhun gyis [79r.1] $/ /grub pa la bya ba'I don to /

/mtha'13 grub14 bzhir bya ba ni15 phyogs bcu16 snod dang bcud kyi 'jig rten thams cad la dbang sgyur bas/
[2] bde ba chen po17 rgyas pa'I don 18/
/dkyil nas lha mo lnga19 phyir lta ba 20/
/lha mo ni dbying21 nyid kyi22 nye ba'i23 byin rlabs te /
/sems [3] dpa's24 dbyings nyid dang nye ba25 mthong ***** ba'i don 26/
sems dpa' lnga nang 27 lta ba ni ****/
de dbyings28 nyid nye ba la lta ba'I don29 [4] sems dpa' gzhan phyir lta ba nI30 ***** dbyings31 nyid thams
cad du32 sa rdo lta bu33 ***** bems po 34 ma yin gyi / /
[5] mkhyen pa'I ye shes dang ldan zhing /
sems dpa' nyid kyang dbyings las myi35 gzhan pa'i tha tshig go /
sems ma gzhan nang [6] du lta ba ni /
mkhyen pa'I ye shes 36 dngos po la stsogs37 pa thams cad du /
/38rnam par rtog pa ma yin gyi /
dbyings nyid 39 ldan [79v.1] zhing /
/ dbyings nyid kyang40 mkhyen pa'I ye shes las myi gzhan41 pa'i tha tshig go /
/rnam gcig du42 sems dpa' phyir lta ba ni43 [2] snod dang bcud kyi 'jig rten thams cad dbang du bsgyur44 ba'i
don 45/
sems ma nang du lta ba ni46 snod dang bcud kyi 'jig rten thams [3] cad dbang du 'dus pa'i don 47/ *****
/sems dpa' dang [Gt316] sems ma phyi nang ***** du gcig la gcig gzigs [4] pa ni48 mnyes par byed pa'I
tha tshig **** go /

1 ste:

Gt te | 2 por: Gt po | 3 spros: Gt sbros | 4 myed: Gt med | 5 /: Gt omits | 6 byin rlabs: Gt byin gyis brlabs | 7 stsogs: Gt sogs
ngang: Gt dang (the Tshal pa Kanjur texts, which include this entire commentarial chapter in full, are divided here; either
reading could fit) | 9 du: Gt tu | 10 /: Gt omits | 11 myed: Gt med | 12 /: Gt omits | 13 mtha': 'a subscribed; Gt mthar | 14 grub: Gt gru
(Gt's reading is supported by the root texts and no doubt correct here) | 15 Gt inserts: / | 16 bcu: Gt bcu'i | 17 po: Gt pos | 18 Gt
inserts: to | 19 lnga: Gt kha | 20 Gt inserts: ni | 21 dbying: Gt dbyings | 22 kyi: Gt omits | 23 ba'i: Gt bar | 24 dpa's: Gt dpa' | 25 ba: Gt
ba'i | 26 Gt inserts: to | 27 Gt inserts: du (Gt's reading is supported by the root texts) | 28 de dbyings: Gt chos (the Tshal pa Kanjur
texts support the Dunhuang manuscript here) | 29 Gt inserts: to/ | 30 Gt inserts: / | 31 dbyings: Gt chos (again, the Tshal pa Kanjur
texts support the Dunhuang manuscript here) | 32 Gt inserts: / | 33 bu: Gt bur | 34 Gt inserts: ni | 35 myi: Gt mi | 36 Gt inserts: dang
(the Tshal pa Kanjur texts support the Dunhuang manuscript here) | 37 stsogs: Gt sogs | 38 du / /: Gt omits | 39 Gt inserts: dang
40 kyang: Gt omits | 41 myi gzhan: Gt mi zhan | 42 gcig du: Gt cig tu na | 43 Gt inserts: / | 44 bsgyur: Gt sgyur | 45 Gt inserts: to
46 Gt inserts: / | 47 Gt inserts: to | 48 Gt inserts: /

344

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/khro bo dang khro mo ril phyir ***** lta ba ni /

phyi ma'i49 dngos po dang /
[5] /50 rgol ba la stsogs51 pa 'joms shing 'dul ba'i don 52/
de la dbyings ni dper 53 nyi ma'i dkyil 'khor lta bu 'o/
/mkhyen [6] pa'I ye shes ni /54
nyi ma'i 'od snang zhing gsal ba lta bu 'o55 /
de 'i byin rlabs56 ni dro zhing tsha ba57 lta bu 'o/
/phyi 'i dngos po [80r.1] $/ /la stsogs58 pa 'joms shing 'dul ba ni59 rtsi shing skam pa dang60 tshig61 pa lta bu
ste62 /
/ lha mo dang sems dpa'63 ni64 chos [2] kyi dbyings nyid rig pa'I don dbyings65 la snang ba 'o /
/ lha dang sems dpa' ni66 ngang67 gyis /68
dngos po dang69 rgol ba'i rang bzhin lhun gyis [3] 'joms shing /
/ dbyings su sbyor ***** bar 'dul ba ni/
khro bo dang khro mo las ***** rnam bzhI rang bzhin lhun gyis70 [4] mdzad cing /
/thugs rjes 'jig **** rten skyong ba'I don to /
/ 'jig rten ***** skyong ba la stsogs71 pa ni / /
[5] byin kyi rlabs72 las73 pho nya ltar74 gnas pa yin /
/ rnam gcig du75 'jig rten gyi lha76 klu77 mthu bo che 78 'khor du 'dus par79 bstan pa'i don80/
/bkra shis [6] brgyad ni thams cad la dbang sgyur zhing /81
'khor ba la 82 spyod kyang83 nyon mongs pa'i84 'ching bar myi 'gyur ba dang /
/mya ngan [Gt317] las 'das [80v.1] pa la spyod kyang /
/ nyan thos kyI85 zhi ba phyogs gcig par myi86 'gyur te /
/gang gis kyang zil gyis myi87 non pa'i88 bkra shIs so / /
[2] de la dpal 89 be'u ni90 mkhyen pa'i ye shes dang ldan pa'i91 bkra shis so /
/'khor lo ni92 thams cad93 la dbang sgyur pa94 ba dang /
chos kyI 'khor [3] lo95 bskor ba'I96 bkra shis so / *****
/rIn po che ni97 zad myi98 shes pa'i gter99 dang ***** ldan pa'i100 bkra shis so / /
[4] pad ma ni101 'khor bas ma zin pa'i bkra ***** shis so /
/gdugs ni102 nyon mongs 103 myi ***** gdungs pa'i104 bkra shis so / /
[5] bum pa ni105 bdud rtsi 'i ro106 mchog dang ldan ba'i107 bkra shis so /
/dung ni108 rgol ba thams cad spa 'gong bar109 sgrogs pa'i bkra shis so /
/nya ni110 [6] lam rgyud lnga 'i sems can dang 'jig rten thams cad kyi rgyun/

49 phyi ma'i: Gt phyi'i (the Tshal pa Kanjur texts support Gt here) | 50 / /: Gt omits | 51 stsogs: Gt sogs | 52 Gt inserts: to | 53 Gt
inserts: na | 54 /: Gt omits | 55 bu 'o: Gt bu'i | 56 rlabs: Gt brlabs | 57 tsha ba: the Tshal pa Kanjur texts give 'jam pa (the Dunhuang
and Tenjur reading seems more appropriate, especially given the following example) | 58 stsogs: Gt sogs | 59 Gt inserts: / | 60 skam
pa dang: Gt skyes shing (Gt is presumably in error here) | 61 tshig: Gt 'tshig | 62 bu ste: Gt bu'o | 63 sems dpa': the Tshal pa Kanjur
texts give, sems ma (sems ma would seem more appropriate here.) | 64 Gt inserts: / | 65 dbyings: the Tshal pa Kanjur texts give,
sems | 66 Gt (and Qt127b.2-3) inserts: /dbyings la snang ba'i don mthong ba'o/ /dbyings nyid la snang ba dang mthong ba'i
byin rlabs ni/; note that the Tshal pa Kanjur TZ editions which include this section also include a version of these lines. It is most
likely that the Dunhuang manuscript has an omission here (perhaps an eyeskip from "ni"). | 67 ngang: QtGt rang bzhin (the Tshal
pa Kanjur texts support the Dunhuang manuscript here) | 68 gyis /: Gt gyi | 69 Gt inserts: / | 70 lhun gyis: the Tshal pa Kanjur texts
give, du | 71 stsogs: Gt sogs | 72 byin kyi rlabs: Gt byin rlabs | 73 las: Gt omits | 74 ltar: Gt lta bu | 75 du: Gt tu | 76 lha: inserted
beneath the line with a small ink stroke above the line marking the position | 77 klu: Gt omits (the Tshal pa Kanjur texts also
omit) | 78 Gt inserts: dang mthu mo che (this seems most probably a kind of dittography) | 79 par: Gt pa'i | 80 bstan pa'i don: Gt
rtags so | 81 /: Gt omits | 82 Gt inserts: longs | 83 Gt inserts: / | 84 pa'i: Gt pas | 85 kyI: Gt omits | 86 myi: Gt mi | 87 myi: Gt mi
88 pa'i: Gt pas (the Tshal pa Kanjur texts give, pa'i phyir) | 89 Gt inserts: gyi | 90 Gt inserts: / | 91 pa'i: Gt pas | 92 Gt inserts: /
93 thams cad: Gt thabs | 94 pa: Gt omits | 95 lo: Gt la | 96 ba'I: Gt bar | 97 Gt inserts: / | 98 myi: Gt mi | 99 pa'i gter: the Tshal pa
Kanjur texts give, pa | 100 pa'i: Gt pas | 101 Gt inserts: / | 102 Gt inserts: / | 103 Gt inserts: pas | 104 myi gdungs pa'i: Gt mi gdugs pas
105 Gt inserts: / | 106 bdud rtsi 'i ro: Gt ro (the Tshal pa Kanjur texts give, bdud rtsi) | 107 ba'i: Gt pas | 108 Gt inserts: / | 109 spa 'gong
bar: Gt skong bas sgra (the Tshal pa Kanjur texts have small variants, but essentially support spa gong sgra) | 110 Gt inserts: /

Chapter 41

345

/111thugs rjes chen pos sim par byed pa'i112 bkra shis so / /
[81r.1] $/ /gling113 bzhi ni rigs lnga 'i byin rlabs114 yin pa 'am /
/lam rgyud115 lnga 'i sems can thams cad116 la/
/ de bzhin gshegs pa'i117 [2] rang bzhin yod par bstan pa 'am /
/thugs rjes118 de dag gi don mdzad pa 'I tha tshig go /
/ rnam 119 gcig du na120 thams cad dbang du 'dus [3] pa'I don no121 /
/ de la gtso bo dang122 gtso ***** mo 'i123 'khor lnga lnga dang bcas pa ni124 ci yang **** myi125 mdzad do /
/ sems dpa' dang [4] sems ma nyi shu ni126 kun la bu gcig **** bzhin no /
/de ltar byams pa'i shugs ***** kyis127 khro bo dang khro mo 'khor dang [5] bcas pa ni /
phyi nang thams cad kyi dngos po la /
/rngon pas [Gt318] ri dags mthong ba bzhin no128 /
/de ltar zhe sdang gyi129 shugs kyis130 thams cad [6] dbyings nyid la sbyor ro /
/ 'byor na 'khor ba'i chos la spyod131 kyang132 longs spyod chen po ste/
nyon mongs pa133 'ching bar myi134 'gyur bas / /
[81v.1] de ltar bsgrub135 pa ni 'dod chags so /
/chos nyid136 kyi dbyings ngang gis gsal ba las137 phrag dog go /
/chos nyid spros pa myed138 pas [2] gti mug go /
/chos nyid de lta bus khams gsum las rgyal bas nga rgyal lo /
/de la gtso bo139 140 lnga 'i byin rlabs nI141 thams [3] cad la142 khyab mod kyi /
/ 143dbus nas144 ***** mdun nas145 mthong ka146 yan cad147 / *****
/148rgyab mjug do yan chad do / /
[4] myi149 'khrugs pa'i 'khor lnga 'i byin ***** brlabs150 ni /
/ phyag g.yas dang rtsib151 ***** log152 g.yas so /
/rin cen 'byung [5] nas153 kyi 'khor lnga 'i byin rlabs ni /
/ phyag g.yon dang rtsib154 logs g.yon no /
/'od dpag myed155 kyi 'khor lnga 'i byin rlabs ni /
[6] zhabs g.yon nas rked pa g.yon pa156 man cad157 do /
/don yod par158 grub pa'i159 byin rlabs ni /
/160zhabs g.yon161 pa nas /
rked pa g.yon [82r.1] $/ / pa162 man cad163 do /
/khro bo dang khro mo thams cad ni164 rigs so so 'i byin rlabs165 bzhin no /
111 'jig

rten thams cad kyi rgyun/ /: Gt /ngan song kun las sgrol ba'i | 112 chen pos sim par byed pa'i: Gt thabs mkhas pa'i (in
this and the above line, the Tenjur has a rather different reading from the Dunhuang manuscript, equally plausible, although
perhaps the imagery drawn on in the Dunhuang manuscript's use of the word, rgyun (stream) may seem slightly more appropriate
for the symbol of the fish. The Tshal pa Kanjur reading of these two lines is even more different, not even sharing the "/thugs rjes"
element, nor the expected ending of bkra shis so: nya ni lam rgyud lnga'i sems can gyi don mdzad kyang nyam nga ba med
pa'o/) | 113 gling: the Tshal pa Kanjur texts give, klong | 114 byin rlabs: Gt byin brlabs | 115 rgyud: Gt omits | 116 thams cad: Gt
omits (the Tshal pa Kanjur texts also omit) | 117 pa'i: Gt par | 118 thugs rjes: Gt thugs rje | 119 Gt inserts: pa | 120 du na: Gt tu
121 no: Gt to | 122 gtso bo dang: the Tshal pa Kanjur texts omit | 123 mo 'i: Gt mo | 124 Gt inserts: / | 125 myi: Gt mi | 126 Gt inserts: /
127 Gt inserts: / | 128 no: Gt du ni | 129 gyi: Gt gi | 130 Gt inserts: / | 131 spyod: Gt spyad | 132 Gt inserts: / | 133 pa: Gt pas | 134 myi:
Gt mi | 135 bsgrub: Gt grub | 136 chos nyid: Gt unclear (the Tshal pa Kanjur texts give, chos) | 137 gsal ba las: Gt gsang bas (Gt's
reading seems preferable here and is shared by the Tshal pa Kanjur texts) | 138 myed: Gt med | 139 bo: Gt bo'i | 140 Gt (and the
Tshal pa Kanjur texts) insert: 'khor (presumably, an omission in the Dunhuang manuscript) | 141 byin rlabs nI: Gt byin brlabs ni/
142 la: Gt omits | 143 Gt inserts: nye ba ni (the Tshal pa Kanjur texts insert: nye bar ni) | 144 dbus nas: the Tshal pa Kanjur texts
give, dbu dang/ (see TZ edition, p.222) | 145 mdun nas: Gt 'dun dang/ | 146 mthong ka: Gt thod ka | 147 yan cad: Gt yan chad | 148 /
/: Gt omits | 149 myi: Gt mi | 150 byin brlabs: Gt byin rlabs | 151 rtsib: Gt rtsibs | 152 log: Gt logs | 153 rin cen 'byung nas: Gt rin chen
'byung gnas | 154 rtsib: Gt rtsibs | 155 myed: Gt med | 156 g.yon pa: Gt omits (Qt also omits) | 157 man cad: Gt man chad | 158 don
yod par: Gt don yong | 159 Note that the Tshal pa Kanjur root texts here give, don yod grub pa 'khor lnga'i, which is consistent
with the consecrations for the other deities. | 160 / /: Gt omits | 161 g.yon: Gt g.yas (the Tshal pa Kanjur root texts, and Qt, agree
with Gt here, and g.yas is surely more likely) | 162 g.yon pa: Gt g.yas (the Tshal pa Kanjur root texts agree with the Dunhuang
Manuscript here, but Gt's reading, also given in Qt, would seem more appropriate and fits with the associations given in Chapter
7) | 163 man cad: Gt man chad | 164 Gt inserts: / | 165 byin rlabs: Gt byin brlabs

346

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/de ltar 'khor lnga po gcig la yang [2] lnga lngar166 'gyur te /
/sra ba 'I bya ba byed pa ni167 rdo rje 'i rigs so /
/dro ba'i bya 168 byed pa thams cad ni169 rin po che 'i rigs so /
/gsher ba'i bya [3] ba byed pa thams cad ni170 pad ma'I171 [Gt319] rigs ***** so/
/bskyod pa'i bya ba byed pa thams cad ****** ni172 las kyi rigs so /
/de rnams [4] cha173 mnyam pa las [s]pyir grub pa ni174 / *****
dngos su bzhugs175 pa de bzhin gshegs ***** pa'i rigs so /
/thams cad bdag [5] nyid ma 'i176 'khor lnga 'i byin rlabs ni /
dbus su gyur pa177 thams cad do178 /
/sra ba'i rang bzhin myed179 pa ni /
yangs pa'i spyan gyi180 'khor lnga 'i [6] byin rlabs181 so/
/dro ba'i rang bzhin myed182 pa ni nam ka 'i183 'khor lnga 'i byin rlabs184 so /
/ gsher ba'i rang bzhin myed185 pa ni / /
[82v.1] gos dkar mo 'i186 'khor lnga 'i byin rlabs187 so /
/ bskyod pa'i rang bzhin myed188 pa ni /
dam tshig189 sgrol ma'i 'khor lnga 'i byin rlabs190 so/
[2] don du na191 bya ba byed nus pa ni sems dpa' 'o/
/de nyid192 rang bzhin myed193 pa ni194 sems ma 'o/
/de ltar rig pa195 ni khro bo dang khro mo 'o / /
[3] de la bsdu zhing spro ba 196 thams cad197 kyang198 ***** de ltar bya 'o/
/phyi nang gi chos thams **** cad la yang199 de ltar blta 'o200 / /
[Here, the Tenjur version of TZComm, and the South Central, Bathang NGB, Hemis, Bathang and Tawang
versions of TZ recognise a chapter ending, which has been lost in the Dunhuang manuscript and the Tshal pa
Kanjur texts. The Bhutanese texts jump from the opening of the chapter above up to here, resuming at the
beginning of the following chapter. Here, we present the Tenjur version.]
[Qt128b.7; Gt319.4; Nt227.2]
/'phags pa thabs kyi zhags pa padmo 'phreng las/
dbang chen bsdus pa dkyil 'khor gyi le'u ste/
bzhi bcu rtsa gcig pa'o//

166 lnga

lngar: Gt lngar lngar | 167 ni: Gt thams cad ni/ (the Tshal pa Kanjur texts support Gt here) | 168 Gt inserts: ba | 169 Gt
inserts: / | 170 Gt inserts: / | 171 pad ma'I: Gt padma'i | 172 Gt inserts: / | 173 cha: Gt omits | 174 [s]pyir grub pa ni: the sa head letter of
spyir is not clearly written, but seems to be intended; Gt omits (the Tshal pa Kanjur texts support the Dunhuang manuscript here)
175 bzhugs: Gt gzhag | 176 Gt inserts: / | 177 gyur pa: Gt 'gyur ba | 178 do: Gt de | 179 myed: Gt med | 180 gyi: Gt / | 181 byin rlabs: Gt
byin brlabs | 182 myed: Gt med | 183 nam ka 'i: Gt /ma ma ki/ (the Tshal pa Kanjur texts support ma ma ki, and this is the more
appropriate reading) | 184 byin rlabs: Gt byin brlabs | 185 myed: Gt med | 186 'i: Gt omits | 187 byin rlabs: Gt byin brlabs | 188 myed:
Gt med | 189 dam tshig: Gt omits (the Tshal pa Kanjur texts also omit, although dam tshig sgrol ma is clearly the full name of the
female consort of Amoghasiddhi given in Chapter 7 of TZComm above, Dunhuang manuscript f.28r, line 1 and 6) | 190 byin rlabs:
Gt byin brlabs | 191 na: Gt omits | 192 nyid: Gt bzhin du (the Tshal pa Kanjur texts support Gt here) | 193 myed: Gt med | 194 Gt
inserts: / | 195 pa: Gt pa'i | 196 Gt inserts: yang (the Tshal pa Kanjur texts support Gt here, although the Dunhuang manuscript gives
kyang below) | 197 cad seems to have an accidental indistinct na ro above it | 198 kyang: Gt omits (the Tshal pa Kanjur texts
support Gt here; see note 196 above) | 199 la yang: Gt kyang | 200 blta 'o: Gt brtag go (the Tshal pa Kanjur texts support the
Dunhuang manuscript here)

Chapter 42
[Ms82v.4] [Gt319.5]
de skad ces bcom ldan 'das1 dpal ***** rna par snang mdzad kyis bka'2 stsal pa ***** la3 4/
/rdo rje gzhu nu5 sde 6 'bum [5] phrag du ma yid rangs te7/
/ 'di skad ces bstod do /
/h/8
bskal pa'i9 mye10 ltar 'bar ba yi /
/'od gzer11 'du 'phro12 klong dkyil na/
[6] khro bo khro mo rnams bzhugs pa /
/sku [Gt320] ni cir yang 13 sgyur14 ba'i gzugs /
h dbu skra kham gnag15 ral pa can /
/gyen du 'khyil [83r.1] $/ /ba16 srid rtser 'bar /
/myi17 mnyam khro gnyer sdang myig18 ni /
/log par lta ba 'dul la brtson19 /
/h rdo rje mche ba20 [2] gtsigs pas ni /
/skye shi rtsa ba21 gcod par mdzad22 /
/sku gsung thugs kyi byin rlabs23 ni /
/phyag ca24 sna tshogs cir yang [3] snang /
/h thugs rje25 byin ***** rlabs26 sprin ltar 'khrigs27/ *****
/byams pa'i 'od gzer28 [4] glog ltar 'gyu29 /
**** /rngam pa'i sgra skad 'brug ltar sgrog******s / ***** /
[5] sprul pa'i thog thog30 char bzhin 'bebs /
/h sngun31 gyi thugs dam chen po ni /
32 /bdud dpung ma lus 'dul [6] mdzad pa33 /
/bgegs rnams thams cad 'joms mdzad cing /
/dam tshig can la dngos grub stsol/
[83v.1] /h thugs rje chen po 'i byin rlabs34 kyis /
/bdag la byin kyis brlab35 mdzad cing /
/bdag gyi36 lhag pa'i bsam pa [2] rnams/
/grub par byin gyis brlab du37 gsol /
/h 'khyil cing rgyan38 yang 'khyil pa ['aya?]ng39 /
/sems kyi [3] chos pas40 sku myi41 mnga' / ****
1 bcom

ldan 'das: Gt omits (the Tshal pa Kanjur texts agree with Gt's omission) | 2 bka': 'a subscribed | 3 pa la: Gt ba | 4 highlighting
is uncertain in this line, but it appears to have been lightly applied. | 5 gzhu nu: Gt gzhon nu'i (a spelling error or archaic spelling
in the Dunhuang manuscript) | 6 Note that the South Central NGB and the local Kanjurs all insert 'khor here | 7 yid rangs te: Gt yi
rang ste | 8 /: Gt omits | 9 pa'i: Gt pa | 10 mye: Gt me | 11 'od gzer: Gt 'od zer | 12 'phro: Gt 'phro'i | 13 Gt inserts: ma nges (none of the
root texts support Gt here) | 14 sgyur: Gt 'gyur | 15 gnag: Gt nag | 16 ba: Gt pas | 17 myi: Gt mi | 18 sdang myig: Gt sdang dmigs
19 la brtson: Gt ba'i mtshan (the root texts support Gt here) | 20 rdo rje mche ba: Gt mche ba zang yag (the root texts support Gt)
21 rtsa ba: Gt rtsad nas | 22 par mdzad: Gt pa'i tshul (the root texts support Gt here) | 23 byin rlabs: Gt byin brlabs | 24 phyag ca: Gt
phyag cha | 25 thugs rje: Gt thugs rje'i | 26 byin rlabs: Gt byin brlabs | 27 'khrigs: Gt gtibs | 28 'od gzer: Gt 'od zer | 29 ltar 'gyu: Gt
bzhin 'gyur | 30 thog thog: Gt thog chen (the root texts support the Dunhuang manuscript here) | 31 sngun: Gt sngon | 32 Gt inserts:
/klan ka chen po bdud la chol/ /rkyal ka chen po bgegs la mdzad/ (found in the root texts also; this seems to be a clear
omission in the Dunhuang manuscript, although the next two tshig rkang could be an alternative reading of them. However, it is
quite possible that these next two tshig rkang, found in both the Dunhuang manuscript and Gt (but not the root texts), were a
commentarial elaboration which were mistaken by the Dunhuang manuscript copyist as the lines of the verse, while the verse lines
were then omitted in error. Gt's inclusion of both sets of lines may be correct for the commentary, but in the root text, we expect
to continue the series of 4 line verses.) | 33 mdzad pa: Gt ba dang | 34 byin rlabs: Gt byin brlabs | 35 byin kyis brlab: Gt byin gyis
brlabs | 36 gyi: Gt gi | 37 du: Gt tu | 38 rgyan: Gt brgyan | 39 pa ['aya?]ng: it appears that yang has been amended to 'ang; Gt ba la
40 pas: Gt la | 41 myi: Gt mi

348

Edition of the Commentary, the 'Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa

/sgra las 'das pa'i sbyor ba ste42 / *****

/bsam dang bral ba'i ye shes [4] lags /
/h srid gsum 'jom*****s mdzad43 lha la phyag 'tshal lo / ****
/mnyam nyid ngang gis khro la [5] phyag 'tshal lo /
/khro bas longs spyod mdzad la phyag 'tshal lo /
[Gt321] /rdo rje 'du 'phro khro44 [6] la phyag 'tshal lo /45 i
zhes bstod de46 /
/nyid kyi thugs kar zhugs nas myi47 snang bar gyurd48 to / ii
/mnyam las 'phros49 te
[beneath the line] pad ma sam ba bhas rang gz[or?] byas pa + + ma yin bar ston

[84r.1] $/ /byung50 ba'I don51/

/skyes bu gang gis52 rig pa de53 /
/ngag gis54 ci55 skad brjod pa'i sgra /
/thams cad ma lus56 tan tra zhes 57/
[2] 'og myin bla myed58 gnas mchog du59 /
/mgon po bdag nyid chen po yis /
/rdo rje 'khor lo bskor pa60 na /
/ljags kyi dbang po [3] bkram las61 gsungs /
/de las brte****n te mgon po 'i62/
/sku gsung63 thugs **** kyi gsang ba rnams /
/rnal 'byor [4] sgrub pas64 rtogs bya 'i phyir / *****
/ma nor tsam du bshad pa yin // ********** //
[5] $/ /dngos grub mchog brnyes ya mtshan chen po 'i65/
[beneath the line] slobs dpon shan ting gar bas brtags nas ma nor nas/ sam ba bha la stod pa 'o/

/ 'jig rten ngam gyur66 pad ma67 rgyal po yis68 /

/de bzhin gshegs pa'i man ngag [6] gsang chen rnams /
/klung69 nas bkrol mdzad de la phyag 'tshal lo //
//
[84v.1] $/ // *****
// ***** // *****//
[3] // ***** // ***** // *****// ** //
[4] $/ /'phags pa thabs kyis70 ****** zhags pa pad ma71 'phreng
don bsdus pa'i75 'grel pa/ [5] rdzogs s+ho76 //
:
//

72

las73 rtog pa'I rgyal ***** po 'i74

[beneath the line] kam cu pa bo'u ko gis bris// ***** // ***** //

Note that Klong chen pa cites the three lines from /mnyam nyid... up to here in his Phyogs bcu mun sel; see our TZ edition,
Chapter 42 note i).
ii This line is cited exactly in Chapter 22 of Klong chen pa's Phyogs bcu'i mun sel (bDud 'joms bKa' ma Volume La: 618-619); see
our TZ edition, Chapter 42 note ii. See also G. Dorje 1988: 1295- 6.
42 ste:

Gt de | 43 mdzad: Gt pa'i | 44 khro: Gt mdzad (the Tshal pa Kanjur and Bhutanese texts agree with the Dunhuang manuscript
here, but Gt's reading is supported by the South Central NGB and local Kanjur texts) | 45 /: Gt omits | 46 de: Gt do | 47 myi: Gt mi
48 gyurd: Gt gyur | 49 'phros: Gt spros | 50 byung: Gt 'byung | 51 don: Gt las | 52 gis: Gt zhig | 53 de: Gt des | 54 ngag gis: Gt gang
zhig | 55 ci: Gt ji | 56 ma lus: Gt omits | 57 Gt inserts: pa ni | 58 'og myin bla myed: Gt 'og min bla med | 59 du: Gt tu | 60 pa: Gt ba
61 las: Gt pas | 62 'i: Gt yis | 63 gsung: Gt gsungs | 64 sgrub pas: Gt bsgrub pa | 65 'i: Gt yis | 66 ngam gyur: Gt ma 'gyur | 67 pad ma:
Gt padma'i | 68 yis: Gt las | 69 klung: Gt klong | 70 kyis: Gt kyi | 71 pad ma: Gt padmo'i | 72 Gt inserts: ba | 73 las: there is a small line
going through sa, but it is not clear whether deletion is intended; Gt las/ | 74 'i: Gt omits | 75 Gt inserts: le'u ste/ bzhi bcu rtsa gnyis
pa'o// //'phags pa thabs kyi zhabs pa padmo'i 'phreng gi (see TZ edition p.227-228 for variants on the final chapter title and the
final title of the text) | 76 s+ho: Gt so

APPENDIX
THE DEITIES OF THE PEACEFUL MAALA IN THE THABS ZHAGS TRADITION
The Principal Deity and his immediate Circle
Gender
and
number

TZ Ch.9
Mantra List1
(name as given in our edition,
without all variants noted;
Sanskrit equivalents in brackets,
with the
Sarvatathgatatattvasagraha
[STTS] equivalents where
appropriate)2

TZ Ch.10
Mudr List3

TZComm Ch.74
Deity List (each
name is given
twice)

1. male

badzra bai ro tsa na (STTS:

Vairocana)

rnam par snang

mdzad

rnam par snang

mdzad (Ch.6:
be ro tsa na)

TZComm
Ch.9
Dunhuang
ms.
annotations
(where
different from
TZComm
Ch.7)

Using the numbering of the list given here, the order of mantras given in the mantra list of Chapter 9 is as follows: male deities
number 1-25; the principal females of groups 1, 2, 3, 4 and 5 (female deities number 1, 6, 11, 16, 21); the sems mas of the first,
second, third, fourth and fifth groups in order (female deities number 2-5, 7-10, 12-15, 17-20, 22-25). For some slight
uncertainties, see the entries in the table.
The peaceful deity list of the Thabs zhags tradition, consisting of twenty-five male and twenty-five female deities, incorporates
a version of the thirty-seven deities of the STTS. Besides the fact that the overlap is not exact, there are also some gender
reversals. The male primary bodhisattvas of Vairocana's group correspond to female deities in the STTS, while the four male
doorkeepers of the STTS correspond to female deities (sems ma rather than doorkeepers) in the Thabs zhags. The additional
deities given in the Thabs zhags tradition are all female. There are also some apparent parallels with mantras of the rGyud
gsang ba snying po tradition (see table entries below).
The deities are not listed in the root text of Chapter 7, which only gives the outline of five groups of five deities with their
consorts. Thus, the mudr list of Chapter 10, together with the transliterated mantra list in Chapter 9, give us the root text's
understanding of the list of deities. This may have been to some extent tidied up by TZComm's commentary on Chapter 7, and
even possibly in the commentarial tradition's transmission of the mantra list in Chapter 9 (see the variants on the mantras
represented by the commentarial tradition descent of the Dunhuang ms., Bhutan NGB, and especially the Tshal pa Kanjur
versions), so the Chapter 10 mudr list, giving the Tibetan names of the deities, is of great significance. Unfortunately, it
appears to be somewhat corrupted or incomplete, and not in a clear and consistent order throughout. Using the numbering of
the list given here, the order of mudrs given in Chapter 10 is as follows: the vajra palms mudr; the samaya mudr; the mudr
for the principal male deity number 1; the mudr of union; the mudrs for the buddhas of the four directions (male deities
number 6, 11, 16, 21); the bodhisattvas of the second group (male deities number 7, 8, 9, 10); the bodhisattvas of the third
group (male deities number 12, 13, 14, 15); the bodhisattvas of the fourth group (male deities number 17, 18, 19, 20); the
bodhisattvas of the fifth group (male deities number 22, 23, 24, 25); probably the principal females of the third, second, fifth
and first groups (female deities number 11, 6, 21, 1, see the entries in the table); the sems mas of the second group (female
deities number 7, 8, 9, 10); the sems mas of the third group (female deities number 12, 13, 14, 15); the sems mas of the fourth
group (female deities number 17, 18, 19, 20); the sems mas of the fifth group (female deities number 22, 23, 24, 25); possibly
the sems mas of the first group (female deities number 2, 3, 4, 5), or even the male bodhisattva deities of this group (see the
entries in the table). We are thus left without any mention of the principal female of group 4; and the male and female
bodhisattva deities of group 1 seem only to have one set of mudrs between them.
Unfortunately, we are missing the Tenjur version of this section of TZComm, so the list relies solely on the Dunhuang
manuscript. Moreover, we are missing the Tenjur witness also for the root text of Chapter 9 (the entire mantra list) and for
most of Chapter 10, so the Tenjur includes only the final four mudrs. However, the male and female principal deities of the
peaceful maala are listed also in TZComm Chapter 41, witnessed in the Dunhuang manuscript, the Tenjur, and the Tshal pa
Kanjur texts, while TZComm also lists the five male buddhas in Chapters 2, 5, and twice in Chapter 6.

350

Appendix

Most probably,
kun tu bzang mo.
(Alternatively,
perhaps the sbyor
ba'i rgya mchog
applies to her?)

1. female

badzra shwa ra dha tu (uncertain;

Vairocana's consort is usually
rendered Vajradhtv var. While
TZ's phonetic suggests a reversed
variant form of this, perhaps
Vajre varadhtu, TZComm's
translation suggests Vi ve var.)

2. male5

badzra sa twa, with variant in

Dunhuang/Tshalpa
Kanjur/Bhutan NGB: sa twa
badzra (Vajrasattva or
Sattvavajra; STTS:
Sattvavajr/Vajrapramit)

2. female

sti ra badzra (Sthiravajr)

rdo rje sems ma

(uncertain; in this
context, it seems
to indicate the
consort of rdo rje
sems dpa',
although the
name seems a
close parallel to
the STTS
goddess; see
previous item.)6

brtan ma
rdo rje

badzra rad na, with variant in

Dunhuang/Tshalpa
Kanjur/Bhutan NGB: rad
na/ratna badzra (Vajraratna or
Ratnavajra; STTS:
Ratnavajr/Ratnapramit)

Perhaps missing,
but the male deity
names seem to
occur in the
female deity list
(see next item).

rin cen
rdo rje

3. male

thams cad bdag

nyid ma (also
given in
TZComm Chapter
41, but kun du/tu
bzang mo is
given as the
principal female
in Chapter 5)

Perhaps missing, rdo rje sems dpa'/

but the male deity sems dpa' rdo rje
names seem to
(both names are
occur in the
given)
female deity list
(see next item).

sems dpa' rdo

rje

Note that the four males and four females surrounding the central deity pair are in some sense the primary bodhisattva pairs for
each of the surrounding four families; in TZ's Chapter 7, the central group are even described as gtso bo lnga dang gtso mo
lnga, rather than sems dpa' lnga dang sems ma lnga, as each of the other groups. Four further male bodhisattvas with the same
names as the four male deities, or with reversed forms of the names, then appear as the first of the retinue bodhisattvas of each
group (male numbers 7, 12, 17 and 22 below).
There is also a mudr for kun tu bzang mo or kun tu bzang po, although neither are given in the mantra list in TZ's Chapter 9,
nor the deity list in TZComm's Chapter 7. Assuming kun tu bzang mo is intended, she is mentioned as the principal female
deity in TZComm's Chapter 5, apparently taking the place of thams cad bdag nyid ma, and she probably is also Vairocana's
consort in the mudr list (see above). In TZComm's Chapter 6, however, she apparently takes the place of brtan ma rdo rje
(see our edition of TZComm, Chapter 5, Dunhuang ms.14r, and note iii; Chapter 6, Dunhuang ms.16v, and note ii).

351

The Deities of the Maala in the Thabs zhags tradition

3. female

dzwa la ba dzri (Jvlavajr or

Jvalavajr)

rdo rje rin chen

ma (uncertain; in
this context, it
seems to indicate
the consort of rdo
rje rin chen,
although the
name seems a
close parallel to
the STTS
goddess; see
previous item.)

'bar ma
rdo rje

4. male

dharma badzra (Dharmavajra;

STTS:
Dharmavajr/Dharmapramit)

Perhaps missing,
but the male deity
names seem to
occur in the
female deity list
(see next item).

chos rdo rje

chos kyi
rdo rje

sang ha ra badzra
(Saharaavajr)

rdo rje chos ma

(uncertain; in this
context, it seems
to indicate the
consort of rdo rje
chos, although
the name seems a
close parallel to
the STTS
goddess; see
previous item.)

sdud ma
rdo rje/
sdud pa
rdo rje

bsdud ma
rdo rje

5. male

badzra kar ma, with variant in

Dunhuang/Tshalpa
Kanjur/Bhutan NGB: karma
badzra (Vajrakarma or
Karmavajra; STTS:
Karmavajr/Karmapramit)

Perhaps missing,
but the male deity
names seem to
occur in the
female deity list
(see next item).

las kyi
rdo rje/
las rdo rje

5. female

tsad tsa la badzra (uncertain;

possibly, Cacalavajr or
Sacalavajr?)

rdo rje las ma

(uncertain; in this
context, it seems
to indicate the
consort of rdo rje
las, although the
name seems a
close parallel to
the STTS
goddess; see
previous item.)

bskyod ma
rdo rje

4. female

352

Appendix

Akobhya's Circle
Gender
and
number

TZ Ch.9
Mantra List
(name as given in
our edition, without all variants
noted)

TZ Ch.10
Mudr List

TZComm Ch.7
Deity List (each
name is given
twice)

TZComm
Ch.9
Dunhuang
ms.
annotations
(where
different from
TZComm
Ch.7)

6. male

badzra a ko bya
(STTS: Akobhya)

mi bskyod pa

myi bskyod pa
(myi khrug[s] pa
in TZComm
Chapters 2, 5, 6,
and 41; and Ch.6:
ag sho bhya)

myi 'khrug pa

badzra bud dha lo tsa na,

(Buddhalocan)

rdo rje spyan?

yangs pa'i spyan

7. male

badzra sa twa
(STTS: Vajrasattva)

rdo rje
sems dpa'

rdo rje
sems dpa'

7. female

badzra la sye
(STTS: Vajralsy)

rdo rje sgeg mo

rdo rje
sgreg mo

8. male

badzra ra tsa
(STTS: Vajrarja)

rdo rje rgyal po

rdo rje
rgyal po

badzra dhu pe
(STTS: Vajradhp)

rdo rje
bdug pa

rdo rje bdug pa

ma/ rdo rje bdug
pa

badzra ra ga
(STTS: Vajrarga)

rdo rje 'dod pa

rdo rje
'dod pa

9. female

badzra a ghu sha

(STTS: Vajrku a)

rdo rje
lcags kyu

rdo rje
kyo ga ma

10. male

badzra sa du
(STTS: Vajrasdhu)

rdo rje legs pa

rdo rje
legs pa

badzra de'i sh re de bya

(uncertain; possibly Dvearati?
The mantra, dvearati, features in
the rGyud gsang ba snying po,
associated with Buddhalocan; see
Dorje 1988: 691.)

rdo rje sgrol ma

(probably in
error; the
Dunhuang ms.
gives rdo rje
sgril ma)

rdo rje
sgril ma

6. female

8. female

9. male

10. female

353

The Deities of the Maala in the Thabs zhags tradition

Ratnasambhava's Circle
Gender and
number

TZ Ch.9
Mantra List
(name as given in our edition,
without all variants noted)

TZ Ch.10
Mudr List

TZComm Ch.7
Deity List (each
name is given
twice)

badzra rad na sam bha ba (STTS:

Ratnasambhava)

rin chen
'byung ba

rin cen 'byung

gnas (rin cen
'byung ldan in
TZComm
Chapters 2 and 5;
rin cen dpal and
rad na sam bha
ba in Chapter 6)

badzra ma ma ki (Mmak)

rdo rje
'dzin ma?

rin cen spyan

(rdo rje 'dzin ma
is noted as an
alternative name;
nam ka is given
in the Dunhuang
TZComm Chapter
41, and m ma k
in the other
editions)

12. male

badzra rad na
(STTS: Vajraratna)

rdo rje rin

chen

rdo rje rin cen

12. female

badzra ma le
(STTS: Vajraml)

rdo rje 'phreng

ba

rdo rje 'phreng

ba

badzra su rya (Vajrasrya; STTS:

Vajratejas)

rdo rje
nyi ma

rdo rje
gzi brjid

badzra pu pe
(STTS: Vajrapup)

rdo rje
me tog ma

rdo rje
me tog ma

badzra ke tu, with variant in the

Tshal pa Kanjur:
badzra dhwa dza
(STTS: Vajraketu)

rdo rje
rgyal mtshan

rdo rje
rgyal mtshan

11. male

11. female

13. male

13. female

14. male

TZComm
Ch.9
Dunhuang
ms.
annotations
(where
different from
TZComm
Ch.7)

rin cen spyan

(without
alternative)

rdo rje nyi ma

354
14. female

15. male

15. female

Appendix

badzra pa sha (STTS: Vajrap a)

rdo rje
zhags pa ma

rdo rje
zhags pa ma/
rdo rje
'gas pa ma (but
'gas is most
likely a scribal
error for zhags)

badzra ha sa
(STTS: Vajrahsa)

rdo rje
bzhad pa

rdo rje
bzhad pa

badzra su ki, with variant in

Dunhuang/Tshalpa Kanjur/Bhutan
NGB: badzra rad na ra ti
(Vajrasukh? The variant,
presumably indicating
Vajraratnarati, might parallel the
rGyud gsang ba snying po's
mohrati, explained as signifying
Mmak, and thus, the ratna
family; see Dorje 1988: 691.)

rdo rje
bde ba ma

rdo rje bde ba/

rdo rje bde ba ma

rdo rje
zhags pa ma

Amitbha's Circle
Gender and
number

TZ Ch.9
Mantra List
(name as given in our edition,
without all variants noted)

TZ Ch.10
Mudr List

TZComm Ch.7
Deity List (each
name is given
twice)

16. male

badzra a mi ta ba (Amitbha;
STTS: Loke vararja)

snang ba
mtha' yas

snang ba mtha'
yas ('od dpag
myed in
TZComm Ch.
41; and Ch.6: a
myi dha ba)

badzra ban da ra ba si ni,

(Paravsin)

missing?

gos dkar ma
(gos dkar mo in
TZComm
Chapter 41)

badzra dharma
(STTS: Vajradharma)

rdo rje chos

rdo rje chos

16. female

17. male

TZComm Ch.9
Dunhuang ms.
annotations
(where
different from
TZComm Ch.7)

gos dkar mo

355

The Deities of the Maala in the Thabs zhags tradition

17. female

badzra gir ti
(STTS: Vajragt)

rdo rje
dbyings ma
(=dbyangs
ma?)7

rdo rje dbyings

len ma/ rdo rje
dbyangs len ma

badzra t ka
(STTS: Vajratka)

rdo rje
rnon po

rdo rje
rnon po

badzra a lo ke
(STTS: Vajrlok)

rdo rje
mar me

rdo rje mye

sgron ma

19. male

badzra tsa kra (Vajracakra; STTS:

Vajrahetu)

rdo rje
'khor lo

rdo rje
'khor lo

19. female

badzra spo ta (STTS: Vajrasphoa)

rdo rje lcags

sgrog ma

rdo rje lcags

sgrog ma

badzra pa sha
(STTS: Vajrabha)

rdo rje
smra ba

rdo rje
smra ba

badzra ra ga ra ti (Possibly
Vajrargarati, or Vajrarokarati?
Although rgarati would not seem
a clear equivalent for the Tibetan
name, it features in the rGyud
gsang ba snying po, associated
with Paravsin; see Dorje
1988: 692.)

rdo rje bkrag

gsal ma

rdo rje gsal

bkra ma

18. male

18. female

20. male

20. female

rdo rje
dbyangs len ma

rdo rje mye

mar ma

It seems that this may not be a one-off error in the archetype; rdo rje dbyings len ma is given on one occasion in TZComm
Chapter 7, and rdo rje dbyings is given in the Dunhuang manuscript's marginal notes (TZComm Chapter 5, 14r.3), apparently
in the context of referring to the goddesses of the senses. These are listed and more explicitly specified as the sense goddesses
in Chapter 6. Apart from the slight anomoly of kun tu bzang mo, they clearly correspond to the second female [according to
our list] of each of the five buddha families, and there the spelling for the goddess of sound is given appropriately as rdo rje
dbyangs.

356

Appendix

Amoghasiddhi's Circle
Gender and
number

TZ Ch.9
Mantra List
(name as given in our edition,
without all variants noted)

TZ Ch.10
Mudr List

TZComm Ch.7
Deity List (each
name is given
twice)

TZComm Ch.9
Dunhuang ms.
annotations
(where
different from
TZComm Ch.7)

badzra a mo ka si ti (STTS:
Amoghasiddhi)

don yod
grub pa

don yod pa grub

pa/ don yod pa
grub pa (gdon
myi za bar grub
pa in TZComm
Chapters 2, 5
and 6; and Ch.6:
a mo ga sid dhI)

don yod
grub pa

badzra sa ma ya ta ra,
(Samayatr)

rdo rje
sgrol ma?

dam tshig sgrol

ma

22. male

badzra karma
(STTS: Vajrakarma)

rdo rje las

rdo rje las

22. female

badzra nir ti
(STTS: Vajranty)

rdo rje gar

rdo rje gar byed

ma

23. male

badzra raksha
(STTS: Vajraraka)

rdo rje
srung ba

rdo rje bsrung

ba

badzra gan dhe

(STTS: Vajragandh)

rdo rje dri

chab ma

rdo rje byug pa

ma

badzra yaka
(STTS: Vajrayaka)

rdo rje gnod

sbyin

rdo rje gnod

sbyin

badzra ghan da
(Vajragha?)

rdo rje
dril bu

rdo rje dril

'khrol ma

badzra san dhi, with variant in the

Tshal pa Kanjur:
badzra mui
(STTS: Vajrasandhi)

rdo rje
khu tshur

rdo rje
khu tshur

21. male

21. female

23. female

24. male

24. female

25. male

rdo rje
srung ba

357

The Deities of the Maala in the Thabs zhags tradition

25. female

badzra ra ti
(STTS: Vajrve a; here, the
transcription of Vajrarati does not
seem to fit with the Tibetan name,
but this mantra features in the
rGyud gsang ba snying po,
associated with Tr; see Dorje
1988: 692.)

rdo rje
'bebs pa ma

rdo rje 'bebs pa

ma/ rdo rje
'bebs ma

rdo rje
'phebs ma

358

Appendix

THE DEITIES OF THE WRATHFUL MAALA IN THE THABS ZHAGS TRADITION8

Deity
type,
Number,
Gender
and
Direction

TZ Ch.12/139

TZ Ch.14
Mantra List10
(name as given in
our edition, without
all variants noted)11

TZ Ch.16
Mudr List12

TZComm Ch.12/13
Deity List13

gtso bo
male
Centre

khrag 'thung
chen po

shr he ru ka ( r
heruka)

khrag 'thung
chen po

shr he ru ka

gtso mo
female
Centre

rdo rje srin mo

chen mo/ 'jigs
byed chen mo

N/A

N/A

gtso mo ral pa
gcig ma (Tenjur:
gtso mo
ral gcig ma)

9
10

11

12

13

TZComm
Ch.14/16
Dunhuang
ms.
annotations
(where
different
from
TZComm
Ch.12/13)

This list gives all the main wrathful deities of the maala and their immediate emanations. It does not include further
wrathful attendant and messenger deities mentioned at various points in TZ and TZComm, such as the groups of goddesses
given in Chapter 11 (see the notes to our TZComm edition of Chapter 11), or the many further 'phra men ma referred to in
TZComm Chapters 13, 18 and 19, and hinted at in the root text for those chapters.
Only the principal male and female wrathful deity pair are specified by name in TZ Chapters 12 and 13.
The mantra list for the male deities in TZ Chapter 14 starts with the principal male deity and proceeds with the khro bo bcu in
the order we give here, from 1-10. This is the standard order for these deities, corresponding to the Above, followed by the
directions in clockwise order through the points of the compass, starting with the East, and concluding with the Below. Note
that this order is not conformed to elsewhere in the text and commentary. The mantras for the female wrathful deities in
Chapter 15 give only series of seed syllables for the deities rather than names, so we have no transliterated list of the female
deities' names.
Note that we are missing the Tenjur witness for both the mantra list and the mudr list in Chapters 14 and 16 respectively,
although we do have the Tenjur for the deity lists in TZComm Chapters 12 and 13.
The mudr list for the male deities in TZ Chapter 16 starts with the principal male deity and proceeds with the khro bo bcu in
the following order: H (number 1, Above), sTobs chen (number 10, Below), and then the standard order for the deities at
the points of the compass from the East (number 2-9).
The descriptions of the male deities in TZComm Chapter 12 begin with the principal male deity and proceed with the khro bo
bcu in the following order: the standard order for the deities at the points of the compass from the East (number 2-9); sTobs po
che (number 10, Below), and finally, Hkara (number 1, Above). They are given in a further different order in TZComm
Chapter 13, but that list relates the deities to the different parts of the body, so the order reflects the body parts rather than the
deities. Note that the female khro mo bcu are simply listed in TZComm Chapter 13, without specifying their consorts or
directions, so we have assumed that these correspond to the standard lists. The list is not, however, given in quite the usual
order. In terms of the consorts' standard directional associations, the list starts with the East (and for this consort alone, the
eastern direction is specified), works through the points of the compass in clockwise order up to the North-west, then proceeds
through the Above, the North-east, the Below, and the North. The male attendants (rdo rje phra men) are listed at the end of
Chapter 12, and assuming that they fit with the standard directions for these deities, the order of the list corresponds precisely
to the order of the khro bo bcu given in this chapter, that is, the attendant associated with the Above is given at the end rather
than at the beginning of the list, as is usual. The female attendants ('phra men ma), listed in Chapter 13, are given in the same
order, with the attendant associated with the Above given at the end of the list. The consistency in the order of these two lists
would seem to confirm that these male and female animal-headed deities correspond to the standard list, found for example, in
virtually all Vajraklaya texts of any length, and we can have confidence in their placement with each specific wrathful male
and female as given in our list here.

359

The Deities of the Maala in the Thabs zhags tradition

khro bo
bcu 1.
Above

N/A

h ka rdza, with
variant in
Dunhuang/Tshalpa
Kanjur/Bhutan
NGB: h ka ra
(Hkara)

hung ka ra
(Tenjur:
h ka ra)

h ka ra

rdo rje
phra men
male 1.

N/A

N/A

N/A

phag gyi/gi mgo

can

N/A

khro mo
bcu 1.

N/A

N/A

N/A

rdo rje sgra 'byin

ma

N/A

'phra men
ma female
1.

N/A

N/A

N/A

smrigs (Tenjur:
smig) bu'i mgo can

N/A

khro bo
bcu 2. East

N/A

bi dza ya (Vijaya)

rnam par
rgyal ba

rnam par rgyal ba

rdo rje
phra men
male 2.

N/A

N/A

N/A

stag mgo can

(Tenjur:
stag gi mgo can)

N/A

khro mo
bcu 2.

N/A

N/A

N/A

rdo rje bsnyems

(Tenjur:
snyems) ma

N/A

'phra men
ma female
2.

N/A

N/A

N/A

bya rgod mgo can

N/A

khro bo
bcu 3.
South-east

N/A

ni la tan ta
(Nladaa)

dbyig
sngon can

be con
sngon po can

dbyig po
sngon po

rdo rje
phra men
male 3.

N/A

N/A

N/A

g.yag mgo can

(Tenjur:
g.yag gi mgo can)

N/A

khro mo
bcu 3.

N/A

N/A

N/A

rdo rje sder mo

(sder mo is
omitted in the
Dunhuang ms., but
found in the
Tenjur)

N/A

360

Appendix

'phra men
ma female
3.

N/A

N/A

N/A

bya rog mgo can

(Tenjur:
omits, surely in
error)

N/A

khro bo
bcu 4.
South

N/A

ya man ta ka
(Yamntaka)

gshin rje
'joms pa

Ch.12: gshin rje

gshed po; Ch.13:
ya man ta ka
(Tenjur:
ya manta ka)

gshin rje
'joms pa

rdo rje
phra men
male 4.

N/A

N/A

N/A

sha ba'i mgo can

N/A

khro mo
bcu 4.

N/A

N/A

N/A

rdo rje dur khrod

bdag mo

N/A

'phra men
ma female
4.

N/A

N/A

N/A

'ug pa'i mgo can

N/A

khro bo
bcu 5.
Southwest

N/A

a rya a tsa la
( rycala)

mi g.yo mgon
po

myi/mi g.yo
mgon po

myi g.yo ba

rdo rje
phra men
male 5.

N/A

N/A

N/A

gzig mgo can

(Tenjur:
gzig gi mgo can)

N/A

khro mo
bcu 5.

N/A

N/A

N/A

rdo rje gtung

(Tenjur:
gtun) khung ma

N/A

'phra men
ma female
5.

N/A

N/A

N/A

khwa'i mgo can

N/A

khro bo
bcu 6.
West

N/A

ha ya 'gri ba
(Hayagrva)

'gri ba

ha ya gri ba
(Tenjur:
ha ya gr ba)

dpal rta
mgrin

rdo rje
phra men
male 6.

N/A

N/A

N/A

byi la'i mgo can

(Dunhuang gives
bya ba'i mgo can,
surely in error)

N/A

khro mo
bcu 6.

N/A

N/A

N/A

rdo rje gtum mo

N/A

361

The Deities of the Maala in the Thabs zhags tradition

'phra men
ma female
6.

N/A

N/A

N/A

pu shud kyi mgo

can (Dunhuang
gives phub shud
mgo can, a minor
spelling error)

N/A

khro bo
bcu 7.
Northwest

N/A

a pa ra tsi ta
(Aparjita)

gzhan gyis
mi thub

gzhan gis myi/gyis

mi thub pa

rdo rje
phra men
male 7.

N/A

N/A

N/A

spyang ka 'i mgo

can (Tenjur:
spyang ki'i mgo
can)

N/A

khro mo
bcu 7.

N/A

N/A

N/A

rdo rje mda'

snyegs ma (Tenjur:
rdo rje bsnyems
ma) [the standard
name is, rdo rje
mda' snyems ma,
and this is
probably intended]

N/A

'phra men
ma female
7.

N/A

N/A

N/A

khra'i mgo can

N/A

khro bo
bcu 8.
North

N/A

a mri ta kun dha li

(Amtakualin)

bdud rtsi
'khyil pa

bdud rtsi 'khyil pa

rdo rje
phra men
male 8.

N/A

N/A

N/A

seng ge 'i mgo can

(Tenjur omits,
surely in error)

N/A

khro mo
bcu 8.

N/A

N/A

N/A

rdo rje rlung 'byin

ma

N/A

'phra men
ma female
8.

N/A

N/A

N/A

pha wang (Tenjur:

pha bang gi) mgo
can

N/A

khro bo
bcu 9.
North-east

N/A

tre log kya bi dza ya

(Trailokyavijaya)

khams gsum
rgyal ba

khams gsum rnam

par rgyal ba

362

Appendix

rdo rje
phra men
male 9.

N/A

N/A

N/A

dom gyi mgo can

N/A

khro mo
bcu 9.

N/A

N/A

N/A

rdo rje gsod ma

N/A

'phra men
ma female
9.

N/A

N/A

N/A

sre mo 'i (Tenjur:

sre mong gi) mgo
can

N/A

khro bo
bcu 10.
Below

N/A

ma ha ba la
(Mahbala)

stobs chen

stobs po che

rdo rje
phra men
male 10.

N/A

N/A

N/A

dred kyi mgo can

N/A

khro mo
bcu 10.

N/A

N/A

N/A

rdo rje skul byed

ma

N/A

'phra men
ma female
10.

N/A

N/A

N/A

byi ba 'i mgo can

(Tenjur: bya ba'i
mgo can, surely in
error)

N/A

BIBLIOGRAPHY

Dunhuang Tibetan Documents

Dunhuang Tibetan manuscripts held at the British Library, London:
IOL Tib J 321 ('Phags pa thabs kyi zhags pa pad ma 'phreng gi don bsdus pa'i 'grel pa'); IOL Tib J 438.
IDP: The International Dunhuang Project <http://idp.bl.uk/>. Contains digital images of many items from the Stein collection, and a
catalogue (Dalton and van Schaik 2005). In particular, the images of the 'Phags pa thabs kyi zhags pa pad
ma 'phreng gi don bsdus pa'i 'grel pa' are available to download from this site, at:
http://idp.bl.uk/database/oo_loader.a4d?pm=IOL Tib J 321.
The Mellon International Dunhuang Archive < http://www.artstor.org/what-is-artstor/w-html/col-mellon-dunhuang.shtml> Contains
digital images of many of the Dunhuang Tibetan manuscripts held at the Bibliothque nationale, Paris, in the
Pelliot Tibtain (PT) collection.
Old Tibetan Documents Online website <http://otdo.aa.tufs.ac.jp/>.

Editions of the rNying ma'i rgyud 'bum [NGB]

sDe dge [D]: The sDe dge edition of the rNying ma'i rgyud 'bum. Twenty-six volumes, Ka-Ra, plus Dkar chag, Volume A. sDe dge
par khang chen mo. The 'Phags pa thabs kyi zhags pa padmo'i phreng is in Volume Pa: 286r-298r.
mTshams brag [M]: The Mtshams brag manuscript of the Ri ma rgyud 'bum (rgyud 'bum/ mtshams brag dgon pa). 1982. Thimphu:
National Library, Royal Government of Bhutan. Forty-six volumes. Microfiche available from The Institute
for Advanced Studies of World Religions, LMpj 014,862 - 014, 907. An electronic version is available from
the Tibetan Buddhist Resource Centre (http://www.tbrc.org), under the title, rnying ma rgyud 'bum, mtshams
brag dgon pa'i bris ma (W21521).
It is also available online, linked to its catalogue, at
http://www.thlib.org/encyclopedias/literary/canons/ngb/ngbcat.php#cat=tb/0416. The 'Phags pa thabs kyi
zhags pa padmo 'phreng gi don bsdus pa is in Volume Wa: 123-152.
sGang steng [G]: The rNying ma'i rgyud 'bum manuscripts preserved by sGang steng monastery, Bhutan. Forty-six volumes. Digital
images were made under an AHRC funded project at Oxford University. The collection primarily consulted
is the sGang steng b manuscript; more recently, the other manuscript collection held at the monastery - sGang
steng a - has also been photographed as part of a British Library Endangered Archives Research Project
(http://eap.bl.uk/database/overview_project.a4d?projID=EAP039;r=25547). The 'Phags pa thabs kyi zhags
pa padmo 'phreng gi don bsdus pa is in Volume Wa: 51r-65r.
dGra med rtse [Gr]: The rNying ma'i rgyud 'bum manuscripts preserved by dGra med rtse monastery, Bhutan. Forty-six volumes. It
has been photographed as part of a British Library Endangered Archives Research Project
(http://eap.bl.uk/database/overview_project.a4d?projID=EAP105;r=12859). The 'Phags pa thabs kyi zhags
pa padmo 'phreng gi don bsdus pa is in Volume Wa: 46r-59r.
gTing skyes [T]: Ri ma rgyud 'bum Reproduced from the MS preserved at Gti-skyes Dgon-pa-bya Monastery in Tibet, under the
direction of Dingo Khyentse Rimpoche, Thimbu, 1973. (Microfiche of some volumes available from The
Institute for Advanced Studies of World Religions, LMpj 011,825 - 012,584.) An electronic version is
available from the Tibetan Buddhist Resource Centre (http://www.tbrc.org), under the title, rnying ma rgyud
'bum, gting skyes (W21518). The 'Phags pa thabs kyi zhags pa padma phreng gi don bsdus pa is in Volume
Dza: 395-422.

364

Bibliography

Rig 'dzin Tshe dbang nor bu [R]: The Rig 'dzin Tshe dbang nor bu edition of the rNying ma'i rgyud 'bum. Twenty-nine volumes are
held at the British Library, under the classification, "RNYING MA'I RGYUD 'BUM MSS", with the
pressmark, OR15217. Volume Ka is held at the Bodleian Library Oxford at the shelfmark, MS. Tib.a.24(R).
Microfilm is available from The British Library, and the Bodleian Library for Volume Ka. Title folios to
Volume Ga and Volume A are held at the Victoria and Albert Museum, Accession no.s: IM 318-1920 and IM
317-1920. The 'Phags pa thabs kyi zhags pa padma phreng gi don bsdus pa is in Volume Dza: 180r-193r.
Kathmandu [K]: Manuscript edition of the rNying ma'i rgyud 'bum from the Khumbu region, held by The National Archives,
Kathmandu. Microfilm is available through the Nepal Research Centre of the Nepalese-German Manuscript
Cataloguing Project. The short title is ri ma rgyud 'bum, Ms no.22, running no.17, reel AT12/3 AT13/1.
The 'Phags pa thabs kyi zhags pa padmo 'phreng gi don bsdus pa is in Volume Ma: 320r-336r.

bKa' 'gyur and bsTan 'gyur Collections

The sDe dge bKa' 'gyur, the sde-dge mtshal-par bka'-'gyur [Dk]: a facsimile edition of the 18th century redaction of si-tu chos-kyi'byun-gnas prepared under the direction of h.h. the 16th rgyal-dban karma-pa, 1976-1979. 103 volumes.
Delhi, Karmapae Chodhey, Gyalwae Sungrab Partun Khang. An electronic version is available from the
Tibetan Buddhist Resource Center, New York (W22084). The 'Phags pa thabs kyi zhags pa padmo'i phreng
is in Volume rnying rgyud Kha (98): 597-621.
The lHa sa bKa' 'gyur [Hk], 1978. 101 volumes. Microfiche set made from a xylograph completed in the early 20th century, kept
in Rashi Gempil Ling (First Kalmuck Buddhist Temple) in Howell, New Jersey. Stony Brook, N. Y., The
Institute for Advanced Studies of World Religions. An electronic version is available from the Tibetan
Buddhist Resource Center, New York (W26071). The 'Phags pa thabs kyi zhags pa padmo'i phreng is in
Volume rgyud Wa: 472v-492r.
The 'Jang sa tham or Li thang bKa' 'gyur [J], from the private collection of Namkha Drime Rinpoche, Jeerang, Orissa. The 'Phags
pa thabs kyi zhags pa padmo'i phreng is in Volume rgyud 'bum (Rnying rgyud), Wa: 294v-307r. It is
available in prints or copies made from the microfilm held at the Staatsbibliothek, Berlin.
The sNar thang bKa' 'gyur [Nk] and bsTan 'gyur [Nt], Narthang Kanjur, 102 volumes, set at the International Academy of Indian
Culture, New Delhi, scanned by the Tibetan Buddhist Resource Center, New York (W22703). The 'Phags pa
thabs kyi zhags pa padmo'i phreng is in the sNar thang bka' 'gyur Volume rgyud Wa: 816-855. The new
sNar thang bstan 'gyur edition (from the blocks made in 1741-1742), in 225 volumes. Note that the Tibetan
Buddhist Resource Center, New York, have scanned a copy in 225 volumes, preserved at Tibet House, Delhi,
supplemented with pages and volumes from Dharamsala and libraries in the U.S.A. (W22704). The 'Phags
pa thabs kyi zhags pa padmo 'phreng gi don bsdus pa'i 'grel pa is in the sNar thang bstan 'gyur Volume
rgyud Bu (77): 176-228.
The Peking bKa' 'gyur [Qk] and bsTan 'gyur [Qt], reprinted and catalogued in The Tibetan Tripitaka, Peking Edition, kept in the
library of the Otani University, Kyoto, edited by D.T. Suzuki, 1955-1961. Vol. 1-45 Bka-gyur. Vol. 46150 Bstan-gyur. Vol. 151 Dkar-chag. Vol. 152-164 Extra (Btso Kha Pa/Lca Skya). Vol. 165-168
Catalogue. Tokyo, Kyoto: Suzuki Research Foundation. The 'Phags pa thabs kyi zhags pa padmo'i phreng is
in the Bka-gyur Volume rgyud Wa: 299v-313r and the 'Phags pa thabs kyi zhags pa padmo 'phreng gi don
bsdus pa'i 'grel pa is in the Bstan-gyur Volume rgyud 'grel Bu: 101r-129v.
The Urga Kanjur [U], edited by Lokesh Chandra, 1990-1994, from the collection of Prof. Raghuvira. 105 volumes. Delhi:
International Academy of Indian Culture and Aditya Prakashan. An electronic version is available from the
Tibetan Buddhist Resource Center, New York (W29468). The 'Phags pa thabs kyi zhags pa padmo'i phreng
is in Volume rnying rgyud Kha: 597-621.

Bibliography

365

The Golden bsTan 'gyur (gser gyi lag bris ma) [Gt], produced between 1731-1741, currently held at Ganden Monastery; published
in Tianjing 1988, digitally scanned for TBRC, New Delhi 2002. An electronic version is available from the
Tibetan Buddhist Resource Center, New York (W23702). The 'Phags pa thabs kyi zhags pa padmo 'phreng
gi don bsdus pa'i 'grel pa commentary is in Volume rgyud 'grel Bu (78): 243-321.
The Ulan Bator Manuscript Kanjur [V], held in the National Library of Mongolia. 114 volumes (3 missing). We have had access
only to images of a few folios. A digital version is available from the Digital Preservation Society
(http://www.tibet-dps.org/index.htm). The 'Phags pa thabs kyi zhags pa padmo'i phreng is in Volume rnying
rgyud Kha from f.358v.
Local Kanjurs: copies of the Local Kanjurs of Hemis [He], from Hemis Tshoms lha khang, of Bathang [Bth], which is held in the
Newark Museum, New York, and of Tawang, O rgyan gling [Ogl] (of which a microfiche copy is held at
the Central University of Tibetan Studies in Sarnath), have not yet been made available in published form.
The 'Phags pa thabs kyi zhags pa pad mo 'phreng gis don bsdus pa occurs in Volume rgyud A of the Bathang
Kanjur, ff.204r-214r; the 'Phags pa thabs kyi zhags pa pad mo 'phreng gyi don bsdus pa occurs in Volume
rgyud Nya of the Tawang O rgyan gling, ff.367v-381r; and the 'Phags pa thabs kyi zhags pa padmo phreng
gyi don bsdus pa occurs in Volume brgyud ka of the Hemis Kanjur, ff.31r-45v.

Dictionaries in Tibetan (see below for Dictionaries using transliteration)

Bod rgya tshig mdzod chen mo 1985 (Tibetan-Tibetan and Chinese dictionary, Chengdu, Szechuan), Mi rigs dpe skrun khang.
bTsan lha ngag dbang tshul khrims 1997 brDa dkrol gser gyi me long, Beijing, Mi rigs dpe skrun khang.

Other Tibetan Sources

bKa' ma shin tu rgyas pa (snga 'gyur bka' ma) 120 volumes, published by Ka thog mkhan po 'jam dbyangs, Chengdu 1999, scanned
by the Tibetan Buddhist Resource Center, New York (W25983). The 'Phags pa thabs kyi zhags pa padmo
'phreng gi don bsdus pa'i 'grel pa is in Volume Wu (80): 125-236, copied from the Peking bsTan 'gyur.
Klong chen pa dPal gsang ba'i snying po de kho na nyid nges pa'i rgyud kyi 'grel pa phyogs bcu'i mun pa thams cad rnam par sel ba,
bdud 'joms bka' ma (Bdud-'Joms 'Jigs-bral-ye- es-rdo-rje Rin ma Bka' ma rgyas pa, 58 volumes 1982-1987,
published by Dupjung Lama, Kalimpong) Volume La: 5-629.
dKon mchog spyi 'dus (compilation based on the revelation of 'Ja' mtshon snying po, 1585-1656) Yang zab dkon mchog spyi 'dus
dang zhi khro nges don snying po'i phyag len chog sgrigs bklag chog tu bkod pa bla ma dam pa'i zhal lung,
n.d, Kathmandu: Lopen Tashi Tsering (HB) Lama.
'Jam mgon Kong sprul blo gros mtha' yas dpal rdo rje phur pa rtsa ba'i rgyud kyi dum bu'i 'grel pa snying po bsdud pa dpal chen
dgyes pa'i zhal lung, bdud 'joms bka' ma (Bdud-'Joms 'Jigs-bral-ye- es-rdo-rje Rin ma Bka' ma rgyas pa, 58
volumes 1982-1987, published by Dupjung Lama, Kalimpong) Volume Tha: 15-213.
Nyang ral, Nyi ma 'od zer bKa' brgyad bde gshegs 'dus pa'i chos skor, 1979-1980. 1. bka' brgyad bde gshegs 'dus pa las: bde bar
gshegs pa rtsa ba'i rgyud ces bya ba le'u nyi shu pa, Volume Ka: 105-149. 2. bka' brgyad bde gshegs 'dus pa
las: byin brlabs phun sum tshogs pa phur pa'i bsgrub pa bi m las mdzad pa zhi ba'i mchog tan spo ba'i cho
ga, Volume Ta: 351-365. Paro, Lama Ngodrup Kyichu Temple. An electronic version is available from the
Tibetan Buddhist Resource Center, New York (W22247).
Nyang ral, Nyi ma 'od zer Slob dpon padma 'byung gnas kyi skyes rabs chos 'byung nor bu'i phreng ba zhes bya ba, rnam thar zangs
gling ma, two versions cited: 1. Kathmandu National Archives manuscript, reel E2703/10. 2. The Rin chen
gter mdzod chen mo (a reproduction of the Stod-lun Mtshur-phu redaction of 'Jam-mgon Ko-sprul's great
work on the unity of the gter-ma traditions of Tibet), Volume Ka: 1-204. Paro: Ngodrup and Sherab Drimay,

366

Bibliography
Kyichu Monastery, 1976. An electronic version is available from the Tibetan Buddhist Resource Centre,
New York (W20578).

bDud 'joms gling pa gter chos/ bdud 'joms gling pa; sprul pa'i gter chen bdud 'joms gling pa'i zab gter gsang ba'i chos sde, 21
volumes, 2004, Lama Kuenzang Wangdue, Thimphu, Bhutan. An electronic version is available from the
Tibetan Buddhist Resource Center, New York (W28732).
bDud 'joms Rin po che, 'Jigs bral ye shes rdo rje (Dudjom Rinpoche) The collected writings and revelations of H. H. bDud-'joms Rinpo-che 'Jigs bral ye shes rdo rje, 25 volumes, 1979-1985, Dupjung Lama, Kalimpong. An electronic version
is available from the Tibetan Buddhist Resource Centre (http://www.tbrc.org), under the title, bDud 'joms
'jigs bral ye shes rdo rje'i gsung 'bum, W20869 0334-0358. 25 Vols.)
Dharma r, Lo chen. Tshogs chen 'dus pai sgrub thabs dngogs grub char 'bebs, bdud 'joms bka' ma (Bdud-'Joms 'Jigs-bral-ye- esrdo-rje Rin ma Bka' ma rgyas pa, 58 volumes 1982-1987, published by Dupjung Lama, Kalimpong) Volume
Pha: 349-443.
gNubs chen sangs rgyas ye shes sGom gyi gnad gsal bar phye ba bsam gtan mig sgron (rnal 'byor mig gi bsam gtan or bSam gtan
mig sgron: A treatise on bhvan and dhyna and the relationships between the various approaches to
Buddhist contemplative practice). Reproduced from a manuscript made presumably from an Eastern Tibetan
print by 'Khor-gdong gter-sprul 'Chi-med rig-'dzin, Smanrtsis shesrig spendzod, Volume 74, Leh:
Tashigangpa, 1974. An electronic version is available from the Tibetan Buddhist Resource Centre
(W00EGS1016286). The text is also reprinted in the bKa' ma shin tu rgyas pa (see above) Volume Phe, 104:
573-1077, Chengdu, 1999.
'Phang thang ma: dkar chag 'phang thang ma/ sgra sbyor bam po gnyis pa 2003 Beijing: Mi rigs dpe skrun khang.
Mag gsar Kun bzang stobs ldan dbang pa 2003 Phur pa'i rnam bshad he ru ka dpal bzhad pa'i zhal lung (bcom ldan 'das dpal chen
rdo rje gzhon nu'i 'phrin las kyi rnam par bshad pa he ru ka dpal bzhad pa'i zhal lung). sNgags mang zhib
'jug khang (Ngak Mang Institute), Beijing: Mi rigs dpe skrun khang.
Mi pham rgya mtsho 1984-1993 gsang 'grel phyogs bcu'i mun sel gyi spyi don 'od gsal snying po. In 'Jam mgon 'ju mi pham rgya
mtsho'i gsung 'bum rgyas pa sde dge dgon chen par ma. Paro: Lama Ngodrup and Sherab Drimey. Volume
19: 1-272.
An electronic version is available from the Tibetan Buddhist Resource Centre
(http://www.tbrc.org), under the title, bDud 'joms 'jigs bral ye shes rdo rje'i gsung 'bum (W23468), 27 Vols.
Rong zom, Chos kyi bzang po 1976 Rong zom bka' 'bum, reproduced from a manuscript copy of an incomplete print from the Zhe
chen wood blocks, with the detailed dkar chag of 'Jam mgon Mi pham. Thimphu: Kunsang Topgay. An
electronic version is available from the Tibetan Buddhist Resource Centre (W27479-4344).
Sa skya Phur chen: dPal rdo rje gzhon nu sgrub pa'i thabs bklags pas don grub, by Sa skya pa bSod nams rgyal mtshan (1312 1375), dPal sa skya'i chos tshogs, Rajpur, India (Tibetan date given: 992). Two different electronic editions
of the Sa skya Phur chen are available. One is included in Volume 18 of the dpal chen k la ya'i chos skor
phyogs bsgrigs [si khron zhing chen mi rigs zhib 'jug su'o, bod kyi shes rig zhib 'jug khang, khreng tu'u 2002
(http://tbrc.org/#library_work_Object-W24051), p.137ff, and one included in the rGyud sde kun btus,
Volume 16. This itself has two available versions: Lungtok & Gyaltsan, Delhi 1971-1972 [W21295], p.427505; and Sachen International, Kathmandu 2004 [W27883], p.479-568. Both are apparently based on the
sDe dge blocks, although the Sachen International version is an entirely remade computer input version
edition rather than a photographic reproduction.
O rgyan gling pa (gter ston) O rgyan gu ru padma 'byung gnas kyi skyes rabs rnam par thar pa rgyas par bkod pa padma bka'i thang
yig (Padma bka' thang shel brag ma). Tibetan date given: fire monkey (me spre, 1956?). Editor for the new

Bibliography

367

publication, 'Tsho byed rdo rje rgyal mchan (mtshan?); scribe, dBu chen phun tshogs rnam rgyal; blockcarver, dBu chung tshe ring bsam 'grub and others. 515pp. Publication house not given. Printed in India.

Works in other languages

Barron, R. (trans., ed.) 2003 The Autobiography of Jamgn Kongtrul: A Gem of Many Colours. Ithaca, Boulder: Snow Lion.
Barth, F. 2002 "An Anthropology of Knowledge", in Current Anthropology Volume 43, Number 1, February 2002: 1-18. The
Wenner-Gren Foundation for Anthropological Research.
Bdier, J. 1928 "La tradition manuscrite du Lai de l'Ombre: rflexions sur l'art d'diter les anciens textes." Romania 54: 161-196,
321-356.
Canevascini, G. 1993 The Khotanese Saghastra A critical edition. Wiesbaden: Reichert Verl. xviii-304. (Beitrge zur Iranistik
Band 14).
Cantwell, C. 1997 "To meditate upon consciousness as Vajra: Ritual 'killing and liberation' in the rNying-ma-pa tradition", in H.
Krasser, M.T. Much, E. Steinkellner, H. Tauscher (eds) Tibetan Studies: Proceedings of the 7th Seminar of
the International Association for Tibetan Studies, Graz 1995, Vienna: sterreichische Akademie der
Wissenschaften, Vol.I, pp.107-118.
Cantwell, C. 2006 "Variations in Tibetan Buddhist meditations on deities: reflections on the process of generating practices",
unpublished paper presented to the UK Association for Buddhist Studies Conference: Buddhism and Popular
Culture, Lancaster, July 2006.
Cantwell, C. and R. Mayer 2007 The Klaya Nirva Tantra and the Vajra Wrath Tantra: two texts from the Ancient Tantra
Collection, Vienna, The Austrian Academy of Sciences Press.
Cantwell, C. and R. Mayer 2008 Early Tibetan Documents on Phur pa from Dunhuang, Vienna, The Austrian Academy of Sciences
Press.
Cantwell, C. and R. Mayer 2009 "A Noble Noose of Methods, the Lotus Garland Synopsis: Methodological issues in the study of a
Mahyoga text from Dunhuang", Journal of the International Association for Tibetan Studies, Issue 5
(December 2009): 1-51, http://www.thlib.org/?tid=T5696.
Cantwell, C. and R. Mayer 2010a "Continuity and Change in Tibetan Mahyoga Ritual: Some Evidence from the Tabzhag (Thabs
zhags) Manuscript and other Dunhuang Texts", in J. Cabezn (ed.), Tibetan Ritual, New York, OUP: 69-88.
Cantwell, C. and R. Mayer 2010b "Enduring myths: smrang, rabs and ritual in the Dunhuang texts on Padmasambhava", in
Franoise Pommaret and Jean-Luc Achard (eds) Tibetan Studies in Honor of Samten Karmay. Dharamsala,
Amnye Machen Institute: 289-312. [First published in 2008 as: Tibetan Studies in Honor of Samten Karmay,
Revue d'Etudes Tibtaines (Langues et Cultures de l'Aire Tibtaines, CNRS, Paris, available online:
http://www.digitalhimalaya.com/collections/journals/ret/), Vol.15, November 2008: 289-312].
Cantwell, C. and R. Mayer 2011 "The Dunhuang Thabs kyi zhags pa padma 'phreng gi don bsdus pa'i 'grel pa' Manuscript: A Source
for Understanding the Transmission of Mahyoga in Tibet. A Progress Report", in Yoshiro Imaeda, Matthew
T. Kapstein and Tsuguhito Takeuchi (eds), New Studies of the Old Tibetan Documents: Philology, History
and Religion, OTDO Monograph Series Volume III. Tokyo: Research Institute for Languages and Cultures
of Asia and Africa, Tokyo University of Foreign Studies: 273-292.

368

Bibliography

Cantwell, C., R. Mayer and M. Fischer 2002-3 The Rig dzin Tshe dbang nor bu Edition of the rNying mai rgyud bum: An
Illustrated Inventory, CSAC (University of Kent) in association with The British Library.
http://ngb.csac.anthropology.ac.uk/Title_page_main.html.
Chattopadhyaya, Alaka. 1972 Catalogue of Kanjur and Tanjur: Vol. 1: Texts (Indian Titles) in Tanjur. Calcutta: Indo-Tibetan
Studies.
Cppers, C., R. Mayer and M. Walter (eds) forthcoming Tibet after Empire: Culture, Society and Religion between 850-1000.
Proceedings of the conference Between Empire and phyi dar: the fragmentation and reconstruction of society
and religion in post-imperial Tibet, March 1-4, 2011, Lumbini International Research Institute. Lumbini:
LIRI.
Dalton, J.P. 2004 "The Early Development of the Padmasambhava Legend in Tibet: A Study of IOL Tib J 644 and Pelliot tibtain
307" Journal of the American Oriental Society 124, 4: 759-772.
Dalton, J.P. 2011 The Taming of the Demons: Violence and Liberation in Tibetan Buddhism. New Haven and London: Yale
University Press.
Das, S.C. 1976 (1902) A Tibetan-English Dictionary. Delhi: Motilal Banarsidass.
Davidson, R. 1982 "The Litany of Names of Manjusr: Text and Translation of the Majurnmasagti", in M. Strickmann (ed.)
Tantric and Taoist Studies in Honour of R. A. Stein, Mlanges Chinois et Bouddhiques, Vol. XX, Brussels,
Institut Belge des Hautes tudes Chinoises: 1-69.
Davidson, R. 2002 "Gsar ma Apocrypha: the Creation of Orthodoxy, Gray Texts, and the New Revelation", in D. Germano and H.
Eimer (eds) The Many Canons of Tibetan Buddhism, Brill, Leiden: 203-224.
Davidson, R. Indian Esoteric Buddhism:A Social History of the Tantric Movement. New York: Columbia University Press, 2002.
Dietz, S 2002 "The 'Jig rten gzhag pa in the Kanjur Manuscript of the Newark Museum", in D. Germano and H. Eimer (eds) The
Many Canons of Tibetan Buddhism, Brill, Leiden: 13-28.
Dorje, G. 1988 "The Guhyagarbhatantra and its XIVth Century Commentary, phyogs-bcu mun-sel". Unpublished Ph.D thesis,
School of Oriental and African Studies, University of London.
Dorje, G. (trans.) 2005 The Tibetan Book of the Dead, Composed by Padmasambhava, Revealed by Terton Karma Lingpa. London:
Penguin Books.
Doney, L. "The Original bKa' thang Zangs gling ma of mNga' bdag Nyang ral Nyi ma 'od zer (1124-1192)?". Paper presented at the
12th Seminar of the International Association for Tibetan Studies, Vancouver 2010.
Dudjom Rinpoche, Jikdrel Yeshe Dorje. 1991 The Nyingma School of Tibetan Buddhism: Its Fundamentals and History. Volume
One: The Translations (translated and edited by Gyurme Dorje with the collaboration of Matthew Kapstein).
Volume Two: Reference Material (Gyurme Dorje and Matthew Kapstein). Boston: Wisdom Publications.
Eastman, K.W. 1983 "Mahyoga texts at Tun-huang." Bulletin of Institute of Buddhist Cultural Studies 22: 42-60.
Eimer, H. 1997a "A Source for the First Narthang Kanjur: Two Early Sa skya pa Catalogues of the Tantras" in H. Eimer (ed.)
Transmission of the Tibetan Canon: Papers Presented at a Panel of the 7th Seminar of the International
Association for Tibetan Studies, Graz 1995.
Wien: Verlag der sterreichischen Akademie der
Wissenschaften: 11-78.

Bibliography

369

Eimer, H. 1997b "Editorial" in H. Eimer (ed.) Transmission of the Tibetan Canon: Papers Presented at a Panel of the 7th Seminar of
the International Association for Tibetan Studies, Graz 1995. Wien: Verlag der sterreichischen Akademie
der Wissenschaften: vii-x.
Eimer, H. 2002 "Structure of the Tibetan Kanjur" in H. Eimer and D. Germano (eds) The Many Canons of Tibetan Buddhism.
Leiden, Brill: 57-72.
Gray, D.B. 2009 "On the Very Idea of a Tantric Canon: Myth, Politics, and the Formation of the Bka' 'gyur." Journal of the
International
Association
of
Tibetan
Studies,
no.
5
(December
2009):
http://www.thlib.org/collections/texts/jiats/#jiats=/05/gray/
Hackin, J. 1924 Formulaire Sanscrit-Tibtain du XE Sicle. Mission Pelliot en Asie Centrale, Srie Petit in Octavo, Vol. II. Paris:
Librarie orientaliste Paul Geuthner.
Harrison, P. 1992 Druma-kinnara-rja-paripcch-stra. A Critical Edition of the Tibetan Text (Recension A). Tokyo: IIBS.
Harrison, P. and H. Eimer 1997 "Kanjur and Tanjur Sigla: A Proposal for Standardisation" in H. Eimer (ed.) Transmission of the
Tibetan Canon: Papers Presented at a Panel of the 7th Seminar of the International Association for Tibetan
Studies, Graz 1995. Wien: Verlag der sterreichischen Akademie der Wissenschaften: xi-xiv.
Helman-Wazny, A. and S. van Schaik. In press "Witnesses for Tibetan Craftsmanship: Bringing together Paper Analysis,
Palaeography and Codicology in the Examination of the Earliest Tibetan Manuscripts", in Archaeometry.
Oxford, Wiley-Blackwell.
Herrmann-Pfandt, A. 2001 Eine Quellenkunde des esoterischen (tantrischen) Buddhismus in Indien von den Anfngen bis zum 9.
Jahrhundert, unpublished Habilitationsschrift, Philipps-Universitt Marburg/Lahn.
Herrmann-Pfandt, A. 2008 Die lHan kar ma: Ein frher Katalog der ins Tibetische bersetzten buddhistischen Texte. Wien: Verlag
der sterreichischen Akademie der Wissenschaften.
Imaeda, Y. 2008 "The Provenance and Character of the Dunhuang Documents", in Memoirs of the Research Department of The Toyo
Bunko No.66. Tokyo, The Toyo Bunko: 81-102.
Jampa Samten 1994 "Notes on the bKa' 'gyur of O-rgyan-gling, The Family Temple of the Sixth Dalai Lama (1683-1706)", in Per
Kvaerne (ed.) Tibetan Studies: Proceedings of the 6th Seminar of the International Association of Tibetan
Studies, Fagernes 1992, Instituttet for Sammenlignende Kulturforskning, Oslo, 1994: Vol.I: 393-402.
Kaneko, Eiichi 1982 Ko-Tantora zenzh kaidai moku-roku (detailed catalogue of the gTing skyes rNying ma'i rgyud 'bum in Roman
Wylie transcription). Kokusho Kankkai: Tokyo.
Kapstein, M. 2000 The Tibetan Assimilation of Buddhism. New York: Oxford University Press.
Kapstein, M. 2006 "New Light on an old friend: PT 849 Reconsidered", in R. M. Davidson and C. K. Wedemeyer (eds) Tibetan
Buddhist Literature and Praxis: Studies in its formative period, 900-1400, Brill, Leiden: 9-30.
Karmay, S.G. 1988 The Great Perfection. A Philosophical and Meditative Teaching of Tibetan Buddhism. Leiden: E.J. Brill.
Karmay, S.G. 2005 Feast of the Morning Light. Senri Ethnological Reports no. 57, Osaka: National Museum of Ethnology.
Keesing, R 1981 (2nd Edition) Cultural Anthropology: A Contemporary Perspective. Florida; Harcourt Brace.

370

Bibliography

Kuijp, L.W.J. van der 2010 "Faulty Transmissions: Some notes on Tibetan Textual Criticism and the Impact of Xylography", in A.
Chayet, C. Scherrer-Schaub, F. Robin, J-L. Achard (eds) Edition, ditions: l'crit au Tibet, volution et
devenir. Munich: Indus Verlag, Collectanea Himalayica: 441-463.
Martin, D. 2006 Tibskrit Philology, April 2006 edition, 1,226 pages. Jerusalem, available at: http://tibetan-studiesresources.blogspot.com/2006/04/tibskrit-bibliography-of-tibetan.html.
Martin, D. 2007 Tibetan Vocabulary, August 29 2007 version (begun in Bloomington, Indiana, on April 10, 1987). Jerusalem,
available at: http://www.memoware.com/?screen=doc_detail&doc_id=17543&p=contributor_id^!21039~!.
Mayer, R. 1996 A Scripture of the Ancient Tantra Collection: The Phur-pa bcu-gnyis. Oxford: Kiscadale Publications.
Mayer, R. 2004 "Pelliot tibtain 349: A Dunhuang Tibetan Text on rDo rje Phur pa", in Journal of the International Association of
Buddhist Studies, JIABS Vol. 27 no.1: 129-164.
McGann, J.J. 1983 A Critique of Modern Textual Criticism. Chicago: University of Chicago Press.
Meinert, C. 2006 "Between the Profane and the Sacred? On the Context of the Rite of 'Liberation' (sgrol ba)", in M. Zimmermann,
Chiew Hui Ho and P. Pierce (eds) Buddhism and Violence: 99-130. Lumbini: Lumbini International
Research Institute.
Nitartha International 2003 Tibetan-English Dictionary of Buddhist Culture. Version 3 on CD ROM. Kathmandu: Rangjung Yeshe
Publications.
Roerich, Y.N. (edited by Y. Parfionovich and V. Dylykova) 1983-1987 Tibetan-Russian-English dictionary with Sanskrit parallels.
Moscow: USSR Academy of Sciences, Institute of Oriental Studies, Nauka Publishers, Central Department of
Oriental Literature.
Rong Xinjiang 2000 "The Nature of the Dunhuang Library Cave", in Cahiers d'Extrme-Asie 11: 247275.
Sanderson, A. 2009 "The aiva Age The Rise and Dominance of aivism during the Early Medieval Period", in S. Einoo (ed.)
Genesis and Development of Tantrism: 41349. Tokyo: Institute of Oriental Culture, University of Tokyo.
Schaik, S. van 2007 Padmasambhava: the early sources, at earlyTibet.com http://earlytibet.com/2007/06/20/padmasambhava/.
Schaik, S. van 2008 "A Definition of Mahyoga", in Tantric Studies Vol 1, 45-88. Centre for Tantric Studies, Hamburg.
Schaik, S. van. In press "Towards a Tibetan Palaeography: Developing a Typology of Writing Styles in Early Tibet", in J. Quenzer
and J-U. Sobisch (eds) Manuscript Cultures: Mapping the Field. Berlin: de Gruyter, Volume 1 of new series,
Studies in Manuscript Cultures.
Silk, J. 1994 The Heart Stra in Tibetan: A Critical Edition of the Two Recensions contained in the Kanjur. Wien 1994, Arbeitskreis
fr Tibetische und Buddhistische Studien Universitt Wien.
Skilling, P. 1997 "From bKa' bstan bcos to bKa' 'gyur and bsTan 'gyur" in H. Eimer (ed.) Transmission of the Tibetan Canon,
Vienna, The Austrian Academy of Sciences Press: 87-111.
Skilling, P. 2001 "The Batang Manuscript Kanjur in the Newark Museum: A Preliminary Report" in Annual Report of the
International Research Institute for Advanced Buddhology at Soka University, for the Academic Year 2000:
71-92. Tokyo, Soka University, March 2001.

Bibliography
Stanley,

D.P.

2005

The Tibetan Buddhist Canon: The Kangyur (Bka gyur) and Tengyur
The Tibetan and Himalayan Digital Library (THDL),
http://www.thlib.org/encyclopedias/literary/canons/index.php#essay=/stanley/tibcanons/.

371
(Bstan

gyur).

Takeuchi, Tsuguhito forthcoming "Old Tibetan Buddhist Texts from the Post-Tibetan Imperial Period (mid-9 c. to late 10 c.)", in C.
Scherrer-Schaub (ed.) Old Tibetan Studies, dedicated to the memory of R. E. Emmerick. PIATS 10, Leiden:
Brill.
Tao Tong, 2008 "The Silk Roads of the Northern Tibetan Plateau during the Early Middle Ages (from the Han to Tang dynasty) as
Reconstructed from Archaeological and Written Sources." Unpublished Ph.D thesis, The Eberhard-Karls
University of Tubingen.
Tauscher, H. and B. Lain 2008 "Western Tibetan Kanjur tradition", in D. Klimburg-Salter, Liang J., H. Tauscher, Zhou Y. (eds) The
Cultural History of Western Tibet: Recent Research from the China Tibetology Research Center and the
University of Vienna, China Tibetology Publishing House, Beijing, and Vienna: 339-362.
Timpanaro, S. (edited and translated by G.W. Most) 2005 The Genesis of Lachmann's Method. Chicago and London: University of
Chicago Press.
Tomabechi, T. 1999 "Selected Tantra Fragments from Tabo Monastery". In E. Steinkellner and C.A. Scherrer-Schaub (eds), Tabo
Studies II: Manuscripts, Texts, Inscriptions, and the Arts, IsIAO (Serie Orientale Roma, LXXXVII), Rome:
55-98.
Yeshe Tsogyal 1993 The Lotus-Born: The Life Story of Padmasambhava, Revealed by Nyang Ral Nyima ser, Translated by Erik
Pema Kunsang. Boston and London: Shambhala.
Wangchuk, D. 2002 "An Eleventh-Century Defence of the Authenticity of the Guhyagarbha Tantra", in D. Germano & H. Eimer
(eds) The Many Canons of Tibetan Buddhism, Brill, Leiden: 265-291.
Wangdu, P. and H. Diemberger 2000. dBa' bzhed: The Royal Narrative Concerning the Bringing of the Buddha's Doctrine to Tibet.
Vienna: sterreichische Akademie der Wissenschaften.
Zimmermann, M. 1998 "A Second Tibetan Translation of the Tathgatagarbhastra in the Newark Manuscript Kanjur from
Bathang: A Translation of the Early Period (snga dar)", in Transactions of the International Conference of
Eastern Studies 43, Tokyo: 3350.
Zimmermann, M. 2002a. A Buddha Within: The Tathgatagarbhastra. The Earliest Exposition of the Buddha-Nature Teaching in
India. Tokyo: The International Research Institute for Advanced Buddhology, Soka University (Bibliotheca
Philologica et Philosophica Buddhica VI).
Zimmermann, M. 2002b "The Tabo fragments and the Stemma of the Tibetan Tathgatagarbhastra", in D. Germano and H. Eimer
(eds) The Many Canons of Tibetan Buddhism, Brill, Leiden: 177-196.

INDEX
Akobhya, 53, 69, 70, 71, 72, 73, 74, 89
Amitbha, 51, 69, 70, 71, 72, 74
Amoghasiddhi, 51, 69, 70, 71, 72, 74
archaic orthography, 32, 33, 34, 44, 63, 64, 65, 99, 100
authority of lineage. See religious authority
Barth, Fredrik, 23, 24
bDud 'joms gter gsar sngon 'gro, 20
Bdier, Joseph, 18
benefits of the accomplishment, 81
Bhutanese NGB witnesses of TZ, 45
bifidity, 19, 61, 66
bKa' ma, 1, 2, 12, 92
bodhicitta, 69, 71, 72, 79, 80, 89
Bon, 7, 14
boundary variations in TZ, 35
British Library, 1, 12, 27, 100
'bru tsha, 32
bSam gtan mig sgron, 3
bskyed rim, rdzogs rim, 88
Bu ston, 2, 14, 95
Buddhagupta, 95
Buddhasamyoga, 2
buddhavacana, 18, 24
'Bum nag, 76
Byang gter, 17, 27
captivating gtor ma, 79
captivating homa, 79
captivating phur pa, 79
captivation ritual, 79
Chattopadhyaya, Alaka, 1
'Chi med srog thig, 21
Chos rgyal 'Phags pa, 2
citations of various tantras in TZComm, 84
codicology, 32
colophon, 3, 15, 50, 65, 87, 91, 92, 95
conjecture, 17
Cutting off the evil vidy mantras, 73
Dalton, Jacob, 7
Daozhen, 8
Davidson, Ronald, 7, 15
dBang c[h]en bsdus pa, 84
dbu can, 32
dbu med, 16, 32, 93
deities of the peaceful maala, 71
descriptive witnesses, 27, 29, 44, 101
destructive gtor ma, 78
destructive homa, 78
destructive phur pa, 78
destructive ritual of union, 78

dharmadhtu, 68, 69, 70, 71, 72, 75, 76, 77, 78, 79, 80, 81,
82, 85, 88
dharmakya, 69, 70, 71, 81, 95
dharmat, 69, 71, 73, 75, 81, 83
Dietz, Siglinde, 28
distributive theory of knowledge, 22, 23, 24, 25
Doney, Lewis, 93
Dorje, Gyurme, 3, 4, 48, 84, 85, 90
dPal mchog dang po, 84, 85, 86
dPal 'phreng dkar po, 84, 85
Du ma, 84, 85
Dudjom, 2, 8, 20, 21, 90
Dudjom Rinpoche, 'Jigs bral ye shes rdo rje, 2, 8, 21, 90
Dunhuang, 1, 2, 3, 5, 6, 8, 9, 10, 15, 18, 22, 34, 35, 40, 41,
45, 64, 85, 87, 92, 93, 95, 97
Dur khrod khu byug rol pa, 92
Eastman, Kenneth, 87
effigy, 5, 48, 74
eight auspicious symbols, 82
Eimer, Helmut, 2, 18, 25, 27, 28
Ekaja, 4, 76, 78
Empire, Tibetan, 8
empowerment, 5, 69
Esler, Dylan, 3
feast offering, 74
four aspects of sameness, 68
four rites, 5
garua, 74
gCig las spros pa'i tan tra, 84
Giotto Canevascini, 12
Glang chen rab 'bog, 69
Glang po chur 'jug, 84
Glang po rab 'bog, 5, 37, 84, 85
Glog gru, 84, 85
Glog sgra bas brtsegs pa, 84
gNubs chen Sangs rgyas ye shes, 3, 8
'Gos lhas btsas, 18
Grags pa rgyal mtshan, 2
Great Perfection, 4, 90
gSang ba'i rgyud, 84
gter ma, 12, 21, 24, 93, 97
'Gu hya or Gu hya ti la ka, 84
gu ru mtshan brgyad, 94
Guhyasamja, 2, 3, 18, 52
Harrison, Paul, 14, 18, 21, 28
Hayagrva, 4, 76
Hazod, Guntram, 9
Heller, Amy, 9
Helman-Wazny, Agnieszka, 32

Index
Hemis Tshoms lha khang, 26, 27
Herrmann-Pfandt, Adelheid, 1, 95
Highlighting of the Root text lemmata in the Dunhuang
manuscript, 36
Horlemann, Bianca, 9
Hkara, 76, 95
hybrid deities, 75, 76, 77, 78, 79, 80
hypearchetype b, 47
hypearchetype c, 50
hypearchetype d (hypothetical), 54
hypercorrection, 17
Imaeda, Yoshiro, 8
increasing gtor ma, 80
increasing homa, 80
increasing phur pa, 80
increasing ritual of union, 80
increasing rituals, 80
indicative errors, 10, 18, 27, 30, 45, 46, 51, 52, 62, 65, 66
interiorization, 5
Jampa Samten, 2, 28
'Jigs bral ye shes rdo rje, 20
'Jigs byed chen mo, 4, 76
'Jigs med gling pa, 16
Kamala la, 6
Kaneko, Eiichi, 27
Kanjur scholarship, 13
kplika, 3, 7, 8, 9
Kapstein, Matthew, 1, 8
Kar ma ma le, 84, 85
Karmay, Samten, 4, 7, 29, 32, 90
Kathmandu National Archives, 93
Keesing, Roger, 22
Khri Srong lDe'u btsan, 8
Khro bo bcu, 4, 65
Ki la ya bcu gnyis, 84
K la ya bcu gnyis, 85
Klong chen pa, 3, 4, 84, 85, 94
Kong sprul, 'Jam mgon, 13, 24
Kriytantra, 1
Kuijp, L. van der, 20
Lachmannian stemmatics, 18, 25
Laine, Bruno, 27
Las kyi 'phreng ba. See Kar ma ma le
lHan kar ma, 95
lHo brag lHa lung, 29
ma dpe, 17, 24, 61, 67
magical hybrid deities, 74
mahmudr, 71, 74, 82
Mahyna, 14, 19, 21, 68, 88
Mahyoga, 1, 3, 4, 5, 6, 7, 11, 12, 14, 44, 84, 85, 90, 94, 95
Eighteen Tantras, 1, 3, 4, 5, 85

373
Man ngag lta phreng, 11, 87
maala of the Great Captivator, 40, 82
maala of the wrathful male deities, 74
Maju rmitra, 95
mantra table, mantroddhra, 22
mantra table, mantroddhra, 3, 22
mantra table, mantroddhra, 22
mantras, 73, 74, 77
Martin, Dan, 95
Masoretic Bible, 19, 20
McGann, Jerome, 20
Mi pham, 2
mTsho skyes rdo rje, 92, 94
mudrs, 77
naturally existent maala, 63, 70
Nepal Research Centre, 16
Newark Museum, New York, 26, 28
nidna, 3, 21
nirmakya, 69, 70, 71
Nyang ral Nyi ma'i 'od zer, 4, 22, 92, 93, 94, 95, 98
O rgyan gling pa, 94
open recension, 18, 19, 20, 35
ordinary beings (sems can phal ba), 87
pacifying gtor ma, 81
pacifying homa, 81
pacifying phur pa, 81
pacifying ritual, 81
pacifying ritual of union, 81
Padma bKa' thang shel brag ma, 94
Padma 'byung gnas, 93
Padma rgyal po, 83, 93, 94
Padma rGyal po, 92, 93, 94
Padmasambhava, 5, 7, 8, 11, 12, 24, 41, 87, 91, 92, 93, 94,
95, 96, 97
par yig, 29
'Phang thang ma, 92, 95
'phra/phra men ma, 76, 78
Phur chen, 76
phur pa, 5, 13, 15, 22, 24, 37, 78, 79, 81
Phur pa bcu gnyis, 22, 84, 85
phyi dar, 7
Phyogs bcu'i mun sel, 4, 94
Pith Instructions on the Garland of Views, 87, 88, 89, 90, 91
pleasures of the five senses, 69
Prabhahasti, 95
pratyekabuddha, 87
printing, 20
PT 849, 3, 8, 85
Ral gcig ma, 4
Rampant Elephant, 5
rang bzo, 21, 96

374
Ratna Gling pa, 16
Ratnasambhava, 51, 69, 70, 71, 72
rDo rje bkod pa, 84
rDo rje phur bu chos thams cad mya ngan las 'das pa'i rgyud
chen po, 17
rDo rje Phur pa, 13
rDo rje phur pa rtsa ba'i rgyud kyi dum bu, 13
rDo rje sems dpa'i le'u, 37, 84, 86
rDzogs chen, 4, 85
regional features of NGB transmission, 16, 17, 24, 65
religious authority, 20, 24
representational maala, 70
rGyal po'i rtog pa, 84
rGyud gsang ba'i snying po, 1, 4, 6, 12, 20
Ri bo brtsegs pa, 84, 85, 86
Rin chen bzang po, 7, 18
Rin chen gter mdzod chen mo, 93, 94
Rong Xinjiang, 8
Rong zom Chos kyi bzang po, 3, 4, 15
rTse gcig bsdus pa, 84, 85
Sa rba 'bu ta 'i don phyi ma, 84, 85
Sa skya, 2, 76
aiva, 3
samdhi maala, 70
Samantabhadra, 89
Samantabhadr, 71, 89
samaya, 59, 69, 74, 76, 77
Sambhava, 5, 95, 96, 97, 98
sambhogakya, 69, 70, 80
Sanderson, Alexis, 3, 8
Saghastra, 12
Sangs rgyas thams cad dang mnyam par sbyor ba series, 85
ntarakita, 6
ntigarbha, 5, 95, 96, 98
Sarvabuddhasamyogakinjlaavara, 3
Sarvatathgatatattvasagraha, 3, 4
Schfer, Peter, 19
Schaik, Sam van, 32
secret mantra, 68
Seven Syllable Supplication (tshig bdun gsol 'debs), 92
sexual yoga, 71, 72
sGron ma brtsegs pa, 37, 73, 84, 85, 86
sGrub pa bka' brgyad, 4
sGyu 'phrul dra ba, 84, 85
sil bu'i dus, 7, 8
Silk, Jonathan, 14, 18, 19, 21
Sihamukh, 48, 74
Smtijnakrti, 7
snga dar, 6, 7
snying rje rol pa, 84
South Central Tibetan NGB witnesses of TZ, 46

Index
rvaka, 76, 82, 88
r Heruka, 4
rparamdya, 2
Stanley, Phillip, 29, 30
Stein, Rolf, 15
Stein, Sir Aurel, 1
stemma codicum, 10, 43, 99
Stemmatic analysis, 18, 44
Sznt, Pter-Dniel, 7
Tabo, 18
Tang, 7, 9
Tao Tong, 9
Tauscher, Helmut, 27
techniques of alphabetic and verbal calculation, 21
ten bodhisattva levels, 68
Tenjur witnesses to TZ, 46, 47
the great self, 79
the supreme sacred offering, 70
the Ten Perfections, 68
the three maalas, 68, 69, 70, 72
the three samdhis, 88
Them spangs ma, 1, 2, 18, 28
Thini, 16, 27
thirteenth bhmi, 90
three buddha bodies, 69, 71
Three Realms (Sanjie) Monastery, 8
Time of Fragments, 6
Timpanaro, Sebastiano, 18
Tomabechi, T., 18
tradent, 21
transferance of consciousness, 78
Tshal pa Kanjur tradition of TZ, 44
Tshal pa kun dga' rdo rje, 28
Tshogs chen 'dus pa, 4
Two Accumulations, 68
unitary redactorial moments, 13, 19
unitary texts, 13
Vairocana, 3, 5, 39, 45, 51, 53, 56, 68, 69, 70, 71, 72, 73, 74,
83, 89
Vajra Vehicle, 68, 87, 88
vajradhtu maala, 70
Vajragt, 71
Vajraklaya, 4, 20, 21, 22, 75, 76
Vajralsy, 71
Vajraml, 71
Vajranty, 71
Vajrapi, 39, 74
Vajrasattva, 3, 5, 37, 38, 39, 41, 45, 68, 72
Vehicle of Characteristics, 68, 87
Vijaya, 76
Vimalamitra, 95

375

Index
wrathful female maala deities, 76
xylograph, 10, 16, 17, 19, 23, 24, 29, 31, 40, 64
Yoganiruttara, 3, 7
Yogatantra, 4, 95
Yogintantra, 3, 7

Zangs gling ma, 93, 94

Zhags pa'i rgyud phyi ma, 84
Zimmerman, Michael, 27
Zla gsang thig le, 85