M. A. K. Halliday
Continuum
On Grammar
On Grammar
M. A. K. Halliday
Edited by Jonathan Webster
Continuum
The Tower Building, 11 York Road, London, SE1 7NX
370 Lexington Avenue, New York, NY 10017-6503
First published 2002 by Continuum
Reprinted 2003, 2005
M. A. K. Halliday 2002
All rights reserved. No part of this publication may be
reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying,
recording or any information storage or retrieval system,
without permission in writing from the publishers.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the
British Library.
ISBN 0-8264-4944-1 (hardback)
Typeset by SetSystems Ltd, Saffron Walden, Essex
Printed and bound in Great Britain by CPI Bath
CONTENTS
Preface
vii
Acknowledgements
ix
17
21
37
95
106
118
127
155
158
173
contents
196
219
261
289
291
323
352
369
384
Bibliography
419
Index
433
vi
PREFACE
On Grammar
Linguistic Studies of Text and Discourse
On Language and Linguistics
The Language of Early Childhood
The Language of Science
Computational and Quantitative Studies
Studies in English Language
vii
preface
viii
ACKNOWLEDGEMENTS
acknowledgements
INTRODUCTION:
A PERSONAL PERSPECTIVE
on grammar
what I did stress was how much linguistics had in common with other
scholarly pursuits and, when it came to asking questions about
language, I always found myself lining up with the outsiders. It seemed
ne for us, as linguists, to determine the content or domain of our
own discipline sociologists studied society, psychologists studied . . .
whatever psychologists do study, linguists studied language.1 Thats
what we were there for. But it did not seem ne, to me at least, for us
to determine what questions should be asked about that domain. I was
interested in what other people wanted to know about language,
whether scholars in other elds or those with practical problems to be
faced and solved including that undervalued and under-rewarded
group who have to be both scholars and practical problem-solvers,
namely teachers.
There was nothing surprising about this last perspective: not only
had my parents both been teachers, but I myself had taught languages
for thirteen years before transferring myself into a teacher of linguistics.
But even before that, while still at school, I had been trying to nd out
about language because I was keen on literature and wanted to
understand why its language was so effective, what was special about
it. There is no separating ones personal history from the academic
paths one pursues, nor any way of detaching cause from effect in
explaining ones chosen approach to a eld of study. One way or
another, I have always found myself asking the kinds of questions about
language that arose, as it were, from outside language itself.
Of course linguists always have been located, and located themselves,
within some broader context; there is nothing unusual about that. But
at any given moment in space-time, there are likely to be only a few
predominant motifs by which the context of linguistic scholarship is
dened. This may even be legislated from on high, as when Stalin
writing in 1950 (or Chikobava, writing on his behalf) instructed Soviet
linguists to get on with the job of demonstrating the linguistic unity of
the Slav nations reasonably enough, since the Soviet Union had just
taken them all over. Usually it is determined by less overtly political
factors: by particular social movements and demands, or notable
advances in knowledge in some other eld.
The present era provides a noteworthy example of the latter. Since
about 1985 there has been spectacular progress in the eld of neuroscience; the combination of new technology positron-emission
tomography, magnetic resonance imaging and its derivatives with
new insights in evolutionary theory and its contributing disciplines has
transformed the way we understand the human brain, how it has
2
a personal perspective
evolved in the species and how it develops in the individual from birth
(and before) to maturity. And this new understanding has radically
redened the place of language. It is now clear that language and the
brain evolve together, and that these develop together in infancy and
childhood. The development of the brain is the development of the
ability to mean; as in every aspect of human history, so in the ontogeny
of the individual human being the material and the semiotic interpenetrate, as complementary aspects of the characterology of the species
(McCrone; Edelman; Deacon; Dawkins; Jones).
To say this is not to proclaim that the human species is unique in
this respect or that no other species has evolved, or could evolve, a
similar type of higher-order semiotic. On the contrary. The work of
Duane Rumbaugh and Susan Savage-Rumbaugh has brought out the
point that the bonobo chimpanzees can operate with sets of arbitrary
symbols in a way that is analogous to our own system of wording
(lexicogrammar). They lack an analogous vocal apparatus, but that is
beside the point. It is tempting to assume that they have been following
the same evolutionary path and are simply less far advanced along the
way this is the assumption that prompts questions like what age have
they reached, in terms of a human child? But this assumption is
probably wrong, or at least misleading, if it is used to describe an adult
chimpanzee in terms of an immature human; the adult bonobos brain
is fully wired up in terms of the construing of experience, and enacting
of social relations, that constitute bonobo culture. The question can
fairly be asked about bonobos brought up from birth in a human-like
semiotic environment, like Kanzi, and it is too early to say yet whether
Kanzi and the other youngsters development of the power of meaning
tracked that of human children and stopped at a certain level or
whether it was proceeding along a somewhat different route.
It might be argued that such new knowledge about how the brain
functions, and how it evolves and develops, has no signicance for the
way linguists describe and explain language, especially at the inner
strata of lexicogrammar and phonology (wording and sounding). Possibly; although even here it seems to me to set certain constraints and
more importantly perhaps to favour certain explanations over others. It
suggests systems thinking rather than compositional thinking (Matthiessen), grammatical logic rather than formal logic (Sugeno 1995),
fuzzy and probabilistic categories rather than clearly bounded and
deterministic ones. Since the brain is more like a jungle than like a
computer (Edelman 1992), it disfavours representations of grammar and
phonology that are inuenced, however indirectly or subconsciously,
3
on grammar
a personal perspective
on grammar
but you have to know language in order to understand them. Computers are built to a logic derived by design from grammar; they will
have to think grammatically if they are going to advance any further.
And while language is subject, like everything else, to the laws of
physics, the laws of physics are themselves construed in language in a
specially designed form known as mathematics, which evolved as the
language of measurement.
The brain, in other words, is only one of many phenomena that can
serve as the point of vantage from which language is viewed and
explained. It is one that happens to be particularly favourable just at
present, because of the success in brain science. But any other perspective literary, social, physical, logico-philosophical or whatever is
equally valid and language will look somewhat different from each of
these different vantage points. Some will obviously be more relevant
than others for particular research applications: an audiologist, for
example, looks at language as a physical system (i.e. system-&-process),
taking account of the physical properties of the sound wave; and again
there is a special branch of linguistics, speech science, where knowledge
about language as a physical system is one of the central concerns. The
fragmentation of linguistics into a family of subdisciplines reects and
institutionalizes these various angles of approach. If we take it that,
whereas branches of technology deal with different parts of a system,
or different stages of a process, branches in science tend to deal with
different aspects of one of the same system-&-process, then it is in
linguistics that this tendency reaches its furthest point.3
I used to think that language, or at least the core layers of language,
lexicogrammar and phonology, would have to be modelled and
described differently in all these different contexts, at least for purposes
of different applications or different research goals. This was the view
expressed in Syntax and the consumer (Halliday 1964). This approach
was partly taken as a defence against the dominant elite, for whom
linguistics was a branch of theoretical psychology (Chomsky) in
the words of Ross I take it for granted that the goal of linguistics is
[sic] to explicate the difference between the human brain and that of
an animal. I was taken to task for suggesting that there might be more
than one way of modelling and describing a language (Wales).
My problem was, however, that I could not concentrate my vision.
Unlike Sydney Lamb, who chose his point of vantage and then stuck
to it, I was constantly jumping around to see what language looked
like when viewed from the other side. To the extent that I favoured
any one angle, it was the social: language as the creature and creator of
6
a personal perspective
on grammar
the century were technological ones: the invention of the tape recorder
and the evolution of the computer. The tape recorder made it possible
to record natural speech. The computer made it possible to process
large quantities of data. The two together have given us the modern
computerized corpus, with natural speech as a signicant component,
on which we can undertake quantitative analyses on a statistically
signicant scale. As a bonus, the computer enables us to test our
descriptive generalizations, through text generation and analysis (parsing), and to observe and represent sound waves in a wealth of
complementary perspectives.
These resources have transformed (or at least are in process of
transforming) the way language looks from the inside. Patterns are
being revealed that we have known must be there, because there was a
gap where the approaches from the lexical and the grammatical poles
of the lexicogrammar converged, but which we could not see: the
nature of grammatical logic is beginning to be understood; the semogenic (meaning-creating) power of discourse is coming into view, both
in monologic and in dialogic mode; quantitative mechanisms of linguistic change are beginning to appear on the agenda. From all this it
should be possible in the next decade or two to crack the semiotic
code, in the sense of coming fully to understand the relationship
between observed instances of language behaviour and the underlying
system of language something that has eluded us up till now, so that
we have even turned the two into different disciplines, calling only one
of them linguistics and labelling the other pragmatics.
Some people will feel threatened by this new understanding. We
know this because there are those who already do. To bring to light
the systems and processes of society is already threatening enough, as
witness the panic reactions to Bernstein thirty years ago when he
demonstrated how social class structures are transmitted, but semiotic
systems and processes are even nearer the bone. As long as linguists
conned their attention to dead languages, codied texts or sanitized
examples like John kissed Mary and Its cold in here, no one would feel
really at risk. But when grammar extends to the study of the meaningcreating power of everyday real life talk, it starts to become dangerous.
Some people feel worried that the grammarian is someone who knows
what they are going to say next and even if they can be reassured that
that is not what theory is about, it is scarcely less threatening (apparently) to be told what proportion of positive to negative clauses they
are going to use in their speech. And for others, just to be faced with a
record of real life conversation can be unnerving; they feel embarrassed
8
a personal perspective
on grammar
a personal perspective
on grammar
a personal perspective
based on the (theoretical) category of system, for matching up descriptive categories across languages but they are not claimed to be universal,
and no grand hypothesis stands or falls by their universality. The
unity of human language, and its relation to the human brain, is
proclaimed by the multifaceted architecture of the theory.
A volume of typological studies organized around the theoretical
category of metafunction will serve to illustrate this standpoint (Caffarel, Martin and Matthiessen 2002). My own interpretation of the
grammar of modern English will be found in An Introduction to
Functional Grammar (Halliday 1985 and later editions). Other descriptive
papers on English and on Chinese will be presented in Volumes 7 and
8. A theory-based account of the ideational semantics of English is in
Construing Experience through Meaning (Halliday and Matthiessen 1999).
I doubt whether any of the present volumes would have appeared
without the enthusiasm, energy and efciency of my editor, Dr
Jonathan Webster, of the City University of Hong Kong. He brought
the whole project to life, convincing me that it was worthwhile and
convincing the publishers that it could actually come to fruition. It has
been a pleasure being driven along by his momentum.
My thanks also to the publishers, especially to Janet Joyce, who
despite years of my ineffectual attempts to get started never lost patience
with me or faith in the enterprise, and to Robin Fawcett, who set the
whole thing going and provided many rounds of valuable suggestions
and advice.
Notes
1. Psychologists, in fact, study psychology the domain is dened by the
discipline, rather than the other way round. Hence the rather odd locutions
like criminal psychology, meaning the mind-set, or psyche, of criminals,
rather than psychological theories that criminals have devised. I was once
put down rather scathingly by a psychologist for suggesting that their
domain of study might be the human psyche.
2. See Volume 4 in this series.
3. Language is a system of meaning (a semiotic system); and semiotic
systems are of the fourth order of complexity, being also physical and
biological and social. This means that one and the same linguistic phenomenon (whether a language or a single utterance by one speaker) will
appear in all these various guises.
4. I hope it will be clear that I am not seeking either to justify this approach
or to apologize for it. These bits of personal history are brought in simply
to provide a context, to explain the way the papers in these volumes
13
on grammar
wander throughout the highways and byways of language. If there has
been any consistent motif, it has been now how would this (phenomenon
and its explanation) seem to someone who is interested in language for
some other reason, different from the one that prompted me to explore
it?
5. Strictly speaking, of course, it is semiotics; but semiotics has not yet
evolved into a general theory of meaning and it seems likely that, for the
time being at least, the way forward is by extending linguistics into other
semiotic systems. I use meaning rather than the term information (the
term imported from those who work on matter) because information is
only a sub-class of meaning; it is the part that can be measured, whereas,
unlike matter, meaning in general is not open to measurement (though
systemic linguistics offers one way in; see Volume 5).
6. We are of course accustomed to linguistics being dismissed in this offhand
way: linguists always . . . (or . . . never . . .), so you neednt bother yourself with what they write. This is irksome but does little real harm
linguists will go on writing anyway. What I am talking about here is the
assertion that language has no relevance for example to social and
political processes, and to anyones intervention in them. Such assertions
can do a great deal of mischief.
7. I have, alas, no tape recordings of my grandmother, who died in 1959, in
her mid-nineties. She belonged to the last generation, within my own
culture, who spoke unselfconsciously in proverbs. A proverb was a theory
of experience, but it was a commonsense theory, not a designed theory,
and so construed in commonsense grammar, as one of a class of instances
rather than a higher order abstraction. A snatch of dialogue might run like
this (the example is invented):
Harrys no good; hell never carry corn. That business of hisll never
thrive, believe me.
I dont know; he might pull through. And Maggies certainly trying
to buck him up a bit; shes set her mind to that.
She cant change him, however hard she tries. You cant make a silk
purse out of a sows ear. Itd take more than Maggie to make
anything out of him.
A task for the grammatics is to show the relationship between the
proverbial construct and the remainder of the discourse.
8. They constitute, in Firths formulation, a general linguistic theory applicable to particular linguistic descriptions, not a theory of universals for
general linguistic description (Firth 1957: 21).
14
SECTION ONE
EARLY PAPERS ON BASIC CONCEPTS
EDITORS INTRODUCTION
editors introduction
ordered repetition of like events that make up the patterns. There are
both primary and secondary structures, distinguished in terms of delicacy, or
depth of detail. Whereas class involves the grouping of like events by
their occurrence in patterns, system deals with the occurrence of one
rather than another among a number of like events. To help the reader
better understand the application of the categories of grammar, Halliday
presents a framework of categories for the description of another very
familiar kind of patterned activity, namely, eating a meal. Looking back
on this chapter after forty years, Professor Halliday provides some
background from his personal history to help readers better understand
his very careful concern for assigning things to categories:
Struggling with the grammar of Chinese, and then of English, in the
conceptual-categorial frameworks which were then available (traditional
grammar, linguists descriptions of languages, Jespersen and Wang Li,
Firths systemstructure theory, Pike, Fries, Hill, Hockett, etc.), I was
constantly nding that the categories were unclear: you would nd a
label attached to some patch or other, but with no indication of what
kind of category it was supposed to be and the whole battery of
technical statements never added up to a coherent picture of the whole.
I felt I needed to know where I was at any moment and where any
descriptive statement that I made tted in to the overall account.
20
Chapter One
21
Comparative
22
Historical
Comparative
Universal
Historical
Evolutionary
the text other criteria, phonetic or graphic, may contribute and may be
taken, where to do so is compatible with the general aims of simplicity
and comprehensiveness, as the primary or even sole criteria (for
example, punctuation or spacing in a written text, features of intonation
of pause in a spoken text, all of which then enter into the grammatical
description). It is probable that in the description of any language at
least two units will be required: these would be such as could be named
the sentence and the word.3
A descending order of procedure seems preferable not only for the
presentation (where indeed it may be varied for a particular purpose),
but also for the analysis of the grammar, where in such a hierarchic
progression the classication made at the level of each unit will itself
determine the classes that are to be set up for the lower units. One may
begin by establishing, and delimiting the exponents of, that unit (which
we may call the sentence) which, while within the scope of grammatical statement not so extensive as to be incapable of systematic
analysis is yet enabled to operate as the linguistic action of participants
in a situation: which is, in fact, living language and constitutes the
unit of analysis at the contextual level. In the subsequent establishment
and classication of the lower units, the statements made about each
unit will be related to values set up in the structure of the higher unit.4
For each unit there will then be set up systems of classes, formally
established in the grammar and exhaustive, such that statements may be
made which are valid for all exponents of a given unit. These classes
are set up independently of structure: that is to say, a unit (for example
clause) having been established, it is then classied by reference to
various sets of formal criteria (for example presence of, or ordering of,
certain formally dened elements); each set of criteria permits the
establishment of one system of classes (clauseclasses). Such mutually
independent systems of classes of any unit are referred to as dimensions: thus one dimension of clauseclasses might be the aspect
dimension, with a system, say, of two terms, perfective and imperfective.
There may be any number of dimensions of classes for each unit, and
the system of any dimension may admit of a neutral or unmarked
term, but each dimension will by itself form an exhaustive system of
classication such that every exponent of a given unit may be placed
within it. Thus it might be that all words are either red or blue or
yellow and all words are either square or round or neutral in shape.
Two dimensions of classes are implicit in the taxonomic hierarchy of
the unit system: as characteristic of each unit except the lowest there
appears the dimension compound / simple, while for each unit except
26
27
29
While perhaps modern general linguistics would recognize the establishment of categories within the language under description itself as
the basis of a particular description, reference to the forms of another
language, including the language of description, may be made without
infringing the requirements of formal analysis. In any other than a
monolingual description of a language text there arises in any case the
specic problem of the relation of the forms of the language under
description to the language of description. Since language is used to
describe language, if in a formal descriptive grammar it is desired
to exclude from consideration as far as possible all forms that do
not belong to the l.u.d., the nearest approach is that outlined above,
where the aim is achieved through the creation of a metalanguage
whose terms, whatever the context of their previous usage, are to be
taken as dened only with reference to the text under description. All
identication of categories either comparative or universal is thereby
excluded.
It may, however, be desirable in a given instance or for a given
purpose to relate the forms of the l.u.d. to particular forms of the l.o.d.
The two languages may then be seen to impinge on one another at
various points. At one extreme, it is possible to make a descriptive
grammar of the l.o.d. using the same procedures as are applied to the
l.u.d. and subjecting the two descriptions to a systematic comparison
(as envisaged by Allen 1953: 88 ff.). A comparison of the systems of
categories would rst establish what categories were comparable; the
latter would then be compared so as to permit the identication of
terms within the system of each category.
At the other extreme, in dealing with a limited language text it may
be possible to make a complete translation of the text under description,
a translation of the type that may be characterized as contextual one
such as to reproduce, as nearly as possible, the creative effect in the
given situation of the original. This may be useful where the l.o.d. is
not itself described and the linguist does not consider it to be within
his terms of reference to make a description such as would permit a
systematic comparison. It may then be found that certain categories of
the l.u.d. show a regular translation equivalence to certain unsystematized but formally dened elements in the language of the translation
(for example, to one term in the system of clauseclasses in the l.u.d.
might correspond regularly in English translation a verb form in -ing).
Such a form of statement, limited though its application, may some30
Chinese
English
wo
33
tamen
they
22,2(2)3(3)
nmen
ta
women
12(2), 12(2)3(3)
zamen
}
{
}
you
he
she
we
Contextual reference is as follows: 1, speaker; 2, addressee; 22, addressees; 3, other person; 33, other persons.
Since what is under consideration here is the use of comparison in
particular description (and not a comparative study as such), no calculation of degree of relationship is necessary. Identication has been
permitted between terms operating in systems with different numbers
of terms: in any calculation of degree of relationship this point would
of course be crucial. Prima facie, such identication would seem justiable
where the criteria for the identication were contextual. The identication of terms by grammatical criteria in a comparative study poses
separate problems which are touched upon briey in the next section.
31
Descriptive comparison
In a descriptive comparison there is no implication of genetic relationship; but beyond the application to particular description permitted by
comparison of this type (since any l.u.d. may be compared with any
l.o.d.) systematic comparison itself connotes a wider purpose, seen
initially in the establishment of degrees of relationship. Here the grammatical identication of terms across languages is an essential basis of
comparison; and while much can be done on contextual criteria (though
not as much as in the establishment of lexical systems for comparative
purposes) some formal procedure must be found in order that statement
of degrees of relationship may assume a general signicance.
Possibly, however, we should not impose a strict demarcation
between identication of grammatical terms on grammatical and on
non-grammatical (e.g. contextual) criteria. Allen, rightly rejecting both
terminological identication and identication by translation, considers
that identication by formal grammatical criteria (as opposed to the
situationalcontextual criteria which are available for the semantic
identication of lexical items) seems not to be possible (Allen 1953: 99,
100). Some attempt at identication may be made with the Chinese
dialects, where speakers of more than one dialect constantly make such
identications in practice, with or without phonological resemblances:
for example, from the personal pronouns, a Cantonese speaker equates
nei with Pekingese n and, as readily, keudei with Pekingese tamen. This
identication in practice demonstrates the contextual basis of the identication; its validity on grammatical criteria may be tested by reference
to the place of the class pronoun in the system of wordclasses.
When we nd that it is possible to describe both Modern Pekingese
and Modern Cantonese in terms of the same units of sentence, clause,
word and character; that at the word level we can set up for both a
three-term system of classes, verbal, nominal and adverbial, and that
one term in the system of nominal wordclasses, the auxiliary noun,
enters in both into the noun group with identical position and value in
the structure, we can regard the class of auxiliary noun in the two
dialects as comparable, and are then justied in seeking to identify
32
Cantonese
yat zek gau a dog (in general context or given)
yat zek gau one dog (in specic context or new)
zek gau
the dog
The non-identication of the rst and third instances will not prevent
the contextual identication of the second if the total spread within that
sub-system can be shown to be the same.10
Some such procedure, for the identication of terms in grammatical
systems, once established, the eld of application of descriptive comparison is seen to be very wide. A comparison of systems in two
languages such as Chinese and English, which at rst sight seems to be
of little purpose beyond its application to the particular description of
one of the two, provides material for a formal systemic typology when
compared with other such comparisons for, if with Allen relationship
is not of languages but of systems, the signicance of such relationship
is that typology is not of languages but of systems. But without going
so far from the scope of comparative studies as usually envisaged, one
may seek a eld for the application of systematic comparison both in
the comparison of languages considered to be genetically related and in
the study of languages in geographical proximity where there is no
material for genetic groupings.
With regard to the former, any discussion of the nature and implications of genetic relationship may be avoided by the choice of a simple
instance of systematic comparison in the phonology of the dialect areas
of Chinese. In the Modern Chinese dialects, in the system of nasal nals
in the phonological unit the syllable may have two or three terms, or
33
there may be only one nasal nal. If we compare these systems in certain
dialects of the Mandarin, Wu and Yueh dialect groups, we nd that it
is prosodically linked, and identical in number of terms, with the system
of plosive nals or, if the latter is absent, with that of vocalic nals.
Cantonese shows the nal plosive system p/t/k with prosodically
identical nasals m/n/; in Shanghai, where there is a single nal plosive,
the glottal closure, there is one nasal nal, with varying point of contact;
while in Pekingese, where there are no nal plosives, the two terms of
the nal nasal system are prosodically identical with the oral nals, so
that n:::i:u.11 The calculation of the degree of relationship will depend
on the number of nal systems set up for each dialect and the number
of these in which all terms can be identied; but the comparison
immediately suggests the historical interconnection between the disappearance of the nal plosive system and the elimination of one term
from the nal nasal system in the dialects where these have occurred.
With regard to the study of languages in geographical proximity in
an area where no, or only partial, genetic groupings have been established, systematic comparison can determine whether the question of
grammatical afnity is to be posed at all, and if so in what form. The
supercially apparent afnity among the languages of certain areas has
long been the basis of traditional typology. An instance of how it may
be demonstrated or disproved, initially as a function of systems, might
be found in the grammar of East Asian languages. It seems possible to
set up in, for example, Pekingese, Cantonese, Vietnamese, Siamese and
Malay systems of nominal wordclasses which would permit the identication as comparable of certain terms, including the category discussed
with reference to Pekingese and Cantonese above and which may be
called the auxiliary noun. If the place of the auxiliary noun in the system
of wordclasses, and the terms in the various systems of the auxiliary
noun, can be compared as between these languages on the basis of
adequate criteria, we may determine whether or not a formal description
of these languages would reveal for these and other systems anything
that could be regarded as grammatical afnity such as, that is, to
exclude from such afnity any systems set up for other languages in
other areas. Only if such afnity could be established among a signicant
number of systems in the grammar of these languages would it be
possible to raise further questions of general grammatical afnity and
as yet only an interesting speculation of geographical gradation.
Such a systemic comparison may help to resolve the difculty that,
on the one hand, there appears an obvious but unformulated grammatical similarity among the East Asian languages as contrasted, for example,
34
with the Indo-European languages, while on the other hand there exists
the quite justiable scepticism among linguists either as to how this is
to be explained, in comparative historical terms, where there are no
correspondences, or as to whether there is anything to be explained (or,
in systematic terms, to be stated comparatively) at all.
Notes
1. The study of the form of the content by plerematics, as envisaged by
Hjelmslev, would set up particular or comparative (i.e. non-universal)
categories; but the criteria for such categories, while as yet inadequately
dened, would also be such as to be regarded as formal-linguistic. A
method of classication of words on the basis of universal categories of
relation and description is to be found in the work of Brndal (1948).
2. Seventh International Congress of Linguists, Preliminary Reports, London,
1952, p. 53: Can a purely formal grammatical analysis be carried out on
languages such as Chinese, in which all or nearly all the words are
invariable, and if so, on what principle?
3. Provided the principle of particular formal description is adhered to, the
choice of current terms seems preferable to the creation of new ones.
There is, however, no reason why, especially in the initial stage of the
process of analysis, completely non-committal terms should not be
employed. The practice of some linguists of talking (at least to themselves)
about e.g. red words and blue words can equally well be extended, for
example, to strings and bits.
4. Compare the descending order within the levels of linguistic analysis in
which meaning may be stated, as suggested by Firth (1951: 121). For an
example of the employment of this method of descending analysis in
grammatical statement, cf. Robins (1953) and Halliday (1956).
5. As pointed out by Robins (1953: 109), Firth has indicated how system
and structure require to be kept separate in General Linguistic theory. I
have attempted to follow Firths view of the two as distinct but related
concepts.
6. The linguistic unit of which the written character is the graphic symbol
corresponding to the syllable at the phonological level. The use of the
term character follows the Chinese practice of calling both the linguistic
and the graphic units by the single name tzu (z`).
7. The concept of context of situation here intended is as developed by
Firth in linguistic analysis, where it is best used as a suitable schematic
construct to apply to language events and should be regarded as a group
of related categories at a different level from grammatical analysis but
rather of the same abstract nature (Firth 1952: esp. 7).
8. Professor Allen has kindly drawn my attention to the paper entitled
35
36
Chapter Two
37
derives most of all from the work of J. R. Firth.4 At the same time I
do not of course imply that I think Professor Firth would necessarily
have found himself in accord with all the views expressed, which in
some places depart from his own; nor do I underestimate the debt to
my present colleagues and the many others whose work I have,
obviously, drawn on.5
No excuse is needed, I think, for a discussion of General Linguistic
theory. While what has made linguistics fashionable has been, as with
other subjects, the discovery that it has applications, these applications
rest on many years of work by people who were simply seekers after
knowledge. It would not help the subject if the success of these
applications led us into thinking that the theoretical problems were
solved and the basic issues closed.
Starting-point
1.5 The primary levels are form, substance and context. The substance
is the material of language: phonic (audible noises) or graphic (visible
marks). The form is the organization of the substance into meaningful
events: meaning is a concept, and a technical term, of the theory (see
below, 1.8). The context is the relation of the form to non-linguistic
features of the situations in which language operates, and to linguistic
features other than those of the item under attention: these being
together extratextual features.
1.6 The complete framework of levels requires certain further subdivisions and additions, and is as follows:
(a) Substance may be either phonic or graphic.
(b) If substance is phonic, it is related to form by phonology.
(c) If substance is graphic, it is related to form by orthography (or
graphology),9 either
(i) if the script is lexical, then directly, or
(ii) if the script is phonological, then via phonology.
(d) Form is in fact two related levels, grammar and lexis.
(e) Context is in fact (like phonology) an interlevel relating form to
extratextual features.
1.7 The study of phonic substance belongs to a distinct but related
body of theory, that of General Phonetics. Since phonology relates
form and phonic substance, it is the place where linguistics and
phonetics interpenetrate. Linguistics and phonetics together make up
the linguistic sciences.10
39
1.8 Language has formal meaning and contextual meaning. Formal meaning is the information of information theory, though (i) it can be
stated without being quantied and was in fact formulated in linguistics
independently of the development of information theory as a means of
quantifying it,11 and (ii) formal meaning in lexis cannot be quantied
until a method is found for measuring the information of non-nite
(open) sets (see below, 2.1 and 8.2). The formal meaning of an item
is its operation in the network of formal relations.
1.9 Contextual meaning, which is an extension of the popular and
traditional linguistic notion of meaning, is quite distinct from formal
meaning and has nothing whatever to do with information.12 The
contextual meaning of an item is its relation to extratextual features;
but this is not a direct relation of the item as such, but of the item in
its place in linguistic form: contextual meaning is therefore logically
dependent on formal meaning.13
1.10 It follows from 1.8 and 1.9 that, in description, formal criteria
are crucial, taking precedence over contextual criteria; and that the
statement of formal meaning logically precedes the statement of contextual meaning.14
1.11 Finally, it is necessary to distinguish not only between theory
and description but also between description and presentation. Presentation, the way the linguist expounds the description, varies with
purpose, and relative merit is judged by reference to the specic
purpose intended. Description depends on the theory; theoretical
validity is demanded, and relative merit is judged by reference to
comprehensiveness and delicacy.15
Grammar
Unit
when they turn up. For English, for the two units between sentence
and word the terms clause and phrase are generally used. It is at the
rank of the phrase that there is most confusion because there are here
the greatest difculties in the description of English; one reason is
that in English this unit carries a fundamental class division (see below,
5), so fundamental that it is useful to have two names for this unit in
order to be able to talk about it: I propose to call it the group, but to
make a class distinction within it between group and phrase. Below the
word, English has one unit, called by the general name for the unit of
lowest rank, the morpheme.29
So in the description of English the sentence30 consists of one or
more complete clauses, the clause of one or more complete groups, the
group of one or more complete words and the word of one or more
complete morphemes. The descriptive meaning of consists of, and
the possibilities of rankshift (including recursive rankshift), are stated as
and where applicable. One distinction that is often useful is between a
member of a unit that consists of only one member of the unit next
below and one that consists of more than one; the former may be
called simple and the latter compound, but if this is done the terms must
be kept rigorously to this, and no other, use.31
3.4 The theory requires that each unit should be fully identiable in
description. This means that, if the description is textual, every item of
the text is accounted for at all ranks, through the various links of the
exponence chain which involve, of course, the remaining theoretical
categories. If the description is exemplicatory, exactly the same is
implied, except that the description proceeds from category to exponent instead of from exponent to category.
It will be clear from the discussion in the next sections that there
can be no question of independent identication of the exponents of
the different units, since criteria of any given unit always involve
reference to others, and therefore indirectly to all the others. A clause
can only be identied as a clause if a sentence can be identied as a
sentence and a group as a group, and so on up and down the line. For
this reason description is not and can never be unidirectional: it is
essential to shunt, and shunting is a descriptive method that is
imposed on description by theory.
45
Structure
4.1 The unit being the category of pattern-carrier, what is the nature
of the patterns it carries? In terms once again of language as activity,
and therefore in linear progression, the patterns take the form of the
repetition of like events. Likeness, at whatever degree of abstraction, is
of course a cline, ranging from having everything in common to
having nothing in common. The commonplace that no two events
are ever identical, that the same thing can never happen twice, is of no
relevance whatever to linguistics; as soon as description starts, however
little the generalization involved, absolute identity is a necessary
hypothesis, which is then built into the theory, as one endpoint of the
likeness cline. Likeness, including absolute identity, is of course redened for each level and each category.
In grammar the category set up to account for likeness between
events in successivity is the structure.32 If the relation between events
in successivity is syntagmatic, the structure is the highest abstraction of
patterns of syntagmatic relations. The scale used for talking about it,
and for its graphic display, will most naturally be the orthographic
scale: to those of us brought up on the roman alphabet this happens to
run horizontally from left to right, which is enough reason for adopting
this version of the scale. But, as in the case of the unit, it must be
stressed that linear progression itself is a feature of substance. A structure
is made up of elements which are graphically represented as being in
linear progression; but the theoretical relation among them is one of
order. Order may, but does not necessarily, have as its realization
sequence, the formal relation carried by linear progression; sequence is
at a lower degree of abstraction than order and is one possible formal
exponent of it.33
4.2 A structure is thus an arrangement of elements ordered in places.
Places are distinguished by order alone: a structure XXX consists of
three places. Different elements, on the other hand, are distinguished
by some relation other than that of order: a structure XYZ consists of
three elements which are (and must be, to form a structure) placeordered, though they can be listed (X, Y, Z) as an inventory of
elements making up the particular structure.34 A structure is always a
structure of a given unit.
Each unit may display a range of possible structures, and the only
theoretical restriction is that each unit must carry at least one structure
that consists of more than one place.35 Each place and each element in
46
the structure of a given unit is dened with reference to the unit next
below. Each place is the place of operation of one member of the unit
next below, considered as one occurrence. Each element represents the
potentiality of operation of a member of one grouping of members of
the unit next below, considered as one itemgrouping.36 It follows
from this that the lowest unit has no structure;37 if it carried structure,
there would be another unit below it.
4.3 In description, structures are stated as linear arrangements of
symbols, each symbol (occurrence) standing for one place and each
different symbol (item) standing for one element. Since elements of
structure exist only at this degree of abstraction, the relation stands
for means simply is shorthand for, like that of an initial: U stands
for United. In a few cases traditional names exist which can usefully
serve as names for elements of structure, with the initial letter as the
descriptive symbol. In the statement of English clause structure, for
example, four elements are needed, for which the widely accepted
terms subject, predicator, complement and adjunct are appropriate.38
These yield four distinct symbols, so that S, P, C, A would be the
inventory of elements of English clause structure. All clause structures
can then be stated as combinations of these four in different places:
SAPA, ASP, SPC, ASPCC, etc. For one type of group we have the
names modier, head, qualier, giving an inventory M, H, Q: here, if
the total range of possible structures is H, MH, HQ, MHQ, these
possibilities can be stated in a single formula, where parentheses indicate
may or may not be present, as (M) H (Q).39
In other cases, no names come ready to hand; names can be imported
or coined, or arbitrary symbols chosen colours, for example, have
advantages over letters in presentation, though there are not enough of
them and they have to be redened in description for each unit. It is
tempting sometimes to derive the symbols from the name given to the
grouping of members of the unit next below which operates at the
given element (as if one were to put V instead of P because what
operates at P is the verbal group); but it is important to avoid
identifying this grouping, which belongs to a different category as well
as a different rank, with the element itself therefore if this method is
to be used at all it must be used all the time and a statement made to
cover it.40
There are some instances where an element of structure is identied
as such solely by reference to formal sequence: where the element is
dened by place stated as absolute or relative position in sequence. It
47
structures can be stated as SPCA, SAPA, ASP, etc.41 This displays the
contrast between this situation, where S is crucially dened by position
relative to P, and realized sequences of elements which are not,
however, dened by sequence, which may be indicated by simple
linearity of the symbols.42
4.4 In the consideration of the places and elements of structure of
each unit, which of course vary from language to language and from
unit to unit within a language, a new scale enters, that of delicacy (see
below, 7.4). This is depth of detail, and is a cline running from a xed
point at one end (least delicate, or primary) to that undened but
theoretically crucial point (probably statistically denable) where distinctions are so ne that they cease to be distinctions at all, like a river
followed up from the mouth, each of whose tributaries ends in a
moorland bog. Primary structures are those which distinguish the
minimum number of elements necessary to account comprehensively
for the operation in the structure of the given unit of members of the
unit next below: necessary, that is, for the identication of every item
at all ranks. (M)H(Q), and the various possible combinations of S, P,
C, A, are primary structures: one cannot account for all words in group
structure, or all groups in clause structure, with fewer than these
elements or places.
Subsequent more delicate differentiations are then stated as secondary
structures. These are still structures of the same unit, not of the unit
next below; they take account of ner distinctions recognizable at the
same rank.43 Rank and delicacy are different scales of abstraction:
primary group structures differ in rank from primary clause structures,
but are at the same degree of delicacy; while primary and secondary
clause structures differ in delicacy but not in rank.
As the description increases in delicacy the network of grammatical
relations becomes more complex. The interaction of criteria makes the
relation between categories, and between category and exponent,
increasingly one of more / less rather than either / or. It becomes
necessary to weight criteria and to make statements in terms of
probabilities. With more delicate secondary structures, different combinations of elements, and their relation to groupings of the unit next
below, have to be stated as more and less probable.44 The concept of
48
Class
XYZ, YZ and XYZY, the primary classes of the unit next below are
class operating at X, class operating at Y and class operating at
Z. If, however, there is a further restriction such that in XYZY,
which will now be (secondarily) rewritten XYaZYb, only a section of
the members of the class at Y can operate at Ya and only another
section (not necessarily mutually exclusive) at Yb, this yields as secondary classes class operating at Ya and class operating at Yb. Secondly,
with increased delicacy the elements of primary structure will be
differentiated into secondary elements. A primary structure generalized
as X . . . nYZ, of which XXXXYZ is an instance, shows a generalized
relation of X to (say) Y; but there may be internal relations within X
. . . n such that XXXX is rewritten pqrs. These will yield secondary
classes class operating at p, class operating at q, etc.
In the second place, more delicate classes appear whenever a
restriction is found which differentiates among the members of a
primary class. There may be a relation of mutual determination, or
concord, between two classes; each divides into two sections such
that a member of one section of one class is always accompanied by a
member of one section of the other class. Thus if the primary class at
X is 1 and that at Y is 2, a structure XY must have as its exponent
either 1.1+2.1 or 1.2+2.2. Secondary classes arrived at in this way in
description may be referred to distinctively as sub-classes, to indicate
that they are derived by differentiation from primary classes without
reference to secondary structures; but it is important to state that there
is no theoretical difference here. The relation between structure and
class is a two-way relation, and there is no question of discovering
one before the other. In any given instance there may be descriptive
reasons for stating the one without the other; but all structures
presuppose classes and all classes presuppose structures.
5.3 What is theoretically determined is the relation between structure and class on the one hand and unit on the other. Class, like
structure, is linked to unit: a class is always a class of (members of) a
given unit: and the classstructure relation is constant a class is always
dened with reference to the structure of the unit next above, and
structure with reference to classes of the unit next below. A class is not
a grouping of members of a given unit which are alike in their own
structure. In other words, by reference to the rank scale, classes are
derived from above (or downwards) and not from below (or
upwards).
The distinction between downward and upward movement on the
50
avoiding confusion with elements of structure. This is not only theoretically desirable, because of their different status; it has descriptive
value in that the theoretical one / one relation between elements and
classes allows for instances where two different elements of structure,
standing in different relation to each other or to a third, yield primary
classes the membership of which is coextensive: these then form a
single primary class derivable simultaneously from two elements of
structure.51 If letters are used for elements of structure, and gures for
classes, the relation between the two can be demonstrated by the use
of a colour code.
System
6.1 Up to this point the theory has accounted for three aspects of
formal patterning: the varying stretches that carry patterns, the ordered
repetition of like events that makes up the patterns and the grouping
of like events by their occurrence in patterns. What remains to be
accounted for is the occurrence of one rather than another from among
a number of like events.
The category set up for this purpose is the system.52 This falls under
the general denition of system given above (2.1). But this does not
yet state its place in grammatical theory, its relation to the other
fundamental categories.
The class is a grouping of items identied by operation in structure:
that is, what enters into grammatical relations of structure is not the
item itself considered as a formal realization53 but the class, which is
not a list of formal items but an abstraction from them. By increase
in delicacy, the primary class is broken down into secondary classes of
the same rank. This set of secondary classes now stands in the relation
of exponent to an element of primary structure of the unit next above.
This gives a system of classes. If class 1 is the primary class (say of
the group) operating at X in (clause) structure, and this has secondary
classes 1.1, 1.2 and 1.3, then 1.1, 1.2 and 1.3 form a system of classes
operating at X. X is now shown to presuppose a choice a choice that
is implied by the nature of the class (as a grouping of items) but that is
displayed rst still in abstraction, by reference to the category of class
itself.54
6.2 Systems of secondary classes thus allow the description to remain
at a high degree of abstraction while displaying at each step, each
increase in delicacy, a more nely differentiated range of choice. This
52
one never by itself entails a shift in either of the others. The reason
why rank is often confused with other scales is that there are cases
where a shift in rank does accompany a shift in something else; but this
is always by virtue of the logical relations among the categories
involved. The fact that by moving from structure to class, which is (or
can be) a move on the exponence scale, one also moves one step down
the rank scale, is due to the specic relation between the categories of
class and structure, and not to any inherent interdetermination between
exponence and rank. The descriptive relevance of keeping the scales
distinct is that it is important to be able to display what happens if one
shifts on one scale, keeping the other two constant.66
7.3 Exponence is the scale which relates the categories of the theory,
which are categories of the highest degree of abstraction, to the data.67
Since categories stand in different relations to the data, it might seem
necessary to recognize four different scales of exponence, one leading
from each category. In fact, however, exponence can be regarded as a
single scale.
In the rst place, each category can be linked directly by exponence
to the formal item: it is in fact a requirement of the theory that any
descriptive category should be able to be so linked. This may be stated
by way of exemplication, as when we say the old man is (an example
of ) an exponent of S in clause structure. This is, however, not a
description of the element S, since by relating it to its exponent at a
stage when it was not necessary to do so we should have lost generality
(cf. above, 6.3). So instead of throwing up the grammatical sponge and
moving out to lexis while this is still avoidable, the description takes
successive steps down the exponence scale, changing rank where
necessary, until (at the degree of delicacy chosen) it is brought
unavoidably face to face with the formal item.
In the second place, therefore, the step by step move from any one
category to the data can proceed via any or all of the other categories.
This is then a move down the exponence scale, and at each step, given
that delicacy is constant, one category is replaced by another (either
with or without change of rank, according to which category is
replaced by which). While therefore the categories are distinct, they
are interrelated in such a way that the relation of exponence has the
status of a single scale.
The ultimate exponent in form is the formal item. This has then to
be related, in turn, to the substance. But this relationship, though it
may also be called exponence, entails a new scale, in which the
57
nature of the abstraction is different and the formal item is now at the
other end it is itself the abstraction, and is not in any way delimited
or categorized by grammar.68
When grammar reaches the formal item, either it has said all there is
formally to be said about it or it hands over to lexis. The formal item
is the boundary of grammar on the exponence scale. It is not of course
the boundary on the rank scale: whenever the formal item is anything
other than a single morpheme (whether in closed system like seeing
(that), or in open set like pickup) the grammar can be taken further
down in rank, since it can state the structures in terms of elements
whose exponents are words and morphemes. But seeing (that) enters
into a system at group rank, while pickup emerges from the grammar as
a word, though being a lexical item it would not necessarily be an
exponent of any whole grammatical unit. Once it has been taken over
by lexis, the grammatical categories, and the grammatical exponence
scale, no longer impinge on it (see below, 8.2).
7.4 Delicacy is the scale of differentiation, or depth in detail. It is a
cline, whose limit at one end is the primary degree in the categories of
structure and class. In the theory, the other limit is the point beyond
which no further grammatical relations obtain: where there are no
criteria for further secondary structures, or systems of secondary classes
or formal items. In description, delicacy is a variable: one may choose
to describe a language without going beyond the primary degree, still
being comprehensive in rank and exponence and making use of all the
categories of the theory. Each subsequent increase in delicacy delays
the move to the exponents (cf. above, 7.3) and thus increases the
grammaticalness of the description. The limit of delicacy is set by the
means at ones disposal.
In well-described languages, such as English, any extension in
delicacy beyond what is already known requires either or both of largescale textual studies with frequency counts and complex secondary
classication based on multiple criteria, criteria which often cut across
each other and may have to be variably weighted. And, as suggested
above (6.3), a point will perhaps be reached where probabilities are so
even as to cease to be signicant69 and classes so delicately differentiated
that the description will have to decide on crucial criteria and ignore
the others,70 thus setting its own limits.
Delicacy is distinct from rank and the limit of delicacy applies at the
rank of all units, for example differentiation of clause structures and of
classes of the group. At one stage, therefore, it becomes a limit on the
58
Lexis
8.1 This section is intended merely to bring lexis into relation with
grammar, not to discuss the theory of lexis as such. As has been pointed
out (above, 3.3. and 6.3), there is no one / one correspondence in
exponence between the item which enters into lexical relations and
any one of the grammatical units. It is for this reason that the term
lexical item is used in preference to word, word being reserved as the
name for a grammatical unit, that unit whose exponents, more than
those of any other unit, are lexical items.
Not only may the lexical item be coextensive with more than one
different grammatical unit; it may not be coextensive with any grammatical unit at all, and may indeed cut right across the rank hierarchy.
Moreover, since the abstraction involved is quite different, what is for
lexis the same lexical item (that is, different occurrences of the same
formal item) may be a number of different grammatical items, so it is
59
not true that one lexical item always has the same relation to the rank
hierarchy. So that, in English, (i) a lexical item may be a morpheme,
word or group (at least); (ii) a lexical item may be assigned to no rank,
being for example more than a word but less than a whole group, or
even both more and less than a word part of one word plus the
whole of another, sometimes discontinuously; and (iii) one and the
same lexical item may in different occurrences cover any range of the
possibilities under (i) and (ii).72
This does not mean that lexical items cannot be identied in
grammar; it means that they are not identied by rank. They are
identied, as has been suggested (above, 6.3), by their being unaccounted for in systems. But it is an additional, descriptive reason
(additional, that is, to the theoretical one that lexical items lend
themselves to different relations of abstraction) for keeping grammar
and lexis apart. When the two have been described separately, the next
stage is to relate them; and it is here that the complex relation between
lexical item and grammatical unit must be accounted for. This is exactly
parallel to what was said above (7.1) about grammar and phonology;
and, of course, it applies equally to phonology and lexis, where, after
separate description, is displayed the relationship between the lexical
item and the categories of phonology.
8.2 The task of lexis can be summed up, by illustration, as that it has
to account for the likelihood of wingless green insects and for the, by
contrast, unlikelihood of colourless green ideas.73 As in grammar, we shall
expect language to work by contrasting more likely with less likely
rather than possible with impossible; but, as has often been pointed
out,74 this particular type of likelihood is not accounted for by
grammar, at least not by grammar of the delicacy it has yet attained. It
is, however, too often assumed that what cannot be stated grammatically cannot be stated formally: that what is not grammar is semantics, and here, some would add, linguistics gives up.75 But the view
that the only formal linguistics is grammar might be described as a
colourless green idea that sleeps furiously between the sheets of
linguistic theory, preventing the bed from being made. What are
needed are theoretical categories for the formal description of lexis.
It seems that two fundamental categories are needed, which we may
call collocation and set.76 The rst basic distinction between these and
the categories of grammar is that in lexis there are no scales of rank and
exponence. There is no hierarchy of units; therefore no rank scale.
There is only one degree of abstraction a set is a set of formal items
60
61
An analogy
9.1 Eating, like talking, is patterned activity, and the daily menu may
be made to yield an analogy with linguistic form. Being an analogy, it
is limited in relevance; its purpose is to throw light on, and suggest
problems of, the categories of grammar by relating these to an activity
which is familiar and for much of which a terminology is ready to
hand.
The presentation of a framework of categories for the description of
eating might proceed as follows:
Units:
Daily menu
Meal
Course
Helping
Mouthful
Unit: Daily menu
Elements of primary structure
Primary structures
Exponents of these elements
(primary classes of unit
meal)
Secondary structures
Exponents of secondary
elements (systems of
secondary classes of unit
meal)
System of sub-classes of unit
meal
E, M, L, S (early, main,
light, snack)
EML EMLS (conated as
EML(S))
E: 1 (breakfast)
M: 2 (dinner)
L: 3 (no names available; see
secondary classes)
S: 4
ELaSaM ELaM EMLbSb
EMSaLc
La: 3.1 (lunch)
Lb: 3.2 (high tea)
Lc: 3.3 (supper)
Sa: 4.1 (afternoon tea)
Sb: 4.2 (nightcap)
E: 1.1 (English breakfast)
1.2 (continental breakfast)
Passing to the rank of the meal, we will follow through the class
dinner:
62
F, S, M, W, Z (rst,
second, main,
sweet, savoury)
MW MWZ MZW FMW
FMWZ FMZW FSMW
FSMWZ FSMZW
(conated as (F(S)MW(Z))
F: 1 (antipasta)
S: 2 (sh)
M: 3 (entree)
W: 4 (dessert)
Z: 5 (cheese*)
(various, involving secondary
elements
Fad, Ma,b, Wac)
Fa: 1.1 (soup)
Fb: 1.2 (hors doeuvres)
Fc: 1.3 (fruit)
Fd: 1.4 (fruit juice)
Ma: 3.1 (meat dish)
Mb: 3.2 (poultry dish)
Wa: 4.1 (fruit*)
Wb: 4.2 (pudding)
Wc: 4.3 (ice cream*)
Fa: 1.11 (clear soup*)
1.12 (thick soup*)
S: 2.01 (grilled sh*)
2.02 (fried sh*)
2.03 (poached sh*)
Wb: 4.21 (steamed pudding*)
4.22 (milk pudding*)
63
At the rank of the course, the primary class entree has secondary
classes meat dish and poultry dish. Each of these two secondary
classes carries a grammatical system whose terms are formal items. But
this system accounts only for simple structures of the class entree,
those made up of only one member of the unit helping. The class
entree also displays compound structures, whose additional elements
have as exponents the (various secondary classes of the) classes cereal
and vegetable. We will glance briey at these:
Unit: Course, Class: entree
Elements of primary structure
Primary structures
Exponents of these elements
(primary classes of unit
helping)
Secondary structures
Exponents of secondary
elements (systems of
secondary classes of unit
helping)
J, T, A (joint, staple,
adjunct)
J JT JA JTA (conated as
J((T) (A)))
J: 1 (esh)
T: 2 (cereal)
A: 3 (vegetable)
(various, involving among
others secondary
elements Ja,b, Ta,b, Aa,b)
Ja: 1.1 (meat)
Jb: 1.2 (poultry)
Ta: 2.1 (potato)
Tb: 2.2 (rice)
Aa: 3.1 (green vegetable*)
Ab: 3.2 (root vegetable*)
10
10.1 From the point of view of the present theory, there are seven
features of what is here labelled Bloomeldian method which would
perhaps justify critical comment.81 These are:
(1)
(2)
(3)
(4)
(5)
(6)
(7)
The confusion of level with rank takes a specic form: the relation
between different units at one level (morpheme . . . sentence) is
conated with the relation between two different levels, grammar and
phonology. Ranged on a single scale (from phoneme to utterance)
are (i) the move between phonology and grammar (which, moreover,
always goes from the interlevel to the level; cf. below, 10.3) and (ii)
the move between units (which, moreover, always goes from the
lowest to the highest; cf. below, 10.5) in (a) phonology and (b)
grammar. In description, this leads to unwanted complexity and thus
weakens the power of the grammar: one must be free to recognize
grammatical units whose exponents in substance both overlap with and
completely fail to coincide with the units carrying phonological contrasts.82 In the theory, there is a confusion of abstractions: the abstrac65
tion involved in relating one unit to another at the same level is quite
different from that involved in relating one level to another. The sense
in which a sentence consists of morphemes is stated in description
with reference to its denition in the theory, and is totally different
from the sense in which a morpheme consists of phonemes indeed
it is doubtful if there is any meaningful sense in which a morpheme
consists of phonemes. (The contrastive use of morph and morpheme is designed to build in an extra stage to account for the two
kinds of abstraction. But it does not get over the rst (descriptive)
difculty; nor in fact does it solve the theoretical one, since morph and
phoneme differ in the extent and kind of formal determination underlying their phonological abstraction.83)
10.2 The relation among the units also tends to be confused with
the relation between a category and its exponent(s). It is assumed that
in moving up the rank scale, from morpheme to sentence, one is also
moving up the exponence scale. It is true (as said above, 6.2 and 7.3)
that in comprehensive description, in order to display the full grammaticalness of language, one takes the nal step on the exponence
scale at the lowest rank possible (though this, as already shown, is
by no means always the morpheme),84 and this is probably the reason
for the confusion of rank with exponence. But the scales of rank
and exponence are again different dimensions of abstraction, and
one can link any unit directly to its formal exponent (and through
this to its exponent in substance): the relation of an exponent of the
unit sentence to the category of sentence is exactly the same as
that of an exponent of the unit morpheme to the category of
morpheme.
10.3 The conation of levels referred to is the conation of grammar
and phonology, which follows, though is distinct, from the confusion
of level with rank (above, 10.1). The theoretical basis of this criticism
is complex but crucial. Any distinction in substance may (i) be a free
variant, in which case it is formally meaningless, or it may (ii) carry a
distinction in form, with meaning either at the grammatical or at the
lexical level.85 All formal distinctions presuppose some distinction in
substance,86 and once a distinction in substance is shown to carry a
formal distinction it must be accounted for in phonology.87 But no
relation whatever is presupposed between the categories required to
state the distinction in form (grammar or lexis) and the categories
required to state phonologically the distinction in substance which
66
carries it.88 For example: the units may not be coextensive, and a
variable relation of phonological unit to grammatical unit may be the
very thing that carries a formal system (cf. above, 10.1 n. 82); or a
single system in grammar may be carried by different phonological
distinctions, say two of its terms by tone and a third by addition of a
segment; or a phonological system, such as tone associated with a given
unit (recognized as phonological because it carries some formal distinction), may carry different formal distinctions, part grammatical and part
lexical, or terms in different grammatical systems.89
The categories required by the grammar, and the criteria for these,
should come from within grammar. They are set up to provide a
description that is comprehensive, consistent and maximally powerful.
In the denition of, for example, the unit clause, the requirement is
that it should yield classes and structures which make possible the
description of the sentence, the group and so on in terms of their
structures and classes: hence the mutual denition of all units and of all
grammatical categories (and, procedurally, until a description is comprehensive (primary delicacy at all ranks) all parts of it remain subject
to revision). A grammatical category is not required to be identiable
by reference to a particular feature of substance stated phonologically:
it merely carries the potentiality of being stated in phonological terms
through a long chain of exponence.90
(The starting point here of course is the theory of grammar, so that
what is being considered, and objected to, is the identication of
grammatical categories on phonological criteria. For linguistic
theory as a whole, the question must also be formulated the other way
round: do we derive phonological categories from formal ones? For
example, in the last case mentioned in the rst paragraph of this section,
would we state one phonological system of tone or more than one?
The nature of phonology as an interlevel suggests that its categories
should be derived from those of form; and this is usually done in
prosodic phonology, as developed by Firth and others, though it is not
an inherent requirement of prosodic method.91 Inter-level dependence
in this direction is theoretically justiable, since the role of phonology
is to account for the formally meaningful organization of phonic
substance. But it has descriptive dangers: rst, that a system carried
only by variable relation between grammatical and phonological unit
may be missed, and second, that unless a comprehensive formal
description has rst been made, formal features may be distorted into a
phonological mould the phonologist may take the word class
verb as a phonological unit, but he (or someone) must have described
67
the grammar rst, or his word and verb may not turn out to be
the grammarians word and verb.)
To be precise, then, what is being criticized here has again both a
theoretical and a descriptive aspect. In the theory, it is the use of
phonologically stated features as crucial criteria for grammatical
categories, as when supra-segmental phonemes are used as criteria for
the category of phrase.92 A phrase is a phrase because it operates in
the structure of the unit above it and has its own structures in terms of
the unit below it. It has then to be related to the phonological
categories which are arrived at by a different process of abstraction on
the basis of the minimal requirement that some formal distinction is
always involved. If at any point this yields a one / one relation of
categories, so much the better: if the phrase turns out to be exponentially coextensive with the tone group, the latter can be used as a
recognition signal for the former. But it remains a signal, not a criterion.
In description, the trouble arises when phonological features are used
as grammatical criteria even when they clash. If, for example, the
segment which carries tone contrasts, or is bounded by juncture
features, will not work as a grammatical unit, then tone and juncture
are no use even as recognition signals; so to dene, say, the clause
by reference to tone or juncture one has to set up a phonological
system in which any feature in substance can be an allophone of any
term in the system. There will be clash; and if it is recognized in the
rst place that there is nothing at all surprising when, say, units carrying
formal patterns do not coincide with those carrying patterns of the
organization of substance, then the search for one / one phonological
identication signals of grammatical categories, such as a phonological
statement of clause boundary, can be abandoned as being without
prot.93
10.4 In both grammar and phonology the smallest units, morpheme
and phoneme, are often assigned a special status distinct from that of
any of the other units. Since the description usually proceeds unidirectionally upwards (consistently in phonology, and in grammar up to
about group rank; cf. below, 10.5) these units are treated rst: they
are then rated (uniquely) fundamental, and in phonology suprasegmental features are given phoneme status. The morpheme
becomes a grammatical brick (here the analogy does yield an interesting metaphor!) which is used to build the larger units; and grammar
or at least that part of it that can be handled in this way becomes
morphemics, as phonology becomes phonemics.
68
Redundancy is assigned to what is displayed as multiple exponence: either in form, where more than one formal item is said to be
the exponent of one grammatical category, or in substance, where a
distinction is carried by what is said to be more than one phonetic
feature. But neither of these is at all clearcut. Formal redundancy
occurs where there is concord, but no criteria are available for
identifying the two prerequisites of concord: that there is more than
one exponent as opposed to one, and that these are exponents of
the same category.101 Discontinuous morphemes, for example, may
sometimes be clearly recognizable, though at others it is impossible to
say how many exponents are present;102 but the question is irrelevant,
since where the description does recognize concord this concord is
itself the exponent of a distinct category of relation that is different
from the category of which the form is exponent, and that has its own
formal meaning.103 Redundancy in substance appears when formal or
phonological distinctions are related to contrastive features.104 Here
precisely the same problem arises, since it is not possible to give
rigorous criteria for deciding what is one phonetic feature and what
is more than one.105 Each time a new parameter or a further degree
of differentiation is introduced into the phonetic statement, all its
precursors are thereby made redundant.
In extreme cases this redundancy becomes completely articial,
since it is simply inserted by the description. This happens when a
contrast (or system) is assigned to a unit lower than that to which it is
appropriate, and may result therefore from overemphasis on the lowest
unit. This tends to happen more in phonology, when the phoneme is
made to carry contrasts appropriate to a higher unit; one of the merits
of prosodic phonology is that it avoids this error.106 But it is not
unknown in grammar, where it may also arise from the use of
morphological instead of syntactic criteria for classes.107 In such cases
redundancy can only refer to the loss of power in the grammar
brought about by such a description: it already follows from the theory
that the appropriate unit for the assignment of any feature is the
highest unit that can carry it without requiring the statement to be
made twice. The best description therefore can be thought of as that
which minimalizes articial redundancy. But at the same time those
instances where what is called redundancy is an articial product of
the description are not essentially different from, but are merely
extreme cases of, the multiple exponence in form and substance to
which the same name is applied.108
What is of doubtful validity here is the implication that there are
71
Notes
1. It is in no way to deny the fundamental importance of Chomskys work
(1957) and elsewhere, if we suggest that the readiness of linguists who
had previously worked in the Bloomeldian tradition to abandon
these methods in favour of Chomskys is in part due to their lack of
72
2.
3.
4.
5.
6.
7.
8.
theoretical foundation. The point of view adopted here is that transformation-generation is a type of description which, like other types,
depends on but does not replace a theory.
Even Chomsky (1961) seems to imply that a textual study cannot be
theoretical. But a grammar of one short text may be based on theory;
and any theory-based grammar, transformational or not, can be stated in
generative terms.
Those linguists who have followed up the work of Firth have always
tended to give more weight to textual description than have those
following Bloomeld, since for the former meaning and the statement
of meaning have always been integrated in the theory. Cf. Firth (1957a):
The object of linguistic analysis as here understood is to make statements of meaning so that we may see how we use language to live
(p. 23; cf. also p. 11).
Professor Firth died on 4 December 1960. I had just completed this
paper and was planning to show it to him on the following day.
Although he had not seen it and was in no way directly responsible for
any of the opinions formulated here, the inuence of his teaching and
of his great scholarship will, I hope, be clearly apparent. See especially
Firth (1955; 1957a; 1957b, chapters 9, 10, 1416; 1957c).
Of major importance to me have been discussions, both on linguistic
theory as a whole and on the specic subjects mentioned, with J. C.
Catford, J. O. Ellis, A. McIntosh (lexis and delicacy the latter concept
is of his origination), J. M. Sinclair (English grammar) and J. P. Thorne
(logical structure of linguistic theory, and the work of Chomsky).
As used by Firth (1957b: 225). Here text refers to the event under
description, whether it appears as corpus (textual description), example
(exemplicatory) or terminal string (transformativegenerative).
The set of these abstractions, constituting the body of descriptive
method, might be regarded as a calculus, since its function is to relate
the theory to the data. It is important to distinguish between calculus
(description) and theory; also between description and the set of generalizations and hypotheses by which the theory was arrived at in the rst
place. The latter precede the theory and are not susceptible of rigorization; though we may distinguish the logical stages of observation
generalizationhypothetizationtheory, keeping Hjelmslevs (1953: 8)
distinction between hypothesis and theory; cf. Allen (1953: 53).
Here we are concerned with the stages, once the theory is formulated,
of theorydescriptiontext.
Since the theory is a theory of how language works, it does not matter
whether the levels are considered levels of language or levels of linguistics (theory or description): it comes to the same thing. Cf. Firth
(1957b): We must expect therefore that linguistic science will also nd
it necessary to postulate the maintenance of linguistic patterns and
73
9.
10.
11.
12.
13.
14.
74
15.
16.
17.
18.
75
19.
20.
21.
22.
me
noi
te
voi
lei
loro
me
noi
te
voi
(The distinctions made in written Italian are ignored, since they would
not affect the point.) The difference in format meaning is a function of
the different number of terms: in system one me excludes ve others, in
system two only three. In contextual meaning only terms of the second
group are affected.
Cf. Firth (1957a): Moreover, these and other technical words are given
their meaning by the restricted language of the theory, and by
application of the theory in quoted works (p. 2). This is true of
descriptive categories too: noun can no more be dened in a glossary
than structure.
I should therefore agree with Palmer (1958) that linguistic levels do not
form a hierarchy. His view is that there are levels, but only in the
widest sense, and that these are related in specic, but different, ways.
The set of relationships cannot be regarded as a hierarchy, except in the
loosest sense of the word. Palmer, however, appears to reintroduce
procedural hierarchy when he says, The procedure is not from phonetics via phonology to grammar, but from grammar via phonology to
phonetics, though with the reminder that the phonetic statement is the
basis, i.e. the ultimate justication for the analysis (p. 240). I would
rather say that there is order among the levels, determined by their
interrelations, but (a) no hierarchy, in the dened sense of the word,
and (b) no procedural direction. Unfortunately Palmer excludes this use
of order. There is a statable order of levels . . . and, therefore, a
hierarchy (pp. 2312, in reference to Hockett).
Immediate Constituent analysis, for example, yields a hierarchy that is
not a taxonomy: it does not full criterion (ii). (It may not always full
(i): cf. Hockett (1957): There must be also at least a few utterances in
which the hierarchical structure is ambiguous, since otherwise the
hierarchical structure would in every case be determined by form, and
order, and hence not a primitive (p. 391).)
The theory thus leads to polysystemic-ness in description both
syntagmatically and paradigmatically. Syntagmatic polysystemic statement follows from the linking of classes and systems to places in structure
(see below, 46), so that the question how can we prove that the b of
beak and the b of cab are occurrences of one and the same phoneme?
(Ebeling: 1960: 17) is regarded as an unreal one; cf. Henderson (1951:
132); Carnochan (1952: 78); Robins (1953: 96); Firth (1955: 93; 1957b:
76
23.
24.
25.
26.
27.
28.
77
morpheme
rank
30. Statistical work on grammar may yield a further unit, above the sentence:
it will then be possible to set up sentence classes, and account for
sequences of them, by reference to this higher unit. Similarly in
phonology we need a unit in English above the tone group to account
for sequences of different tones. The grammatical and phonological
paragraph (and perhaps paraphone?) is probably within reach of a
team of linguist, statistician, programmer and computer; cf. Firth
(1957a): Attention must rst be paid to the longer elements of text
such as the paragraph . . . (p. 18); Harris (1952); for Hill (1958: 406),
and others, this is stylistics, but in the present theory it would come
within exactly the same general framework of categories.
31. The simple / compound opposition is thus one of structure. It may,
of course, happen that a given realization yields simple membership all
the way up and down the rank scale. Yes may be (i.e. may be an
exponent of) one sentence which is one clause which is one group
which is one word which is one morpheme.
32. Cf. Robins (1953: 109); Firth (1957a, esp. 17, 30; and 1955, esp. 89,
91); Halliday (1959: 49).
33. Cf. Firth (1957a): Elements of structure . . . share a mutual expectancy
in an order which is not merely a sequence (p. 17). Since sequence is a
variable, and may or may not be an exponent of structure, we nd
difference in sequence without difference in structure (cf. below, 4.3
n. 42), or difference in structure without difference in sequence. I am
indebted to J. M. Sinclair for a recent conversational example of the
latter: orthographically, The man came(,) from the Gas Board. Phonologically (relevant units: tone group, bounded by //, and foot, by / these
are unit boundaries and have nothing to do with juncture): what was
said was (tonic syllable underlined):
// 1 the / man_ / came // 1 from the / Gas / Board //
Grammatically, one clause, structure SP; exponent of P came, of S the
78
34.
35.
36.
37.
79
38.
39.
40.
41.
42.
80
81
Most delicate
Highest
abstraction
Scale of
Exponence
Primary
structure
Secondary
structure
Primary
class
Secondary
class
Data
51. Or nearly coextensive: the criteria for the setting up of one primary class
or two are descriptive. For example, in English clause structure S and C
are different elements standing in different relation to P. There is a high
degree of overlap between their exponents: one primary class (class
nominal of unit group) can be set up as exponent of both S and C.
The lack of exact coextensiveness will be stated by secondary elements
and classes, to account for (for example) the occurrence of the old hall,
the old town, the old town hall, this hall / town / town hall is old, this is a hall /
town / town hall, and the non-occurrence of this old is a hall, this is an old
or this hall is town.
82
SYSTEM
1 (at X)
2 (at Y)
3 (at Z)
Secondary structures
X Ya Z Yc and
X Yb Z Yc
Secondary classes
Primary classes
55. As already stressed (above, 2.2), the order of presentation here is for
convenience of exposition; the relations among the theoretical categories
do not involve logical precedence.
56. Since the categories are set up precisely to account for the data that are
stated as their exponents, this is not surprising. Cf. Firth (1957a): A
theory derives its usefulness and validity from the aggregate of experience
to which it must continually refer in renewal of connexion (p. 1). The
relation of category to exponent can be generalized as one of abstraction;
one endpoint of this relation may be any one of four categories, but
there is no scale of abstraction among the categories their relation to
each other is such that the move from any category to its exponent may
be made either directly or via any or all of the other categories. As said
below (6.2), the route may involve rankshift; but this does not mean
that rank is to be equated with exponence or that there is any distinction
between different units as regards the kind or degree of their relation to
their exponents. (That is to say, even if one chooses to move from
clause to exponent via group, this does not mean that the group is
in any sense nearer to the data than the clause; indeed, the move from
83
group
etc.
structures . . .
(system of ) classes . . .
structures
(system of) classes . . .
etc.
exponents
59. Without in any way affecting the syntactic nature of the class.
60. Cf. Garvin (1957): Morphemes of limited membership class should be
listed in the grammar and morphemes which belong to classes of
unlimited membership should be exemplied in the grammar and listed
in the dictionary (p. 55).
61. Except in the sense that the description will always try a move down
the rank scale as a possible way of extending its power (remaining in
grammar). But wherever the lexical item is greater than a morpheme,
its further analysis by grammar into morphemes will leave its lexical
relations unaccounted for. For example, in the train left ten minutes late,
but made it up, made up is a discontinuous verbal group analysed as two
words, one (made) of two morphemes, the other simple; but it enters
into an open set qua lexical item make up, which itself is here assigned
to no grammatical unit.
62. See below, 8.
63. Regressive structures can of course be regarded as forming a scale; but
their description does not require the introduction of a separate scale
into the theory. Cf. above, 3.2 n. 26.
64. Cf. Allen (1956), from which the following is taken: It frequently
occurs that an appropriate bit of the corresponding phonological
statement (or of the orthography) is used as a label for the grammatical
unit in question . . . The price of using such labels is constant vigilance
. . . Where the phonological analysis permits of alternatives, that alternative is to be chosen which is most congruent with the grammatical
84
65.
66.
67.
68.
85
69.
70.
71.
72.
Primary structure:
Exponent of P: class:
of unit:
formal item:
(ii)
SPC
verbal
Group
ran up
Down the rank scale, verbal groups ran and ran up:
Primary structure:
Exponent of F: class:
of unit:
formal item:
(i)
F
verb nite
word
ran
(ii)
Fpo
verb nite
Word
ran
In (i), the lexical item is ran, exponent of both the unit group and the
unit word. If after further analysis ran is a compound word with two
elements of structure whose exponents are morphemes, then the lexical
86
87
88
Substance
[a] / [c]
Form
- (free variant)
singular / plural
(grammar)
cat / dog (lexis)
89
Phonology
(Unit)
phoneme:
4 tone
group:
(Contrast)
addition of segment /t/
replacement of /v/ by
/d/
change in sequence
shift of unit boundary (so
re-distribution of strong
and weak syllables
change of tonicity)
replacement of one term
in intonation system by
another
Form
phonology
with the dotted line representing the (logically) nal stage in which two
separate statements of abstraction are related. Cf. Robins (1959: 103):
(grammatical distinctions are) not deducible from the phonetic shapes,
as such, of the words concerned nor from phonological rules based on
these shapes and the phonetic categories involved in them . . . While
both phonological and grammatical categories are abstraction from the
phonic material of utterance, their relation to the phonic material is
entirely different.
89. For example the phonological system of intonation in English, operating
at the rank of the phonological unit tone group, carries a number of
different grammatical systems operating at different units in the
grammar.
90. Thus, the clause in English is dened by its operation in sentence
90
91.
92.
93.
94.
95.
structure, and by its own classes and their structures. The difference
between two instances, such as an afrmative clause he saw them and an
interrogative clause did he see them?, is of course ultimately statable in
terms of substance; but grammar is not grammar if it tries to dene the
class system in this way or even to state the difference phonologically
at all. No linguist would ever try to state the grammar of clauseclasses
by reference to phonology; yet the attempt to dene the unit clause
by reference to phonological features such as juncture is no less objectionable and leads, not surprisingly, to a phonology in which any
substantial feature is a possible exponent of any term in the phonological
system!
A phonological description will, in this view, be prosodic if (i) it
incorporates a rank scale, with a hierarchy of units to which contrasts
are assigned, and (ii) it is polysystemic, so that, for example, the /t/ in
10.3 n. 87, above is not the phoneme /t/ of the English language
(no such entity will be postulated) but a phoneme identied by reference
to its place in the structure of the unit concerned; this would still be
true whether the phonological units concerned are (partly) taken over
from grammar or are set up independently in the phonology. Firth
stresses the very different nature of the phoneme in a description of
this type, and prefers to use the distinct term phonematic unit
(cf. 1957b, Chapter 9, passim).
Cf. Haas reference to a structure of a number of pyramids, all inverted
(1960: 267).
I personally feel that English requires a totally different set of phonological units not derived from the grammatical units. Intonation in English
needs a carrier unit tone group to display the (phonological) system of
intonation; this system, and the terms in it, can then be related to the
grammar. The attempt to describe intonation in a framework of the
intonation of the clause, of the group, etc. is complicated and may
lead to a misunderstanding of the operation of the intonation system.
But the attempt is not theoretically sinful, as would be the attempt to
describe the grammar of the tone group.
As in Hills description of the English personal pronouns (1958:
pp. 1458). This playing games, or party linguistics, is again linked
to the confusion of levels. Cf. Haas (1960: 273).
I should thus agree with Robins (1959): The morpheme must be
recognized as the minimal element of grammatical structure; but this
does not imply that it is the most suitable element to bear the assignment
of all the grammatical functions fullled by the word into whose
composition it enters (pp. 1278). I would not follow Robins, however, when he says that In many ways . . . the word is a unique entity
in grammar, and not just a stage in the progression from morpheme to
utterance (p. 137). Robins rejects morpheme-based grammar but
91
96.
97.
98.
99.
100.
101.
102.
103.
92
104.
105.
106.
107.
108.
93
109.
110.
111.
112.
113.
94
Chapter Three
95
other same certain, etc.), together with other items forming an open set
(i.e. that cannot be so reduced: Johns, etc., including compound ones
as in the railway companys property).
This class of deictic may be variously subdivided along both axes. On
the one hand, there are certain sets whose members can occur in
combination, as in all my other friends; there are in fact three such
secondary groupings, the members occurring respectively in rst, second
and third place in a maximum sequence. This gives three secondary
chain classes which may be called predeictic (for example all), deictic (for example my) and postdeictic (for example other). Within each
of these three classes, choices are made. There are many ways of
describing these, according to what are taken to be the principal
dimensions. The deictic, for example, may be specic / nonspecic (my / every); selective / non-selective (my / the); and, as a
further subdivision of the class formed by the intersection of specic
and selective, it may be possessive / demonstrative (my / this).
These and various other systems eventually yield, by their subdivisions
and intersections, one-member classes: thus my can be uniquely classied
as deictic: specic, selective, possessive: personal: rst person.
Secondary classes regularly cut across each other. The systems of
specication and selection, for example, form a matrix as follows:
Specic
Selective
this / these
that / those
which
what
my your
our
their
his / her
its
Johns (etc.)
whose
the
Non-selective
Non-specic
both all
every each
no neither
a
some any
either another
deictics. But the patterns they display are typical in their complexity: a
given class breaks down by simple subdivision into a system of more
delicate classes, but the same original class will also subdivide in a
number of different ways, so that many dimensions of classication
intersect with one another. Any given item, to be fully identied, may
require to be simultaneously classied on all such dimensions. In this
way it can be assigned to a microclass, this representing its value in
respect of all the properties which have been found relevant to the way
it patterns in the language. There will be, of course, a very large
number of such microclasses: for example, in a computational study of
English phrasal verbs (items like take up, put on) which is being
carried out at the moment, 557 such items were found to yield 125
microclasses.
Up to this point I have been concerned only with place-ordered
structures. These are sometimes thought of as being the normal type of
linguistic structure. By a place-ordered structure I mean one composed
of a limited number of different elements occurring nonrecursively.
Such a structure may be fully class-dening, in the sense that to each
element corresponds a distinct class of lower rank: for example the
clause structure Subject + Predicator, with classes respectively nominal group and verbal group, as in my friends have arrived; or it
may be only partially class-dening, where two or more elements are
expounded by the same class but differentiated in sequence. In this
type of structure, there is no constant relation between successive (or
otherwise paired) elements: for example, in the structure Subject +
Predicator + Complement (for example John saw Mary, my friends have
invited me) it is not true that Subject is to Predicator as Predicator is to
Complement.
This is not the only type of structure found in language, and there
seems no particular reason for assuming it to be the norm, especially in
its pure form. Language also exhibits a different kind of structure, the
recursive or depth-ordered structure. Here, as the name implies,
an element of structure, or a combination of elements, is repeated in
depth, a series of such elements (or combinations) thus forming a
progression. It is doubtful whether one should set a theoretical limit to
the degree of depth in recursion; rather there appears to be some
logarithmic scale of diminishing frequency, so that the number of
observations one would expect to have to make before recording a
depth of, say, ten would be extremely high. Spoken English seems to
tolerate greater depth in recursion, or at least to tolerate it more readily,
than written English; and this may be true of language generally. The
101
do
Pe
his homework
Ce
pipe
g
support
b
strap
a
in
[qa
the garden
[ca
the bridge
[cg
over
[qd
in front of
[qb
the house
[cb
near
[qg
the river
[cd]
8
6) Nominal group structure and clause structure, clause only rankshifted ([[ ]] = boundary of rankshifted clause):
this is the farmer
in the garden; but this class can also operate as Adjunct in the clause,
being thus not rankshifted, as in he sat in the garden. Where a class may
occur either rankshifted or not, often ambiguity may arise, as in take
that chair in the garden: is the chair already in the garden (= take that
chair which is in the garden; rankshift) or not (= take that chair and
put it in the garden; no rankshift)?
Some of the most complex problems in the description of a
language arise where the same structure combines place-ordered with
recursive elements. Perhaps the most striking example of this in English is the nominal group, which is a troublesome mixture of the two
types; the earlier elements are largely place-ordered, recursive elements
being increasingly tolerated as one approaches the Head (the head
being, e.g., houses in the two old stone houses by the river) and continuing,
by rankshift, thereafter. The element immediately following the
Determiner, which may be called Ordinative and denes the classes
of cardinal and ordinal numeral, together with superlatives, is already
marginally recursive (e.g. the rst three second best hotels); with later
modifying elements preceding the Head this potentiality is greatly
increased, giving items such as example 4 above and familiar also in
the language of headlines: holiday coach death crash inquiry verdict. Moreover, the linear succession of the items does not act as a constant in
showing the depth relation: compare 5-millimetre perspex boxes (gamma
beta alpha) with 6-inch perspex boxes (beta beta alpha). The description
of the wordclasses entering into the structure of the English nominal
group is extremely complicated if one treats it as a simple placeordered structure, with classes dened for each possible position, as
the various attempts to do so have already shown. On the other hand
it is unsatisfactory to treat the whole thing as recursive in structure
and to recognize no secondary classes beyond the primary wordclass
noun. The facts of the language here lie in between the extremes
of these two types of structure, and the best description seems to be
one which takes this into account.
These seem to me to be some of the problems that arise in intralinguistic classication at the level of grammar. Such problems are
probably most acute at this level, but similar ones arise also at other
levels, notably phonology, which has other additional classication
problems of its own. Lexical classication is rather a different matter,
and there are reasons for preferring not to use the term class in
talking about lexis; but this subject would require a separate paper. The
eld of classication as a whole is one where linguists can learn much
from disciplines faced with similar or related problems. At the same
104
Notes
1. I am assuming here the more abstract view of grammatical categories
such as morpheme; cf. Palmer (1964b: 2327).
2. For interdisciplinary purposes I have used here the terms chain axis and
choice axis in place of their less self-explanatory technical equivalents
syntagmatic axis and paradigmatic axis.
3. Strictly speaking the ultimate exponent is a token (occurrence) of the
type (formal item) fetch. If, however, we conne ourselves to the level
of grammar we can regard the formal item as the ultimate exponent.
4. Another way of drawing the same distinction between grammar and lexis
is to say that grammar is deterministic by contrast with lexis which is
probabilistic; in the sense that in grammar one can distinguish what is
possible from what is impossible (before assigning probabilities, if one
wishes, to what is possible), whereas in lexis one can only distinguish
between what is more and what is less probable.
5. In fact the verse does contain rankshifted nominal groups, as prepositional
complement in adverbial groups (for example that lay in the house), but
these have been ignored in the illustration for the sake of simplicity.
6. This restriction to the rank immediately above implies a particular
model of grammar more specically than do most of the other points made
here. A more generally valid formulation would be enter into a structure
of higher rank.
105
Chapter Four
106
This may be illustrated from the example its John who has seen the play.
Leaving aside variation that is immaterial to the discussion, there would
seem to be three possible representations of its structure:
(1) it Subject, is Predicator, John who has seen the play Complement
(2) it . . . who has seen the play Subject, is Predicator, John
Complement
(3) its John who Subject, has seen Predicator, the play Complement
(1) would presumably be an attempt merely to state the simplest
sequence of classes in the syntagm, although it could be shown to be
unsatisfactory even on class-distributional grounds. (2) is distributionally
acceptable and would account adequately for the syntagmatic relations;
but it fails to account for the paradigmatic relations in that it does not
show the relatedness of this clause to John has seen the play, etc. If the
structural description is required to show the paradigmatic as well as
the syntagmatic relations of the grammar we need some representation
such as (3) in which John is the Subject. This leads to complexity in
the realization, since a nominalization of the form its John who seems
to add no new insight elsewhere in the grammar. A more serious
difculty arises in relation to the element Subject in English, which
is a complex element within which it is possible to distinguish three
components, or features; each of these may contrast independently of
the other two, although there is a general, and generalizable, tendency
to co-variation among them.
The three contrasts can be seen independently in (i) John has seen the
play, with tonic on play, versus, respectively, (ii) the play has been seen by
John (Subject as actor versus Subject as goal); (iii) the play John has seen (=
the play, John has seen, but . . ., Subject as theme versus Subject
nonthematic); (iv) John has seen the play (with John tonic; Subject as
given versus Subject as new). Each of these three is related paradigmatically to the original item, and each of them contrasts with it in
respect of one feature only. By a further contrast, that of unpredicated
theme versus predicated theme, (iv) is related to (v) its John who has
seen the play with John tonic.6 Thus, despite the difference in constituent
structure, (v) differs from (iv) in respect of only one feature. Such
patterns, where different complexes of (paradigmatic) features may be
combined in what is syntagmatically one and the same element of
structure, here the Subject, involve some complexity for a structural
description; if they were handled in systemic terms, the structure need
represent only their realization in syntagmatic relations. We could then
adopt a form of structural representation such as (2) above.
113
Notes
1. This paper was rst presented at a meeting of the Linguistics Association,
Newcastle, March 1965. I am grateful to R. M. W. Dixon and R. D.
Huddleston for their subsequent valuable comments and suggestions.
2. Hass (1966: 131) points out that Firths position here has been misinterpreted; this may be partly due to his use of the term element of structure
as a functional term.
3. Firth, perhaps somewhat confusingly, reserved the term order precisely
for this non-linear relation among the components (elements) of a
structure, contrasting it with sequence.
4. I use Lambs term realization instead of the earlier exponence. Lambs
term is more widely known; it also corresponds closely to my own use,
whereas as Palmer (1964b) pointed out my use of exponence differed
materially from that of Firth.
5. This may be interpreted either as if there is at least one set of conditions
under which both could occur or as if both could occur under the given
set of conditions. It is the latter interpretation which I take to be the basis
of (one aspect of) Firths polysystemic approach. Firth himself was
inconsistent in referring to a system of wordclasses noun, verb, etc.
6. Since the subject normally has the feature given, that of new being
realized in other elements, the realization of the feature new in the
subject is often accompanied by its predication as in its John who has seen
the play. This explains why its John who has seen the play usually, though
not obligatorily, has John and not play tonic, while the opposite is true of
John has seen the play.
117
Chapter Five
118
that is my fault rather than his. I plead guilty unreservedly to the charge
of inadequate documentation of my own work, although the relevant
bibliography of others writings is rather longer than Matthews seems
to suggest.
Some of the points raised in my paper, Some notes on deep
grammar (see above, Chapter 4), are I think relevant to Matthews
comments; more important in this connection is Huddleston (1965),
written without knowledge of Matthews article. Here I shall try to
take up with reasonable brevity some of the issues as Matthews himself
sees them.
Two assertions seem crucial to Matthews argument. (1) A rank-free
constituency grammar is to be preferred to a rank constituency grammar. (2) All formulations of rank grammars are either theoretically
insignicant or empirically unsound: unsound if they differ materially
from rank-free grammars and insignicant if they do not. (1) follows
from (2), except that, if they do not differ materially, this by itself gives
us no reason for preferring either. Let us examine each of these in turn.
By a rank grammar I mean one which species and labels a xed
number of layers in the hierarchy of constituents, such that any
constituent, and any constitute, can be assigned to one or other of the
specied layers, or ranks. The European linguistic tradition, by its use
of terms such as sentence and clause, has always implied the
possibility of such a grammar, although, as is well known, the absence
of criteria regulating the necessary modications of the simple constituency relation led to various difculties. I return below to the
concept of rank as suggested in such formulations as clause used as a
word and the like.
A rank grammar is, as Matthews observes, a hypothesis about the
nature of language. This leads us to ask, rst, whether it can be falsied,
and second, whether it is worth making in the rst place. Like many
other hypotheses, both in linguistics and in other disciplines, its
empirical falsication, given that it cannot be logically disproved, is
unlikely to take the form of the discovery of a clear counter-example
in this case, of a language which it is impossible to describe in rank
constituent terms. It would be likely to take the form rather of
demonstrating that the hypothesis prompts no interesting new questions
and leads to unnecessarily complex or otherwise unsatisfactory accounts
of the facts. Neither this limitation on the conditions of its falsication
nor the fact that disagreements may arise (as in all subjects) over
whether a given hypothesis has in fact been shown to be inadequate of
themselves deprive it of interest. It seems to me that this is a hypothesis
119
Notes
1. In this connection I nd the concept of a neo-Firthian , especially one
committed to certain statements, rather extraordinary. Must we all be
labelled in this way? There is an important principle at stake here: that a
scholar is responsible for what he says and writes, not for what others say
and write. If I express agreement with something another linguist has put
forward this neither makes that linguist responsible for my views nor
commits me to acceptance of the whole of his. I may be wrong, but I
feel that there are undesirable limitations on this principle inherent in
Matthews rst two paragraphs.
2. I am sorry if my own formulation was unclear here. But in that case the
whole of this part of Matthews argument rests on one purely terminological point since even if I was appearing to suggest that such an item
should be labelled a word, no further use was being made of such a
suggestion.
126
TABLE 1
Tonality: Distribution of utterance into tone groups (location of tone
group boundaries)
Tonicity: Distribution of tone group into tonic and pretonic (location
of tonic foot)
Tone (primary; pitch movement on tonic)
Tone (secondary)
This paper was written between May and August 1964 and formed the substance of a
course on the description of English at the University of Indiana. First published in System
and Function in Language, 1976, edited by G. R. Kress, London: Oxford University Press,
pp. 10135.
127
Symbols
// tone group boundary (always
also foot boundary)
/ foot boundary
tonic syllable
silent ictus
.
. . . pause
128
appendix
TABLE 2
Tonicity: location of information focus
tonic = nal lexical item (neutral)
tonic = pre-nal item or nal grammatical item (contrastive)
Tone (assuming tonality neutral):
Place of clause in sentence structure: nal main 1, non-nal
coordinate 3, non-nal subordinate 4
Declarative clauses
reservation: 1 unreserved, 4 reserved
involvement: 1 neutral, 3 uninvolved, 5 involved
agreement: 1 neutral, 3 conrmatory, 2 contradictory
information: 1 one information point, 13 two information points
key: 1 neutral, 1 + strong, 1 mild
Interrogative clauses, WH-type
key: 1 neutral, 2 (with nal tonic) mild
relation to previous utterance: 1 unrelated, 2 (with WH-tonic) echo
Interrogative clauses, yes/no type
key: 2 neutral, 1 strong
involvement: 2 neutral, 3 uninvolved, 5 involved
place in alternative question: 2 rst alternative, 1 second alternative
specication of point of query: 2 unspecied, 2 specied
Imperative clauses
key (positive): 1 neutral, 3 moderate, 13 mild
key (negative): 3 neutral, 1 strong, 13 mild
force: 1 neutral, 4 compromising, 5 insistent
function: 1 etc. command, 2 question
Moodless clauses (also as declarative)
function: 1 answer etc., 2 question, 3 warning, 5 exclamation
In the section under tone, the headings (e.g. reservation) indicate
the nature of the choice, the entries under each heading representing
the terms in the choice with their appropriate tone. Thus 1 unreserved,
4 reserved means in this choice tone 4 indicates reservation, by
contrast with one 1 which indicates no reservation. Secondary tones
are indicated where relevant.
129
Systems of tone
appendix
2
2.1
131
2.2
Declarative clause
132
appendix
2.3
Interrogative 1
133
2.4
Interrogative 2
134
appendix
2.5
Imperative
135
2.6
Transitivity
136
appendix
2.7
Theme
137
2.8
Key to examples
Examples
//1 who said / that //
//2 who said / that //
//1 who was it / said that //
//2 who was it / said that //
//2 it was / who said / that //
//1 who did you / see //
//2 who did you / see //
9 10
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
0
1
0
1
1
0
1
0
0
0
1
1
1
1
1
1
0
0
0
0
0
0
1
1
1
0
0
138
0
0
1
-
appendix
//2 who was it you / saw //
//2 where did he / go //
//2 where was it he / went //
//1 who said / that //
//2 who said / that //
//1 who was it / said / that //
//2 who was it / said / that //
//1 who did you / see //
//1 who was it you / saw //
//2 where did he / go //
//1 where / was it he / went //
//1 who wants / what //
//2 whos / going to sit / where //
//1 which shall I / put / where //
//1 John said / what //
//2 John said / what //
//1 put it / where //
//2 John put / what there //
2.9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
0
1
1
1
0
0
0
0
0
0
0
0
(1)
(1)
(1)
(1)
(1)
(1)
(1)
0
1
0
1
0
0
1
0
0
1
0
-
1
1
1
0
1
0
1
0
0
1
0
1
1
0
0
1
0
0
0
0
1
1
1
1
0
0
0
1
1
0
(0)
(0)
(0)
(0)
0
1
1
1
0
1
1
0
0
1
0
1
0
1
0
0
1
1
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
-
Key to examples
Examples
//2 . did / who do it //
//1 . did / who do it //
//2 did he do / what //
//1 did he go / where //
//1 who said / what //
//1 who was it / said / what //
//1 what did / who say //
//2 where did / who go //
//2 who went / where //
//2 where was it / who went //
139
0
0
0
0
1
1
1
1
1
1
1
1
0
0
-
0
1
0
1
-
0
0
0
0
0
1
0
0
0
1
0
0
0
1
1
1
0
1
1
0
0
Key to examples
Examples
//1 ask / John //
//1 you ask / John //
//1 you ask / John //
//1 lets ask / John //
//1 lets ask / John //
//1 do ask / John //
//13 do ask / John //
//1 you / do ask / John //
//1 do lets ask / John //
//13 do lets ask / John //
//3 dont ask / John //
//13 dont ask / John //
//4 dont / you ask / John //
//1 dont / lets ask / John //
// 1 lets / not ask / John //
//13 dont lets ask / John //
//13 lets not ask / John //
//1 you / ask / John //
//5 dont lets / ask John //
//53 . lets / ask / John //
//1 you ask //4 John
0
0
1
0
1
0
1
1
0
1
0
1
1
1
1
1
1
1
1
0
-
0
0
0
0
0
0
0
0
0
0
0
1
1
-
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
0
1
0
0
0
0
0
0
0
1
1
1
1
1
0
0
0
0
0
0
1
1
0
0
0
1
1
0
0
0
1
1
1
1
0
1
1
0
0
1
1
0
0
1
0
0
1
1
1
0
0
1
1
0
0
1
-
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
1
-
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
140
appendix
Key to examples
Examples
1
//1 John / saw the / play / yesterday //
0
0
//13 John / saw the / play / yesterday //
0
//4 John //1 saw the / play / yesterday //
//1 John / saw the / play / yesterday //
0
0
//1 John / saw the / play / yesterday //
0
//13 John / saw it / yesterday the / play //
//13 . he / saw the / play / yesterday / John // 0
//4 . it / wasnt / John that / saw the / play /
yesterday //
0
//13 . it was / John that / saw the / play /
0
yesterday //
//4 . it / wasnt the / play John / saw /
1
yesterday //
//13 . it was the / play John / saw / yesterday // 1
0
//4 John he //1 saw the / play / yesterday //
1
//1 yesterday / John / saw the / play //
1
//4 yesterday //1 John / saw the / play //
1
//1 . the / play / John saw / yesterday //
//4 . the / play //1 John saw / yesterday //
1
//4 . it / wasnt / yesterday John / saw the /
play //
1
2
-
3
0
0
0
0
0
0
0
4
0
0
0
0
0
0
0
5
-
6
0
0
0
0
0
0
1
7
0
0
0
0
0
1
0
8
0
0
0
1
1
0
0
9
00
01
10
00
00
01
01
- 0 1 1 0 0 1 00
- 0 1 0 0 0 1 01
1
1
0
0
1
1
0
0
1
0
0
0
0
1
1
0
0
0
0
0
1
0
-
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
00
01
10
00
10
00
10
0 0 1 1 0 0 1 00
141
142
appendix
143
Verbal group
7.1
te
Verbal group 1
td
tg
tb
past in
present in
future in
past in
144
future in
ta
past
present
future
past
present
future
past
present
future
past
present
future
past
1
2
3
4
5
6
7
8
9
10
11
12
13
appendix
present in
7.2
present in
past in
present in
future in
future in
past in
past in
future in
past in
present in
past in
future in
present in
future in
past in
past in
future in
past in
Verbal group 2
took/did take
takes/does take
will take
had taken
has taken
will have taken
was taking
is taking
will be taking
was taking
is going to take
will be going to have taken
was going to have taken
is going to have taken
will be going to have taken
145
present
future
past
present
future
past
present
future
past
present
future
past
present
future
past
present
future
past
present
future
past
present
future
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
7.3
Verbal group 3
146
appendix
7.4
Verbal group 4
had taken
took
would like
(1)
(1)
would have taken
had been taking
was taking
would be taking
had been going to take
was going to take
would be going to take
had been going to have taken
was going to have taken
would be going to have taken
(7)
(7)
would have been taking
had been going to be taking
was going to be taking
would be going to be taking
(10)
(10)
would have been going to take
(13)
(13)
would have been going to have taken
had been going to have been taking
was going to have been taking
would be going to have been taking
(19)
(19)
would have been going to be taking
(28)
(28)
would have been going to have taking
147
8
8.1
Nominal group 1
Principal systems
148
appendix
8.2
Determiners
149
8.3
Determiners
150
appendix
8.4
Quantiers
151
SECTION TWO
WORDCLAUSETEXT
EDITORS INTRODUCTION
wordclausetext
editors introducton
157
Chapter Six
158
wordclausetext
and third, strong (but not powerful) is a member of a class entering into
this relation with a class of which tea is a member. It would be hoped
that such classes would reappear elsewhere in the grammar dened on
other criteria. Argument, car and tea will, for example, already have been
distinguished on other grounds on the lines of abstract, concrete
inanimate and mass; but these groupings are not applicable here,
since we can have a strong table and powerful whisky, while a strong device
is at least questionable.
The same patterns do reappear: he argued strongly, I dont deny the
strength of his argument, his argument was strengthened by other factors.
Strongly and strength are paralleled by powerfully and power, strengthened
by made more powerful. The same restrictions have to be stated, to
account for the power (but not the strength) of his car and the strength (but
not the power) of her tea. But these involve different structures; elsewhere
in the grammar strong, strongly, strength and strengthened have been
recognized as different items and assigned to different classes, so that
the strong of his argument has been excluded on equal terms with the
strong of his car. Strong and powerful, on the other hand, have been
assigned to the same class, so that we should expect to nd a powerful
car paralleled by a strong car. The classes set up to account for the
patterns under discussion either will cut across the primary dimension
of grammatical classication or will need to be restated for each primary
class.
But the added complexity involved in either of these solutions does
not seem to be matched by a gain in descriptive power, since for the
patterns in question the differences of (primary) class and of structure
are irrelevant. Strong, strongly, strength and strengthened can all be
regarded for this present purpose as the same item; and a strong argument,
he argued strongly, the strength of his argument and his argument was
strengthened all as instances of one and the same syntagmatic relation.
What is abstracted is an item strong, having the scatter strong, strongly,
strength, strengthened, which collocates with items argue (argument) and
tea; and an item power (powerful, powerfully) which collocates with argue
and car. It can be predicted that, if a high-powered car is acceptable, this
will be matched by a high-powered argument but not by high-powered tea.
It might also be predicted, though with less assurance, that a weak
argument and weak tea are acceptable, but that a weak car is not.
As far as the collocational relation of strong and argue is concerned, it
is not merely the particular grammatical relation into which these items
enter that is irrelevant; it may also be irrelevant whether they enter
into any grammatical relation with each other or not. They may be in
161
wordclausetext
wordclausetext
kind of pattern and attempt to subsume all formal relations within it:
some grammatical models, as has been noted, envisage that it is the
grammars task to distinguish strong from powerful as well as to distinguish a from the and past from present; while a lexicographical
model in which a and the, as well as strong and powerful, are entered in
the dictionary and described by means of citations could be regarded as
in a similar way attempting to subsume grammar under lexis. Even
where the model recognizes two distinct kinds of pattern, these still
represent different properties of the total phenomenon of language, not
properties of different parts of the phenomenon; all formal items enter
into patterns of both kinds. They are grammatical items when described
grammatically, as entering (via classes) into closed systems and ordered
structures, and lexical items when described lexically, as entering into
open sets and linear collocations. So in a strong cup of tea the grammar
recognizes (leaving aside its higher rank status, for example as a single
formal item expounding the unit group) ve items of rank word
assignable to classes, which in turn expound elements in structures and
terms in systems; and the lexis recognizes potentially ve lexical items
assignable to sets.
But, to take a further step, the formal items themselves vary in
respect of which of the two kinds of pattern, the grammatical or the
lexical, is more signicant for the explanation of restrictions on their
occurrence qua items. The items a and of are structurally restricted, and
are uniquely specied by the grammar in a very few steps in delicacy;
collocationally on the other hand they are largely unrestricted. For the
item strong, however, the grammar can specify uniquely a class (subclass of the adjective) of which it is a member, but not the item itself
within this class; it has no structural restrictions to distinguish it from
other members of the class (and if the members of its scatter strong,
strength, etc., turn out to operate collocationally as a single item then
this conated item is not even speciable qua class member); collocationally, however, it is restricted, and it is this which allows its
specication as a unique item. There might then appear to be a scale
on which items could be ranged from most grammatical to most
lexical, the position of an item on the scale correlating with its overall
frequency ranking. But these are three distinct variables, and there is
no reason to assume a correlation of most grammatical with either
least lexical or most frequent. The most grammatical item is one
which is optimally speciable grammatically: this can be thought of as
reducible to a one-member class by the minimum number of steps in
delicacy. Such an item may or may not be least lexical in the sense
165
wordclausetext
wordclausetext
wordclausetext
the results to be interesting; and if they are, this is yet another indication
of the great insight into the nature of language that is so characteristic
of J. R. Firths contribution to linguistic studies.
Notes
1. See especially Firth (1935) (reprinted in Firth 1957b).
2. Firth (1957b: 1956): It must be pointed out that meaning by collocation
is not at all the same thing as contextual meaning, which is the functional
relation of the sentence to the processes of a context of situation in the
context of culture . . . Meaning by collocation is an abstraction at the
syntagmatic level and is not directly concerned with the conceptual or
idea approach to the meaning of words. Compare also Firth (1935) (in
lexical items): This (sc. the lexical) function should not be misnamed
semantic.
3. See Firth (1957a: 12). In the present paper lexical level has been used
in preference to collocational level in order to suggest greater generality
and parallelism with the grammatical level.
4. It is also stated explicitly by Firth (1957a: 12): Collocations of a given
word are statements of the habitual or customary places of that word in
collocational order but not in any other contextual order and emphatically
not in any grammatical order. Note that here order refers to the
mutual expectancy of syntagmatically related categories, such as elements of structure in grammar or phonology, and not to linear sequence:
cf. ibid., pp. 5, 17 and Halliday (1961: 2545 (see above, Chapter 2)).
5. That is, that distinctions are made which involve the recognition of more
nely differentiated syntagmatic relations in the grammar, and that these
in turn dene further sub-classes on various dimensions within previously
dened classes.
6. The place of collocational restrictions in a transformational grammar is
considered by Matthews (1961).
7. For a discussion of the relation between grammatical and lexical patterns
see McIntosh (1961).
8. Sinclair (1966).
9. The implication is, in effect, that wellformedness is best regarded as
lexicogrammaticalness, and that a departure from wellformedness may
be ungrammatical, unlexical or unlexicogrammatical. That the last two
are distinct is suggested by such examples as sandy hair, sandy gold and
sandy desk: sandy desk is unlexical, in that this collocation is unlikely to
occur in any grammatical environment, whereas sandy gold is merely
unlexicogrammatical: there is nothing improbable about golden sand. An
analogous distinction is observable in cliches: in shabby treatment the
mutual expectancy is purely lexical, and is paralleled in they treated him
171
wordclausetext
10.
11.
12.
13.
14.
15.
shabbily, a shabby way to treat him and so on, whereas the collocation faint
praise is restricted to this structure, in the sense that it will not occur with
similar probability under other grammatical conditions.
Compare the methods used to assess the disponibilite of lexical items in the
development of Francais fondamental (Gougenheim, Michea, Rivenc
and Sauvageot 1956).
Roget (1960: 8).
Cf. Firth (1951).
The following text examples (drawn from written work by learners of
English) may be cited in this connection: festive animals, circumspect
beasts, attired with culture, funny art, barren meadows, merry admiration,
the situation of my stockings was a nightmare, lying astray, fashionable
airliner, modern cosy ights, economical experience, delightfully stressed,
serious stupid people, shining values, a wobbly burden, light possibility,
luxurious man, whose skin was bleeding, driving a bicycle, old and
disturbed bits of brick wall, a comprehensive trafc jam, her throat
became sad, my head is puzzled, people touched with assurance, thoughts
are under a strain, a sheer new super car.
This research is being undertaken by Dr. A. R. Meetham and Dr. P. K.
T. Vaswani at the National Physical Laboratory, Teddington, Middlesex.
Sinclair, Jones and Daley (1970).
172
Chapter Seven
First published in New Horizons in Linguistics, 1970, edited by John Lyons. Harmondsworth:
Penguin, pp. 14065.
173
wordclausetext
speakers experience of the real world, including the inner world of his
own consciousness. We may call this the ideational function, though it
may be understood as easily in behavioural as in conceptual terms
(Firth 1968: 91). In serving this function, language also gives structure
to experience, and helps to determine our way of looking at things, so
that it requires some intellectual effort to see them in any other way
than that which our language suggests to us.
2. Language serves to establish and maintain social relations: for the
expression of social roles, which include the communication roles
created by language itself for example the roles of questioner or
respondent, which we take on by asking or answering a question; and
also for getting things done, by means of the interaction between one
person and another. Through this function, which we may refer to as
interpersonal, social groups are delimited, and the individual is identied and reinforced, since by enabling him to interact with others
language also serves in the expression and development of his own
personality.
These two basic functions, to each of which corresponds one broad
division in the grammar of a natural language, are also reected in
Bernsteins studies of educational failure (e.g. Bernstein 1970). Bernsteins work suggests that in order to succeed in the educational system
a child must know how to use language as a means of learning, and
how to use it in personal interaction; these can be seen as specic
requirements on his control of the ideational and interpersonal functions of language.
3. Finally, language has to provide for making links with itself and
with features of the situation in which it is used. We may call this the
textual function, since this is what enables the speaker or writer to
construct texts, or connected passages of discourse that is situationally
relevant; and enables the listener or reader to distinguish a text from a
random set of sentences. One aspect of the textual function is the
establishment of cohesive relations from one sentence to another in a
discourse (Hasan 1968).
All these functions are reected in the structure of the clause. Here
we attempt to show, by reference to English, what a clause is: how it
serves for the realization of a number of very general meanings, or
semantic options, relating to the interpersonal, ideational and textual
functions of language; and how these are expressed through various
congurations of structural roles functional elements such as
process and actor that derive from these basic functions. For a
more detailed exemplication we shall consider an aspect of ideational
175
wordclausetext
Since normally every speech act serves each of the basic functions of
language, the speaker is selecting among all the types of options
simultaneously. Hence the various sets of structural roles are mapped
onto one another, so that the actual structure-forming element in
language is a complex of roles, like a chord in a fugue: for example Sir
Christopher Wren, in the clause Sir Christopher Wren built this gazebo, is
at once Actor and Subject and theme (see, 13 below). Each of these
three represents a value in some conguration some melodic line, so
to speak such as Process plus Actor plus Goal. And all such
congurations are meaningful, since what we have called the basic
functions of language, looked at from another point of view, are simply
different kinds of meaning. So for example there is a difference in
meaning between (1i) and (1ii):
(1i) She would marry Horatio. She loved him.
(1ii) She would marry Horatio. It was Horatio she loved.
The difference concerns the organization of the second clause as a
piece of information, and it derives from the textual function. There is
also a difference between (1i) and (1iii):
(1iii) She would marry Horatio. She did not love him.
But we cannot say that this difference is greater or more meaningful
than that between (1i) and (1ii); it is merely of a different kind. The
speaker does not rst decide to express some content and then go on to
decide what sort of a message to build out of it whether to turn it into
a statement or a question, whether to make it like (1i) or (1ii) and so on.
If he did, the planning of each sentence would be a totally discrete
operation and it would be impossible ever to answer a question that had
actually been asked. Speech acts involve planning that is continuous and
simultaneous in respect of all the functions of language.
Linguistics is not as a rule concerned with the description of
particular speech events on individual occasions (although it is possible
176
wordclausetext
wordclausetext
The three main types of transitivity role process, participant, circumstance correspond, by and large, to the three major word (or word
group) classes found in most languages: verb, noun, adverb. In English,
typically, processes are expressed by verbal groups, participants by
nominal groups and circumstances by adverbial groups the last often
in the form of prepositional phrases. There are also incongruent forms
of expression, with functions of one type expressed by classes primarily
associated with another type, as in (8):
(8) dinner of roast beef was followed by a swim
Here the processes of eating and swimming are expressed by nouns;
the temporal relation between them by the verb follow; and of the two
participants, one is omitted and the other (roast beef ) is made to qualify
dinner (contrast in the evening they ate roast beef and then swam).
The circumstantial functions seem less central to the process than do
the participant functions; this is related to their inability to take on the
role of subject. But this peripheral status is not a feature of all
circumstantial elements, which can be subdivided into an inner and
outer type. Within the function place, in:
(9i) he was throwing stones at the bridge
(9ii) he was throwing stones on the bridge
at the bridge (the inner type) seems more central to the process than
on the bridge: we can say what was he throwing stones at? and not (in this
sense) what was he doing at the bridge? (On the other hand, we can say
180
what was he doing on the bridge? and not what was he throwing stones on?)
However, the sense of inner and outer is contributed to by various
factors not all of which coincide. For example, in (10) the place
element is obligatory in (i) but optional in (ii):
(10i) he put all his jewels in the wash
(10ii) he lost all his jewels in the wash
In (11), there is a difference of clause type; (i) is a relational clause
(see below, Section 7) whereas (ii) is an action clause (Fillmore, from
whom (11) is taken, gives this as an instance of dependency between
functions: the place element is outer if an actor is present and inner
otherwise):
(11i) John keeps his car in the garage
(11ii) John washes his car in the garage
Inherent functions
wordclausetext
within one type there may be different sets, and different alignments,
of participants; this is the function of the system of voice of the choice
between active and passive, though the actual patterns are more
elaborate than this. The options in the voice system (simplifying
somewhat) are (a) middle / non-middle (see next paragraph); if nonmiddle, then (b) active / passive (not exactly equivalent to active
and passive in the verb; see Halliday, 196768: I 39ff., where they
are referred to as operative and receptive); if active, then (c)
plus / minus goal; if passive, then (d) plus / minus actor. The reason
for choosing one rather than another of these options lies in the textual
function of language (see Sections 11 and 12 below); but which options
are available to choose from depends on transitivity.
Voice is concerned with the roles of Actor and Goal (but see below,
Section 8), both as inherent and as actualized roles. A middle clause
is one which has only one inherent participant, which for the moment
we will continue to refer to as the Actor; examples are Hector sneezed,
the cat washed. A non-middle clause is one which has two, an Actor
and a Goal, but one or the other may not be actualized: if active,
there may be no Goal, for example Mary is washing (the clothes), and
if passive, no clear Actor, e.g. the clothes have been washed (by Mary).
All actions are classied into those involving one participant role and
those involving two; there are then different ways of presenting the
situation in those cases where there are two.
The point was made earlier that the notion of participant derives
from the more fundamental concept of syntactic function, or role.
The basic elements of transitivity structure are the various roles associated with processes; and two or more such roles may be combined
in one participant, as in a reexive clause such as John is washing
(himself ) where John is both Actor and Goal at the same time. The
elements that operate as Actor, Goal, etc. also play a part, simultaneously, in other structures of the clause, expressing aspects of the
interpersonal and textual functions of language. The principle of
combining a number of roles in a single complex element of structure
is fundamental to the total organization of language, since it is this that
makes it possible for the various functions of language to be integrated
in one expression. We return to this below, Section 9.
All the clauses so far considered have been concerned with actions or
events, and have involved an actor as inherent role. Let us refer to
182
roles
voice
(verb)
example
Actor
Actor, Goal
Goal, Actor
wordclausetext
voice (15i), and those having the Processer (15ii). In the rst types, the
passive form is much more frequent than the passive in action clauses;
in the second type it is much less so:
(15i) the gift pleased her / she was pleased by (with) the gift
(15ii) she liked the gift / the gift was liked by her
This is because the passive is a means of bringing the element governed
by by into prominence as the focus of information (see below, Section
12); in (15ii) the by element, i.e. her, is the Processer, and in English
this tends to be the given element in the situation (she must have
been referred to already in the text), and thus does not appropriately
carry such prominence.
Mental process clauses express (a) perception, e.g. see, look; (b)
reaction, e.g. like, please; (c) cognition, e.g. believe, convince; (d) verbalization, e.g. say, speak. They are distinct in that the Phenomenon that
which is perceived, reacted to, etc. is not limited, as are the
participants in action clauses, to the class of thing, namely persons,
objects, abstractions and the rest of the phenomena on the plane of
experience.
What is perceived or felt or thought of may be a simple phenomenon of this kind, but it may also be what we might call a metaphenomenon: a fact or a report a phenomenon that has already as it were
been ltered through the medium of language. Here words as well as
things may participate in the process.
For example, in (16) all the processed entities are simple phenomena, or things:
(16i)
(16ii)
(16iii)
(16iv)
We could insert the fact (that) in (17) and the report (that) in (18i); not,
however, in (18ii), which is a clause of verbalization, since such clauses
accept only reports, and reported speech is the meaning of clauses of
184
this type. The difference between fact and report is that a fact is a
representation at the semantic level, where the truth lies in the meaning
(she regretted) that he had gone away; whereas a report is a representation at the lexicogrammatical, or syntactic, level, where the truth lies
in the wording (she said) that he had gone away.
In relational clauses, the process is simply a form of relation
between two roles. One type is the attributive, such as:
(19i) Marguerite is a poet
(19ii) Marguerite looks desperate
where the relation is one of class membership: Marguerite belongs to
the class of poets, . . . the class of people who look desperate. This
is a relation between entities of the same order of abstraction but
differing in generality.
The other type, exemplied by (20):
(20i) Templecombe is the treasurer
(20ii) the treasurer is Templecombe
has two functions, resembling the two terms of an equation, where the
one serves to identify the other, as in x = 2. Here the two entities are
alike in generality but differ in abstraction: the identifying element may
be of a higher order of abstraction, as in (21i), where the treasurer
expresses Templecombes function, or of a lower order, as in (21ii)
where the fat one expresses Templecombes form, how he is to be
recognized:
(21i) (which is Templecombe?) Templecombe is the treasurer
(21ii) (which is Templecombe?) Templecombe is the fat one
(21i) could be interpreted in the sense of (21ii) if the committee were
in view on the platform; there is in fact partial ambiguity between
these two sub-types.
These two major types of relational clause, the attributive and the
equative, differ in various respects. The attributive are non-reversible
(e.g. we can say that man is a poet but not a poet is that man), have the
role Attribute which may be an adjective and is usually indenite,
express class inclusion, are usually questioned by what? or how? and are
expressed by the verbs be, get, turn, keep, remain, seem, sound, look, etc.
The equative are reversible (i.e. have a voice system), have the role
Identier which must be as noun and is usually denite, express class
identity, are usually questioned by who? or which? and are expressed by
the verbs be, equal, represent, resemble, stand for, etc.
185
wordclausetext
The ergative
unreal, since there are verbs like lead which are normally transitive (two
inherent participants) and others like march which are normally intransitive (one inherent participant). But with a large number, especially of
the more frequently used verbs, either form seems equally normal:
there is nothing to choose, as regards the more typical use of the verb
bounce, between he bounced the ball and the ball bounced. In addition there
are a number of verbs which, while themselves clearly transitive or
clearly intransitive, group into pairs differing only in transitivity, so that
Mary put out the re is to the re went out as Polly lit the re is to the re
lit.
It has been pointed out by various linguists (Halliday 196768: 3;
Anderson 1968; Fillmore 1968) that action clauses in English seem to
be organized on an ergative rather than on a transitive (or nominative) basis. This means that, with any action clause, there is associated
one inherent role which is that of the participant affected by the process
in question. Fillmore describes this as the semantically most neutral
function, and labels it the objective; I used the term affected,
which I will retain here. In (23) the recruits has the role of affected in
every case, even through it is Goal (if an ActorGoal analysis is used)
in (23i) and Actor in (23ii); in general, the affected is the Goal in a
transitive and the Actor in an intransitive clause.
We have now turned what was the borderline case, such as (23iii),
into the most central clause type. This is the type in which both middle
(one-participant) and non-middle (two-participant) forms are equally
normal; it may be considered the favourite clause type of Modern
English. The transitive and intransitive types those with non-middle
as norm and with middle as norm respectively are the marginal ones,
and they seem to be becoming more marginal as time goes on.
Hence all the examples in (24i) have the same structure, with a
Process and an Affected. Those in (24ii) also have a Causer (Fillmores
agentive):
(24i) theyre being led
theyre being trained / theyre training
theyre being marched / theyre marching
(24ii) hes leading them
hes training them
hes marching them
These two ways of representing processes, the transitive and the
ergative, are very widely distributed; possibly all languages display one
or the other, or (perhaps always) both, in different mixtures. In English,
187
wordclausetext
the two occur side by side. The transitive system asks does the action
extend beyond the active participant or not?, the ergative, is the
action caused by the affected participant or not? The ergative component is more prominent now than it was in Middle English, and this
appears in various ways, for example, the change from impersonal to
personal forms in mental process clauses (formerly methinks, it likes me).
In the modern form I like, I cannot be explained as an Actor (among
other things we cannot say what he does to jam is like it); but it can be
shown on various grounds to have the function Affected.
As this suggests, the ergative pattern, whereby a Process is accompanied by an obligatory Affected participant and an optional Causer, is
more readily generalizable than that of Actor and Goal. It extends
beyond action clauses to those of mental process, and perhaps even to
clauses of relation as well. We want to say that Paul has the same
function in both (25i) and (25ii):
(25i) Paul fears ghosts
(25ii) ghosts scare Paul
not that they are identical in meaning, but that the transitivity roles
are the same. This is not possible in ActorGoal terms. But in an
ergative system there is considerable evidence for regarding Paul as the
Affected participant in both cases. The ergative, therefore, represents
the more general model of the transitivity patterns of Modern English
that is, of the options available to the speaker of English for talking
about processes of all kinds.
10
Mood
wordclausetext
11
Theme
iv) the Theme is the item outside the brackets, what is inside being
the Rheme:
(28i)
(28ii)
(28iii)
(28iv)
(28v)
I (dont know)
yesterday (we discussed the nancial arrangements)
his spirit (they could not kill)
suddenly (the rope gave way)
people who live in glasshouses (shouldnt throw stones)
As we have seen, Theme, Actor and modal Subject are identical unless
there is good reason for them not to be (cf. (26) above). Where they
are not, the tendency in Modern English is to associate Theme and
modal Subject; and this is the main reason for using the passive. The
passive has precisely the function of dissociating the Actor from this
complex, so that it can either be put in focal position at the end or,
more frequently, omitted, as in (29):
(29i) this gazebo was built by Sir Christopher Wren
(29ii) this gazebo is being restored
The typical theme of declarative clause is thus the modal Subject (or
grammatical subject this gazebo in both cases); in interrogatives,
however, the picture is different. If we ask a question, it is usually
because we want to know the answer, so that the typical Theme of an
interrogative is a request for information. Hence we put rst, in an
interrogative clause, the element that contains this request for information: the polarity-carrying element in a yes / no question, and the
questioning element in a wh- question, as in (30):
(30i) didnt (Sir Christopher Wren build this gazebo?)
(30ii) how many gazebos (did Sir Christopher Wren build?)
In English there is a denite awareness of the meaning expressed by
putting something in rst position in the clause. The Theme is the
point of departure for the message; a paradigm form of it is the
headword in a denition, for example a gazebo in (31):
(31) a gazebo is a pavilion or summerhouse on an eminence, open
for the view
In addition to the selection of a particular element as the theme, the
speaker has other options in thematic structure open to him (Halliday
196768: 2); for example, any clause can be split into two parts by
the use of nominalization, as in
191
wordclausetext
(32) the one who built this gazebo was Sir Christopher Wren
where the Theme is the whole of whichever part comes rst here the
one who built this gazebo.
12
Information structure
193
wordclausetext
13
Conclusion
195
Chapter Eight
Preamble
Figure 1
content1
(realized as)
(realized as)
expression1 = content2
expression2 . . .
196
Figure 2
Now it is typical of semiotic systems that the different strata are not
isomorphic; there is no relation of biuniqueness (oneone correspondence) between one level and the next. This is bound to be the case in
a system such as language, where the coding not only converts elements
of one kind into elements of another kind meanings into wordings
into sounds but also reduces both the size and the inventory of the
basic components. By any usual denition of linguistic units, units of
speech sound are both smaller than and fewer than units of form; and
units of form are both smaller than and fewer than units of meaning.
197
wordclausetext
INTERPERSONAL
TEXTUAL
LOGICAL
formal logic are ultimately derived. These two, the logical and the
experiential, together make up the ideational component in the semantic system: that of meaning in the reective mode.
The second main component, the interpersonal, is language as
interaction: it is meaning in the active mode. Here the semantic system
expresses the speakers intrusion in the speech event: his attitudes,
evaluations and judgements; his expectations and demands; and the
nature of the exchange as he is setting it up the role that he is taking
on himself in the communication process, and the role, or rather the
role choice, that he is assigning to the hearer. This component is
therefore both speaker- and hearer-oriented; it is interpersonal what
Hymes called socio-expressive and represents the speakers own
intrusion into the speech situation.
All discourse involves an ongoing simultaneous selection of meanings
from both these components, which are mapped into a single output
in the realization process. But there is also a third component, which
we are calling the textual, whereby the meanings of the other two
kinds take on relevance to some real context. Here the semantic system
enables the speaker to structure meaning as text, organizing each
element as a piece of information and relating it signicantly to what
has gone before. If the ideational component is language as reection
(the speaker as observer of reality), and the interpersonal component is
language as action (the speaker as intruder in reality), the textual
component is language as relevance (the speaker as relating to the
portion of reality that constitutes the speech situation, the context
within which meanings are being exchanged). The textual component
provides what in modern jargon we might refer to as the ecology of
the text. For example, from the Walrus and the Carpenter, in Alice
Through the Looking-Glass, when the Carpenter says to the Walrus
Cut us another slice!
the ideational meaning is the representation of a material process,
cutting, in which three entities participate: the one who cuts, the thing
that is cut and the one that the thing is cut for; also the place of cut in
the taxonomy of actions and of slice in the taxonomy of things. The
interpersonal meaning is a demand for goods-and-services, I want you
to do something for me, embodied in the selection of the imperative
mood, direct, explicit and without any special modulation. The textual
meaning is the internal organization of this as a message with the focus
on what is demanded, together with its relation to the preceding text
199
wordclausetext
wordclausetext
For the rest of this paper we shall be concerned with what is below
the semantic system: with the question of what kinds of structural
mechanism are typically involved in the realization of these various
components of meaning. The suggestion will be that here too the same
categories are relevant, since they tend to be expressed through
fundamentally different types of structural organization (cf. Mathesius
1964: 245 of Czech original).2
Let us consider the experiential function rst. Here we are concerned with the semantic (linguistic) encoding of experience; particularly our experience of the processes of the external world and of the
internal world of our own consciousnesses. We tend to encode such
experience in terms of congurations of elements each of which has a
special and distinct signicance with respect to the whole. Typically,
we recognize a process itself, and various more or less specialized
participants and circumstantial elements.
For example, suppose there is a ock of birds ying overhead. We
represent this in language as something like There are birds ying: that is,
a process of ying and, separated out from this, an entity that is
doing the ying, namely birds. This is, certainly, one valid way of
encoding it; but it is not the only one we might have said, instead,
Its winging. If we did say this in English we would be treating the
phenomenon as a single unanalysed process, not as a process plus a
participant; this is, after all, what we do with Its raining (although not,
for some reason, with The winds blowing). No doubt it is useful to be
able to talk about birds doing other things than ying, and about ying
being done by other things than birds. Some languages feel the same
about rain: in Chinese one says, liberally translated, Theres rain
falling; and in one south Chinese dialect, a variety of Cantonese, there
are usually two participants in the pluvial process, which is encoded as
The sky is dropping water.
So there is no reason for assuming that each particular process will
always be encoded as just this or that particular conguration of
202
elements. But we can formulate the general principle that this is how
processes are represented in languages. This means that a structure
which represents experiential meanings will tend to have this form: it
will be a conguration, or constellation, of discrete elements, each of
which makes its own distinctive contribution to the whole.
We usually represent this kind of structure linguistically as a functionally labelled constituent structure as shown, for example, in Figure
3.
Figure 3
Figure 4
Figure 5
Figure 6
203
wordclausetext
Figure 7
Figure 8
introduced into the structure in this way are the source of the linguists
classic mode of representation of constituent structure in the form of
trees, or labelled bracketings. But the appropriate bracketing for functionally labelled structures is minimal, not maximal (many ICs rather
than few ICs, in the terms of Hudson 1967). Maximal bracketing
imposes too much structure for a functional grammar: for example, in
four young oysters it is reasonable to recognize young oysters as a
constituent provided the labelled elements are classes, since it is a
nominal group, but in no way does this correspond to any meaningful
functional constituent. The nonlinear representation implies more of a
molecular model of structure, with a taxonomy similar to cell: molecule: atom: subatomic particle.
Experiential meanings are typically realized as elemental structures of
this type. The basic structural mechanism is that of constituency, with
larger units constituted out of layered clusters or bracketed strings of
smaller units, each part having its own specic function with respect to
the whole. We could call this segmental, except that it is better
perhaps to reserve the term segment for an element in the nal
output the syntagm that serves as input to the next realizational
cycle. So let us say that experiential meanings are realized through
some kind of constituent structure.
This expresses the particular way in which we order our experience
of reality when we want to turn it into meaning. The bounded entities
that enter into constituent structures with specic functions like Process, Actor, Goal, Extent or Manner offer a presentation of reality in
terms of things doings by, and happenings to, persons and objects,
in the environment of other persons and objects, with yet other persons
and objects, and also times and places and so on, as attendant circumstances; and including various metathings (facts and reports), which
are complex things that have already been encoded in language and so
acquired a status which enables them to participate in certain types of
204
wordclausetext
place where the greatest pitch movement takes place. (This may be a
falling movement or a rising movement or some kind of complex
movement, depending on which kind of melody it is. It corresponds
to what is sometimes called primary stress, although it is not, in fact,
a stress feature.) Suppose we represent the intonation unit (the tone
group) as bounded by double slash, and intermediate rhythmic units
(the feet) by single slash, with tonic prominence as bold type; the likely
form of the utterance would be:
// why did you / let the / big one / get a/way //
The meaning of tonic prominence is the focus of information; it
signals the climax of what is new in the message. This kind of focal
prominence can be assigned at any point in the clause; it is not realized
by nal position, in the way that thematic prominence is realized by
initial position. But it is typically located at the end, and any other
focus is marked and so explicitly contrastive. In the typical form of
the message, in other words, the speaker puts what is new at the end.
So there is a peak of prominence at the beginning, which is the
Theme; and another peak of prominence, usually at the end, which is
the focus of information or, simply, the New. The two are different in
meaning. The Theme is speaker-oriented; it is the speakers signal of
concern, what it is that he is on about he may even make this
explicit, by starting as far as . . . is concerned. The New is heareroriented (though still, of course, selected by the speaker); it is the
speakers presentation of information as in part already recoverable to
the hearer (the Given) and in part not recoverable (the New). These
two types of prominence are independent of each other. But both
contribute to the texture, to fashioning the fabric of the text.
What these text-forming systems do is to organize discourse into a
succession of message units, quanta of information such that each has
its own internal texture, provided by the two systems of prominence
just mentioned. The message unit corresponds, typically (i.e. in the
unmarked case), to a clause. Hence it is possible in such instances to
represent both thematic and focal prominence as constituent-like
structures of the clause, by recognizing the functional signicance of
the non-prominent part. So Theme contrasts with Rheme, and New
contrasts with Given, as in Figure 9. In fact, the information unit is not
always coextensive with the clause; to return to the Walrus and the
Carpenter, in The moon was shining sulkily (following the earlier The sun
was shining on the sea, shining with all his might)
207
wordclausetext
Figure 9
Figure 10
Figure 11
Given structures of these very general kinds, it is clear that each can
be reduced to some form of constituency; but not all with the same
success. Experiential structures are quite constituent-like; whereas interpersonal ones are not, and the attempt to represent them in constituency terms involves idealizing them to an extent that is tantamount to
209
wordclausetext
Notes: (1) experiential : clause as representation (of process); (2) interpersonal : clause as interaction; (3) textual : clause as message
Figure 12
210
Figure 13
wordclausetext
reference to things. The linguistic structures actually stand as metaphors for the relations between things; and the elements that enter into
and are dened by these relations are identied as Process, Actor, Goal,
Extent, Manner and the like. These in turn are interpreted as roles
occupied by various classes of phenomena, and these classes of
phenomena themselves have names, names like moon and shine and
sulky.
In the logical mode, reality is represented in more abstract terms, in
the form of abstract relations which are independent of and make no
reference to things. No doubt these relations, which taken all together
constitute what we might call the logic of natural languages, have
evolved by a process of generalization out of relations between things;
and some of them, for instance and, are not hard to interpret in
concrete terms (one can lay a set of objects side by side). But unlike
experiential structures, logical structures present themselves in the
semantic system as independent of any particular class or classes of
phenomena. They are not the source of rules about what goes where.
Again we have to deal with a distinction whose boundaries are
fuzzy; there are the usual doubtful cases. More interesting, however, is
the question whether languages differ as to what relations they are
going to treat as logical. It seems to me that they do, although this
argument will depend on our being willing to accept evidence from
below that is, to argue that, because we can identify a particular
type of structure as characteristic of the expression of logical meanings,
wherever we nd that type of structure we shall assume it derives from
the logical component. Since we are claiming these structural manifestations are only tendencies, such an argument is only tenable on the
grounds that the type of structure that is generated by the logical
component is in fact signicantly different from all the other three.
The principle is easy to state: logical structures are recursive. But we
immediately encounter a difculty here, a difculty that is associated
with the use of the term embedding to cover two different types of
structure-forming process.
In one type which I have referred to as rankshift the output of
one network (by the application of realization statements) produces an
element of structure which is a point of entry into the same (or some
higher rank) network. A typical instance of rankshift is nominalization,
where a function in the structure of a clause may be lled, not by a
nominal group (the congruent form) but by something that itself has
the structure of a clause, for example to come and spoil the fun and That
you have wronged me in:
212
Figure 14
wordclausetext
214
the same way, the days of the week do not form a closed system, in
the linguistic sense; they form a lexical set whose members happen to
be limited and xed. In both instances, the inniteness of the set of
natural numbers and the niteness of the set of days of the week are
properties not of the language but of the social system.
Logical structures are different in kind from all the other three. In
the terms of systemic theory, where the other types of structure
particulate (elemental), prosodic and periodic generate simplexes
(clauses, groups, words, information units), logical structures generate
complexes (clause complexes, group complexes, etc.) (Huddleston
1965). The apparent exception is the sentence, which is generated by
logical structures; but this is merely a terminological exception the
sentence is, in fact, simply a name for the clause complex. While the
point of origin of a non-recursive structure is a particular rank each
one is a structure of the clause, or of the group, etc. recursive
structures are in principle rank-free: coordination, apposition, subcategorization are possible at all ranks. The more restricted ones, like tense
and report, are also the ones that are nearer the borderline; they are
only just logical structures. Tense is particularly interesting because it
has only come into this category within the last two to three centuries,
and English appears to be unique in treating tense in this way.
Postscript
wordclausetext
Table 2
Functional mode
Example
logical : paratactic
logical : hypotactic
Textual
Experiential
10
Summary
Notes
1. Written language may be (i) an alternative coding of meanings (ideograms);
(ii) an alternative coding of wordings (characters, as in Chinese); or (iii) an
alternative coding of sounds (syllabaries and alphabets).
2. Mathesiuss observation relates to functional styles (orientation towards
different functions in the use of language), not to functionally derived
components of the system; but it is pertinent nevertheless: The inuence
of functional styles on the lexical and semantic aspects of speech was
stressed especially by Grober, . . . [who] distinguishes the subjective
expression . . . and the objective expression . . . The subjective expression
differs from the objective both quantitatively (inasmuch as it expresses by a
pause, by tone or by gesture what the latter expresses by words; and
further, as it repeats what could be expressed only once) and qualitatively (by
217
wordclausetext
choosing other words than factual names of the things referred to), and,
nally, locally (by placing sentence elements into positions not pertaining to them
in objective speech). Both ways of expression are often combined in actual
speech. (My italics throughout.)
218
Chapter Nine
219
wordclausetext
Text is the process of meaning; and a text is the product of that process.
A text is therefore a semantic entity; it is given to us in clauses, but it
is not made of clauses, in the sense of being a whole of which the
clauses are simply parts. So when we speak of the problem of relating
clause to text as one of getting from micro to macro, this is only one
aspect of the relationship. It is true that texts are, on the whole, larger
than clauses; what is more signicant, however, is that they are one
level of abstraction beyond the clause. The relationship is not so much
one of size as one of overt to covert; the text is realized in clauses. In
scale-and-category terminology (Halliday 1961), the relationship of
clause to text is one of exponence as well as one of rank.
This has consequences for the ways in which the properties of a text
are made manifest. For example, the notion that a text has structure
221
wordclausetext
For at least some registers, perhaps all, it is possible to state the structure
of a text as a conguration of functions (Hasan 1979). A generalized
structural representation is likely to include some elements that are
obligatory and others that are optional; and the sequence in which the
elements occur is likely to be partly determined and partly variable.
Most of the actual formulations of text structure that have been put
forward seem to relate to one broad genre, that of narrative. The
original source of inspiration for these was Propps theory of the folk
tale. The structure of traditional oral narrative has been investigated in
detail within tagmemic and straticational frameworks, on foundations
provided by Longacre and Gleason. A well-known representation of
another kind of narrative is Labov and Waletskys structural formula
for narratives of personal experience:
222
One of the most frequent observations made about texts that are felt to
be defective in some way is that they do not hang together: they
lack coherence. A text has coherence; it forms a unity, a whole that
is more than the sum of its parts.
Coherence is a complex property to which many factors contribute.
One way to approach it is through the category of cohesion, as dened
by Halliday and Hasan (1976). Cohesion is a necessary but not a
sufcient condition of coherence. The different types of cohesive
relation are the fundamental resources out of which coherence is built.
But the mere presence of cohesive ties is not by itself a guarantee of a
223
wordclausetext
wordclausetext
high school. She has shown how the conjunctive relation of specic
instance to general principle, coded in Martins network as:
(internal / comparative) : similarity : nonexhaustive
is the major conjunctive factor giving coherence to these texts. For
example, in one part of a text there occurs the sentence
Odysseus friends and anyone else who heard his story respected him.
Elsewhere in the same text we nd:
Heroic men are very much respected and idolized.
(Either member of such a pair may come rst.) There is lexical cohesion
between individual items the repetition of respect. There is cohesion
between Odysseus and heroic (men), with Odysseus being a hyponym
of hero. But between the two sentences as wholes there is a conjunctive relation itself an extension of hyponymy such that the second
one stands as a general principle of which the rst one offers a specic
instance; and this type of conjunction is a distinguishing feature of the
sort of expository discourse she is investigating.
1.1.3
wordclausetext
1.1.5
wordclausetext
wordclausetext
We should not press the analogy too far. But if it seemed useful to
set up simultaneous structures in a text along these lines we might ask
whether there is the same kind of structural variation as we nd in the
clause, with the eidological structure being particulate (represented
by denable segments of the text) and the ethological being eldlike (represented by overlapping prosodies in the text). (The wavelike periodic movement corresponding to the textual dimension of
clause structure has already been referred to under 1.1.4 above; cf.
further below.) (Cf. Halliday 1979.)
To say that a text resembles a clause in having coherence is not to
say very much, since the coherence in a clause is created by its
structure, whereas coherence in a text largely depends on cohesion.
Cohesion is the resource that takes over, as it were, when grammatical
structure no longer holds (i.e. above the clause complex). We could
point out that cohesion also obtains within clauses; we nd reference,
substitution, ellipsis, conjunction and lexical cohesion all operating
between elements in the same clause, for example:
Ms evening speech caused more fuss than his morning one had.
C
R
L
S E
But this is a supercial similarity. A more signicant analogy can be
found with the notions of cohesive harmony and conjunctive relations
discussed in 1.1.2 above.
Ruqaiya Hasans work showed how lexico-referential motifs enter
into a text not as isolated motifs but as interlocking chains having some
kind of regular functional relationship with each other. But these
functional relationships are relationships within the clause; and this
reects the fact that the elements in these chains themselves cannot
occur as isolated entities. Names have no place in language except in
function with other names; and the functions are dened within the
structure of the clause.
The conjunctive relations discussed by J. R. Martin (1992) are also
derived from relations with the clause. Consider a series such as the
following:
She didnt know the rules. She died.
She didnt know the rules. So she died.
She died, because she didnt know the rules.
She died because of not knowing the rules.
That she died was because she didnt know the rules.
232
That she died was caused by her not knowing the rules.
Her ignorance of the rules caused her death.
The same conjunctive relation, the external causal, can be coded in
very many ways. It can appear as a relationship within the clause,
realized lexicogrammatically; but it can also serve to link segments of
the text, at any distance and of any extent. The kind of coherence that
is achieved by the presence in the text of semantic relationships of the
conjunctive kind is essentially a clause-type coherence, one that is
based on relations dened systematically within the transitivity system
of the clause.
The notion that a text has function is again closely related to an
analogous feature of the clause, also one that is coded in the lexicogrammatical system: that a clause has a speech function, realized by the
mood system. The speech function of the clause in simplest terms, as
statement, question, command, offer, or a minor speech function is
represented by the grammatical categories of declarative, interrogative
and so on; this is the rhetorical function of the clause, and the whole
range of rhetorical functions that we assign to text are simply the
mood of the text. (Cf. Martin (1980) for speech-functional analysis
of dialogue.)
The development of the text again has its analogy in the clause. This
has already been made clear from the example cited in 1.1.4, since
Peter Fries used the themerheme structure of the clause as the source
from which to derive the method of development of the text. We can
generalize this still further by bringing in the notion of information
structure, the givennew movement within the clause.
In its textual aspect, a clause has a wave-like periodic structure
created by the tension between themerheme (where theme is the
prominent element) and givennew (where new is the prominent
element); the result is a pattern of diminuendocrescendo, with a peak
of prominence at each end. There is a balance of development (i) away
from the theme, and (ii) towards the new. But these are separate
movements. They are in phase in the unmarked, default case, where
the theme is selected from what is given, and the news is put into the
rheme. But they can also be out of phase, and this gives an alternative
pattern of texture to the clause. Putting the two out of phase means
locating the new (the focus of information) somewhere other than at
the end of the rheme; this as it were changes the wave shape but does
not disturb the essential periodicity.
This pattern is the method of development of the clause. It is
233
wordclausetext
closely analogous to what takes place in a text; not only over the whole
text but also in structurally dened intermediate units within the text.
The classic movement of a paragraph, beginning with a topic sentence
(from theme to elaboration) and culmination having a high point,
unmarkedly but not obligatorily nal in a climax (from prelude to
main point), is one of the clearest manifestations of the analogy between
clause and text. It is in the clause that this movement is displayed in
the most systematic and clearly motivated form.
Finally a clause can be said to have character in both the generic
and the specic sense. If a text is typied by virtue of its being
organized around the expression of processes of a particular type, the
clause is the unit in which these processes are realized and categorized.
The clause is the locus of the transitivity system: the system of material,
mental and relational processes, together with their numerous subcategories. Thus analogous to the major types or genres of text are the
major types or classes of the clause, each being characterized by the
selection of a dominant process type.
But each clause is also a unique combination, or potentially unique
combination, of features deriving from the different semantic functions,
ideational, interpersonal and textual. Moreover any one or other of
these may be foregrounded: the clause may display an orientation
towards any one, or any combination, of the various systems and their
subsystems. The extract from J. B. Priestleys An Inspector Calls
(Appendix 5, pp. 2579) gave an instance of a clause that is oriented
towards a certain type of modality: interpersonal meanings are highlighted, with the speakers skeptical doubting as the predominant
rhetorical function. The passage cited is a unique interplay between the
exploration of probabilities and the assertion of obligations, and so is
the entire text. No one clause can recapitulate the whole; but all
contribute, and some achieve a remarkable likeness a likeness that is
possible because the systems of the clause not only embody all the
semantic components from which the text is built but do so in a way
that allows an innitely varied, almost text-like balance among them.
Thus the properties that we recognize in a text are also, in a
transformed way, properties that we recognize in a clause. A clause is a
kind of metaphor for a text and a text for a clause. That this is
possible is due to two things: one, that a text is not only (typically)
larger than but also more abstract than a clause; two, that on the other
hand there is no line of arbitrariness between clause and text as there is
between clause and syllable. Hence it is not only in a formal sense that
a text is like a clause. It is no accident that it is possible to illustrate so
234
CLAUSE
conguration of
semantically grammatical
functions
Coherence:
by cohesion
(i) cohesive harmony:
chains interrelate by
function in (semantic)
transitivity
(ii) conjunctive relation:
between messages or
larger parts of text
by structure
(i) names (things)
interrelate by function in
(grammatical) transitivity
(ii) conjunctive relation:
between parts of clause, as
major or minor process
Function:
Structure:
Character:
1.2
generic: selects
favourite process type
as message type
specic: foregrounds one
or more systems
has information
structure:
theme - - - - >
- - - - > focus / new
A functional interpretation
wordclausetext
third person function, Malinowskis narrative function. And the opposition is incorporated into the semantics of natural languages, in the form
of what I have referred to as the interpersonal and ideational
components. (The distinction between rst and second person
language, however, is not a systematic one; the two are simply different
angles on the same interpersonality.) For all human beings, in all social
groups, the environment in which they live has these two validations:
it is something to be acted on, turned into food or shelter or other
needs; and it is something to be thought about, researched and
understood. Language has evolved to serve both these elementary
functions. The reective mode is coded directly as the ideational
element in the semantic system. But since language is symbolic, one
who speaks does not act on reality directly but only through the
intermediary of a listener. Hence the active mode, when translated into
a network of semantic systems, comes to be coded as interpersonal.
While these two functions are given to language from the outside,
as it were, by its role in human situations, in order to full such roles a
language has to have a third semantic component, whereby it is enabled
to latch on to those situations in a systematic way. There must be a
relevance function, a system of meaning potential which allows a text
to cohere with its environment, both the non-linguistic environment
and that part of the environment which consists of what has been said
before. So there is a third component in the semantics of natural
language which only an immanent linguistics will discover, since it has
no transcendent motivation; this is the contextualizing function or
the textual function, as I have called it, because it is what makes text
text, what enables language to be operational in culturally meaningful
environments.
Now a clause is a complex realization of all these three semantic
functions. It has an ideational component, based on transitivity, the
processes, participants and circumstantial elements that make up the
semantics of the real world, and including the onomastic systems that
classify these into nameables of various kinds. It has an interpersonal
component, consisting of mood, modality, person, key and all the
various attitudinal motifs that come to be organized as meaningful
alternatives. And it has a textual component, the functional sentence
perspective (thematic and news-giving systems) and the cohesive
resources of reference, ellipsis and conjunction. Each of these components makes its contribution to the total make-up of the clause. What
we identify as a clause is the joint product of functional-semantic
processes of these three kinds.
237
wordclausetext
succeeding clause. Thus a clause is at one and the same time particle,
eld and wave, as Pike suggested more than twenty years ago (Pike
1959), although the details of this interpretation are not quite the same
as those worked out by Pike.
Now, the signicance of this step in our interpretation lies not only
in establishing that these three distinct patterns of realization go to
make up the English clause, but also in the fact that they appear to be
non-arbitrary; this is clearly important when we come to ask whether
such tendencies are found in every language. The grammar of languages
is a natural grammar; as I expressed it earlier, there is no line of
arbitrariness between semantics and grammar as there is between
grammar and phonology. If the clause is at once particle, eld and
wave this is because the meanings it has to express have different
semiotic contours, to which these three realizational forms correspond
in a natural, non-arbitrary way.
The particular nature of ideational structures reects the relative
discreteness of the phenomena of our experience. Consider cows eat
grass: we know where the cow begins and ends, what eating is and is
not, which part of reality consists of grass and which part consists of
other things. Many of our perceptions are schematized into entities that
are bounded in this way, and the constituent-like form of the wording
reects this fact: the word cow has an outline because the object cow
has an outline. Of course not all experience is like this; indeed I have
always tended to emphasize the unboundedness of many phenomena,
the indeterminacy and the ux; and I share Whorfs view that language
itself, once it has been constituted in this way, strongly inuences the
forms that our perceptions take. Nevertheless there is a basic t between
the discreteness of words and the discreteness of things; otherwise we
should not be able to talk about the things at all, or explain contrastively
those instances where the t of words to things is less than perfect.
By contrast, the interpersonal kind of meaning is a motif that runs
throughout the clause; and this is represented by lexicogrammatical or
phonological motifs that are likewise strung unboundedly throughout.
The speakers attitudes and assessments; his judgements of validity and
probability; his choice of speech function, the mode of exchange in
dialogue such things are not discrete elements that belong at some
particular juncture, but semantic features that inform continuous
stretches of discourse. It is natural that they should be realized not
segmentally but prosodically, by structures (if that term is still appropriate) that are not particulate but eld-like. The linguists tree is an
inappropriate construct for representing structures of this kind.
239
wordclausetext
The Footman, just like the rest of us, favours himself as an unmarked
Theme.
So the patterns of wording in the clause, which is the basic unit of
lexicogrammatical organization, display a variation that derives from
the different kinds of meaning they express; and the structural shape is
in each case a natural product of the semantic functions. A functional
grammar is an interpretation of the primary semiotic purposes that
language has evolved to serve, and of the different ways in which
meanings relating to these different purposes tend to be encoded (and
the patterns just described are only tendencies). When we go on to
observe the developmental processes whereby young children construct
their language, we gain a further insight into the steps by which
grammar may have evolved on the way towards its present form.
wordclausetext
wordclausetext
wordclausetext
Appendices
Appendix 1
from Hasan (1980).
Text A
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Text B
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
wordclausetext
11. and the dogs chased the bear out the room
12. and the boy will sit down in their
----- chair what the bear was
sleeping in
Underlined items are those which enter into lexico-referential chains.
Broken underlining indicates that one item incorporates more than one
token, for example ----they referring to the girl, the boy, the sailor and
the dog. Empty underlining within parenthesis ( ), (-----) indicates a
token or tokens presupposed by ellipsis.
The number of lexico-referential tokens in the two texts is not very
different: 66 in Text A, 56 in Text B. But whereas 43 of those in Text
A (65 per cent) occur in chain interaction, the comparable gure for
Text B is only 20 (36 per cent). Text A thus displays considerably
greater cohesive harmony.
When subjects were asked to judge the coherence of the two texts,
Text A was consistently rated more coherent than Text B.
248
girl
girl
girl
girl
girl
girl
girl
girl
girl
girl
girl
girl
went
got
took
had
teddybear
teddybear
took-to-bed
fell-to-sleep
got-up
washed
combed
washed
brushed
lovely
dirty
teddybear
teddybear
teddybear
teddybear
teddybear
teddybear
speak
teddybear
teddybear
say
speak
words
English
Scottish
All-the-rest
bear
bear
go
come
go-to-sleep
sleep
bear
bear
chuck
chase
bear
bear
chair
chair
room
room
sailor
sailor
Each rectangle corresponds to one chain; subdivisions in the chain are indicated by boxing. Each box contains those
items which are in a constant functional relation (shown by a double-headed arrow) to items in some box in a
different chain; for example in Text A, between girl (4) and the box containing washed . . . brushed there is an actor
action relation; between the latter and teddybear (4), a relation of actiongoal.
Appendix 2
from James R. Martin (forthcoming)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
251
wordclausetext
252
Figure 2
253
wordclausetext
Appendix 3
from Peter Fries (1981).
1. A. The English Constitution that indescribable entity is a
living thing, growing with the growth of men, and assuming
ever-varying forms in accordance with the subtle and complex laws of human character.
2. B. It is the child of wisdom and chance.
3.
C. The wise men of 1688 moulded it into the shape we
know.
4.
C. but the chance that George I could not speak English
gave it one of its essential peculiarities the system of a
cabinet independent of the Crown and subordinate to
the Prime Minister.
5.
C The wisdom of Lord Grey saved it from petrication and
set it upon the path of democracy.
6.
C. Then chance intervened once more.
7.
D. A female sovereign happened to marry an able and
pertinacious man,
8.
D. and it seemed likely that an element which had been
quiescent within it for years the element of irresponsible administrative power was about to
become its predominant characteristic and change
completely the direction of its growth.
9.
C. But what chance gave chance took away.
10.
D. The Consort perished in his prime,
11.
D. and the English Constitution, dropping the dead limb
with hardly a tremor, continued its mysteric life as if
he had never been.
Queen Victoria, Lytton Strachey (p. 192)
[The above] is a well constructed paragraph which contains within it
three lexical systems; the rst concerns living, growing, changing, the
second system concerns wisdom versus chance and the third system
concerns concepts having to do with government. From reading the
paragraph it is clear that the main point of the paragraph is that the
English constitution is living, growing and changing, that the paragraph
is developed via the opposition between wisdom and chance and that
the lexical system having to do with government plays no particular
role within the structure of the paragraph. On examining the paragraph
one nds that the terms having to do with living, growing and
254
Appendix 4
From Robin Melrose (1979).
The remaining eleven sections deal with every aspect of life, regulating
it at every stage and aspect, ordering everything, forcing everything into
a symmetrical pattern: the cities are uniform, married life is strictly
controlled, education is minutely prescribed. Philosophy is conned
within rigid limits, the ne arts somewhat less so (. . .) This planned
paradise is enforced by drastic penal laws. Machinery of government is
paternalistic and pyramidal. It is based on division into families, tribes,
cities and provinces, and, in the case of the different crafts and professions, on units of ten. To each unit of work is assigned its master
(. . .) Each paterfamilias over fty is a senator, each family in turn
provides a tribal chief, each town in turn a city chief. Subordinate
senates of cities are controlled by the Supreme Senate. At the head of
the state is the General.
Totalitarianism, Leonard Schapiro (pp. 878)
255
wordclausetext
Theme:
The remaining eleven sections ; the cities ; married life ; education ;
Philosophy ; This planned paradise ; Machinery of government ; It ;
To each unit of work ; Each paterfamilias over fty ; each family ;
Subordinate senates of cities ; At the head of state
Lexical Cohesion:
Group A :
regulating ; ordering ; forcing ; uniform ; controlled ; prescribed ;
conned ; planned ; enforced ; penal laws ; paternalistic and
pyramidal
This is a particularly clear example of an attributive message. It
begins with a summation, every aspect of life, with the general noun
aspect acting as Head of a nominal group. It is this summation which
determines the Theme of the clauses that follow: thematic prominence
is assigned precisely to aspects of life, so that there is a relationship of
superordinate to hyponym between summation and Theme in the
message, reinforced by Theme in the last ve clauses, which is in a
relationship of hyponym to machinery of government, itself an aspect
of life.
There is no explicit conjunction of interest : the chief conjunctive
relationship is an implicit one, of the internal additive type. More
worthy of attention is the lexical string of Group A. Just as Theme was
determined by the summation, so the lexical string of Group A is
determined by the non-nite clauses dependent on the clause of which
the summation is an element. Together with the three verbs in these
non-nite clauses, nine lexical items of the message proper constitute a
string of synonyms, near-synonyms, and collocates and of these nine,
six function as Complement, and so form a kind of pattern in the
Rheme. Thus it may be seen that in this attributive message the
summation and the clause complex of which it is an element are both
closely related to the message proper that follows: the summation is
hyperonymously linked to the Theme, and the clause complex (or,
more precisely, the non-nite clause) is synonymously connected
with one exception to the Rheme. Or, to put it another way, it is
most often the case that aspects of life are encoded in the Subject,
while regulation is realised in the Complement.
256
Phenomenal
Relational
Type of
Process
material (doing,
happening)
mental (seeing,
feeling, thinking);
verbal (saying)
relational (being
attribute, identity)
Characteristic
Theme
main participant
cognizant / sayer
or phenomenon /
discourse
synonym or
hyponym of
summative element
Typical
Conjunction
external temporal
external or internal
additive or
temporal
internal additive or
adversative
Summative
Element
general noun, of
which message is
meronym
general noun, of
which message is
hyponym
general noun +
expansion, of
which message is
meronym
Appendix 5
from M. A. K. Halliday (1982).
Mrs. Birling: I think weve just about come to an end of this
wretched business.
Gerald:
I dont think so. Excuse me.
(He goes out. They watch him go in silence. We hear
the front door slam.)
Sheila:
(to Inspector) You know, you never showed him that
photograph of her.
Inspector:
No. It wasnt necessary. And I thought it better not
to.
Mrs. Birling: You have a photograph of this girl?
Inspector:
Yes. I think youd better look at it.
Mrs. Birling: I dont see any particular reason why I should
Inspector:
Probably not. But youd better look at it.
Mrs. Birling: Very well.
(He produces the photograph and she looks hard at
it.)
Inspector:
(taking back the photograph) You recognize her?
257
wordclausetext
polarity positive
modality low / (indicative : probability) / (subjective :
congruent)
Clause 2
259
indicative
(MODALITY)
it is
probability
temporality
it either is
it both is
or isnt
and isnt
modal
categorical
positive
high
median
low
imperative
(MODULATION)
do!
obligation
inclination
you do!
me do!
is
do
certain
must be
always
required
must do
desperate
probable
will be
usually
supposed
will do
keen
possible
can be
sometimes
allowed
can do
able
categorical negative
high:
must ought to need has/had to
median: will would shall should
low:
can could may might
isnt
dont
subjective, congruent: must &c. (modal auxiliary verbs)
subjective, explicit:
I think &c. (mental process clauses)
objective:
certain(ly) &c. (modal adverbs, etc.)
Chapter Ten
This chapter presents a brief sketch of a lexicogrammatical text interpreter for Modern English, in terms of systemic-functional grammar.
The grammar is in general neutral between spoken and written English,
but the text used for illustration is taken from spoken language; it is a
discussion among an adult and three nine-year-old schoolgirls. Here is
the text in standard orthography and punctuation (Hasan 1965: 65):
Adult: Do you when you have a small baby in the house, do you
call it it, or do you call it she or he?
Elsie: Well if its just if you dont know what it is I think you
ought to call it it, because you dont know whether youre
calling it a boy or a girl, and if it gets on and if you start
calling it she then you nd out that its a boy you cant
stop yourself cause youve got so used to calling it she.
Lacey: Em Mrs. Siddons says that if if some neighbour has a
new baby next door and you dont know whether its a he
or a she, if you refer to it as it well then the neighbour will
be very offended.
Tilly: Well if its in your family I think you should call it either
he or she or else the poor thing when it grows up wont
know what it is.
Adult: Well what did Mrs. Siddons suggest you should do if . . .
your neighbour has a baby and you dont know whether its
a boy or a girl?
First published in The Handbook of Discourse Analysis, Vol. 2: Dimensions of Discourse, 1985.
London: Academic Press, pp. 2956.
261
wordclausetext
Tilly:
Elsie:
Lacey:
Elsie:
Figure 1
263
wordclausetext
The text is analysed into clause complexes, showing the interdependencies and logical-semantic relations among their constituent (ranking,
nonembedded) clauses.
Clause complex 1
1.1 b
1.2 1
1.3 +2
b ^ a ( 1 ^ +2 )
||| when you have a small baby in the house ||
|| do you call it ||
|| or do you call it she or he ||
Figure 3a
264
265
wordclausetext
266
Theme
Each clause, and each clause complex characterized by rising dependency (b ^ a), is analysed for thematic structure:
Clause Theme
Clause
Textual
Interpersonal
267
Topical
Markedness of
topical theme
wordclausetext
268
Information structure
Figure 5
269
wordclausetext
In the New column, the focal element is outside the brackets. Square
brackets enclose other New material; within this, the items shown
within curved brackets are those previously mentioned.
The New is what the speaker presents as being for the listener to
attend to: this is whats news. It may be previously unknown, or
contrary to expectation, or picked out for special prominence. It is
realized by means of the tonic accent.
Phonologically, spoken discourse in English consists of a linear
succession of tone groups, each characterized by one intonation contour or tone. The tone group, in turn, consists of a tonic segment that
carries the characteristic tone contour: 1, falling; 2, rising; 3, level
(phonetically realized as low rising); 4, fallingrising; 5, risingfalling;
13, falling plus level; 53, risingfalling plus level. The tonic segment
begins with the tonic accent, which embodies the distinctive pitch
movement. Optionally, the tonic segment may be preceded by a
pretonic segment that also forms part of the same tone contour. Both
tonic and pretonic segments display a range of more delicate options
within each tone: wide fall, narrow fall, low pretonic, high pretonic,
and so on.
Grammatically, the tone group realizes a unit of information, which
is one piece of news, so to speak. It consists of a New component
optionally accompanied by a component that is Given. Typically, the
New comes at the end; but unlike thematic structure, information
structure is not realized by the order in which things are arranged, but
by tonic prominence the New is the element containing the tonic
accent. The particular word on which the tonic accent falls is said to
carry the information focus. Anything after the focal element is thereby
marked as Given, while anything preceding it may be Given or may
be New (there are rather subtle intonational and rhythmic variations
serving as signals).
An information unit is not necessarily coextensive with a clause, but
that is its unmarked status: each ranking clause (i.e. independent or
dependent, but not embedded) is typically one information unit. The
principal systematic variants are (1) two clauses / one information unit:
failing dependencies, that is, a ^ b sequences; (2) one clause / two
information units: thematic focus, that is, // Theme // Rheme //
information patterns.
Analysis of the New elements will reveal the main point of the
text. In the example it is to do with babies, what sex they are, and
how they are to be referred to in cases of doubt. Subsidiary points of
270
attention are the babys growing up, the childrens understanding and
obligations, and the adult worlds possible displeasure.
The Theme in a clause is what is prominent for the speaker; it is
what I am on about. The New in an information unit is what is
made prominent (by the speaker) to the listener; it is what you are
being invited to attend to. When clause and information unit are
mapped on to each other, the result is a wavelike movement from
speaker to listener, the diminuendo of the speakers part being as it
were picked up by the crescendo of the listeners part.
The effect of this movement is cumulative over the text as a whole.
The present text is typical in the way that the sequence of Themes
represents the method of development of the dialogue, while the
sequence of News represents the main point of the discussion, with
each speaker contributing her part to the construction of the overall
pattern all unconsciously, of course.
This interplay of two distinct waves of prominence is possible
because ThemeRheme and GivenNew are not (as often conceived)
one single structure, but two distinct structures that interpenetrate. As
a result, they can vary independently, allowing for all possible combinations of the two kinds of texture. In unit 2h, for example, Elsie might
have chosen a different distribution by combining thematic and information prominence (mapping New onto Theme):
cause youve got //.1. so / used to / calling it / she //
This would have strongly highlighted used to, as a marked focus, and
marked calling it she as Given; the effect of the latter would have been
to bring out the repetitive facet of calling it she, thus reinforcing its
cohesion with 2e, but by the same token to deprive it of its status as a
main point for attention. The interaction of the thematic and informational systems is the clause grammars contribution to the creation of
texture in discourse.
Mood
Each nite clause is analysed for mood; its Subject and Finite element
are shown, together with any modality:
271
wordclausetext
Figure 6
Figure 7
wordclausetext
This is typical of logical argument; and it is interspersed with interrogatives as the adult prompts the children to explore further.
There is very little modality in the text. In the Finite element, apart
from one ability form cant, in you cant stop yourself, there are just
three expressions of obligation (ought and two instances of should).
Other than in combination with niteness, there are again only three
modalities, in this case expression of probability: (I suppose and two
instances of I think). As it happens, in two paired instances the two
kinds of modality are associated: where the speaker expresses a judgement of obligation, she qualies it with a judgement of probability, it
may be that it should be so. Thus when the children are making rules,
they are also being tentative about them.
274
Figure 8
wordclausetext
possessive / attributive:
intensive / attributive:
intensive / identifying:
cognitive:
verbal:
have + baby ( 3)
be + male / female ( 5)
call+ baby+ he / she ( 11)
know, nd out ( 8)
say, suggest ( 3)
The remaining six clauses include three characterizing the baby (two
material: grow, get on; one circumstantial: be in + family), one characterizing the neighbour (affective: be offended), and two others, one minor
(greeting: hello) and one a WH- process (do what).
There are three major types of process in English: Type I, doing
(material and behavioural); Type II, sensing / saying (mental and
verbal); Type III, being (relational and existential). They are distinguished in the grammar in various ways; the principal distinctions are
as follows:
Figure 9
276
In the example text, the clauses in (iiii) above are all type III,
relational; the issue is one of being, partly attribution (having a baby,
which in this text means possessing it, not bearing it; being male or
female) and partly identication (being the name of). Of these, (i) and
(ii) are middle there is a Medium (the neighbour, the baby) but no
Agent; (iii) however is effective there is an Agent in the identication
process, always represented as you, but moving from personal you in
clause complex 1 to impersonal you thereafter. The clauses in (iv)
and (v) are type II, mental-verbal; (iv) are cognitive, with the Medium
(Senser) being you, I, Mrs. Siddons or the baby when it grows up; (v)
are verbal, the Medium (Sayer) being Mrs. Siddons. All interactants,
real and hypothetical, are involved in thinking about the problem
including the baby, at some future date, if a solution is not reached
now; and the teacher has put the problem into words.
A summary of process types and the relevant participant functions
follows:
Figure 10
277
wordclausetext
Groups and phrases are analysed with respect to features that are
relevant to the inquiry:
Figure 11
possible combinations. In the full (nite) system of tenses the number is 36; from this is derived, by a neutralizing of certain tenses in
the past series (he said she had arrived corresponds to she arrived, she
has arrived, and she had arrived), the 24-tense sequent system; and
from this in turn, by a parallel neutralization in the future series (to
be about to depart corresponds to will depart, is going to depart, and will
be going to depart), is derived the non-nite system which has just
12. This last is also the system that applies if the Finite verbal element
is a modality (e.g., should, must), since that eliminates the primary
tense choice.
Most of the verbal groups in this text are simple present tense; not
only because of the general nature of many of the propositions, but
because most of the processes are other than doing ones, and
therefore have simple present, not present in present, as their unmarked
choice. Furthermore, even the material ones are dependent (if it gets
on, when it grows up), which again requires simple present.
Nominal Group. The only nominal groups with structure other than
simply Head / Thing are the following:
Figure 12
wordclausetext
Figure 13
280
Figure 14
wordclausetext
Grammatical metaphor
his nose is long. One very common type of these is the use of a
mental process (cognitive) clause to express a modality, such as I think,
I dont believe. It was pointed out above that the tagged form of I think
its broken is I think its broken, isnt it?, not I think its broken, dont I?,
showing that this is a metaphor for its probably broken. We can use one
of these as an example:
|| I dont suppose
mental: cognitive
negative
reworded as:
||| she probably didnt come
material
modalized:probability/median
negative
|||
(single clause)
In cases like this it saves time if the analysis moves directly to the
nonmetaphorical version, since the rewording is quite automatic. In
other instances, however, the principle is as follows:
1. Analyse the clause as it stands.
2. Reword it, in nonmetaphorical form.
3. Analyse the reworded version.
Both analyses gure in the interpretation. Sometimes it takes several
steps in rewording to reach a nonmetaphorical version, and there may
be more than one possible route; all are potentially relevant to an
understanding of the text.
10
Context of situation
wordclausetext
Tenor. Adult and three children: adult (neither teacher nor parent)
interviewer, informal; children self-conscious but relaxed.
Speech roles: adult questioning, children suggesting.
Mode. Informal spontaneous speech. Dialogue: question-andanswer exchanges. Exploratory; hypothetical and logical in
orientation, moving towards (partly humorous) resolution.
These features determine the choice of register: that is, the kinds of
meanings that are likely to be exchanged. Like the rest of the linguistic
system, the patterns are probabilistic: given these features of the context of situation, we can make semantic (and therefore lexicogrammatical) predictions with a signicant probability of being right that,
after all, is precisely what the interactants themselves are doing all the
time.
What makes this possible is what makes it possible for a child to
Figure 15
284
learn the language in the rst place: the systematic relationship between
these categories of the situation and the metafunctions of the content
system. By and large, characteristics of the eld predict experiential
meanings, those of the tenor predict interpersonal meanings, and those
of the mode predict textual meanings.
In analysing a text, we identify those features of the lexicogrammar
which in a text-generation program might reasonably be expected to
be called upon if the situation was represented in this way.
11
Conclusion
Three nal points should be made about an outline of this type. One
is that it is just an outline: Obviously, the analysis under every heading
could be developed much further in delicacy, and other headings could
be added. The guiding principle is to select and develop whatever is
needed for the particular purpose in hand. There are many different
purposes for analysing a text, and the scope and direction of the analysis
will vary accordingly. Often we may want to scrutinize only one or
two features, but to follow them through to a considerable depth.
Secondly, a text analysis is a work of interpretation. There are
relatively few absolute and clearcut categories in language; there are
many tendencies, continuities, and overlaps. Many actual instances can
be analysed in two or more different ways, none of which can be ruled
out as impossible; some may be less sensible than others, and so can be
discarded, but we may still be left with valid alternatives. Especially in
a literary text it is to be expected that we will nd clauses with multiple
grammars; but this is not a distortion of the system it is a richer use
of its natural resources. All analyses may need to gure in the
interpretation.
Thirdly, the lexicogrammatical analysis is only a part of the task. It is
an essential part; all text is made of language, and the central processing
unit of the linguistic system is the lexicogrammar. But just as the
grammatical system does not itself create text text is a semantic
creation, with the grammar functioning largely (though not entirely) as
the automatic realization of the semantic choices so the analysis of
the grammar does not constitute the interpretation of a text. (There
has been some misunderstanding on this point, for example in the use
of cohesion as a method of text analysis. Cohesion is an essential
property of texts, but it is the way the cohesive resources are deployed
that makes the difference between text and non-text, and between one
text and another.)
285
wordclausetext
286
SECTION THREE
CONSTRUING AND ENACTING
EDITORS INTRODUCTION
one is doing; so when one talks about something, one tends to say that
it happened or was done. By means of grammatical metaphor, written
language symbolically distances the act of meaning and its counterpart
in the real world. Each plays complementary roles when it comes to
using language to acquire knowledge and reect on ones experience.
Halliday maintains that this complementarity must gure into any
attempt at developing a linguistic theory of learning.
In How do you mean? (Chapter 13), appearing in Advances in
Systemic Linguistics: Recent Theory and Practice (1992), meaning is taken
as a mode of action occurring at the intersection of the conscious and
material modes of experience. In particular, Halliday examines the
evolution of protolanguage into language, or how the two-dimensional
semiotic that denes the mammalian experience evolves into a semiotic
of a new kind: a stratied, tri-stratal system in which meaning is twicecooked, thus incorporating a stratum of pure content form. The
two dimensions of protolanguage, a minimal semiotic system, include
the inner dimension of reective / active, I think as against I want,
and the outer dimension of intersubjective / objective, you and me
as against he, she, it . The third dimension results from the introduction of grammar, or as Halliday describes it: a purely symbolic mode
of being between these two interfaces. The processes of instantiation
and realization make possible this dynamic open system we call language.
In the two nal works in this section, Grammar and daily life,
which rst appeared in Functional Approaches to Language, Culture and
Cognition (2000), a Festschrift for Sydney Lamb, and Grammar and
grammatics, published as Volume 121 of the series Current Issues in
Linguistic Theory (1996), Halliday describes how grammar enables us,
unconsciously, to construe our reality, and interpret our experience,
while grammatics makes it possible for us to reect consciously on how
this theory of our human experience works. Halliday introduced the
term grammatics, in contradistinction to grammar, to distinguish
between a particular stratum of a natural language and the study of this
stratum. Thinking grammatically, or using grammatics to think
about what grammar thinks about the world, may help us better
understand this grammatical energy or grammatical logic that powers
language and also conditions our attitudes to each other and to the
world around us. To be a linguist, Halliday writes, is inevitably to
be concerned with the human condition, and those who think
grammatically will be better prepared not only to address issues of
social injustice and inequality, but also to contribute to the development of new applications of linguistics such as intelligent computing.
290
Chapter Eleven
291
semantic. The same is true of the pair sentence and discourse. But
with the majority of grammatical terms those for functions, systems
and terms in systems (features) this does not hold. These too have to
be imported into grammar from outside; but in this case they come
not from folk semantics but from outside language altogether. When
categories of this type come to be named, the terms that are introduced
for the purpose interpret the categories by reference to some aspect of
extralinguistic experience.
A typical example would be a complex of categories such as (system)
number: (features) singular / plural. Consider a label such as plural.
This derives from the ideational meaning of the category: it expresses a
relation between that category and the speakers experience of the
world. The term plural is the name of this relationship. But the same
term is also used as the name of the grammatical category which realizes
this relationship; a noun will be said to be plural (in number). And
this can cause problems.
Typically in the history of western linguistics the reasoning has
proceeded as follows. In the morphology we are presented with a
certain category, let us say a two-term system, of the form a : x/y;
thus, all a are either x or y. By inspecting, or perhaps introspecting,
typical contexts in which these forms are used we recognize a redundancy, such that x redounds with one of a set of countable things and
y redounds with two or more. The abstract labels a : x/y are then
replaced by the substantive labels number : singular / plural. This is
of course an idealized model of the process; such abstract labels have
never been used, as far as I know, at least until modern times. But it
helps to bring the issue into focus.
What happens next? The categories that have been labelled in this
way then come to be reied and the question is asked what they mean.
The answer given is: singular means one of a thing, plural means more
than one. In fact, these are denitions of the words singular and plural;
but they are made to serve as denitions of the metawords, the terms
of the metalanguage. Next, the terms are evaluated for their predictive
power: will they correctly predict text from situation, or situation from
text? Given a plural form, will it refer to more than one of a thing?
Given more than one of a thing, will it be referred to by a plural form?
The answer this time is: yes, with a certain degree of probability
high enough for many purposes, but inadequate for some, and disturbing for those who like their categories pure. This then gives rise to a
theory of core meanings: a term of the metalanguage is said to
represent the core meaning of the category. With this defence, in
295
spite of fruit and furniture, scissors and oats (or their counterparts in the
language in question), singular and plural can continue to be used.
This problem that of interpreting a symptom and then labelling its
interpretation is common to all sciences. It arises in any realm of
discourse involving explanation and abstraction. Somehow, a metalanguage has to be created, and created out of natural language, in order
to assign a Value to a Token. I go to my doctor with a swelling
somewhere on my body. He looks at it, and pronounces Youve got
an oedema. What is oedema? Greek for swelling. But what he is
diagnosing is a more abstract swelling it is that which is manifested
by the swelling on my body. An oedema is the Value of which a
swelling is the Token. The use of a different term in this way allows
for stratal diversication: not all swellings realize oedemata, and not all
oedemata are realized as swellings. The relationship is a probabilistic
one and hence invites further, more delicate investigation.
Where do such terms come from? Ideally, they come from a parallel
but distinct semiotic. It should be a natural language, in order to
maintain the non-arbitrariness of the relation between the symptom
and the underlying condition given swelling and oedema we would
predict that, in default of any special circumstances, they will refer to
the same thing. But it should be a different language, or at least a
different sub-language, in order to allow for instances where they do
not. And it should be a higher status code, in order to symbolize a
higher order of abstraction (and also the social value of abstract
knowledge). The ideal source of a metalanguage is thus the high
variant in a diglossia. A word of a natural language that is at one
remove from primary reality, such as ancient Greek, or classical
Chinese, or Sanskrit, is appropriate for symbolizing a phenomenon that
is at one remove from primary observation.
But when it comes to metalinguistic matters, linguistics presents a
special case. It is not just another science. It is language turned back
on itself, to use Firths (very British) expression; or, in Weinreichs
(very American) formulation, language as its own metalanguage. As a
consequence, where other sciences need two terms, we need three:
one for the phenomenon, and two for the metaphenomenon, one
grammatical and the other semantic. To return to the example of
number: we need to be able to say that the grammatical category of
plural typically expresses (realizes) the semantic category of, say,
manifold, which typically expresses (redounds with) more than one of
a thing.
But notice what has happened. The grammatical category of plural
296
was set up in the rst place to account for a morphological phenomenon: suppose this had been in English, then the s / z / iz of cats,
dogs and horses. At this point, therefore, we ought to have come round
in a circle: s / z / iz means s / z / iz. But instead we have
tried to escape from the circle by nding a gloss for s / z / iz
that is, an exact synonym for it, in natural language wording; and that
is an extremely difcult thing to do. We might try glossing it as more
than one, or several, or many; but the trouble is we dont actually say
I like more than one cat, or I like many cat we say I like cats. The
meaning of the s on cats is impossible to gloss in natural language,
except by means of itself. The category is, quite simply, ineffable.
Why should this be so? One hypothesis might be that natural languages
are not good things for glossing with; in that connection, Reddy
remarked, As a metalanguage, English, at least, is its own worst
enemy. We can certainly point to some deplorable habits that English
has, both in its vocabulary and in its grammar. For example, we
frequently use the same lexical item to stand both for the study of a
particular phenomenon and for the phenomenon itself, as when we
talk of someones psychological make-up instead of their psychic make-up.
It can be disastrous for students of linguistics (not to mention the
general public) that grammar is both the name of a stratum in language
and the name for the study of that stratum; and likewise with phonology
and semantics. Not even the conduit metaphor excuses a ragged
polysemy such as these.
Even worse are some of Englishs grammatical pathologies. For our
metalinguistic vocabulary, we usually draw on some parallel semiotic as
already illustrated, bringing in new words so as to be freed from the
accumulated associations of the old ones. (The freedom is often shortlived, since the new term may soon be borrowed into the daily
language, like the psychological above.) But for the grammar of our
metalanguages we are usually content to stick with the everyday forms
of English; and this can lead to serious misconstructions such as the
following, perpetrated by myself, when I wrote some time ago:
the Theme of an English clause is the element that is put in rst
position.
Now I meant this as Value ^ Token, with is meaning is represented
by. But all such clauses in English, if they have the verb be, are
297
298
It is only a very short step from here to the assertion that the Subject
has no meaning. The implication is: whatever it is that is functioning
as Subject in any instance has meaning as actor, or has meaning as
topic; but as Subject it has none the category of Subject has no
meaning in itself. In this view, Subject is a grammatical function whose
only function is to be a grammatical function.
Such a view is enshrined in the terminology, in the term grammatical Subject (used for example in Sweet 1891); this is in contradistinction to logical Subject (i.e. Actor) and psychological Subject (i.e.
Theme). Compare the later Prague school interpretation, with syntactic structure (SubjectPredicate) contrasted with semantic structure
299
One project in which they play an essential part is the Penman project
at the Information Sciences Institute (Mann and Matthiessen 1983;
Mann 1983). In Penman the motive power is provided by the
grammar, Nigel, which is a systemic grammar consisting of a network
of some hundreds of options. At each choice point a Chooser is
activated; the Chooser consults the environment (the Knowledge Base)
for instructions on which way to go. The Choosers questions are
referred to an Inquiry Operator; and they take forms such as the
following:
Is this concept inherently multiple, i.e. a set or collection of things,
or is it unitary?
Is the process one which conceptually has some sort of entity which
causes the process to occur?
Does this represent a concept which the speaker expects the hearer
to nd novel, not previously mentioned or evoked, and thus does
not expect the hearer to identify uniquely by reason of the attention
which it currently holds, its inherent uniqueness in culture, or its
association with an identiable entity?
All such glosses are attempts to get at the grammar beneath the skin;
and they may be supported by a variety of different beliefs. First, it
may be assumed that all grammatical generalizations have some signicance at a higher stratum; or alternatively, that some are simply housekeeping devices and have no semantic function. Secondly, those
grammatical categories that are regarded as semantically signicant may
be thought of as universal, or as particular to the given language, or as
particular to a given register, a functional variety of a language. (These
would represent fairly well the respective views of Jakobson, Hjelmslev
and Firth.) Thirdly, it may be held that every such category has one
meaning that is common to all its manifestations, and the problem is to
nd the right semantic generalization to cover all cases; or alternatively,
that some categories at least are polysemous, so that their meaning
varies in ways that are not predictable from the context (cf. Ikegami
1980: 59). Fourthly, there is a range of beliefs about the place of
grammar, and the need to postulate some higher level semiotic system
(semantics, semology, the conceptual level, etc.) to which grammatical categories can be related in a systematic and in some sense
natural way. Positions taken on these issues may complicate the task
of semantic interpretation: for example, if categories are assumed to be
universal, and yet are established at an insufciently abstract level.
But whatever beliefs are held about them, grammatical categories
301
will remain ineffable. Some of the more recalcitrant ones are categories
that Whorf originally called covert: having no mark other than
distinctive reactances with overtly marked categories. But by no means
all are of this kind there is nothing covert about deniteness in
English, for example. The phenomenon we are concerned with has
more to do with Whorfs follow-up notion of a cryptotype.
Whorf remarks of these that they easily escape notice and may be
hard to dene, and yet may have profound inuence on linguistic
behavior (1956: 92). Among cryptotypes in English Whorf cites
gender, transitivity (of the verb), inherence (of the adjective), and
various more delicate categories, such as that of verbs that may be
phrasalized with up. There is, of course, a connection between the two
senses in which Whorf is using crypto-: a category may be hard to
dene precisely because it is hidden from view. But hidden from whose
view? It is not because they are hidden from the linguist that grammatical categories are hard to dene; once the linguist has found them, the
fact that they had escaped his notice ceases to matter. The signicance
of this concept of a cryptotype is that it is something that escapes the
notice of the speakers of the language.
Franz Boas long ago drew attention to the unconscious nature of
language, contrasting it in this respect with the other meaning systems
of a culture; and although his observations have often been quoted, it
seems to me that their signicance is seldom fully taken into account.
There is a fundamental relationship between the unconsciousness of
language and the nature of its semantic categories. I have often pointed
out, in the many years since I began the study of informal speech, that
it is only in the most spontaneous, un-self-monitored kinds of discourse
that a speaker stretches his semantic resources to the utmost (cf.
Halliday 1966). This does not happen in formal speech; and it certainly
does not happen in writing. It is in unconscious spoken language that
we typically nd the truly complex sentences, with their labyrinths of
hypotaxis and all their projections and expansions, from which, while
we blunder through such sequences often losing ourselves completely
when we are engaged in the planned self-monitoring discourse of an
academic lecture, we emerge in good order and with every node
unravelled provided we are completely unaware of what we are saying
and attending only to whatever it is we are involved in at the time.
(That sentence is best taken orally, at high speed.) Our ability to use
language depends critically on our not being conscious of doing so
which is the truth that every language learner has to discover, and the
contradiction from which every language teacher has to escape.
302
child. Semogenesis begins well before the mother tongue, as the infant
creates his own protolanguage or child tongue: he constructs this
symbolic system, in interaction with those who share in his meanings,
for the twofold purpose of doing and thinking that characterizes all
such systems; and in the same process he also constructs the objects of
his action and his reection. But the child tongue has its limitations,
for both these purposes; so he moves into the adult mode, and takes
over the mother tongue with its ready-made grammatical categories.
The symbols of the mother tongue, which have been around him from
the start, now become his reality, at once a part of, and a key to, the
complex phenomena of his experience. Language and culture are
construed as one.
Does a child, then, know what a Subject is? We cannot ask him;
nor can we set up a test situation to nd out if only because children,
given an unnatural task, will respond with unnatural behavior. (It is
not intended to suggest that there is any contrast here with adults. The
problem lies again in the unconscious nature of linguistic processes,
which adults cannot reproduce experimentally either.) Nevertheless it
is clear, surely, that a child does know what a Subject is, because he
uses one a hundred times a day. We only have to listen to a ve-yearold in ordinary, real life situations, and we will hear the categories of
the grammar that we nd most difcult to explain, deployed in their
appropriate semantic roles.
What we observe, of course, even with a tape recorder on permanent duty, is only a limited set of instances. We have to infer the
system that lies behind them; for language (if I may be allowed to
invert Chomskys famous dictum) is an innite system that generates
only a nite body of text. But what we can observe is already very
convincing.
If I assert that a ve-year-old knows what a Subject is, it is because
I have listened to children for many years, and heard them talking in
clauses which have Subjects. In my own detailed record of one
particular child, there are about 2,500 of them; but since child language
studies became fashionable there has been an abundance of such
material available, if one does not feel one can rely just on ones
listening. Now, any one of these clauses could have had the appropriate
Subject by chance. Moreover, since no linguistic category is chosen in
isolation in choosing the Subject one is always making other choices
besides (and this will apply whatever category is used as illustration)
in any one instance we could always claim that the appropriateness of
the Subject was a consequence of some other choice. But if countless
304
Subject
Actor
Theme
I
I
I
I
I
ball
I
ball
I
I
ball
ball
The speaker keeps the listeners attention by varying the Theme and
the Actor:
1
2
3
4
Theme
Actor
about me (I . . .)
about me (I . . .)
about the ball (the 3rd)
about the ball (by the 4th)
Clause 1 has the speaker in all three functions of prominence: interpersonal (I as Subject), ideational (I as Actor) and textual (I as Theme).
Clause 2 is marked by ideational modesty (this is what happened to
me, not what I did), and clause 3 is marked by textual modesty (now
Ill tell you about the ball, not about myself). In clause 4, the speaker
gains further merit by ending on a doubly modest note, in which he is
neither the Actor nor the Theme.
But in regard to the speech function, the picture is quite different.
The speaker retains himself in the role of Subject throughout. There is
no sign here of interpersonal modesty; the assertion is made to rest on
I every time, and the listeners response, correspondingly, must always
have a you in it Did you?, Were you? In other words, every step in
308
my aunt
Subject
What I have been trying to show with this illustration is that while,
with a category like Subject, it is impossible to answer the question
what does it mean?, this does not signify that it has no meaning. The
problem of ineffability is common to all grammatical categories; there
are various reasons why some may seem less problematic than others,
but it is an illusion to think that any can be exhaustively dened. And
this, as I remarked above, is not because of the shortcomings of natural
language for serving as a metalanguage, real though such shortcomings
are. Rather the converse: it is the very richness of natural language, its
power of distilling the entire collective experience of the culture into a
single manageable, and learnable, code that puts its categories beyond
the reach of our conscious attempts at exegesis.
This leads us back to the question of the Grundbedeutung. The
categories we have been considering have been categories of the
grammar: grammatical systems and structures, and their component
309
features and functions; and since grammar is the central processing unit,
where meanings of different kinds are brought together as wordings,
we expect its categories to be valid for the language as a whole. All
uses of English involve Mood : Subject + Finite, or tense : past /
present / future; and these are assumed to be in some sense the same
thing in all contexts, since otherwise we would not be looking for
denitions of them.
No such constraint gures in our conception of semantics. The
grammar is the grammar: it has internal organization, of a metafunctional kind; it has some special purpose sub-grammars; and it has
considerable indeterminacy but there is such a thing as the grammar
of English. We do not operate with a separate grammar for each
register. No doubt we can also conceive of such a thing as the
semantics of English; but we also feel that (at least at the present state
of our knowledge) it is not counter-productive to envisage a more
restricted domain for semantic generalizations, and to operate with
semantic sub-systems each relevant to a specic universe of discourse.
In principle, the domain of a semantic description may be anything
from the whole language down to a single text. At one end of the
scale, I have found it useful to set up a semantic system relating to just
one dialogue of 35 words long; this was a childadult dialogue, and the
purpose was to explore what meaning potential the child must have in
order to be able to construe such a discourse (see Appendix, p. 313).
Geoffrey Turners (1973) semantic networks dene a rather broader
range of texts, such as motherchild control patterns in specic experimental situation types. More general again is Ruqaiya Hasans (1983)
message function network, which describes spontaneous interaction
between children and parents, for the purpose of investigating the
development of childrens learning patterns. At the other end of the
scale, J. R. Martins (1983a, 1992) conjunction networks are like
grammatical networks in that they are set up for the language as a
whole.
When we describe semantic systems, we are saying what it is that
preselects the grammatical categories: what choices in meaning call on
what features in the grammar for their realization. It is by this process
that the grammatical categories are dened; when this is done, there is
no need to gloss them further. Once the semantic system is made
explicit, it can only be misleading to attach separate semantic descriptions as glosses to the categories of the grammar.
At the same time, if the semantic system is set up only for a restricted
domain, some particular register variety, then the meanings of any
310
4. Perhaps as a combination of the metaphoric and the metonymic we should cite Ecos recent novel The Name of the Rose
(1980), where instead of using language to talk about literature
he turns the tables and uses literature to talk about language
Whether or not language has the property that is sometimes claimed
for it, of being able to interpret all other semiotic systems (and I see no
reason to assume that this is so), there are certainly limitations on the
ability of language to interpret itself. We may have to move outside
language, to some parallel or higher order semiotic which, since it is
not itself language, can be represented in language and then refracted
to become a metalanguage for representing language. All such interpretation is ultimately circular; but in linguistics, we have tended to
operate within circles that are pathologically small. Until we can create
a greater distance between the semiotic object and the metasemiotic,
grammatical categories are bound to remain ineffable.
Appendix
Nigel at 5;4 (from Halliday (1984))
Nigel: Shall I tell you why the North Star stays still?
Father: Yes, do.
Nigel: Because thats where the magnet is, and it gets attracted by the
earth; but the other stars dont so they move around.
//2 shall I / tell you / why the / North / Star / stays / still//
//1 yes // .1 do//
//4 . because / thats //1 where the / magnet / is and it //1 gets at/
tracted //1 by the / earth //4 . but the / other / stars //4 dont so /
/4 they //1 move a/round //
1.
Ideational
1.1
Experiential
tell
you
Sayer
Process: Receiver
Medium Verbal Beneciary
why
Cause Carrier
Medium
Actor
313
still
Process: Range
relational
intensive
Process: material
where
Identied Process:
Token
relational
circumstantial
Medium
it
the magnet is
Identied
Value
they
move
around
Actor
Process: Place
material
Medium
Attribute Carrier
Range
Medium
Process:
relational
circumstantial
gets attracted by the earth the other stars dont [get attracted by the earth]
Goal
Process:
Medium material
Actor
Agent
Goal
Medium
Process: material
Actor
Agent
Process type
Process
Medium
do
stay still
be at relational: identifying
circumstantial: locative
is (represents)
material: middle
4
be at relational: attributive
[[ ]]
circumstantial: locative
Other elements
Location: at
North Star
do to material: effective
attract
do to material: effective
(not) attract
do
move
other stars
1.2
Logical
material: middle
Network: experiential
clause
PROCESS
TYPE
VOICE
[Nigel] 12
[Father] 3
[Nigel] 47
material
relational
verbal
31
32
33
effective
middle
41
42
33.42
31.42
33.42
32by.41 [[3ay.42]]
31.41
attributive
identifying
a
b
intensive
circ.: place
possessive
x
y
z
31.41
31.42
Network: logical
clause
complex
INTERDEPENDENCY
(TAXIS)
paratactic
hypotactic
expansion
LOGICAL-SEMANTIC
RELATION
123...
abg...
alaborating
extending
enhancing
=
+
locution
idea
projection
cause?
magnet be-at
cause:
add (effect): earth attract
earth not attract
contrast:
move
effect:
North Star
North Star
other stars
other stars
2
2.1
Ideational semantics
Experiential
reactants (magnet)
North Star
objects
stars
bodies
other stars
earth
PHENOMENA
move
motion
Moving (do)
not-move
(stay still)
force
alternatively:
moving
(do)
2.2
positive
negative
uncaused (motion)
caused (force)
Logical
contrastive (but)
additive
LOGICAL
SEMANTIC
RELATIONS
positive (and)
effect (so)
causal
unknown (why)
cause
known (because)
316
Interpersonal semantics
COMMODITY
ROLE
(proposal)
goods-&-services
(proposition)
information
give
offer
statement
demand
command
question
give
ROLE IN
EXCHANGE
demand
tentative (shall I? / will you?)
VEIN
COMMODITY
EXCHANGED
goods-&-service
(proposal)
information
(proposition)
Interpersonal
Finite
Subject
Mood
tell you
Residue
do
Finite
mood
why
the North
WH/
Adjunct
Resi-
Subject
Mood
Star
stays
still
Finite
due
that
s
317
it
the other stars
they
gets attracted
dont
move around
Subject
Mood
Finite
Residue
Network: interpersonal
declarative
indicative
01
yes/no 02
interrog
major
clause
WH- 03
Mood
1st person
imperative
1st/2nd persons 05
2nd person
MOOD
PROJECTION
ELLIPSIS
04
direct
11
indirect
12
full
21
06
elliptical 22
[Nigel] 1
[Father] 3
[Nigel] 2
02.11.21
06.11.22
03.12.21
318
statement
Textual
3.1
Intertopical
personal
Theme
tell you
why
Rheme
Intertopical
personal
Theme
Rheme
because
that
and
it
but
dont
so
they
move around
structural
Theme
topical
Rheme
(topical) [speaker]
319
New
by the earth
Focus
Focus
Given
New
New
New
New Given
Focus
dont
Focus
New
New
Focus
New
unmarked
TOPICAL
THEME
TONICITY
marked
320
so they
Focus
move around
New
marked focus
INTERPERSONAL
THEME
TEXTUAL
THEME
3.2
{
{
mood nite
WH-
modality
TONALITY
STATUS
vocative
continuative
structural
conjunctive
contrastive
Cohesive
stays still
S A
S H
the other stars
|R
1
move around they
SA:
SH:
C:
E:
R:
Key:
synonymy antonyms
synonymy cohyponyms
collocation
ellipsis
reference
dont
1:
Process + Medium
2:
Process + Location
3:
Process + Agent
columns: lexical chain
boxes:
referential chain
321
4.
4.1
Field
Context of culture (the system)
4.2
5.
5.1
Tenor
Context of culture (the system)
5.2
Child (5 years) and parent interacting: child (i) makes explicit and (ii)
seeks approval for interaction to impart knowledge:
(a) displaying knowledge (boasting)
(b) seeking conrmation.
Parent approves: child proceeds to do so (would have done so anyway).
322
Chapter Twelve
323
Already half a century earlier Franz Boas (1911) had stressed the
unconscious character of language, unique (as he saw it) among the
phenomena of human culture. Boas observation was to be understood
in its contemporary context as a characterization of the language system
(langue); not that, writing in 1911, he could have read Saussures Course
in General Linguistics, any more than Mathesius could have done; but
the unconscious was in the air, so to speak, and playing a critical role
in the conception of systems as regularities underlying human behaviour. But Boas may also have had in mind the unconsciousness of the
behaviour itself: the act of speaking (acte de parole) as an unconscious
act. The lack of conscious awareness of the underlying system, and
the difculty that people have in bringing it to consciousness, are things
which language shares with other semiotic systems for example, social
systems like that of kinship; what is unusual about language is the
extent to which even the manifestation of the system, the actual
process of meaning, remains hidden from observation, by performer
and receiver alike. In that respect talking is more like dancing, or even
running, than it is like playing chess. Speaker and listeners are of course
aware that the speaker is speaking; but they are typically not aware of
what he is saying, and if asked to recall it, not only the listeners but
also the speaker will ordinarily offer a paraphrase, something that is
true to the meaning but not by any means true to the wording. To
focus attention on the wording of language is something that has to be
learnt for example if you are studying linguistics; it can be a difcult
and somewhat threatening task.
About 30 years ago, as a result of being asked to teach English
intonation to foreign students, I began observing natural spontaneous
discourse in English; and from the start I was struck by a curious fact.
Not only were people unconscious of what they themselves were
saying; they would often deny, not just that they had said something I
had observed them to say, but also that they ever could say it. For
example, I noticed the utterance itllve been going tove been being tested
every day for the past fortnight soon, where the verbal group will have been
325
going to have been being tested makes ve serial tense choices, present in
past in future in past in future, and is also passive. This passed quite
unnoticed by both the speaker and the person it was addressed to; yet
at the time it was being seriously questioned whether a simple verb
form like has been being tested, which one can hear about once a week,
could ever occur in English. Five-term tense forms are, predictably,
very rare one can in fact make a reasonable guess as to how rare, on
the basis of observed frequencies of two- and three-term tense forms
together with the constraints of the tense system; but they are provided
for within the resources of the spoken language. Another instance I
observed was they said theyd been going tove been paying me all this time,
only the funds just kept on not coming through.
Other things I noted regularly included present in present participial
non-nites like being cooking in I never heard you come in it must have
been with being cooking; marked thematic elements with reprise pronoun,
as in that poor child I couldnt get him out of my mind; and relatives
reaching into dependent clauses, such as thats the noise which when you
say it to a horse the horse goes faster. These are all systematic features that
people are unaware that they incorporate in their speech, and often
deny having said even when they are pointed out; or at least reject as
unsystematic after I didnt say it, the next line of defence is well
it was a mistake. But of course it was not a mistake; it was a regular
product of the system of spoken English.
But perhaps the most unexpected feature of those early observations
was the complexity of some of the sentence structures. Here are two
examples from recordings made at the time:
(i) Its very interesting, because it fairly soon is established when youre
meeting with somebody what kind of conversation youre having: for
example, you may know and tune in pretty quickly to the fact that
youre there as the support, perhaps, in the listening capacity that
youre there, in fact, to help the other person sort their ideas; and
therefore your remarks, in that particular type of conversation, are aimed
at drawing out the other person, or in some way assisting them, by
reecting them, to draw their ideas out, and you may tune in to this, or
you may be given this role and refuse it, refuse to accept it, which may
again alter the nature of your conversation.
(ii) The other man who kicks is the full-back, who usually receives the ball
way behind the rest of his team, either near his line or when somebodys
done what the stand-off in the rst example was doing, kicked over the
defenders; the full-back should be able then to pick it up, and his job is
326
usually to kick for touch nearly always for touch because hes miles
behind the rest of his side, and before he can do anything else with the
ball hes got to run up into them, before he can pass it, because he cant
pass the ball forward, and if he kicks it forward to another of his side
the other mans automatically off-side.
And you get a penalty for that, do you, the other side?
Depending on whether its kicking or passing forward. Passing forward
no, its a scrum. If you kick it forward and somebody else picks it up
that will be a penalty.
And if not, if the other side picks
If the other side picks it up thats all right; but the trouble is this is in
fact tactics again, because you dont want to put the ball into the hands
of the other side if you can avoid it because its the side that has
possession, as in most games of course, is at an advantage.
Examples such as these were noteworthy in two respects. One was
that they embodied patterns of parataxis (combining with equal status)
and hypotaxis (combining with unequal status) between clauses which
could run to considerable length and depth. The other was that they
were remarkably well formed: although the speaker seemed to be
running through a maze, he did not get lost, but emerged at the end
with all brackets closed and all structural promises fullled. And this
drew attention to a third property which I found interesting: that while
the listeners had absorbed these passages quite unconsciously and
without effort, they were difcult to follow in writing.
Lexical density
These two examples have been around for a long time; so let me turn
to some recent specimens taken from recordings made by Guenter
Plum to whom I am indebted for drawing them to my attention. In
these spontaneous narratives Plum regularly nds sequences such as the
following:
1A I had to wait, I had to wait till it was born and till it got to about eight
or ten weeks of age, then I bought my rst dachshund, a black-and-tan
bitch puppy, as they told me I should have bought a bitch puppy to
start off with, because if she wasnt a hundred percent good I could
choose a top champion dog to mate her to, and then produce something
that was good, which would be in my own kennel prex.
327
This displays the same kind of mobility that the earlier observations
had suggested was typically associated with natural, unselfconscious
speech which is what it was. I asked myself how I would have
expressed this in writing, and came up with two rewordings; the rst
(1B) was fairly informal, as I might have told it in a letter to a friend:
1B I had to wait till it was born and had got to about eight or ten weeks of
age; that was when I bought my rst dachshund, a black-and-tan bitch
puppy. By all accounts I should have bought a bitch puppy at the start,
because if she wasnt a hundred percent good I could mate her with a
top champion dog and produce a good offspring which would carry my
own kennel prex.
My second rewording (1C) was a more formal written variant:
1C Some eight or ten weeks after the birth saw my rst acquisition of a
dachshund, a black-and-tan bitch puppy. It seems that a bitch puppy
would have been the appropriate initial purchase, because of the
possibility of mating an imperfect specimen with a top champion dog,
the improved offspring then carrying my own kennel prex.
The aim was to produce a set of related passages of text differing
along one dimension, which could be recognized as going from most
likely to be spoken to most likely to be written. How such variation
actually correlates with difference in the medium is of course problematic; the relationship is a complicated one, both because written /
spoken is not a simple dichotomy there are many mixed and
intermediate types and because the whole space taken up by such
variation is by now highly coded: in any given instance the wording
used is as much the product of stylistic conventions in the language as
of choices made by individual speakers and writers. Here I am simply
moving along a continuum which anyone familiar with English usage
can readily interpret in terms of spoken and written poles.
The kind of difference that we nd among these three variants is
one that is often referred to as a difference of texture, and this familiar
rhetorical metaphor is a very appropriate one: it is as if they were the
product of a different weave, with bres of a different yarn. But when
we look behind these traditional metaphors, at the forms of language
they are describing, we nd that much of the difference can be
accounted for as the effect of two related lexicosyntactic variables. The
written version has a much higher lexical density; at the same time, it
has a much simpler sentential structure. Let us examine these concepts
in turn.
328
1A
1B
1C
(1)
Lexical
items
(2)
Running
words
(1:2)
(3)
Clauses
(1:3)
23
26
25
83
68
55
1:3.6
1:2.6
1:2.2
13
8
4
1.8:1
3.3:1
6.3:1
grammatical words goes down; and the number of clauses goes down
even more.
Let us attempt a similar rewording the other way round, this time
beginning with a passage of formal written English taken from Scientic
American:
2A Private civil actions at law have a special signicance in that they provide
an outlet for efforts by independent citizens. Such actions offer a means
whereby the multiple initiatives of the private citizens, individually or
in groups, can be brought to bear on technology assessment, the
internalization of costs and environmental protection. They constitute a
channel through which the diverse interests, outlooks and moods of the
general public can be given expression.
The current popular concern over the environment has stimulated
private civil actions of two main types.
2B is my attempt at a somewhat less written version; while 2C is in
another step nearer to speech:
2B Private civil actions at law are especially signicant because they can be
brought by independent citizens, so enabling them to nd an outlet for
their efforts. By bringing these actions, either as individuals or in groups,
private citizens can regularly take the initiative in assessing technology,
internalizing costs and protecting the environment. Through the use of
these actions as a channel, the general public are able to express all their
various interests, their outlooks, and their moods.
Because people are currently concerned about the environment, they
have been bringing numerous private civil actions, which have been
mainly of two types.
2C One thing is especially signicant, and that is that people should be
able to bring private civil actions at law, because by doing this
independent citizens can become involved. By bringing these actions,
whether they are acting as individuals or in groups, private citizens can
keep on taking the initiative; they can help to assess technology, they
can help to internalize costs, and they can help to protect the environment. The general public, who want all kinds of different things, and
who think and feel in all kinds of different ways, can express all these
wants and thoughts and feelings by bringing civil actions at law.
At present, people are concerned about the environment; so they have
been bringing quite a few private civil actions, which have been mainly
of two kinds.
330
2A
2B
2C
(1)
Lexical
items
(2)
Running
words
(1:2)
(3)
Clauses
(1:3)
48
48
51
87
101
132
1:1.8
1:2.1
1:2.6
5
12
17
9.6:1
4.0:1
3.0:1
Grammatical intricacy
Figure 1
Figure 2
Let us return to Text 1, in its original spoken form (Text 1A). This
consisted of 13 clauses. However, these 13 clauses were not strung out
end to end; they were constructed into a small number of clause
complexes of mixed paratactic and hypotactic construction: arguably
just one clause complex throughout. Here is its interpretation as one
clause complex:
Figure 3
Other examples from the same source but from different speakers
show similar patterns; there are, obviously, individual differences
(including perhaps in the preference for one or other type of interdependency), but the same free-owing intricacy is noticeable all the
time, as in Texts 46:
4 Roy was always interested in dogs and unfortunately hed never had the
opportunity to have a dog of his own, just because of circumstances
where he lived and what not, and so I bought him a Shepherd pup, which
was supposedly, you know, pure-bred Shepherd, but unfortunately people
sold it because it didnt have papers with it, so it was a pup.
5 Now how I got a German Shepherd was that I worked with a veterinary
surgeon, as Ive told you before, and there used to be a lady that brought
her Shepherds along to the clinic and I used to admire them greatly, and
she said, Well, she said, if you get married Ill give you one as a
wedding present, so immediately I bustled around looking for someone to
marry so I could get a Shepherd given to me for a wedding present, you
see, so thats how that worked out well, not quite! However I got my
Shepherd and he was my rst dog, mainly because when I was a youngster
I always wanted a dog but I lived with grandparents who wouldnt have
dogs or cats and I was a very frustrated animal lover at that stage of the
game, so as soon as I got out on my own I sort of went completely berserk!
6 So we rang up the breeder, and she sort of tried to describe the dog to us,
which was very hard to do over the phone, so we went over to have a look
to see what they were like, and we bought Sheba, because at that stage
Bob was away a lot on semitrailers with the army and it used to get quite
334
bad with the exercises youd have prowlers and perverts through the
married quarters, so if we, you know, got a dog, which we could do
because it didnt matter what sort of dog anyone had, itd bark and they
wouldnt bother us.
Types of complexity
Two distinct points need to be made here, and both of them run
counter to received attitudes towards spoken language. One is that
speech is not, in any general sense, simpler than writing; if anything,
it is more complex. There are, of course, many different kinds of
complexity, and we have already noted one measure lexical density
whereby speech will appear as the simpler of the two. But the
patterns we have been illustrating, which are the patterns of the
organization of the clause complex, referred to above as grammatical
intricacy, would seem to be at least as central to any conception of
complexity; and in this respect, speech appears as the more complex.
The syntactic complexity expected in writing, with which Deborah
Tannen (1982) introduces her discussion of oral and literate strategies,
does not turn out to be a characteristic of written discourse.
Of course, there are many other variables. Some writers achieve
considerable intricacy in the structure of the clause complex; it can be
learnt and consciously developed as a style. Some forms of spoken
discourse, on the other hand, militate against it: rapid-re dialogue
presents no scope for lengthy interdependencies complex semantic
patterns can be construed between interactants, but usually without
being realized in syntactic terms. And the categories of written and
spoken are themselves highly indeterminate they may refer to the
medium in which a text was originally produced, or the medium for
which it was intended, or in which it is performed in a particular
instance; or not to the medium at all, but to other properties of a text
which are seen as characteristic of the medium. So it is important to
indicate specically which variable of discourse is being referred to,
when one variety is being said to display some distinctive characteristic.
My point here is to question the assumption that written language is
syntactically more complex than spoken, and to suggest that, as far as
one particular kind of syntactic complexity is concerned the intricacy
(I do not want to call it structure because that assumes a particular
interpretation) of the sentence or clause complex this is more a
characteristic of the most unconscious spontaneous uses of language.
The more natural, un-self-monitored the discourse, the more intricate
335
In other words: when you speak, you cannot destroy your earlier
drafts. If we were to represent written language in a way that is
comparable to such representations of spoken language, we should be
including in the text every preliminary scrap of manuscript or typescript, with all the crossings out, misspellings, redraftings and periods of
338
silent thought; this would then tell us what the writer actually wrote.
Figure 4 is a specimen.
for most of a century. But for many purposes the discarded rst
attempts are merely trivial; they clutter up the text, making it hard to
read, and impart to it a spurious air of quaintness. What is much more
serious, however, is that transcribing spoken discourse in this way gives
a false account of what it is really like. It may seem a harmless piece of
self-gratication for a few academics to present spoken language as a
pathological phenomenon; one might argue that they deceive nobody
but themselves. But unfortunately this is not the way. Just when we
are seeing real collaboration between linguists and educators, and the
conception of language in education is at last gaining ground as a
eld of training and research, it seems we are determined to put the
clock back to a time when spoken language was not to be taken
seriously and could have no place in the theory and practice of
education.
Let us recapitulate the argument. Speech and writing as forms of
discourse are typically associated with the two modal points on the
continuum from most spontaneous to most self-monitored language:
spontaneous discourse is usually spoken, self-monitored discourse is
usually written. We can therefore conveniently label these two modal
points spoken and written language. Spoken and written language
do not differ in their systematicity: each is equally highly organized,
regular, and productive of coherent discourse. (This is clearly implied
once we recognize them both as language.)
Discourse in either medium can be characterized by hesitation,
revision, change of direction, and other similar features; these tend to
arise when attention is being paid to the process of text production.
Since highly monitored discourse is typically written, these features are
actually more characteristic of writing than of speech; but because most
written text becomes public only in its nal, edited form, the hesitations
and discards are lost and the reader is shielded from seeing the process
at work. Where they are likely to remain in is precisely where they
occur least, in the more spontaneous kinds of writing such as personal
letters. (Not all discourse features that are regarded as pathological, or
assigned negative value, are of this self-monitoring kind. One form of
discourse that has received a lot of critical attention is casual conversation, where the well-recognized characteristics are those of turntaking, such as interruptions and overlaps. But the strictly linguistic
deviations of casual conversation are mainly systematic features that
would not seem deviant if we had a grammar that took into account
the specically spoken resources of the linguistic system.)
Spoken and written language do differ, however, in their preferred
340
|||
||
[[ ]]
||| Thus the sympathetic induction of people into a proper and deep
understanding of [[what Christianity is about ]] should not be bracketed
341
simply with the evangelizing aim [[to which I referred earlier]]. ||| It
is not absolutely incompatible with that aim, however, for the following
reason.||| [[What counts as indoctrination and the like]] depends upon
a number of criteria,|| to do with the degree [[to which a teacher fails to
mention alternative beliefs]] , the tone of voice [[used]], the lack of
sympathy for the criticism [[levelled at Christianity or Humanism]]
and so on. |||A dogmatic teacher or lecturer differs from an open
one.||| The non-dogmatic teacher may be tepid;|| the open one may
be fervent.||| Fervour and indifference are not functions of closedness
and openness.|||
To see how this lexical density is achieved, we can look at the rst
clause. After the cohesive thus, it begins with a nominal group the
sympathetic induction of people into a proper and deep understanding of what
Christianity is about. The Head is induction; the Postmodier consists of
a series of alternating embedded prepositional phrases and nominal
groups, mainly one inside the other, and ending with an embedded
clause:
| the sympathetic induction [ of [ people ] ]] [ into [ a proper and deep
understanding [ of [[ [ what ] < Christianity | is > about ]] ] ] ] |
group or phrase boundary |
embedded group or phrase [ ]
enclosed elements
<>
(the prepositional phrase what . . . about is discontinuous, the items
Christianity and is being enclosed within it)
This nominal group contains a large amount of lexical information;
and if we take this passage as a whole we nd that out of the 52 lexical
items the only ones that do not occur in nominal groups are bracketed,
simply, depends, do, and differ. It is a characteristic of written discourse
that most of the lexical information is encoded in nominal form: that
is, in nominal groups, with their structure potential of Head (typically
a noun or adjective), Premodier (typically adjectives and nouns), and
Postmodier (typically embedded phrases and clauses, which then have
further nominal groups inside them).
Not every instance of a nominal group has a complex structure, of
course; the remaining ones in this passage range from:
| the lack [ of [sympathy [ for [ the criticisms [[ levelled | at [ Christianity
or Humanism ] ]] ] ] ] ] |
342
344
Grammatical metaphor
processes, one mental and one material, have been dressed up as one
which is neither. This coding of a semantic relation between two
processes as if it was the single process is very common in writing; the
sentence immediately preceding Text 2A contained another example
of the same thing, here with the verb leads to:
A successful tort action leads to a judgment of damages or an injunction
against the defendant company.
But this is just one type of a more general phenomenon, something
that I call grammatical metaphor (Halliday 1985, Chapter 10). Written
language tends to display a high degree of grammatical metaphor, and
this is perhaps its single most distinctive characteristic.
Here are three further examples of grammatical metaphor taken
from various written sources, together with suggested rewordings
which are less metaphorical:
Issue of the specially-coded credit cards will be subject to normal credit
checking procedures.
Credit cards have been specially coded and will be issued only when
credit has been checked in the normal way.
Strong Christmas sales were vital to the health of the retail industry,
particularly in the present depressed climate.
Unless many goods were sold at Christmas the retail industry would
not be healthy, particularly when the economy is depressed as it is
now.
He also credits his former big size with much of his career success.
He also believes that he was successful in his career mainly because he
used to be big.
In all these examples nominalization plays a signicant part, as it
does in many types of grammatical metaphor; so it is perhaps worth
stressing that nominalization is well motivated in English. It is not
simply a ritual feature that has evolved to make written language more
ambiguous or obscure; like the passive, which is another feature whose
functions are widely misunderstood, nominalization is an important
resource for organizing information. Take the example youth protest
mounted, which is not a headline but a complete sentence from a feature
article. We might reword this as more and more young people protested, or
young people protested more and more; but the only way to get the
combination of youth and protest as the Theme of the clause is by means
347
351
Chapter Thirteen
I realize that the title might well prompt someone to ask, How do
you mean, How do you mean?? I could have written, How are
meanings made? although I prefer the more personalized version.
The question is meant theoretically; but, like so many theoretical
questions, it becomes relevant in practice the moment we want to
intervene in the processes we are trying to understand. And some
processes of meaning are involved in more or less everything we do.
I shall need to talk about two fundamental relationships, those of
realization and instantiation; so let me begin by distinguishing these
two. Instantiation I take to be the move between the system and the
instance; it is an intrastratal relationship that is, it does not involve a
move between strata. The wording ne words butter no parsnips is an
instance, or an instantiation, of a clause. Realization, on the other hand,
is prototypically an interstratal relationship; meanings are realized as
wordings, wordings realized as sound (or soundings). We often use the
term to refer to any move which constitutes a link in the realizational
chain, even one that does not by itself cross a stratal boundary (for
example, features realized as structures); but the phenomenon of realization only exists as a property of a stratied system. To anticipate the
discussion a little, I shall assume that realization may be formalized as
metaredundancy, as this is dened by Jay Lemke (1985). Instantiation I
shall dene by making reference to the observer; it is variation in the
observers time depth. Firths concept of exponence is the product of
these two relations: his exponent is both instantiation and realization.1
First published in Advances in Systemic Linguistics: Recent Theory and Practice, 1992, edited by
Martin Davies and Louise Ravelli. London: Pinter, pp. 2035.
352
form of
consciousness
action
reection
1st/2nd person
regulatory
interactional
3rd person
instrumental
personal
domain
of experience
p, q, r (l, m, n a, b, c)
(p, q, r l, m, n) a, b, c
Figure 2 Metaredundancy
interpretation with a dynamic one. This leads us into the other critical
concept, that of instantiation.
Consider the notion of climate. A climate is a reasonably stable
system; there are kinds of climate, such as tropical and polar, and these
persist, and they differ in systematic ways. Yet we are all very concerned
about changes in the climate, and the consequences of global warming.
What does it mean to say the climate is changing? Climate is instantiated
in the form of weather: todays temperature, humidity, direction and
speed of wind, etc., in central Scotland are instances of climatic
phenomena. As such they may be more, or less, typical: todays
maximum is so many degrees higher, or lower, than average meaning
the average at this place, at this time of year and at this time of day.
The average is a statement of the probabilities: there is a 70 per cent
chance, let us say, that the temperature will fall within such a range.
The probability is a feature of the system (the climate); but it is no
more, and no less, than the pattern set up by the instances (the weather),
and each instance, no matter how minutely, perturbs these probabilities
and so changes the system (or else keeps it as it is, which is just the
limiting case of changing it).
The climate and the weather are not two different phenomena. They
are the same phenomenon seen by two different observers, standing at
different distances different time depths. To the climate observer, the
weather looks like random unpredictable ripples; to the weather
observer, the climate is a vague and unreal outline. So it is also with
language;8 language as system, and language as instance. They are not
two different phenomena; they are the same phenomenon as seen by
different observers. The system is the pattern formed by the instances;
and each instance represents an exchange with the environment an
incursion into the system in which every level of language is involved.
The system is permeable because each instance redounds with the
context of situation, and so perturbs the system in interaction with
the environment. Thus both realization and instantiation are involved
in the evolution of language as a dynamic open system.
Now the relation of system to instance is in fact a cline, a continuous
zoom; and wherever we focus the zoom we can take a look into history.
But to know what kind of history, we have to keep a record of which
end we started from. To the system observer, history takes the form of
evolution; the system changes by evolving, with selection (in the sense
of natural selection) by the material conditions of the environment.
This is seen most clearly, perhaps, in the evolution of particular sub359
Here for the rst time Nigel is selecting in two systems of meaning at
once, and by this token the initial move into grammar has been made.
Through the second year of life this new stratied system will gradually
replace the protolinguistic one, and all meanings (except for a few
protolanguage remnants that persist into adult life like hi! and ah! and
yum! and ouch!) will come to be stratally and metafunctionally complex.
So in Nigels rst exemplar, just cited, we have (i) proto-metafunctions
(proto-ideational different persons; proto-interpersonal seeking /
nding), and (ii) proto-strata, with the meaning rst construed as
wording (the ideational as contrasting names; the interpersonal as
contrasting mood) and then (re)construed as sounding (names as
articulation, mood as intonation). At this second interface the child can
now combine the segmental and the prosodic choices, in this way both
realizing and also iconically symbolizing the two different modes of
meaning that are combined at the rst interface. The resources for
making meaning are now in place.
It is probably not a coincidence that, as the ideational grammar
evolved, so in the system of transitivity the eld of processes was
construed into different process types along precisely the lines that (if
my understanding is right) went into the making of meaning in the rst
place. If meaning arises out of the impact of the conscious and the
material, as mutually contradictory forms of experience, then it is not
surprising that when experience is construed semantically, these two
types of process, the material and the conscious, should come to be
systematically distinguished. But there is a further twist. The semogenic
process, as we saw, involves setting up a relationship between systems
such that one is the realization of the other that is, they stand to each
other in a relation of Token and Value. This TokenValue relationship
is set up at both interfaces, and it is also what makes it possible to prise
the two apart and wedge in a grammar in between. Here then we nd
the third of the kinds of process construed by the grammar: the relational
process, based on identifying a Token with a Value. The grammar of
natural language, in its ideational metafunction, is a theory of human
experience; thus it may reasonably be expected to take as its point of
departure the very set of contrasts from which its own potential is
ultimately derived.
Let me return once again, nally, to the suggestion that meaning is a
mode of action engendered at the intersection of the material (or
phenomenal) and the conscious, as complementary modes of experience.
Now, the effect of this impact is to construe order. By the act of
meaning, consciousness imposes order on the phenomena of experience.
364
366
Notes
1. For Firths concept of exponence see especially his Synopsis of Linguistic
Theory in Firth (1957a).
2. I am speaking here of phylogenesis; but the process is recapitulated in the
growth of the individual, where it can be observed in the form of
behaviour. A child experiences certain phenomena as out there as lying
beyond the boundary between me and non-me: some perturbation seen
or heard, like a ock of birds taking off, or a bus going past, or a coloured
light ashing. At the same time, he also experiences a phenomenon that
is in here: his own consciousness of being curious, or pleased, or
frightened. At rst these two experiences remain detached; but then
(perhaps as a result of his success in grasping an object that is in his line of
sight in Trevarthens terms, when pre-reaching becomes reaching,
typically at about four months) a spark ies between them by which the
material is projected on to the conscious as Im curious about that, I like
that and so on. Now, more or less from birth the child has been able to
address others and to recognize that he is being addressed (Catherine
Batesons proto-conversation). The projection of the material on to the
conscious mode of experience maps readily on to this ability to address an
other; and the result is an act of meaning such as Nigels very highpitched squeak, which he rst produced at ve months, shortly after he
had learnt to reach and grasp.
3. Other microfunctions were added as the protolanguage evolved by degrees
into the mother tongue; but these were the original four. See Halliday
(1975, 1978).
4. At rst labelled, somewhat misleadingly, the level of context. See the
discussion of levels in Halliday, McIntosh and Strevens (1964). See also
Ellis (1966).
5. Based on giving and demanding that is, on exchange. Initially this meant
the exchange of goods-and-services; but eventually, by a remarkable
dialectic in which the medium of exchange became itself the commodity
exchanged, it extended to giving and demanding information. By this
step, meaning evolved from being an ancillary of other activities to being
a form of activity in its own right.
6. See the chapters entitled Towards a model of the instructional process,
The formal analysis of instruction and Action, context and meaning in
Lemke (1984).
7. It is impossible to have metaphor in a protolanguage at all, unless one
chooses to call metaphor (or perhaps proto-metaphor) what is taking
place when, for example, Nigel transfers a particular sign [gg gg gg]
from Im sleepy to lets pretend Im going to sleep. See Halliday (1975:
Chapter 2).
8. The analogy should not, of course, be pressed too far. Specically, while
367
9.
10.
11.
12.
13.
14.
368
Chapter Fourteen
Let me rst say what I mean by grammar in the title of the paper. I
mean the lexicogrammatical stratum of a natural language as traditionally understood, comprising its syntax and vocabulary, together with
any morphology the language may display: Lambs lexical system, in
his current (1992: Chapter 5) three-level architecture in commonsense terms, the resources of wording in which the meanings of a
language are construed. And here I have in mind particularly the
evolved, spontaneous grammar that construes the discourse of daily life.
This is not to exclude from the picture the elaborated grammars of
scientic and other metalanguages; but these can only be understood as
what they are: an outgrowth, supported by design, of the original
grammar that is learnt at mothers knee and on fathers shoulders.
Now English is not very efcient at creating technical nomenclature,
since it tends to confuse the study of a phenomenon with the
phenomenon itself. So while the term grammar is commonly used
in the way in which I have dened it, to mean the wording system,
the central processing unit of a natural language, it is also used
indiscriminately to mean the study of that system: grammar2 meaning
the study of grammar1. Since the study of language is called linguistics, I have been calling the study of grammar grammatics in order
to make the distinction clearer. A grammatics is thus a theory for
explaining grammar.
But is not a grammar itself also a theory? Clearly it is. A grammar is
First published in Functional Approaches to Language, Culture and Cognition, 2000, edited by
Teun A. van Dijk. Amsterdam: John Benjamins, pp. 22137.
369
in the way that it has been done countless other times by countless
other people, so it is reasonable to talk about the condition being
construed by the grammar.
In I have a headache the grammar construes a kind of thing, called an
ache; it then uses a part of the body to classify this thing, setting up a
taxonomy of aches including stomachache, backache and various others.
(Not all the parts of the body are allowed to ache, however; you cannot
have a footache or a thighache.) The grammar then sets up a conguration
of possession between the ache and some conscious being, in this case
the speaker I. The speaker becomes the owner of one specimen of that
complex class of things. It is not a prototypical form of possession; the
possessor does not want the thing possessed but cannot get rid of it
cannot give it away, or put it back where it came from. Why then
does the grammar not favour my head aches; or my heads aching? in
which the aching is a process, a state of being, rather than a thing, and
the entity involved in that state of being is my head rather than me.
The grammar has no trouble in constructing the clause my head aches;
yet it is not the most usual way in which the experience is worded.
Why is I have a headache preferred instead?
In English, as in many other languages (though not all), there is a
particular meaning associated with being the rst element in the clause.
What is put rst is being instated by the speaker as the theme of the
coming message; it is the setting for the information that follows (Fries
1995). This pattern of the clause, a structure of Theme + Rheme,
was apparently identied by the earliest rhetorical grammarians of
ancient Greece, the sophists, who seem to have recognized in the
thematic organization of the clause a potent resource for constructing
legal and political discourse. In modern times it was rst investigated in
detail by Mathesius, the founder of the Prague school; it is a particularly
prominent feature of English, appearing not only in the clause but also
as a fractal pattern in both smaller and larger structures inside word
groups, both nominal and verbal, on the one hand and extending over
a nexus of clauses on the other. The following example, taken from
natural conversation, shows thematic predication of a whole clause
complex (from Svartvik and Quirk 1980: 304):
. . . in my last year at college I said to myself: You want to do
applied chemistry, right? What industries are now just being born
which will blossom in the next quarter of a century, which is going
to be my working lifetime? And I said Plastics, sure as the nose
on your face. Im going to get into this. . . .
371
Im dazzled, you know . . . Its being able to see your working life
will span a period in which so-&-so is the topmost industry which I
nd so dazzling.
Now if I say my head aches, the rst element in that clause is my head:
I have constructed a message in which my head is enunciated as Theme.
My head is instated as what I want to elaborate on. But it isnt; Im the
one thats suffering, so the Theme of the clause should more appropriately be me. How does the grammar accommodate this alternative?
Most naturally, by making me the Subject, since there is a strong
association of these two functions in English. The ache becomes a
thing separated from myself, something that I possess, with my head
identied as its location: I have an ache in my head. Better still, if my
head is used as a classier, the ache and its location become a single
complex thing; and this now occupies the culminative position in the
clause: I have a headache. The ow of information here is very different
from that of my heads aching.
If this was just a feature of the grammar of localized aches and pains,
it might remain a curiosity, a special effect rather than a principle. But
this pattern has evolved in English as the prototypical form for
construing bodily qualities and states; rather than her hair is long, his
throat is sore, we tend to say she has long hair, he has a sore throat, putting
the person rather than the body part into the thematic role.1 And in
certain other languages where initial position is thematic we also
regularly nd the person, rather than the body part, lodged at the
beginning of the clause. The overall patterns are of course different: in
particular, there may be no strong bond between Theme and Subject,
and this makes it clear that the relevant function is that of Theme. We
can give examples from Chinese, Russian and French. In Chinese it is
possible to say wodi tou te`ng my head aches, where as in the English
wodi tou my head is a single element in the clause and so functions as
the Theme. The preferred form, however, is wo tou te`ng me the head
aches, where the head is detached from the personal pronoun; wo
me and tou head are now independent elements in the clause and
only the rst one, wo, is thematic. Again, this is the typical pattern for
all such expressions in Chinese: ta toufa chang her the hair (is) long, ta
houlong to`ng him the throat (is) sore and so on. In Russian, likewise,
one can say moja golova bolit my head aches; but this also is not the
preferred form. Russian however displays a different pattern: u menja
golova bolit at me the head aches, where again it is the me that has
thematic status. In French instead of ma tete me fait mal my head is
372
(1984b) Ways of saying, ways of meaning, where she shows how the
grammar of Urdu construes experience as collectively shared; and in
Martins (1988) account of grammatical conspiracies in Tagalog. If
we are comparing the different realities of one language with
another, it is the syndrome rather than the single feature that is likely
to be signicant.
Side by side with such frames of consistency, however, there are also
frames of inconsistency: regions where the grammar construes a pattern
out of tensions and contradictions where the different voices of
experience conict. To put this another way, the grammars theory of
experience embodies complementarity as well as concurrence. Metaphorically the grammar is representing the fact that human experience
is too complex, and has too many parameters, to be construed from
any one angle alone. It is the combination of these two perspectives
concurrence and complementarity that is the salient characteristic of
the grammar of daily life.
Let me rst try to illustrate the complementarity, and then use this
as a point of departure for exploring concurrence, looking at a more
general syndrome of features within which the earlier, more particular
example might be located. Many grammars (perhaps all) make a rather
clear distinction between the two fundamental modes of human
experience referred to above: between what we experience as taking
place in the world outside of ourselves and what we experience as
processes of our own consciousness seeing and hearing, liking,
disliking, fearing, hoping, thinking, knowing, understanding and the
like. In English, the conscious or mental processes differ from the
other, material kind in various respects: (1) they have a less exact
present time; (2) they presume a conscious being taking part; (3) they
do not fall within the scope of doing, and (4) they can project that
is, they can construe any meaning as taking place in someones
consciousness (as direct or indirect thought). In addition, these inner
processes display another feature not found with the grammar of
processes of the external, material type: they are bi-directional. Processes of consciousness can be construed with the conscious participant,
the Senser, either as object (active Complement), as in it frightens me,
or as active Subject, as in I fear it; likewise it pleases/convinces/strikes me,
I like/believe/notice it, and so on. These are two different and in fact
contradictory constructions of the same class of phenomena. Inner
experience is complex and difcult to interpret; the grammar offers
two complementary models, one with the Senser in the more active
role (by analogy with material processes), one with the Senser appearing
374
to be acted upon. Each of these brings out different agnate forms; the
grammar of daily life, in English, accommodates both.
In late Middle to early Modern English the common verbs of
consciousness such as like and think changed their allegiance from the
one pattern to the other: from it likes/thinks (to) me to I like it, I
think so. This happened at about the same time as the emergence of the
pattern discussed earlier: I have a headache, etc. For very general
processes of consciousness the grammar came to favour the type of
construction in which the Senser, the participant credited with consciousness, was the Theme. What was explained above as a preference
for a person rather than a part of the body as the starting point for
bodily states and conditions is part of a broader picture whereby the
grammar of all inner processes and physiological states tended to orient
the message towards the human, or humanlike, participant perhaps
with I, the individual self, as the prototypical member of this class.
This in turn leads us to another feature. At the same period of
history another shift took place affecting processes of the external kind,
those experienced as happening out there. In earlier English the
grammatical Subject in such processes had been overwhelmingly the
active participant, whether human or not (in fact the distinction
between human and non-human, or conscious and non-conscious,
plays no part in the construction of these processes of the external
world). Thus in an arrow pierced his eye the arrow was the natural Subject,
and remained Subject even if the narrative required the thing acted on
to function as Theme. To use a constructed example, the pattern was
that of:
The king fell to the ground; his eye an arrow had pierced.
with the
position
between
different
pattern:
The king fell to the ground; his eye had been pierced by an arrow.
This change led to an increase in the frequency of passive verbs, which
was followed by a change in the tense system as passive tenses caught
up with the active ones; and various other changes took place besides.
What this new alignment of grammatical forces amounted to was that
relatively less prominence was being given to the structure of the
experience which partner is the doer and which the done-to, so to
375
me, you say I gave a lot of money to my brother. The preposition to makes
explicit your brothers role as a participant in the process, and is added
just in those positions which are prominent in the information ow
(likewise if the brother appears as a marked Theme: to my brother I
gave a lot of money). It is precisely this same principle which adds by to
the Actor when the clause is passive: his eye had been pierced by an arrow/
by an arrow his eye had been pierced.
All the features I have sketched in here are features of the grammar
of daily life: some more global, some more local, but all of them
characteristic of unconscious, spontaneous, everyday linguistic encounters. These, and others that could be added, form a syndrome, a
concurrence of related developments, that has helped to shape the
meaning potential of Modern English, giving the language its characteristic avour that certain cut, in Sapirs terms, which makes each
language unique. What all these have in common is that they tend
towards giving greater prominence to the organization of discourse as
a ow of information, making more explicit how each element is to
be construed as part of a message. As a corollary to this, less prominence
is given to the experiential patterning, much of which is in fact left
implicit once the concern with the message begins to take over. Most
of these effects are fairly recent in history; they reect the changing
social conditions of the language over the past ve hundred years. Or
rather: they do not reect them they help to bring them about.
These features in the grammar construe the kind of discourse that can
be addressed to a stranger, who does not necessarily share the same
expectations and norms of interaction. They can be written down in a
book that is going to be printed in thousands of copies and read by
people who have never met the author and do not even know who he
is. In other words, they are features of a standard language: a form of
discourse in which the ow of information will typically be rendered
explicit rather than being taken for granted. (Interestingly, many of
these changes appear not to have taken place in the surviving British
rural dialects.)
Effects like these are not the result of sudden catastrophic changes.
They are trends and tendencies in a long process of evolution; and at
any given time they are quantitative changes in the relative frequency
with which this or that pattern is selected from within the system. The
grammatics is thus a theory of probabilities, in which possible/impossible is only a special case of more and less probable and a rather
uninteresting case, because meaning is a product of choice and when
something becomes impossible there is no more choice. So, for
377
Actor
Goal
Agent
Medium
-er (process) (type X)
-ee (process)
er (process) -ee
-er (process)
-ee
a
x
a
x
381
Notes
1. Notice on the other hand that in the interrogative this pressure is much
less strong: we readily say does your head ache? is your throat sore? as well as
have you got a headache/a sore throat? This is because in the interrogative the
grammar preempts the thematic slot to signal that the clause is, in fact, a
question, by putting at the beginning the part of the verb that selects for
yes or no, the Finite operator, does/is: does your head ache? signals my
message is concerned with whether it does or not. As a result there is
relatively little thematic weight left over; the difference in information
ow between is your throat sore? and have you got a sore throat? is very much
less noticeable than that between the agnate declarative pair my throats sore
and Ive got a sore throat, where the full thematic weight is felt on either my
throat or I.
2. Thus the grammar signals metaphorically that meaning is a social process.
We might put this together with the recent neurobiological nding by
Robin Dunbar (1992), that species living in large social groups have
proportionally larger cortices. Dunbars explanation is that large group
sizes demand greater social cohesion and hence more advanced skills for
communicating . . ..
383
Chapter Fifteen
The problem
Most of us are familiar with the feeling that there must be something
odd about linguistics. We recognize this as a problem in the interpersonal sphere because as linguists, probably more than other professionals, we are always being required to explain and justify our existence.
This suggests, however, that others see it as a problem in the ideational
sphere.
The problem seems to arise from something like the following. All
systematic knowledge takes the form of language about some
phenomenon; but whereas the natural sciences are language about
nature, and the social sciences are language about society, linguistics is
language about language language turned back on itself , in Firths
often quoted formulation. So, leaving aside the moral indignation some
people seem to feel, as if linguistics was a form of intellectual incest,
there is a real problem involved in drawing the boundary: where does
language end and linguistics begin? How does one keep apart the
object language from the metalanguage the phenomenon itself from
the theoretical study of that phenomenon?
The discursive evidence rather suggests that we dont, at least not
very consistently. For example, the adjective linguistic means both of
language, as in linguistic variation, and of linguistics as in linguistic
association (we never know, in fact, whether to call our professional
bodies linguistic associations or linguistics associations). But a situation
First published in Functional Descriptions: Theory in Practice, 1996, edited by Ruqaiya Hasan,
Carmel Cloran and David G. Butt. Amsterdam: John Benjamins, pp. 138.
384
In fact the ambiguity that I myself rst became aware of, as a teacher
of linguistics (and before that, as a teacher of languages), was that
embodied in the term grammar. Here the slippage is in the opposite
direction to that of psychology: grammar, the name of the phenomenon
(as in the grammar of English), slides over to become the name of the
study of the phenomenon (as in a grammar of English). This was already
confusion enough; it was made worse by the popular use of the term
to mean rules of linguistic etiquette (for example bad grammar). As a
way of getting round part of the problem I started using the term
385
Dening grammar
+ life
= biological
S
1
+ value
= social
+ meaning
= semiotics1
[primary]
S
2
S
3
+ grammar
= semiotic2
[higher order, i.e.
grammatico-semantic]
S
4.1
4.2
The grammar is thus the latest part of human language to have evolved;
and it is likewise the last part to develop in the growth of the individual
child. It emerges through deconstructing the original sign and reconstructing with the content plane split into two distinct strata, semantics
and lexicogrammar. Such a system (a higher-order semiotic organized
around a grammar) is therefore said to be stratied (Lamb 1964;
1992; Martin 1992; 1993).
A stratied semiotic has the unique property of being able to create
meaning. A primary semiotic, such as an infants protolanguage,
means by a process of reection: its meanings are given, like here I
am!, Im in pain, lets be together!, thats nice; and hence they
cannot modify each other or change in the course of unfolding. By
389
these two modes of meaning together into a single current, such that
everything we say (or write, or listen to, or read) means in both
these functions at once. Thus every instance of semiotic practice
every act of meaning involves both talking about the world and
acting on those who are in it. Either of these sets of phenomena may
of course be purely imaginary; that in itself is a demonstration of the
constitutive power of a grammar.
Grammar as theory
81 systems in it, Mann was probably not far wrong when he estimated
off the cuff that it would generate somewhere between 108 and 109
clause types.
Of course there are lots of mistakes in these complex networks, and
the only way to test them is by programming them and setting them
to generate at random. It is not difcult to generate the paradigm of
selection expressions from a reasonably small network (already in 1966
Henrici developed a program for this purpose; cf. Halliday and Martin
1981), where you can inspect the output and see where it has gone
wrong. But even if the program could list half a billion expressions it
would take a little while to check them over. As far as their overall
capacity is concerned, however, they are probably not orders of
magnitude out.
It has been objected that the human brain could not possibly process
a grammar that size, or run through all the alternative options whenever
its owner said or listened to a clause. I am not sure this is so impossible.
But in any case it is irrelevant. For one thing, this is a purely abstract
model; for another thing, the number of choice points encountered in
generating or parsing a clause is actually rather small in the network
of the verbal group it took only 28 systems to produce some 70,000
selection expressions, and in any one pass the maximum number of
systems encountered would be even less probably under half the
total, in a representative network. In other words, in selecting one out
of half a billion clause types the speaker/listener would be traversing at
the most about forty choice points. So although the system network is
not a model of neural processes, there is nothing impossible about a
grammar of this complexity that is, where the complexity is such that
it can be modelled in this way, as the product of the intersection of a
not very large number of choices each of which by itself is extremely
simple.
like those associated with the recent move into nanotechnology (engineering the very small); but the expansion may take place anywhere in
the lexicogrammar, as new wording, in any form. The grammar is not
simply tagging along behind; technological developments, like other
historical processes, are simultaneously both material and semiotic the
two modes are interdependent. Early on in his researches into science
and technology in China, Needham noted how in the medieval period,
when there was no adequate institutional mechanism for keeping new
meanings alive, the same material advances were sometimes made two
or three times over, without anyone realizing that the same technology
had been developed before (Needham 1958).
On the other hand, grammars grow by increasing the delicacy in
their construction of existing domains. (This has been referred to by
various metaphors: rening the grid or mesh, sharpening the focus,
increasing the granularity and so on. I shall retain the term delicacy,
rst suggested by Angus McIntosh in 1959.) This is a complex notion;
it is not equivalent to subcategorizing, which is simply the limiting case
although also the one that is likely to be the most easily recognized.
The grammar does construct strict taxonomies: fruit is a kind of food, a
berry is a kind of fruit, a raspberry is a kind of berry, a wild raspberry is a
kind of raspberry; these are typically hyponymic and can always be
extended further, with new words or new compositions of words in a
grammatical structure, like the nominal group in English and many
other languages. But greater delicacy is often achieved by intersecting
semantic features in new combinations; and this is less open to casual
inspection, except in isolated instances which happen to be in some
way striking (like certain politically correct expressions in presentday English). The massive semantic innovations brought about by
computing, word processing, networking, multimedia, the information
superhighway and the like, although in part construing these activities
as new technological domains, more typically constitute them as new
conjunctions of existing meanings, as a glance at any one of thousands
of current periodicals will reveal. On a somewhat less dramatic scale,
we are all aware of the much more elaborate variations in the discourse
of environmental pollution and destruction than were available a
generation ago. Even a seemingly transparent piece of wording such as
smoke-free construes a new conuence of meanings; indeed the whole
semogenic potential of -free as a derivational morpheme has recently
been transformed. (Similar expansions have happened with -wise and
-hood.)
There is a special case of this second heading perhaps even a third
396
categories must make sense as a whole. And this means that it needs to
be founded on compromise. The grammar of every natural language is
a massive exercise in compromise, accommodating multiple perspectives that are different and often contradictory.
Such compromise demands a considerable degree of indeterminacy
in the system.
10
Indeterminacy in grammar
should provide typical examples, for example the brake should be on,
meaning both ought to be and probably is. There is then the further
indeterminacy between an ambiguity and a blend, because a wording
which is clearly ambiguous in one context may be blended when it
occurs in another. A metaphor is the limiting case of a blend.
Complementarities are found in those regions of (typically experiential) semantic space where some domain of experience is construed
in two mutually contradictory ways. An obvious example in English is
in the grammar of mental processes, where there is a regular complementarity between the like type (I like it; cf. notice, enjoy, believe, fear,
admire, forget, resent . . . ) and the please type (it pleases me; cf. strike,
delight, convince, frighten, impress, escape, annoy . . .). The feature of
complementarities is that two conicting proportionalities are set up,
the implication being that this is a complex domain of experience
which can be construed in different ways: here, in a process of
consciousness the conscious being is on the one hand doing, with
some phenomenon dening the scope of the deed, and on the other
hand being done to with the phenomenon functioning as the doer.
All languages (presumably) embody complementarities; but not always
in the same regions of semantic space (note for example the striking
complementarity of tense and aspect in Russian). One favourite domain
is causation and agency, often manifested in the complementarity of
transitive and ergative construals.
Strictly speaking probability is not a fuzzy concept; but probability
in grammar adds indeterminacy to the denition of a category. Consider the network of the English verbal group in Figure 2 above. As an
exercise in grammatics this network is incomplete, in that there are
distinctions made by the grammar that the network fails to show: in
that sense, as already suggested, no network ever can be complete. But
it is incomplete also in another sense: it does not show probabilities. If
you are generating from that network, you are as likely to come up
with wont be taken as with took; whereas in real life positive is
signicantly more likely than negative, active than passive, and past
than future. Similarly a typical dictionary does not tell you that go is
more likely than walk and walk is more likely than stroll, though you
might guess it from the relative length of the entries. A grammar is an
inherently probabilistic system, in which an important part of the
meaning of any feature is its probability relative to other features with
which it is mutually dening. Furthermore the critical factor in register
variation is probabilistic: the extent to which local probabilities depart
from the global patterns of the language as a whole; for example a
400
11
In the last few sections I have picked out certain features of natural
language grammars which a theory of grammar a grammatics is
designed to account for. The purpose of doing this was to provide a
context for asking the questions: how does the grammatics face up to
this kind of requirement? Given that every theory is, in some sense, a
lexicogrammatical metaphor for what it is theorizing, is there anything
different about a theory where what it is theorizing is also a
lexicogrammar?
There is (as far as I can see) no way of formally testing a grammar in
its role as a theory of human experience: there are no extrinsic criteria
for measuring its excellence of t. We can of course seek to evaluate
the grammar by asking how well it works; and whatever language we
choose it clearly does grammars have made it possible for humanity
to survive and prosper. They have transmitted the wisdom of accumulated experience from one generation to the next, and enabled us to
interact in highly complex ways with our environment. (At the same
time, it seems to me, grammars can have quite pernicious side-effects,
now that we have suddenly crossed the barrier from being dominated
by that environment to being in control of it, and therefore also
responsible for it; cf. Halliday 1993). I suspect that the same holds true
for the grammatics as a theory of grammar: we can evaluate such a
theory, by seeing how far it helps in solving problems where language
is centrally involved (problems in education, in health, in information
management and so on); but we cannot test it for being right or wrong.
(This point was made by Hjelmslev many years ago, as the general
distinction between a theory and a hypothesis.) By the same token a
grammatics can also have its negative effects, if it becomes reductionist
or pathologically one-sided.
The special quality of a theory of grammar, I think, is the nature of
the metaphoric relationship that it sets up with its object of enquiry. If
we consider just those features of language brought into the discussion
401
above the size (and growth) of the grammar, its trinocular perspective,
and its fuzz how does the grammatics handle these various parameters? To put this in very general terms: how do we construe the
grammatics so as to be able to manage the complexity of language?
It seems to me that there are certain matching properties. The
grammatics copes with the immense size of the grammar, and its
propensity for growing bigger, by orienting itself along the paradigmatic axis, and by building into this orientation a variable delicacy; this
ensures that the grammar will be viewed comprehensively, and that
however closely we focus on any one typological or topological domain
this will always be contextualized in terms of the meaning potential of
the grammar as a whole. It copes with the trinocular vision of the
grammar by also adopting a trinocular perspective, based on the stratal
organization of the grammar itself. And it copes with the indeterminacy
of the grammar by also being indeterminate, so that the categories of
the theory of grammar are like the categories that the grammar itself
construes.
Theories in other elds, concerned with non-semiotic systems, begin
by generalizing and abstracting; but they then take off, as it were, to
become semiotic constructs in their own right, related only very
indirectly and obliquely to observations from experience. The prototype of such a theory is a mathematical model; and one can theorize
grammatics in this way, construing it as a formal system. But a
grammatics does not need to be self-contained in this same manner. It
is, as theory, a semiotic construct; but this does not create any
disjunction between it and what it is theorizing it remains permeable
at all points on its surface. The grammatics thus retains a mimetic
character: it explains the grammar by mimicking its crucial properties.
One could say that it is based on grammatical logic rather than on
mathematical logic. In some respects this will appear as a weakness: it
will lack the rigour of a mathematical theory. But in other respects it
can be a source of strength. It is likely to be more relevant to
understanding other semiotic systems: not only verbal art, but also
other, non-verbal art forms, as demonstrated by OTooles masterly
interpretation of painting, architecture and sculpture in terms of systemic grammatics, referred to already (OToole 1994). And the new
eld of intelligent computing, associated with the work of Sugeno,
and explicitly dened by him as computing with (natural) language,
requires a theory that celebrates indeterminacy (it is a development of
fuzzy computing) and that allows full play to the interface between
wording and meaning (see section 20 below).
402
In the next few sections I will make a few observations about these
matching properties of the grammatics, as they seem to me to emerge
in a systemic perspective.
12
When many years ago I rst tried to describe grammar privileging the
paradigmatic axis of representation (the system in Firths framework
of system and structure), the immediate reasons related to the theoretical and practical tasks that faced a grammatics at the time (the middle
1960s): computational (machine translation), educational (rst and
second language teaching; language across the curriculum); sociological
(language and cultural transmission, in Bernsteins theoretical framework, for example Bernstein (1971)); functional-variational (development of register theory) and textual (stylistics and analysis of spoken
discourse). All these tasks had in common a strong orientation towards
meaning, and demanded an approach which stretched the grammar in
the direction of semantics. There were perhaps ve main
considerations.
i: The paradigmatic representation frees the grammar from the
constraints of structure; structure, obviously, is still to be
accounted for (a point sometimes overlooked when people draw
networks, as Fawcett (1988) has thoughtfully pointed out), but
structural considerations no longer determine the construal of
the lexicogrammatical space. The place of any feature in the
grammar can be determined from the same level, as a function
of its relationship to other features: its line-up in a system, and
the interdependency between that system and others.
ii: Secondly, and by the same token, there is no distinction made,
in a paradigmatic representation, between describing some feature and relating it to other features: describing anything consists
precisely in relating it to everything else.
iii: Thirdly, the paradigmatic mode of description models language
as a resource, not as an inventory; it denes the notion of
meaning potential and provides an interpretation of the
system in the other, Saussurean sense but without setting up
a duality between a langue and a parole.
iv: Fourthly, it motivates and makes sense of the probabilistic
modelling of grammar. Probability can only be understood as
the relative probabilities of the terms in a (closed) system.
403
13
A note on delicacy
never was assumed, except perhaps among a very few linguists, that a
function word like of has only one location in the terrain described
by the grammatics. These exceptional cases challenge the implicit
generalization that the orthographic form always denes a type
within the wording.
A more explicit principle could be formulated: that, as far as the
grammatics is concerned, the endpoint in delicacy is dened by what
is systemic: the point where proportionalities no longer continue to
hold. As long as we can predict that a : a :: b : b :: . . . , we are still
dealing with types, construed as distinct categories for purposes of
grammatical description.
In practice, of course, we are nowhere near this endpoint in writing
our systemic grammars. (I nd it disturbing when the very sketchy
description of English grammar contained in Halliday (1994) is taken
as some kind of endpoint. Every paragraph in it needs to be expanded
into a book, or perhaps some more appropriate form of hypertext; then
we will be starting to see inside the grammar and be able to rewrite
the introductory sketch!) We are only now beginning to get access to
a reasonable quantity of data. This has been the major problem for
linguistics: probably no other dened sphere of intellectual activity has
ever been so top-heavy, so much theory built overhead with so little
data to support it. The trouble was that until there were rst of all tape
recorders and then computers, it was impossible to assemble the data a
grammarian needs. Since grammars are very big, and very complex, an
effective grammatics depends on having accessible a very large corpus
of diverse texts, with a solid foundation in spontaneous spoken
language; together with the sophisticated software that turns it into an
effective source of information.
14
15
Trinocular vision
16
Indeterminacy in grammatics
categories used for construing the grammar things like noun and
subject and aspect and hypotaxis and phrase are also like everyday
terms: they impose discontinuity. Either something is a noun or it is a
verb (or . . .); we cannot decide to construe it as a nerb. But, in turn,
each one of these itself denotes a fuzzy set. And, thirdly, the same
resources exist, if in a somewhat fancier form, for making the indeterminacy explicit: verbal noun, pseudo-passive, underlying subject, and so
on.
What then about the specic construction of indeterminacy in the
overall edice constructed by such categories? Here we see rather
clearly the grammatics as complexity management. On the one hand,
it has specic strategies for defuzzifying for imposing discontinuity
on the relations between one category and another; for example, for
digitalizing the grammars clines (to return to the example of force,
cited in section 10, it can establish criteria for recognizing a small,
discrete set of contrasting degrees of force). A system network is a case
in point: qualitative relationships both within and between systems may
be ironed out, so that (i) the system is construed simply as a or b
(or . . .), without probabilities, and (ii) one system is either dependent
on or independent of another, with no degrees of partial association.
But, at the same time, the grammatics exploits the various types of
indeterminacy as resources for managing the complexity. I have already
suggested that the concept of lexicogrammmar (itself a cline from
most grammatical to most lexical) embodies a complementarity in
which lexis and grammar compete as theoretical models of the whole.
There are many blends of different types of structure, for example the
English nominal group construed both as multivariate (congurational)
and as univariate (iterative) but without ambiguity between them. And
the two most fundamental relationships in the grammatics, realization
and instantiation, are both examples of indeterminacy.
I have said that a grammar is a theory of human experience. But that
does not mean, on the other hand, that it is not also part of that
experience; it is. We will not be surprised, therefore, if we nd that its
own complexity comes to be managed in ways that are analogous to
the ways in which it itself manages the complexity of the rest. In the
last resort, we are only seeing how the grammar construes itself.
17
the relationship between the strata; the verb realize faces upwards,
such that the lower stratum realizes the higher one. (Realization is
also extended to refer to the intrastratal relation between a systemic
feature and its structural (or other) manifestation.) Instantiation is the
relationship between the system and the instance; the instance is said to
instantiate the system.
It can be said that, in the elements of a primary semiotic (signs), the
signier realizes the signied; but this relationship is unproblematic:
although the sign may undergo complex transformations of one kind
or another, there is no intermediate structure between the two (no
distinct stratum of grammar). With a higher order semiotic, where a
grammar intervenes, this opens up the possibility of many different
types of realization. It is not necessary to spell these out here; they are
enumerated and discussed in many places (for example Berry 1977;
Fawcett 1980; Martin 1984; Hasan 1987; Matthiessen 1988; Eggins
1994).
But there is another opening-up effect which is relevant to the
present topic: this concerns the nature and location of the stratal
boundary between the grammar and the semantics. This is, of course,
a construct of the grammatics; many fundamental aspects of language
can be explained if one models them in stratal terms, such as metaphor
(and indeed rhetorical resources in general), the epigenetic nature of
childrens language development, and metafunctional unity and diversity, among others. But this does not force us to locate the boundary at
any particular place. One can, in fact, map it on to the boundary
between system and structure, as Fawcett does (system as semantics,
structure as lexicogrammar); whereas I have found it more valuable to
set up two distinct strata of paradigmatic (systemic) organization. But
the point is that the boundary is indeterminate it can be shifted; and
this indeterminacy enables us to extend the stratal model outside
language proper so as to model the relationship of a language to its
cultural and situational environments.
Instantiation is the relationship which denes what is usually thought
of as a fact in the sense of a physical fact, a social fact and so on.
Facts are not given; they are constructed by the theorist, out of the
dialectic between observation and theory. This has always been a
problem area for linguistics: whereas the concept of a physical principle
became clear once the experimental method had been established a
law of nature was a theoretical abstraction constructed mathematically by the experimenter the concept of a linguistic principle has
proved much more difcult to elucidate.
411
18
It is safe to say that neither of these concepts has yet been thoroughly
explored. Problems arise with instantiation, for example, in using the
corpus as data for describing a grammar (why a special category of
corpus grammar?); in relating features of discourse to systemic
patterns in grammar (why a separate discipline of pragmatics?); and
in construing intermediate categories (such as Bernsteins code,
which remains elusive (like global warming!) from whichever end it is
observed which is what makes it so powerful as an agency of cultural
reproduction). (See Francis 1993 for the concept of corpus grammar;
Martin 1992 for showing that there can be a system-based theory of
text; Bernstein 1990 for code; Hasan 1989; 1992b for interpretation of
coding orientation; and also Sadovnik 1995 for discussion of Bernsteins
ideas).
As far as realization is concerned, Lemke has theorized this power412
(In that respect the original term ascriptive, which I had used earlier
to name this type of process, might better have been retained, rather
than being replaced by attributive.) Here too, then, the grammar is
construing a signicant aspect of human experience the perception
of a phenomenon as an instance of a general class in terms of a
property of language itself, where each act of meaning is an instance of
the systemic meaning potential.
Of course, the boot is really on the other foot: the grammatics is
parasitic on the grammar, not the other way around. It is because of
the existence of clause types such as those exemplied above that we
are able to model the linguistic system in the way we do. The
grammatics evolves (or rather one should say the grammatics is
evolved, to suggest that it is a partially designed system) as a metaphoric transformation of the grammar itself. This is a further aspect of
the special character of grammatics: while all theories are made of
grammar (to the extent that they can be construed in natural language),
one which is a grammar about a grammar has the distinctive metaphoric
property of being a theory about itself.
19
Centricity
Since the grammatics is a theory about a logo system, it is logocentric, or rather perhaps semocentric: its task is to put semiotic
systems in the centre of attention. In the same way, biological sciences
are bio-centric: biased towards living things; and so on. I think it is
also a valid goal to explore the relevance of grammatics to semiotic
systems other than language, and even to systems of other types. The
grammatics is also totalizing, because that is the job of a theory. Of
course, it focuses on the micro as well as on the macro the semiotic
weather as well as the semiotic climate; but that again is a feature of
any theoretical activity.
It has always been a problem for linguists to discover what are the
properties of human language as such, and what are features specic to
a given language. The problem is compounded by the fact that there is
more than one way of incorporating the distinction (wherever it is
drawn) into ones descriptive practice. Firth articulated the difference
between two approaches: what is being sketched here is a general
linguistic theory applicable to particular linguistic descriptions, not a theory of
universals for general linguistic description (Firth 1957: 21; Firths emphasis). I have preferred to avoid talking about universals because it
seems to me that this term usually refers to descriptive categories being
414
treated as if they were theoretical ones. As I see it, the theory models
what is being treated as universal to human language; the description
models each language sui generis, because that is the way to avoid
misrepresenting it.
Thus while the theory as a whole is logocentric, the description of
each language is what we might call glottocentric: it privileges the
language concerned. The description of English is anglocentric, that
of Chinese sinocentric, that of French gallocentric and so on. (Note
that the theory is not anglocentric; the description of English is.) This
is not an easy aim to achieve, since it involves asking oneself the
question: how would I describe this language as if English (or other
languages that might get used as a descriptive model) did not exist?
But it is important if we are to avoid the anglocentric descriptions that
have dominated much of linguistics during the second half of the
century.
In practice, of course, English does exist, and it has been extensively
described; so inevitably people tend to think in terms of categories set
up for English or for other relatively well-described languages. I have
suggested elsewhere some considerations which seem to me relevant to
descriptive practice (Halliday 1992). As far as my own personal history
is concerned, I worked rst of all for many years on the grammar of
Chinese; I mention this here because when I started working on
English people told me I was making English look like Chinese! (It
seems ironic that, now that systemic theory is being widely applied to
Chinese studies, the work of mine most often cited as point of reference
is the descriptive grammar of English.)
In my view an important corollary of the characterological approach
(that is, each language being described in its own terms) is that each
language is described in its own tongue. The protocol version of the
grammar of English is that written in English; the protocol version of
the grammar of Chinese is that written in Chinese; and so on. The
principle of each language its own metalanguage is important,
because all descriptive terminology carries with it a load of semantic
baggage from its use in the daily language, or in other technical and
scientic discourses; and this semantic baggage has some metalinguistic
value. This applies particularly, perhaps, to the use of theoretical terms
as metacategories in the description; words such as (the equivalents of)
option, selection, rank, delicacy are likely to have quite signicant (but
variable) loadings.
But the principle also helps to guard against transferring categories
inappropriately. Even if descriptive terms have been translated from
415
English (or Russian, or other source) in the rst place, once they are
translated they get relocated in the semantic terrain of the new
language, and it becomes easier to avoid carrying over the connotations
that went with the original. So if, say, the term subject or theme appears
in a description of Chinese written in English, its status is as a translation
equivalent of the denitive term in Chinese. Perhaps one should point
out, in this connection, that there can be no general answer to the
question how much alike two things have to be for them to be called
by the same name!
20
inferencing with fuzzy sets and fuzzy matching processes. But to use
natural language requires a grammatics: that is, a way of modelling
natural language that makes sense in this particular context. Systemic
theory has been used extensively in computational linguistics; and the
Penman nigel grammar, and Fawcetts communal grammar, are
among the most comprehensive grammars yet to appear in computational form (Matthiessen 1991a; Matthiessen and Bateman 1992; Fawcett and Tucker 1990; Fawcett, Tucker and Lin 1993). But, more
importantly perhaps, systemic grammatics is not uncomfortable with
fuzziness. That is, no doubt, one of the main criticisms that has been
made of it; but it is an essential property that a grammatics must have
if it is to have any value for intelligent computing. This is an exciting
new eld of application; if it prospers, then any grammarian privileged
to interact with Sugenos enterprise will learn a lot about human
language, as we always do from applications to real-life challenging
tasks.
Note
1. This is not to question the semiotic achievements of the bonobo chimpanzees (cf. Introduction, p. 3). The issue is whether their construal of human
language is an equivalent stratied system with a lexicogrammar at the
language is an equivalent stratied system with a lexicogrammar at the
core.
417
BIBLIOGRAPHY
419
bibliography
Montgomery (eds), Studies in Discourse Analysis. London: Routledge &
Kegan Paul.
Berry, M. (1981b) Towards layers of exchange structure for directive
exchanges, Network 2.
Bloomeld, L. (1933) Language. Chicago: University of Chicago Press.
Boas, F. (1911) Introduction to the Handbook of American Indian Languages.
Vol. 40.1, Bulletin of the Bureau of American Ethnology. Washington, D.C.:
Smithsonian Institution. Reprinted in C. I. J. M. Stuart, Report of the
Fifteenth Annual (First International) Round Table Meeting on Linguistics and
Language Study (Monograph Series in Languages and Linguistics 17).
Brndal, V. (1948) Les Parties du discours, trans. by Naert. Copenhagen:
Munksgaard.
Buhler, K. (1934) Sprachtheorie: Die Darstellungfunktion der Sprache. Jena: G.
Fischer.
Bull. W. E. (1960) Time, Tense and the Verb. Berkeley: University of California
Press (University of California Publications in Linguistics 19).
Carnochan, J. (1952) Glottalization in Hausa, Transactions of the Philological
Society.
Carroll, J. B. (ed.) (1956) Language, Thought and Reality: Selected Writings of
Benjamin Lee Whorf. Cambridge, MA.: M.I.T. Press.
Carter, M. G. (ed.) (1981) Arab Linguistics: An Introductory Classical Text, with
Translation and Notes. Amsterdam: J. Benjamins (Studies in the History of
Linguistics 24).
Chafe, W. L. (1982) Integration and involvement in speaking, writing
and oral literature, in D. Tannen (ed.), Spoken and Written Language:
Exploring Orality and Literacy. Norwood, N.J.: Ablex (Advances in Discourse Processes 9).
Chandola, A. (1970) Some systems of musical scales and linguistic principles,
Semiotica 2.
Chao, Y. R. (1968) A Grammar of Spoken Chinese. Berkeley and Los Angeles:
University of California Press.
Chavarria-Aguilar, O. L. (1954) Transfer grammar, in Lectures in Linguistics.
Poona: Deccan College Post-graduate and Research Institute (Deccan
College Handbook Series 5).
Chomsky, N. (1957) Syntactic Structures. The Hague: Mouton ( Janua Linguarum 4).
Chomsky, N. (1961) Generative grammar, Word 17.
Chomsky, N. (1964) Topics in the theory of generative grammar, in T. A.
Sebeok (ed.), Theoretical Foundations. Vol. 3, Current Trends in Linguistics.
The Hague: Mouton.
Colby, B. N. (1973) A partial grammar of Eskimo folktales, American
Anthropologist 75.
420
bibliography
Colby, B. N. and Colby, L. M. (1980) The Daykeeper: The Life and Discourse
of an Ixil Diviner. Cambridge, MA.: Harvard University Press.
Danes, F. (1966) A three-level approach to syntax, Travaux linguistiques de
Prague 1.
Davidse, K. (1992a) A semiotic approach to relational clauses, Occasional
Papers in Systemic Linguistics 6.
Davidse, K. (1992b) Transitivity/ergativity: the Janus-headed grammar of
actions and events, in M. Davies and L. Ravelli (eds), Advances in Systemic
Linguistics: Recent Theory and Practice. London: Pinter.
Derewianka, B. (1995) Language development in the transition from childhood to
adolescence: the role of grammatical metaphor. Unpublished PhD dissertation,
Macquarie University.
Doughty, P. S., Pearce, J. J. and Thornton, G. M. (1971) Language in Use.
London: Edward Arnold (for the Schools Council).
Dunbar, R. (1992) Summary in Secret life of the brain, New Scientist 1850
(Supplement 4).
Ebeling, C. L. (1960) Linguistic Units. The Hague: Mouton ( Janua Linguarum
12).
Eco, U. (1983) The Name of the Rose, trans. by W. Weaver. London: Secker
& Warburg. (Italian original Il nome della rosa. Milano: Bompiani.)
Edelman, G. M. (1992) Bright Air, Brilliant Fire: On the Matter of the Mind.
New York: Basic Books.
Eggins, S. (1994) An Introduction to Systemic Functional Linguistics. London:
Pinter.
Eggins, S. and Martin, J. R. (1997) Genres and registers of discourse, in T.
A. van Dijk (ed.), Discourse: A Multidisciplinary Introduction. London: Sage.
Eiler, M. A. (1979) Meaning and choice in writing about literature: a study of
cohesion in the expository texts of ninth graders. Unpublished PhD dissertation,
Illinois Institute of Technology.
Ellis, J. M. (1993) Language, Thought and Logic. Evanston Il.: Northwestern
University Press.
Ellis, J. O. (1966a) Towards a General Comparative Linguistics. The Hague:
Mouton (Janua Linguarum).
Ellis, J. O. (1966b) On contextual meaning, in C. E. Bazell et al. (eds) In
Memory of J. R. Firth. London: Longman.
Fawcett, R. P. (1980) Cognitive Linguistics and Social Interaction. Heidelberg and
University of Exeter: Julius Groos (Exeter Linguistic Studies 3).
Fawcett, R. P. (1988) What makes a good system network good?, in J. D.
Benson and W. S. Greaves (eds), Systemic Functional Approaches to Discourse.
Norwood, N.J.: Ablex.
Fawcett, R. P. and Tucker, G. H. (1990) Demonstration of GENESYS: a
very large semantically based systemic functional grammar, Proceedings of
421
bibliography
the Thirteenth International Conference on Computational Linguistics, Helsinki,
Vol. 1.
Fawcett, R. P., Tucker, G. H. and Lin, Y. Q. (1993) How a systemic
functional grammar works: the role of realisation, in H. Horacek and M.
Zock (eds), New Concepts in Natural Language Generation. London: Pinter.
Fillmore, C. J. (1968) The case for case, in E. Bach and R. T. Harms (eds),
Universals in Language. New York: Holt, Rinehart & Winston.
Firbas, J. (1959) Thoughts on the communicative function of the verb in
English, German and Czech, Brno Studies in English 1. 3963.
Firbas, J. (1964) On dening the theme in Functional Sentence Analysis,
Travaux linguistique de Prague 1. 26780.
Firth, J. R. (1951) General linguistics and descriptive grammar, Transactions
of the Philological Society. Reprinted in J. R. Firth 1957b.
Firth, J. R. (1948) Sounds and prosidies, in Transactions of the Philological
Society.
Firth, J. R. (1951) Modes of meaning, in J. R. Firth 1957b.
Firth, J. R. (1952) Personality and language in society, Sociological Review
42. 2.
Firth, J. R. (1955) Structural linguistics, Transactions of the Philological Society.
Reprinted in F. R. Palmer (ed.), Selected Papers of J. R. Firth 19521959.
London: Longman (Longmans Linguistic Library).
Firth, J. R. (1957a) A synopsis of linguistic theory, in J. R. Firth et al.
(eds), Studies in Linguistic Analysis. Oxford: Blackwell (Special Volume of
the Philological Society). Reprinted in F. R. Palmer (ed.), Selected Papers
of J. R. Firth 19521959. London: Longman (Longmans Linguistic
Library).
Firth, J. R. (1957b) Papers in Linguistics 19341951. London: Oxford University Press.
Firth, J. R. (1957c) Applications of general linguistics, Transactions of the
Philological Society. Reprinted in F. R. Palmer (ed.), Selected Papers of J. R.
Firth 19521959. London: Longman (Longmans Linguistic Library).
Firth, J. R. (1957d) Ethnographic analysis and language with reference to
Malinowsis views, in R. W. Firth (ed.), Man and Culture: An Evaluation of
the Work of Bronislaw Malinowski. London: Routledge & Kegan Paul.
Reprinted in F. R. Palmer (ed.), Selected Papers of J. R. Firth 19521959.
London: Longmans (Longmans Linguistic Library).
Firth, J. R. (1968) Selected Papers of J. R. Firth 19521959 (ed. F. R. Palmer).
London: Longman (Longmans Linguistic Library).
Francis, G. (1993) A corpus-driven approach to grammar: principles,
methods and examples, in M. Baker, G. Francis and E. Tognini-Bonelli
(eds), Text and Technology: In Honour of John Sinclair. Philadephia: J.
Benjamins.
422
bibliography
Fries, P. H. (1981) On the status of theme in English: arguments from
discourse, Forum Linguisticum 6.1. Reprinted in J. S. Peto and E. Soozer
(eds), Micro and Macro Connexity of Texts. Hamburg: Buske.
Fries, P. H. (1995) Patterns of information in initial position in English, in
P. H. Fries and M. Gregory (eds), Discourse in Society: Systemic Functional
Perspective. Norwood, N.J.: Ablex.
Garvin, P. (1957) Report of the Seventh Annual Round Table Meeting on Linguistics
and Language Study. Washington, D.C.: Georgetown University Press
(Languages and Linguistics 9).
Gougenheim, R. et al. (1956) Lelaboration du Francais elementaire. Paris: Didier.
Gregory, M. J. (1967) Aspects of varieties differentiation, Journal of Linguistics
3. 2.
Haas, W. (1954) On dening linguistic units, Transactions of the Philological
Society.
Haas, W. (1957) Zero in linguistic description, in J. R. Firth et al. (eds),
Studies in Linguistic Analysis. Oxford: Blackwell (Special Volume of the
Philological Society).
Haas, W. (1960) Linguistic structures, Word 16.2.
Haas, W. (1966) Linguistic relevance, in C. E. Bazell et al. (eds), In Memory
of J. R. Firth. London: Longman (Longmans Linguistic Library).
Halliday, M. A. K. (1956) Grammatical categories in Modern Chinese,
Transactions of the Philological Society.
Halliday, M. A. K. (1958) Review of Watmough: Language: A Modern
Synthesis, Archivum Linguisticum 10. 2.
Halliday, M. A. K. (1959a). Review of Herdan: Language as Choice and Chance,
Archivum Linguisticum 11. 2.
Halliday, M. A. K. (1959b) The Language of the Chinese: Secret History of the
Mongols. Oxford: Blackwell (Philological Society Publications 17).
Halliday, M. A. K. (1961) Categories of the theory of grammar, Word 17.3:
24292.
Halliday, M. A. K. (1964) Syntax and the consumer, in C. I. J. M. Stuart
(ed.), Report of the Fifteenth Annual (First International) Round Table Meeting
on Linguistics and Language Studies. Washington, D.C.: Georgetown University Press (Monograph Series on Languages and Linguistics 15).
Halliday, M. A. K. (1966) Patterns in words, The Listener 75.1920.
Halliday, M. A. K. (196768) Notes on transitivity and theme in English
13, Journal of Linguistics 3 and 4.
Halliday, M. A. K. (1975a) Language as social semiotic: towards a general
sociolinguistic theory, in A. Makkai and V. B. Makkai (eds), The First
LACUS Forum. Columbia, S.C.: Hornbeam Press.
Halliday, M. A. K. (1975b) Learning How to Mean: Explorations in the
Development of Language. London: Edward Arnold.
423
bibliography
Halliday, M. A. K. (1977) Text as semantic choice in social contexts, in A.
van Dijk and J. Peto (eds), Grammars and Descriptions. Berlin: de Gruyter.
Halliday, M. A. K. (1978) Meaning and the construction of reality in early
childhood, in H. L. Pick and E. Saltzman (eds), Modes of Perceiving and
Processing of Information. Hillsdale, N.J.: Erlbaum.
Halliday, M. A. K. (1979) Modes of meaning and modes of expresson: types
of grammatical structure, and their determination by different semantic
functions, in D. J. Allerton, E. Carney and D. Holdcroft (eds), Function
and Context in Linguistic Analysis: Essays Offered to William Haas. Cambridge:
Cambridge University Press.
Halliday, M. A. K. (1982) The de-automatization of grammar: from Priestleys An Inspector Calls, in J. Anderson (ed.), Language Form and Linguistic
Variation: Papers Dedicated to Angus McIntosh. Amsterdam: J. Benjamins
(Current Issues in Linguistic Theory 15).
Halliday, M. A. K. (1984) Listening to Nigel. Sydney: Linguistics Department,
Sydney University (mimeo).
Halliday, M. A. K. (1985) An Introduction to Functional Grammar, 1st edn.
London: Edward Arnold (2nd edn 1994).
Halliday, M. A. K. (1990) New ways of meaning: a challenge to applied
linguistics, Journal of Applied Linguistics 6. Reprinted in M. Putz (ed.),
Thirty Years of Linguistic Evolution. Amsterdam: J. Benjamins; and in M. A.
K. Halliday, Language in a Changing World. Canberra: Applied Linguistics
Association of Australia.
Halliday, M. A. K. (1991a) Towards probabilistic interpretations, in E.
Ventola (ed.), Functional and Systemic Linguistics: Approaches and Uses. Berlin:
Mouton de Gruyter.
Halliday, M. A. K. (1991b) Corpus studies and probabilistic grammar, in K.
Aijmer and B. Altenberg (eds), English Corpus Linguistics: Studies in Honour
of Jan Svartvik. London: Longman.
Halliday, M. A. K. (1992) Systemic grammar and the concept of a science
of language , Waiguoyu ( Journal of Foreign Languages, Shanghai International Studies University) 2.78. Reprinted in Network 19.
Halliday, M. A. K. (1993). Reprint of M. A. K. Halliday 1990.
Halliday, M. A. K. (1994) Introduction to Functional Grammar, 2nd rev. edn.
London: Edward Arnold.
Halliday, M. A. K. (1995) Computing meanings: some reections on past
experience and present prospects. Paper presented to the second conference of the Pacic Association of Computational Linguistics (PACLING
II), University of Queensland, Brisbane.
Halliday, M. A. K. and Hasan, R. (1976) Cohesion in English. London:
Longman.
Halliday, M. A. K. and Hasan, R. (eds) (1980) Text and Context: Aspects of
424
bibliography
Language in a Social-Semiotic Perspective. Tokyo: Sophia University Graduate
School of Linguistics (Sophia Linguistica 6).
Halliday, M. A. K. and James, Z. L. (1993) A quantitative study of polarity
and primary tense in the English nite clause, in J. McH. Sinclair, M.
Hoey and G. Fox (eds), Techniques of Description: Spoken and Written
Discourse. A Festschrift for Malcolm Coulthard. London and New York:
Routledge.
Halliday, M. A. K. and Martin, J. R. (eds) (1981) Readings in Systemic
Linguistics. London: Batsford Academic.
Halliday, M. A. K. and Martin J. R. (1993) Writing Science: Literary and
Discursive Power. London: Falmer Press.
Halliday, M. A. K. and Matthiessen, C. (1999) Construing Experience through
Meaning: A Language-based Approach to Cognition. London: Continuum.
Halliday, M. A. K., McIntosh, A. and Strevens, P. (1964) The Linguistic
Sciences and Language Teaching London: Longman (Longmans Linguistics
Library).
Halliday, M. A. K. and Poole M. (1978) Notes on Talking Shop: Demands on
Language for Use with the Film Australia Production. Lindeld: Australian Film
Commission.
Harris, Z. S. (1951) Methods in Structural Linguistics. Chicago: University of
Chicago Press.
Harris, Z. S. (1952) Discourse analysis, Language 28.1,4.
Harris, Z. S. (1954) Transfer grammar, International Journal of Linguistics 20.
Hasan, R. (1965) The Language of Nine-year-old Children. Leeds: Nufeld
Foreign Languages Teaching Materials Project, Child Language Survey
(Transcript No. 2b).
Hasan, R. (1968) Grammatical Cohesion in Spoken and Written English Vol. 1.
London: Longman.
Hasan, R. (1979) On the notion of text, in J. Peto (ed.), Text Versus
Sentence: Basic Questions of Text Linguistics. Hamburg: Helmut Buske Verlag.
Hasan, R. (1980) The texture of a text, in M. A. K. Halliday and R. Hasan
(eds), 1980.
Hasan, R. (1983) Coherence and cohesive harmony, in J. Flood (ed.),
Understanding Reading Comprehension. Newark, DE.: International Reading
Association.
Hasan, R. (1984a) The structure of the nursery tale: an essay in text typology,
in L. Coveri (ed.), Linguistica Testuale: Proceedings of the 15th Internatonal
Congress of the Italian Linguistics Society. Rom: Bulzoni.
Hasan, R. (1984b) Ways of saying: ways of meaning, in R. P. Fawcett, M.
A. K. Halliday, S. M. Lamb and A. Makkai (eds), Language as Social
Semiotic. Vol. 1, The Semiotics of Culture and Language. London and Dover:
Pinter (Open Linguistics Series).
425
bibliography
Hasan, R. (1984c) Coherence and cohesive harmony, in J. Flood (ed.),
Understanding Reading Comprehension. Newark, DE.: International Reading
Association.
Hasan, R. (1985) Lending and borrowing: from grammar to lexis, in J. E.
Clark (ed.), The Cultivated Australian: Festschrift for Arthur Delbridge. Hamburg: Helmut (Buske Beitrage zur Phonetik und Linguistik 48).
Hasan, R. (1986) The ontogenesis of ideology: an interpretation of mother
child talk, in T. Threadgold et al. (eds), Semiotics, Ideology, Language.
Sydney: Sydney Association for Studies in Society and Culture (Sydney
Studies in Society and Culture 3).
Hasan, R. (1987) The grammarians dream: lexis as most delicate grammar,
in M. A. K. Halliday and R. P. Fawcett (eds), Theory and Description.
Vol. 1, New Developments in Systemic Linguistics. London: Pinter.
Hasan, R. (1989) Semantic variation and sociolinguistics, Australian Journal of
Linguistics 9.2.
Hasan, R. (1992a) Rationality in everyday talk: from process to system, in J.
Svartvik (ed.), Directions in Corpus Linguistics: Proceedings of Nobel Symposium 82,
Stockholm, 48 August 1991. Berlin: de Gruyter (Trends in Linguistics 65).
Hasan, R. (1992b) Meaning in sociolinguistic theory, in K. Bolton and H.
Kwok (eds), Sociolinguistics Today: International Perspectives. London:
Routledge.
Hasan, R. (1995) The conception of context in text, in P. H. Fries and M.
J. Gregory (eds), Meaning and Choice in Language. Vol. 2, Discourse in Society:
Systemic Functional Perspectives. Norwood: Ablex (Advances in Discourse
Processes 50).
Henderson, E. (1951) The phonology of loanwords in some southeast Asian
languages, Transactions of the Philological Society.
Hildyard, A. and Olson, D. R. (1982) On the comprehension and memory
of oral vs. written discourse, in D. Tannen (ed.), Spoken and Written
Language: Exploring Orality and Literacy. Norwood, N.J.: Ablex (Advances
in Discourse Processes 9).
Hill, A. A. (1958) Introduction to Linguistic Structures. New York: Harcourt,
Brace & World.
Hjelmslev, L. (1953) Prolegomena to a Theory of Language, trans. by F. J.
Whiteld. Bloomington: Indiana University Press (Danish original Omkring
Sprogteoriens Grundlaeggelse. Copenhagen: Munksgaard).
Hjelmslev, L. (1961) Prolegomena to a Theory of Language, trans. by F. J.
Whiteld. Madison, Wl.: University of Wisconsin Press (Danish original
Omkring Sprogteoriens Grundlaeggelse. Copenhagen: Munksgaard).
Hockett, C. F. (1955) Manual of Phonology. Baltimore: Waverly Press (Memoirs of the International Journal of Anthropological Linguistics 2, Indiana
University Publications of Anthropology and Linguistics).
426
bibliography
Hockett, C. F. (1957) Two models of grammatical description, in M. Joos
(ed.), Readings in Linguistics. Washington: American Council of Learned
Societies.
Hockett, C. F. (1958) Course in Modern Linguistics. New York: Macmillan.
Hockett, C. F. (1961) Linguistic elements and their relations, Language 37.
Hockett, C. F. (1965) Sound change, Language 41.
Huddleston, R. D. (1965) Rank and depth, Language 41.
Huddleston, R. D., Hudson, R. A., Winter, E. O. and Henrici, A. (1968)
Sentence and Clause in Scientic English. London: University College London,
Communication Research Centre (O.S.T.I. Programme in the Linguistic
Properties of Scientic English, Final Report).
Hudson, R. A. (1967) Constituency in a systematic description of the English
clause, Lingua 18. 22550.
Hymes, D. H. (1971) Competence and performance in linguistic theory, in
R. Huxley and E. Ingram (eds), Language Acquisition: Models and Methods.
London and New York: Academic Press.
Ikegami, Y. (1980) On some categories for describing the semolexemic
structure, in J. E. Copeland and P. W. Davis (eds), Papers in Cognitivestraticational Linguistics. Houston: Rice University (Rice University Studies,
662).
Jakobson, R. (1963) Implications of language universals for linguistics, in J.
Greenberg (ed.), Universals of Language. Cambridge, MA.: MIT Press.
Jespersen, O. (190943) A Modern English Grammar. London: Allen & Unwin.
Johnston, T. (1990) A study of the grammar of Auslan (Australian sign).
Unpublished PhD dissertation, University of Sydney.
Kluckhohn, C. and Leighton, D. (1962) The Navaho (The National History
Library). Garden City: Doubleday.
Kress, G. R. (1976) Halliday: System and Function in Language. Selected Papers.
London: Oxford University Press.
Lamb, S. M. (1964a) The sememic approach to structural semantics, American
Anthropologist 66.
Lamb, S. M. (1964b) On alternation, transformation, realization and stratication, in C. I. J. M. Stuart (ed.), Report of the Fifteenth Annual (First
International) Round Table Meeting on Linguistics and Language Studies. Washington, D.C.: Georgetown University Press (Monograph Series on
Languages and Linguistics 17).
Lamb, S. M. (1988) Discussion in M. A. K. Halliday, S. Lamb and J. Regan
(eds), In Retrospect: Using Language and Knowing How. Claremont, CA.:
Claremont Graduate School (Issues in Communication 12).
Lamb, S. M. (1992) Outlines of a Cognitive Theory of Language: A Work in
Progress (pre-publication draft).
Lemke, J. L. (1984) Semiotics and Education. Toronto: Victoria University
427
bibliography
(Toronto Semiotic Circle Monographs, Working Papers and Prepublications 2).
Lemke. J. L. (1985) Ideology, intertextuality and register, in J. D. Benson
and W. S. Greaves (eds), Systemic Perspectives on Discourse. Norwood, N.J.:
Ablex.
Lemke, J. L. (1990) Talking Science: Language, Learning and Values. Norwood,
N.J.: Ablex.
Lemke, J. L. (1993) Discourse, dynamics, and social change, Cultural Dynamics 6.12 (Language as Cultural Dynamic).
Linde, C. and Labov, W. (1975) Spatial networks as a site for the study of
language and thought, Language 51.
Longacre, R. E. (1964) Grammar Discovery Procedures: A Field Manual. The
Hague: Mouton (Janua Linguarum 33).
Lyons, J. (1963) Structural Semantics: An Analysis of Part of the Vocabulary of
Plato. Oxford: B. Blackwell (Philological Society Publications 20).
Lyons, J. (1968) Introduction to Theoretical Linguistics. London: Cambridge
University Press.
Malinowski, B. (1935) The Language of Magic and Garden. Vol. 2, Coral Gardens
and their Magic. New York: American Book Co. Reprinted as The Language
of Magic and Gardening. Bloomington: Indiana University Press (Indiana
University Studies in the History and Theory of Linguistics).
Mann, W. C. (1983) Inquiry semantics: a functional semantics of natural
language grammar, Proceedings of the First Annual Conference: Association for
Computational Linguistics. European Chapter. Marina del Rey: Information
Sciences Institute, University of Southern California (ISI/RR 83105).
Mann, W. C. and Matthiessen, C. M. I. M. (1983) Nigel: A Systemic Grammar
for Text Generation. Marina del Rey: Information Sciences Institute,
University of Southern California (ISI/RR 83105).
Martin, J. R. (1980) How many speech acts?, University of East Anglia Papers
in Linguistics 1415.
Martin, J. R. (1983a) Conjunction: the logic of English text, in J. S. Peto
and E. Sozer (eds), Micro and Macro Connexity of Texts. Hamburg: Buske.
Martin, J. R. (1983b) The development of register, in R. O. Freedle and
J. Fine (eds), Developmental Issues in Discourse. Norwood, N.J.: Ablex.
Martin, J. R. (1984a) Process and text: two aspects of human semiosis, in
J. D. Benson and W. S. Greaves (eds), Systemic Perspectives on Discourse:
Selected Theoretical Papers from the Ninth International Systemic Workshop.
Norwood, N.J.: Ablex.
Martin, J. R. (1984b) Functional components in a grammar: a review of
deployable recognition criteria, Nottingham Linguistic Circular 13.
Martin, J. R. (1988) Grammatical conspiracies in Tagalog: family, face and
fate with regard to Benjamin Lee Whorf , in J. D. Benson, M. J.
428
bibliography
Cummings and W. S. Greaves (eds), Linguistics in a Systemic Perspective.
Amsterdam: J. Benjamins (Current Issues in Linguistic Theory 39).
Martin, J. R. (1992) English Text: System and Structure. Amsterdam: J.
Benjamins.
Martin, J. R. (1993) The model, in M. A. K. Halliday and J. R. Martin
(eds), 1993.
Martin, J. R. (forthcoming), Lexical cohesion, eld and genre: parcelling
experience and discourse goals, in J. E. Copeland (ed.), Text Semantics and
Discourse Semantics: Proceedings of the Second Rice Symposium in Linguistics and
Semiotics. Houston: Rice University.
Martin, J. R. and Matthiessen, C. (1992) Systemic typology and topology,
in F. Christie (ed.), Literacy in Social Processes. Darwin: Northern Territory
University, Centre for Studies in Language Education.
Mathesius, V. (1928) On linguistic characteriology with illustrations from
Modern English, in Actes du premier congre`s international de linguistes a` la
Haye. Reprinted in J. Vachek (ed.), A Prague School Reader in Linguistics.
Bloomington: Indiana University Press.
Mathesius, V. (1964) On the potentiality of the phenomena of language,
trans. by J. Vachek, in J. Vachek (ed.), A Prague School Reader in Linguistics.
Bloomington: Indiana University Press. (Czech original O potencialnosti
jevu jazykovych, Proceedings of the Philosophy and History Section, Czech
Academy of Sciences.)
Matthews, P. H. (1961) Transformational grammer, Archivum Linguisticum
13.2.
Matthews, P. H. (1966) The concept of rank in neo-Firthian linguistics ,
Journal of Linguistics 2.1: 10118.
Matthiessen, C. (1988) Representational issues in systemic functional grammar, in J. D. Benson and W. S. Greaves (eds), Systemic Functional Approaches
to Discourse. Norwood, N.J.: Ablex.
Matthiessen, C. (1991a) Language on language: the grammar of semiosis,
Social Semiotics 1.2.
Matthiessen, C. (1991b) Lexico (grammatical) choice in text generation, in C.
L. Paris, W. R. Swartout and W. C. Mann (eds), Natural Language Generation
in Articial Intelligence and Computational Linguistics. Dordrecht: Kluwer.
Matthiessen, C. (1991c) Developments in systemic functional theory: a
discussion of the dimensions of change, Network 17.
Matthiessen, C. (1992) Interpreting the textual metafunction, in M. Davies
and L. Ravelli (eds), Advances in Systemic Linguistics: Recent Theory and
Practice. London: Pinter.
Matthiessen, C. (1993a) The object of study in cognitive science in relation
to its construal and enactment in language, in M. A. K. Halliday (ed.),
Cultural Dynamics 6.12 (Language as Cultural Dynamic).
429
bibliography
Matthiessen, C. (1993b) Register in the round: diversity in a unied theory
of register analysis, in M. Ghadessy (ed.), Register Analysis: Theory and
Practice. London: Pinter.
Matthiessen, C. (1995) Theme as an enabling resource in ideational knowledge construction, in M. Ghadessy (ed.), Thematic Development in English
Texts. London: Pinter.
Matthiessen, C. and Bateman, J. (1992) Systemic Linguistics and Text Generation:
Experiences from Japanese and English. London: Pinter.
Matthiessen, C. and Nesbitt, C. (1996) On the idea of theory-neutral
descriptions, in R. Hasan, C. Cloran and D. G. Butt (eds), Functional
Descriptions: Theory into Practice. Amsterdam: J. Benjamins.
McIntosh, A. (1956) The analysis of written Middle English, Transactions of
the Philological Society.
McIntosh, A. (1961) Patterns and ranges, Language 37.3.
Melrose, R. (1979) General word and particular text: a study of cohesion in academic
writing. Unpublished Phd dissertation, University of Sydney.
Mitchell, T. F. (1957) The language of buying and selling in Cyrenaica:
a situational statement, Hesperis 26. Reprinted in T. F. Mitchell, Principles of Firthian Linguistics. London: Longman (Longmans Linguistics
Library 19).
Mitchell, T. F. (1958) Syntagmatic relations in liguistic analysis, Transactions
of the Philological Society.
Mukarovsky, J. (1977) The Word and Verbal Art: Essays, trans. by J. Burbank
and P. Steiner. New Haven, CT.: Yale University Press.
Mullie, J. (1937) The Structural Principles of the Chinese Language, Vol. 2.
Beijing, Peitang Lazarist Press.
Needham, J. (1958) The translation of old Chinese scientic and technical
texts, in A. H. Smith (ed.), Aspects of Translation. London: Secker &
Warburg (Studies in Communication 2).
Nesbitt, C. and Plum, G. (1988) Probabilities in a systemic grammar: the
clause complex in English, in R. P. Fawcett and D. J. Young (eds).
Nida, E. (1960) A Synopsis of English Syntax. Norman: University of
Oklahoma, Summer Institute of Linguistics (Linguistic Series 4).
Nystrand, M. (1987) The Structure of Written Communication: Studies in Reciprocity between Writers and Readers. London: Academic Press.
Oldenburg, J. (1987) From child tongue to mother tongue: a case study of language
development in the rst two and a half years. Unpublished PhD dissertation,
University of Sydney.
OToole, M. (1994) The Language of Displayed Art. London: Leicester University Press.
Painter, C. (1984) Into the Mother Tongue: A Case Study in Early Language
Development. London: Pinter (Open Linguistics Series).
430
bibliography
Painter, C. (1996) The development of language as a resource for thinking: a
linguistic view of learning, in R. Hasan and G. Williams (eds), Literacy in
Society. London: Longman.
Palmer, F. R. (1958a) Linguistic hierarcy, Lingua 7.
Palmer, F. R. (1958b) Comparative statement and Ethiopian semitic,
Transactions of the Philological Society.
Palmer, F. R. (1964a) Sequence and order , in C. I. J. M. Stuart (ed.),
Report of the Fifteenth Annual (First International) Round Table Meeting on
Linguistics and Language Studies. Washington, D.C.: Georgetown University
Press (Monograph Series on Languages and Linguistics 17).
Palmer, F. R. (1964b) Grammatical categories and their phonetic exponents,
in H. G. Lunt (ed.), Proceedings of the Ninth International Congress of Linguists.
The Hague: Mouton.
Pike, K. L. (1959) Language as particle, wave and eld, Texas Quarterly 2.
Quirk, R. (1959) Review of W. Nelson Francis: The structure of American
English , Archivum Linguisticum 11.
Reddy, M. J. (1979) The conduit metaphor: a case of frame conict in our
language about language, in A. Ortony (ed.), Metaphor and Thought.
Cambridge: Cambridge University Press.
Robins, R. H. (1953) Formal divisions in Sundanese, Transactions of the
Philological Society.
Robins, R. H. (1957a) Vowel nasality in Sundanese, in J. R. Firth et al.
(eds), Studies in Linguistic Analysis. Oxford: Blackwell (Special Volume of
the Philological Society).
Robins, R. H. (1957b) Aspects of Prosodic Analysis. Vol. 1, Proceedings of the
University of Durham Philosophical Society 1. Durham: Durham University
Philosophical Society (series B).
Robins, R. H. (1959) Status of grammar, Lingua.
Robins, R. H. (1961) Syntactic analysis, Archivum Linguisticum 13.
Roget, P. M. (1960) Thesaurus of English Words and Phrases. London: Longman.
Russell, B. (1925) ABC of Relativity. London: Allen & Unwin (3rd revised
edn 1977).
Sadovnik, A. R. (ed.) (1995) Knowledge and Pedagogy: The Sociology of Basil
Bernstein. Norwood, N.J.: Ablex.
Schapiro, L. (1972) Totalitarianism. London: Pall Mall Press.
Sinclair, J. McH. (1966) Beginning the study of lexis, in C. E. Bazell et al. (eds),
In Memory of J. R. Firth. London: Longman (Longmans Linguistic Library).
Sinclair, J. McH. (1991) Corpus, Concordance, Collocation. Oxford: Oxford
University Press.
Sinclair, J. McH. (1992) Trust the text: the implications are daunting, in M.
Davies and L. Ravelli (eds), Advances in Systemic Linguistics: Recent Theory
and Practice. London: Pinter.
431
bibliography
Sinclair, J. McH. and Coulthard, R. M. (1975) Towards an Analysis of Discourse:
The English Used by Teachers and Pupils. London: Oxford University Press.
Sinclair, J. McH., Jones, S. and Daley, R. (1970) English Lexical Studies.
Birmingham: University of Birmingham, School of English (Report to the
ofce of Scientic and Technical Information).
Smart, N. (1968) Secular Education and the Logic of Religion. London: Faber &
Faber.
Sugeno, M. (1995) Intelligent fuzzy computing. Paper presented to the
second conference of the Pacic Association of Computational Linguistics
(PACLING II), University of Queensland, Brisbane.
Svartvik, J. and Quirk, R. (eds) (1980) A Corpus of English Conversation. Lund:
C. W. K. Gleerup (Lund Studies in English 56).
Svoboda, A. (1968) The hierarchy of communication units and elds as
illustrated by English attributive constructions, Brno Studies in English 7.
49101.
Sweet, H. (1891) A New English Grammar: Logical and Historical. Oxford:
Clarendon.
Tannen, D. (1982) Oral and literate strategies in spoken and written
narrative, Language 58.
Taylor, C. V. (1979) English for Migrant Education Curriculum. Canberra:
Education Development and Research Committee (Report 18).
Thibault, P. J. (1992) Grammar, ethics and understanding: functionalist reason
and clause as exchange, Social Semiotics 2.1.
Turner, G. J. (1973) Social class and childrens language of control at age ve
and age seven, in B. Bernstein (ed.).
Ure, J. (1971) Lexical density and register differentiation, in G. E. Perren
and J. L. M. Trim (eds), Applications of Linguistics: Selected Papers of the
Second International Congress of Applied Linguistics, Cambridge 1969. London:
Cambridge University Press.
Ure, J. N. and Ellis, J. O. (1979) Register in descriptive linguistics and
linguistic sociology, in O. U. Villegas (ed.), Issues in Sociolinguistics. The
Hague: Mouton.
Vachek, J. (1966) The Linguistic School of Prague. Bloomington: Indiana
University Press.
van Leeuwen, T. (1988) Music and ideology: notes towards a socio-semiotics
of mass media music, Sydney Association for Studies in Society and Culture
Working Papers 2:1/2.
Winter, E. (1977) A clauserelational approach to English texts: a study of
some predictive lexical items in written discourse, Instructional Science 6.
Yngve, V. H. (1960) A model and an hypothesis for language structure,
Proceedings of the American Philosophical Society 104.5.
432
INDEX
433
index
category (cont.)
335, 343, 351, 3546, 385, 397,
399400, 408, 410, 412
causal
consequential 225
external 233
relation 357
causality
vs. redundancy 357
causation 400
causative 186, 225
Cause 238, 305
chain 45, 67, 95, 97100, 208, 224,
2325, 2446, 292, 352, 357, 360
chain-exhausting 166
chain-exhaustive 120
chaos 353, 3656
Chinese 2533, 167, 202, 294301,
3723, 385, 41516
box 43
choice 33, 513, 959, 100, 11618,
163, 174, 182, 192, 198200, 228,
23940, 262, 268, 279, 283, 301,
3047, 310, 328, 348, 356, 364,
377, 380, 395, 404
chooser 301
points 395
Chomsky, N. 4, 6, 72, 106, 112, 304
class(es) 2434, 41, 45, 4968, 95104,
10613, 1235, 15968, 1845,
212, 222, 234, 294, 307, 371,
3745, 397, 41214
class-dening 101
class-sequence 109
class-structure 50
class-type 108
natural 391
clause 2632, 4558, 6870, 95104,
10915, 1214, 17594, 20517,
21922, 22847, 26283, 2978,
305, 3089, 32948, 352, 3717,
3945, 414
clause-classes 26
clause-nal 238, 252
clause-initial 238
clause-like entities 221
clause-to-text analogy 241
434
index
creolized 362
crescendo 228, 240, 271
criteria 2434, 40, 42, 45, 48, 556, 58,
61, 6772, 967, 107, 114, 119,
161, 1668, 170, 1789, 3412,
398401, 410
(situational-) contextual 323
cryptogrammatic 407
crypotype 302
crypotypic 360
crystalline 303, 336, 350
Davidse, K. 380, 413
declarative 109, 111, 189, 233, 268, 273,
305
deep
grammar 116
structure 106, 116
denite 98, 185, 299300, 307
deictic 55, 99101, 243
deicticity 272
postdeictic 100
predeictic 100
delicacy 401, 4870, 989, 11415,
1589, 165, 223, 285, 293, 396,
4026, 415
degree of 48, 54, 57, 69, 99, 159, 223,
404
most delicate grammar 49, 54, 59, 405
depth 48, 58, 1014, 107, 120, 285, 327,
333
depth-ordered 97
determiner 99, 104
diachronic 23, 324
synchronic-diachronic 22
dialect 7, 324, 202, 377
British rural dialects 377
Wu and Yueh dialect groups 34
dialogue 22533, 23940, 271, 283, 308,
325, 335, 341
dictionary 54, 158, 160, 165, 186,
3923, 400, 407
diglossia 296
dimension(s) 22, 423, 1001, 159, 161,
164, 188, 194, 222, 225, 232, 241,
328, 35562, 391
of abstraction 59, 66, 96
of choice 356
of classes 269
of realization 365
diminuendo 228, 240, 270
diminuendo-crescendo 233, 243
discourse 78, 1011, 175, 189, 1934,
199209, 2258, 239, 2423,
2457, 261, 268, 2703, 282, 285,
292, 294, 296, 3023, 311, 3245,
32931, 33550, 3656, 369, 371,
377, 3867, 391, 3967, 407, 412,
415
spoken 270, 324, 331, 3357, 340,
403
spontaneous 325, 33740
written 331, 3356, 3406, 34850
ecolinguistics 9
Eggins, S. 411, 413
eidological 232
eidon 231
eidos 231
ellipsis 181, 225, 232, 237, 2812
Ellis, J. 174, 227, 391
embedding 126, 3434
embedded 227, 270, 280, 329, 3412
empirical(ly) 119, 242
enact 3, 5, 356, 382, 390, 392, 407
encode 202
encoded 202, 204, 220, 2401, 307,
342, 346
encoding 202, 235, 2923
endocentric word groups 243
English 33, 448, 545, 58, 60, 70,
96104, 11315, 120, 160, 167,
17592, 202, 206, 20910, 21417,
22846, 266, 270, 276, 279, 282,
297312, 32330, 343, 34750,
362, 36982, 385, 392400, 4057,
410, 41316
modern 1878, 191, 261, 3757
spoken 70, 101, 326, 343, 350
written 101, 261, 330, 343, 350
equative 183, 185, 243, 300
ergative 289, 1868
and transitive 3801, 400
see also voice
435
index
ergativity 312
Eskimo folk tales 231
ethnographic 173, 226, 231, 236, 386
ethnographer 230
ethological 227, 232, 236
ethos 4, 231
evolution 65, 303, 349, 35562, 370,
3778, 390, 392, 413
evolutionary semogenesis 362
exemplicatory 45, 70, 72
existential 243, 275
exophoric 201
experiential 198217, 224, 276, 284,
377, 379, 390, 400, 405
exponence 41, 45, 5372, 97, 2212,
293
Firths concept of 352
scale 5761, 66
see also rank
exponent(s) 237, 4572, 97, 116, 221
expressive 227, 236
Extent 204, 212, 238
extralinguistic 31, 236, 295
extratextual 3940
linguistic 111
rhetorical 22634
speech 2339, 268, 273, 307
syntactic 107
textual 1756, 182, 1935, 237, 273
functional
categories 209, 300
component(s) 2001, 211, 215, 241
element(s) 175, 225, 242, 262
environment 110, 122
interpretation 200, 235
labels 107, 203
schemata 236
semantic(s) 209, 237, 311
tenor 227, 231
theories 226, 2356
variety 301, 307
variational 403
fuzzy 3, 210, 417
computing 402, 416
set(s) 40910, 417
Field
eld, tenor, mode 201, 217, 221,
22731, 243, 2834, 364
particle, eld and wave 20911, 241
Fillmore, C. 1789, 181, 187, 194
Firth, J. R. 7, 12, 21, 24, 25, 38, 67, 106,
109, 110, 158, 1701, 1745, 210,
219, 262, 296, 301, 307, 311, 384,
387, 398, 403, 414
foreground 235
foregrounded 234
foregrounding 230, 349, 379, 409
free/bound 278
French 372, 373, 381, 415
Fries, Peter 228, 233, 2456, 371
function 12, 10716, 122, 173, 2024,
21247, 262, 281, 2924, 299308,
32950, 37682, 389416
experiential 202, 224
ideational 17593
imaginative 227
interpersonal 272, 382
gender 302
generative 37
genetic 22, 324
genitive 293, 300, 336
genre(s) 209, 222, 230, 234, 242, 312,
357, 413
gesture 354, 356, 366
given
and new 29, 1901
Given 1924, 207, 209, 238, 2701, 376
Gleason, H. A. 222, 244
Goal 113, 176, 17883, 1868, 2034,
212, 300, 312, 376, 379, 381
Goal/Medium 305
goods and services 199, 273
grammar
descriptive 30, 163, 415
lexis 3767, 165, 379, 404, 410
phonology 5668, 220, 239
protocol version 415
rank 1217
semantics 220, 239, 306, 324
systemic(-functional) 2612, 332
theory 41, 44, 67, 370, 4012, 416
see also grammatics
436
index
grammatics 11, 3656, 36978, 384417
trinocular 4029
graphology 39
Greek 2946
ancient 292, 365
ancient Greece 371
alphabet 104
Hasan, R. 11, 175, 2225, 229, 231, 242,
244, 261, 285, 351, 404, 406,
41113
hearer 199, 205, 207, 301
hearer-oriented 199, 207, 240
hesitation 205, 3378, 340
hierarchy 256, 424, 56, 5960,
11011, 115, 11924, 166, 213,
228, 242
Hjemslev, L. 4, 5, 12, 106, 109, 110,
112, 219, 236, 262, 301, 312, 354,
401
Hockett, C. 106, 112, 219, 221
Huddleston, R. D. 120, 125, 215, 349
Hudson, R. 204, 349
hyponomy 12, 226
hyponym 226, 282
hyponymic 396
hypotaxis 107, 266, 302, 327, 3323,
333, 3434, 362, 410
hypotactic 21317, 242, 266, 266,
282, 333
see also parataxis
iconicity 312
ideational 21011, 21617, 22744, 268,
298, 308, 311, 3489, 356, 364,
382, 384, 392, 401
component 186, 198200, 208, 237,
242
features 243
function 175, 177, 189, 193
meaning 177, 188, 193, 199, 201, 217,
229, 295
semantics 13
structure 231, 2412, 244
voice 230
indenite 185, 299
437
index
language 113, 2130, 3772, 95104,
10617, 11925, 15871, 17395,
197217, 21920, 22646, 261,
2825, 291313, 32351, 35366,
36983, 384417
child language development 34, 310,
411
language-in-action 229, 329
teaching 7, 170, 403
natural 175, 198, 237, 296309, 364,
369, 389417
spoken 4, 7, 10, 261, 3026, 32348,
366, 406
written 43, 227, 32351
langue 44, 2367, 403, 412
Latin 25, 95
Lemke, J. L. 3067, 312, 352, 356, 358,
370, 382, 387, 391, 41213
lexical density 229, 32732, 335, 3412
lexicogrammar 3, 6, 812, 163, 1689,
185, 21721, 231, 233, 239, 241,
243, 246, 2612, 2815, 294, 298,
341, 345, 3578, 369, 3789, 387,
389, 391, 393, 3968, 401, 4034,
408, 41112
lexicoreferential 224, 232, 244
lexico semantic 246
lexico syntactic 328
lexicogrammatical entities 241
lexicogrammatical system 345
see also wording(s)
lexis 37, 3943, 49, 5461, 67, 98, 104,
158, 16270, 379, 404, 410
lexis-based 407
lexis-driven 407
linguistics 112, 2130, 3740, 42, 46,
51, 56, 60, 68, 72, 95, 98, 105,
11819, 123, 15860, 170, 1747,
219, 2367, 243, 292, 2948, 307,
31113, 3245, 351, 365, 3846,
406, 411, 41517
listener 175, 18990, 1923, 209, 227,
237, 240, 243, 2703, 338, 344,
350
listener-prominence 240
literate 323, 325, 405
literature 2, 5, 37, 230, 313, 323
Location 238
logic 3, 6, 8, 1989, 212, 399, 402, 416
logical-semantic 262, 264, 266, 2812
logo-centric 414
logo-genetic 360
Longacre, R. E. 106, 222, 242
Lyons, J. 114, 178, 194
McIntosh, A. 227, 396
Malinowski, B. (Kasper) 173, 2267,
2367, 262, 311
Mann, W. C. 301, 3945
Manner 204, 212, 238
Martin, J. 13, 223, 2257, 2313, 242,
285, 310, 312, 344, 357, 3734,
389, 392, 395, 397, 41213
Mathesius, V. 190, 202, 311, 3245, 371
Matthiessen, C. 3, 13, 301, 365, 370,
373, 390, 397, 407, 411, 413,
41617
meaning 3, 913, 23, 3941, 45, 61, 66,
71, 958, 121, 170, 17493,
196207, 2214, 227, 231, 23747,
262, 276, 2945, 297312, 323,
325, 346, 348, 35066, 36974,
38693, 397414
act of 201, 348, 354, 356, 391, 414
choices in 307, 310
contextual 40, 61
formal 40, 71
meaning-creating 8, 354
meaning-making 397
meaning-oriented 350
Medium 238, 274, 277, 3801
Melrose, R. 2278, 242
metafunction 1213, 268, 284, 31011,
3568, 3645, 382, 3902, 411,
413
metalanguage 30, 293300, 30913, 354,
366, 369, 378, 384, 415
metaphor
clause as 222, 234, 245
conduit 2934, 297
grammatical 12, 280, 282, 34551,
35860, 397, 4001
rhetorical 328
metaredundancy 352, 3568, 362
438
index
microclass 101, 159
Mitchell, J. F. 205, 223
modal 216, 242, 2456, 300, 340, 399
subject 1901, 194
modality 200, 205, 215, 230, 234, 2378,
242, 245, 268, 2714, 27882, 399
visual 312
mode 221, 227, 229, 283
active 199, 237
experiential 202
interpersonal 205
logical 21115
metaphoric 397
reective 199, 237
textual 206
model 15967
bricks-&-mortar 404
modier 47, 108
modier-head 1089
mood 18990, 194, 199200, 205, 215,
230, 233, 237, 242, 262, 268,
2713, 364
morpheme 45, 5369, 71, 956, 103,
1245, 164, 21920, 295, 298
morphemics 69
439
index
person 70, 230, 2367, 268, 337, 353,
378
phatic 226
phoneme 656, 68, 71, 96, 21920, 222
supra-segmental 68
phonetic 26, 71
phonetics 39, 324
phonetic/kinetic 354
phonic 25, 39, 67
phonology 3, 6, 33, 379, 56, 60, 6571,
104, 120, 206, 21920, 239, 262,
297, 324, 357, 387, 389, 408
Chinese 292
diachronic 324
phonological 323, 39, 55, 6671,
11415, 208, 213, 231, 239, 293,
324, 338, 345, 357
phonological-lexical 22
phonological-morphological 22
prosodic 67, 71
see also grammar
phrasal verb 101, 376
phrase 45, 68, 121, 1256, 180, 206,
263, 278, 298, 336, 342, 410
phylogenetic 355, 360
Pike, K. 106, 174, 211, 239, 242
Pinyin 33
plosive 34
Plum, G. 327, 362
plural 295, 307
polarity 12, 61, 230, 279, 363, 398
polarity carrying element 189, 266
polysemy 294, 297, 386, 405
possession 327, 371, 373
possessive 100, 275
postmodier 342
pragmatic 10, 11, 412
Prague School 174, 190, 262, 292, 299,
311, 371
predicate 102, 188, 194, 2989
predicator 47, 99, 1012, 113, 238
preposition 193, 3767
prepositional 178, 180, 193, 263, 298, 342
present 70, 102, 110, 2745, 326, 379
Priestley, J. B. 234, 245
process 1878, 2035, 212, 224, 238,
244, 274, 305, 346
440
index
rhema 292, 294
rheme 18891, 205, 226, 231, 236, 264,
266, 269, 300, 371, 376
rhetorical 191, 199, 221, 2246, 2313,
2404, 283, 328, 371, 411
rhythm 114, 216, 2623, 338, 348, 405
rhythmic 207, 269
Rumbaugh and Rumbaugh 3, 4
Russell, Bertrand 349
Russian 3723, 392, 400, 416
Sanskrit 296
Sapir, E. 219, 246, 262, 311, 377
Saussure, F. 4, 220, 262, 325, 403, 412
science 56, 9, 296, 396
language(s) of 12, 3656
linguistic 212, 65
technology and 291, 392, 395
semantic 1112, 32, 114, 159, 170, 175,
185, 196216, 22046, 2837,
294312, 32450, 354, 3738,
389416
component(s) 200, 23441, 389
system 197216, 237, 31011, 345
semiotic 112, 109, 196247, 296313,
325, 35366, 37082, 387414
higher-order semiotics 3, 389, 392
space 35365
system(s) 8, 12, 109, 1967, 298,
301, 311, 313, 325, 3567,
3703, 3878, 392, 397, 402, 406,
414
transformation 390, 408
semogenic 8, 227, 3036, 35464,
3927
semogenesis 35762
Sinclair, J. McH. 168, 170, 242, 407
Slav 2
sociolinguistics 10
Soviet 2
speaker 31, 1746, 18892, 199200,
2057, 2267, 229, 236, 240, 243,
2704, 30111, 3256, 3378, 344,
371, 376
speaker-now 273
speaker-oriented 207, 240
speaker-prominent 2403
441
index
tense 25, 112, 215, 2734, 27980, 300,
310, 326, 365, 375, 379, 3934,
400
text 712, 2230, 38, 45, 49, 109, 166,
169, 175, 184, 192, 199201,
20711, 21747, 26187, 2945,
30011, 324, 32835, 33847, 350,
3603, 3867, 394, 412
text-forming 2067
text-generation 284
text-like 234
text-linguistic 222
texture 2078, 211, 2245, 233, 271,
281
Theme 111, 113, 1904, 2069, 216,
228, 230, 2335, 23846, 2668,
2703, 297300, 3039, 347, 357,
3718, 416
Theme-Rheme 209, 228, 233, 262,
271
(un)predicated 111
Token 2923, 2968, 364, 366, 413
tone 678, 164, 192, 205, 270
concord 266
contour 269, 399
group(s) 55, 192, 2078, 266, 269
key 203
tonic 11314, 126, 192, 270
accent 206, 208, 270
prominence 2067, 270
segment 270
topic 266
and comment 299
sentence 211, 228, 234, 243
Topic 300
transcription 2623, 345
transitivity 17693, 200, 215, 224, 229,
2337, 243, 246, 263, 268, 274,
276, 302, 305, 312, 364, 382
typological 13, 120, 397, 402
typology 9, 334, 167, 388, 399
442