Number and Numbers

'Breathtaking.Badiou ...announcesanewepochinphilosophy.
'
5LAvCJ Z|ZEK
o
Alain Badiou Number and Numbers
Number and Numbers
A|ain badiou
Trans|ated Dy HoDin Mackay
polity
First published in French as Le Nombre cI les lIombres < Editions du Seuil,
1990.
This English edition ' Polity Press, 2008
Reprinted in 2008
Polity Press
65 Bridge Street
Cambridge CB2 1 UR, UK.
Polity Press
350 Main Street
Malden, MA 02148, USA
All rights reserved. Except for the quotation of short passages for the purpose
of criticism and review, no part of this publication may be reproduced, stored
in a retrieval system, or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without the prior permission
of the publisher.
ISBN-13: 978-07456-3878-2
ISBN-13: 978-07456-3879-9 (pb)
A catalogue rewrd for this book is available from the British Library.
Typeset in 10.5 on 12 pt Sabon
by SNP Best-set Typesetter Ltd., Hong Kong
Printed and bound in the United States by Odyssey Press Inc.,
Gonic, New Hampshire
For further information on Polity, visit our website: www.polity.co.uk
This book is supported by the French Ministry for Foreign Affairs, as part of
the Burgess programme headed for the French Embassy in London by the
Institut Fran . ais du Royaume-Uni.
www.frenchbooknews.com

ri 5u|dCdS

Contents
Translator's Preface vii
0 Number Must Be Thought
Genealogies: Frege, Dedekind, Peano,
Cantor 5
I Greek Number and Modern Number 7
2 Frege 16
3 Additional Note on a Contemporary
Usage of Frege 24
4 Dedekind 31
5 Peano 46
6 Cantor: 'Well-Ordered ness' and the
Ordinals 52
2 Concepts: Natural Multiplicities 59
7 Transitive Multiplicities 61
8 Von Neumann Ordinals 67
9 Succession and Limit. The Infnite 73
10 Recurrence, or Induction 83
II Natural Whole Numbers 93
vi CONTENTS
3 Ontology of Number: Defnition, Order,
Cuts, Types 99
12 The Concept of Number: An Evental
Nomination 101
ADDITIONAL NOTES ON SETS OF ORDINALS 112
13 Diference and Order of Numbers 115
14 The Concept of Sub-Number 131
15 Cuts: The Fundamental Theorem 139
16 The Numberless Enchantment of the
Place of Number 156
4 Operational Dimensions 181
17 Natural Interlude 183
18 Algebra of Numbers 198
Conclusion 209
19 In Conclusion: From Number to Trans-Being 211
Notes 215
Index 231
Translator's Preface
Al ai n Badi ou' s Number and Numbers, frst publ i shed two years after
his Being and Event, is far from being the speci al ist work its title
might suggest. I n fact, it recapi tul ates and deepens Being and Event's
explosion of the pretexts upon whi ch the ' phi l osophy of mathemati cs'
is reduced to a theoretical ghetto; and thei r ki nshi p to those reacti on
ary modes of thought that systematical l y obscure the most pressing
questi ons for contemporary phi losophy. Neither does Number and
Numbers balk at suggesti ng that even the greatest efforts on the part
of number-theorists themselves have fal len short of the properly
radical i mport of the question of number. Badi ou' s astoni shi ng anal y
ses in the historical section of the book uncover the i nextricabl e bond
between phi l osophi cal assumpti ons and mathematical approaches to
the problem in these supposedly ' merely techni cal ' works. The ai m
of Number and Numbers, then, i s certai nl y not to moul d the unwi l l
i ng reader i nto a cal culating machi ne, or a ' phi l osopher of mathemat
ics' : its exhortation is that we ( mathematici ans, phi l osophers, subj ects
under Capi tal ) systematical l y thi nk number out of the techni cal ,
procedural contai nment of which its quoti di an tyranny, and the
abysmal fear it stri kes into the heart of the non-mathematici an, are
but symptoms. Symptoms, needless to say, whose expressi on withi n
the situation of phi l osophy i s a pronounced di staste for number-as
phi l osopheme - whence its recogni sable absence i n much ' conti nental
phi l osophy' , except where it i s pi l loried as the very nemesi s of the
ontological vocati on. So i f the ' return of the numeri cal repressed'
proposed here wi l l , by defni ti on, excite a symptomatic resi stance, for
viii TRANSLATOR'S PREFACE
Badi ou it al one can cl ear the way for the proper task of phi losophy;
as a worki ng-through of the mathematical ontol ogy presented in
Being and Eve1Zt, Number and Numbers is a thorough conceptual
apprenticeshi p preparatory to the thi nki ng of the event.
For the great thi nkers of number-theory at the end of the ni ne
teenth century, the way to an ontological understanding of number
was obscured by cal cul atory and operational aspects. Today, accord
ing to Badi ou, the pol i ti cal domi nation of number under capitalism
demands that the proj ect be taken up anew: onl y i f contemporary
phi l osophy rigorousl y thi nks through number can it hope to cut
through the apparently dense and i mpenetrable capi tal i st fabric of
numeri cal rel ati ons, to thi nk the event that can ' subtract' the subj ect
from that ' onti c' skei n wi thout recourse to an anti-mathematical
romantici sm.
Whi l st thi s doubtless demands ' one more effort' on the part of the
non-mathematici an, it woul d be a peevi sh student of phi losophy who,
understanding the stakes and contemplating the conceptual vista
opened up, saw thi s as an unreasonable demand - especi al l y when
Badi ou offers to those l acki ng in mathematical knowl edge the rare
pri vi lege of taki ng a meticul ousl y navigated conceptual shortcut to
the heart of the matter.
Badi ou' s remarkabl e book compri ses a number of di fferent works
- a radi cal phi l osophi cal treatise, a contri buti on to number-theory, a
document in the hi story of mathematics, a congenial textbook and a
subtl e and subversi ve exercise in pol i ti cal theory - whose intricate
interdependencies defy any order of priority. The translator' s task i s
to reproduce, wi th a foreign tongue, that uni que voice that can
compel us to ' count as one' these di sparate fgures. In negotiating this
chal l enge, I have sought to pri ori ti se cl arity over adherence to any
rigid scheme of transl ati on, except where mathematical termi nol ogy
demands consi stent usage, or where an orthodoxy is clearly al ready
i n force wi thi n extant transl ati ons of Badi ou' s work. In the latter case,
my references have been Ol iver Feltham' s l andmark translation of
Being and Event,
I
wi th which I have sought to harmonise key terms,
Peter Hal l ward' s i nval uable A Subject to Truth/ and Ray Brassier
and Al berto Toscano' s col lecti on of Badi ou' s Theoretical Writings.3
Apart from these, i n transl ati ng chapters 2 and 3 I referred cl osely
to Sam Gi l lespi e and Justin Cl emens' translation i n UMBR( a) , Science
and Truth (2000). Fi nal l y, whi l st seeking al so to mai ntai n conti nuity
with l ong-standi ng Engl i sh transl ati ons of number-theoretical works,
some cl assi cs in thei r own right, occasi onal l y the rigour of Badi ou' s
thi nki ng has demanded a re-eval uation of thei r chosen transl ations
for key terms. 4 Translators also fnd themselves obliged to arbitrate
TRANSLATOR'S PREFACE ix
between a fdelity to Badiou' s in many ways admi rabl e i ndi fference
to the pedantic apparatus of schol arl y citati on, and the temptation
to pin down the al l usi ons and quotations distri buted throughout hi s
work. Badi ou' s selection of texts is so di scerning, however, that it i s
hardl y a chore to return to them. Havi ng thus had frequent recourse
to the texts touched on i n Number and Numbers ( particularl y i n the
frst, hi storical part) , I have seen no reason not to add citations where
appropri ate.
One presumes that those sel f-consci ous styles of phi l osophi cal
writing that necessitate l aboured ci rcuml ocuti ons or termi nol ogi cal
preci osi ty on the part of a transl ator woul d for Badi ou fal l under the
sign of ' modern sophistry' , taken to task herei n, as el sewhere i n his
work. Nevertheless, the aspi ration to universal conceptual transpar
ency does not precl ude consideration of Badi ou as styl i st: frstly, as
Ol iver Feltham has remarked, Badiou' s sentences utilise subj ect/verb
order i n a characteristic way, and I have retai ned hi s tensi l e syntax
whenever doing so does not j eopardi se comprehension i n transl ati on.
Perhaps j ust as i mportantl y, Badi ou does not achieve the deft and
good-humoured development of such extremel y rich and complex
conceptual structures as are found in Number and Numbers wi thout
a generous and searching l abour on behal f of the reader, not to
mention a talent for suspense. Al though the l ater secti ons of Number
and Numbers may seem daunting, I hope to have reproduced Badi ou' s
confdent, meticul ous, but never stuffy mode of exposi ti on so as to
ease the way as much as possi bl e. In fact, in contrast to hi s own
occasi onal l y chilly edicts, I woul d venture to suggest that here, ' i n his
element' , Badiou al l ows hi msel f a certai n enthusi asm. One certai nl y
does not accompany hi m on thi s odyssey wi thout al so devel opi ng a
taste for the ' bitter j oy' of Number.
This translation sl owl y came to frui ti on on the basi s of a somewhat
i mpul si ve deci si on; it may not have survi ved to compl eti on wi thout
the enthusi asm and ai d of an i nternati onal l y di spersed group of
friends and acquaintances, actual and vi rtual , wi th whom I shared
the work in progress. I would like to extend my thanks to those who
hel ped by poi nti ng out errors and offering advice on the evol vi ng
manuscript: Ani ndya Bhattacharyya, Ray Brassier, Michael Carr,
Howard Caygi l l , Thomas Duzer, Zachary L. Fraser, Peter Hal l ward,
Armelle Menard Seymour, Reza Negarestani , Robi n Newton, Ni na
Power, Manuel a Tecusan, Al berto Toscano, Kei th Ti l ford, Davi d
Sneek, and Dami an Veal . My thanks al so to Al ai n Badi ou for hi s
generous hel p and encouragement, and to the Institution and Staff of
the Bodlei an, Tayl or Institution, and Radcl i ffe Science Li braries i n
Oxford. Part of my work on the transl ati on was undertaken whi lst
X TRANSLATOR'S PREFACE
in recei pt of a studentshi p from the Centre for Research in Modern
European Phi l osophy at Mi ddlesex University, London.
My greatest debt of gratitude i s to Ruth, without whose l ove and
understandi ng my battles wi th i ncomprehensi on coul d not even be
staged; and to Donal d, a great i nspi rati on, for whom the i nfnite j oys
of number sti l l l i e ahead.
Robi n Mackay
o
Number Must Be Thought
0. 1 . A paradox: we l i ve in the era of number' s despoti sm; thought
yields to the l aw of denumerable multi pl ici ti es; and yet ( unl ess perhaps
thi s very default, thi s fai l ing, i s onl y the obscure obverse of a concept
less submi ssi on) we have at our di sposal no recent, active idea of
what number i s. An i mmense effort has been made on this point, but
its l abours were essenti al l y over by the begi nni ng of the twentieth
century: they are those of Dedeki nd, Frege, Cantor, and Peano. The
factual i mpact of number onl y escorts a si lence of the concept. How
can we grasp today the question posed by Dedeki nd i n hi s 1 888
treatise, Was sind und was sollen die Zahlen?l We know very wel l
what numbers are for: they serve, stri ctl y speaki ng, for everything,
they provide a norm for Al l . But we sti l l don' t know what they are,
or else we repeat what the great thi nkers of the late ni neteenth
century - anti ci pati ng, no doubt, the extent of thei r future j uri sdi cti on
- sai d they were.
0. 2. That number must rule, that the i mperative must be: ' count ! ' -
who doubts thi s today? And not in the sense of that maxi m whi ch,
as Dedeki nd knew, demands the use of the ori gi nal Greek when
retraced: &d 0 av8Quo; &QL81'LEL2 - because it prescri bes, for
thought, its si ngul ar condition in the matheme. For, under the current
empire of number, it is not a question of thought, but of real iti es.
0. 3. Firstly, number governs our conception of the pol i tical , wi th
the currency - consensual , though it enfeebles every pol i tics of the
2 NUMBER MUST BE THOUGHT
thi nkable - of suffrage, of opi nion pol l s, of the maj ority. Every ' pol iti
cal ' convocati on, whether general or l ocal , i n pol l i ng-booth or parl ia
ment, muni ci pal or internati onal , is settled with a count. And every
opi ni on is gauged by the incessant enumeration of the fai thful (even
if such an enumeration makes of every fdel ity an i nfdel i ty) . What
counts - in the sense of what is val ued - is that which is counted.
Conversel y, everythi ng that can be numbered must be val ued. 'Politi
cal Science' refnes numbers into sub-numbers, compares sequences
of numbers, its onl y obj ect being shifts in voting patterns - that is,
changes, usual l y mi nute, i n the tabul ati on of numbers. Political
' thought' i s numeri cal exegesi s.
0. 4. Number governs the quasi -total ity of the ' human sciences'
( although thi s euphemi sm can barel y di sgui se the fact that what is
cal led ' science' here is a techni cal apparatus whose pragmatic basis
i s governmental ) . Statistics i nvades the entire domai n of these disci
plines. The bureaucrati sati on of knowl edges is above all an infnite
excrescence of numberi ng.
At the begi nni ng of the twentieth century, soci ol ogy unvei led its
proper dignity - its audacity, even - in the wi l l to submi t the fgure
of communi tari an bonds to number. It sought to extend to the social
body and to representati on the Gal i l ean processes of l iteral i sation
and mathematisati on. But ulti mately it succumbed to an anarchic
devel opment of thi s enterpri se. It is now repl ete with pi ti ful enumera
ti ons that serve onl y to val i date the obvi ous or to establ i sh parlia
mentary opportuni ti es.
History has drawn massi vel y upon stati sti cal techni que and is -
even, in fact above al l , under the auspices of academic Marxism -
becoming a di achroni c soci ol ogy. It has lost that whi ch alone had
characterised it, si nce the Greek and Latin hi stori ans, as a discipline
of thought: its conscious subordi nati on to the real of pol itics. Even
when i t passes through the di fferent phases of reaction to number -
economi sm, soci ol ogi sm - it does so onl y to fal l into thei r si mple
i nverse: bi ography, hi stori ci si ng psychol ogi sm.
And medicine itself, apart from i ts pure and si mpl e reduction to its
scientifc Other ( mol ecul ar biology) , is a disorderly accumul ation of
empi ri cal facts, a huge web of bl i ndl y tested numerical correlations.
These are ' sciences' of men made into numbers, to the saturation
poi nt of al l possi ble correspondences between these numbers and
other numbers, whatever they might be.
0. 5. Number governs cultural representati ons. Of course, there is
televi si on, vi ewi ng fgures, adverti si ng. But that's not the most
NUMBER MUST BE THOUGHT 1
important thi ng. It is in its very essence that the cultural fabric is
woven by number alone. A ' cul tural fact' i s a numeri cal fact. And,
conversely, whatever produces number can be cul tural l y l ocated; that
which has no number wi l l have no name either. Art, whi ch deal s wi th
number onl y in so far as there i s a thinking of number, i s a cultural l y
unpronounceable word.
0. 6. Obvi ousl y, number governs the economy; and there, without a
doubt, we fnd what Loui s Al thusser woul d have cal l ed the 'determi
nati on in the l ast instance' of i ts supremacy. The ideology of modern
parl i amentary societies, if they have one, is not humani sm, law, or
the subj ect. It i s number, the countabl e, countabi l ity. Every citizen is
expected to be cogni sant of foreign trade fgures, of the fexi bi l ity of
the exchange rate, of fuctuations in stock pri ces. These fgures are
presented as the real to whi ch other fgures refer: governmental
fgures, votes and opi ni on pol l s. Our so-cal led ' si tuati on' i s the i nter
section of economi c numeri cal ity and the numeri cal ity of opi ni on.
France ( or any other nati on) can onl y be represented on the bal ance
sheet of an i mport-export business. The onl y image of a country i s
thi s i nextricabl e heap of numbers i n whi ch, we are tol d, its power i s
vested, and whi ch, we hope, i s deemed worthy by those who record
its mood.
0. 7. Number i nforms our souls. What i s it to exist, i f not to give a
favourable account of onesel f? In America, one starts by saying how
much one earns, an i dentifcation that i s at l east honest. Our ol d
country is more cunni ng. But sti l l , you don' t have to l ook far to di s
cover numerical topi cs that everyone can identify wi th. No one can
present themselves as an i ndi vi dual wi thout stating i n what way they
count, for whom or for what they are real l y counted. Our soul has
the col d transparency of the fgures i n whi ch it i s resol ved.
0. 8. Marx: 'the icy water of egoti sti cal cal cul ati on' .
3
And how! To
the point where the Ego of egoism is but a numeri cal web, so that
the 'egoti stical cal cul ati on' becomes the cipher of a ci pher.
0. 9. But we don' t know what a number i s, so we don' t know what
we are.
0. 10. Must we stop with Frege, Dedeki nd, Cantor or Peano? Hasn' t
anything happened in the thi nki ng of number? Is there onl y the
exorbitant extent of its soci al and subj ective reign? And what sort
of innocent culpability can be attri buted to these thi nkers ? To what
4 NUMBER MUST BE THOUGHT
extent does thei r i dea of number prefgure thi s anarchic reign? Did
they think number, or the future of general i sed numerical ity? Isn' t
another idea of number necessary, i n order for us to turn thought
back agai nst the despoti sm of number, in order that the Subj ect might
be subtracted from it ? And has mathematics si mpl y stood by si lently
during the comprehensi ve soci al i ntegration of number, over which
it formerly had monopol y? This is what I wish to exami ne.
I
Genealogies: Frege, Dedekind,
Peano, Cantor
I
Greek Number and
Modern Number
1 . 1 . The Greek thi nkers of number related it back to the One, whi ch,
as we can sti l l see in Eucl i d' s Elements, 1 was considered not to be a
number. From the supra-numeric being of the One, uni ty is derived.
And a number is a col lecti on of uni ts, an addi ti on. Underl yi ng thi s
conception i s a problematic that stretches from the Eleatics through
to the Neoplatonists: that of the processi on of the Mul ti pl e from the
One. Number i s the schema of thi s processi on.
1 . 2. The modern col l apse of the Greek thi nki ng of number proceeds
from three fundamental causes.
The frst is the i rrupti on of the probl em of the i nfni te - i nel uctable
from the moment when, with di fferenti al cal cul us, we deal with
the real ity of series of numbers which, although we may consider
their l imit, cannot be assigned any termi nus. How can the l i mit of
such a series be thought as number through the sol e concept of
a col lecti on of units ? A seri es tends towards a l i mit: it i s not the
addi ti on of its terms or its uni ts. It cannot be thought as a processi on
of the One.
The second cause is that, i f the enti re edi fce of number i s sup
ported by the being of the One, whi ch i s itsel f beyond being, it is
i mpossi bl e to introduce, wi thout some radi cal subversi on, that other
pri nci pl e
-
that ontological stoppi ng point of number - whi ch is zero,
or the void. It coul d be, certai nl y - and Neoplatoni st specul ati on
appeals to such a thesi s - that the i neffabl e and archi -transcendent
character of the One can be marked by zero. But then the probl em
8 GENEALOGIES: FREGE. DEDEKIND. PEANO. CANTOR
comes back to numeri cal one: how to number unity, if the One that
supports it is voi d? This probl em is so complex that, as we shal l see,
it remai ns today the key to a modern thi nki ng of number.
The thi rd reason, and the most contemporary one, i s the pure and
si mpl e di sl ocati on of the i dea of a being of the One. We fnd oursel ves
under the j uri sdiction of an epoch that obliges us to hol d that being
is essenti al l y multi pl e. Consequentl y, number cannot proceed from
the supposit ion of a transcendent being of the One.
1 . 3. The modern thi nking of number thus found itsel f compel led to
forge a mathematics subtracted from thi s supposi ti on. In so doing, it
took three di fferent paths:
Frege' s approach, and that of Russel l ( which we wi l l cal l , for
brevity, the l ogi ci st approach) , seeks to 'extract' number from a pure
consideration of the l aws of thought itsel f. Number, according to this
poi nt of view, is a uni versal trait
2
of the concept, deducible from
absol utel y origi nal pri nci pl es ( pri nci pl es without which thought in
general woul d be i mpossi ble) .
Peano' s and Hi l bert' s approach ( let' s cal l thi s the formal i st
approach) construes the numerical fel d as an operati onal feld, on
the basi s of certai n si ngul ar axi oms. Thi s time, number occupies no
parti cul ar posi ti on as regards the l aws of thought. It is a system of
rul e-governed operations, specifed in Peano' s axi oms by way of a
transl uci d notati onal practice, enti rel y transparent to the material
gaze. The space of numeri cal signs is si mpl y the most ' originary' of
mathemati cs proper ( preceded onl y by purel y l ogical cal culations) .
We mi ght say that here the concept of number is enti rely mathema
tised, i n the sense that it is conceived as existing only i n the course
of its usage: the essence of number is cal culation.
The approach of Dedeki nd and Cantor, and then of Zermelo, von
Neumann and Godel ( which we shal l cal l the set-theoretical or 'pla
tonising' approach) determi nes number as a particul ar case of the
hi erarchy of sets. The ful crum, absol utely antecedent to al l construc
ti on, is the empty set; and ' at the other end' , so to speak, nothing
prevents the exami nati on of i nfnite numbers. The concept of number
is thus referred back to an ontology of the pure mul ti pl e, whose great
Ideas are the cl assi cal axi oms of set theory. In thi s context, ' bei ng a
number' is a particular predicate, the deci si on to consider as such
certai n cl asses of sets ( the ordi nal s, or the cardi nal s, or the elements
of the conti nuum, etc. ) with certai n di sti ncti ve properties. The essence
of number i s that it i s a pure mul ti pl e endowed with certai n proper
ties relating to its i nternal order. Number is, before being made
avai l able for cal cul ati on ( operati ons wi l l be defned ' on' sets of
GREEK NUMBER AND MODERN NUMBER 9
pre-existing numbers ) . Here we are deal i ng with an ontologisation of
number.
1 . 4. My own approach will be as fol lows:
( a) The logicist perspective must be abandoned for reasons of i nter
nal consistency: it cannot sati sfy the requi rements of thought,
and especi al ly of phi losophi cal thought.
( b) The axiomatic, or operational , thesi s is the thesi s most ' prone'
to the i deol ogi cal soci al i sati on of number: it ci rcumscri bes the
question of a thi nking of number as such withi n an ultimately
technical proj ect.
(c) The set-theoretical thesis i s the strongest. Even so, we must draw
far more radi cal consequences than those that have prevai led up
to the present. Thi s book tri es to fol l ow the thread of these
conseq uences.
1. 5. Whence my pl an: To exami ne the theses of Frege, Dedeki nd
and Peano. To establ i sh mysel f wi thi n the set-theoretical concepti on.
To radical i se it. To demonstrate ( a most i mportant poi nt) that
i n the framework of thi s radical i sation we wi l l rediscover also
( but not onl y) ' our' fami l i ar numbers: whol e numbers, rational
numbers, real numbers, al l , fnal l y, thought outside of ordi nary
operati onal mani pul ati ons, as subspecies of a unique concept of
number, itsel f statutori l y inscri bed wi thi n the ontology of the pure
multi pl e.
1 . 6. Mathemati cs has al ready proposed thi s rei nterpretati on, as
mi ght be expected, but only in a recessive corner of itsel f, bl i nd to
the essence of its own thought: the theory of surreal numbers, i nvented
at the begi nni ng of the 1970s by J. H. Conway ( On Numbers and
Games, 1976),3 taken up frstly by D. E. Knuth ( Surreal Numbers,
1974),4 and then by Harry Gonshor i n hi s canoni cal book ( An Intro
duction to the Theory of Surreal Numbers, 1986).5 Any i nterest we
might have in the techni cal detai l s of thi s theory wi l l be here strictly
subordi nated to the matter i n hand: establ i shi ng a thi nking of number
that, by fxi ng the latter' s status as a form of the thi nki ng of Being,
can free us from i t suffciently for an event, al ways trans-numeric, to
summon us, whether thi s event be pol i tical , artistic, scientifc or
amorous. Li miting the gl ory of number to the i mportant, but not
excl usive, glory of Being, and thereby demonstrating that what pro
ceeds from an event i n terms of truth-fdelity can never be, has never
been, counted.
1 0 GENEALOGIES: FREGE. DEDEKIND. PEANO. CANTOR
1 . 7. None of the modern thi nkers of number ( I understand by thi s,
I repeat, those who, between Bol zano and Godel , tri ed to pi n
down the idea of number at the j uncture of phi l osophy and the
l ogico-mathematical ) have been abl e to offer a unifed concept
of number. Customari l y we speak of ' number' with respect to
natural whol e numbers,6 ' relative' ( posi ti ve and negative) whole
numbers, rational numbers ( the ' fracti ons' ) , real numbers ( those
that number the l i near continuum) and, fnal l y, compl ex numbers
and quaterni ons. We also speak of number i n a more di rectly set
theoretical sense when designating types of well -orderedness (the
ordi nal s) and pure quantiti es of any multi pl e whatsoever, incl uding
i nfnite quantities ( the cardi nals ) . We might expect that a concept
of number woul d subsume all of these cases, or at least the more
' cl assi cal ' among them, that is to say, the whol e natural numbers (the
most obvious schema of di screte ' stepwi se' enumerati on) and the real
numbers ( the schema of the conti nuum) . But thi s is not at al l
the case.
1 . 8. The Greeks cl earl y reserved the concept of number for whole
numbers, which was quite i n keepi ng with thei r concepti on of the
composi ti on of number on the basis of the One, since onl y natural
whole numbers can be represented as col lections of units. To treat of
the conti nuum, they used geometrical denominati ons, such as the
rel ati ons between si zes or measurements. So thei r powerful concep
ti on was marked through and through by that division of mathemati
cal di sci pl ines on the basis of whether they treat of one or the other
of what were hel d by the Greeks to be the two possible types of
obj ect: numbers ( from which arithmetic proceeds ) and fgures ( from
whi ch, geometry) . Thi s di vi si on refers, i t seems to me, to the two
orientations whose unity is di alectical l y effectuated by effecti ve,
or materi al i st, thought: the algebraic orientati on, whi ch works by
composi ng, connecting, combi ni ng elements; and the topological
orientation, which works by perceiving proximities, contours and
approxi mati ons, and whose poi nt of departure i s not elementary
belongi ngs but i ncl usi on, the part, the subset.7 This di vi si on is sti l l
wel l -founded. Wi thi n the di sci pl i ne of mathematics itsel f, the two
maj or di vi si ons of Bourbaki ' s great treati se, once the general onto
logical framework of set theory i s set out, deal with ' algebraic struc
tures' and 'topol ogi cal structures' . 8 And the val idity of this arrangement
subtends al l di alecti cal thought.
1 . 9. It i s neverthel ess cl ear that, ever si nce the seventeenth century,
it has no longer been possi ble to pl ace any suffciently sophi sticated
GREEK NUMBER AND MODERN NUMBER I I
mathematical concept excl usi vel y on one side of the opposition arith
metic/geometry. The triple chal lenge of the i nfni te, of zero and of
the termination of the idea of the One disperses the idea of number,
shreds it into a refned di al ectic of geometry and arithmetic, of the
topological and the algebraic. Cartesi an analytic geometry radical ly
subverts the di stinction from the very outset, and what we know
today as ' number-theory' had to cal l on the most compl ex resources
of 'geometry' , i n the extremely broad sense i n which the l atter has
been understood in recent decades. Moderns therefore can no l onger
accept the concept of number as the obj ect whose provenance is
foundational (the idea of the One) and whose domai n i s prescri bed
( arithmetic) . ' Number' i s said i n many senses. But which of these
senses constitutes a concept, al l owi ng somethi ng si ngul ar to be
proposed to thought under this name?
1 . 10. The response to thi s questi on, in the work of the thi nkers I
have mentioned, is altogether ambi guous and exhi bits no unani mity
whatsoever. Dedeki nd, for exampl e, can legitimately be named as the
frst one to have, with the notion of the cut, convi nci ngl y 'generated'
the real numbers from the rati onal s. 9 But when he poses the questi on:
'What are numbers ? ' he responds wi th a general theory of ordi nal s
whi ch certai nl y, as a particul ar case, mi ght found the status of whol e
numbers, but which cannot be appl ied di rectl y to real numbers. 1O I n
whi ch case, what gives us the right to say that real numbers are
' numbers ' ? Si mi l arl y, in The Foundations of Arithmetic" Frege offers
a penetrating critique of al l previous defni ti ons ( i ncl udi ng the Greek
defnition of number as a ' set of uni ts' ) '
z
and proposes a concept of
'cardi nal number' that i n effect subsumes - on the basis of certain
arguable premises, to whi ch I shall later return - cardinals i n the set
theoretical sense, of whi ch natural whol e numbers represent the fnite
case. But at the same ti me he excl udes ordinal s, to say nothi ng of
rational numbers, real numbers or compl ex numbers. To use one of
his favourite expressi ons, such numbers do not ' fal l under the
[Fregean] concept' of number. Fi nal l y, it is cl ear that Peano' s axi om
ati c defnes whol e numbers and them al one, as a rule-governed opera
tional domai n. Real numbers can certai nl y be defned di rectly wi th a
speci al axiomatic ( that of a complete, total l y ordered Archi medean
feld) . But, i f the essence of ' number' resides i n the speci fcity of the
statements constituting these axi omatics, then, given that these state
ments are entirely di ssi mi l ar i n the case of the axi omatic of whol e
numbers and of that of real numbers, i t woul d seem that, i n respect
of their concept, whole numbers and real numbers have nothi ng
In common.
1 2 GENEALOGIES: FREGE. DEDEKIND. PEA NO. CANTOR
1 . 1 1 . It is as i f, chal l enged to propose a concept of number that can
endure the modern ordeal of the defection of the One, our thi nkers
reserve the concept for one of its ' i ncarnati ons' ( ordi nal , cardi nal ,
whol e, real . . . ) , wi thout bei ng abl e to account for the fact that the
idea and the word ' number' are used for all of these cases. More
parti cul arl y, they prove i ncapabl e of defning any uni fed approach,
any common ground, for discrete numeration ( whole numbers ) , con
ti nuous numeration ( real numbers ) and 'general ' , or set-theoretical,
numeration ( ordinal s and cardinal s) . And yet it was precisely the
problem of the conti nuum, the dialectic of the discrete and the con
ti nuous, which, saturating and subverting the ancient opposition
between ari thmetic and geometry, compel led the moderns to rethink
the idea of number. In thi s sense thei r work, admi rabl e as it i s in so
many ways, is a fai l ure.
1 . 12. The anarchy thus engendered ( and I cannot take thi s anarchy
to be i nnocent of the unthi nki ng despotism of number) i s so much
the greater i n so far as the methods put to work in each case are
total l y di sparate:
( a) Natural whol e numbers can be determined either by means of
a speci al axi omati c, at whose heart i s the pri nci pl e of recurrence
( Peano) , or by means of a particul ar ( fnite) case of a theory of
ordi nal s, i n whi ch the pri nci pl e of recurrence becomes a theorem
( Dedeki nd) .
( b) To engender negative numbers, algebraic mani pul ati ons must
be i ntroduced that do not bear on the ' bei ng' of number, but
on its operati onal arrangement, on structures ( symmetricisation
of additi on) .
( c) These mani pul ati ons can be repeated t o obtain rational numbers
( symmetrici sati on of mul ti pl icati on) .
( d) Onl y a fundamental rupture, marked thi s time by a shi ft towards
the topol ogical , can found the passage to real numbers ( consid
eration of i nfnite s ubsets of the set of rati onal s, cuts or Cauchy
seq uences ) .
( e) We return to algebra to construct the fel d of complex numbers
( algebrai c cl osure of the Real Field, adj unction of the ' i deal '
el ement i = , or di rect operati onal axi omatisation on pai rs
of real numbers ) .
( f) Ordi nal s are i ntroduced through the consi deration of types of
order ( Cantor) , or through the use of the concept of transitivity
( von Neumann) .
GREEK NUMBER AND MODERN NUMBER 1 3
( g) The cardi nal s are treated through a total l y di fferent procedure,
that of bi uni vocal correspondence. 1
3
1 . 1 3. This arsenal of procedures was hi storical l y deployed according
to overl appi ng l i nes whi ch passed from the Greeks, the Arab algebra
ists and those of Renai ssance Ital y, through al l the founders of modern
anal ysi s, down to the 'structural i sts' of modern algebra and the set
theoretical creations of Dedeki nd and Cantor. How are we to extract
from it a clear and uni vocal i dea of number, whether we thi nk i t as
a type of being or as an operational concept? All that the thi nkers of
number have been able to do is to demonstrate the intellectual pro
cedures that l ead us to each species of ' number' . But, i n doi ng so,
they left number as such i n the shadow of its name. They remai ned
di stant from that ' uni que number whi ch cannot be any other' ,
I
4
whose stel l ar i nsurrection Mal larme proposed.
1 . 14. The questi on, then, is as fol l ows: is there a concept of number
capabl e of subsumi ng, under a si ngl e type of being answering to a
uni form procedure, at least natural whol e numbers, rational numbers,
real numbers and ordi nal numbers, whether fnite or i nfnite? And
does it even make sense to speak of a number without at once speci
fying whi ch singular, i rreduci bl e apparatus i t belongs to? The answer
is yes. Thi s i s precisely what i s made possi bl e by the margi nal theory,
which I propose to make phi l osophi cal l y central , of ' surreal
numbers' .
Thi s theory offers us t he true contemporary concept of number,
and i n doi ng so it overcomes the i mpasse of the thi nki ng of number
in its modern-cl assi cal form, that of Dedeki nd, Frege and Cantor. On
i ts basi s, and as the resul t of a long l abour of thought, we can prevai l
over the bl i nd despotism of the numeri cal unthought.
1 . 1 5. We must speak not of a si ngle age of the modern thi nki ng of
number, but of what one might cal l , taki ng up an expressi on Natacha
Mi chel appl i es to l i terature, the ' frst moderni ty' of the thi nki ng of
number.
I
S The names of thi s frst modernity are not those of Proust
and Joyce, but those of Bolzano, Frege, Cantor, Dedeki nd and Peano.
I am attempting the passage to a second modernity.
1 . 16. I have sai d that the three chal lenges to which a modern doc
trine of number must address itself are those of the i nfnite, of zero
and of the absence of any grounding by the One. I f we compare Frege
and Dedeki nd - so close on so many poi nts - on thi s matter, we
i mmediately note that the order in which they arrange thei r responses
to these chal lenges di ffers i n an essenti al respect:
On the infnite Dedeki nd, with admi rable profundity, begins with the
infnite, whi ch he determines with a cel ebrated positive property: ' A
system S i s sai d to be i nfnite when it is similar to a proper part of
i tsel f. '
1
6 And he undertakes i mmediately to ' prove' that such an inf
nite system exi sts. The fnite will be determined onl y subsequentl y,
and i t will be the fnite that is the negation of the i nfnite ( i n which
regard Dedeki nd' s numeri cal di al ectic has somethi ng of the Hegelian
about i t) .
1
7 Frege, on the other hand, begi ns with the fnite, with
natural whole numbers, of which the i nfnite will be the ' prol onga
ti on' or the recol lection i n the concept.
I
S
On zero Dedeki nd abhors the voi d and its mark, and says so quite
expl i ci tl y: '[W]e intend here for certai n reasons whol l y to exclude the
empty system which contains no elements at al l . '
1
9 Whereas Frege
makes the statement 'zero is a number,
2
0 the rock of his whole
edifce.
On the One There is no trace of any pri vi leging of the One in Frege
( preci sel y because he starts audaci ousl y with zero) . So one
-
rather
than the One - comes onl y i n second pl ace, as that which fal l s under
the concept ' i dentical to zero' (the one and onl y obj ect that fal l s under
the concept bei ng zero i tsel f, we are entitled to say that the extension
of thi s concept i s one) . Dedeki nd, on the other hand, retai ns the idea
that we shoul d ' begi n' wi th one: 'the base-element 1 i s cal led the
base-number of the number-series N' .
2
1
And, correlativel y, Dedeki nd
fal l s back without hesi tati on on the idea of an absol ute AW
2
of
thought, an i dea that coul d not appear as such in Frege' s formal i sm:
' My own real m of thoughts, i . e. the total i ty S of al l things, whi ch can
be obj ects of my thought, is i nfni te. '
2
3 Thus we see that, i n retai ni ng
the rights of the One, the Al l is supposed, because the Al l is that
which, necessari l y, proceeds from the One, once the One is.
1 . 1 7. These divergences of order are no mere technical matter. They
relate, for each of these thi nkers, to the respective centre of gravity
of their conception of number and - as we shal l see - to the si mul ta
neous stoppi ng pui nt and foundi ng poi nt of thei r thought: the infnite
and exi stence for Dedeki nd, zero and the concept for Frege.
1 . 1 8. The passage to a second modernity of the thi nki ng of number
obl iges thought to retur to zero, to the i nfnite and to the One. A
GREEK NUMBER AND MODERN NUMBER 1 5
total di ssi pation of the One, an ontol ogi cal deci si on as to the being
of the void and that which marks it, a l avi shi ng without measure of
infnities: such are the parameters of such a passage. Unbi ndi ng from
the One del i vers us to the unicity of the void and to the di ssemi nation
of the i nfnite.
2
Frege
2. 1 . Frege
1
mai ntai ns that pure thought engenders number. Like
Mal l arme, al bei t wi thout the effect of Chance, Frege thi nks that
'every thought emits a dicethrow' .
2
What is cal led Frege' s ' logicism'
runs very deep: number is not a si ngul ar form of being, or a particul ar
property of thi ngs. It is neither empi ri cal nor transcendent. Nor is i t,
on the other hand, a constitutive category; it is deduced from the
concept. It is, i n Frege' s own words, a trait of the concept.
3
2. 2. The pi votal property that permits the transition from pure
concept to number is that of a concept' s extension. What does thi s
mean? Gi ven any concept whatsoever, an obj ect ' fal l s' under thi s
concept i f it i s a ' truth-case' of thi s concept, if the statement that
attri butes to thi s obj ect the property compri sed in the concept is a
true statement. In other words, if the obj ect satisfes the concept. Note
that everythi ng ori gi nates with the truth-val ue of statements, which
is thei r denotation ( truth or fal sity) . It coul d be sai d that, i f the
concept generates number, it does so only i n so far as there is truth.
Number is in thi s sense the index of truth, not an i ndex of being.
But the idea of extensi on i s rami fed and obscure.
2. 3. Gi ven a concept, by the extensi on of that concept we mean all
the truth-cases ( al l obj ects qua truth-cases) that fal l under thi s concept.
Every concept has an extensi on.
Now, take two concepts C
1
and C2 We wi l l cal l them equinumer
ate4 if there exists a bi univocal correspondence associating, obj ect for
FREGE 1 7
obj ect, that whi ch fal l s under concept C
1
wi th that whi ch fal l s under
concept C2 That i s, i f a bi univocal correspondence can be defned
between the extensi on of C
1
and the extensi on of C2
It is cl ear that Frege favours a ' cardi nal ' defnition of number; that
he is not overly concerned with the structural order of that which fal l s
under the concept. And in fact thi s essential tool of bi uni vocity i s
characteristic of al l attempts to ' number' the mul ti pl e i n itsel f, the pure
mul tipl e subtracted from al l structural consi derations. To say that two
concepts are equi numerate is to say that they have the 'same quantity' ,
that thei r extensions are the same size, abstracting from any con
sideration as to what the obj ects are that fal l under those concepts.
2. 4. Number consi sts in marking equi numeracy, the quantitative
identity of concepts. Whence the famous defni ti on: 'The number
which belongs to the concept C i s the extensi on of the concept "equi
numerate to concept C". ' 5 Whi ch means: every concept C generates
a number
-
namely, the set of concepts equi numerate to C, havi ng
the 'same pure quantity' , the same quanti ty of extensi on, as C. Note
that a number, grasped in its being, al ways designates a set of con
cepts, namel y al l those that sati sfy the statement ' i s a concept equi
numerate to C' .
2. 5. The concept of number i s constructed through the fol l owi ng
progressi on:
Concept Truth Obj ects that fal l under the concept ( that
satisfy the statement attri buting the concept to the obj ect)
Extension of the concept ( al l truth-cases of the concept) Equi
numeracy of two concepts ( vi a bi univocal correspondence of thei r
extensi ons) Concepts that fal l under the concept of equi nu
me racy to a gi ven concept C ( that sati sfy the statement ' i s equi nu
merate to C' ) The extensi on of equi numeracy-to-C ( the set of
concepts from the precedi ng stage) The number that bel ongs
to concept C ( number i s thus the name for the extensi on of
equi numeracy-to-C) .
From a si mpl i fed and operati onal poi nt of view, i t coul d al so
be sai d that, starting from the concept, we are able to pass through
the obj ect on condition that there i s truth; that we then compare
concepts, and that number names a set of concepts that have in
common a property made possi bl e and defned by thi s comparison
( equi numeracy) .
2.6. To rediscover the ' ordi nary' , fami l iar numbers on the basis of
thi s pure conceptual i sm regul ated by truth al one, Frege begins with
his admi rabl e deduction of zero: zero i s the number belonging to the
concept ' not i denti cal to itsel f' . 6 Si nce every obj ect is identical to itsel f,
the extensi on of the concept ' not identical to itsel f' is empty. It
fol l ows that zero is the set of concepts whose extensi on is empty and
whi ch, by vi rtue of thi s, are equi numerate to the concept 'not
identical to itsel f' . Whi ch is preci sel y to say that zero i s that number
bel ongi ng to every concept whose extensi on is empty, i s zero.
I have i ndicated in 1 . 1 7 the passage to the number 1: 'One' i s the
number that bel ongs to the concept ' i dentical to zero
,
. 7 It i s interest
ing to note that Frege emphasises, with regard to 1, that it has no
' intuiti ve' or empirical pri vi lege, any more than it is a transcendent
foundati on: 'The defni ti on of 1 does not presuppose, for its obj ective
legitimacy, any matter of observed fact.
,
8 There can be no doubt that
Frege parti ci pates i n the great modern process of the destitution of
the One.
The engendering of the sequence of numbers beyond 1 poses only
techni cal probl ems, which are resol ved, in passi ng from n to n + 1 ,
by constructing between the extensi ons o f corresponding concepts a
correl ati on such that the ' remai nder' is exactly 1 - which has al ready
been defned.
2. 7. Thus the deduction of number as a consequence of the concept
appears to have been accompl i shed. More exactly: from the triplet
conceptltruth!obj ect, and from the si ngl e formal operator of bi uni vo
cal correspondence, number emerges as an instance of pure thought,
or an i ntegral ly logical producti on; thought must presuppose itsel f,
in the form of a concept susceptible to having truth-cases ( and there
fore endowed with an extensi on) . In so doi ng, thought presupposes
number.
2. 8. Why choose particul arly the concept ' not identical to itself' to
ground zero? Any concept coul d be chosen so long as one is sure it
has an empty extensi on, that no thi nkable obj ect coul d have the
property it designates. For exampl e ' square ci rcl e' - a concept which
i n fact Frege decl ares is ' not so black as [it is] pai nted
,
. 9 Si nce we
seek an entirely conceptual determi nation of number, the arbitrary
nature of this choice of concept i s a l i ttle embarrassing. Frege i s quite
aware of thi s, si nce he writes: 'I coul d have used for the defnition of
nought any other concept under whi ch no obj ect fal l s. "
o
But, to
obvi ate his own obj ecti on, he i nvokes Lei bni z: the Pri nci pl e of Iden
tity, which says that every obj ect i s identical to itsel f, has the merit
FREGE 1 9
of being ' purel y logical ' .
1 1
Purely l ogical ? But we understood that it
was a matter of legitimating l ogico-mathemati cal categories ( speci f
cal ly, number) on the sol e basi s of the l aws of pure thought. Isn' t
there a ri sk of ci rcul arity i f a l ogi cal rul e is requi red ri ght at the
outset ? Now, equal ity i s one of the l ogical , or operati onal , predicates
that requi re grounding ( namel y, equal ity between numbers ) . It might
be sai d, of course, that ' i dentical to itsel f' shoul d not be confused
with 'equal to itsel f' . But i f ' i dentity' must here i ndeed be careful l y
di stinguished from the l ogi cal predicate of equal i ty, it i s neverthel ess
equal l y cl ear that the statement ' every obj ect is i dentical to i tsel f' i s
not a ' purel y logical ' statement. It is an onto-logical statement. And,
qua ontologi cal statement, it is i mmediatel y di sputabl e: no Hegel i an,
for exampl e, woul d admi t the uni versal val idity of the pri nci pl e of
identity. For our hypotheti cal Hegel i an, the extensi on of the concept
' not identical to itsel f' is anythi ng but empty!
2. 9. The purely a priori determi nati on of a concept certain to have
an empty extensi on is an i mpossi bl e task wi thout powerful pri or
ontol ogical axi oms. The i mpasse that Frege meets here i s that of an
unchecked doctrine of the object. For, from the poi nt of view of the
pure concept, what is an ' obj ect' in general , any obj ect whatsoever,
taken from the total Universe of obj ects ? And why i s the obj ect
requi red to be identical to itsel f, when the concept i s not even requi red
to be non-contradictory in order to be l egi ti mate, as i ndicated
by Frege' s positive regard for concepts of the ' square ci rcl e' type,
which, he stresses, are concepts l i ke any other? Why woul d the
law of the being of obj ects be more stringent than the l aw of the
being of concepts ? Doubtless it would be so if one were to accept
Leibnizian ontology, for whi ch exi stent obj ects obey an other
principle than do thi nkabl e obj ects, the Pri nci pl e of Suffcient
Reason. It thus appears that the deduction of number on the basis
of the concept i s not so much universal , or 'purely l ogical ' , as it is
Lei bnizi an.
2. 10. To posit as sel f-evident that the extensi on of a concept i s thi s
or that ( for exampl e, that the extensi on of the concept ' not identical
to itsel f' is empty) i s tantamount to supposi ng that we can move
unproblematical l y from concept to exi stence, given that the extensi on
of a concept bri ngs i nto pl ay the ' obj ects ' that fal l under thi s concept.
A general ised ontol ogical argument is at work here, and it i s thi s very
argument that subtends the deduction of number on the basi s of the
concept al one: number belongs to the concept through the mediation
of the thinkable objects that fall under the concept.
2. 1 1 . The pri nci pal thought-content of Russel l ' s paradox, com
municated to Frege i n 1903, is its undermi ni ng of every pretensi on
to l egi sl ate over existence on the basi s of the concept al one, and
especi al l y over the exi stence of the extensi on of concepts. Russell
presents a concept (in Frege' s sense) - the concept 'to be a set that is
not an element of itself' - whi ch i s surel y a wholly proper concept
( more so, truth be told, than ' not identical to itself' ) , but one none
theless whose extension does not exist. It is actual l y contradictory to
suppose that ' obj ects' - i n this i nstance, sets - that ' fal l under this
concept' themselves form a set. 1
2
And, i f they 40 not form a set, then
no bi uni vocal correspondence whatsoever can be defned for them.
So thi s 'extensi on' does not s ustain equi numeracy, and consequently
no number belongs to the concept 'set that i s not an element of
itsel f' .
The advent, to the concept, of an i nnumerable rui ns Frege' s
general deducti on. And, taki ng i nto account the fact that the para
doxi cal concept in question i s quite ordi nary ( for exampl e, the concept
i s val i d for all the sets customari l y used by mathematici ans: they are
not elements of themselves ) , we mi ght wel l suspect that there exist
many other concepts to which no number belongs. In fact, it is impos
sible to predict a pri ori the extent of the disaster. Even the concept
' not i denti cal to itsel f' coul d well turn out not to have any existent
extensi on, whi ch i s somethi ng entirely di fferent from havi ng an empty
extensi on. Let' s add that Russel l ' s paradox is purely logi cal , that is
to say, i t i s preci sel y proven: to admit the existence of a set of all
those sets that do not bel ong to themselves undermines deductive
l anguage by i ntroducing a formal contradi ction ( the equivalence
between a proposi ti on and its negation) .
2. 1 2. A sort of ' repai r' was proposed by Zermel o.
1
3
It consi sts in
sayi ng that we can concl ude from the concept the exi stence of its
extensi on on condition that we operate within an already given exis
tence. Gi ven a concept C and a domain of existing objects, we can
say that there exi sts, in this existing domain, the set of obj ects that
fal l under thi s concept - i . e. the extensi on of the concept. Obvi ousl y,
thi s extensi on is rel ati ve to a domai n speci fed i n advance and does
not exi st ' i n itself' . Thi s i s a maj or ontological transformation: withi n
thi s new framework it i s not possi bl e to move from concept to exi s
tence ( and thus to number) ; we can onl y move to an exi stence that
i s somehow carved out of a pre-given exi stence. We can 'separate' i n
a gi ven domai n those obj ects wi thi n it that val i date the property
exposed by the concept. Thi s i s why Zermel o' s pri nci pl e, whi ch dras
tical l y l i mits the rights of the concept and of l anguage over existence,
FREGE 21
is cal led the Axiom of Separati on. And it does i ndeed seem that
accepting thi s axi om safeguards us agai nst the i nconsi stency-effects
of Russel l -type paradoxes.
2. 1 3. Russel l ' s paradox is not paradoxi cal in the sl ightest. It is a
materi al i st argument, because it demonstrates that mul ti ple-being is
anterior to the statements that affect it. It i s i mpossi bl e, says the
' paradox' , to accord to l anguage and to the concept the right of
unfettered legislation over exi stence. Even supposi ng that there i s a
transcendental function of l anguage, it supposes al so the avai l abi l ity
of some pri or existent, the power of thi s function being si mpl y that
of carving out or del i miting extensi ons of the concept wi thi n thi s
specifed existent.
2. 14. Can we, in assumi ng Zermel o' s axiom, save the Fregean con
struction of number? Once agai n, everythi ng turns on the questi on
of zero. I might proceed i n the fol l owing way: given a del i mited
domain of obj ects, whose existence i s somehow external l y guaran
teed, I wi l l cal l ' zero' ( or ' empty set' , whi ch is the same thi ng) that
which detaches, or separates, wi thi n thi s domai n, the concept ' not
identical to itsel f' , or any other such concept under which I can assure
mysel f that no obj ects of the domai n fal l . As we are deal i ng with a
limited domai n, and not, as in Frege' s constructi on, wi th ' al l obj ects'
( a formulation that led to the i mpasse of a Lei bni zi an choice wi thout
criteri a) , there i s a chance of my fnding such a concept. If, for
example, I take a set of bl ack obj ects, I wi l l cal l ' zero' that which
separates in thi s set the concept 'to be white'. The rest of the con
struction wi l l fol l ow.
2. 1 5. But what domai n of obj ects coul d I start wi th, for whi ch it can
be guaranteed that these obj ects pertai n to pure thought, that they
are ' purely logical ' ? Recal l that Frege i ntends to construct a concept
of number that i s, according to his own expressi on, ' not . . . either
anything sensi ble or a property of an external thing' ,
1
4
and that he
emphasises on several occasi ons that number i s subtracted from the
representabl e. Establ i shing that number is a producti on of thought,
deducing it from the abstract attributes of the concept i n general -
this cannot be achieved usi ng black and white ohj ects . The question
then becomes: what exi stent can I assure mysel f of, outside of any
experience? Is the axi om ' somethi ng exi sts' an axi om of pure thought
and, supposing that i t i s, can I discern any property of whi ch it
is certain that it does not belong i n any way to this exi stent
'something' ?
2. 16. A ' purely logical ' demonstration of existence, for thought, of
a nondescri pt obj ect, a poi nt of being, an ' obj ect = x': the statement
'every x i s equal to itsel f' is an axi om of logic with equal ity. Now,
the uni versal rules of frst-order logic, a logic val i d for every domai n
of obj ects, al low us to deduce, from the statement ' every x is equal
to x' , the statement 'there exists an x that i s equal to x' ( subordina
ti on of the exi stenti al quantifer to the universal quantifer) . 15 There
fore, there exists x (at least that x which is equal to itsel f) .
Thus we can demonstrate wi thi n the framework of set theory, frst
of all, by purel y l ogi cal means, that a set exists. Then we can separate
the empty set wi thi n that exi stent whose exi stence has been proved,
by uti l i si ng a property that no element can sati sfy ( for example, ' i s
not equal to itself' ) . We have respected Zermel o' s axi om, since we
have operated wi thi n a pri or exi stent, but we have succeeded i n
engenderi ng zero.
2. 1 7. It i s quite obvi ous, I thi nk, that this ' proof' is an unconvi nci ng
artifce, a l ogi cal sleight of hand. From the uni versal postul ate of
sel f-equal i ty ( whi ch we might possi bl y accept as an abstract law, or
a l aw of the concept) , who coul d reasonabl y i nfer that there exists
somethi ng rather than nothing? If the uni verse were absol utely void,
it would remai n logi cal l y admi ssi bl e that, supposi ng that somethi ng
exi sted ( which woul d not be the case) , it woul d have to be equal to
itsel f. The statement ' every x is equal to x' woul d be val i d, but there
woul d be no x, so the statement ' there exi sts an x equal to itsel f'
woul d not be val i d.
The passage from universal statement to assertion of exi stence i s
an exorbi tant ri ght, whi ch the concept cannot arrogate to itself. It is
not possible to el icit exi stence on the basi s of a uni versal law that
could be upheld j ust as well i n absol ute nothingness ( consi der for
exampl e the statement ' the nothi ng is identical to itsel f' ) . And, si nce
no exi stent obj ect can be deduced from pure thought, you cannot
separate zero therei n. Zermel o does not save Frege.
2. 1 8. The exi stence of zero, or of the empty set, and therefore the
exi stence of numbers, i s in no way deduci bl e from the concept, or
from l anguage. 'Zero exi sts' i s inevitabl y a frst assertion; the very
one that fxes an exi stence from whi ch all others wi l l proceed. Far
from it being the case that Zermel o' s axi om, combi ned with Frege' s
logici sm, al l ows us to engender zero and then the chai n of numbers,
it is on the contrary the absol utely i naugural exi stence of zero ( as
empty set) that ensures the possi bi l ity of separating any extensi on of
a concept whatsoever. Number comes frst here: it is that point of
FREGE 23
being upon whi ch the exerci se of the concept depends. Number, as
number of nothi ng, or zero, s utures every text to i ts l atent bei ng. The
void is not a production of thought, because it i s from its exi stence
that thought proceeds, i n as much as ' it i s the same thi ng to thi nk
and to be' .
1
6 I n thi s sense, i t i s the concept that comes from number,
and not the other way around.
2. 19. Frege' s attempt i s uni que i n certai n regards: i t i s not a matter
of creating new intra-mathematical concepts (as wi l l be the case i n
Dedeki nd and Cantor) , but of el uci dati ng - with the sol e resource of
rigorous anal ysi s - what, among the possi ble obj ects of thought,
si ngul ari ses those whi ch fal l under the concept of number. I n thi s
respect, my own efforts fol l ow along the same l i nes. We si mpl y need
to remove the obstacles by reframi ng the i nvestigation according to
new parameters. Above al l , i t must be shown that thought i s not
constituted by concepts and statements alone, but al so by deci si ons
that engage it wi thi n the epoch of its exercise.
3
Additional Note on a
Contemporary Usage of Frege
- -
3. 1 . Jacques-Al ai n Mi l l er, in a 1965 lecture entitled ' Suture' and
subtitled ' Elements of the l ogic of the si gni fer' ,
1
put forward a reprise
of Frege' s construction of number. Hi s text founds a certain regime
of compati bi l ity between structural i sm and the Lacanian theory of
the subj ect. 1 am mysel f peri odi cal l y brought back to thi s foundation,
2
al bei t onl y on condi ti on of di srupting it somewhat. Twenty-fve years
l ater, ' I am here; 1 am sti l l here' .3
3. 2. Mi l l er puts the fol l owing question to Frege: ' What is it that
functions in the sequence of whol e numbers ? ,
4
And the response to
thi s question - a response, might 1 say, forceful l y extorted out of
Frege - i s that 'in the process of the constitution of the sequence, the
function of the sub;ect, unrecognised,
s
i s operative' . 6
3. 3. If we take thi s response seri ousl y, it means that, in the l ast
i nstance, i n the proper mode of its mi scogni ti on, it is the function of
that subj ect whose concept Lacan' s teachi ng communicates to us that
consti tutes, i f not the essence, at least the process of engenderment
(the 'genesis of progressi on' , says Mi l lerf of number.
Obvi ousl y such a radical thesi s cannot be ignored. Radical, it
woul d seem at frst gl ance, wi th regard to Frege' s doctri ne, which
dedicates a specifc argument to the refutation of the idea that number
might be ' subj ective, 8 ( al though it i s true that, for Frege, ' subj ective'
means 'caught up i n representation' , whi ch obvi ousl y does not match
the Lacani an function of the subj ect) , Radi cal al so wi th regard to my
A CONTEMPORARY USAGE OF FREGE 25
own thesi s, si nce I hol d that number i s a form of bei ng, and that, far
from being subtended by the function of the subj ect, it is on the
contrary on the basi s of number, and especi al l y of that frst number
being that is the void ( or zero) , that the function of the subj ect
recei ves its smal l share of bei ng.
3. 4. We wi l l not undertake here to exami ne the i mportance of thi s
text - the frst great Lacani an text not to be written by Lacan hi msel f
-
for the doctri ne of the si gni fer, nor to expl ore what analogy it
employs to i l l umi nate the i mportance - at the ti me, sti l l l ittle appreci
ated - of al l that the master taught us as to the subj ect' s being com
pri sed i n the effects of a chai n. We seek to exami ne excl usi vel y what
Mi l ler' s text assumes and proposes with regard to the thinking of
number as such.
3. 5. Mi l ler's demonstrati on is organised as fol l ows:

To found zero, Frege ( as we saw i n 2. 6) summons to hi s ai d the
concept ' not identi cal to itsel f' . No obj ect fal l s under this concept.
On this poi nt, Mi l ler emphasises - even compounds - Frege' s refer
ence to Lei bni z. To suppose that an obj ect could be not be identical
to itsel f, or that it coul d be non-substitutabl e for itsel f, woul d be
entirely to subvert truth. I n order to be true, a statement bearing upon
obj ect A must suppose the in variance of A in each occurrence of the
statement, 'each ti me' the statement is made. The pri nci ple ' A i s A'
is a l aw of any possi ble truth. And reci procal l y, i n order that truth
be saved, it is cruci al that no obj ect shoul d fal l under the concept
' not identical to itself' . Whence zero, whi ch numbers the extensi on
of such a concept.

Number i s thus shown to issue from the concept al one, on condi
tion of truth. But thi s demonstration i s consistent onl y because it has
been able to invoke in thought an object non- identical to itself, even
i f onl y to di scharge it in the inscription of zero. Thus, Mi l ler writes,
'the 0 whi ch is inscribed i n the place of the number consummates the
excl usion of thi s obj ect' . 9
To say that ' no obj ect' fal l s under the concept ' not identical to
itself' is to make thi s obj ect vani sh as soon as it is invoked, in thi s
nothing the onl y subsi sti ng trace of whi ch wi l l be, preci sel y, the mark
zero: ' Our purpose has been, ' Mi l l er concl udes, 'to recognize in the
zero number the suturing stand-i n for the l ack' .
1 0

What is i t that comes to lack thus ? What ' obj ect' can have as a
stand-in for its own absence the frst numerical mark; and support,
26 GENEALOGIES: FREGE, DEDEKIND, PEANO, CANTOR
in relation to the whule chai n of numbers, the uni nscri babl e pl ace of
that whi ch appears onl y in order to vani sh? What i s it that i nsi sts
between numbers ? We must certai nl y agree that no ' obj ect' can, even
by fai l ure or default, fal l i n that empty pl ace that assigns non
sel f-i dentity, But there does exist (or here, more preci sel y, ek-si st)
preci sel y that whi ch is not obj ect, that whi ch i s proper to the non
obj ect, the obj ect as i mpossi bi l ity of the obj ect: the subj ect. 'The
i mpossi bl e obj ect, which the di scourse of logic summons as the not
identical with itsel f and then rej ects . , . i n order to constitute itself as
what it i s, whi ch it summons and rej ects wanting to know nothing
of it, we name thi s obj ect, in so far as it functions as the excess whi ch
operates i n the sequence of numbers, the subj ect. '
|
j
3. 6. We must meticul ousl y di sti ngui sh between that which Mi l ler
assumes from Frege and that whi ch he deci phers i n Frege' s work on
hi s own account, I wi l l proceed i n three stages.
3. 7. IIK51 51AGL Mi l ler takes as hi s starting poi nt the proposition
of Lei bniz-Frege according to which salva veritatel
2
demands that al l
obj ects shoul d be identical to themselves. The whol e l iteral i sation of
the real towards which Lei bni z worked al l hi s l i fe, and to which
Frege' s i deography i s the undoubted hei r, i s i n fact surreptiti ousl y
assumed here. In thi s regard, Mi l ler i s indeed right to equate, along
with Lei bni z, ' identi cal to itself' and ' substitutabl e' , thus denoting an
equi valence between the obj ect and the l etter. For what coul d it mean
to speak of the substitutabi l ity of an obj ect? Onl y the l etter i s entirely
substitutabl e for itself. ' A i s A' is a pri nci pl e of letters, not of obj ects.
To be identi fabl e at a remove from itself, and subj ect to questions
of substitutabi l ity, the obj ect must fall under the authority of the
letter,
1
3
which al one renders it over to cal cul ati on. I f A is not identical
at all moments to A, truth (or rather veridical ity) as calculation
col l apses.
The l atent hypothesi s i s therefore that truth is of the order of cal
culation. It is onl y on thi s supposi ti on that, frstly, the obj ect has
to be represented as a l etter; and, secondl y, that the non-self-identity
of the obj ect-letter radical l y subverts truth. And i f truth is of the
order of calcul ati on, then zero - which numbers the excl usi on
of the non-sel f-identical ( the subj ect) - i s itsel f nothing but a letter,
the letter O. The concl usi on then fol l ows straightforwardly that
zero i s the i nert stand-i n for lack, and that what 'drives' the sequence
of numbers as a product of marks - a repetition in which is articu
l ated the mi scogni ti on of that which i nsi sts - is the function of the
subj ect.
More si mpl y: i f truth i s saved onl y by uphol di ng the pri nci pl e of
identity, then the obj ect emerges i n the fel d of truth onl y as a letter
amenable to cal cul ati on. And, i f thi s i s the case, number can sustai n
itsel f onl y as the repetition of that whi ch i nsi sts i n l acki ng, whi ch i s
necessarily the non-obj ect ( or the non-letter, whi ch i s the same thi ng) ,
the pl ace where ' nothing can be wri tten' \
4
- i n short, the subj ect.
3. 8. No one is obl iged to be a Lei bni zi an, even i f we must recognise
in thi s phi l osophy the archetype of one of the three great orientations
in thought, the constructi vist or nomi nal i st orientation ( the other two
being the transcendent and the generic) . \ 5 As an advocate of the
generic orientati on, I decl are that, for truth to be saved, one must
precisely abolish those two great maxi ms of Lei bni zi an thought, the
Princi pl e of Non-Contradiction and the Princi pl e of Indi scerni bl es.
3. 9. A truth supposes that the situation of whi ch it i s the truth attai ns
non-self-i dentity: thi s non-sel f-identity i s i ndicated by the si tuati on' s
being supplemented by an 'extra' multi pl e, one whose belonging or
non-belonging to the si tuati on i s, however, i ntri nsi cal l y undeci dabl e.
I have named thi s supplement 'event' , and it i s al ways from an event
that a truth-process originates. Now, when the undeci dabl e event
must be decided wi thi n the si tuati on, that si tuati on necessari l y under
goes a vaci l lation as to its identity.
3. 10. The process of a truth - puncturing the strata of knowledge
harboured by the situation - inscri bes itsel f wi thi n the si tuati on as
i ndi scerni bl e i nfnity, whi ch no thesaurus of establ i shed l anguage has
the power to designate.
Let' s say si mpl y that zero, or the void, has nothing i n itsel f to do
with the sal vation of truth, which i s at play i n the ' l aboured' correl a
ti on between the undeci dabi l ity of the event and the i ndi scerni bi l ity
of its result within the si tuati on. No more so than it is possi bl e to
refer truth to the power of the letter, si nce the exi stence of a truth is
precisely that to whi ch no inscription can attest. The statement 'truth
i s' - far from guaranteeing that no obj ect fal l s under the concept of
' not identical to itsel f' and that therefore zero is the number of that
concept - i nstead al lows us thi s threefol d concl usi on:
-
there exists an obj ect that has attained ' non-sel f-i dentity' ( unde
ci dabi l ity of the event) ;
-
there exist an i nfnity of obj ects that do not fal l under any concept
( indi scerni bi l ity of a truth) ;
- number is not a category of truth.
3. 1 1 . 5LCONl 51AGL What is the strategy of Mi l ler's text ? And
what role does number as such play wi thi n it? Is i t real l y about
argui ng that the function of the subj ect is i mpl icated - as a mi scog
ni sed foundation - in the essence of number? Thi s i s undoubtedly
what is stated i n all cl arity by the formul a I have al ready ci ted above:
' I n the process of the consti tuti on of the sequence [ of numbers 1 . . . the
function of the subj ect . . . i s operative, '
1
6 More preci sel y, onl y the
function of the subj ect - that which zero, as number, marks in the
pl ace of lack, holding the pl ace of its revocation - is capable of
expl ai ni ng what, in the sequence of numbers, functions as iteration
or repetiti on: being excl uded, the subj ect (the non-sel f-identical )
i ncl udes itsel f through the very i nsistence of marks, incessantl y repeat
ing the ' one more step' , frstly from 0 to 1 ( ' the 0 counts for 1 ' , notes
Mi l l er) , then i ndefni tel y, from n to n + 1 : ' i ts [the subj ect' s - in the
Lacani an sense 1 exclusion from the feld of number i s i dentifed with
repetiti on' .
1
7
3. 12. Other passages of Mi l ler' s text are more equivocal , i ndicating
an anal ogi cal reading. For exampl e: ' If the sequence of numbers,
metonymy of the zero, begi ns with its metaphor, i f the zero member
of the sequence as number i s only the stand-in suturing the absence
( of the absol ute zero) which moves beneath the chai n accordi ng to
the alternation of a representation and an excl usi on
-
then what is
there to stop us from recogni si ng i n the restored rel ati on of the zero
to the sequence of numbers the most el ementary articulation of the
subj ect' s rel ati on to the si gni fyi ng chai n? , 1 8 The word ' recogni si ng' is
compatible wi th the idea that the Fregean doctrine of number pro
poses a ' matri x' ( the title of another arti cl e by Mi l ler on the same
question)
1
9 that is i somorphic with ( maxi mum case) or si mi l ar to
( mi ni mum case) , but i n any case not identical to, the relation of the
subj ect to the signi fyi ng chai n. Frege' s doctrine woul d then be a per
ti nent analogon of Lacani an l ogi c: to whi ch we woul d have no reply,
si nce in that case Mi l l er' s text would not be a text about number. It
woul d be doubly not so: frstly because it would speak, not of number,
but of Frege' s doctrine of number ( without taki ng any position on
the val i dity or consistency of that doctrine) ; and secondly because it
would present the sequence of numbers as a di dactic vector for the
logic of the si gni fer, and not as an effective exampl e of an i mplica
ti on of the function of the subj ect i n the sequence of numbers.
3. 13. This critical evasi on assumes that two conditions are met: that
there i s, between the doctrine of number and that of the signifer,
i somorphi sm or si mi l arity, and not i dentity or exempl i fcati on; and
that Mi l ler does not account for the val i dity of the Fregean doctrine
of number.
3. 14. On thi s last poi nt, where, to my mi nd ( that i s, to one who is
concerned with the thi nki ng of number as such) , everythi ng hangs i n
the bal ance, Mi l l er mai ntai ns the suspense at every step. He speaks
of ' Frege' s system' wi thout our bei ng abl e to decide whether or not,
in hi s opi ni on, the l atter is an actual l y accompl i shed theory of number,
theory entirely defensi bl e i n essence. It is stri ki ng that at no poi nt
in thi s very subtle and intricate exercise are the i mmanent probl ems
of ' Frege' s system' ever rai sed - in parti cul ar, those that I highl ighted
above with regard to zero, the impact of Russel l ' s paradox, Zermel o' s
axi om and, ultimately, the relation between l anguage and existence.
It thus remai ns possible to bel ieve that the i somorphi sm si gni fer/
number operates between, on the one hand, Lacan and, on the other,
Frege reduced to a si ngul ar theory whose inconsistency i s of no con
sequence with regard to the anal ogi cal goal s pursued.
3. 1 5. Evi dentl y, it remai ns to be seen whether thi s i nconsi stency i sn' t,
as a resul t, transferred to the other pole of the analogy, that i s, to
the l ogic of the signifer. The risk i s not i nconsequenti al , given that
Mi l ler pl aces the latter i n a foundi ng posi ti on with regard to l ogi c
tout court - presumabl y i ncl udi ng Frege' s doctri ne: ' The frst [the
l ogi c of the signifer] treats of the emergence of the second [the l ogi c
of l ogicians] , and shoul d be concei ved of as the l ogic of the ori gi n of
l ogic. '
2
0 But what happens i f the compl eti on of thi s process of ori gi
nati on is i nduced, through the theme of the subj ect, by a scheme
( Frege' s) marred by inconsi stency? But thi s is not my probl em. Given
the conditions I have l ai d out, i f the text i s not about number, then
we are fni shed here.
3. 16. 1HIKl 51AGL There remai ns, however, an i ncontestabl e
degree of adherence on Mi l ler' s part to a common representation of
number, wherein number i s conceived of as i n some way intuitive,
and which I cannot accept. Thi s concerns the idea - central , si nce it
is precisely here that the subj ect makes itsel f known as the cause of
repetition - according to whi ch number is grasped as a ' functi oni ng' ,
or in the 'genesi s of the progressi on' . Thi s i s the i mage of a number
that is 'constructed' iteratively, on the basi s of that poi nt of puncture
that is denoted by zero. Thi s dynami cal theme, whi ch woul d have us
see number as passage, as sel f-producti on, as engenderment, i s omni
present i n Mi l ler' s text. The anal ysi s centres precisely on the ' passage'
from 0 to 1 , or on the ' paradox of engendering' n + 1 from n.
3. 1 7. Thi s i mage of number as iteration and passage precl udes any
orderly di scussi on of the essence of number. Even i f we can only
traverse the numeric domain accordi ng to certai n laws of progression,
of whi ch successi on is the most common ( but not the onl y one, far
from i t) , why must it fol l ow that these l aws are constitutive of the
being of number? It i s easy to see why we have to ' pass' from one
number to the next, or from a sequence of numbers to its l i mit. But
it i s, to say the least, i mprudent thereby to concl ude that number is
defned or constituted by such passage. It mi ght well be ( and thi s is
my thesi s) that number does not pass, that it is i mmemori al l y deployed
i n a swarmi ng
2
1
coextensive with its being. And we wi l l see that, j ust
as these labori ous passages onl y govern our passage through thi s
deployment, i n the same way it is l i kel y that we remai n i gnorant of,
have at the present ti me no use for, or no access to, the greater part
of those numbers that our thought can conceive of as existent.
3. 1 8. The ' constructivist' thesi s, which makes of iteration, succes
si on, passage the very essence of number, leads to the concl usi on that
very few numbers exi st, since here ' exi st' has no sense apart from
that effectively supported by some such passage. Certai nl y, i ntui ti on
ists assume thi s i mpoveri shed perspecti ve. Even a semi -i ntui ti oni st
l i ke Borel
22
thought that the great maj ority of natural whol e numbers
' don' t exi st' except as a fcti onal and i naccessi bl e mass. So it might
wel l be that the Lei bni zi an choice that Mi l ler borrows from Frege is
doubled by a l atent i ntui ti oni st choice.
We must recogni se that i ntui ti oni st logic and the l ogic of the si gni
fer have more than a l ittle i n common, i f onl y because the former
expressly i nvokes the subj ect ( the ' mathematici an-subj ect' ) as part of
its machi nery. But i n my opinion such a choice woul d represent an
additional reason not to enter i nto a doctri ne of number whose
overal l effect is to make the place of number, measured by the opera
tional i ntui ti on of a subj ect, i nexorabl y fnite. For the domain of
number i s rather an ontological prescri pti on i ncommensurable with
any subj ect and i mmersed i n the i nfnity of i nfniti es.
3. 1 9. The probl em now becomes: how to thi nk number whi lst admit
ting, agai nst Lei bni z, that there are real i ndiscerni bl es; against the
i ntui ti onists, that number persists and does not pass; and agai nst the
foundati onal use of the subj ecti ve theme, that number exceeds al l
fnitude ?
4
Dedekind
4. 1. Dedekind I introduces hi s concept of number wi thi n the frame
work of what we woul d today cal l a ' na'ive' theory of sets. ' Na'ive'
because a theory of mul ti pl icities is advanced that recapitulates
various presupposi ti ons about thi ngs and about thought. ' Na'ive'
meaning, in fact: phi l osophical .
Dedeki nd states expl i ci tl y, in the openi ng of hi s text The Nature
and Meaning of Numbers, that he understands 'by thing every obj ect
of our thought' ;
2
and, a l ittle later, that, when di fferent things are
' for some reason consi dered from a common poi nt of view, associ ated
in the mi nd, we say that they form a system S' .
3
A system i n Dede
ki nd' s sense is therefore quite si mpl y a set i n Cantor' s sense. The
space of Dedeki nd' s work i s not the concept ( as i n Frege) , but,
di rectly, the pure mul ti ple, a col lection that counts for one (as a
system) obj ects of thought.
4. 2. Dedeki nd develops a conception of number that ( l i ke Cantor' s)
is essenti al l y ordinal. We saw ( compare 2. 3) that Frege' s conception
was essential l y cardinal ( proceeding vi a bi uni vocal correspondences
between extensi ons of concepts ) . What i s the signi fcance of this di s
ti nction? In the ordi nal view, number i s thought as a l i nk i n a chai n,
it is an element of a total order. I n the cardi nal view, it i s rather the
mark of a ' pure quantity' obtai ned through the abstraction of domai ns
of obj ects havi ng 'the same quantity' . The ordi nal number is thought
according to the schema of a sequence, the cardi nal number, accord
ing to that of a measurement.
4. 3. Dedeki nd affrms that i nfnite number ( the totality of whole
numbers, for exampl e) precedes, i n constructi on, fnite number ( each
whol e number, its successor, and so on) . Thus the exi stence of an
i nfnite ( i ndetermi nate) system, and then the particul ar existence of
N ( the set of natural whole numbers ) form the contents of the para
graphs numbered 66 and 72 i n Dedeki nd' s text, whereas a result as
apparently elementary as 'every number n i s di fferent from the fol
l owi ng number n
"
comes onl y i n paragraph 8 1 .
Dedeki nd i s a true modern. He knows that the i nfnite is simpler
than the fnite, that it is the most general attri bute of being, an intu
ition from which Pascal had al ready drawn radical consequences -
and was the frst to do so - as to the si te of the subj ect.
4. 4. Dedeki nd frst of all i nvites us to accept the phi l osophi cal concept
of ' system' , or any mul ti pl icity whatsoever ( compare 4. 1 ) . The pri n
ci pal operator wi l l then be, as i n Frege ( 2. 3 ) , the i dea of bi uni vocal
correspondence between two systems. Dedeki nd, however, will make
use of it in a total l y di fferent way than di d Frege.
Let' s note in passi ng that the bi uni vocal correspondence, bi j ection,
is the key notion of all the thi nkers of number of this epoch. It orga
ni ses Frege's thought, Cantor' s and Dedeki nd' s.
4. 5. Dedeki nd cal l s the functi on, or correspondence, a 'transforma
ti on' ,
4
and what we would cal l a bi j ective function or a bi univocal
correspondence he cal l s a ' si mi l ar transformation' .
s
In any case, we
are deal i ng with a function ( which makes every element of a set ( or
system) S' correspond to an element ( and one onl y) of a set S, i n such
a way that:
- to two di fferent elements 5 1 and 52 of S wi l l correspond two di f
ferent elements (( s d and (( 52 ) of S/;
- every element of S
'
is the correspondent, through (, of an element
of S.
A distinct ( today we woul d s ay i nj ecti ve) functi on is a function
that compl i es onl y with the frst condi ti on:
Evidently, such functi ons can be defned ' i n' a system S, rather
than ' between' a system S and another system S/. Functi ons ( or
transformati ons) of thi s type make every el ement of S correspond
to an el ement of S (ei ther another element or the same one: the
DEDEKIND 33
function coul d be the function of identity, at least for the element i n
question) .
4. 6. Take, then, a system S, an appl ication ( ( not necessari l y one of
l i keness or a bi univocal one) of S t o itsel f, and s, an element of S. We
wi l l cal l the chain of the element s for the appl ication (, the set of
val ues of the function obtai ned by iterating it starting from s. So the
chai n of s for (i s the set whose elements are: s, (( s ) , (( (( s) ) , (( (((( s) ) ) , . . . ,
etc.
We are not necessari l y deal i ng here with an i nfnite iteration: it
could very well be that, at a certai n stage, the val ues thus obtai ned
would repeat themselves. Thi s i s evi dentl y the case i f S is fnite, si nce
the possi bl e val ues, which are the elements of S (the appl ication (
operates from S wi thi n S) , wi l l be exhausted after a fnite number of
stages. But it woul d al so be the case were one to come across a val ue
p of the function ( where, for p, ( is identical . Because then (( p) = p,
and therefore (( ((p) ) = ((p) = p. The function halts at p.
4. 7. We wi l l say that a system N is ( thi s i s Dedeki nd' s expressi on)
simply in{nite6 i f there exi sts a transformation ( of N wi thi n N that
complies wi th the three fol l owi ng condi ti ons:
1 The appl ication ( of N wi thi n N is a di stinct appl icati on (d.
4. 5) .
2 N is the chai n of one of its elements, which Dedeki nd denotes by
1 , and whi ch he cal l s the base-element of N.
3 The base-element 1 is not the correspondent through ( of any
element of N. I n other words, for any n which i s part of N,
f( n) . 1 : the function ( never ' returns' to 1 .
We can form a si mpl e enough i mage of such an N. We 'start' with
the element 1. We know (condi ti on 3 ) that (( 1 ) is an element of N
di fferent from 1 . Next we see that (((( 1 ) ) is di fferent from 1 ( which
is never a val ue for f. But, equal l y, (((( 1 ) ) i s di fferent from (( 1 ) . In
fact, the function ( ( condi ti on 1) is a di stinct transformati on - so two
di fferent elements must correspond, through (, to di fferent el ements.
From the fact that 1 is di fferent from (( 1 ) it fol lows that (( 1 ) is di f
ferent from (( (( 1 ) ) . More general l y, every element obtained through
the iteration of function ( wi l l be di fferent from all those that ' pre
ceded' it. And, since N (condi ti on 2) is nothing other than the chai n
thus formed, N wi l l be composed of an ' i nfnity' ( i n the i ntuitive
sense) of elements, all di fferent, ordered by function (, i n the sense
that each element ' appears' through an additional step of the
process that begi ns with 1 and is conti nued by repeatedly appl yi ng
operation f.
4. 8. The ' system' N thus defned i s the place of number. Why?
Because al l the usual ' numerical ' mani pul ati ons can be defned on the
elements n of such a set N.
By vi rtue of t he function f, we can pass without di ffculty t o the
concept of the ' successor' of a number: i f n i s a number, fI n) i s its
successor. It i s here that Dedeki nd' s ' ordi nal ' orientation comes into
effect: function f, vi a the medi ati on of the concept of the chai n, is
that whi ch defnes N as the space of a total order. The frst ' poi nt'
of thi s order i s obvi ousl y 1 . For phi losophi cal reasons ( compare
1 . 1 7) , Dedeki nd prefers a denotation begi nning with 1 to one begin
ni ng wi th 0; ' 1 ' denotes the frst link of a chain, whereas zero is
' cardi nal ' i n its very being: it marks l ack, the cl ass of al l empty
extensi ons.
Wi th 1 and the operation of successi on it wi l l be easy to obtai n,
frstl y, the pri mi ti ve theorems concerning the structure of the
order of numbers, and then the defnition of arithmetical operati ons,
addi ti on and mul ti pli cati on. On the sol e basi s of the concepts
of ' system' ( or set) and of ' si mi l ar transformati on' (or bi uni vocal
correspondence) , the ' natural ' ki ngdom of numeri cal ity will be
rediscovered.
4. 9. A system N, structured by a function f which compl ies with the
three condi ti ons above ( 4. 7) wi l l be cal led ' a system of numbers' , a
pl ace of the set of numbers. To ci te Dedeki nd: 7
I f, i n t he consi derati on of a si mpl y i nfni te system N, set i n order by
a transformati on t, we enti rel y neglect the speci al character of the el e
ments, si mpl y retai ni ng thei r di sti ngui shabi l i ty and taki ng i nto account
onl y the rel ati ons to one another i n whi ch they are pl aced by the order
setti ng transformati on t, then are these el ements called natura/ numbers
or ordinal numbers or si mpl y numbers, and the base-el ement 1 i s cal led
the base-number of the number-series N. Wi th reference to this freei ng
the elements from every other content ( abstracti on) , we are j usti fed i n
cal l i ng numbers a free creati on of the human mi nd.
The enthusi astic tone leaves no room for doubt: Dedeki nd i s con
sci ous of having, with hi s purel y functi onal and ordi nal engendering
of ' system' S, torn number away from any form of external j urisdic
ti on, i n the di rection of pure thought. Thi s was al ready the tone, and
these the stakes, of the ' procl amation' that appeared i n the Preface
to the frst edi ti on of hi s pamphlet: 'In speaki ng of arithmetic ( algebra,
DEDEKIND 35
anal ysi s) as a part of l ogic, I mean to i mpl y that I consi der the number
concept to be entirely i ndependent of the notions or i ntui ti ons of
space and ti me, that I consi der it more as an immediate resul t of the
l aws of thought. ' Thi s i s a text that, as wi l l be appreci ated, l ends
i tsel f to a Kanti an i nterpretati on: the whol e probl em for modern
thinkers of number is to navigate within the triangle Pl ato-Kant
Lei bni z. 8 In defning, not ' a' number, but N, the si mpl y i nfnite
'system' of numbers, Dedeki nd consi ders, with legiti mate pri de, that
he has establ ished hi mself, by means of the power of thought alone,
in the intel l i gi bl e pl ace of numerical i ty.
4. 10. I nformed by Frege' s di ffcul ti es, whi ch do not concern hi s
concept of zero and of number, but the transi ti on from concept to
existence or the j uri sdiction of l anguage over being, we ask: does a
system of numbers, a ' si mpl y i nfnite' system N, exist? Or wi l l some
unsuspected ' paradoxes' come to temper, for us, Dedeki nd' s intel
lectual enthusi asm?
4. 1 1 . Dedeki nd is evidently concerned about the exi stence of hi s
system of number. In order to establ i sh i t, he proceeds i n three
steps:
1 Intrinsic defni ti on, wi th no recourse to phi l osophy or to i ntui ti on,
of what an i nfnite system ( or set) i s.
2 Demonstration ( thi s, as we shal l see, highl y specul ati ve) of the
existence of an i nfnite system.
3 Demonstration of the fact that al l i nfnite systems 'contain as
a proper part a si mpl y i nfnite system N' .
These three steps permit the fol l owi ng concl usi on t o be drawn:
si nce at l east one i nfnite system exists, and every i nfnite system has
as a subsystem an N - a si mpl y i nfnite system or ' place of number'
- then thi s pl ace exi sts. Which is to say: number exists . The i dea that
' arithmetic shoul d be a part of logic, 9 si gni fes that, by means of the
conceptual work of pure thought al one, I can guarantee the consi s
tency of an intel l igi bl e place of numerical ity, and the effective exi s
tence of such a pl ace.
4. 12. Dedeki nd' s defni ti on of an i nfnite set is remarkabl e. He
himself was very proud of it, and with good reason. He notes that
'the defnition of the i nfnite . . . forms the core of my whole i nvestiga
tion. Al l other attempts that have come to my knowl edge to di sti n
gui sh the i nfnite from the fnite seem to me to have met with so l i ttle
success that I thi nk I may be permitted to forgo any cri ti que of
them, '
l
0
Thi s defni ti on of the i nfnite systematises a remark al ready made
by Gal i leo: there i s a bi uni vocal correspondence between the whole
numbers and the numbers that are their squares. Suffce to say, f( n)
= n
2
However, the square numbers constitute a proper part of the
whole numbers (a proper part of a set is what we call a part that i s
di fferent from the whol e, a trul y ' parti al ' part ) . It seems, therefore,
in exami ning i ntui ti vel y i nfnite sets, that there exist bi univocal cor
respondences between the sets as a whol e and one of thei r proper
parts. Thi s part, then, has 'as many' elements as the set itself. Gal i leo
concl uded that it was absurd to try to conceive of actual i nfnite sets.
Since an i nfnite set is ' as large' ( contai ns ' as many' elements ) as one
of its proper parts, the statement 'the whol e is greater than the part'
is apparently false i n the case of i nfnite total ities. Now, this statement
i s an axi om of Eucl i d' s Elements, and Gal i leo did not think it coul d
be renounced.
Dedeki nd audaci ousl y transforms thi s paradox into the defnition
of i nfnite sets: 'A system S is sai d to be infnite when it is si mi l ar to
a proper part of itsel f. In the contrar} case, S is said to be a fnite
system, '
l
l
( Remember that, i n Dedeki nd' s termi nol ogy, 'system'
means set, and the si mi l ari ty of two systems means that a bi uni vocal
correspondence exists between them) .
4. 1 3. The most stri ki ng aspect of . Dedeki nd' s defnition i s that it
determi nes i nfnity positively, and subordinates the fnite negatively.
This i s its especi al l y modern accent, such as i s al most always found
in Dedeki nd. An i nfnite system has a property of an existenti al
nature: there exists a bi uni vocal correspondence between it and one
of its proper parts. The fnite i s that for which such a property does
not obtain. The fnite is si mpl y that whi ch is not i nfnite, and al l the
posi tive si mpl ici ty of thought hi nges on the i nfnite. This intrepid
total secul ari sation of the i nfnite i s a gesture whose vi rtues we ( inept
parti sans of ' fnitude' , wherein our rel igi ous dependence can sti l l be
read) have not yet exhausted.
4. 14. The third poi nt of Dedeki nd' s approach ( that every i nfnite
system contai ns as one of its parts a system of type N, a pl ace of
number, see 4. 1 1 ) i s a perfectl y elegant proof.
Suppose that a system S is i nfni te. Then, given the defniti on of
i nfnite systems, there exists a bi uni vocal correspondence f between
S and one of its proper parts S'. In other words a bi j ective function
f that makes every element of 5 correspond to an element of 5
'
. Si nce
DEDEKIND 37
5
'
is a proper part of 5, there is at least one element of 5 that is not
in the part 5
'
( otherwi se 5 = 5
'
, and 5
'
i s not a ' proper' part ) . We
choose such an element, and call it 1 . Consider the chain of 1 for the
function f ( for ' chain' d. 4. 6) . We know that:

f is a di stinct ( i nj ective) transformati on, or functi on, si nce it is
precisely the bi univocal correspondence between 5 and 5
'
, and al l
bi uni vocal correspondence i s di stinct;

1 certai nl y does not correspond through f to any other term of
the chai n, si nce we have chosen 1 from outside of 5
'
, and f only
makes elements of S
'
correspond to elements of 5. An element s
such that f( s ) = 1 therefore cannot exist in the chai n. In the chai n,
the function never ' returns' to 1 .
The chai n o f 1 for f in 5 i s, then, a si mpl y i nfnite set N: it compl ies
with the three conditions set for such an N in 4. 7 above.
We are thereby assured that, if an infnite system 5 exists, then an
N, a pl ace of number, al so exi sts as part of that 5. Dedeki nd' s thesi s
i s ultimately as fol lows: if the infnite exists, number exists. Thi s poi nt
( taki ng account of the ordi nal defni ti on of number as the chai n of
1 for a si mi l ar transformati on, and of the defni ti on of the i nfni te)
is exactly demonstrated.
4. 1 5. But does the i nfnite exist ? There l i es the whol e questi on. Thi s
is point two of Dedeki nd' s approach, where we see that, for Dede
kind, the infnite, upon which the existence of number depends,
occupies the place which for Frege is occupied by zero.
4. 16. To construct the proof upon whi ch henceforth al l wi l l rest ( the
consi stency and the existence of an infnite system or set) , Dedeki nd
bri skl y canvasses al l hi s i ni ti al phi l osophi cal presupposi ti ons ( the
thi ng as obj ect of thought) . Of course, these presupposi ti ons al ready
quietly prop up the very idea of a 'system' ( col lection of anythi ng
whatsoever) . But, sei zed by the superbl y smooth surface of the sub
sequent defniti ons ( chai n, si mpl y i nfnite set) and proofs, we had the
time to l et thi s fragi l ity sl i p from our mi nds. We coul d do no better
than to cite here Dedeki nd' s ' proof' of what i s put forward bl ithel y
as the 'theorem' of paragraph 66: 1
2
66. Theorem: There exi st i nfni te systems.
Proof My own real m of thoughts, i . e. the total i ty S of al l thi ngs, whi ch
can be obj ects of my thought, i s i nfni te. For, i f 5 si gni fes an el ement
of S, then i s the thought 5
'
, that 5 can be an obj ect of my thought, i tsel f
an el ement of S. If we regard thi s as transform (( s) of the el ement s,
then has the transformati on ( of S, thus determi ned, the property that
the transform S' i s a part of S; and S' i s certai nl y a proper part of S,
because there are el ements i n S ( e. g. my own ego) whi ch are di fferent
from every such thought s' and therefore are not contai ned i n S'.
Fi nal l y i t i s clear that, i f Sl and S2 are di fferent el ements of S, their
transforms 51
'
and sz' are al so di fferent, that therefore the transforma
ti on ( i s a di sti nct ( si mi l ar) transformati on. Hence S i s i nfnite, which
was to be proved.
4. 1 7. Once our stupor di ssi pates ( but it is of the same order as
that which gri ps us i n reading the frst propositions of Spi noza' s
Ethics ) , we must proceed to a close exami nation of thi s proof of
exi stence.
4. 1 8. Some technical speci fcs: The force of the proof lies i n the
consi derati on of the correspondence between an ' obj ect of my thought'
and the thought 'thi s i s an obj ect of my thought' - that i s to say, the
correspondence between a thought and the thought of that thought,
or refection - as a functi on operating between elements of the set of
my possi bl e thoughts ( i n fact, we may as wel l i dentify a ' possi bl e
obj ect of my thought' wi th one possi bl e thought) . Thi s function is
' di sti nct' ( we woul d now say i nj ecti ve) , because it possesses the prop
erty ( which bi uni vocal correspondences al so possess) that two di s
tinct el ements always correspond vi a the function to two distinct
elements. Given two thoughts whose obj ects di sti ngui sh them from
each other, the two thoughts of these thoughts are di stinct ( they al so
have di sti nct obj ects, si nce they thi nk of di sti nct thoughts ) . Conse
quently there i s a bi uni vocal correspondence between thoughts in
general and thoughts of the type ' thought of a thought' . Or, i f you
prefer, there i s such a correspondence between thoughts whose obj ect
is anythi ng whatsoever and thoughts whose obj ect i s a thought. Now
thi s second set forms a proper part of the set of al l possi bl e thoughts,
si nce there are thoughts which are not thoughts of thoughts: the
stri ki ng exampl e Dedeki nd gives i s what he cal l s 'the ego' . Thus the
set of al l my possi bl e thoughts, being in bi uni vocal correspondence
with one of its proper parts, i s i nfnite.
4. 19. Dedeki nd' s approach is a singular combination of Descartes '
Cogito and the idea of the idea in Spinoza.
The starting poi nt is the very space of the Cogito, as 'closed' con
fguration of all possible thoughts, exi stenti al point of pure thought.
It i s cl ai med ( but only the Cog ito assures us of thi s) that something
like the set of all my possi bl e thoughts exists.
DEDEKIND 39
From Spinoza' s causal ' seri al i sm' ( regardless of whether or not he
fgured i n Dedeki nd' s hi storical sources ) are taken both the exi stence
of a ' paral l el i sm' which al lows us to i denti fy si mple i deas by way of
their obj ect ( Spinoza says : through the body of which the idea i s an
idea) , and the exi stence of a refexive redoubl i ng, which secures the
exi stence of 'complex' ideas, whose obj ect is no longer a body, but
another idea. For Spinoza, as for Dedeki nd, thi s process of refexive
redoubl ing must go to infnity. An idea of an idea ( or the thought of
a thought of an obj ect) is an idea. So there exi sts an idea that i s the
i dea of the i dea of the i dea of a body, and so on.
Al l of these themes have to be i n pl ace i n order for Dedeki nd to
be abl e to conclude the exi stence of an i nfnite system. There must
be a ci rcumscri bed ' pl ace' , representable under the si gn of the One,
of the set of my possi bl e thoughts. We recogni se here the soul , the
'thi nking thi ng' as paradigmatical l y establ i shed by Descartes, i n its
exi stence and essence ( pure thought) , in the Cogito. An idea must be
identi fabl e through its obj ect, so that two di fferent ideas correspond
to two di fferent obj ects: thi s al one authorises the bi uni vocal character
of the correspondence. And, ul ti mately, it must be that the refexive
process goes to infnity, since, i f it di d not, there would exi st thoughts
with no correspondent through the functi on, thoughts for which
there were no thoughts of those thoughts. Thi s would rui n the argu
ment, since it woul d no l onger be establ i shed that to every element
of the set of my possible thoughts S there corresponds an el ement of
the set of my refexive thoughts S' . Ultimatel y - above al l , I woul d
say - there must be at l east one thought that is not refexive, that is
not a thought of a thought. Thi s al one guarantees that S
'
, the set of
refexive thoughts, i s a proper part of S, the set of my possi ble
thoughts. Thi s time, we recognise in thi s fxed poi nt of di fference the
Cogito as such - what Dedeki nd cal l s ' my own ego' . That which does
not allow itself to be thought as thought of a thought is the act of
thinking itself the '[ think '. The ' I thi nk' is non-decomposable; i t is
i mpossi ble to grasp it as a thought of another thought, si nce every
other thought presupposes it.
It is therefore no exaggeration to say that for Dedeki nd, ultimately,
number exists in so far as there i s the Cogito as pure poi nt of exis
tence, underlying al l refection ( speci fcal l y, there i s an ' I thi nk that 1
think' ) , but itsel f situated outsi de of al l refecti on. The exi stenti al
foundation of the i nfnite, and therefore of number, i s what Sartre
cal l s the ' pre-refexive Cogito' .
And here we di scover a variant of Jacques-Al ai n Mi l ler' s thesi s:
what subtends number is the subject. The di fference i s that, whereas
for Mi l ler it i s the ' process of engendering' of number that requi res
the function of the subj ect, for Dedeki nd it is the existence of the
i nfnite as its pl ace. The Fregean programme of the conceptual deduc
ti on of zero and the Dedeki ndi an programme of the structural deduc
tion of the i nfnite l ead back to the same point: the subj ect, whether
as i nsi stence of lack or as pure point of exi stence. To the Lacani an
subj ect can be ascri bed the genesis of zero, to the Cartesi an subj ect:
the exi stence of the i nfnite. As if two of the three great modern chal
lenges of thi nking number ( zero, the i nfnite, the downfal l of the
One) , once the thi rd i s assumed in the gui se of a theory of sets,
can onl y be resolved through a radi cal empl oyment of that great
philosophical category of modernity: the subj ect.
4. 20. I could si mpl y say that, j ust as I am not enough of a Leibnizian
to fol l ow Frege, I am equal l y neither Cartesi an nor Spi nozist enough
to fol l ow Dedeki nd.
4. 21 . Against Dedeki nd' s Spi nozi sm: Far from the i dea of an i nfnite
recurrence of the thought of a thought of a thought of a thought of
a thought, and so on, being abl e to found the exi stence of the pl ace
of number, it presupposes it. In fact, we have no experience of thi s
type. Onl y the exi stence - and consequently the thought
-
of the
sequence of numbers al lows us to represent, and to make a numerical
fction of, a refecti on which refects itsel f endlessly. The very possi bi l
i ty of stating a ' thought' at, say, the fourth or ffh level of refection
obvi ousl y relies on an abstract knowl edge of numbers as a condi ti on.
As to the i dea of a refection that 'goes to i nfnity' , thi s obvi ousl y
contai ns preci sel y what we are tryi ng to demonstrate, namel y the
effect of i nfni ty in thought: an effect whose only known medi um is
the mathemati cs of number.
4. 22. As regards questi ons of existence, Spi noza hi msel f made certai n
not to proceed as Dedeki nd does. He never sought to i nfer the exi s
tence of the i nfnite from the recurrence of i deas. It i s, rather, precisely
because he postulated an i nfni te substance that he was abl e to estab
lish that the sequence that goes from the idea of a body to ideas of
ideas of i deas, and so on, i s i nfnite. For Spi noza, and he i s quite j usti
fed i n this, the exi stence of the i nfnite is an axiom. Hi s problem is
rather ' on the other si de' , the si de of the body ( or, in Dedeki nd' s
terms, that of the obj ect ) . For, i f there is a rigorous paral lel i sm
between the chai n of ideas and the chai n of bodies, then there must
be, corresponding to the idea of an idea, the ' body of a body' , and
we are unable to grasp what the reality of such a thi ng might be.
Dedeki nd evades this probl em because the place of thi nking he
DEDEKIND 41
postulates assumes Cartesi an cl osure: t he corporeal exteri or, the
extensive attribute, does not i ntervene i n it. But, in seeki ng to draw
from Spi nozist recurrence a concl usi ve ( and non-axi omatic) thesi s on
the infnite, he produces onl y a vi ci ous ci rcl e.
4. 23. Against Dedeki nd' s Cartesi ani sm: It is essenti al to the proof
that every thought can be the obj ect of a thought. Thi s theme i s
i ncontestabl y Cartesi an: the ' I thi nk' subtends the bei ng of i deas i n
general as a ' materi al ' of thought, and it is cl ear that there i s no idea
that cannot be a thi nkable idea, that is to say ( since we are speaki ng
of the set of my possi bl e thoughts ) vi rtual l y actual i sabl e as obj ect of
my thought. But obvi ousl y thi s excl udes the possi bi l ity that ' i t'
I 3
could thi nk wi thout my thi nki ng that I thi nk that thought, and
wi thout it bei ng even possible to do so. Dedeki nd is Cartesi an in hi s
excl usi on of the unconsci ous, whi ch, since Freud, we know to thi nk,
and to thi nk in such a way that some of i ts thoughts can be defned
precisely as those that 1 cannot thi nk. ' Unconsci ous thoughts' are
precisely those unable, at least di rectly, to become obj ects of my
thought.
More general l y, it i s doubtful , for a contemporary phi l osopher,
whether true thoughts, those that are i ncl uded in a generi c procedure
of truth, coul d ever be exposed as such in the fgure of thei r refecti on.
This woul d be to i magi ne that thei r transl ati on onto the fgure of
knowledge ( whi ch is the fgure of refecti on) is coextensive with them.
Now the most sol i d i dea of contemporary phi l osophy is preci sel y not
to understand the process of truth except as a gap i n knowledge. If
'thought' means: i nstance of the subj ect in a truth-procedure, then
there i s no thought of thi s thought, because it contai ns no knowl edge.
Dedeki nd' s approach founders on the unconsci ous, and does not hol d
frml y enough to the di stinction between knowledge and truth.
4. 24. Descartes hi msel f i s more prudent than Dedeki nd. He makes
certain not to i nfer the i nfnite from refecti on, or from the Cogito as
such. He does not consi der, i n provi ng the exi stence of God, the
totality of my possi bl e thoughts, as Dedeki nd does. On the contrary,
he singul arises an idea, the idea of God; his l ocal argument mi ght be
contrasted with Dedeki nd' s gl obal , or set-theoretical , argument. Des
cartes' problem i s el sewhere, it is a Fregean probl em: how do we pass
from concept to existence? For thi s, an argument positing a di spro
portion between the idea and its place is necessary: the idea of the
infnite i s without common measure with its pl ace, whi ch i s my soul
-
or, i n Dedeki nd' s terms, the set of my possi ble thoughts; because
this pl ace, grasped in its substanti al being, i s fnite. The si ngul ar i dea
of the i nfnite must therefore 'come from el sewhere' ; it must come
from a real i nfnity.
We can see how, in the end, Descartes' and Dedeki nd' s positions
are reversed. For Dedeki nd, it i s the pl ace that i s i nfnite, because it
must support refection ( the capacity of the Cogito) in its going to
i nfni ty. For Descartes, it is the exteri or of the pl ace ( God) that is
i nfnite, si nce the pl ace of my thought, guaranteed i n its being by the
Cogito, is fnite, and is therefore not capable of supporting alone the
idea of the i nfnite. But, i n seeking to break with the fni tude of the
place, Dedeki nd forgets that this place could well be nothing but a
scene fabricated by an Other pl ace, or that thought coul d wel l fnd
its pri nci pl e onl y i n a presupposition of i nfnite number, of which it
woul d be the fnite and i rrefexi ve moment.
4. 25. Immanent cri ti que: Dedeki nd' s starting point is 'the real m of
all possible obj ects of my thought' , which he i mmediately decides to
call system S. But can this domain be considered as a system, that is
to say, a set? Do the ' possi bl e obj ects of my thought' form a set, a
consistent multi pl icity, whi ch can be counted as one ( leaving asi de
the thorny question of knowi ng what carries out thi s accounting of
my thoughts ) ? Isn' t it rather an i nconsi stent mul ti pl icity, in so far as
its total recol lection i s, for thought itsel f, preci sel y impossi ble? If one
admits the Lacani an identifcati on of the impossible and the real ,
woul dn' t the ' system' of all possi bl e obj ects of my thoughts be the
real of thought, i n the gui se of the impossi bi l ity of its counti ng-for
one? Before establ i shing that the ' real m of all possible obj ects of my
thought' i s an i nfnite system, then, we must establ i sh that it i s a
system (a set) at al l .
4. 26. In the same way in whi ch Russel l ' s paradox comes to spoi l
Frege' s deri vati on of number on the basi s of the concept, the ' paradox'
of the set of all sets - a descendant of the former - comes to break
Dedeki nd' s deduction of the exi stence of the i nfnite, and conse
quently the deduction of the exi stence of N, the ' simply infnite' set
whi ch is the pl ace of number. Conceptual l y set out by Dedeki nd with
i mpeccable i nferences, the pl ace of number does not stand the test of
consistency, whi ch i s al so that of existence.
4. 27. Reasoni ng 'a la Dedeki nd' : Any system whatsoever (a set) ,
grasped i n abstracti on from the si ngul arity of i ts obj ects or, as Dede
ki nd says, thought uni quel y according to 'that which di sti ngui shes'
these obj ects ( thus, thei r si mpl e belonging to a system and its laws ) ,
DEDEKIND 43
is obviously a possi bl e obj ect of my thought. Consequentl y, within
the supposed system S of all possi bl e obj ects of my thought must
fgure, as a subsystem ( subset) , the system of all systems, the set of
al l sets. By vi rtue of thi s fact, thi s system of al l systems i s i tsel f a
possi bl e obj ect of my thought. Or, in si mpl ifed terms, the system of
al l systems is a thought.
Now, thi s is an impossible situati on. In fact, a fundamental prin
ci ple of Dedeki nd' s demonstration has it that every thought gives rise
to a thought of thi s thought, which i s di fferent from the ori gi nal
thought. So i f there exi sts a thought of the set of al l sets, there must
exi st a thought of thi s thought, which is in S, the set of al l my possi ble
thoughts. S is then larger than the set of al l sets, si nce it contai ns at
least one element (the thought of the set of al l sets) that does not
fgure in the set of all sets. Whi ch cannot be, si nce S is a set, and
therefore must fgure as an el ement in the set of al l sets.
Or, once agai n: considered as a set or system, S, the domai n of al l
the possi bl e obj ects of my thought, i s an element of the set of al l sets .
Considered in its seri al or refexive capacity, S overfows the set of
al l sets, si nce it contai ns the thought of that thought whi ch is the set
of al l sets. S i s thus at once i nside ( or ' smal l er than' ) and outsi de ( or
' l arger than' ) one of i ts el ements: the thought of the set of al l sets.
We must concl ude then, excl udi ng l ogi cal i nconsistency, ei ther that
the set of al l sets, the system of al l systems, is not a possible ob;ect
of my thought, even though we have j ust thought it; or, more reason
abl y, that the domai n of al l possi bl e obj ects of my thought is not a
system, or a set. But, in that case, it cannot be used to support the
proof of the exi stence of an i nfnite system.
4.28. Reasoning more mathematical l y now: Suppose that the set of
al l sets exists ( which i mpl ies necessari l y the exi stence as set of the
domai n of al l possi bl e obj ects of my thought) . Then, since it i s a set,
we can separate ( Zermel o' s axi om, 2. 1 2) , as an exi stent set, all of the
elements that have a certai n property i n common. Take the property
'not being an element of itsel f' . By means of separation this time, and
therefore wi th the guarantee of existence al ready i n pl ace, we ' cut
out' from the set of al l sets, which we suppose to exi st, the set of al l
the sets which do not bel ong to themselves. Thi s set then exi sts, whi ch
Russel l ' s paradox tel l s us i s i mpossi bl e ( admitting the existence of the
set of al l sets whi ch do not bel ong to themselves leads di rectly to a
formal contradiction, d. 2. 1 1 ) . So it is i mpossi bl e that the set of al l
sets shoul d exist, and a forti ori that the domai n of al l my possi bl e
thoughts coul d be a set.
4. 29. Dedeki nd' s attempt ul ti mately fai l s at the same point as did
Frege' s: i n the transi ti on from concept to assertion of existence. And
at the root of the affai r i s the same thing: Frege and Dedeki nd both
seek to deduce from ' pure logic' , or thought as such, not j ust the
operati onal rul es of number, but the fact of its existence for thought.
Now, j ust like the empty set, or zero, the infnite will not be deduced:
we have to decide its exi stence axi omatical l y, whi ch comes down to
admitting that one takes thi s exi stence, not for a construction of
thought, but for a fact of Being.
The site of number, whether we approach it, l i ke Frege, ' from
bel ow' , on the si de of pure l ack, or, like Dedeki nd, ' from above' ,
from the si de of i nfnity, cannot be establ i shed logical l y, by the pres
sure of thought alone upon itsel f. There has to be a pure and si mple
acknowledgement of its exi stence: the Axi om of the Empty Set founds
zero, and, as a resul t of thi s, the fni te cardi nal s exist. The Axiom of
Infnity founds the exi stence of the i nfnite ordi nal s, and from there
we can return to the exi stence of fnite ordi nal s. The chal l enges posed
to the moderns by the thi nking of number cannot be met through a
deduction, but onl y through a decision. And what subtends thi s deci
si on, as to its veri di cal i ty, relates nei ther to i ntui ti on nor to proof. It
relates to the deci si on' s conformity to that whi ch being qua being
prescri bes to us. From the fact that the One i s not, it fol lows, with
regard to zero and the i nfnite, that nothing can be sai d other than:
they are.
4. 30. Nevertheless, we must give Dedeki nd i mmense credi t for three
cruci al ideas.
The frst is that the best approach to number i s a general theory
of the pure multi pl e, and therefore a theory of sets. This approach,
an ontological one, entirely di sti ngui shes hi m from the conceptual or
l ogi ci st approach, as found in Frege.
The second i s that, wi thi n thi s framework, we must proceed in
' ordi nal ' fashi on, erecting a sort of uni versal series where number
will come to be grasped. Certai nl y, the theory of ordi nal s must be
removed from its overdependence on the idea of order, sti l l very much
present i n Dedeki nd. Because, as I obj ected to Jacques-Al ai n Mi l ler,
there is no reason to presume that the being of number will be await
ing us al ong the ordered route that we propose to it. The concept of
the ordi nal must be sti l l further ontol ogi sed, rendered less opera
ti onal , less purely serial .
The thi rd of Dedeki nd' s great i nspi red ideas i s that, to construct
a moder thi nki ng of number, a non-Greek thi nki ng, we must begin
with the i nfnite. The fact that it is vai n to try to gi ve thi s beginni ng
DEDEKIND 45
the form of a proof of exi stence i s ul ti mately a secondary matter,
compared to the idea of the begi nni ng itsel f. It i s truly paradigmatic
to have understood that, i n order to thi nk fni te number, natural
whol e number, it i s necessary frst to thi nk, and to bring i nto exi s
tence - by way of a deci si on that respects the hi stori al nature of bei ng,
in so far as our epoch i s that of the secul ari sati on of the i nfnite ( of
whi ch its numericisation i s the frst i nstance) - i nfnite number.
On these three poi nts, Dedeki nd i s trul y the cl osest compani on,
and i n certain respects the ancestor, of the father - sti l l mi sunderstood
-
of the great laws of our thought: Cantor.
5
Peano
5. 1 . Peano' s work is certai nl y not comparabl e in profundity or in
novelty either to Frege' s or to Dedeki nd' s. Hi s success l ies more i n
the cl arifcation of a symbol i sm, i n the frm assurance of the connec
tion between logic and mathemati cs, and i n a real talent for discern
ing and denoting the perti nent axi oms. One cannot speak of number
wi thout tackl i ng the famous 'Peano axi oms' at thei r source; they have
become the reference text for any ki nd of formal introduction on the
natural whole numbers.
5. 2. Even though, from the very begi nning of hi s Principles of Arith
metic,
1
- written, del i ci ousl y, in Latin - Peano speaks of ' questions
that pertai n to the foundations of mathematics' , whi ch he says have
not recei ved a ' sati sfactory sol ution' ,
2
the approach he adopts i s not
so much a fundamental meditation as a 'technici sati on' of proce
dures, wi th a view to establ i shing a sort of consensus on mani pulation
( something i n whi ch, in fact, he succeeds perfectly) . This i s the sense
in which we ought to understand the phrase: 'The di ffculty has its
main source i n the ambi guity of l anguage. '
3
To expound number in
the cl arity of a language - an artifci al cl arity, certai nl y, but legi bl e
and i ndubi tabl e - thi s i s what i s at stake in Peano' s work.
5. 3. In terms of its content, the approach is model l ed on Dedeki nd' s.
We ' start' from an i ni ti al term, whi ch, as with Dedeki nd, i s not
zero but one. We ' put to work' the successor function ( denoted in
Peano accordi ng to the addi ti ve i ntui ti on: the successor of n i s written
PEANO 47
n + 1 ) . We rel y heavi l y on i nducti on, or reasoni ng by recurrence. But,
whereas Dedeki nd, who works i n a set-theoreti cal framework, deduces
the val i dity of thi s procedure, in Peano it is treated purel y and si mpl y
as an axi om. We deci de that:
-
if 1 possesses a property,
-
and if it is true that, when n possesses a property, then n + 1 al so
possesses it,
-
then, all numbers n possess the property.
Armed with thi s inductive pri nci pl e and with purely logical axi oms
whose presentation he has cl ari fed, Peano can defne al l the cl assi cal
structures of the domai n of whol e numbers: total order and algebraic
operations ( addition, mul ti pl ication) .
5. 4. The axi om of induction, or of recurrence, marks the di fference
i n thi nki ng between Peano and Dedeki nd on the cruci al i ssue of the
infnite. Treated as a simple operati onal pri nci pl e, recurrence actual l y
permits legi sl ati on over an i nfnite total ity without making mention
of its infnity.
It is cl ear that there i s an i nfnity of whol e numbers . To speak of
' al l ' these numbers therefore means to speak of an actual i nfnity. But
in Peano' s axi omatic apparatus, this i nfnity is not i ntroduced as such.
The axiom of recurrence permits us, from a verifcation ( 1 possesses
the property) and an i mpl icative proof ( if n possesses the property,
then n + 1 also possesses it), to conclude that 'all numbers possess
the property' , without having to i nqui re as to the extensi on of thi s
' al l ' . The universal quantifer here masks the thought of an actual
infnity: the i nfnite remai ns a l atent form, i nscri bed i n the quanti fer
without being released into thought.
Thus Peano introduces the concept of number wi thout transgress
ing the ol d prohi bition on actual i nfnity, a prohi bi ti on that sti l l hangs
over our thought even as the l atter is summoned to its abol i ti on by
the modern i nj unction of being. Peano' s axi omatic evades the i nfnite,
or explicit mention of the i nfnite.
For Dedeki nd, on the other hand, not onl y the concept of the
infnite, but also its exi stence, i s absol utely cruci al . Dedeki nd says
thi s explicitly i n a letter to Keferstei n:
4
After the essenti al nature of the si mpl y i nfni te system, whose abstract
type i s the number sequence N, had been recogni zed in my anal ysi s
. . . the question arose: does such a system exist at al l i n the real m of
our i deas? Wi thout a logical proof of exi stence i t woul d always remai n
doubtful whether the noti on of such a system mi ght not perhaps
contai n i nternal contradi cti ons. Hence the need for such proofs.
5. 5. Peano does not broach questions of exi stence. When a system
of axi oms is appl ied to operati onal arrangements, we wi l l be able, if
necessary, to enqui re as to that system' s coherence; we need not
speculate on the being of that whi ch i s interrogated. The vocabul ary
of the 'thi ng' , or obj ect, common to Frege and Dedekind ( even i f it
i s a matter of ' mental thi ngs' i n the sense of Husserl ' s noematic cor
relate) i s dropped i n Peano' s work, in favour of a somewhat 'post
moder' sensi bi l ity where the sign reigns. For example, he writes: ' I
have denoted by si gns all i deas that occur i n the pri nci ples of arith
metic, so that every proposition is stated onl y by means of these
si gns.
,
5 If the l atent model in Dedeki nd and of Frege i s phi losophical
( ' phi l osophy as rigorous sci ence' ) ,6 i n Peano it is di rectly algebrai c:
' With these notati ons, every proposi ti on assumes the form and the
preci si on that equati ons have i n algebra . . . the procedures are si mi l ar
to those used i n sol vi ng equati ons. ' 7
The ' economy of number' proposed by Peano is an economy of
signs whose paradigm i s al gebrai c, whose transparency is consensual ,
and whose operati onal effectiveness i s therefore not i n doubt. He thus
partici pates forceful l y i n that movement of thought, victorious today,
that wrests mathematics from its anti que phi l osophi cal pedestal and
represents it to us as a grammar of signs where all that matters is the
making expl i ci t of the code. Peano prepares the way from afar - by
el i mi nating al l idea of a being of number, and, even more so, that of
number as being - for Carnap' s maj or theses, which reduce mathe
matics, treated as a ' formal l anguage' (as opposed to empirical lan
guages ) , not to a science ( because according to thi s conception every
science must have an ' obj ect' ) , but to the syntax of the sciences. Peano
i s i nscri bed i n the twentieth century's general movement of thought
- forged, i n fact, at the end of the ni neteenth century
-
whose char
acteristic gesture is the destitution of Platoni sm, in the guise of that
which had always been its basti on: mathematics, and especi al l y the
Idea of number.
5. 6. We see here, as i f i n the pangs of its bi rth, the real ori gi n of
what Lyotard calls the ' l i nguistic turn' i n Western phi losophy, and
what I cal l the rei gn of the great modern sophi stry: i f it is true that
mathematics, the highest expression of pure thought, i n the fnal
anal ysi s consists of nothing but syntactical apparatuses, grammars of
signs, then a forti ori all thought fal l s under the constitutive rule of
l anguage.
PEANO 49
It is certai n that, for Plato, the subordi nation of l anguage to ' thi ngs
themselves' , as deal t wi th for exampl e i n the Cratylus, has as its
horizon of certitude the ontol ogi cal vocati on of the matheme. There
is no uphol di ng the pure empi re of the sign i f number, which we
i ndicate with j ust a si mpl e stroke, is, as Pl ato thought, a form of
Bei ng. Conversel y, if number i s nothi ng but a grammar of speci al
signs, ruled by axi oms wi th no foundati on in thought, then it becomes
probable that phi losophy must be, frst and foremost (as i n Deleuze's
reading of Ni etzsche' s ' di agnostics' ) , a thi nking of the force of si gns.
Either truth or the arbitrariness of the sign and the diversity of syn
tactical games: thi s is the central choice for contemporary phi losophy.
Number occupies a strategic posi ti on i n t hi s confict, because i t i s
si multaneously the most general ised basi s of thought and that whi ch
demands most abruptly the question of its bei ng.
Peano' s axi omatic, poor in thought but strong i n its effects, a
grammar whi ch subdues number, the organi si ng pri nci pl e of an oper
ational consensus, a deft mediation of the i nfnite into the fni tude of
signs, represents something of a l ucky fnd, a gi ft, for modern
sophistry.
5. 7. Every purel y axi omatic procedure introduces undefned signs,
which can onl y be presented by codifng thei r usage in axi oms. Peano
is hardl y economi cal with these ' pri mi ti ve' si gns: there are four, i n
fact ( you are remi nded that set theory has recourse to one singl e
primitive sign E , bel ongi ng, whi ch denotes presentation as such) :
Among the si gns of ari thmeti c, those that can be expressed by other
signs of ari thmetic together wi th the si gns of logic represent the i deas
that we can defne. Thus, I have defned all signs except four . . . I f, as
I thi nk, these cannot be reduced any further, i t i s not possi bl e to defne
the i deas expressed by them through i deas assumed to be known
previ ousl y. 8
These four i rreduci bl e si gns are: 9
1 The sign N, whi ch 'means number (positive whole number} ' .
2 The sign 1 , whi ch ' means unity' .
3 The sign a + 1 , whi ch ' means the successor of a' .
4 The sign whi ch ' means is equal to' .
Peano thus explicitly renounces al l defnition of number, of suc
cession, and of 1 . (The case of the sign mi ght be treated separatel y:
i t i s in poi nt of fact a logical si gn, not an arithmetical one. Peano
hi msel f writes: 'We consi der thi s sign as new, although it has the form
of a sign of 10gic. ' )
J
O Evi dently thi s i s the price to be pai d for opera
tional transparency. Where Frege musters all thought towards
attempting to understand the revol uti onary statement 'zero i s a
number' , Peano si mpl y notes ( it is the frst axi om of hi s system) : 1 E
N, a formal correl ati on between two undefned signs that ' means'
( but according to what doctrine of meani ng? ) that 1 i s a number.
Where Dedeki nd generates the pl ace of number as the space of pos
si ble empl oyment, or the real l y existing i nfnite chai n, of a bi univocal
functi on, Peano notes:
l l
a E N a + 1 E N, an i mpl ication that
involves three undefned signs, and which 'means' that, i f a i s a
number, its successor is al so a number. The force of the letter is here
at the mercy of meani ng. And the effect is not one of obscurity,
but rather one of an excessi ve l i mpi dity, a cumbersome levity of
the trace.
5. 8. In the poem, the obscure is born of that which, as a breaki ng
open of the si gni fer at the l i mits of l anguage, di ssemi nates the letter.
In Peano' s pure axi omatic, the retreat of sense i ssues from the fact
that the force of the letter is turned back upon itself, and that it i s
onl y from outside that thought can come to it. Peano wi shes to put
off any confrontation with the l atent poem the absence of whi ch
number - astral fgure of bei ng ( 'col d with neglect and di suse, a
Constel l ati on' ) 1 2 - unfai l i ngl y instigates and the effect of which Frege
and Dedeki nd unconsci ousl y preserve in the desperate attempt to
conj ure forth into Presence now zero, now the i nfnite.
5. 9. Peano' s axiomatic is a shi ni ng success story of the tendency of
our ti mes to see nothing in number except for a network of opera
tions, a mani pul able logic of the sign. Number, Peano thinks, makes
signs about the sign, or is the Sign of si gns.
From thi s poi nt of view, Peano i s as one with the idea that the
uni verse of science reaches its apex i n the forgetti ng of being, homog
enous with the reabsorption of numerical ity i nto the unthought of
techni cal wi l l . Number i s trul y machinic. Thus it can be mai ntained
that the success of Peano' s axi omatic participates i n the great move
ment that has given up the matheme to modern sophistry, by unbi nd
i ng it from al l ontology and by s ituating it wi thi n the sol e resources
of l anguage.
5. 10. It wi l l be a great revenge upon thi s operation to di scover, with
Skolem and then Robi nson, 1
3
the semantic limits of the grammar of
si gns to whi ch Peano had reduced the concept of number. We know
PEANO 51
today that such an axi omatic admits of ' non-standard' model s, whose
proper being di ffers greatly from all that we i ntui ti vel y understand
by the idea of natural whole number. So that Peano' s system admits
of models where there exist ' i nfnitely l arge' numbers, or models
whose type of i nfnity exceeds the denumerabl e. Peano arithmetic i s
suscepti ble to ' pathological ' interpretati ons; it does not have the
power to establ i sh a uni vocal thought wi thi n the machi ni sm of si gns.
Every attempt to reduce the matheme to the sol e spati a I i sed evidence
of a syntax of signs runs aground on the obscure prodigal ity of being
in the forms of the mul ti pl e.
5 . 1 1 . The essence of number wi l l not be spoken, ei ther as si mple
force of counting and of its rules, or as sovereignty of graphi sms. We
must pass into it through a meditation on its bei ng.
N is not an ' undefned' predicate, but the i nfnite place of exercise
of that which succeeds the voi d (or zero) , the exi stenti al seal whi ch
stri kes there where i t
1
4
i nsi sts on succeedi ng.
What ' begi ns' is not the 1 as opaque sign of ' uni ty' , but zero as
suture of al l l anguage to the being of the si tuati on whose l anguage
it i s.
Successi on i s not the addi ti ve codi ng of a + 1 , but a si ngul ar di s
position of certain numbers which are successors rather than thei r
succeeding, and whi ch are marked i n thei r bei ng by thi s di spositi on.
We must know al so that zero and the i nfni te are precisely that which
does not succeed, and that they are so i n thei r very bei ng, i n di fferent
ways; although both are located, by vi rtue of this fact, on the shores
of a Nothingness.
Number is neither that which counts, nor that wi th which we
count. This regime of numerical i ty organises the forgetting of number.
To think number requires an overturni ng: it is because it i s an unfath
omable form of being that number prescri bes to us that feeble form
of its approximation that i s counti ng. Peano presents the i nscri ption
of number, whi ch i s our i nfrmity, our fnitude, as the condi ti on of
i ts being. But there are more things, i nfnitely more, i n the kingdom
of Number, than are dreamt of i n Peano' s arithmeti c.
6
Cantor: 'Wel l -Orderedness'
and the Ordinals
6. 1 . The ordi nal s represent the general ontological horizon of numer
i cal i ry. Fol l owi ng the el uci dati on of the concept of the ordi nal , with
which we shal l presently occupy oursel ves, this pri nci ple will govern
everythi ng that fol lows, and it i s wel l sai d that in thi s sense Cantor
is the true founder of the contemporary thi nking of number. In fact,
Cantor
l
consi dered that the theory of ordi nal s constituted the very
heart of hi s discovery. Today, the working mathematician, for whom
it suffces that there are sets and numbers and who does not worry
at all about what they are, thi nks of the ordi nal s rather as something
of a curi osi ty. We must see i n this mild disdain one of the forms of
submi ssi on of the mathematici an, i n so far as he or she is excl usively
working, to the i mperatives of social numerical ity. Speci al ists i n
mathemati cal l ogi c or set theory are doubtless an excepti on, even i f
they themsel ves often regret thi s excepti on: i n spite of themselves,
they are cl osest to the i nj unction of Bei ng, and for them the ordi nal s
are essenti al .
6. 2. I have sai d, i n connection with Dedeki nd, that, i n our present
phi l osophi cal di scourse, we must assume as complete an 'ontologisa
tion' of the ordi nal s as possi bl e. The presentation of this concept
by Dedeki nd or Cantor relates it essenti al l y to the notion of weI l
orderedness - something sti l l very cl ose to a si mpl e seri al or
operati onal i ntui ti on of number.
6. 3. Every school boy knows that, given two di fferent whole numbers,
one of them is larger and the other smal ler. And he knows al so that,
CANTOR: 'WELL-ORDERED NESS' AND THE ORDINALS 53
given a ' bunch' of numbers, there is one and one onl y that is the
smal lest of the bunch.
From thi s seri al knowledge, if one abstracts out its general proper
ties, the concept of the wel l -ordered set can be devel oped.
6. 4. A ' wel l -ordered' set i s a set for whi ch:

between the elements of the set, there is a relation of total order;
given two el ements, e and e
'
, if < denotes the order-rel ati on, then
either e < e
'
, e
'
< e, or e e
'
; no two el ements are ' non-comparabl e'
by thi s relation;

given any non-empty part of the set so ordered, there is a smallest
element of thi s part ( an element of thi s part that is smal l er than
al l the others ) . If P is the part considered, there exists p, whi ch
belongs to P and for whi ch, for every other p
'
belonging to P, P
< p
'
Thi s element p wi l l be cal led the mi ni mal element of P.
If an element p i s mi ni mal for a part P, it alone possesses that
property. For, if there were another, a p
'
di fferent from p, then, because
the order is total , either p < p
'
and p
'
woul d not be mi ni mal , or p
'
<
p and p would not be mi ni mal . So we can speak without hesitation
of the ' mi ni mal element' of a part P of a wel l -ordered set.
We can see that the general concept of the well -ordered set is
merely a sort of extrapol ati on from what the schoolboy observes i n
the most fami l i ar numbers: the natural whol e numbers.
6. 5. A good i mage of a wel l -ordered set i s as fol l ows. Let E be such
a set. ' Start' with the smal lest el ement of E, whi ch, gi ven condi ti on
2 above, must exist. Cal l this element 1 . Consi der the part of E
obtained by removing 1 , the part ( E - 1 ) . It too has a mi ni mal
element, which comes in a certai n sense straight after 1 . Cal l thi s
element 2. Consider the part of E obtained by removing 1 and 2 to
be the part ( E - ( 1 , 2) ) . It has a mi ni mal element, cal l it 3, and so on.
A wel l -ordered set presents itsel f l i ke a chai n, so that every l i nk of
the chai n follows ( ' fol l ows' meani ng: comes j ust after in the relation
of total order) onl y one other, wel l determi ned ( it i s the mi ni mal
element of what remains) .
6. 6. Cantor' s stroke of geni us was t o refuse t o l i mi t thi s i mage to
the fnite, and thereby to introduce infnite numerati ons. He had the
fol l owing idea: If I suppose the exi stence - beyond that sequence
1 ,2, 3, . . . , n, n + 1 , . . . - of a whole number which i s the ' frst' wel l
ordered set, the matrix of al l others, an ' i nfnite ordi nal number'
and decl are it l arger than al l the numbers that precede it, then what
prevents me from continuing? I can very wel l treat 0 as the mi ni mal
element of a wel l -ordered set that comes i n some sense afer the set
of al l the whol e numbers. And I can then consi der the ' numbers'
0 + 1, 0 + 2, . . . , 0 + n, . . . , etc. I will arrive eventual l y at 0 + 0,
and will continue once agai n. No stoppi ng poi nt i s prescribed to me,
so that I have a sort of total series, each term of which is the possible
measure of every exi stent sequence. Thi s term indicates to me that,
however many came before it, it numbers every series of the
same length.
6. 7. Al l ow me to cal l ordinal the measure of the length of a wel l
ordered set, from its mi ni mal el ement to its 'end' . The ' enti re' sequence
of ordi nal s woul d then provi de us with a scale of measurement for
such lengths. Each ordi nal woul d represent a possi bl e structure of
wel l -orderedness, determi ned by the way in which the el ements
succeed each other, and by the total number of these elements. This
i s why we say that an ordinal, whether fnite (the ordi nal s which
come before 0, and whi ch are qui te si mply the natural whol e numbers )
or i nfnite ( those ordi nal s which come after 0) , numbers a 'type of
wel l -orderedness' .
6. 8. To give a technical groundi ng for thi s idea, we wi l l consider the
cl ass of wel l -ordered sets that are i somorphi c to one of the sets among
them ( and therefore i somorphi c to each other) . What shoul d we
understand by thi s?
Take two wel l -ordered sets, E and E', < the order-relation of E,
and <
the order-rel ati on of E

'
. I wi l l say that E and E' are isomorphic
i f there exi sts a bi univocal correspondence ( ( d. 4. 5 ) between E and
E
'
, s uch that, when el < e2, i n E, then (( el l <
(( e2 ) in E'
.
We can see that ( proj ects the order of E i nto the order of E', and,
what' s more, si nce ( is bi univocal , there are ' as many' elements i n E
'
as in E. We can therefore say that E and E', consi dered strictly from
the poi nt of view of thei r wel l -orderedness, and abstracted from the
si ngul arity of thei r elements, are identical : the ' morphi sm' ( form) of
thei r wel l -orderedness is ' i so' ( the same) , as the correspondence (
assures us.
In fact, each cl ass of wel l -ordered sets i somorphic to each other
represents a wel l -orderedness, that wel l -orderedness common to al l
sets of that cl ass. It i s this wel l -orderedness that can be represented
by an ordi nal .
Thus an ordi nal is the mark of a possi ble fgure ( a form, a
morphi sm) of wel l -orderedness, i somorphic to al l the sets that
CANTOR: 'WELL-ORDEREDNESS' AND THE ORDINALS 55
take that form. An ordi nal i s the number or the cipher of a
well-orderedness.
6. 9. Thi s concepti on, al ready moving strongly in the di rection of
determi ni ng a horizon of being for al l number i n the form of a uni
versal scal e of measurement for forms of wel l -orderedness, neverthe
l ess presents some seri ous di ffculti es; the frst among them techni cal ,
the remainder phi l osophical .
6. 10. The techni cal di ffculties are three i n number, three questi ons
whi ch must be answered:
1 Which is the frst term in the total series of ordi nal s, the i ni ti al
l i nk that ' anchors' the whol e chai n? Thi s is the conceptual ques
ti on of zero or the empty set, whi ch alone is able to number
sequences of no length, sequences with no elements, the well
orderedness that orders nothing. This i s the question that caught
out Frege.
2 What exactly is the procedure of thought that al l ows us to suppose
a beyond of the sequence of fnite whole numbers ? What is the
gesture by whi ch we pass beyond the fnite, and decl are the
frst ordinal whi ch will not be a natural whole number, the frst
mark of a wel l -orderedness that descri bes the structure of a non
fnite set? Thi s i s the exi stenti al question of the i nfnite, upon
which Dedeki nd foundered.
3 Does the universal series of ordi nal s - the scale of measurement
of al l length, whether fnite or i nfnite, the total ity of specifcations
of wel l -orderedness - exist i n the set-theoretical framework? Isn' t
it - l i ke the ' system of al l the possi bl e obj ects of my thought'
introduced by Dedeki nd - an inconsistent total i ty, one that
thought cannot take as one of its possi bl e obj ects ? Thi s i s the
question of counting for one an ' absol ute' total ity. It is thus
the problem of the defection of the One as soon as we cl ai m to
' count' the universe of di scourse.
And so, once agai n, we fnd ourselves returned to the three chal
lenges of the modern thi nking of number: zero, the i nfnite and the
non-being of the One.
6. 1 1 . It rapidly turns out that the thi rd probl em admits of no positive
sol ution. Somethi ng that was at one time put forward as a ' paradox' ,
the Bural i -Forti paradox, can actual l y be proved: the ordi nal s do not
form a set, they cannot be col l ected i n a multi ple that can be counted
for one. The i dea of ' al l ' the ordi nal s is inconsistent, i mpossi bl e; it
i s, to thi s extent, the real of the horizon of the being of number.
This proof i s very cl osely rel ated to that which refutes Dedeki nd' s
attempt to prove the exi stence of an i nfnite set ( compare 4. 28) : the
set of ' al l ' the ordi nal s must itsel f be an ordi nal , and thus it would
be inside itself ( since i t is a set of all the ordi nal s) and outside itself
( since it i s not counted i n the sequence it total i ses ) . We are therefore
prohi bited to speak of a ' set of ordinal s' with no further qual i fcation.
Whi ch is preci sel y to say: ' bei ng an ordi nal ' is a property with no
extension. It i s possi bl e to confrm that a certai n obj ect is an ordinal
( possesses the property) , but not to count for one all the obj ects that
have this property.
6. 12. I have sai d enough, in my cri ti que of Frege and Dedeki nd, for
the treatment of probl ems 1 and 2 ( 6. 10) to be antici pated: the exis
tence of zero, or the empty set, and the existence of an i nfnite set
can in no way be deduced from ' purely logical ' presupposi ti ons. They
are axi omatic deci si ons, taken under the constraints of the hi storial
i nj unction of bei ng. The worl d of modern thought is nothing other
than the effect of this i nj uncti on. Begi nni ng in the Renaissance, by
way of a rupture with the Greek cosmos,
2
it became necessary,
i n order to be able to thi nk at al l i n accordance with our pre
understanding of ontological exigency, to assume:

that the proper mode under whi ch every situation ' that i s' is
sutured to its being i s not Presence, the dehiscence of that which
pro-poses itself wi thi n its l i mi ts, but pure subtracti on, the unqual i
fabl e voi d. In that form of being which i s number, thi s can be
stated as fol l ows: 'zero exists' , or, in a style more homogenous
with Cantor's ontological creation: ' a set exi sts which has no
el ements' ;

that, in thei r quasi -total i ty, and by way of a rupture with the
medi aeval tradition which reserves this attri bute for God alone,
si tuati on-beings are i nfnite; so that, far from being a predicate
whose force i s that of the sacred, the i nfnite i s a banal determi na
ti on of bei ng, such as it proffers i tsel f as pure mul ti pl icity under
the l aw of a count-for-one. In that form of being which is number,
this can be stated as fol l ows: 'an i nfnite set exists'; or, more
techni cal l y: ' an ordi nal exi sts whi ch is not a natural whole
number' . Or, i n other words, O r exi sts' .
6. 1 3. One had to wai t practi cal l y unti l the begi nni ng of the twentieth
century before these deci si ons rel ati ng to zero and to the i nfnite
CANTOR: 'WELL-ORDEREDNESS' AND THE ORDINALS 57
would be recognised i n themselves ( under the names of the Axi om
of the Empty Set and the Axiom of I nfni ty) , al though they had been
operative i n thought for three hundred years. But thi s i s not surpri s
ing. We can observe a veritabl e philosophical desperation constantly
putting these i mperatives i nto reverse, whether through the intellec
tual derel iction of the theme of fnitude or through nostal gi a for the
Greek ground of Presence. It is true that, when we are deal i ng with
pure decl arations, decided i n themselves, these decl arati ons exhi bit
the fragil ity of thei r hi storicity. No argument can support them.
What' s more, certai n truth procedures, in particul ar pol itics, art and
love, are not yet capable of sustaining such axi oms, and so in many
ways remain Greek. They cl i ng to Presence ( art and love) , conti nual l y
recusing the statement 'zero i s the proper numeric name of being' i n
order to give tri bute to the obsolete rights of the One. Or ( pol i tics )
they manage fnitude, corrodi ng day after day the statement 'the
situation is i nfni te' , i n order to val ori se the corrupted authority
of practicalities.
6. 14. The two axi oms of the voi d and of the i nfnite structure the
enti re thi nki ng of number. The pure void i s that which supports there
being number, and the i nfnite, that by whi ch it is affrmed that
number is the measure of the thi nking of every situati on. The fact
that thi s is a matter of axi oms and not of theorems means that the
exi stence of zero and of the i nfnite i s prescri bed to thought by bei ng,
in order that thought might exist in the ontological epoch of such an
existence.
In thi s sense, the current strength of reactive, archai c and rel i gi ous
wi l l s is necessari l y accompanied by an i rremedi abl e opacity of number
-
whi ch, not ceasi ng to rule over us, since thi s i s the epochal l aw of
being, nevertheless becomes unthi nkable for us. Number may exist
as form of being but, as a resul t of the total secul ari sati on of the voi d
and of the i nfnite, thought can no l onger exist i n the form and wi th
the force that the epoch prescribes to it. So number wi l l now mani fest
itsel f, without l i mit, as tyranny.
6. 1 5. The pri nci pal phi losophi cal di ffculty of the Cantori an concept
of the ordi nal s i s as fol l ows. In the presentati ons which bind it to
the concept of wel l -orderedness, the theory of ordinal s rather
seems to 'general i se' the intuition of natural whole number that
al lows us to think the being of number. It draws its authority
from that which it cl ai ms to el ucidate. The i dea of wel l -orderedness
in effect does not so much found the concept of number as deduce
it from the l acunary and fnite experience of numerical immediacy,
whi ch I i ncarnated ( i n 6. 3) in the sympathetic fgure of the
school boy.
I f we trul y wi sh to establ ish the bei ng of number as the form of
the pure mul ti pl e, to remove it from the schoolroom ( which means
al so to subtract the concept from its ambient numeri cal ity) , we must
di stance ourselves from operati onal and seri al mani pulati ons. These
mani pul ati ons, so tangi bl e in Peano, proj ect onto the screen of modern
i nfnity the quasi -sensi bl e i mage of our domestic numbers, the 1 , fol
l owed by 2, which precedes 3, and then the rest. The establ i shing of
the correct distance between thought and countable mani pul ati ons i s
preci sel y what I cal l the ontol ogi sati on of the concept of number.
From the poi nt at which we presentl y fnd ourselves, it takes on
the form of a most precise task: the ontologisation of the ' universal '
series of t he ordi nal s. To proceed, we must abandon the idea of
wel l -orderedness and thi nk ordinati on, ordi nal ity, i n an intrinsic
fashi on.
It i s not as a measure of order, nor of di sorder, that the concept
of number presents i tsel f to thought. We demand an i mmanent deter
mi nation of its being. And so for us the question now formulates
itsel f as fol l ows: which predicate of the pure mul ti pl e, that can be
grasped outside of al l serial engenderment, founds numerical ity? We
do not want to count; we want to thi nk the count.
2
Concepts: Natural Multiplicities
7
Transitive Multiplicities
7. 1 . What permits the abandonment of every primitive bond between
number and order or seri al ity is the concept of the transitive set.
Only this structural - and essenti al l y ontol ogi cal - operator enables
an i ntri nsi c determi nati on of number as a fgure of natural bei ng. In
vi rtue of it, we are no longer trapped i n the quandaries of the deduc
tion of the concept ( Frege) , of the subj ect as causal ity of l ack i n seri al
engenderment ( Mi l l er) , of the exi stence of the i nfnite ( Dedeki nd) , or
of the 'school boy' intuition of wel l -orderedness ( Cantor) .
7.2. Although thi s concept might seem at frst gl ance rather mysteri
ous, i ts lack of rel ati on to any intuitive i dea of number is to my eyes
a great vi rtue. It proves that i n it we grasp something that breaks the
ci rcle of an ontological elucidation of number enti rely transparent in
its pure and si mpl e presupposition. We have seen that thi s ci rcl e
recurs in Frege and i n Dedeki nd, and that the Cantori an conception
of ordinals as types of wel l -orderedness i s sti l l compl iant wi th it. But
we shal l see that the legitimacy for philosophical thought of the
concept of transitivity l eaves no room for doubt.
7. 3. To understand what a transitive set i s, it i s essenti al to penetrate
the distinction - of which it woul d not be an exaggeration to say that
it supports al l post-Cantori an mathematics - between an element' s
belonging to a set and the inclusion of a part. Thi s distinction is
rudimentary, but it i mpl ies such profound consequences that for a
long time it remained obscure.
62 CONCEPTS: NATURAL MULTIPLICITIES
7. 4. A set is ' made out of el ements' , is the ' col lecti on' ( i n my l an
guage, the count-for-one) of its elements.
Take the set E, and let e be one of the elements from which it
' makes' a set: we denote thi s by e E E, and we say that e belongs to
E, E bei ng the sign for belonging.
I f you now 'gather together" many el ements of E, they form a part
of E. Taki ng E' as the set of these elements, E' is a part of E. This is
denoted by E' C E, and we say that E' i s i ncl uded in E, C being the
sign of i ncl usi on.
Every el ement of a part E' of E is an element of E. I n fact thi s is
the defni ti on of a part: E' i s included i n E when al l the el ements that
belong to E
'
also belong to E. So we see that inclusion i s defned i n
terms of bel onging, which is the onl y ' pri mitive' si gn of set theory.
The cl assi c ( mi sl eadi ng) i mage is drawn l i ke thi s:
I n it we can see that E
'
i s a part of E, that e, is at once ( as is every
element of E
'
) an element of E' and an el ement of E, and that
e2 i s an element of 'the whol e' E, but not of the part E' . We al so say
that e2 bel ongs to the difference of E and E', which is denoted by
E-E
'
.
7. 5. Is it possi ble for an element that belongs to set E also to be a
part of that set, al so to be included? Thi s seems total l y bizarre, above
al l if we refer to the i mage above. But thi s sentiment mi sses the most
i mportant point, whi ch is that an element of a set can obvi ousl y be
itsel f a set ( and even that thi s is al ways the case) . Consequently, if e
belongs to E, and e is a set, the question arises whether an element
of e i s or is not, in its turn, an element of E. If al l the elements of e
are al so el ements of E, then e, whi ch is an element of E, is al so a part
of E. It bel ongs to E and is i ncl uded in E.
7. 6. Suppose for exampl e that V i s the set of l i ving bei ngs. My cat
belongs to thi s set. But a cat is composed of cel l s, whi ch one might
say are themselves all l i ving beings. So my cat i s at once a l i vi ng being
TRANSITIVE MULTIPLICITIES 63
and a set of l i ving beings. He belongs to V ( qua one, this l i vi ng cat) ,
and he is a part of V - he is incl uded i n V ( qua group of l i ving
cel l s) .
7. 7. Forget cats. Consider the three fol l owi ng ' obj ects' :
- the obj ect e] ;
- the obj ect ez ;
- the obj ect which is the 'gathering together' of the frst two, and
which we denote by ( e] ,e2 ) ' This i s cal led the pair of e] and ez .
Form a set from these three obj ects. In the same way, we denote
it by: ( e] ,e2, ( e] ,e2 ) ) ' Thi s is cal led the triplet of e] and e2 and the pai r
( e] ,ez ) . We wi l l denote it by T. Note that the three elements that
belong to this triplet are e] , e2, and ( e] ,ez ) .
Si nce e] and ez belong t o T, i f I 'gather them together' , I obtai n a
part of T. Thus, the pai r ( e] ,ez ) , which is the 'gathering together' of
these two elements of T, i s included i n T. But i n addi ti on we can see
that it i s an element of it, that it al so belongs to it. Thus we have
constructed a very si mpl e case of a set of whi ch an element i s al so a
part. In set T, the pai r ( e] ,ez ) is si multaneousl y in a posi ti on of bel ong
ing and of i ncl usi on.
7. 8. We know, from a famous theorem of Cantor' s, that there are
more parts than elements in any set E whatsoever. Thi s is what I cal l
the excess of i ncl usi on over belonging, a l aw of bei ng qua being
whose consequences for thought are immense, since it affects the
fundamental categories that i nform the couplets One/Mul ti ple and
WholelPart. It is therefore i mpossi bl e that every part shoul d be an
element, that everything that is i ncluded shoul d al so bel ong: there are
always parts that are not elements.
But we can put the question from the other di recti on: si nce we can
see that it is possi bl e in cenai n cases ( for exampl e my cat for the set
V of l i vi ng beings, or the pair ( e] ,ez ) for our tri pl et T) for an element
to be a part, is it possi bl e for all elements to be parts, for everythi ng
that belongs to the set to be i ncl uded? Thi s i s not the case for T: the
element e] taken al one, for exampl e, i s not a part of T.
Can we produce a non-empi ri cal exampl e ( because my V, my cat
and its cel l s are rational l y suspect) of a set al l of whose el ements
would be parts ?
7. 9. Let's retrace our steps a l ittle, back to the empty set. We have
proposed ( i n 2. 1 8) the axi om ' a set exi sts whi ch has no el ements' ,
that i s, a set to whi ch nothing belongs . We are going to give to this
set, the ' empty' rock of the whole edifce of multiple-being, a proper
name, the name ' 0' .
The fol l owi ng, extremely subtle, remark must be made: the empty
set is a part of every set; 0 is incl uded in E whatever E might be. Why?
Because, if a set F is not a part of E, it is because there are elements
of F that are not elements of E ( i f every element of F is an element of
E, then by defnition F is a part of E) . Now 0 has no elements. So, it
i s i mpossi bl e for it not to be a part of E. The empty set is ' universal l y'
incl uded, because nothing i n it can prevent or deny such i ncl usi on.
To put it another way: to demonstrate that F is not a part of E
requi res that we pi ck out, within F, at least one element: that element
whi ch, not being an element of E, proves that F cannot be i ncl uded
' enti rel y' wi thi n E. Now the void does not tolerate any di fferentiation
of this sort. It i s i n-di fferent, and, because of this, it i s i ncl uded in
every multi pl i ci ty.
7. 10. Consider the two fol l owing ' obj ects ' :
- the empty set, 0;
- the set whose one and onl y element is the empty set, which is
cal led the singleton of the empty set, and i s denoted by ( 0) .
Note wel l that thi s second obj ect i s different from the empty set
itsel f. In fact, the empty set has no elements, whereas the si ngleton
has one element - preci sel y, the empty set. The si ngleton of the
void ' counts for one' the void, whereas the empty set does not count
anything ( thi s i ndicates a subtle distinction between 'does not
count anythi ng' , whi ch is what 0 does, and ' counts nothing' , which
is what ( 0) does. Pl ato al ready played on thi s di sti nction i n the
Parmenides ) .
7. 1 1 . An addi ti onal remark as regards si ngl etons ( si ngletons
' i n general ' , not the parti cul ar singleton of the empty set) : Take a set
E and one of its elements e ( so e E E) . The si ngl eton of e, written ( e) ,
is a part of E: ( e) C E.
What i s the si ngleton of e, i n fact ? It i s the set whose uni que
element is e. Consequently, if e is an element of E, ' al l ' the elements
of the si ngleton (e) - namel y the single element e - are elements of
E, and so (e) i s i ncl uded i n E.
7. 12. ' Gather together' our two ohj ects, the empty set denoted by 0
and the si ngleton of the empty set, denoted by ( 0) . We obtai n the
TRANSITIVE MULTIPLICITIES 65
pai r ( 0, ( 0) ) , whi ch we wi l l denote by D. Thi s time, the two elements
of the pair D are also parts; everything that belongs to D is al so
included in D. In fact, the frst element, 0, the empty set, i s i ncl uded
in any set whatsoever ( see 7. 9) . Speci fcall y, it is a part of the pai r
D. But, what' s more, si nce 0 is an element of D, i ts si ngleton ( 0) , i s
a part of D ( 7. 1 1 ) . But ( 0) i s preci sel y the second el ement of D. Thus
thi s element i s al so included i n D. The set D i s such that every element
of it i s also a part; everything that belongs to D is i ncl uded in D.
7. 1 3. As predicted by Cantor' s theorem, there are parts of D that are
not elements of D. For exampl e, the si ngleton of the element ( 0) of
D is a part of D, as is every si ngleton of an element ( 7. 1 1 ) . We can
write this ' si ngleton of the si ngleton' as ( ( 0) ) . Now, thi s obj ect is not
one of the two elements of D.
7. 14. An i mportant defnition: we say that a set T is transitive i f it
i s l i ke the set D that we have j ust bui lt: i f all of its el ements are al so
parts, i f everything that belongs to it is al so i ncl uded i n i t, i f, wherever
it i s the case that t E T, it i s also the case that t C T.
7. 1 5. Transitive sets exist, wi thout a doubt. Perhaps V, the set of
l i ving beings; certai nl y the set ( 0, ( 0) ) , whi ch is transparent, transl u
cent even, constructed as i t is from the voi d ( the pai r of the voi d and
the singleton of the voi d, the voi d as such and the voi d as one) .
7. 16. Modernity i s defned by the fact that the One i s not ( Nietzsche
said that ' God i s dead' , but for hi m the One of Li fe took the pl ace
of the deceased) . So, for we moderns ( or ' free spi rits' ) , the Multi pl e
without-One i s the last word on being qua being. Now the thought
of the pure mul ti pl e, of the mul ti pl e consi dered i n itsel f, without
consideration of what it i s the multiple of (so: without consi deration
of any object whatsoever) , is cal l ed: ' mathematical set theory' . There
fore every maj or concept of thi s theory can be understood as a
concept of modern ontol ogy.
What does ontology di scover in the concept of the transitive
set?
7. 1 7. Belongi ng is an ontological function of presentation, i ndicating
that which is presented in the count-for-one of a mul ti pl e. Incl usi on
i s the ontological function of representation, i ndicating multi pl es re
counted as parts in the framework of a representati on. A most i mpor
tant problem (the problem of the state of a situation) i s determi ned
by the relation between presentation and representati on.
Now, a transitive set represents the maximum possible equi l i brium
between belonging and i ncl usi on, the element and the part, E and C.
Transiti vity thus expresses the superi or type of ontological stabi li ty;
the strongest correl ati on between presentation and representati on.
There i s always an excess of parts over el ements ( Cantor's theorem)
,
there al ways exist parts of a set whi ch are not elements of that set.
Thus we obtai n the maxi mal correspondence between belonging and
i ncl usi on precisely when every element is a part: when the set con
si dered i s transi ti ve.
Thi s strong i nternal frame of the transi ti ve set (the fact that every
thi ng that it presents i n the multi ple that it is, it represents a second
time i n the form of i nclusi on ) , this equi l i bri um, this maxi mal stabi l ity,
has led me to say that transitive sets are ' normal ' , taking ' normal ' in
the doubl e sense of non-pathological , stable, strongly equi l i briated,
that is to say: not exposed to the disequi l i bri um between presentation
and representati on, a di sequi l i bri um whose effective form i s the
evental caesura; and submitted to a norm, that of a maxi mal l y
extended correspondence between the two maj or categories of
ontological i mmanence: belonging and i ncl usi on.
7. 1 8. The concept of transitive mul ti pl icity wi l l constitute the normal
basi s for the thi nki ng of number. Transitivity i s at once that which
makes of number a section
2
taken from the equi l i brated fabric of
being and that whi ch provi des the norm for thi s secti on.
8
Von Neumann Ordi nals
8. 1 . Let' s consi der more cl osely set D, i ntroduced in 7. 12, written as
( 0, ( 0) ) , whi ch is the pai r of the voi d and the si ngleton of the voi d.
We know that set D i s transi ti ve: i ts two elements, 0 and ( 0) ,
are al so parts of D. We can make a further remark here: these two
elements are also transitive sets.

That ( 0) is transitive is sel f-evident: the onl y element of the si ngl e
ton (0) i s O. Now, 0 is a ' uni versal ' part i ncl uded i n every set,
and, in particul ar, it i s i ncl uded in the set ( 0) . So the uni que
element of ( 0) is al so a part of ( 0) , and consequently ( 0) is a
transitive set.

That 0, the empty set, is transitive results from its negative ' poros
ity' to every property, whi ch al ready makes it a part of any set
whatsoever ( compare 7. 9) : a transitive set i s one al l of whose ele
ments are al so parts. Thus a set that i s not transitive has at least
one element that is not a part. Now 0 has no elements. So it
cannot not be transitive. And, so, it i s.
With our set D we have constructed not onl y a transitive set, but
a transitive set of transitive sets: thi s transitive set 'gathers together'
transi tive sets. Both 0, ( 0) and thei r pai r ( 0, ( 0) ) , are transi ti ve.
8. 2. A truly fundamental defni ti on: A set is an ordinal (in von
Neumann's sense) ' i f it i s l i ke D, that i s, i f it is transitive and al l of
its elements are transitive.
8. 3. Thi s defnition completes the techni cal part of the ontologisation
of the concept of the ordinal . We are no longer deal i ng with wel l
orderedness, wi th the i mage of the sequence of natural whol e numbers,
or with an operati onal status. Our concept i s purely i mmanent. It
descri bes a certain i nternal structural form of the ordi nal , a form that
connects together i n a singul ar fashi on the two cruci al ontological
operators belonging and i ncl usi on, E and Co
Set D, which we have used as an exempl ary case, i s therefore an
ordi nal . We can l i ft a corner of the veil on its i dentity: it i s the number
Two. Moreover, thi s Two al l ows us to affrm that von Neumann
ordi nal s exist.
8. 4. Before depl oyi ng this new concept of the ordi nal , let's begi n with
a frst exami nati on of the status of its defni ti on and of the reasons
why the ordi nal s constitute the absol ute ontological horizon of all
numbers.
8. 5. I have i ndicated ( 7. 1 6) that a transitive set is the ontological
schema of the ' normal ' mul ti pl e. Taking into account the fact that
the excess of representation over presentation is i rremedi abl e, tran
si ti vi ty represents the maxi mal equi l i bri um between the two.
Now, not only is an ordi nal transitive, but all of its elements are
al so transitive. An ordi nal di ssemi nates to the interi or of a multiple
that normal ity whi ch characterises it. It i s a normal ity of normal ities,
an equi l i bri um of equi l i bri a.
A trul y remarkabl e property results from thi s, whi ch i s that every
element of an ordinal is an ordinal.
Take an ordi naF W, and an element of that ordi nal x (so that
x E W) . W being an ordinal , al l of its elements are transitive, so x is
transi ti ve. For the same reason ( the ordi nal ity of W) W i s itsel f transi
ti ve, so x, an element of W, is al so a part of W: x C W. As a result,
all the el ements of x are elements of W. And, j ust as all the elements
of W are transitive, the same fol l ows for all the elements of x. The
set x i s thus a transitive set al l of whose elements are transitive: it is
an ordinal .
8. 6. I f transitivity i s a property of stabi l ity, thi s ti me we di scover a
complementary property of homogeneity: that whi ch makes up the
internal multi pl e of an ordi nal , the elements belonging to it, are al l
ordi nal s. An ordi nal is the count-for-one of a mul ti pl icity of
ordinal s.
Because of thi s homogenous and stabl e ' fabric' of ordi nal multi
pl icity, I have been led to say that ordi nal s are the ontological schema
VON NEUMANN ORDINALS 69
of the natural multiple. I call ' natural ' ( by way of opposi ti on to
mul ti pl icities that are unstabl e, heterogeneous, historical, and whi ch
are thus exposed to the evental caesura) preci sel y that whi ch i s exem
pl i fed by the underlyi ng mul ti ple-being as thought by mathemati cs:
a maxi mal consistency, an i mmanent stabi l ity wi thout l acuna, and a
perfect homogeneity, in so far as that of whi ch thi s multi ple-bei ng is
composed is of the same type as itsel f.
We therefore posit, once and for al l , that an ordi nal is the i ndex
of the being of a natural mul ti pl icity.
8. 7. I f it i s true that the ordi nal s constitute the great ontol ogi cal
'ground' of number, then we can al so say that number i s a fgure of
natural being, or that number proceeds from Nature. Wi th the caveat,
however, that ' Nature' refers here to nothi ng sensi ble, to no experi
ence: 'Nature' is an ontologi cal category, a category of the thought
of the pure multi pl e, or set theory.
8. 8. Must we say si multaneousl y that ordi nal s ' are numbers ' ? Such
would indeed be the i dea of Cantor, who thought to achi eve by way
of the ordinals an i nfnite prolongation of the sequence of whole
numbers. But for us, who have yet to propose any concept of number,
this would be begging the questi on. We wi l l see, after havi ng defned
what I call Number (the capi tal i sation is not for the sake of maj esty,
but to designate a concept that subsumes all species of number,
known or unknown) , that the ordi nal s, though playing a deci sive role
in this defni ti on, are onl y the representable amongst numbers, in the
numerical swarming which being lavishes on the ground of Nature.
The ordi nal s wi l l thus be at once the i nstrument of our access to
number, of our thi nki ng of number, and, al beit l ost i n a profusi on
of Numbers that exceeds them in every way, they wi l l be represent
able or fgurable as themselves, too, being Numbers .
8. 9. The empty set, 0, is an ordinal . We have seen above that it i s
transitive ( 8. 1 ) . It fol lows that al l i ts el ements are al so transitive:
havi ng no elements, how coul d it have an element that was not transi
tive? Contrary to al l intuition, zero, or the voi d, i s a natural ontologi
cal donation. The voi d, whi ch sutures all l anguage and all thought
to being, i s al so the point of nature where number is anchored.
8. 10. Von Neumann ordi nal s have two cruci al properties:
1 They are total l y ordered by the fundamental ontological relation
belonging, the sign of mul ti ple-presentati on. That is to say that,
given two ordi nal s WI and W
2
, either the former belongs to the
l atter ( W
I
E W
2
) , or the other way around (W
2
E Wd, or they
are identical ( WI = W2 ) .
2 They obey a pri nci pl e of mi ni mal ity: gi ven any property P what
soever, if an ordi nal possesses thi s property, then there exists a
smal lest ordi nal to possess it. Order is al ways belonging: if you
have an ordi nal W such that it possesses the property P (if the
statement P( W) is true) , then there exi sts an ordi nal WI which
has the property and whi ch i s the smal lest to have it (if W2 E W b
W2 doesn't have the property) .
These two properties are natural . The frst expresses the universal
i ntrication of those stable and homogenous multi pl icities that are
natural mul ti pl icities ( see 8. 6) : thought in their being, two natural
mul ti pl es - two ordi nal s, then - cannot be i ndependent. Either one
is i n the presentation of the other, or vice versa. Nature does not
tolerate i ndi fference or di sconnecti on. The second property expresses
the ' atomi c' or, if you l i ke, ' quantum' character of nature. I f a prop
erty appl i es to some natural multi pl e, then there is always a natural
multi pl e that i s the mi ni mal support of that property.
Taken together, these two properties reunite the global status
of nature with its local status. Even though Nature
3
does not exist
( there is no set of all the ordi nal s, see 6. 1 1 ) , there i s a sort of unity
of pl an, of gl obal interdependence, between natural mul ti pl es:
the presentati on of whi ch they are the schema i s al ways 'embedded' .
And, al though there are no uni que and i ndiscerni bl e components
of nature like the Ancients' atoms ( unless one considers the void
as such) , there i s an exceptional local point for every property
that obtai ns for the ' regi ons' of nature: the mi ni mal support of this
property.
This arti cul ation of the global and the local l ends its ontol ogical
framework to every Physi cs.
8. 1 1 . The two cruci al properties ( total order and mi ni mal ity) can
both be proved on the basi s of von Neumann' s defni ti on of the
ordi nal s.
These proofs depend upon a key pri nci pl e of set theory ( ontol ogy
of the multi ple) : the Axi om of Foundati on.
4
Thi s axi om says that
every situation ( every pure mul ti pl e) comprises at least one term ( one
element) that has ' nothing in common' with the situation, i n the
sense that nothing of that which composes the term ( no element of
the el ement) i s presented in the situation ( bel ongs to the original
multi pl e ) .
VON NEUMANN ORDINALS 71
8. 12. Let' s return to the exampl e of my cat ( 7. 6) . It i s an element of
the set of l iving beings, and it is composed of cel l s that are in turn
elements of this set, if one grants that they are l i vi ng organi sms. But
i f we decompose a cel l into molecules, then into atoms, we eventual l y
reach purely physi cal elements that don' t belong to the set of l i vi ng
bei ngs. There is a certai n term ( perhaps the cel l , in fact) which
belongs to the set of l i ving beings, but none of whose el ements
belongs to the set of l i ving beings, because those elements all i nvol ve
onl y ' inert' physico-chemi cal materi al ity. Of thi s term, whi ch belongs
to the set but none of whose el ements belongs to it, we can say
that it grounds the set, or that it is a fundamental term of the set.
'Fundamental ' meani ng that on one si de of the term, we break through
that whi ch it constitutes; we leave the origi nal set, we exceed its
presentative capacity.
8. 13. Once more, let' s leave l i ving beings, cats, cel l s and atoms
behind. Consider the si ngl eton of the si ngl eton of the voi d, that i s,
the set whose uni que element i s the si ngleton of the voi d, and whi ch
is written as ( ( 0) ) . The element ( 0) of thi s set has as its onl y element
the void, O. Now the voi d is not an element of the origi nal set ( ( 0) ) ,
whose onl y element i s ( 0) , because the void 0 and the si ngl eton
of the void ( 0) are di fferent sets. So ( 0) represents, in ( ( 0) ) , a local
foundati on-point: it has no element in common with the original set
( ( 0) ) . That whi ch it presents qua multiple - that i s, 0 - i s not presented
by ( ( 0) ) , in the presentation in which i t fgures.
The Axiom of Foundation tel l s us that thi s si tuati on i s a l aw of
being: every multi pl e i s founded, every multi pl e comprises at least
one element whi ch presents nothing that the mul ti pl e itsel f
presents.
8. 14. The Axiom of Foundation has a remarkabl e consequence,
which i s that no set can belong to itself, that no multiple fgures i n
i ts own presentati on, that no mul ti pl e counts itsel f as one. I n thi s
sense, being knows nothing of refection.
Take a set E whi ch is an element of itsel f: E E E. Consi der the
singleton of thi s set, ( E) . The only el ement of thi s singleton i s E. So
E must found ( E) . But thi s i s i mpossi bl e, since E bel ongs to E, and
thus has i n common with ( E) that element whi ch i s itsel f. Si nce the
axiom of foundation is a l aw of being, we must rej ect the ori gi nal
hypothesi s: there does not exist any set that i s an el ement of itsel f.
8. 15. Returning to the cruci al properties of the ordi nal s: They can
be proved, once the axi om of foundation i s assumed. I will do so
here for the pri nci pl e of mi ni mal i ty. For the pri nci pl e of total order
through belonging, see the note. 5
Take an ordi nal WI which possesses property P. If it i s mi ni mal ,
al l i s wel l . Suppose that it i s not. In that case, there exi st ordi nal s
smal l er than WI ( and whi ch therefore bel ong to WI , since the order
in question i s bel ongi ng) and whi ch also possess the property. Con
si der the set E of these ordi nal s ( 'gathering together' all those which
possess property P and belong to Wd. Set E obeys the Axi om of
Foundati on. So there i s an element Wz of E whi ch i s an ordi nal ( si nce
E is a set of ordi nal s) that possesses property P ( si nce al l the elements
of E possess it) and that has no element i n common with E.
But, si nce W I i s an ordi nal , it is transitive. So W 2
,
which belongs
to it, i s also a part of it: the elements of W2 are all elements of WI .
If an element of W2 possesses property P, then, since it is an element
of WI ! it must belong to E ( since E i s the set of all the elements of
WI possessing property Pl . Which cannot be, because W2 founds E
and therefore has no element i n common with E. Consequently,
no element of W2 has the property P, and Wz i s mi ni mal for this
property. QED.
8. 16. Thus i s knitted the ontological fabri c from whi ch the numbers
wi l l be cut out. 6 Homogenous, intricate, rooted in the voi d, local l y
mi ni mi sabl e for every property, it i s very much what we coul d cal l a
horizonal structure.
9
Successi on and Li mi t.
The I nfni te
$4
9. 1 . In chapter 6, when we spoke of Dedeki nd' s and Cantor' s
approaches to the notion of the ordi nal ( on the basi s of wel l -ordered
ness ) , we saw that the whole problem was that after one ordinal
comes another, well -determined, and that thi s series can be pursued
indefnitel y. We also saw that it was not at all the same thing to ' pass'
from n to n + 1 ( its successor) as to pass from ' al l ' the natural
numbers to their beyond, which is the i nfnite ordi nal . In the latter
case, there is mani festly a shi ft, the punctuation of a ' passage to the
l i mi t' .
In the ontologised concept of the ordi nal s which von Neumann
proposed and to whi ch we dedicated chapter 8, do we once more
fnd thi s di alectic between si mpl e successi on and the ' l eap' to the
infnite? And, more general l y, how does the thorny i ssue of the exi s
tence of an i nfnite mul ti pl e present itsel f i n thi s new context ?
9. 2. Let' s appl y ourselves frstly to the concept of successi on.
We must take care here. The i mage of successi on, of ' passage' to
the next, is so vi vi dl y present in the i mmedi ate representation of
number that it i s often thought to be constitutive of i ts essence. I
reproached J. A. Mi l ler ( see 3. 17) precisely for reduci ng the probl em
of number to the determination of that whi ch i nsists in its succes
si onal engenderment. I hel d that the l aw of the seri al passage across
the numeric domai n, a l aw which i s i mposed on us, does not coincide
with the ontological i mmanence of number as singular form of the
mul ti pl e.
Consequently, if we fnd the idea of succession once again in von
Neumann' s conception of the ordi nal s, it too must yi el d to the process
of ontologi sati on. Our goal wi l l be to discover, not so much a pri n
ci ple of passage as an i ntri nsi c qual ifcation of that which succeeds,
as opposed to that whi ch does not. What counts for us i s not succes
si on, but the being of the successor. The repetitive monotony of
Peano' s +1 does not concern us any longer: what we want to thi nk
is the proper being of that which can onl y be attained in the modal ity
of the additi onal step.
9. 3. Let's consi der an ordi nal W, i n von Neumann' s sense ( a transi
ti ve set al l of whose el ements are transitive) .
A set, then, whose elements are:
- al l of the el ements of W;
- W itsel f.
So, to everything that composes the multi pl e W, we ' add' one
supplementary element, namely W itsel f. And it i s i ndeed a question
of the adj unction of a new element, si nce we know ( it is a conse
quence of the axiom of foundati on, compare 8. 14) that W i s never
an element of itsel f.
A non-operati onal form of +1 can be seen emerging here: it is not
a matter of an extrinsic additi on, of an external ' pl us' , but of a sort
of i mmanent torsi on, whi ch ' completes' the interior multiple of W
with the count-for-one of that mul ti pl e, a count whose name is pre
ci sel y W. The +1 consi sts here i n extending the rule of the assembl y
of sets to what had heretofore been the pri nci pl e of this assembl y,
that i s, the unifcation of the set W, whi ch i s thereafter al igned with
its own elements, counti ng along with them.
9. 4. An exampl e of the procedure. We have demonstrated that set
D, which i s written ( 0, ( 0) ) , and whi ch is the pai r of the voi d and the
si ngleton of the voi d, is an ordi nal ( it i s transiti ve and al l its elements
are transitive) . Our non-operati onal defni ti on of +1 consists in
formi ng the set of the three fol l owing elements: the two elements of
D and D itsel f. We write thi s as ( 0, ( 0) , ( 0, ( 0) ) ) ( the ' whol e' D i s found
in the thi rd position) . Call thi s tri plet T. We can now demonstrate
that:

T is transitive. Its frst el ement, 0, is a uni versal part, and so it
must be a part of T; its second element, ( 0) , is the si ngleton of its
frst el ement, o. So it i s al so a part of T ( see 7. 1 1 ) . Its thi rd element
SUCCESSION AND LIMIT. THE INFINITE 75
( 0, ( 0) ) is nothi ng but the 'gathering together' , the formi ng i nto a
pai r, of these frst two. So it is al so a part. So every element of T
is a part, and T is transitive.

Al l the elements of T are transi ti ve. Gi ven that we have shown
that D i s an ordi nal , we have dul y shown that its el ements, 0 and
( 0) , are al so transi ti ve. We have equal l y demonstrated that it itsel f,
( 0, ( 0) ) , is transitive. And these are preci sel y the three elements
of T.
So T, obtained by ' adj oi ni ng' D to the el ements of D, i s a von
Neumann ordinal : a transitive set al l of whose elements are
transitive.
9. 5. The reasoning we j ust fol l owed can easily be general i sed. For
any ordi nal W whatsoever, everything wi l l fol low j ust as for T: the
set obtained in adj oi ning W itsel f, as an element, to W' s elements is
an ordi nal .
We 'step' from W to a new ordi nal by adj oi ning to W' s elements
a single additional element ( thi s, now, al lows us to l i ft a corner of
the vei l on the i dentity of our exampl e T: j ust as D was two - I woul d
l i ke to say the being of number Two - T is none other than the
number Three) .
The fact that one steps from W t o a new ordi nal , whose
elements are those of W supplemented by the one-name of thei r
assembl y, by way of a sort of i mmanent +1 , j ustifes the fol l owi ng
defnition: we will call the ordinal obtained by joining W to the
elements of W, the successor of the ordinal W, and will denote it
by S(W).
So, in our exampl e, T ( three) i s the successor of D ( two) .
9. 6. The idea of the ' passage' from two t o three, or from W t o S( W) ,
i s, i n truth, purel y metaphorical . In fact, from the start there are
fgures of a mul ti ple-being, D and T, and what we have defned i s a
relation whose sol e purpose is to faci l itate for us the intel l igi ble
passage through their exi stences. Finitude demands the bi ndi ng of the
un-binding of being. We therefore thi nk, i n the successi on T = S( D) ,
a relation whose basi s i s , i n truth, i mmanent: T has the structural
property, veri fable in its ontological composi ti on, of bei ng the suc
cessor of D, and it i s no more than a necessary i l l usi on to represent
T as being constructed or defned by the relation S, which connects
it external l y to D.
A more rigorous phi l osophi cal approach consi sts i n exami ni ng the
ordi nals i n themselves and i n asking ourselves whether they possess
the property of succeeding. For exampl e, T possesses the property of
succeeding D, recognisable in itself from the fact that D is an element
of T, and, what' s more - as we shal l see - that D i s an element that
can be i mmanently distingui shed ( it is ' maxi mal ' in T) .
We wi l l cal l successor ordinal an ordi nal that possesses the prop
erty of succeedi ng.
So T i s a successor ordi nal .
9. 7. It mi ght be obj ected that the property ' succeeds W' i s sti l l latent
in the i ntri nsi c concept of successor, and therefore that we have fai led
to establ i sh oursel ves i n the ontological unbi ndi ng. This obj ection
can be al levi ated.
Let's consi der an ordi nal W having the fol lowing, purely i mma
nent, property: amongst the elements of W, there i s one element, say
wI . of which all the other elements of W are elements: i f Wz i s an
el ement of W di fferent from wI , then Wz E WI . I say that W i s neces
sari l y a successor ordi nal ( i n fact, it succeeds wd.
For i f thi s situation obtai ns, i t i s because W' s elements are:
- on the one hand the element WI ;
- on the other, elements whi ch, l i ke wz, are el ements of WI .
But, i n real ity, all the el ements of WI are elements of W. For we
know that belonging, E , is a total order over the ordi nal s ( see 8. 10) .
Now, al l the el ements of an ordi nal are ordi nal s ( 8. 5) ; specifcal l y,
al l the elements of W are ordinal s. WI i s therefore an ordi nal , and i t
fol l ows that the elements of WI are al l ordi nal s. These elements are
connected to ordi nal WI and W by the relation of total order that i s
belonging: i f W E WI , si nce WI E W, then w E W (transitivity of the
order-rel ati on) .
Thus W i s composed of al l the el ements of WI , and WI itself: W is
by defni ti on the successor of WI .
Let' s agree to cal l the maximal element of a n ordi nal the
element of that ordi nal which is l i ke WI for W: all the other
elements of the ordi nal belong to the maxi mal element. The
reasoni ng above now permits us to make the fol l owing defniti on:
An ordi1zal will be called a successor if it possesses a maximal
element.
Here we are in possessi on of a total l y intrinsic defnition of the
successor ordi nal . The si ngul ar exi stence of an ' i nternal ' maxi mum,
located sol el y through the exami nation of the mul ti pl e structure of
the ordinal , of the fabric of el ementary bel onging at its heart, al lows
us to decide whether i t i s a successor or not.
9. S. Si nce we now have an i mmanent, non-rel ati onal and non-seri al
concept of ' what a successor i s' , we can pose the questi on: Are there
ordi nal s that are not successors ?
9. 9. The empty set, 0, is an ordi nal that is not a successor. It obvi
ousl y cannot succeed anythi ng, since it has no el ements and, to
succeed, it must have at least one element, namely the ordi nal that it
succeeds.
Or, stayi ng cl oser to the i mmanent characterisati on: to be a suc
cessor, must have a maxi mal element. Havi ng no el ements, it
cannot be a successor.
Once agai n, we di scover the voi d' s function as ontol ogi cal anchor:
purely decided in its being, it i s not inferable and, i n particul ar, it
cannot succeed: the voi d i s itself on the edge of the void, there i s no
way it coul d fol l ow from bei ng, of whi ch it is the ori gi nal poi nt.
9. 10. Al l the ordi nal s that we have used in our exampl es, apart from
the void, are successors. Thus ( 0) ( which i s the number 1 ) is the suc
cessor of 0. The number 2, whose bei ng i s ( 0, ( 0) ) , and whi ch is
composed of the voi d and 1 , is the successor of 1 . And our T (the
number 3) , whi ch i s composed of the voi d, 1, and 2 and i s written
( 0, ( 0) , ( 0, ( 0) ) ) , i s the successor of 2. It is cl ear that we can conti nue,
and wi l l thereby obtain 4, 5, and, fnal l y, any of the natural whol e
numbers, all of which are successor ordinals.
9. 1 1 . Does thi s mean that we have at our di sposal a thi nking of
natural whol e number? Not yet. We can say that 1, then 2, then 3 ,
etc. , i f we thi nk each i n its mul ti ple-bei ng, are natural whol e numbers .
But, without being able to determi ne the place of thei r deployment,
it is i mpossi bl e for us to pass beyond this case- by-case designation
and to propose a general concept of whol e number. As Dedeki nd
perceived, such a concept necessitates a detour through the i nfnite,
si nce it is within the i nfnite that the fnite i nsi sts. The only thing that
we can say with certainty is that whole numbers are successor ordi
nals. But this i s certai nl y not a suffcient characteri sati on of them:
there mi ght well be other successors that are not whol e numbers,
perhaps successors that are not even fnite sets.
9. 12. The question becomes : are there any other non-successor
ordi nal s apart from the void?
Let' s cal l these non-successor ordi nal s di fferent from ( wi thout
yet knowing whether they exi st) limit ordinals. We ask once more:
do l i mi t ordi nal s exi st?
We are not yet in a position to decide upon thi s questi on. But we
can prove that, if they do, they are structural l y very di fferent from
successor ordi nal s.
9. 1 3. No ordi nal can come in between an ordi nal Wand i ts successor
S( W) . By this we mean that, gi ven that the order-rel ati on between
ordi nal s is that of belonging, no ordi nal W I exi sts such that we have
the sequence W E WI E S( W) .
We know i n fact that W i s the maxi mal element of S( W) ( see 9. 7) .
Consequentl y, every element of S( W) t hat i s di fferent from W belongs
to W. Now, our supposed WI belongs to S( W) . Therefore one of two
things must appl y:

ei ther WI is identical to W. But thi s is i mpossi ble, because we
have supposed that W E WI . whi ch woul d give us W E W. But
we know ( 8. 1 4) that no set can be an element of itsel f;

or W
I
is an element of W. But then it woul d not be possi ble that
W E WI . si nce WI E W.
It can be seen that ordi nal successi on is the schema of the 'one
more step' , understood as that which hol lows out a void between the
i ni ti al state and the fnal state. Between the ordi nal W and its succes
sor S( W) , there i s nothing. Meani ng: nothi ng natural , no ordi nal . We
coul d al so say that a successor ordi nal del i mits, j ust ' behi nd' itsel f, a
gap where nothi ng can be establ i shed. In thi s sense, rather than suc
ceedi ng, a successor ordinal begins: it has no attachment, no conti nu
i ty, wi th that whi ch precedes i t. The successor ordi nal opens up for
thought a begi nni ng in bei ng.
9. 14. A l i mit ordi nal , i f such a thi ng exists, is a di fferent case alto
gether. The defnition of such an ordi nal is, please note, purely nega
tive: it is not a successor; that i s all that we know of it for the moment.
We can also say: i t does not possess a maxi mal element. But the
consequences of thi s lack are consi derabl e.
Take L, a supposed l i mit ordinal , and WI , an el ement of thi s
ordinal . Si nce WI i s not maxi mal , there certai nl y exi sts an element Wl
of L which is l arger than it: so we have the chai n: WI E Wz E L. But,
si nce i n its turn Wz i s not maxi mal , there exi sts a W
3
such that
WI E Wz E W
3
E L. And so on.
Thus, when an ordi nal belongs to a l i mit ordi nal , a thi rd party is
i ntercal ated into the rel ati on of bel ongi ng, and, as thi s process has
no stopping poi nt, as there i s no maxi mal element, it can be sai d that,
between any el ement W of a l i mi t ordi nal L and L itsel f, there i s
always an ' infni ty' , i n the i ntuitive sense, of intermedi ate ordi nal s.
So it is i n a strong sense that the l i mit ordi nal does not succeed. No
ordi nal is the l ast to belong to it, the 'closest' to it. A l i mit ordi nal is
al ways equal l y ' far' from al l the ordi nal s that belong to it. Between
the element w of L and L, there i s an i nfnite distance where i nter
mediaries swarm.
The result is that, contrary to what is the case for a successor ordinal,
a limit ordinal does not hollow out any empty space behind itself. No
matter how 'close' to L you imagine an element w to be, the space
between w and L is infnitely populated with ordi nal s. The limit ordi nal
L is therefore in a relation of adherence to that which precedes it; an
infnity of ordinals 'cements' it in place, stops up every possi ble gap.
I f the successor ordi nal is the ontological and natural schema of
radical begi nni ng, the l i mi t ordi nal i s that of the insensible result, of
transformation wi thout gaps, of i nfnite conti nuity. Whi ch i s to say
that every acti on, every wi l l , i s placed either under the sign of the
successor, or under the si gn of the limit. Nature here furni shes us
with the ontologi cal substructure of the ol d problem of revol uti on
( tabula rasa, empty space) and of reform ( insensi bl e, consensual and
painless gradations) .
9. 1 5. There i s another way t o i ndicate the di fference between
successors and l imi ts ( which are for us the predicates of natural
mul ti ple-bei ng) .
The union of a set E is the set constituted by t he elements of the
elements of E. Thi s i s rel ated to a very i mportant operator of the
ontology of the mul ti pl e, the operator of dissemination. The uni on
of E ' breaks open' the elements of E and col l ects al l the products of
this breaking-open, al l the elements contained i n the elements whose
counting-for-one E assures.
An exampl e: take our canoni cal exampl e of three, the set T that
makes a triplet of the void, the si ngleton of the voi d and the pai r of
the voi d and i ts si ngleton. It i s wri tten ( 0, ( 0) , ( 0, ( 0) ) ) . What i s the
uni on of T?
The frst element of T i s 0, whi ch has no el ements. It therefore
donates no elements to the uni on. The second element is ( 0) , whose
si ngl e el ement i s O. Thi s l atter element wi l l feature i n the uni on.
Final ly the thi rd element i s ( 0, ( 0) ) , whose two el ements are 0 ( which
we al ready have) and ( 0) . So i n the end the uni on of T, the set of the
elements of i ts elements, i s composed of 0 and ( 0) : it i s the pai r ( 0, ( 0) ) .
That i s t o say, our D, or the number two. The di ssemi nati on of three
is no other than two. We state i n passing (this wi l l be cl arifed i n
9. 1 8) that the union of T is ' smal ler' than T itsel f.
9. 1 6. The posi ti on of ordi nal s with regard to uni on is most pecul i ar.
Gi ven that an ordi nal W i s transitive, al l its elements are al so parts.
And this means that the elements of the elements of W, which are
also the elements of the parts of W, are themselves elements of W.
In the uni on of an ordi nal we fnd nothi ng but the elements of that
ordi nal . That is to say that the union of an ordi nal is a part of the
ordinal. If we denote the set ' uni on of E' by uE, then, for every
ordi nal , uW C W.
Thi s property i s characteristical l y natural : the i nternal homogene
ity of an ordi nal is such that di ssemi nati on, breaki ng open that which
it composes, never produces anything other than a part of itself. Dis
semi nati on, when it i s appl ied to a natural multi pl e, del i vers only a
' shard' of that mul ti pl e. Nature, stabl e and homogenous, can never
' escape' its proper constituents through dissemi nation. Or: i n nature
there is no non-natural ground.
9. 1 7. That the union of an ordi nal shoul d be a part of that ordi nal ,
or that the elements of i ts elements shoul d be el ements, bri ngs us to
the questi on: are they all ? Do we ultimately fnd not even a ' parti al '
part ( or proper part, compare 4. 12) , but onl y the ordinal we began
with? It coul d wel l be that every element can be found as element of
an element, since the internal fabric of an ordi nal is entirely intri
cated. In that case, uW = W. Not onl y woul d dissemination return
only natural materi al s, but i t woul d restore the i ni ti al totality. The
di ssemi nati on of a natural set woul d be a tautological operati on.
Whi ch is to say that it woul d be absol utely in vai n: we coul d then
concl ude that nature does not allow itself to be disseminated.
9. 1 8. Thi s seductive thesi s i s verifed in the case of limit ordinals, if
such a case exi sts.
Take any element WI whatsoever of a l i mi t ordi nal L. We have
shown (in 9. 1 4) that between WI and L necessari l y comes an inter
cal ated element W2, in such a fashi on that we always have ( whatever
the element wd the chai n WI E W2 E L. But, in addi ti on, when we
di ssemi nate L the element WI wi l l be found agai n in the uni on, as an
element of W2 . Consequently, every element of L features in uL, the
uni on of L. And, as we have seen, conversely ( 9. 1 5) , that every
element of uL is an element of L ( since uL C L) , it onl y remains to
concl ude that the el ements of L and those of uL are exactly the same.
Whi ch is to say that L i s identical to uL.
To di ssemi nati on, the l i mit ordi nal opposes its i nfnite seI f
coalescence. It is exempl ari l y natural , in so far as, in bei ng ' di ssected' ,
i ts el ements do not al ter. It is its own di ssemi nation.
9. 1 9. A successor ordi nal , on the other hand, resi sts bei ng identifed
with its dissemi nation. It remai ns in excess of its uni on.
Let' s consider a successor ordi nal W. By defni ti on it has a maxi mal
element WI ' Now it i s i mpossi bl e that thi s el ement shoul d be found
i n the uni on of W. If i t were found, that woul d mean that it was the
element of another element, Wz, of W: so WI E Wz, and WI woul d not
be maxi mal . The maximal element WI necessarily makes the differ
ence between Wand uW. There i s at least one element of a successor
ordi nal that bl ocks the pure and si mpl e di ssemi nati ve restorati on of
i ts mul ti ple-bei ng. A successor, unl i ke a l i mi t, i s 'contracted' , altered,
by dissemi nati on.
9. 20. In my view, thi s contrast is of the greatest phi l osophi cal i mpor
tance. The prevai l ing idea i s that what happens ' at the l i mit' i s more
complex, and also more obscure, than that which i s in play in a suc
cession, or i n a si mpl e 'one more step' . For a l ong ti me phi l osophi cal
speculation has fostered a sacral i sation of the l i mi t. What I have
cal led elsewhere
I
the ' suture' of phi l osophy to the poem rests l argely
upon thi s sacral isation. The Hei deggeri an theme of the Open, of the
deposition of a cl osure, i s the modern form of the assumpti on of
the l i mi t as a wrenchi ng away from counting, from techni que, from
the succession of discoveries, from the seri al ity of Reason. There is
an aura of the l imit, and an unbei ng of successi on. The ' heart come
from another age' aspi res ( and thi s horizon-effect is onl y captured,
so i t seems, by the poem) to a movement across those ' infnite
meadows where al l time stands sti l l ' .
z
What the ontology of the mul ti pl e ( based in a contemporary
Platonism) teaches us i s, on the contrary, that the di ffculty resi des
in successi on, and that there, al so, resides resistance. Every true
test for thought originates i n the l ocal i sabl e necessity of an addi
ti onal step, of an unbroachable begi nni ng, whi ch i s nei ther fused
through the i nfnite repl eni shment of that which precedes it, nor
identical to i ts di ssemi nati on. To understand and endure the test
of the addi ti onal step, such is the true necessity of ti me. The l i mi t
is a recapitulation of that whi ch composes it, i ts ' profundi ty' i s
fal laci ous; it i s in vi rtue of its havi ng no gaps that the l i mi t ordi nal ,
or any mul ti pl icity ' at the l i mi ts' , attracts the evocative and hol l ow
power of such a ' profundity' . The empty space of the successor
is more redoubtabl e, it i s truly profound. There i s nothi ng more to
thi nk in the l i mi t than in that whi ch precedes it. But i n the successor
there is a crossi ng. The audacity of thought i s not to repeat ' to the
l i mi t' that which i s al ready entirely retai ned wi thi n the situation
which the limit limits; the audacity of thought consi sts in crossi ng a
space where nothing is given. We must learn once more how to
s ucceed.
9. 21 . Basi cal l y what i s di ffcul t in the limit i s not what it gives us to
thi nk, but its existence. And what is di ffcul t i n succession i s not its
exi stence ( as soon as the void i s guaranteed, it fol l ows i nel uctably)
but that which begi ns in thought with thi s exi stence.
And so, speaking of the l i mi t ordi nal , the question returns, ever
more i nsi stent: do l i mit ordi nal s exist ? On condition of the existence
of the voi d, there i s 1 , and 2, and 3 . . . , al l successors. But a l i mi t
ordinal ?
The reader wi l l have real i sed: we fnd ourselves on t he verge of the
deci si on on the infnite. No hope of proving the existence of a si ngle
l i mit ordi nal . We must make the great modern decl arati on: the inf
ni te exists, and, what i s more, it exi sts i n a wholly banal sense, being
neither revealed ( rel i gi on) , nor proved ( mediaeval metaphysics) , but
being si mpl y decided, under the i nj unction of being, i n the form of
number. All our preparati ons amount onl y to saying, to being able
to say, that the i nfnite can be thought in the form of number. We
know it, at least for that which fal l s wi thi n the natural ontological
horizon of number: the ordi nal s. That is i nfnite whi ch, not being
voi d, meanwhi l e does not succeed. It i s ti me to announce the
fol l owi ng:
Axi om of I nfni ty. P l i mi t ordinal exists.
1 0
Recurrence, or I nduction
10. 1 . A momentary pause to begin with: l et' s recapitulate what the
ordi nal s give us to thi nk as regards being qua being, from the view
point of a phi losophy i nformed by mathematical ontology.
10. 2. The ordi nal s are, because of the i nteral stabi l ity of their
multiple-bei ng (the maximal identity between belonging and i ncl u
si on, between ' frst' presentation through the multi pl e, as element,
and re-presentati on through i ncl usi on, as part) and the total homo
geneity of thei r internal composi ti on ( every element of an ordi nal i s
an ordinal ) , the ontological schema of natural mul ti pl ici ty.
1 0. 3. The ordi nal s do not constitute a set: no mul ti pl e-form can
total i se them. There exist pure natural multi pl es, but Nature does not
exist. Or, in Lacani an terms: Nature i s not-al l , j ust as i s being qua
being, si nce no set of al l sets exi sts either.
lOA. The anchori ng of the ordinals in being as such i s twofol d.
The absol utely i ni ti al poi nt that assures the chai n of ordi nal s of
i ts bei ng i s the empty set 0, decided axi omatical l y as secul arised form,
or number-form, of Nothingness. Thi s form i s nothing other than
the situation-name of being qua bei ng, the suture of every si tua
tion-being, and of every l anguage, to thei r l atent bei ng. The empty
set being an ordi nal , and therefore a natural mul ti ple, we might say:
the poi nt of bei ng of every si tuati on is natural . Materi al i sm i s founded
upon thi s statement.
10. 5. The poi nt-l i mit that ' restarts' the existence of the ordi nal s
beyond Greek number ( the fnite natural whol e numbers; on Greek
number, see chapter 1 ) is the frst i nfnite set, (, deci ded axiomatically
as a secul arised form - and thus enti rel y subtracted from the One - of
i nfnite mul ti pl icity.
From thi s poi nt of view, the ordi nal s represent the modern scale
of measurement ( conformi ng to the two cruci al deci si ons of modern
thought) of natural multi pl ici ty. They say that nothi ngness i s a form
of natural and numerabl e being, and that the infnite, far from being
retained in the One of a God, is omni present in nature, and, beyond
that, i n every s ituati on-bei ng.
1 0. 6. Our passage through the ordi nal s ( or the l i mi ts of our repre
sentation of them) arranges them according to an untotal i sable
sequence. Thi s sequence ' starts' with O. It continues through the
natural whole numbers ( 1 ,2, . . . ,n,n + 1 , . . . , etc. ) , numbers whose
form of being is composed of the void ( i n the forms ( 0) , ( 0, ( 0) ) , ( 0, ( 0) ,
( 0, ( 0) ) ) , . . . , etc. ) . I t i s continued by an i nfnite ( re)commencement,
guaranteed by the axi om ' a l i mi t ordi nal exi sts' , which authorises the
i nscri pti on, beyond the sequence of natural whol e numbers, of (, the
frst i nfnite ordi nal . This recommencement opens a new series of
successi ons: (, ( + 1 , . . . , ( + n, . . . , etc. Thi s series i s cl osed beyond
itsel f by a second l i mi t ordi nal , ( + (, which inaugurates a new series
of successi ons, and so on. Thus we achieve the representation of a
series of ordi nal s, depl oyed with no concei vabl e stoppi ng point,
which transits within the i nfnite ( beyond () j ust as i n the fnite.
1 0. 7. The orderi ng pri nci pl e of thi s sequence is i n fact belonging
i tsel f: given two ordi nal s WI and W2, then WI E W2, or W2 E WI .
or WI W2 Bel onging, a uni que ontol ogical relation because it
governs the thi nking of multi pl e-being as such, i s also that which
total l y orders the series of ordi nal s. So that, i f W i s an ordi nal and
S( W) its successor, then W E S( W) . So that, i f n is a natural whol e
number ( a fnite ordi nal ) and n
'
a ' l arger' whol e number, then n E
n
'
. And so that, for any natural whol e number n whatsoever, n E (
( the frst i nfnite ordi nal ) , etc.
1 0. 8. There are three types of ordi nal ( after the modern deci si ons
whi ch impose the voi d and the i nfnite) :
1 The empty set, 0, is the i naugural poi nt of being.
2 The successor ordi nal s adj oi n to thei r predecessor one element,
namel y that predecessor itsel f. The successor of W is cal led S( W) .
RECURRENCE. OR INDUCTION 85
W is the maxi mal element i n S( W) , and the presence of a maxi mal
element al lows us to characterise successors i n a purel y i mmanent
( non-seri al ) fashi on. Successor ordinals give us a numerical schema
for what it means to say ' one more step' . Thi s step consists al ways
i n supplementing al l that one has at one' s di sposal , wi th a uni que
mark for that al l . To take ' one more step' comes down to maki ng
one of al l of the gi ven multipl icity, and adj oi ni ng that one to i t.
The new si tuati on i s ' maxi mal i sed' : it contai ns one term that
domi nates al l the others.
3 The l i mit ordi nal s have no maxi mal internal element. They mark
the beyond proper to a series wi thout stoppi ng poi nt. They do
not succeed any particul ar ordi nal , but it can be sai d that they
succeed all the ordinals of the sequence of whi ch they are the
l i mi t. No ordi nal i n thi s sequence i s ' cl oser' to the limit ordi nal
than any other. For a thi rd ordi nal , and ul ti mately an ' i nfni ty'
( in the intuitive sense of a series wi th no stoppi ng poi nt) of ordi
nal s, wi l l intercal ate themselves ( according to the order-rel ati on,
whi ch i s belonging) between every ordi nal of the sequence and
the l i mi t ordi nal . The l i mi t ordi nal adheres to everythi ng that
precedes it. Thi s i s speci fcal l y i ndicated by its identity wi th i ts
own di ssemi nati on ( L uL) . The l i mi t totalises the sequence, but
does not distinguish any particul ar ordi nal within i t.
10. 9. Just as a l i mi t ordi nal i s structural l y di fferent from a successor
ordi nal ( as regards the i nternal maxi mum, and as regards dissemi na
tion) , so the ' passage to the l i mi t' is an operation of thought entirely
di fferent from ' taki ng one more step' .
Succession i s, i n general , a more difcult l ocal operation than the
global operation of passage to the l i mi t. Successi on gives us more to
think about than does the l i mit. The wi despread view to the contrary
stems from the fact that, not being ' absolutely modern' , we sti l l tend
to sacral i se the i nfnite and the l i mi t, which is to say: retai n them sti l l
in the form of the One. A secul ari sed thought, subtracted from the
One and the sacred, recognises that the most redoubtable problems
are local problems, probl ems of the type: ' How to succeed? ' , ' How
to take one more step? ' .
10. 1 0. The space of the ordi nal s al lows us t o defne t he i nfnite and
the fni te. An ordi nal is fnite i f, in the chai n of order governed by
belonging, it comes before 0. It i s i nfnite i f i t comes after 0 ( i ncl ud
ing 0 itsel f) .
We wi l l fnd that, j ust as Dedeki nd' s intuition suggested, onl y the
exi stence of an i nfnite ordi nal permi ts us to defne the fni te. Modern
thought says that the frst si tuati on, the banal situation, is the i nfnite.
The fnite is a secondary si tuati on, very speci al , very si ngul ar,
extremel y rare. The obsessi on with ' fnitude' is a remnant of the
tyranny of the sacred. The 'death of God' does not deliver us to fni
tude, but to the omni present i nfnitude of si tuati ons, and, correla
ti vel y, to the i nfnity of the thi nkabl e.
1 0. 1 1 . The fnal synthetic recapitulation of the fact that the ordinals
give us to think bei ng qua being, i n its natural proposi ti on, is
complete. Now we must turn towards our capacity to traverse and
to master rational l y this donation of being. One way to do so is
simply to proceed, i n thi s boundl ess fabric, to the carvi ng-out of
Number.
10. 12. It i s a bl essi ng for our subj ective fnitude that the authority
properly wi thout measure - of natural mul ti pl icities al l ows that diag
onal of passage, or of j udgement, which is reasoni ng by recurrence,
also cal led complete induction and, i n the case of i nfni te ordi nal s,
' transfnite i nduction' . In fact thi s al one al lows us, i n treati ng of an
i nfnite domain ( and even, i f we consider the ordi nal s, one that is
i nfnitely i nfni te) , to antici pate the mome1tt of conclusion.
Suppose that we wish to show that all ordi nal s possess a certain
property P. Or that we wi sh to establ i sh rational l y, with a proof, a
universal statement of the type: ' For al l x, if x is an ordinal , then
P( x) , . How can this be achieved? It i s certai nl y i mpossi bl e to confrm
case by case that it i s so: the task would be i nfnitely infnite. Neither
i s it possi bl e to consi der the 'set of ordi nal s' , si nce such a set does
not exist. The ' al l of the ordi nal s' , i mpl i ed in the uni versal quantifer
of the statement ' for all x', cannot be converted into 'all the elements
bel onging to the set of ordinal s' . Such a set i s i nconsi stent ( see 6. 1 1 ) .
It is preci sel y the al l evi ati on of thi s i mpasse that is the busi ness of
reasoni ng by recurrence.
1 0. 1 3. Reasoni ng by recurrence combines one verifcation and the
demonstration of one i mpl ication. Once in possessi on of these two
moments, the structure proper to the ordi nal s authori ses the uni versal
concl usi on.
Take property P. We begin by confrming that the empty set 0 pos
sesses thi s property; we test P for the ' case' of O. If the empty set does
not possess the property P, it is poi ntless to pursue the i nvestigation.
If one ordinal , 0, does not have property P, it i s certai nl y false that
all ordi nal s do. Suppose, then, that the statement P( O) is true, that
the test i n the case of 0 i s posi ti ve.
RECURRENCE. OR INDUCTION 87
We wi l l now try to prove the fol l owing i mpl icati on: i f al l the
ordi nal s that precede some ordinal W ( according to the total ordering
of the ordi nal s, whi ch i s belonging) have the property P, then W al so
has i t.
Note that thi s i mpl ication does not tel l us that an ordi nal wi th
property P exists. It remai ns i n the hypothetical register, according
to the general pattern: ' if x i s so, then what fol l ows x i s so' . The
impl ication is real l y universal , it does not speci fy any ordinal W. It
says onl y that, for every ordi nal W, supposi ng that those whi ch
precede it i n the chai n of ordi nal s sati sfy P, one i s compel led to admi t
that W satisfes i t al so.
It is usual l y necessary to divide thi s demonstration ( suppos ing that
it is possi bl e, which obvi ousl y depends on property P) , by treating
the case where we suppose W to be a successor separatel y from the
case where we suppose it to be a l i mit ( si nce W is any ordi nal what
soever, it coul d be one or the other) . Reasoni ng by recurrence, as we
saw i n the central i mpl ication that constitutes it, strongly bi nds that
which is the case for an ordi nal W to that whi ch i s that case for the
ordi nal s that precede i t. Now the rel ati onshi p of a l i mi t ordi nal to
the anterior ordinal s ( one of i nfnite adherence) di ffers radical l y from
that of a successor ( whi ch, between itsel f and its predecessor, cl ears
an empty space) . Because of thi s, the procedures of thought and of
proof put i nto pl ay i n the two cases are usual l y heterogeneous. And,
as we mi ght expect, gi ven the phi l osophy of thi s heterogeneity ( d.
9. 19) , it is general l y the case of the successor that is the most
di ffcult.
Assume that we have verifed the truth of P( O) , and that we have
proved the i mpl ication ' i f, for every ordi nal w that precedes W ( that
belongs to W: order is belonging) , it i s the case that P( w) , then it i s
al so the case that P( W) ' . We can concl ude that all ordi nal s sati sfy P,
in spite of the fact that thi s ' al l ' not onl y al l udes to an i nfnitely inf
ni te immensity of multi pl es, but that, even so, it does not make an
All. It is truly the i nfnite and i nconsi stency ' conquered word by
word' .
10. 14. What authorises such a passage to ' al l ' , such an ambiti ous
'moment of concl usi on' ? The authori sati on i s granted us by a funda
mental property of the ordi nal s as ontol ogical schema of the natural
multiple: thei r 'atomi sti c' character, the exi stence, for every property
P, of a minimal support for this property as soon as one ordi nal pos
sesses it. ( See 8. 10 and 8. 1 5 ) .
I f the concl usi on were false - i f i t were not the case that al l ordi nal s
possess property P - that woul d mean that there was at l east one
ordi nal which di d not possess property P. Thi s ordi nal would then
possess the property not-P, not-P meaning si mpl y 'not possessing
property P, bei ng a non-P' .
But, i f there exi sts an ordi nal that possesses property not-P, there
exists a smallest ordinal which possesses this property 1ZOt-P, by
vi rtue of the atomistic pri nci pl e, the pri nci pl e of mi ni mal ity. And,
si nce it i s the smallest to possess property not-P, al l those which are
smal ler than it must possess property P.
We coul d obj ect: these ordi nal s ' smal ler than it' may not exi st,
because i t is possi ble that the mi ni mal ordi nal for the property not-P
is the voi d, whi ch is not preceded by anythi ng. But no: since ( frst
moment of our procedure) we have verifed preci sel y that 0 possesses
the property P, the mi ni mal ordi nal for not-P cannot be o. Thus it
does make sense to speak of ordi nal s smal l er than it; they exist, and
must al l possess property P.
Now our central i mpl icati on, supposed proved, sai d exactly that,
i f all the ordinals smaller than a gi ven ordi nal possess property P,
then that ordi nal al so possesses it. We have reached a formal contra
diction: that the supposed mi ni mal not-P must be a P. It i s necessary
then to concl ude that this l atter does not exist and that therefore al l
ordi nal s do possess property P.
Thus the ontol ogi cal substructure of natural muti pl ici ties comes
to found the legitimacy of recurrence. Our verifcation ( the case of
0) and our demonstration ( i f P( w) for all w such that w E W, then
P( W) al so) , i f it i s possi bl e ( whi ch depends on P . . . and on our math
ematical know-how) , authorises the concl usi on for 'all ordi nal s' .
1 0. 1 5. We have remarked, in studying Peano' s axiomatic ( see 5. 3)
that reasoni ng by recurrence i s a fundamental given of serial numeri
cal ity, of which the natural whole numbers are an exampl e. It i s quite
natural that it shoul d extend to that ' universal series' composed by
the ordi nal s. But the great di fference is that, whereas in Peano the
pri nci pl e of i nducti on or recurrence is an axi omatic form or a formal
di sposi ti on, here, si nce it i s founded i n being ( i n the theory of the
pure mul ti pl e) , it i s a theorem that is, a property deducible from
the ordi nal s.
It i s of the essence of the natural mul ti pl e, which escapes al l total i s
ing thought, to submit itsel f nonetheless to that intellectual ' capture'
whi ch is the inductive schema. Here, once more, being i s found to be
amenabl e to thought i n that form of Number which is the concl usi on
for ' al l ' , proceedi ng both from the verifcation for one onl y ( here, 0)
and from a general procedure whi ch transfers the property of what
comes ' before' ( predecessor or endless series, dependi ng on whether
RECURRENCE, OR INDUCTION 89
it is a case of a successor ordi nal or a limit ordi nal ) to what comes
' after' . Number i s that which accords bei ng to thought, i n spite of
the i rremedi abl e excess of the former over the l atter.
10. 16. Reasoni ng by recurrence is a proof-procedure for universal
statements concerning ordi nal s. It al l ows us to concl ude. But there is
a more i mportant usage of recurrence, or of transfnite i nducti on, one
which al lows us to attain the concept. This is inductive defnition.
Suppose that the ai m of our thi nki ng is not to prove that thi s or
that type of mul ti pl e, for exampl e ordi nal s, has property P, but to
defne property P i n a way that would al low us then to test it on
mul ti pl es. A wel l -known di ffculty i n such a case is that we don't
know in advance whether a property defned in language is 'applica
ble' to a pure multiple without inconsistency resulting. We have seen,
for example (in 2. 1 1 ) , that the property ' not bei ng an element of
itself' does not appl y to any exi sti ng set, and that its perfect formal
correctness does not alter the fact that, handl ed wi thout care, i t l eads
to the rui n, by way of inconsi stency, of al l formal thought. But how
can we introduce l i mi tations and guarantees, i f l anguage al one cannot
support them? The procedure of defnition by recurrence, or i nductive
defnition, answers thi s question.
1 0. 1 7. What wi l l found the legitimacy of the procedure thi s ti me i s
the fact that, with the ordinal s, we have at our di sposal a sort of
universal scale, whi ch al lows us to defne property P at successive
levels, without exposing ourselves to that danger of i nconsistency that
attends on any supposi ti on of an Al l . Inductive defniti on is a ramif
cation of the concept: property P wi l l not be defned 'in general ' , but
al ways as i ndexed to a certai n level , and the operators of thi s index
ation will be the ordinal s. Here, once agai n, being comes to the aid
of fnitude, in assuri ng for our thought, which the domai n of being
as pure mul ti pl e exceeds on all sides, that it can proceed in steps, in
fragments.
10. 1 8. In conformity with the typology of ordinal s, whi ch di sti n
guishes three types ( the voi d, successors, l i mits ) , our procedure is
divided into three.
1 We frst defne explicitly, with a statement, level 0 of the pro
perty. An expl i ci t defni ti on assumes that we have a property -
say, Q - already defned, and that we can affrm that level 0 of
P - say Po - is equivalent to Q. We woul d then have: Po( x) H
Q(x) .
90 CONCEPTS: NATURAL MULTI PLICITIES
2 We then say that, if level w of P is defned, Pw, then level S( w) ,
that i s, PS( w) ' is defned through an expl ici t procedure to be i ndi
cated. To say that Pw i s defned is to say that there is a property
- cal l it R - al ready defned such that P w i s equivalent to it, so
Pw(x) H R( x) . The exi stence of an expl i ci t procedure enabl i ng us
to pass from the defni ti on of P w to that of PS(wl means that there
is a function f that assures the passage of R ( which defnes Pw) to
a property f( R) which wi l l defne PS( w) ' Fi nal l y, we can say that 'x
has the property PS( w) ' means 'x has the property f( R) ' , or that f,
whi ch permits the ' passage' from the defnition of P w to that of
PS( w) ' is an expl ici t operation on R, fxed once and for al l .
3 Final l y, we wi l l say that, if al l the level s of P below a l i mi t ordi nal
L have been defned, say: PO, Ph ,Pn, Pn+ h , then l evel L of
P, say for exampl e Pw, i s defned by a ' recol lecti on' that can be
expl i cated by that whi ch defnes al l the levels anterior to it ( in
thi s process, uni on or di ssemi nati on general l y pl ays a deci si ve
role, for reasons gi ven in 9. 1 7) . Usual l y we have something l i ke:
for a gi ven x, Pdx) i s true, i f there exi sts a level bel ow L, cal l it
w, where w E L, for which P w(x) i s true. The limit level , i n con
formity wi th its essence, wi l l assume all the i nferior levels and
wi l l not i ntroduce anythi ng new.
Thus we will have at our disposal not j ust a single concept P, but an
infnite and infnitely ramifed family of concepts, from Po, explicitly
defned, up to the more considerable ordinal indexations P w, passing
through Pn,PO' Prn, etc. We will then be able to say that concept P, as
unique concept, is defned by transfnite induction, in the following
sense: for a given x, P(x) will be true i f and only if there exists an ordinal
W such that x possesses the property at level W. We would have the
following equivalence: P(x) H 'there exists a W such that Pw(x) ' .
So the i nductive mastery of the concept passes by way of i ts ordi nal
rami fcati on, and by way of the equivalence between ' the concept P
hol ds for x' and 'the concept P hol ds for x at level W of that concept' .
Thi s equi valence avoids all mentioning of the All. It tests the property
P not ' i n general ' , but on one level , thus freeing it from paradoxes of
i nconsi stency.
1 0. 1 9. I shal l give a most i nteresting example; its i nterest i s both
i ntri nsic ( it sheds a keen l ight on the general structure of the theory
of the pure mul ti ple, or ontol ogy: it proves that, thought in their
being qua being, mul ti ples are strati fed) and methodol ogical ( we will
see clearly the functioning of levels i n the defnition of the concept) .
RECURRENCE, OR INDUCTION 91
The underl yi ng i dea i s t o defne, for each multi ple, an ontol ogi cal
rank, i ndexed on the ordi nal s, whi ch measures its ' di stance' , i n a
certai n sense, from that i ni ti al suture which is the empty set. We
could al so say that the rank is a measure of the complexity of a
set, of the immanent intrication of the i nstances of the voi d that
constitute i t.
Natural l y, it is i mpossi bl e to speak of ' al l ' sets: to do that i t woul d
be necessary to col l ect them as the elements of a set of al l sets, whi ch
would be i nconsi stent. The prudent, gradual approach of the i nduc
tive procedure i s i ndi spensabl e here.
The two i mportant operati ons of set theory which al low one to
'step' from one set to another are:
1 Uni on, or the set of elements of elements of the i ni ti al set; the
operation of di ssemi nati on, which we have al ready met ( compare
9. 1 5) . Given a set E, we denote its union by uE.
2 The set of parts, whi ch consi sts of 'gathering together' to make
one al l the parts of the i ni ti al set, al l that i s i ncl uded i n that set
(on belonging and incl usi on, see 7. 3) . We denote by pI E) the set
of the parts of E. Note that the elements of pI E) are the parts of
E: i f e E PI E) , then e C E.
We wi l l construct the hierarchy of ranks by means of these two
operations . The property we wi l l try to defne through transfnite
i nduction, according to the method expl ai ned i n 10. 18, wi l l be
denoted by R( x) , to be read as: 'x possesses a rank' ( or: 'x i s wel l
founded' ) . Our three steps wi l l be as fol l ows:
1 Explicit defnition of the property at lellel O. We propose that
R
o
(x) is not true for any x, i n other words that R
o
( x) is equi val ent
to x E O.
2 Uniform treatment of successive levels. We posi t that Rs( w) ( x) i s
true i f and onl y i f x belongs to the set of parts of the set consti
tuted by al l the whi ch sati sfy Rw. In other words, the rank at
successor level S( w) is the set of parts of the rank defned for the
level whi ch the predecessor w i ndexes. This can be written as
fol l ows: Rs( _ (x) H ( ( y E x) Rw(Y) ) : i f x sati sfes Rs( w) , the el e
ments of x sati sfy Rw, and consequently x is a part of the set of
sets whi ch sati sfy Rw. We coul d al so wri te, denoting by Rw the
set of x for whi ch Rw( x) is true:
3 Uniform treatment of limit levels. As woul d be expected, it i s
uni on that is at work here. We wi l l say that Rdx) i s true i f x is
of a rank whose index is smaller than L, that i s, i f there exists
a W E L for which Rw( x) i s true. Thus the rank RL recol lects all
the elements of the ranks below i t; it is the union of these ranks.
With the same conventions as above, we can write: (x E Rd
H x E uRw for all W smal ler than L.
Property R is thereby total l y defned by i nduction. We wi l l say that
x possesses a rank, or that R(x) ( without i ndex) i s true, i f a ( successor
or l i mi t) ordi nal W exi sts for which Rw( x) is true. This property
' means' that one arrives at the complexity of x, begi nni ng from 0
( which defnes level Ro of the property) , through the successive
empl oyment of union and of passage to the parts, an employment
whose ' l ength' i s measured by an ordi nal : the smal lest ordi nal N for
whi ch Rw( x) is true.
1 0. 20. That thi s procedure real l y ' works' , that it makes sense ul ti
matel y to speak of the property R, however, i s not sel f-evident. The
generosity of natural bei ng consi sts i n the fact that one can prove the
effectivity of thi s rami fed determi nati on of the concept. '
Thus thought proceeds in its passage through being, under the
uni versal l y intricated and hierarchised rule of Nature, which doesn' t
exi st, but provides measurabl e steps. Number is accessi bl e to us
through the l aw of such a passage, at the same ti me as it sets the
condi ti ons - as we saw with the ordi nal s - for this passage itsel f.
Number is that through whi ch being organises thought.
I I
Natural Whole Numbers
1 1 . 1 . The ordi nal s di rectly give us the Greek numbers: natural whol e
numbers. We are even i n a position to attach a new, non-Greek,
legitimacy to the adj ective ' natural ' whi ch mathematici ans, with the
symptomatic subtlety of thei r nomi nati ons, adj oi n to the civil status
of these numbers: they are ' natural s' by virtue of the fact that, wi thi n
the real m of the fni te, they coi ncide purel y and si mply wi th the
ordi nal s, whi ch consti tute the ontological schema of the pure natural
mul ti pl e.
For it i s ' natural ' to identify, i n its being, the place of number ( that
is, of whole number) - a pl ace whose exi stence Dedeki nd vai nl y tried
to secure on the basi s of the consideration of 'all the possi bl e obj ects
of my thought' - with the frst i nfnite ordi nal 0, whose exi stence we
decide, under the modern i nj unction of being, as we enounce the
axiom ' a l i mi t ordi nal exists' .
1 1 . 2. To say that 0 is the pl ace of whol e number has a precise set
theoretical meaning: what ' occupi es' the pl ace is that which bel ongs
to it. Now, not onl y do all ordi nal s that precede a given ordi nal
belong to i t; they constitute all the elements of that i ni ti al ordi nal .
I n fact, we know that total order over the ordi nal s is real l y belong
ing ( see. 8. 10) . And, consequently, an ordi nal smal ler than a given
ordi nal W i s precisely an ordinal that belongs to W. The image of an
ordi nal ( for example, one l arger than () is as fol l ows:
O e l e 2 e . . . e n e n + l e . . . e o e o + l e . . . e W
where al l the numbers in the chai n of belonging constitute precisely
the el ements of W. Vi sual ised like this, the ordi nal W appears as a
sequence of ' embedded' ordi nal s, whose ' length' is exactly W. There
are W l i nks i n the chai n i n order to arrive at W. We might also see
an ordi nal W, contai ni ng exactly W ordi nal s ( al l those that precede
i t) , as the number of that of which it is the name. Which i s another
way of saying that it is identifed with the pl ace where its predecessors
i nsi st, being the recol lecti on of that i nsi stence.
Thus the defnition of natural whole numbers is enti rely l i mpi d:
an ordi nal is a natural whol e number i f it i s an el ement of the
frst l i mit ordi nal c. In whi ch case, the structure of the pl ace of
number is:
O e l e 2 e . . . e n e n e n + l e . . . e c
But we must take care to note that c itsel f, whi ch is the name of
the pl ace, is not a part of it, si nce no set belongs to itself ( d. 8. 14) .
The place of whol e number, c, i s not an element of that place, i t is
not a whole number. As is the {rst l i mit ordi nal , it fol lows that al l
whol e numbers, except the empty set 0 of course, are successors.
1 1 . 3. An attentive reader might obj ect as fol l ows: I say that c i s the
frst l i mit ordi nal . But am I sure that a ' frst' l i mi t ordi nal exi sts ? The
Axi om of Infnity ( 9. 20) says onl y: 'a l i mit ordi nal exists' , it does not
speci fy that this ordi nal i s 'the frst' . What authorises us to call c the
' frst l i mit ordi nal ' , or frst i nfnite ordi nal ? It coul d well be that, as
soon as I announce that ' a limit ordi nal exists' , a multitude of them
appear, none of which is ' frst' . There coul d be an i nfnite descending
chai n of such ordi nal s, j ust like the descendi ng chai n of negative
numbers whi ch, it is clear, has no frst term: no whole negative
number is 'the smal lest' , j ust as no whole posi ti ve number i s 'the
l argest' (this second point i n fact comes back to saying that c, the
beyond and the pl ace of the series of positive numbers, is a l i mit
ordi nal ) .
But i f I cannot unequi vocal l y determi ne and fx the frst l i mi t
ordi nal , then what becomes of my defnition of whol e numbers ?
1 1 . 4. We can overcome thi s obj ecti on, once more, thanks to that
great pri nci pl e of natural multi pl es that i s mi ni mal i ty. We know that,
gi ven a property P, if an ordi nal exi sts that sati sfes that property,
then there is one and onl y one mi ni mal ordi nal that satisfes i t. Take
the property ' bei ng a l i mi t ordi nal ' . There certai nl y exi sts an ordi nal
that sati sfes it, si nce the Axi om of the Infnite says precisely that.
NATURAL WHOLE NUMBERS 95
Thus, there exi sts one and onl y one l i mit ordi nal that i s mi ni mal for
this property. Consequently we can speak wi thout hesitation of a ' frst
l imit ordi nal ' , or of the ' smal lest l i mi t ordi nal ' , and it i s to this uni que
ordi nal that we give the proper name 0. There i s therefore no ambi
guity i n our defnition of natural whol e numbers.
1 1 . 5. We must never lose sight of the fact that notations of the type
t ,2,n, etc. are ciphers, i n the sense of codes, whi ch serve to designate
multiples fabricated from the voi d alone. We have known for a long
time ( al ready in 8. 3) that t is in real ity the si ngleton of the voi d, that
i s, ( 0) , that two i s the pai r of the voi d and the si ngl eton of the voi d,
that i s, ( 0, ( 0) ) , that three is the tri pl et of the voi d, the si ngleton of
the void, and the pai r of the voi d and si ngleton of the voi d, that i s,
( 0, ( 0) , ( 0, ( 0) ) ) , etc. To exhi bi t further thi s weaving of the voi d with
itsel f, let's also write down the real being of the ci pher 4:
( 0, ( 0) , ( 0, ( 0) ) , ( 0, ( 0) , ( 0, ( 0) ) ) .
Evidently 4 i s a set of four el ements, i n the order 0, then ( 0) , then
( 0, ( 0) ) , then ( 0, ( 0) , ( 0, ( 0) ) ) . These four elements are none other than
zero, 1 , 2 and 3. The elements of a whole number comprise preci sel y
al l those numbers that precede it, whi ch is not surpri si ng, since we
have shown above that thi s is the i nnermost structure of every ordi nal
( 1 1. 2) . We coul d write: 4 ( 0, 1 ,2, 3 ) . And, as we have sai d, to pass
from 3 to 4 ( as from any n to n + 1 ) , we ' adj oi n' to the el ements of
3 ( or of n) the number 3 itsel f ( or the number n) . Whi ch i s not sur
pri si ng, si nce thi s is the general defnition of successi on in the ordi nal s
( 9. 6) .
I t would obvi ousl y be i mpossi bl e to use the procedure of succes
sion to 'step' from some whol e n, no matter how l arge, to the frst
limit ordi nal 0. This i s because 0, let us repeat, i s not a whol e
number, it is the pl ace of such numbers. An i mportant l aw of thought
emerges here ( one whi ch, we might say in passi ng, the Hegel i an fgure
of Absolute Knowledge, supposed to be the ' l ast' fgure of Consci ous
ness, contravenes ) , whi ch states that the place of succession does not
itself succeed.
1 1 . 6. Once we have at our di sposal the place of natural whol e
numbers, their multi ple-being whi ch weaves the voi d through the
fnite, and the law of successi on as l aw of our passage through these
numbers, we ' redi scover' the classical operations ( addi ti on and mul
ti pl ication for exampl e) through si mpl e techni cal mani pulati ons that
ari se from the general pri nci ples of i nductive defni ti on, or defni ti on
by recurrence, expl ained and legitimated on the basi s of natural bei ng
in chapter 1 0. It is time to give a new exampl e.
1 1 . 7. Take a given number, say tor exampl e 4. We want to defne
through induction a functi on F whose outcome will be as fol lows:
for any number n whatsoever ( therefore for ever whol e number, and
there i s an i nfnity of them) , F( n) is equal to the sum 4 + n. To achieve
thi s, we have at our di sposal onl y one operator: ordi nal succession,
si nce the onl y thi ng we know i s that al l the whol e numbers except 0
are successors. We wi l l proceed exactly according to the schema
expl ai ned i n 1 0. 1 8, except that we will not have to worry ourselves
about the case of l i mi t ordi nal s ( since there are none before () . We
wi l l , as before, use S( n) to denote the successor ordi nal of the whole
number n.
1 We wi l l frst state: F( O) = 4 ( an expl i ci tl y given val ue, the underly
ing intuition bei ng that 4 + 0 = 4) .
2 Then we wi l l proceed to the successi onal i nduction by positing:
F( S( n) ) = S( F( n) ) . A regul ar and uni form relation between the
val ue of the functi on for S( n) and its val ue for n, a relation that
uses only what we al ready know; the operation of succession,
defned i n general on the ordi nal s. The underl yi ng i ntui ti on i s that
4 + ( n + 1 ) = ( 4 + n) + 1, to return to the usual ' calculating' nota
ti on where the successor of n i s denoted by n + 1 .
The val ue of the function i s defned entirely by these two equa
tions. Say, for example, that I wanted to cal cul ate F( 2) . I woul d have
the fol lowi ng mechanical sequence:
F( O) = 4
F( 1 ) = ( F( S( O) ) = ( S( F( O) ) = S( 4) = 5
F( 2) = ( F( S( 1 ) ) = ( S( F( 1 ) ) = S( 5) = 6
We can see cl earl y that such a schema is a true defnition of addi
tion, through the use of recurrence, on the basi s of the operation of
successi on al one. Once we have obtained thi s general i nductive
schema of addi ti on, mul ti pl ication can be si mi l arl y defned. Take the
function to be defned, P( n) , whose val ue i s n mul ti pl ied by 4. We
begi n the induction thi s time with 1 and not with 0, stating that if
F(n) is as above ( defning 4 + n i nductively) :
P( 1 ) = 4 ( gui di ng i ntui ti on : 4 X 1 = 4)
P( S( n) ) = F( P( n) ) ( guidi ng i ntui ti on : 4 X ( n + 1 ) = 4 + ( 4 X n) )
NATURAL WHOLE NUMBERS 97
These technical manoeuvres are of no di rect interest. They serve onl y
to convince us that whol e numbers thought i n thei r bei ng ( ordinal s
that precede 0, fabricated from fnite combi nati ons of the voi d) are
i ndeed also the same ones with which we count and recount wi thout
respite, as the epoch prescri bes us to do.
1 1 . 8. The phi l osophico-mathematical reconstruction of whol e
numbers i s now complete. They do not derive from the concept
( Frege) , nor can thei r place be i nferred from our possi bl e thoughts
( Dedeki nd) , nor is thei r l aw l i mited to that of an arbitrari l y axi oma
tised operati onal feld ( Peano) . They are, rather, in the retroaction of
a deci si on on the i nfnite, that part of number whi ch being provides
to us i n its natural and fnite fgure.
The whole numbers are Nature itsel f, i n so far as it i s exposed to
thought only to the l i mited extent of its capacity for fnitude. Agai n,
thi s exposition i s possi ble onl y on condi ti on of a poi nt of i nfnity, the
l i mi t ordi nal 0, the exi stenti al guarantee of whol e number. Thi s poi nt
of infnity is i mmense i n relation to the whol e numbers, si nce, sub
tracted from successoral repeti ti on, it constitutes the place of thei r
total exercise, a pl ace wi thout i nternal l i mits ( successi on can always
continue) . Nevertheless, it is mi nute in relation to the profusion of
natural i nfnite bei ng beyond its frst term 0. Whol e number is the
form of being of the fnite ' almost nothi ng' depl oyed by being qua
being between the voi d and the frst i nfni ty.
1 1 . 9. It is onl y i n an anti ci pati on without sol i d foundati on, and in
homage to thei r antiquity, that we cal l the natural whol e numbers
' numbers' . We have al ready remarked ( 8. 8) that, sti l l without a
general concept of number at our di sposal , it woul d be i l legitimate
to say that the ordi nal s were numbers. Now, the whol e numbers are
none other than the ordi nal s. And number, or rather Number, qual i
fes a type of being of the pure mul ti pl e which exceeds the ordi nal s.
Until we have made sense of thi s type, i n such a way that it becomes
appl i cable to all species of number ( whol e, relative, rational , real ,
ordi nal , cardi nal ) , we can onl y speak of ' number' in a sense sti l l
insuffciently l i berated from its operati onal i ntui ti on, or from the
hi storical heredity of thi s si gni fer.
But our preparations are complete. The homage pai d to the Greek
numbers i s onl y the l ast act of a vast introducti on, genealogical and
then conceptual . Now it i s necessary to defne Number.
3
Ontology of Number:
Defnition, Order, Cuts, Types
1 2
The Concept of Number:
An Evental Nomination
12. 1 . The frst part of thi s book was historical and critical ( a study
of the great enterprises of the past) . The second was constructive and
conceptual ( the determi nati on of the ordi nal s as schema of natural
mul ti pl icity, on the basi s of the concept of transi ti ve sets ) . In thi s
thi rd part, we are goi ng to proceed regressi vel y, axi omatical l y: we
shall begi n with a general defnition of Number, a remarkabl y si mpl e
defnition involving onl y the concept ' ordi nal ' . Then, by way of
increasingly specifc determi nati ons, we shal l address the essenti al
attributes of the resulting concept of Number: total order, the process
of cutting, and fnal l y - i n the l ast pl ace only - operati ons . In so
doing, we shal l demonstrate how al l of our traditi onal numbers ( the
whol es, the rational s, the real s, and the ordi nal s themsel ves, con
cei ved and handled as Numbers ) are onl y particular cases of the
general concept.
In my view, the three most i mportant aspects of these proceedings
are as fol l ows:
1 Consi derations of order and operations ari se from the i ntri nsi c,
or ontol ogi cal , defni ti on of Number. Number i s therefore not
itsel f an operati onal concept, it is a particul ar fgure of the pure
mul ti ple, which can be thought in a structural and i mmanent
fashi on. The operati onal di mensi ons are only subsequent
traits. Number is not constructed; on the contrary, its very
being makes possi ble all of the constructi ons in whi ch we
engage it.
1 02 ONTOLOGY: DEFINITION, ORDER, CUTS, TYPES
2 The ordi nal s constitute the base materi al for the defnition of
Number, its natural ontol ogi cal horizon. But, taken in al l their
general ity, Numbers are ' non-natural ' deducti ons from thi s natural
materi al .
3 Our tradi ti onal numbers are onl y very specifc cases, which cer
tai nl y fal l under the general and uni fed concept of Number but
by no means exhaust it. There remai ns an i nnumerable immensity
of Numbers we have not yet thought or used.
12. 2. Defnition: A Number is the conj oint givenness of an ordinal
and a part of that ordinal.
A Number will be denoted by the letter N, fol l owed by i ndices to
disti ngui sh between several di fferent Numbers.
In other words, a number N i s constituted by:
- an ordi nal W;
- a subset F i ncl uded in thi s ordi nal , such that F e W.
The ordi nal wi l l be cal l ed the matter of Number, which we wi l l
denote by M( N) .
The part of the ordi nal wi l l be cal led the form of the Number,
whi ch we wi l l denote by F( N) .
That part of the matter whi ch i s not i n the form, that i s, those ele
ments of the ordi nal W whi ch are not i n the part F( N) , constitute the
residue of the Number. We denote this by R( N) . The resi due is equal
to the matter minus the form, and therefore to the set M( N) - F( N) .
It i s cl ear that, i f we add together the form and the residue, we
end up with the whol e of the matter. Therefore, using u to stand for
uni on ( see 9. 1 5) : F( N) u R( N) M( N) .
Si nce a Number i s entirely determi ned by its matter ( an ordinal )
and i ts form ( a part of that ordi nal ) , it wi l l often be convenient to
write it as a pai r ( M( N) , F( N) ) , with the convention that the ordinal
matter i s wri tten to the l eft, and the form to the right.
12. 3. In exami ni ng this defni ti on, the reader must observe a number
of precauti ons.
( 1) We are deal ing wi th a pure defni ti on, a priori for the moment:
it i s of no use, nor is i t possi bl e, to try to 'recogni se' straightaway, in
thi s defni ti on, any of our fami l i ar numbers.
I wi l l gi ve an exampl e: take as matter the ordi nal 1 ( whose onto
logical composi ti on i s ( 0) , the si ngleton of the void) , and, as form,
CONCEPT OF NUMBER: AN EVENTAL NOMINATION 1 03
the ordi nal 0, the void, which is of course a part of 1 , as it is a part
of every set ( 7. 9) . Using the above conventi on, we have the Number
N = ( 1 ,0) . Al l we know i s that, according to the defniti on ( the result
of an ordi nal and a part of that ordi nal ) , N is a Number. The si gns
' 1 ' and ' 0' do not di rectly refer to any Number, si nce we have not
yet even establ i shed that we are deal i ng wi th Numbers. In fact, what
these si gns 1 and 0 are goi ng to indicate here - each on i ts own
account, a matter and a form - cannot be understood as Numbers,
since we do not discern in thei r writing the fundamental dual given
ness of al l Number: a matter and a form. A Number must i nvolve
two marks, that of its matter and that of its form: now ' 1 ' is onl y
one mark, as i s ' 0' . It woul d therefore be i l l usory to ' recognise' i n the
Number ( 1 , 0) any fami l i ar number whatsoever, on the pretext that
one ' recognises' 1 or O. At the moment, we have i n ( 1 ,0) onl y an
abstract example of a Number, conformi ng to the concept of Number
gi ven i n the defnition.
Just as the prisoners in Pl ato' s cave once again make the descent
from the Idea (of Number) back down to the empi ri cal ( numbers ) ,
we wi l l demonstrate much later, i n chapter 1 6, that the Number ( 1 ,0)
is the true concept of the fami l i ar negative number -1 . But at thi s
stage it i s essenti al that the reader consider the exampl es as si mpl e
clarifcations of the defnition of Number, and not seek to reconnect
them to the cavernous empi ri cal domai n of numbers.
( 2) The matter of a Number i s an ordi nal ; we have said enough about
ordinals for there to be no mystery about thi s. On the other hand, the
form of a particul ar Number is constrained only to be a part of that
ordinal, a set included in the ordi nal . The general concept of a part (or
subset) is somewhat i ndeterminate and, when the matter happens to
be infnite, offers no foothold for intuition. In particular, note:
- that thi s part might be empty ( compare the exampl e above) ;
-
that thi s part might be the entire ordi nal ; if we take for matter
the l i mi t ordi nal 0, and for form thi s same ordi nal ( which is a
'total part' of itsel f) , we obtai n a whol l y permi ssi bl e Number
(conforming to the defniti on) , which is written N ( 0, () ; we
wi l l see i n chapter 1 6 that there are excellent reasons to all ow
that thi s Number i s none other than the ordi nal 0 itsel f, but at
the moment thi s is not at all obvi ous;
- that thi s part does not necessari l y have to be contai ned or con
nected as one; it coul d be dispersed, l acunary, composed of scat
tered elements, and so on; for example, i f we take as matter the
1 04 ONTOLOGY: DEFINITION. ORDER. CUTS. TYPES
l i mi t ordi nal 0, we can take as form the set constituted by the
whole numbers 3, 587 and 1 1 65. These three fnite ordi nal s are
al l elements of 0, and therefore, taken together, they form a part
of o. We wi l l have a quite permi ssi ble number N = ( 0, ( 3, 587, 1 1 65) ) ,
whose form has three completely separate elements.
12. 4. These formal possi bi l ities make a vi sual i sati on of Number di f
fcult. We can i magi ne spati al designs somehow l i ke thi s:
ordinal-mailer
residue
l) Number whose form i s connected
ordinal-mailer
/
residue
3)Number whose form i s void
ordinal-matter
2) Number whose form is di spersed
ordinal-mailer
/
4)Number whose form takes up the
whole of the mailer
But doubtless the si mplest way is to have recourse to a l i near
arrangement ( see below) . This fguration i s based on ordinal l inearity,
conceived as a universal series from which the being of Number i s
deducted.
A l i ne segment, whose supposed origin i s the ordi nal 0, represents
the ' ordi nal axis' . We mark with an asterisk "" upon this axis the
matter of the Number, an ordi nal W. We mark with an emboldening
of the l i ne the form of the Number, part of its matter. The rest, left
unchanged, represents the resi due. If we want to represent a particu
l ar ordinal , we can do thi s wi th a l ittle ci rcl e on the arrow, wi th the
name above or bel ow. With these conventions, a Number wi l l look
l i ke thi s:
form W ^ ordi nal -mater
*
reside ordi nal s
Once agai n, this type of drawing can aid comprehensi on, but can
also be an encumbrance. Its pri nci pal fai l ing, which it shares with the
famous ' Venn di agrams' used to teach schoolchi l dren operations on
sets ( uni on, intersecti on, etc. ) , is that it habituates one to i magi ni ng
that a part of a set is a sort of conti nuous whol e, a compact nei gh
bourhood. Now the sol e prerequi site of a part i s that it shoul d
contai n onl y elements of the set of whi ch it i s a part. These elements
might very wel l be highly di spersed, scattered to the far regi ons of
the initial set, and the vi sual schema of a part, to i ndicate this di sper
sion, must be abl e to be punctured, fragmented, dismembered. The
unfortunate thi ng is that the drawing then loses any i ntui ti ve val ue
it might have had: one si mply gets the i mpressi on that there are many
parts. In l ooki ng at my l i nes and their embol deni ngs, one must always
keep i n view, conceptual l y, that there is no reason for the form of
a Number to be a conti nuous segment, but that it coul d wel l be
di spersed throughout the ful l extent of the ordi nal-matter, as coul d
the residue.
For example, the Number mentioned above, which has for matter
the limit ordinal ( and for form the triplet ( 3, 587, 1 1 65 ) , must be
represented somehow l i ke this ( with the addi ti onal compl icati on that
the infnity of ( i s not truly ' commensurabl e' i n a drawi ng) :
3

bd

form
1 1 bb r

*

ordi nal s
12. 5 The fol l owing section i s enti rel y dedicated to a phi l osophical
elucidation of our defni ti on.
We wi l l begi n wi th the capi tal N wi th which I furni sh Number.
In al l attempts undertaken to determi ne the concept of number,
the probl ems of termi nol ogy bring the wei ght of the event to bear
upon the researcher.
Take for exampl e the appel l ation ' i rrati onal numbers' . It is truly
astoni shi ng to fnd such a designation at the heart of mathematical
rational i ty. The doctrine of 'cuts' forged by Dedeki nd i s nothi ng other
than the determi nati on - whol l y rati onal and demonstrative - of the
concept of i rrati onal number. But exactly the same coul d be said for
the theory of proporti ons in Eucl i d' s Elements. It is cl ear, then, that
' i rrational ' , i n these mathematical texts whose rati onal ity is transpar
ent, paradigmatic even, no longer has any signifcation.
We might say that what makes i tsel f known here i s a symptom of
the radical di fference between nomination and signifcation. A signi
fcati on i s al ways di stri buted through the l anguage of a situation, the
l anguage of establ i shed and transmi tted know ledges. A nomi nation,
on the other hand, emerges from the very i nabi l ity of signifcation to
fx an event, to decide upon its occurrence, at the moment when this
event - which supplements the situation with an incalcul abl e hazard
- is on the edge of its di sappearance. A nomi nation is a ' poetic' i nven
ti on, a new signi fer, which affxes to l anguage that for whi ch nothing
can prepare i t. A nomi nati on, once the event that sustains it is gone
forever, remai ns, i n the void of si gni fcati ons.
Now, at the moment of the great Greek cri si s of number, when
the arri val of that at once i nevitabl e and enigmatic event made it
known that certain rel ati onshi ps ( those, for exampl e, of the diagonal
of a square and i ts side) cannot be ' numbered' wi thi n the code of
exi sti ng numbers, the word alogos arri ved, saturating and exceedi ng
the mathemati cal si tuati on. Thi s word designates that whi ch, havi ng
no logos, nonetheless must be decided as number. It inscri bes in a
new si tuati on of thought a nomi nation wi thout si gni fcati on: that of
a number which i s not a number.
Si nce that ti me, the word has l odged itsel f, without alteration,
i n mathemati cal l anguage. It traverses transl ati ons, negl i gi bl e but
subsi stent. Our word ' i rrati onal ' i s unmi ndful of the i mport of
the nomi nati on alogos to the same extent that the word ' rational '
retai ns l ittle of the Greek logos. And, above al l , thi s nomination
has ended up taking on a univocal si gni fcati on. But the contrast
remai ns, and one can reactivate it - as I do - in between si gni fcation
and that whi ch, i n the word that i mparts it, contradicts it explicitly.
For thi s contrast i s the trace wi thi n language of a foundational
truth -event.
It can easi l y be shown that the same appl ies for ' real ' numbers,
or for ' i magi nary' numbers. Even Cantor' s reason for cal l ing the
ordi nal s begi nni ng wi th r ' transfnite' numbers becomes less and less
obscure for us now, connected as it i s to his mi ndful ness of offendi ng
the sanctity of the Infnite wi th hi s inventi on.
The frequency in number-theory of a gap between the trace of a
nomination and the sediments of si gni fcation i ndicates that the thi nk
ing of number is a true evental site: it represents in mathematics a
zone of si ngul ar precarity and sensitivity, struck regul arl y by the
excess of an event that l anguage and establ i shed know ledges consi der
destitute of signifcation, and whose desti ny can oil y be sustained by
means of a poetic and supernumerary nomi nati on.
And thi s is because number i s, amongst the forms of bei ng, that
one whi ch opens onto our thought by way of its organi sation ( see
10. 20) . Which means that everythi ng excessi ve that thought encoun
ters in number, everything that interrupts the regi me of its bei ng
by way of an evental caesura, has i mmediate di sorgani si ng effects
for thought.
12. 6. My doctrine of Number, even i f my termi nol ogy and the echo
I give i t in phi l osophi cal thought are very di fferent thi ngs, i s neverthe
less substanti al l y that of ' surreal numbers' invented by J. H. Conway
in the seventies ( see 1 . 7) .
1
I make no claim at all to having produced
anything new of a strictly mathematical order. Why, then, change
' surreal number' to j ust ' Number' , with a capital N?
I t i s basical l y a poetical di sagreement. The nomi nation proposed
by Conway seems to me rather too narrow; let's say that it belongs
to an oneiric genre ( ' surreal ' obvi ousl y suggesting ' surreal i st' ) , whereas
the excessive nature of the di scovery i n my view demands the maj estic
genre of the epic, somethi ng capabl e of conveyi ng the unanti ci pated
royal arri val of Number as such.
More techni cal l y, it seems to me that ' surreal ' remai ns caught
withi n the noti on - al l too highl y charged wi th meani ngs - of a con
tinuity through successive widenings. The adj ecti ve ' surreal ' seems to
suggest itsel f because these new numbers 'contai n' the real numbers
( as they contai n the ordinal s ) ; as i f the new space conquered was an
extension of the ol d. In hi s book, Gonshor ( see 1 . 7) , seeking to make
propaganda for the surreal s, decl ares that 'we now know the exciting
fact that the surreals form a feld contai ning both the real s and the
ordi nal s. '
2
But what i s exciting i n the discovery, at least for the phi
losopher, goes wel l beyond thi s algebraic col lecti on of real s and
ordi nal s. It rel ates rather to a complete reinterpretati on of the very
idea of number, to the possi bi l ity of fnal l y thi nki ng number as a
unifed fgure of mul ti pl e-bei ng. That real s and ordi nal s ari se wi thi n
thi s fgure i s the least of the matter, a si mple consequence. And al l
the more so given that, along with real s and ordi nal s, the mi snamed
' surreal s' contain an i nfnitel y i nfnite throng of numbers whose exi s
tence no one has conceived of before, and whi ch retroacti vel y make
our hi storical numbers seem l i ke a mi ni scul e deduction from al l those
abundant varieties of numeri cal being. To give j ust one example:
surreal numbers permit a complete doctrine not onl y of i nfnitesi mal
numbers, but of an i nfnity of i nfnitely smal l numbers, descri bi ng a
' downwards' numeri cal swarmi ng j ust as vast as that whi ch the ordi
nals descri be ' upwards' .
To us e a pol i ti cal i mage: t he nomi nati on ' surreal ' seems t o me to
be marked by that cauti on, by that attachment to ol d si gni fcations,
that characterises a certai n 'reformi st' reserve when confronted with
the event. Now, I thi nk - I wager - that we must adopt the language
of rupture here, the ' revol uti onary' l anguage. I wi l l say therefore that
what takes place here i s nothing less than the advent to our thought
of Number.
Ul ti mately the capi tal i sati on of Number does not so much di sti n
gui sh the genera from the speci es subsumed to it ( whol e numbers,
rational numbers, real numbers, ordi nal numbers, i nfnitesimal
numbers, etc. ) - al though i t does indeed activate such a distinction
as it emphasises the gap between a nomi nation ( here at last is Number)
and the di verse si gni fcati ons that, havi ng once been nomi nations
themsel ves, have become the names of numbers.
12. 7. Making thus our wager on the word Number, let us try to
legiti mi se the defni ti on: ' A Number is constituted by the conj oi nt
givenness of an ordi nal and a part of that ordi nal . '
The ordi nal s are the ontologi cal schema of the natural mul ti pl e.
An ordi nal i s a consi stent natural unity, counted for one in the onto
l ogi cal si tuati on ( set theory) . These uni ties ( i n the non-numerical
sense of the pure and si mpl e consi stency of the mul ti pl e, of the 'gath
ering together' of the mul ti pl es that consti tute it, or belong to its
presentation) provide the material of Number, that on the basis of
which there i s Number, or more preci sel y that wi thi n which Number
operates a section.
3
The si mplest way to think about this i s to con
sider that a Number extracts a form from its natural ordinal materi al ,
as a part, piece or fragment of it, a consi stent unit of thi s materi al :
an ordinal .
12. 8. Because of thei r anti quity, thei r uni versal ity, thei r si mpl icity
( which i n fact masks a formi dabl e complexity i n the detai l ) , the
natural whole numbers wi l l be our gui de. We have seen ( chapter 1 1 )
that, thought according to thei r being, natural whol e numbers are
nothi ng but a parti cul ar section in the i nfnitely i nfnite domain of
ordi nal s: the section that retai ns onl y the initial poi nt of being of thi s
domain ( the voi d) and the ' frst' successi ons, bounded external l y by
the frst l i mi t ordi nal . Or that the natural whol e numbers extract
and i sol ate, in the boundless fabric of natural mul ti pl es, onl y that
which i s fnite.
Why not continue in the same way? It i s certai nl y more rati onal
uni formly to attach the concept of Number to the ordi nal s in the
mode of a secti on, than to deploy an anarchi cal selecti on of di sparate
procedures ( algebraic, topol ogical , set-theoreti cal . . . , see 1 . 1 3 ) .
Of course, we must be sure thi s i s possible. ' Possi bl e' meani ng
what ? That in thi s way we can fnd our fami l i ar numbers. It woul d
certai nl y be arbitrary si mpl y to i mpose, in the name of ontological
si mpl icity, a concept of Number which woul d not subsume either the
rational numbers or the real numbers. But i f Number, as a secti on in
a natural multiple, defnes whole numbers as wel l as rational ( or
fracti onal ) numbers, whol e negative numbers as wel l as real numbers,
infnitesi mal s as wel l as ordi nal s, then nothi ng, i n my view, can
prevai l against both the mathemati cal unity and the phi l osophi cal
novelty of such a concept.
Moreover, the properly ontological si mpl icity of the idea of 'secti on'
confrms that our wager i s good. To say that a Number is constituted
on the one hand by an ordi nal ( which i s the signature of the Number' s
belonging to the natural form of presentation) , on the other by a part
of that ordi nal ( which is the section as ' formati on' in the natural
materi al ) i s to defne Number by putti ng to work onl y the most el e
mentary, ' basi c' categories of the ontol ogy of the mul ti pl e.
12. 9. Number wi l l then appear as the medi ati on between Nature' s
i nfnite prodigal ity of forms of being and that whi ch we are i n a
position to traverse and to measure. It is that whi ch, at least in a
l i mi ted domai n of its exi stence, accords our thought the capaci ty to
grasp and measure being qua natural bei ng. Something whi ch every
physics confrms.
12. 10. There i s no doubt that Aristotl e' s language ( Matter and Form)
is the most eloquent one for transcri bi ng the idea of Number. In
particul ar, it affords us the advantage of i nstal l i ng oursel ves withi n
materi al ist metaphors. Thi s is no negl i gi bl e advantage when we know
that, si nce Pl ato, on account of its apparent mystery, Number has
been at the heart of al l i deal i st representations of Nature. Up to, and
i ncl udi ng, what it has become under the l aw of Capital , what it is
today, as I recounted at the beginni ng of this book: the unthought
basis of the ideology of the countabl e.
I 1 0 ONTOLOGY: DEFINITION. ORDER. CUTS. TYPES
Si nce the section of Number al ways operates upon an ordi nal , it
can be sai d that, given any Number whatsoever, there always exists
an ordi nal that i s its matter. ' Matter' here has a very precise meaning.
On the one hand, the ordi nal is the ' basi s' of Number, that from
whi ch its form is sectioned. Thus it proceeds from one ordi nal , from
whi ch an extracti on is made, that there shoul d be a Number qua
pri nci pl e of thi s extracti on. On the other hand, we know that al l the
elements of an ordi nal are ordi nal s ( see 8. 5 ) . If the numerical ity of
Number, what it secti ons, its form, i s a part of an ordi nal , then, si nce
all the el ements of a part of a set are obvi ousl y el ements of that set,
that which secti ons a Number must al so be entirely composed of
ordinals. ' Matter' thi s ti me means frst matter. When we speak of the
constituents of the numeri cal secti on, we are speaki ng excl usi vel y of
ordi nal s. It is an ordi nal that i s sectioned, and the elements of the
secti on are al so ordi nal s. With regard to the categories of its matter,
Number i s natural through and through.
1 2. 1 1 . The Ari stotel i an metaphor i s easi l y extended: we say that the
product of the numeri cal secti on, i n the ordi nal that indicates its
natural provenance and furni shes its matter, i s the form of the Number.
Number itself is rather the gesture of secti oni ng, which i s why it is
represented by the pair of its matter (an ordi nal ) and its form ( a part
of that ordi nal ) . But in the form is concentrated that by vi rtue of
which Number escapes its natural prescripti on, or at least might
escape it. Because the form, being any part of an ordi nal whatsoever,
bri ngs forth, withi n a natural uni ty, a mul ti pl e whi ch i n general is
not natural .
The form i s, si mpl y, a set of ordi nal s taken from among the ele
ments of an ordi nal . This deducti on di stinguishes a part of the matter.
Now, al though every ordi nal is a set of ordi nal s ( i n fact, the set of
ordi nal s whi ch precedes it, 1 1 . 2) , not every set of ordinals is neces
sarily an ordinal. An ordi nal has no hol es; all ordi nal s that precede
it belong to it, from the voi d 0 right up to itsel f. Thi s is, moreover,
why an ordi nal is the name of its own ' length' . If, on the other hand,
you take any set whatsoever of ordi nal s, there is a good chance that
a great many ordinals will be mi ssing, that the set will be ful l of hol es.
It wi l l therefore not itsel f be an ordinal . Consequentl y, the form of
a Number i s usual l y not an ordi nal ; onl y its matter is. As might be
expected in a materi al ist phi l osophy, it i s matter that i s homogenous,
non-l acunary, regul ar, and form that is hol ey, i rregul ar, non-natural .
Wi th the form of a Number we general l y transgress the l i mits of
natural being, even i f its materi al is always extracted from within
those l i mits.
CONCEPT OF NUMBER: AN EVENTAL NOMINATION I I I
12. 12. If the form is a part extracted from an ordi nal by the section
which i s Number - a ( usual l y non-natural ) subset of a natural
set - then it leaves a remai nder; there is something l i ke the l eftover
cuttings from the scul pting of the form in the ordi nal-matter. This
remainder is made up of those elements of the i niti al ordi nal that are
not elements of the form of the Number, the portion of the matter
that is not taken up in the form. We cal l this the residue of the
Number.
Just l ike the form, the residue of a Number is a multi pl e made of
ordi nal s. And, again j ust l i ke the form, it i s usual l y not an ordi nal ( it
would be somewhat paradoxi cal if the resi due was natural ; it is so,
nevertheless, i n the specifc case where the form cuts all ordi nal s out
of the ordinal-matter without excepti on, starti ng from ordi nal W) .
The residue is obtained by the si mple difference between the Matter
and the Form.
It might be obj ected that, in that case, form and residue are
interchangeabl e. And, i n a certai n sense, that i s the case. Contem
porary art has bl i ndl y thought thi s ambiguity i n the composi ti on
of Number, by exhi biting as new works the residue of works of
art whose form i s outdated. What wi l l ultimately di scri mi nate
between the resi due and the form of a given Number, though, wi l l
relate to the l aw of order over Numbers, a l aw we shal l study i n
chapter 1 3 .
Note once more that taki ng the form and residue together - the
union of the form and the resi due - restores the matter, that i s, the
ordi nal we began with. The set-theoretical triplet of matter, form and
residue i s al l there is to the numerical secti on.
12. 1 3. Armed wi th these remarks, we can outl i ne our programme of
investigation i nto the concept of Number:

Study what it is that makes the di fference between two Numbers,
and understand the l aw of order that seri al ises them and wi thout
which we woul d not, fnite subj ects that we are, have any hope
of progressing in our knowledge of them.

Reconstitute algebra, the operati onal di mensi on of Number ( addi
tion, mul ti pl icati on, etc. ) , wi thout whi ch, constrained above al l
as we are by the ideology of the countabl e, no one woul d bel i eve
that Number i s a number. We wi l l always hol d frm to the poi nt
that the bei ng of Number precedes operations, that Number i s
above al l a thi nking, on the basi s of Nature, of a section that
extracts a form from a natural unity thi nkable as the matter of
Number.
1 1 2 ONTOLOGY: DEFINITION. ORDER. CUTS. TYPES

Fi nd agai n, in the i nfnitely i nfnite swarmi ng of Numbers, in the
i ncredi bl e prodigal ity of being in numerical form, our historical
numbers : natural whole numbers, relative whole numbers ( nega
tive numbers) , rati onal s ( fracti onal numbers ) , real s, ordi nal s . . .

Show that there exi st i nfnitely more Numbers than we can know
or can handl e, that our hi storical numericality i s most i mpover
i shed compared to the excess of being in Numbers.

Make sure, in thi s way, both that Number opens an authentic
space for thought and that this thought expl ai ns in terms of effec
tive operations onl y a mi nute part of al l the types of Numbers of
whi ch multi pl e-being - as coupled to thought by set-theoretical
ontology - i s capable.
12. 14. This programme accompl i shed, we wi l l taste the bitter joy of
Number, i n both its thi nkable and its unthi nkable aspects. Number
wi l l be entrusted to being, and we wi l l be abl e to turn ourselves
toward the numberless effects of the event.
Additional Notes on Sets of Ordi nals
Nt . The concept of Number makes central use of the concept of ' part
of an ordinal ' ; that is to say, of the concept of an arbitary set of
ordi nal s extracted from a gi ven ordi nal . Some remarks must be made
concerning the correct treatment of the notion 'set of ordi nal s' , which
incorporates that of a ' part of an ordi nal ' , si nce all the el ements of
an ordinal are ordi nal s.
N2. For a set of ordi nal s to be an ordi nal it i s necessary and suffcient,
as we have noted, that it shoul d have no holes, that no ordi nal shoul d
di srupt the chai n of bel ongi ng that bi nds the ordi nal s to each other
up to the ordi nal under consi derati on.
N3. Si nce belonging i s a total order over the ordi nal s, every set of
ordinals is totally ordered by belonging. And thi s i s the case whether
or not it has hol es. If X i s a set of ordi nal s and Xl and Xz are elements
of this set, then i t i s always the case either that Xj E Xz, or Xz E Xl ,
or Xl Xz. Thus the form and the residue of a Number are total l y
ordered by bel ongi ng, j ust as its matter i s. What makes the form and
the resi due unnatural are the holes i n them, not their order. The
uni versal intrication of natural presentation prescri bes its law to all
the components of a Number. But what subtracts most Numbers
CONCEPT OF NUMBER: AN EVENTAL NOMINATION 1 1 3
from the strictly natural domai n of bei ng resides i n the l acks that
affect their form ( and therefore thei r resi due) . A Number i s non
natural i n so far as its natural fabric i s perforated.
N4. Every set of ordinal s has a mi ni mal element. Thi s results once
more from that i mportant l aw of natural mul ti ples, the pri nci pl e of
mi ni mal ity ( see 8. 10) . Take X, a defned set of ordi nal s; and P, the
property ' belonging to X' . If there exi sts any ordi nal that possesses
the property ( it i s suffcient for thi s that X should not be empty) , then
there exists a smal lest ordi nal that possesses it. It i s this smal lest
ordi nal that is the minimal element of X: it belongs to all the ordi nal s
of X, but no ordi nal of X belongs to i t.
The exi stence of a mi ni mal el ement has nothi ng to do wi th whether
or not the set has any hol es. Therefore one can al ways speak of the
mi ni mal element of the form of a Number, or of the mi ni mal element
of its resi due. As to the mi ni mal element of its matter, thi s i s always
the empty set 0, si nce the matter is an ordinal .
N5. We must be very careful on the other hand to observe that a
given set of ordinals does not always have a maximal element. We
have al ready seen that a l i mit ordi nal ( which is a set of ordi nal s) has
no maxi mal element ( 9. 1 4) . A fortiori any set whatsoever of ordi nal s
may very wel l be i nfnitely ' open' , contai ni ng no el ement that domi
nates al l the others.
N6. However, there al ways exi sts an upper bound of a set X of
ordi nal s. By ' upper bound' we understand the smal lest ordi nal to be
l arger than every ordinal in X. Here agai n, the existence of an upper
bound is guaranteed by the pri nci pl e of mi ni mal ity. Let P be the
property ' being l arger than al l the ordi nal s that belong to X' . There
certai nl y exists an ordi nal that possesses thi s property, unless X is
equivalent to the set of all the ordi nal s, which i s i nconsi stent. There
fore there exists a smal lest ordi nal whi ch possesses the property P; it
is the smal lest ordi nal to be l arger than al l the elements of X, and
thus i t is the upper bound of X. We wi l l denote thi s upper bound by
sup( X) .
N7. If a proper part of an ordi nal ( a part whi ch is not t he ordi nal
itsel f, a truly partial part) is an ordi nal , then it belongs to the i ni ti al
ordinal .
We have known for a long time that the converse pri nci ple i s part
of the defni ti on of ordi nal s. They are transitive, and so every ordi nal
that belongs to an ordi nal is al so a part of it. We now want to show
1 1 4 ONTOLOGY: DEFINITION, ORDER, CUTS, TYPES
that every ordinal which is a proper part of an ordi nal belongs to it.
Thi s comes back to sayi ng that, between ordi nal s, the order of belong
ing is equivalent to the order of inclusion.
Suppose an ordi nal WI is a part of ordi nal W2: WI C W2 Si nce
bel ongi ng is a total order over the ordi nal s, and si nce WI is di fferent
from W2 ( it is a proper part of W2 ) , there are two possi bi l i ties:
1 Either WI E W2, and the theorem is true, the ordi nal WI which
i s i ncl uded i n W 2 belongs to it, the part i s al so an element.
2 Or W2 E WI ' But, since WI is transitive, that woul d mean that
W2 C WI ' Now we know that WI C W2 If one set is i ncl uded in
another, and the other i ncl uded i n it, then they are equal , as is
i ntuitively obvi ous, and as the reader can prove in one line. Now,
WI cannot be equal to W2, since it i s a proper part of it. Thus the
frst case must hol d, and the theorem i s proved.
So it is the same thi ng, when deal i ng with ordi nal s, to say that one
belongs to the other, and to say that one i s i ncl uded in the other. In
other words, i f an ordi nal represents (as a part) another ordinal , then
it also presents it (as an element) . Which does not prevent an ordi nal
from havi ng some parts which are not elements. These parts wi l l
si mpl y not be ordi nal s ei ther. This woul d be the case, for example,
with holey, l acunary sets, sets which begi n i n the middle of an ordi nal
chai n or onl y present separated elements, etc. In fact it i s general l y
the case wi th the form of a Number.
If, however, the form of a Number i s an ordi nal , then it fol lows
from the preceding arguments that not only i s it a part of the matter
(the i ni ti al ordi nal ) , but al so an element of it. Then the form is of a
pecul i ar ki nd, l i ke a ' corpuscl e' of matter. In such cases, Number is
less a representati on extracted from Nature than a si mpl e natural
presentati on.
1 3
Diference and Order
of Numbers
1 3. 1 . A Number is enti rel y determined by its matter ( the ordi nal
from whi ch i ts form is extracted) and i ts form. The resi due i s obtai ned
by taking the di fference between the matter and the form. Because
of this, it i s often convenient, as we have said, to wri te a Number i n
the form N ( W, F( N) ) , where W is the ordi nal-matter and F( N) the
form. The resi due R( N) , i s equal to W - F( N) .
Gi ven these conditions, how can we thi nk the difference between
two Numbers ? It is natural to posi t that they are identical i f they
have the same matter and the same form. I f they are not identi cal ,
thi s coul d be:

because they do not have the same matter. Take WI the ordi nal
matter of one, and Wz the ordi nal-matter of the other. Two ordi nal s
l i ke WI and Wz are ordered by belonging, but al so, as we have seen
( N7), by i ncl usi on: ei ther WI C Wz, or Wz C WI ' Thus we can say
that in this case what di fferentiates WI from Wz i s the set Wz - WI ,
or WI - W2 Since al l elements of an ordi nal are ordinal s, we can al so
say that what di fferenti ates WI from Wz - and thus the numbers NI
and N2, of which these ordi nal s are the matters - are the ordi nals
which are elements of W2 but not of WI ( i f WI C W2 ) or elements of
WI but not of W2 ( i f Wz C Wd;

because, having the same matter, they do not have the same form.
In thi s case, there are elements ( and therefore ordinal s, since the frst
matter of a Number i s composed of ordinals i n its three components,
matter, form and resi due) that are in the form of one but not in the
form of the other. But, si nce the matter is the same, every element of
the form of one whi ch is not in the form of the other is in its residue:
i f W E F( N d and W E F( Nz ) , then W E R( Nz ) . What di fferentiates
the two Numbers Nt and Nz is the set of ordi nal s that are i n the form
of one and in the residue of the other.
We can see then that the di fference between two Numbers can be
understood in terms of ordi nal s. If an ordi nal is in the matter of one
and not i n that of the other, or i f it i s i n the form of one and in
the resi due of the other, it makes a difference between the two
Numbers.
1 3. 2. Take any two Numbers whatsoever. We wi l l say that an ordi nal
w discriminates between these two Numbers i f it i s i n the matter of
one and not i n that of the other, or i f it is i n the form of one and i n
the resi due of the other ( whi ch i mpl i es that it i s in the matter of both,
si nce form and residue are both parts of matter) .
1 3. 3. Let' s take an exampl e: Let Nt be the Number ( 2, ( 0) ) whose
ordi nal-matter is 2 and whose form i s ( 0) . It is certai nl y a Number,
si nce 2 is an ordi nal (it is the fnite ordi nal whose being i s ( 0, ( 0) ) , see
1 1 . 5 ) and the si ngleton of 0, denoted by ( 0) , i s a part of that ordi nal
( 7. 1 1 ) . Thi s Number Nt has, for matter, the ordi nal 2, and, for form,
the part ( 0) .
Now let N2 be the Number ( 0, 2) . Once again i t i s a Number,
si nce 0 i s an ordi nal (the frst l i mi t ordinal ) and the ordi nal 2,
whi ch i s an element of 0, is al so a part of it ( transi ti vity of ordinal s) .
Thi s Number N2 has for i ts matter 0 and for i ts form the part
( 0, ( 0) ) 2.
The ordi nal 0 does not discriminate between t he Numbers Nt and
Nz Indeed, 0 i s certai nl y not in the matter of Nt ( which i s 2, a fnite
successor ordi nal ) , but nei ther is it in the matter of N2, because thi s
matter i s 0, and we know that no set bel ongs to itsel f: it cannot be
that 0 E 0.
The ordi nal 0 ( the empty set) does not discriminate between the
Numbers Nt and Nz either. In fact, it i s in the form of both. The
form of Nt i s the si ngleton ( 0) , of whi ch 0 is the only element. So 0
is an element of thi s form. And, on the other hand, 0 i s an element
of the ordi nal 2, which i s the form of Nz Thus 0 i s also i n the form
of N2
However, the ordi nal ( 0) ( whi ch is the whol e number 1 ) does
discriminate between the Numbers Nt and Nz: ( 0) is an element of
DIFFERENCE AND ORDER OF NUMBERS 1 1 7
the ordi nal 2, and thus belongs to the form of Nz But it cannot
belong to the form of N" whi ch is preci sel y ( 0) , si nce the sel f-belong
i ng (0) E (0) is i mpossi bl e. Given that ( 0) is an element of the matter
of N, ( which is the ordi nal 2) , since it i s not in its form, it must be
in its residue.
13. 4. Gi ven two Numbers and any ordinal whatsoever, it is al ways
possi ble to say whether thi s ordi nal di scri mi nates between the two
Numbers or not. I f N, and Nz are Numbers, the property ' di scri mi
nati ng between N, and N2' is wel l -defned.
But if there is an ordi nal that di scri mi nates between N, and Nz
( that i s, i f N, and N2 are di fferent) , then i n vi rtue of the pri nci pl e of
mi ni mal ity - whi ch we have constantl y made use of because i t i s a
fundamental law of natural mul ti pl es ( see 8. 10) - there is one uni que
smal lest ordi nal whi ch di scri mi nates between them. Or, i f you l i ke,
a mi ni mal ordi nal for the property ' di scri mi nating between the
Numbers N, and Nz' .
13. 5. An extremely important defnition: The smallest ordinal to
discriminate between two Numbers is called their discrminant.
The interesting thi ng about the concept of di scri mi nant is the fol
l owi ng: it bri ngs the i dea of the di fference between two Numbers
down to a matter of one single ordinal. Thi s ' mi ni mal point' of
di fferentiation al lows a local rather than gl obal treatment of the
comparison between two Numbers. The exi stence of a di scri mi nant
suffces for us to concl ude that two Numbers are di fferent.
1 3. 6. One more exampl e. Take the two Numbers N, and Nz from
the example above ( 1 3. 3) , N, = ( 2, ( 0) ) and Nz = ( 0,2) . What i s thei r
discriminant ?

We have seen that 0 does not discri mi nate between N, and Nz

We have al so seen that ( 0) di scri mi nates between them. Since the
onl y ordi nal smaller than ( 0) i s 0, whi ch does not di scri mi nate
between N, and Nz, ( 0) i s defnitely the smal lest ordi nal that
discri mi nates between them. We therefore say that ( 0) i s the di s
cri mi nant of the Numbers ( 2, ( 0) ) and ( 0,2) .
Note the l ocation of the di scri mi nant: i t i s i n the matter of the
two Numbers N, and Nz, but is i n the form of Nz and i n the residue
of N, .
Now consi der the fol lowing two Numbers ( S( W) denotes the
successor of the ordi nal W, see 9. 5 ) :
I 1 8 ONTOLOGY: DEFINITION. ORDER. CUTS. TYPES

N3 ( S( r) , r) . Its matter is the ordi nal S( r) , its form r itself. The
l atter i s a part of S( r) , si nce every ordi nal i s an element of
its successor, and every element of an ordi nal i s a part of it
( transi ti vi ty) .

N4 ( S( S( r) ) ,r) . Its matter is the successor of the successor of r,
its form is ro
What i s the di scri mi nant of N3 and N4 ? These two Numbers have
the same form, that is r, but extracted from di fferent matters, S( r)
and S( S( r) ) . In summary, in these numhers everythi ng is exactly the
same up to the ordinal S(r) . Thi s ordi nal i s i n the matter of N4, si nce
S( r) E S( S( r) ) , but it is not i n the matter of N
3
, si nce S( r) E S( r) .
Thus the ordi nal S( r) is the smal lest ordi nal to make a di fference
between N
3
and N4; it i s the di scri mi nant of these two Numbers.
Note once agai n the location of the di scri mi nant of N. and N
4
:
S( r) is not in the matter of N3, but is in that of N4 Meanwhi le,
this time it is not in the form of N4, which i s ro It i s therefore i n
its resi due.
The comhi nati on between the ordi nal punctual ity of the
di scri mi nant and its location in the Numbers compared wi l l give
us the key to the concept of order in the boundless domain of
Numbers.
1 3. 7. Let' s give al l of thi s a sl ightl y stricter form.
The location of an ordinal w with regard to Number N, written
L( w,N) , is the posi ti on that it occupies with regard to the three
di mensi ons of the numerical section carried out by the Number N:
matter, form, resi due. There are obvi ousl y three locations:
1 Ei ther the ordi nal w is not an element of the ordi nal W which is
the matter of the Number N. In thi s case we say that it is located
' outsi de the matter' and we posi t that: L( w, N) oM( N) .
2 Or the ordi nal w is in the matter W and belongs to the form of
the Number. We then posi t L( w,N) F( N) .
3 Or the ordi nal w is in the matter W, but belongs to the resi due
of the Number: We then posi t L( w, N) R( N) .
When there is no ambiguity as t o the number N in question, we
mi ght si mpl y use the notation L( w) R, si gni fyi ng that the location
of w ( for the number i n questi on, of course) i s its belonging to the
resi due (of that number) .
Given a number N, every ordi nal can be l ocated for N so long as
we al l ow t he location ' outsi de t he matter' .
DIFFERENCE AND ORDER OF NUMBERS 1 1 9
When an ordi nal di scri mi nates between two numbers Nt and N2
( see 1 3. 2) , it is very si mpl y because its location in the two Numbers
is not the same. The table of possi bl e locations for an ordi nal W which
discriminates between the two Numbers i s as fol l ows ( usi ng oM, F
and R to denote the locations) :
L ( w,Nt l L ( w, Nz)
F R
F oM
R F
R oM
oM F
oM R
The discri mi nant of Nt and N2, being the smal lest ordi nal to di s
criminate between them, necessari l y responds to one of the ' pai rs' of
locati ons i ndicated i n the tabl e. For exampl e, i f it is i n the resi due of
Nt . it must be i n the form of Nz or outside the matter of N2, etc.
13. 8. Defnition of order over Numbers
Take two Numbers Nt and N2 and thei r discri mi nant w ( i f neces
sary, reread 1 3. 4-1 3. 6, given that the concept of di scri mi nant is
central ) .
We say that Nt is smal l er t han N2, written Nt < N2 , i f t he l ocati on
of the discri mi nant w for the Numbers Nt and N2 satisfes one of the
three foll owi ng cases:
1 Either L( w,Nd = R( Nd, and L( w,N2 ) = F( N2 ) : the discriminant
is in the residue of NJ and in the form of N2
2 Or L( w,N t l = oM( N t l , and L( w,N2 ) = F( N2 ) : the discriminant is
outside the matter of NJ and in the form of N2
3 Or L ( w,Nt l = R( Nt l and L( w,N2 ) = oM( N2 ) : the discriminant is
in the residue of NJ and outside the matter of N2
Compare these three cases careful l y with what the tabl e in 1 3. 7
indicates as t o the possi bl e locations of the di scri mi nant of two
Numbers.
1 3. 9. It is not i mmediately evident that the rel ati on Nt < N2 i s one
of order. But, even before establ i shing that this i s the case, we can
reveal the characteristics of thi s rel ati on.
The di scri mi nant gathers into one poi nt ( one ordinal ) the concept
of di fference between two Numbers. The order introduced here
depends on the location of thi s poi nt, and therefore on a sort of
topology of difference. Since, i n the gesture of secti oni ng that consti
tutes every Number N, the ' posi ti ve' numericality - that whi ch thi s
gesture extracts from matter - is the form, we wi l l always consider
that, if the di scri mi nant of two Numbers i s i n the form of one, this
number i s ' l arger' than the other. In the other, of course, the discri mi
nant wi l l ei ther be i n the resi due or outside the matter.
Conversel y, the resi due of a Number is the purely passive resul t of
the secti on of its form, the unintenti onal remai nder of the numeric
gesture. It i s that which Number as gesture leaves to matter. I f the
di scri mi nant of two Numbers is i n the resi due of one of them, we
wi l l al ways consider this Number to be ' smal ler' than the other.
In the other, the di scri mi nant wi l l be i n the form, or outside the
matter.
1 3. 10. An apparently paradoxi cal consequence of thi s conception,
whi ch determi nes all order on the basi s of the active superiority
of form - thought as the numeri cal ity of Number - over residue -
thought as passi ve i nverse - is that a number NI is sai d to be smal l er
than Nz i f thei r di scri mi nant is i n the resi due of N I and outside the
matter of N2 The ' paradox' results from the assumption that the
posi ti on ' outside matter' i s completely unaffected by the numerical
gesture, being neither in its form nor i n its resi due. Isn't it even more
passi ve then, even less i nvol ved in the numerical extraction of the
form, than an ordi nal which is i n the residue, and which therefore at
least fgures in the matter of the Number? Isn't the l ocation oM a
fgure of nothing in rel ation to Numher, an ontological l y ' i nferior'
fgure to the passive fgure of the resi due ?
1 3. 1 1 . Thi s sense of ' paradox' mi sses an essenti al point, whi ch i s
that the 'outside matter' position includes the matter itself, si nce an
ordinal W i s not an element of itsel f. There i s no reason to suppose
that the matter i s ' i ndi fferent' to the gesture of Number: it is its
pri mary ' gi ven' , that on the basi s of which there i s Number; the
natural mul ti pl e whose being is exposed to the numerical secti on.
And it is always the matter itsel f of one of the two Numbers that is
at stake when the di scri mi nant i s located ' oM' for one of them.
I f the di scri mi nant of NI and Nz i s, say, outside the matter of Nz,
then it i s i n the matter of N I ( i n its form, or in its residue) . I f not - i f
it were outside the matter of both Numbers - it coul d not discrimi
nate between them. Therefore the di scri mi nant must indeed be the
DIFFERENCE AND ORDER OF NUMBERS 1 21
smal lest ordi nal in the matter of N, and outside the matter of N2.
Evidently thi s means that the ordinal WI which is the matter of N,
is larger than the ordinal W2 which is the matter of N2 If not, there
could be no ordi nal in W, that was not in Wz, si nce the elements of
an ordi nal are al l the ordi nal s that precede i t ( see 1 1 . 2) . Thi s means
that W2 E W, ( the order-rel ati on over the ordi nals i s belonging) . But
Wz itsel f i s the smal lest ordinal that does not belong to Wz, si nce al l
the ordi nal s smal ler than Wz are, preci sel y, the elements of Wz . And
so, ultimately, Wz i s i n W, and i s the smal lest ordi nal not to be i n
Wz. It i s the smal lest ordi nal to be in W, ( the matter of Nt l , and not
i n Wz ( the matter of N2 ) , and therefore outsi de the matter of N2 . Wz,
the matter of Nz, is the discriminant of NI and N2
Thi s demonstration has a general val i dity: whenever we say that
N, i s ' smal ler' than N2, or that N, < N2, because the di scri mi nant of
N, and N2 is in the resi due of N, and outside the matter of N2, this
also means that the discriminant in question is the ordinal-matter of
N2 And this relation is legitimate because the matter of a Number,
the one-ordi nal in which the numerical section operates, i s a pri mor
dial donation of bei ng ontological l y superi or to the passi vity of the
residue.
It is therefore phi losophi cal l y wel l -founded to put the locations
in the order R < oM < F: the form, affrmative numerical i ty of the
secti on, i s superior to what i s outsi de matter, which i s i tsel f superior
to the passivity of the residue because i n real ity thi s ' outside of
matter' is the matter itsel f, i ntegral l y counted for one as an ordi nal .
The relation N, < N2 founded on otr three cases ( the di scri mi nant
in R( Nt l and F( N2) ; the discriminant in R( Nd and oM( N2 ) ; the di s
cri mi nant in oM( Nd and i n F( Nz ) ) descri bes a hi erarchy, founded i n
the being of Number as the sectioni ng of a form in natural matter.
13. 12. What remai ns now is to establ i sh that the relation N, < Nz i s
trul y an order-relati on, in the mathemati cal sense of the term: that it
serialises Numbers. Thi s amounts to respondi ng positively to three
questi ons:
1 Is the relation total? ' Or: given any two di fferent Numbers N,
and Nz, i s i t always the case that N, < Nz or N2 < N ?
2 Is the relation non-refexive? Or: is it i mpossi bl e that Nt < Nt ?
3 Is the relation transitive? Or: from the rel ati ons Nt < N2 and
Nz 5 N3 does it necessari l y fol l ow that N, < N
3
?
If we prove these three poi nts, we wi l l have brought the
phi l osophi cal legitimacy establ i shed i n 1 3. 1 1 to coi nci de with
mathemati cal ( ontological ) legi ti macy. Or, rather, with regard to the
order of Numbers, we wi l l have obtained the si tuation in which we
have been conti nuousl y striving to remai n: where what is sai d under
the sign of the phi losophi cal statement ' mathematics i s ontology'
remai ns i n harmony with what i s said under the sign of mathematical
i nferences themsel ves. Or where the interpretation of mathematics as
science of being qua being draws its contact with the real from the
effective thoughts of such a science.
1 3. 1 3. The relation < is total
Look one l ast time at the tabl e of cases of i nequal ity for N 1 < N2
Thi s tabl e fxes the l ocation of the discri mi nant of Nl and N2
Case 1
Case 2
Case 3
R
oM
R
F
F
oM
To demonstrate that the relation is total is to show that, given two
di fferent Numbers, one of them i s al ways ' smal ler' than the other.
Take two randomly selected numbers N3 and N4, and N the
ordinal which i s their di scri mi nant. Three cases are possi bl e:
1 The di scri mi nant N is in the resi due of N
3
. Then:
( a) ei ther i t i s i n the form of N4, and ( see the tabl e) N
3
< N4;
( b) or i t i s outside the matter of N4, and ( idem) N3 < N
4.
2 The di scri mi nant N is in the form of N3 . Then:
( a) either i t i s i n the residue of N4, and ( idem) N4 < N
3
;
( b) or i t i s outside the matter of N4, and ( idem) N4 < N
3
.
3 The di scri mi nant is outside the matter of N
3
. Then:
( a) either i t i s i n the form of N4, and ( idem) N
3
< N4;
( b) or i t i s i n the residue of N4, and ( idem) N4 < N3 .
Havi ng exhausti vel y enumerated al l the possi bl e cases, we see that
the relation < between Numbers N3 and N4 is al ways defned. The
rel ati on real l y is total i n the domai n of Numbers, there are no two
di fferent Numbers not related by <.
1 3. 14. It is good to get into the habit of thi nking through the inequal
ities between Numbers more rapidly. For exampl e we coul d say: if
the di scri mi nant N i s in the resi due of one of the two Numbers, the
table (cases 1 and 3, whi ch are the onl y possi bi l ities ) shows that thi s
Number i s smal l er than the other. I f N is i n the form of one of the
two Numbers, the tabl e (cases 1 and 2, the onl y possi bi l ities ) shows
that thi s number i s l arger than the other. Apparently we have l eft to
one si de the case where N i s outsi de the matter of one of the Numbers.
Not so, because then it woul d necessari l y be in the residue or the
form of the other ( i f it was outside the matter of both, it would not
discriminate between them) , and we are referred back to one of the
preceding cases.
To compare two numbers according to the relation < we therefore
proceed as fol lows: frstly we check whether the di scri mi nant i s in
the form of one of them: if so, we concl ude i mmedi atel y that thi s
Number is the l argest. I f not, we check whether it i s in the resi due
of one of the two: if so, we conclude that that Number is the smal l
est. The work is done, no other case i s possi bl e.
1 3. 1 5. The relation < is irefexive
This point is tri vi al . It cannot be that N 1 < Nt . si nce the relation i s
founded on the existence and l ocati on of a discri mi nant, whi ch cannot
exi st ' between' N, and i tsel f.
1 3. 1 6. To exerci se ourselves in the compari son of Numbers usi ng the
< relation, let' s take up the examples from 1 3. 6 once more. We had,
adopting the notation by the pai r of matter and form, the four
fol l owing Numbers:
N, ( 2, ( 0) )
N2 ( 0, 2)
N3 ( S( o) ,o)
N
4
( S( S( o) ) ,o)
The di scri mi nant of N, and N2 i s ( 0) . It i s in the resi due of N"
and i n the form of N2. S o N, < N2 .
The discriminant of N3 and N
4
is S( o) . It is outside the matter of
N3, and in the resi due of N
4
. So N
4
< N3.
The di scri mi nant of N3 and N, is ( 0) , whi ch i s in the residue of N,
and i n the form of N3. S o N, < N3 .
The discri mi nant of N
4
and N2 is 2 ( why? ) . It i s i n the resi due of
N2 and i n the form of N
4
. So N2 < N
4
.
The reader can study the remai ni ng compari sons on thei r own
account.
It will be remarked that it i s not simply because the matter of a
Number is ' larger' that that Number is l arger. Thus N
4
has for
ordi nal-matter the successor of the successor of 0, which is l arger
than the successor of 0, the matter of N
3
. Nevertheless, N4 < N .
What is sti l l more remarkabl e i n t hi s exampl e is that the form of
N
3
and of N4 i s the same ( it i s () . Thus we have t he fol lowing ' l aw' :
if the form stays the same whilst the matter grows, the Number gets
smaller. It is quite straightforward to demonstrate the general
case. Take a Number NI ( W"X) and a Number Nz ( Wz, X) , where
WI E Wz; and X the same set of ordi nal s ( which is a common part
of WI and Wz ) . The di scri mi nant of these two Numbers cannot be
found in the form of ei ther of them, si nce they have the same form,
X: an element of X cannot di scri mi nate between them (it even has a
location for N I and Nz, namel y the form) . It is therefore in the resi due
of one, and outside the matter of the other. It is cl ear that thi s
di scri mi nant i s none other than WI , which i s the smal lest ordi nal not
to belong to WI , and which is in Wz, since WI E W2 Now WI is
necessari l y i n the resi due of Nz ( since it belongs to its matter, but not
to X, its form) , and outside the matter of NI . Therefore it is indeed
the case that Nz < NI .
Thi s process suggests a compari son between the Number ( W, X)
and the rel ati on . We know that such a relation diminishes when
its denomi nator W grows. But be warned: this i s only a di stant
analogy, because means nothi ng here.
Al l the same we can show, i nspi red by thi s anal ogy, that, if the
matter remains the same whilst the form gets larger - so that the old
form X is included in the new form x' - then the Number gets larger.
Thi s ti me, it is the enl argement of the ' numerator' that enl arges the
' rel ati on' . I leave the detai l s of the demonstration of this to the reader.
Suffce to say that the discriminant i s the smal lest ordi nal to belong
to X' and not to X; so it i s i n the form of the second Number and
i n the residue of the frst, and therefore the second i s l arger.
These observations are phi l osophi cal l y wel l -founded. What does
it mean, i n fact, to produce the same form from a l arger matter? That
the gesture of the numeri cal section did not manage to extract from
a vast matter ( that of a l arger ordi nal ) any more of a form than could
have been obtained with a smal ler matter. The gesture was thus less
concentrated, less elegant, less effecti ve. It i s qui te l egiti mate that the
Number which marks this gesture shoul d be hel d for i nferior. The
converse also fol l ows: to obtain a more widely depl oyed form, con
tai ni ng all the elements of the frst and more, with the same i ni ti al
matter, requi res a more effcient gesture of secti oni ng. It is quite right
that this shoul d be marked by a superi or Number.
The rel ati on < does i ndeed express i n the mathematical feld the
ontological ly rati onal di sposi ti ons of the compari son of Numbers.
1 3. 1 7. The relation < is tansitive
This is a question of proving that, given three Numbers, Nb N2 and
NJ, i f NJ < N2 and N2 < N
3
, then NJ < NJ . Obvi ousl y everything
hi nges on the location of the discriminants. We shall write the
di scri mi nant of NJ and N2 as w( 1 ,2) , that of N2 and NJ as w( 2, 3 ) ,
and that of NJ and NJ as w( 1 , 3) .
( a ) Fi rst step. An ordi nal smal ler than w( 1 ,2) and w( 2, 3 ) does not
di scri mi nate between NJ and NJ.
The di scri mi nant of two Numbers i s the smallest ordi nal that di s
cri mi nates between those two Numbers ( i n the order of ordinal s,
whi ch is belonging) .
If an ordi nal W is smal l er than w( 1 ,2) , i t doesn' t di scri mi nate
between Nl and N2 . Its location ( F, R, or oM) i s the same i n NJ and
i n N2. Equal l y, i f it i s smal l er than w( 2, 3 ) , it doesn' t di scri mi nate
between N2 or NJ either - its location i s the same i n N2 and NJ.
Ultimately, therefore, i ts location must be the same i n Nb i n N2 and
in N3 , and it does not discri mi nate between NJ and N3
( b) Concl usi on of the frst step: w( t , 3) , whi ch obvi ousl y di scri
mi nates between N
l
and NJ, cannot be smal l er than w( 1 ,2)
and w( 2, 3 ) . I t i s therefore at least equal to the smaller of the two.
( c) Second step. The smal lest of the two ordi nal s w( 1 ,2) and w( 2, 3 )
di scri minates between NJ and NJ .
For convenience of exposi ti on, we wi l l suppose that the smal lest
is w( 1 ,2) ( the reasoni ng would be exactly the same i f it was w( 2, 3 ) ;
confrming thi s woul d be an excel l ent exerci se for the reader) . Si nce
w( 1 ,2) di scri mi nates between NJ and N2 , its l ocation i n NJ i s di fferent
from its location in N2 . But, since it is smal l er than w( 2, 3 ) , it does
not discri mi nate between N2 and N
3
, since w( 2, 3 ) - di scri mi nant of
N2 and N
3
- is the smal lest ordi nal that di scri mi nates between these
two Numbers. Therefore the l ocation of w( 1 ,2) in N2 and NJ i s the
same. Si nce its location in N2 di ffers from that in N b if it i s the same
in N
3
as i n N2, i ts l ocati on i n NJ al so di ffers from its l ocation i n N
1

So w( 1 ,2) discri mi nates between Nl and NJ .
( d) Third step. w( 1 , 3) , the di scri mi nant of N
l
and NJ, is actual l y
equal to the smallest of the ordinals w(1 ,2) and w(2, 3).
We have seen that w( 1 , 3) must be at l east equal to the smal lest of
the two ordi nal s w( 1 ,2) and w( 2, 3 ) ( frst step) . We supposed w( 1 ,2)
t o be the smal lest. Thus w( t ,3 ) i s at least equal t o w( 1 ,2) . Now w( 1 , 2)
di scrimi nates between N
l
and N3 ( second step) . Si nce w( 1 , 3 ) i s the
di scri mi nant of N, and N
3
, and thus the smal lest ordi nal to discrimi
nate between them, and since i t cannot be smaller than w( l ,2) , which
di scri mi nates between N, and N2, it is equal to w(1 , 2) . So w( l , 3) =
w( l ,2) .
( e) An aside: i f we were t o suppose the opposite hypothesis, that
w( 2, 3 ) i s smal ler, we would fnd that w( l , 3 ) = w( 2, 3 ) , for the same
reasons.
( f) Fourth step, the concl usive step. We have discovered that w( l ,3 )
= w( l ,2) . Thi s can be expressed a s fol lows: w( 1 ,2) , discriminant of
N, and N2 , i s al so the discriminant of N, and N3.
Now, we know that N, < N2 . So we know there are two possible
locations for the discriminant w( l ,2) in N" the smal ler of the two
Numbers:
1 Either w( l ,2) is in the resi due of Nt . But then, si nce it is al so the
di scri mi nant of Nt and N
3
, its posi ti on in the resi due of Nt leads
us to concl ude that Nt < N
3
(on this poi nt, see 13. 14) .
2 Or w( l ,2) is outside the matter of Nt . It must then be in the form
of N2, for the usual reason that Nt < N2 . But w( l , 2) , which is
smal l er than w( 2, 3 ) , does not di scri mi nate between N2 and N3.
Therefore it i s al so i n the form of N
3
. And, si nce it i s the di scri mi
nant of N, and N
3
, bei ng outsi de the matter of Nt and in the form
of N
3
, once agai n Nt < N3 .
So we have proved that, i f Nt < N2 and N2 < N
3
, then N, < N3.
We have even discovered, as a bonus, a sti l l fner resul t: the discri mi
nant of Nt and N
3
i s equal to the smal lest of the di scri mi nants of Nt
and N2, and of N2 and N3.
13. 18. Di alogue wi th a tenaci ous reader, on the subj ect of the preced
ing demonstration:
THE READER: You suppose from start to fni sh that the di scri mi nant
of N, and N3 exists. It' s not so obvi ous. I can well see that the dis
cri mi nant of Nt and N2 exi sts, si nce we know that N, < N2 . The same
for that of N2 and N
3
. But it could wel l be that i n the end N, NJ ,
and in that case there woul d be no di scri mi nant w( l , 3) . The relation
woul d be ci rcul ar: Nt < N2 < N, .
ME: But that' s absurd: I f the di scri mi nant of N, and N2 i s l ocated in
N, and N2 i n such a way that N, < N2 , it cannot be the case that
N2 < N, . Therefore N, is necessari l y di fferent from N
3
, and thei r
di scri mi nant exists.
READER: Okay, you' ve got me. But I' m sti l l not sati sfed. In your
second step, you suppose that one of the two di scri mi nants
w( 1 ,2) and w( 2, 3 ) is the smal l er of the two. But surel y i t coul d quite
easily be the case that there i s no smal lest of the two; for thi s to be
the case i t suffces that they be equal. And, so that you don' t try to
pul l the wool over my eyes, I ' l l gi ve an exampl e. Take these three
numbers:
1 N, is the number ( 2, ( 1 ) ) , which has for its matter 2 and for its
form the si ngleton of 1 . I know ( I' ve read you saying so j ust now)
that the ordi nal 1 is an el ement of the ordi nal 2 ( see 1 1 . 5) , and
that the si ngleton of an el ement i s a part ( 7. 10) . Here we have
the pair of an ordi nal and of a part of that ordi nal , so i t' s a
Number ( 12. 1) .
2 N2 is the Number ( 0, 0) , whi ch has for matter the empty set, and
for form the empty set. It' s sti l l a Number though ! Because 0 is
an ordi nal , whi ch serves for the matter, and 0 i s a uni versal part
of every set ( see 7. 9) , i ncl uding 0 itself, whi ch i s, as I know, a set.
So 0 i s fne as the form too.
3 N3 is the Number ( 2, 1 ) . You can' t refuse me thi s, because 2 is an
ordi nal , and 1 , being an element of the ordi nal 2 ( fol l owing your
1 1 . 5, as al ways ) , i s al so a part of it, since every ordi nal i s
transitive.
Now, let's see, what do I have? The di scri mi nant of N, and N2 is
0: i t' s in the resi due of N" si nce 0 is an element of the matter 2, but
does not fgure i n the form, the si ngleton ( 1 ) , whose onl y element is
1. And i t' s outsi de the matter of N2, since thi s matter i s 0, of which
o cannot be an element. Therefore N, < Nz
The di scri mi nant of N2 and N
3
is also 0, which is outside the
matter of N2, as we can see, and which i s i n the form of N3, si nce
o E 1 . From thi s we concl ude
t
hat N2 < N
3
.
Here is a concrete test case where N, < N2, where N2 < N3, and
where, neverthel ess, w( 1 ,2) and w( 2, 3 ) , to use your notation from
the begi nni ng of 1 3. 1 7, are equal . Therefore nei ther i s smal ler than
the other, and your chai n of i nference i s broken.
ME: Very shrewd! You wi l l have to al l ow me al l the same that i n the
end, in your exampl e, transi ti vity is confrmed. Because the discri mi
nant of N, and N3 i s sti l l 0, which is i n the resi due of N, and in the
form of N
3
. So it i s sti l l the case that N, < N
3
.
READER: I make an obj ection on a point of princi pl e, and you respond
with an empi ri cal remark! My exampl e rui ns your general argument,
which rests on the fact that one can always di scern the smal lest of
the di scri mi nants of N, and Nz and of N2 and N
3
. I have shown you
a case where this cannot be done. The fact that transitivity sti l l works
for my exampl e might j ust be chance, si nce it now seems you have
yet to prove anythi ng.
ME: You al l ow my frst step, al l the same: that w( 1 , 3 ) cannot be
smal ler than w( 1 ,2) and w( 2, 3 ) ?
READER: With the caveat that the ' and' seems somewhat suspect to
me, si nce it might relate two equal di scri mi nants. See my example:
you woul d be sayi ng that ' w( 1 , 3 ) cannot be smal ler than 0 and 0' ,
which i s l udi crous.
ME: Unl ess it was smal ler than 0 . . . But anyhow - i f, as in your
exampl e, w( 1 ,2) is equal to w( 2, 3 ) , do you admit that w( 1 ,3) cannot
be smal l er than w( 1 , 2) alias w( 2, 3 ) ? Because no ordinal smal ler than
thi s common discriminant can discrimi nate between N" N2 and N3.
READER: Obvi ousl y.
ME: But w( 1 ,2) di scri mi nates between N, and N2 - i ts location in N,
i sn' t the same as in N2 ?
READER: No; how coul d it be?
ME: And it al so di scri mi nates ( goi ng by the name of w( 2, 3 ) , to which
it i s equal ) between N2 and N
3
- its location i s not the same in these
two?
READER: That' s exactly what I sai d.
ME: Let's l ook at these locations more closel y. Si nce N, < N2, w( 1 ,2)
must either be in the residue or outside the matter of N, . But can it
be outside the matter ?
READER: ( after some ti me thi nki ng) No. Because, if it were outside
the matter of N" it woul d have to be in the form of Nz, si nce N, <
Nz But, as it is al so the di scri mi nant of Nz and N
3
, and N2 < N3, it
cannot be in the form of Nz, as expl ai ned in 1 3. 1 4. So it is defnitely
in the resi due of N, and . . .
ME: outside the matter of N2, because not in its form. But where
is it l ocated in N3 ?
READER: ( after some ti me thi nki ng) In the form. Because t hi s w( 1 ,2) ,
which i s al so w( 2, 3 ) , i s the discri mi nant of N2 and N
3
. Being outsi de
the matter of N2, si nce N2 < N3 , it is in the form of N3 .
ME: Perfect! Here i s a w( 1 ,2) which w( 1 , 3 ) cannot be l ess than, and
which i s found in the residue of Nt and i n the form of N
3
. Therefore
it discrimi nates between Nt and N3 . That i s to say . . .
READER: Okay, I get it. Al ready identical to w( 2, 3 ) , it must al so be
identical to w( 1 ,3 ) . And thi s i dentity means that Nt < N
3
, si nce thi s
common di scri mi nant ( of NI and N2, of N2 and N
3
, and of Nt and
N3) is i n the residue of Nt and i n the form of N3 That works.
ME: It' s j ust as your exampl e says: 0 was the common di scri mi nant
of the couples N t -N2 and NrN3 . It is al so the di scri mi nant of Nt
and N
3
. And it i s l ocated in the resi due of N h outside the matter
of N2, and i n the form of N3. Which gives us the sequence: Nt < N2
< N
3

READER: You must admi t that you' ve had to add quite a bi t to your
ori gi nal account.
ME: It is a subsecti on of the argument, the principle i s the same. But
in mathematics one cannot ski p over anythi ng, for the reason that
one never knows what one i s ski ppi ng over.
13. 19. Si nce the rel ati on < i s total , i rrefexive and transitive, it real l y
is an order-relation in the mathematical sense. We have entirely j usti
fed our sayi ng ' Nt is smaller than N2 ' when it is confrmed, by means
of the location of the di scri mi nant of N I and N2, that the relation
Nt < N2 is val i d.
Thus the uni verse of Numbers - even i f it i s, as we shal l see,
borderless, saturated to an i nexpressi hl e degree, of a density wi th
regard to whi ch the cel ebrated ' conti nuum' is thi n and l acunary - can
nevertheless be comprehended whol l y under the seri al l aw of an
order.
The additional fact that this order can be designated solel y by the
examination of the location of an ordi nal ( the discri mi nant) with
regard to three possible si tes ( F, R and oM) i ndicates a si mpl icity that
is reassuring as to our capacity to think the uni verse of Numbers.
It is stri ki ng that, gi ven its combi nation of a logic of mi ni mal ity
( the di scri mi nant, the smal lest ordi nal to mark the di fference of two
Numbers ) and a l ogic of posi ti ons (the three components of the
numeri cal section) , this order appears to be al l ied with lexicographi
cal order. I n fact, i t i s presented as such i n purel y mathematical
expositions.
2
Now, lexicographi cal order, whi ch organi ses words by recourse to
an al phabet of the phonic or scri ptural unities that compose them,
touches on the distincti on, so i mportant i n Lacan, between the signi
fer and the letter.
3
In real ity, Number i s i ndeed l i ke a signifer, whose
internal ' posi ti ons' are the three locations - matter, form and resi due
- and whose l etters are the ordi nal s. Thi s al one permits us to organise
somethi ng as anarchi c as sets of any ordi nal s whatsoever, ordinal
' words ' .
If Number i s t he medi um i n which Nature, grasped in its being,
opens itself to our thought, this i s, without doubt - as the order of
Numbers testi fes - because, in the secti on it carri es out, we fnd,
under the si mple form of one and three, that di alectic of the position
and of the letter whi ch has been recogni sed, si nce Gal i leo, as the true
terrai n of materi al i sm. Nature consents to its profusi on within the
fction of a wri ti ng system; and we must recognise i n Number the
most inscribed i nstance of bei ng:
two fngers
snap in the abyss, i n
scri bbl ebooks
a worl d rushes up, thi s depends
on you.
4
1 4
The Concept of Sub- Number
14. 1 . The concept of substructure, and even ( i n category theory) I
that of the sub-obj ect, is fundamental for al l areas of contemporary
mathematics. We know the extreme i mportance of the determi nati on
of subgroups of a group, subs paces of a topol ogi cal space, etc. A
good many of the most profound mathemati cal theorems of recent
years are theorems of decomposi ti on or of presentati on: proving that
a structure can be presented as a composi ti on of ( possi bl y si mpler)
substructures, or that a structure i s decomposabl e i nto a sequence of,
or as a product of, pre-defned substructures. The elegance of thought
reaches its highest point when one manages completel y to ' resol ve' a
presented axi omatic structure into substructures that are of the same
type, but si mpler. Fi nite group theory offers some spectacul arl y
accompl ished exampl es of such resol uti on.
The underlying i dea i s as fol l ows: si nce the ' materi al ' of mathemat
ics is the pure or undi fferenti ated mul ti ple, structures are inevitabl y
homogenous with structured sets . Mathematical ontol ogy i s uni tary:
there aren' t, on the one hand, pre-given ' obj ects', on the other, struc
tural rel ati ons i nto whi ch these obj ects enter. Everything can poten
ti al l y be reduced to a mul ti ple wi thout qual ity, made of the voi d
alone. Gi ven thi s fact, it i s inevitabl e that the exerci se of thought
should consist in reducing compl ex mul ti pl icities to si mpler mul ti
pl icities, through the medi um of the axi omatic defniti on of si mpl e
and complex. The concept of structure organi ses thi s medi um: it di s
tinguishes elementary confgurations from more i ntricate confgura
tions. Ul ti mately the strategic stakes of the thi nki ng of being qua
1 3 2 ONTOLOGY: DEFINITION, ORDER, CUTS, TYPES
bei ng are to discern - given that every mul ti pl e is a multi pl e of mul
tiples ( One having no bei ng) - which multi ples a presented multiple
assures, in its turn, the presentation of. Whence theorems of decom
position, resol uti on, or presentati on.
What a mathemati ci an cal l s an ' obj ect' i s nothi ng but a multipl icity
wi thi n which sub- multi pl icities are i ntricated, often i n a very opaque
fashi on. The obj ect is a packet of mul ti pl es, whose i ntrication is an
obstacle to thought, and wi thi n which must be separated, as far as
possi ble, the multi pl es-regi ons whose presentati onal combi nation i s
assured by the ' obj ect' . The ' obj ective' i l l usi on, what we might cal l
the phantasm of the obj ect, relates to the i ni ti al di stance between the
entanglement of mul ti pl es and our separative access to this entangle
ment through the medi um of l anguage. Concepts, axiomatically
i ntroduced, determi ne types of structures, whi ch are the operators of
separation and al low us to exhi bit such an ' obj ect' as an arti cul ation
of substructures, indicating the l atency of sub-mul ti pl ici ties i n their
relation to the medi um of l anguage.
That a structure can be resol ved into substructures according to
vari ous operators of combi nati on ( embedded sequences of subgroups,
fnite or i nfnite products of compact spaces, etc. ) i s the defnitive
mark, in the inscribed strategy of thought, of the fact that what it
confronts is being qua bei ng, in the fgure of an i nfnite entanglement
of pure mul ti pl iciti es. A mathematici an wi l l say that he has 'thought
the obj ect' (or ' understood the probl em' ) when he has mapped the
l i nked i mmanence of the substructures whose presentati ve bond is
detai ned, i ni ti al l y i n an opaque fashi on, by the ' obj ect' . So it is al so
a question of the decomposi ti on of the obj ect, a putting to death of
the phantasm of the obj ect, which is an obj ect onl y in so far as it
resists, through its consti tutive entanglement, its resolution into the
specifc di versity of structures. Thi nking hy means of substructures
deposes the object and returns toward bei ng.
14. 2. In its commonl y accepted usage, the concept of number is not
a concept of the structural tpe. One doesn' t speak of ' numerical
structure' as one speaks of the structure of groups or of vector space.
What i s cal led ' number-theory' today i s an i nconsi stent set whose
centre of gravity i s i n fact a certai n area of al gebra: ri ng theory and
ideals theory. I n particul ar, no concept of sub-number exists, si nce
' number' doesn' t designate a type of structure.
Consequentl y, since the Greeks, the concept of number has been
the pri nci pal redoubt of a real ist, even empiricist, vi si on of mathemat
ics. Either number i s taken for a ' given' entity, or taken as proof that
mathematical nomi nati ons have a strictly symbol ic or operational
THE CONCEPT OF SUB-NUMBER 1 3 3
val ue. There is a c1 osi ng-i n-on-itsel f of the entity ' number' , which
is l i nked to its purely algebraic mani pul ati on. Certai nl y, numbers
are combined accordi ng to algebraic rules. But it does not at
all fol l ow from 7 + 5 12 ( whether this statement be analytic or
synthetic) that 7 and 5 are ' substructures' of 1 2. The most tenaci ous
i l l usion of obj ectivism resides in the convi cti on that 7, 5 and 12 are
non-decomposabl e marks, whose serial engenderment assures thei r
consistency.
It woul d therefore be a great victory for an ontological vi si on of
mathematics to establ i sh the structural character of number, to
unbi nd it from i ts empi ri cal punctual ity, to extract it from the si mple
form of the obj ect. Thi s programme, which woul d make of the predi
cate ' number' a reputable type of pure mul ti pl e, woul d fnd its most
si gni fcant moment i n the determi nati on of the concept of sub-number.
This concept woul d al ign numeri cal ity with the great structural cat
egories of mathematical thought (group, fel d, space . . . ) ; categories
by means of whi ch thought separates and unbi nds the intrications
of the pure mul ti pl e.
14. 3. The set-theoretical presentation of the concept of Number,
such as we have worked it through above, authori ses a strict defni
ti on of the sub-Numbers of a given Number. Better sti l l : as we shal l
show step by step, a Number is defned in a univocal manner by
its sub-Numbers. There exi sts a presentation of Number on the basi s
of the Numbers that are i mmanent to i t. Thus Number i n its turn
admits of theorems of decomposi ti on or of presentati on. It is
structuralised.
14. 4. The concept of sub-Number
The general idea of the sub-Number i s very si mpl e: we obtai n a sub
Number of a gi ven Number i f we ' partition
,
2 thi s Number at a poi nt
of i ts matter and keep everythi ng that comes ' before' thi s partiti on.
Si nce the matter of a Number i s an ordi nal , a ' poi nt' of partition i s
an element of thi s ordi nal , and thus a smal ler ordi nal . What there i s
'before the parti ti on' is constituted by the ordi nal s smal l er than the
one that defnes the partition. But the ordi nal s smal ler than a given
ordi nal are precisely the elements of that ordi nal . Consequently, i f N
is the point of the partiti on, then what comes before it, being consti
tuted by al l the elements of N i s nothing other than N itsel f. By
parti tioning at point N we obtain a new ordi nal-matter, which is N
a matter evidently more l i mited than that from which it was cut.
But, it wi l l be asked, what happens to the form, the numerical
section from the matter? Here, once agai n, the i dea i s very si mpl e: as
the form of the new Number, we keep preci sel y those ordi nal s that
are i n the form of the partitioned Number and which are ' before' the
partition. A sub-Number wi l l truly be a segment of a Number, up to
poi nt w, retai ni ng up to w (that is, between 0 and w) al l the charac
teristics of the partitioned Number.
Let' s give al l of thi s a more preci se form. Take a Number
N, ( W) , F( Nd) and an ordi nal w which is an element of WI ( i . e. is
in the matter of Nd. We partition NI at poi nt w, retaining only
ordi nal s that are lower than w, wi thout changing the rest at all: the
elements of the form of the new Number wi l l be those of F( Nd that
are lower than w. We thus make use of a property possessed by every
set of ordi nal s ( and therefore by the form of every Number) ( see N3) :
because i t i s composed of ordi nal s, its elements are ordered by the
rel ati on E It i s therefore entirely proper to speak of ' al l the ordi nal s
of F( Nd smal ler than the ordi nal w' . The diagram ( compare 12. 4)
shows thi s:
W W
1
Number :

l
l
l
l
Sub-Number : ---.:- -
w
We wri te Elw for the segment up to poi nt w of a set E of ordi nal s
of whi ch w is an el ement. Elw contai ns onl y elements of E l ower than
w ( but not w i tsel f, pl ease note ! ) . The Number obtained by the parti
ti on of N I
,
and which, by extensi on of our notati on, we will call N I lw
( which means that w must be in the matter W, of N) , that w E Wd,
wi l l have as its code: ( w, F( N dlw) . Its matter i s w - the poi nt at which
it i s partitioned, an ordinal that comes ' before' WI - and its form is
composed of al l the ordi nal s i n the form of N I whi ch are smal ler than
w. By the same token, its residue i s composed of all the ordi nal s
smal l er than w whi ch are i n the residue of NI .
We shoul d note that thi s Number ( w, F( Ndlw) i s exactly ' l i ke' NI
up to the ordinal w (exclusive) : in fact, up to w, any ordi nal that is
i n the form of NI i s in the form of ( w, F( Ndlw) too, and an ordi nal
that i s in the resi due of the former is al so i n the resi due of the latter.
The new Number obtai ned through partition i s, i n short, the ' i niti al
segment' of N) , an exact copy of the ' begi nning' of NI .
Take two Numbers N, and N2. If there exi sts an ordi nal w such
that NI Niw, where NI partiti ons N2 at point w, then we say that
N, is a sub-Number of N2 Or, alternativel y: a sub-Number of N, is
a segment N "w of N , .
14. 5. One sub-Number of N, - and one onl y - can be defned for
every ordi nal w in the matter of N, : therefore for every el ement of
W
I
' There exist exactly W, sub-Numbers of N" si nce an ordi nal
'counts' the ordi nal s that precede it. General l y speaki ng, a Number
admi ts of as many sub-Numbers as there are ordi nal s i n its matter.
14. 6. Take N, lw, a sub-Number of N, . 1t i s cl ear ( see the defni ti ons
and the diagram) that w i s the di scri mi nant of N"w and N" si nce,
up to w, these two Numbers are identical . Now, the matter of N"w
is w. So w is outsi de the matter of N, lw. The order-rel ati on between
N, and its sub-Number N, lw wi l l therefore depend enti rel y upon the
location of the ordi nal w in the Number N, : whether w i s i n its form
or in its resi due.
There are therefore two types of sub-Numbers for a given Number
N, :
1 Sub-Numbers N, lw2 where W2 - which is at once thei r matter and
the di scri mi nant of themselves and N, -is in the form ofN1 These
sub-Numbers are smaller than the Number N, (the di scri mi nant
W2 is outside the matter of N "W2 and in the form of N, ) .
2 Sub-Numbers N, lw
3
where W
3
is in the residue of N, . These sub
Numbers are larger than the Number N, ( the di scri mi nant W
3
i s
in the resi due of N, and outside the matter of N "W
3
) '
A sub-Number N "W2 of the frst type wi l l be cal led a low sub
Number. A sub-Number of the second type wi l l be cal led a high
sub-Number. The fol l owing diagram shows a low sub-Number and
a high sub-Number:
W,
Number N, , .. _,.-I-
I
I
Sub-Number N, /w
2:

(low) I
W
2
I
I
Sub-Number N, /w
a
:
-
(hi gh)
W
a
Note that there are evidently as many l ows as there are elements
i n the form of N I ( W2 must be i n the form) , and that there are as
many highs as there are elements in the residue of NI ( W3 must be in
the resi due) .
The low set of Number NJ , denoted by Lo( Nd, i s the set of low
sub-Numbers of N
1
The symmetrical case ( the set of hi gh sub
Numbers) i s denoted by Hi ( Nd, to be read ' high set of Number
N
1
' .
14. 7. The cruci al poi nt, then, i s the fol l owi ng. Take a Number N,
i ts l ow set Lo( N) and i ts high set Hi ( N) . N i s the one unique Number
of minimal matter to be situated ' between' the sets of Numbers which
are its high and its l ow sets.
This can be stated precisely as fol l ows:
1 N is si tuated ' between' Lo( N) and Hi ( N) in the sense that it i s
l arger than al l the Numbers of one and smal ler than al l the
Numbers of the other.
2 Al l the other Numbers si tuated between Lo( N) and Hi ( N) have a
greater matter than those of N. N is therefore the onl y Number
of mi ni mal matter to occupy the interval between its low set and
its high set. Thus a Number N is a ' cut' between its low set and
its hi gh set, a cut defned ' up to matter
,
.
3
The two sets of sub
Numbers Lo( N) and Hi ( N) defne N itself by way of location
( between the two) and material mi ni mal ity.
14. 8. The statement that N i s between i ts high set and its l ow set is
quite tri vi al , si nce by defni ti on al l the low sub-Numbers are smal l er
than N and al l the high sub-Numbers are l arger than N. The problem
i s to establ i sh that N i s of mi ni mal matter between the Numbers thus
si tuated, and that it i s the onl y one to have thi s matter.
1 4. 9. Principal lemma
Take NJ , a Number, and N2, another Number, smal ler than NI and
of l esser matter than NI ( so that N2 < NI and M( N2) < M( N
1
) ) . Then
either N2 i s a Number from the l ow set of NJ , or there exi sts a
Number from the l ow set of NI situated between N2 and N
1

Let w be the di scri mi nant of N I and N2 . Since we suppose the
matter of N2 to be lower than that of N J , and si nce N2 < N J
,
w is
necessari l y i n the form of NI ( it cannot be i n the resi due of N2 and
outsi de the matter of NJ , because then i t would be i n the matter of
N2 and outsi de the matter of N J , which possi bi l ity i s excluded by the
fact that M( N2 ) < M( Nd) . Consi der the sub-Number N
1
/w. Since w
is in the form of N" it is a sub-Number from the low set of N, ( it is
smal ler than N d.
Up to, but excl uding, w, N2 and N, are identical . If the di scri mi
nant w is outside the matter of N2 and therefore equal to i ts matter,
N2 is none other than the sub-Number Ndw, and i s therefore a sub
Number from the l ow set of N, . If w i s i n the residue of N2, then N2
is smal l er than the sub-Number N , Iw, because the discri mi nant of N2
and Ndw is necessari l y w - N2 being identical to N, up to the ordi nal
w
( excl usi ve) , and therefore al so i denti cal to N , Iw, whi ch i s a parti
tion at w of N" up to w ( excl usi ve) . Now, w is outside the matter
of N ,Iw, so we must suppose that it is i n the resi due of N2 . So N2 <
Ndw.
Thus it is " establ i shed that N2 is i ndeed ei ther a Number from the
low set of N, or smal l er than a Number from the low set of N, .
14. 10. An absol utely symmetrical chai n o f reasoni ng woul d prove
that, if N, < N2 and N2 is of a lesser matter than N" then either N2
is a Number from the high set of N" or else there exi sts a Number
from the high set of N, si tuated between N, and N2.
14. 1 1 . Concl usi on: for every number lower than ( or, respectively,
higher than) N, and of lesser matter than N" it i s the case either that
it i s a Number from the l ow set ( or, respecti vel y, the hi gh set) of N"
or else that a Number from the l ow set ( or high set) can be i nterca
lated between it and N, . It is therefore i mpossi bl e for any of these
numbers to be situated ' between' Lo( Nd and Hi ( Nd ( to be higher
than every element of Lo and lower than every element of Hi) whi lst
at the same ti me being of lesser matter than N, . The resul t is that N"
which i s indeed situated between its low set and its high set, i s of
mi ni mal matter wi th regard to al l Numbers thus situated.
14. 12. We wi l l now demonstrate that NJ is the only Number of
mi ni mal matter situated between i ts l ow set and its hi gh set.
Suppose there exi sted another Number N2, situated between the
low set and high set of N" and of the same matter as N , . Such a
Number coul d be represented as fol l ows ( with some abuse of our
notation) :
Si nce N2 i s of the same matter as N" the discri mi nant w of N, and
N2 is necessari l y in the residue of N2 and i n the form of N, . Thi s
means that the sub-Number Ndw i s i n the l ow set of N, . Now thi s
sub-Number, N, lw, is mani festly l arger than N2 ( thei r di scri mi nant,
once again, i s w, which i s i n the resi due of N2 and outside the matter
of Ndw) . Thus it cannot be the case that N2 i s l arger than every
Number in the l ow set of NI .
H we had the arrangement:
- we coul d demonstrate in the same way that there must exist a
Number from the high set of NI whi ch is smal ler than Nz ( a good
exerci se) . .
It fol l ows that N, real l y i s the onl y Number of mi ni mal matter to
be si tuated between Lo( Nd and Hi ( Nd.
NI is identifed, ' up to matter' - as t he uni que mi ni mal element of
that matter, once the ' between' position has been fxed - by the cut
of two sets of Numbers, the l ow set and the high set. We shal l write:
N, = Lo( Nd/Hi ( Nd. We shal l call the cut Lo( Nd/Hi ( Nd the canoni
cal presentation of N I
'
1 4. 1 3. A remarkabl e characteristic of the canoni cal representation
of NI i s that all the el ements of Lo and of Hi are sub-Numbers of
NI . Every number can be represented on the basis of Numbers
deducted from lesser matters than thei r own.
The canonical presentation is a frami ng
4
of Number from above
and bel ow, real i sed by means of more tightly controlled sections than
those carried out by Number.
Every Number i s a cut wi thi n sets of sub-Numbers, every Number
operates at the l i mi t of two series of Numbers subordinate and i mma
nent to i t.
Wi th thi s, the structural i sati on of the concept of Number is com
plete. Not only can a Number be l ocated as a secti on cut from natural
mul ti pl ici ti es, but this secti on can itsel f be presented as a point of
cutting between two seri es of secti ons of the same type. A Number
i s preci sel y thi nkable as the hi nge of its sub-Numbers. Number, so
far from being a si mple entity, answers to theorems of decomposi ti on:
it i s a structure local i sable i n thought as a point of articulation of i ts
substructures.
A Number exhi bits, as a one-result, its i mmanent numerical
determi nati ons.
1 5
Cuts: The Fundamental
Theorem
1 5. 1 . And so, let us penetrate into the swarmi ng of Numbers.
A frst remark, concerning what mi ght be cal led the number of
Numbers: thi s number i s preci sel y not a Number, it i s not even a
consistent multi pl i ci ty. Numbers are numberless.
In fact, given that a Number is the pai r of an ordinal and of a part
of that ordi nal , not onl y are there at least as many di fferent Numbers
as there are di fferent ordi nal s, but there are many more, even i f this
'more' fickers beyond the frontiers of sense. For each ordi nal , there
are as many di fferent Numbers as there are di fferent parts of that
ordi nal : i f W is an ordi nal , serving as the matter of certai n Numbers,
there wi l l be p( W) (the set of parts of W) forms - each one vi rtual ly
extractable by means of a numerical secti on from thi s matter.
Now we al ready know that the ordi nal s do not constitute a set.
' Al l ' the ordinals cannot be counted for one i n a set-theoretical recol
lecti on. I n other words, the ordinal s form an i nconsi stent mul ti pl icity.
Consequently, the same goes for Numbers.
But, what i s more, for any gi ven mul ti pl e whatsoever, we cannot
know exactl y what the quantity of the set of its parts i s. Certai nl y,
we know ( Cantor' s theorem) that it must be l arger than that of the
i ni ti al set: it i s always the case that P( W) W. But ' how much' l arger?
I t has been proven ( by Godel and Cohen' s theorems) that the amount
of thi s excess i s undeci dabl e on the basi s of the fundamental axi oms
of set theory. In fact it is coherent wi thi n these axi oms to say that
p( W) is ' i mmensel y' l arger than W; and it i s al so coherent to say that
it i s ' mi ni mal l y' l arger. '
Ulti matel y, for every ordi nal , there are always more possi ble
Numbers for whi ch it is the matter than there are elements of the
ordi nal itsel f. And, as thi ngs currentl y stand, the extent of this ' more'
can onl y be decided. Whi ch is to say that the number of Numbers is
an i nconsi stency of i nconsi stenci es.
The si mpl est way to put it is to say: Number is coextensive with
Being. It i nconsi sts, is dissemi nated and profused j ust l i ke the pure
mul ti pl e, the general form of being qua being.
1 5 . 2. Thi s i nconsi stent swarmi ng of Numbers gives us to antici pate
the di ffculties that ari se with regard to the identifcation of a specifc
Number pi cked ' from the crowd' . Every Number i s cemented into
the throng of those that pack in tightly, on its right ( Numbers l arger
than it) and on its l eft ( smal ler Numbers ) . No Number si mpl y,
uncompl icatedly succeeds any other. Every microzone of the numeri
cal domain teems wi th a numberless horde of Numbers. The numeri
cal topol ogy is pecul i arl y dense. And thi s is the problem: is it possible
to identi fy a Number as opposed to sets of Numbers ? Or must we
consign oursel ves, when we consi der series of Numbers, i nfnite sets
of Numbers, to being unable to attach to them, uni vocal l y, any spe
ci fc Number? Does the numberless throng of Numbers necessarily
lead us into ' those i ndefnite regi ons of the swel l where all real ity is
di ssol ved' ?
2
Thi s is where trans-numeric i nconsi stency summons us to thi nk
the cut. I s it possi bl e, i n a fabric so dense that nothing any longer
numbers it, to cut at a specifc point? Can one determine, by cutting,
a singular Number?
1 5. 3. Thi s probl em i s not in the least bi t academic, nor is it relevant
solely to the thi nking of Number. We are told every day how
' the complexity of modern society' prevents us from making any cut,
any i nterventi on. Contemporary conservatism no longer argues from
the sacredness of the establ i shed order, but from its densi ty. Every
l ocal cut, it says, is real l y a 'tear i n the soci al fabric' . Leave natural
l aws (the market, appetite, domi nati on) to operate - because it i s
i mpossi ble to i nterrupt them at any poi nt. Every poi nt i s too depen
dent on all the others to permit the preci sion of an i nterrupting
cut.
Thi nki ng the cut i n the hyper-dense, cl osel y kni tted fabric of
Numbers will al low us to concl ude that such arguments are fal l a
ci ous. Every poi nt separates dense sets of Numbers, every Number is
the pl ace of a cut, and, conversel y, every cut prescri bes one Number
and one onl y. Not ' i ndefnite regi ons' , but 'a Constel l ation' .
3
CUTS: THE FUNDAMENTAL THEOREM 1 41
15. 4. This probl em al so has a compl ex phi l osophi cal geneal ogy: that
of the di alectic between continuous magni tude and discrete magni
tude. I f the being of the continuum is grasped in i ts i nti mate coal es
cence, so that it is not constituted from distingui shabl e poi nts, but
rather from compl icated ' nei ghbourhoods' , it must be thought as
di sj oi nt from di screte quantity, whi ch enumerates successive marks.
Up to, and i ncl udi ng, Hegel , this opposi ti on, which subsumes and
underwrites that between geometry and ari thmetic, remai ns i n the
position of an enigmatic real for the phi l osophy of quantity. In Kant,
sti l l , it ul ti matel y supports the dual ity of forms of sensi bi l i ty: Space
is the transcendental fgure of the conti nuous, Time - from which
proceeds number - that of discrete successi on.
The most profound concept of the cut, a concept that pl ays an
immense rol e i n modern thought,
4
di spl aces and refounds the di alecti
cal schema whi ch consi ders the couplet di screte/conti nuous to be the
founding contradiction of the quantitative. Thi s concept bri ngs forth
a si ngul arity - and therefore a basi s for distinction - in the fabric of
the conti nuous, in the dense stuff of i nfnitel y smal l neighbourhoods.
Overturni ng the customary order of thought, it shows how a certai n
sort of i nterruption of the conti nuum defnes a type of di screteness.
Rather than sayi ng that the conti nuum i s composed of poi nts, it
determines points wi thi n the continuum, and even defnes punctual i ty
on the basi s of a cut in the conti nuum. The concept of cut substitutes,
for a problematic of composition, a problematic of completion: a
point comes to ' fl l i n' a j uncture, or an impercepti bl e l acuna, in a
pre-given conti nui ty.
15. 5. Dedeki nd5 invented the concept of the cut i n order to defne
irrati onal numbers.
He begi ns with rati onal numbers. We know that a positive rati onal
number is of the form ~ p and q bei ng natural whol e numbers. The
rational numbers provi
d
e our pri mary i mage of conti nuity owing to
the fact that thei r order i s dense. A dense order i s an order such that
between two ordered elements i s al ways i ntercal ated a thi rd - and,
by reiteration of thi s property, an i nfnity of elements. If we take the
rational number 0 ( which i s rational because i t can al so be expressed
as any fraction %) and the rational number t, then 0 < t. But the
numbers
t
, t,
t, etc. - and an i nfni ty of numbers of the form . -
intercal ate themselves between 0 and
t
.
Density does not di rectl y express a quantitative property: the ratio
nal numbers are an infnity of the type bel onging to the countabl e,
an infnity no greater than that of the natural whol e numbers, and
the l atter, bei ng none other than the fnite ordi nal s, do not present a
dense order: there is no natural whol e number between n and n + 1 .
Density i s real l y a topological property of order: excl uding the si mple
idea of ' another step' , of the wel l -determined follower of a term, it
proposes instead a sort of general coalescence i n which every term
'sticks' to an i nfnity of neighbours. The density of an order is a
topol ogi cal property, whereas successi on is an algebraic property.
Density is ' quasi -conti nuous' , one can approach a rational number
as cl osel y as one wi shes through other rati onal s. One even gets the
feel i ng that, between two rational numbers and, more generally,
between two terms of a dense order, there is no pl ace for numbers,
or terms, of another type, since the whole interval , no matter how
smal l it i s, is al ready popul ated with an i nfnity of rati onal s, or an
i nfnity of terms of the dense order.
Now, i t i s preci sel y i n this quasi -conti nui ty of rati onal s that
Dedeki nd wi l l , by means of the cut, defne additional ' poi nts' that
will complete the apparently uncompl etabl e densi ty of the rati onal s
and obtai n a ' true' conti nuum, through i nterruptions in thei r
quasi -conti nuity.
We will return in greater detai l to thi s procedure in chapter 1 6.
But schematical l y: Dedeki nd considers di sj oi nt sets of rational
numbers RI and R2, for whi ch every element of RI i s less than every
element of R2 and whi ch, R2 havi ng no rati onal internal maxi mum
nor RI any rati onal internal mi ni mum, are two ' open' sets, one high,
the other l ow. Dedeki nd then identifes a real number as occupying
the pl ace of a cut between RI and Rz This real number will be both
the upper l i mit of RI and the lower l i mit of Rz The density of the
order of rati onal s pl ays an essenti al role i n thi s constructi on, once it
i s understood that density and the cut, far from bei ng excl usi ve, are
pai red together in thought.
It must be noted strai ghtaway that thi s procedure seeks to defne
real numbers, the rati onal numbers being supposed to be known. The
Dedeki nd cut i s whol l y an operation of completi on: where there is
nothing, no rati onal number, the name of somethi ng 'extra' comes
forth. The real number defned by the cut R/R2 fl l s i n that which,
thought purel y from the poi nt of view of rati onal s, is a void in the
density, and thus a void to whi ch nothing attests. This is why the cut
founds a new species of numbers, which ' compl ete' the i ni ti al density
and retroactively i ndi cate that this density was not so dense that gaps
could not be di scovered therei n.
1 5. 6. We cannot hope to ' complete' the inconsi stent domain of
Numbers, nor to found, outside Number, a hyper-number which
would name the i nvi si bl e l acunae i n i t. Our Numbers are uncomplet-
abl e, being coextensi ve wi th Being ( see 1 5 . 1 ) . All the Numbers are
al ready there. What could a cut mean i n such condi ti ons ?
Nevertheless, there is a very strong concept of the cut for Numbers.
This concept hol ds ' up to matter' , l i ke that of the si ngul ar el ement
separative of a Number and i ts sub-Numbers, of i ts identity as cut
between its low set and its hi gh set ( see chapter 1 4) .
This concept of the cut i s presented i n the fol l owi ng theorem
whi ch, articul ating the i nconsi stent swarmi ng of Numbers with the
preci si on and the unity of a punctual cut, wel l deserves the name of
fundamental theorem of the ontolog of Number:
Given two sets of Numbers, denoted by B ( for 'from bel ow' ) and
A ( for ' from above' ) , such that every Number of set B i s smal l er
than every Number of set A ( i n the order of Numbers, of course) ,
there always exists one unique Number N of minimal matter
situated 'between' B and A. ' Situated between' means that N
is l arger than every element of B and smal ler than every el ement
of A.
The Number N i s evidently not the onl y one between B and A.
The numerical swarming i s such, the density i s so consi derabl e, that
such a sol itude would be unthi nkabl e. But it i s the only number to
be found with its matter. All the others have a l arger matter, i n a
rigorous sense, si nce matters are ordinal s: the ordi nal-matter of N is
mi ni mal for the property ' i s the matter of a Number situated between
the sets of Numbers B and A' .
It wi l l not surprise us at al l to fnd mi ni mal ity here: it i s a classic
organisational principle of ordinal s. What is surpri si ng i s:
-
that such a Number shoul d exist;
- that it shoul d be unique.
Its exi stence founds the pri nci pl e of the cut. If two sets of Numbers
are l ike B and A ( every Number of B being smal ler than every
Number of A) , then one can sti l l speak of what exists ' between' B
and A and is neither of B nor of A, in spite of the prodigi ous density
of the order of Numbers. It is thus possi bl e to make a cut in the
hyper-dense fabric of thi s order.
Uni queness ( up to matter, which is to say uni queness of the
Number-ut of mi ni mal matter) founds the pri nci pl e of identifcati on,
the persistence of the count-for-one even where al l is coalescent, i n
dense nei ghborhoods. A cut designates one Number, and designates
it on the basi s of sets of Numbers. We will hold that no complexi ty,
even one pushed to the poi nt of i nconsi stency, no density, even one
pushed to the fnest i nfnitesi mal proxi mi ti es, can authorise the pro
hi bi ti on agai nst cutting at a poi nt.
1 5. 7. The rest of thi s chapter is dedicated to the demonstration of
the fundamental theorem, the onl y theorem i n thi s book that is a
l ittle complex. 6
I do not, moreover, intend to gi ve al l the detai l s of the proof.
However, we are at the heart of the mathematics of Number, and
what must be put i nto play i n order to thi nk the cut i s of a conceptual
i nterest far surpassi ng mathematical ontology. All truth-procedures
proceed via a cut, and here we have the abstract model of every
strategy of cutting. The intel lectual effort demanded of the reader wi l l
lead hi m or her, I am qui te sure, to beatitude i n the Spi nozist sense.
1 5. 8. Upper bound of set of Numbers
Si nce we are engaged i n i nvestigati ons whose character i s topological ,
and si nce in particul ar we are wondering how to fnd Numbers l arger
( or, respectively, smal ler) than a given set of Numbers, let' s begin
with the si mplest concept, that of an upper bound: given a set of
Numbers, does it make sense to speak of a ' unique' Number l arger
that al l those i n the set ?
Once more we must, in view of the prol i feration of Numbers, avai l
ourselves of a concept ' up to matter' . We wi l l prove the following:
i f B is a set of Numbers, then there exi sts a Number N whi ch is the
unique Number of mi ni mal matter to be l arger than al l the Numbers
i n set B. We wi l l cal l thi s N the upper bound of B. Right away the
upper bound exhi bits a surpri si ng characteristic: it is always a Number
written (W,W) - that i s, a Number whose form i s its whole
matter. 7
1 5. 9. Take B, a set of Numbers. Consider the ordi nal defned as
fol l ows: ' the smal lest ordi nal W such that, for every Number N of
set B, there exi sts a WI E W which i s either in the resi due or outside
the matter of N' .
One such ordi nal W exists, because B is a set, and is therefore
consistent. If W di d not exist, that woul d mean that all ordinals
woul d fal l into the form of at least one Number N of B. But ' al l
ordi nal s' is an i nconsi stent mul ti pl icity, and consequently B would
also be an i nconsi stent mul ti pl ici ty, and would not be able to be
thought as a set.
That there shoul d exi st such a ' smal lest' W results from the prop
erty of mi ni mal ity that characterises the ordi nal s.
W having been specifed, now consider the Number ( W,W) . Thi s
number is larger than every Number i n set B. In fact, by the defni ti on
of W, for every Number N of B there exists a WI E W whi ch i s in i ts
resi due or outside its matter. Now, as the form of ( W, W) is W, every
WI E W i s i n the form of ( W,W) . The discri mi nant of a Number N
and of ( W, W) is necessari l y the smal lest WI E W that is in the residue
or outside the matter of N. And, since thi s WI i s i n the form of ( W,W) ,
the residue ( or outside-matter)/form relation demands that ( W, W)
shoul d be l arger than N.
Si nce a Number l arger than every Number i n B exists - namel y
( W, W) - one of mi ni mal matter must exist, i n vi rtue of the ordi nal s'
property of mi ni mal ity.
Therefore, let ( WI ,X) be a Number of mi ni mal matter for the
property ' being the matter of a Number l arger than al l the Numbers
in B'. Its form X is i n fact equal to WI .
For, if X di ffered from WI - i f, that i s, the form of the Number
was not its whol e matter - that woul d mean that there existed at l east
one ordi nal W2 E W which was i n the resi due. Consi der then the
sub-Number of ( WI ,X) obtained by partition at W2 - that i s, the sub
Number ( WZ, X/W2 ) . Si nce Wz i s i n the resi due of ( W"X) , the sub
Number ( WZ,X/W2 ) i s in the high set of ( W I , X) ( see 14. 7) . It is there
fore l arger than ( WI ,X) , and a fortiori l arger than every Number i n
B, si nce thi s is al ready the case for ( W"X) .
But that i s i mpossi ble, because t he number ( WZ, X/W2 ) i s of lesser
matter than the Number ( WI ' X) . Now, we supposed that ( WI , X)
was of mi ni mal matter for Numbers higher than every number
in B.
Our i ni ti al hypothesis must be rej ected: there does not exist i n
( WI ,X) any element that is i n the residue, whi ch is to say that the
form occupies the whol e matter, and that the Number must be
written ( W"Wd.
There exists therefore one Number only of mi ni mal matter that i s
higher than al l the Numbers of set B: i t i s the Number ( W"Wd, where
W I is thi s mi ni mal matter.
We can thus legitimately speak of the upper bound of a set of
Numbers. Al ready the theme of unicity comes to inscri be itsel f as bar,
or caesura, in the hyper-dense swarmi ng of Numbers.
1 5. 1 0. Lower bound of set of Numbers
Reasoning totally symmetrical wi th that empl oyed for the upper
bound will permit us to defne the uni que Number of mi ni mal matter
that is smal ler than a set A of Numbers . This wi l l be the lower bound
of the set A. We wi l l see that, thi s time, thi s Number i s wri tten ( W2, O) :
its form is voi d, the numeri cal section does not extract anything from
the matter Wz. 8
Let A be a set of Numbers, and l et W be the ordi nal defned thus:
' the smal lest ordi nal such that, for ever Number N of A, there exists
a WI E W whi ch i s either in the form, or outside the matter of N' .
Thi s ordi nal exists necessari l y, because A i s a set, and in vi rtue of the
pri nci pl e of mi ni mal ity ( see above) .
The Number ( W, O) , whose matter i s W and whose form i s the
void, is smal ler than every Number of A. In fact, al l ordi nal s WI E
W are in the residue of ( W,O) . Now, for every Number N of A, by
the defni ti on of W, there exists a WI E W which is either in the
form of N, or outsi de the matter of N. The smal lest such WI is the
di scri mi nant of N and of ( W, O) , and its location means that ( W, O)
< N.
Hence there exi sts a Number smal l er than every Number in A, and
- by the pri nci pl e of mi ni mal ity - there exists at l east one of mi ni mal
matter, say (W2,X) .
I t i s easy t o prove that X is necessari l y the empty set. If it were
not, that woul d mean that there exi sted a W3 E W 2 whi ch was in
the form of ( Wz, X) . But then the sub-Number of ( Wz, X) obtained
by partition at W3, that i s, ( W3 , Xlw3 ) , would be in the low set of
(Wz,X) ( see 14. 7) . It would then be smaller than ( Wz,X) , and therefore
smal ler than every Number in A, al though of lesser matter than
( Wz, X) : which i s i mpossi bl e i n view of the mi ni mal i ty of Wz for thi s
posi ti on.
Therefore, there exi sts one uni que Number of mi ni mal matter that
i s smal ler than every Number in A. It i s the Number (Wz, O) , where
W2 is thi s mi ni mal matter. The Number ( W2, 0) is the lower bound
(up to matter) of set A.
1 5. 1 1 . Fundamental theorem, frst part: Existence
' Exi stence' means here: exi stence of at least one Number situated
between two sets of Numbers B and A, whi ch, i n an abuse of our
usual notati on, we shal l general l y write as B < N < A.
Take B and A, two sets of Numbers such that every Number of B
is smal ler than every Number of A. Our techni que wi l l consist in
constructing, between B and A, step by step - that is to say, ordi nal
by ordi nal - starting from 0, a Number N ' suspended' at every step
i n such a way as to assure us that nothi ng up to the ordi nal W in
questi on, which i s to say, for every step taken i n the procedure - can
force the Number N to be smal l er than a Number of B, or l arger than
a Number of A. We might al so say that we are going to construct
N from its sub-Numbers of i ntersecting matter, by ' choosi ng' to put
an ordi nal W i n the form or i n the resi due of the Number N under
constructi on, dependi ng on the relati onshi p between the segment of
the procedure of N whi ch goes from to W, and the vari ous sub
Numbers in B and i n A.
The underlying i dea is that the construction of a cut makes nec
essary a local domi nati on of the substructures i mpl icated in the
course of thi s constructi on. Thi s i s a general l aw of practice, at
least in so far as the latter ai ms at effects of cutting ( foundati onal
interrupti ons) .
This techni que boasts the very great interest of highl ighti ng the
l i nk between cutting and a sort of procedure of neutralisation. So
that N can sl i de i n between the Numbers of B and the Numbers of
A, we are goi ng to remai n mi ndful of the fact that the pri nci pl e of
order, at every poi nt of N, ' neutral i ses' the di scri mi nati on between
the Numbers of B and the Numbers of A. The great di ffculty being
to know when to stop ourselves, when to fx the matter of the
Number N, whi ch we woul d have traversed, all the whi le postponi ng
its closure.
In all domai ns of thought, to proceed with a precise cut i n a
densely ordered fabric is to cal cul ate a prudent tactics of i nser
tion step by step, and then to ri sk a stopping poi nt whi ch wi l l
i rreversi bl y fx the intermedi ary term. The cut thus combines the
neutrality of the i nterval and the abruptness of the interruption.
This is why great strategies of thought must al ways attain a
mastery both of the patience whi ch, poi nt by poi nt, opens and
enlarges a l acuna, and of the i mpatience which comes to seal and to
name i ts existence from thi s moment forward, wi thout return or
recourse.
1 5. 12. So we begin from the ordi nal 0, and we traverse the ordi nal s,
assigni ng to each a val ue f( W) - the val ues bei ng F ( for form) , R ( for
residue) , or M ( for matter) . The val ue M can obvi ously only be given
once, and last of al l , because the Number N that we want to construct
has only one matter. For an ordi nal W, i f f( W) F, we wi l l put W i n
the form of the Number N under constructi on; if f( W) R, we wi l l
put it in the residue. So l ong as we have not assigned the val ue M,
the sub-Numbers are sti l l ' under construction' . The procedure
amounts to fxing a l ocation-status for each ordi nal W, so that the
sub-Number NIW, as the procedure continues, wi l l appear retroac
tively as never constrai ni ng N to be l arger than any Number of A,
or smal l er than any Number of B.
1 48 ONTOLOGY: DEFI NITION. ORDER. CUTS. TYPES
The strategic patience of the construction of a cut consists in insert
ing addi ti onal local val ues without compromi si ng the chances of a
global cut. It is a work that proceeds point by poi nt, but is retroac
tively deci ded as an i rreversi ble and general caesura.
We wi l l denote by Nb and Na, with i ndices i f need be, the Numbers
of B and A. Ni wi l l designate the interval lic Number, the Number we
wi sh to construct between B and A.
If, for a given Number Nb (or respectively Na) , the val ues attrib
uted by f to ordinals smaller than an ordi nal W ( val ues of the type
F or R, whi ch the ordi nal s of the Ni under construction take) are
exactly those whi ch l ocate these ordi nal s in Nb (or respectively in
Na) , then we say that W identifes Ni and Nb ( or respectively Na) .
W' s i denti fyi ng Nb ( or Na) and Ni means that, in every case, no
ordi nal smal ler than W can discri mi nate between Nb ( or Na) and Ni.
In particular, the di scri mi nant of Nb, or Na, and the segment of Ni
under construction (a segment which ranges from 0 to W excl usive)
cannot fgure in the ordi nal s i nferi or to W. Whi ch amounts to saying
- and thi s i s the most tractabl e form of the rel ation of identifcation
at ordi nal point W - that, up to W, the ' sub-Number' Ni/W is identi
cal to the sub-Number Nb/W ( respectively Na/) .
We wi l l denote by Id. ( W, Nb) the fact that W i dentifes Ni and Nb.
And the same thi ng for Na. Al l the whi l e we shoul d keep i n mi nd
that Id. ( W, Nb) means that Ni/ Nb/W.
The strategic idea is to construct an Ni ' neutral i sed' for order, by
making sure, each time one comes to ' the end' of a series of ordi nal s
whi ch identify Nb ( or respectively Na) and the Ni under constructi on,
that the choice of a val ue for f( W) wi l l not be abl e to compromise
our chances of positing a hypothetical ly completed Ni, which would
be interval l ic between B and A. We must j ust make sure that no
ordi nal comes to be in the posi ti on of an unfavourabl e discri mi nant
forcing Ni to be smal ler than a Number of B, or l arger than one of
A. The prudence of the cut consi sts here in never ri ski ng l osi ng the
chance to take up an i nterval l i c positi on. Conserve its chances, that
i s the maxi m of the ' step by step' phase of the construction of
a cut.
1 5. 1 3. We wi l l posit the fol l owi ng rul es - rules of construction of Ni
for the ordinal s starting from 0:
K L 1 : I f I d. ( W, Nb) and W i s the matter of Nb, then f( W) F.
We put the ordi nal W i n the form of Ni whenever, at the end of
an Nb/W i denti cal to Ni/, W i s the matter of Nb. So, usi ng a
bl ack square to denote a belonging to the form:
Nb -
T
Ni "
Nbl
Id!m
Nil
r
K L 2: If Id. (W, Na) and W is the matter of Na, then f( W) R.
The diagram shoul d be cl ear, marki ng with I a belonging to the
residue:
Ni
Nil
v
W
Id!m
Na
~
Nal
K L 3: If rules 1 and 2 do not appl y for a given W, but i nstead

we have an Nb such that Id. ( W, Nb) with W i n the form of Nb,
then f( W) F. I f cases 1 and 2 do not appl y, we put W in the form
of Ni each time that, at the end of an NblW identi cal to NilW,
W is in the form of Nb:
Nb --
Ni ------- ---

-----

-

W
K L 4: If rul es 1 and 2 do not appl y, and we have an Na such
that Id. (W,Na) with W in the resi due of Na, then f( W) R:
N -
V
' -
w
Id;m
Na !
l
/ W
r
K L 5: If none of the frst four rul es appl y, it must be the
case that, for the W consi dered, no Nb such that Id, ( W, Nb) has
W for matter or i n its form, and that no Na such that I d, (W,Na)
has W for matter or in its resi due. Under such condi ti ons, at
poi nt W, i f there exi sts an Nb for whi ch Id.( W,Nb), W i s in
the residue of Nb, and i f it is the case that Id. ( W,Na) , W is in
the form of Na. We then say that f( W) = M, which completes
the construction of Ni.
As j usti fcati on for thi s rul e, note the fol lowi ng: si nce al l Nb ( or
respectively Na) where W is not thus l ocated - so, W in the resi due
( or respectively i n the form) - are such that W does not identify
them with Ni, then, for these Nb ( or Na) , it is the case that NbfW .
NifW ( or respectively NafW . NifW) . In other words, these Nb and
these Na have already been di scri mi nated, before ordi nal W, by the
process Ni. The only Nb and Na not to have al ready been discrimi
nated are those where W i s i n the residue (or respectively i n the
form) .
Gi ven thi s remark, we can state that rule 5 prescribes witb com
plete justifcation tbe decision of closure of the process Ni . We can
posi t: f( W) M, thereby fxi ng W as the ordi nal-matter of Ni, and
therefore as that pl ace where the process of the construction of Ni
ends.
If W i s the matter of Ni, it i s located outsi de the matter for that
Ni supposed closed in W. Now, W does not di scri mi nate Ni from
Nb where W is in the resi due, or from Na where W is in the form.
The location for Ni will remain ' between' B and A, since the schema
of the order-relation is precisely R < oM < F. We will have:
Nb -
/
":
Ni .
Idm
..
..
,
-
..
Na "
W
Cl osure is enti rely possi bl e, si nce, beyond ordi nal W, all Nb and
Na are di scri mi nated by Ni ( before W through rules 1 to 4, at point
W by rule 5) ; and our rules reflect the fact that this di scri mi nation
always goes in the di rection Nb < Ni < Na.
This regul ati on, however, merits i mmediate exami nati on.
15. 14. It is essenti al to confrm that our rul es do not contradict one
another.
Take for example rules 1 and 2. If by some mi schance i t shoul d
happen that at the same time Id. ( W, Nb) and Id. ( W, Na) , wi th W the
matter both of Nb and of Na, then W would have to be pl aced
si multaneousl y i n the form and i n the resi due of Ni . . .
But such a case cannot ari se. Because, i f W i s the matter of Nb
and of Na, since every Number of B is smal l er than every Number
of A, it i s the case that Nb < Na. And, si nce they have the same matter
W, their discri mi nant must be less than W, which i s to say that there
is at least one ordi nal WI E W whi ch doesn' t have the same location
in Nb and i n Na. It is therefore not possible for sub-Numbers NblW
and NalW to be identical . Thi s means, moreover, that, if both
Id. ( W, Nb) and Id. ( W, Ba) , thei r common identity must be NilW. So
rules 1 and 2 are compati bl e.
But take rules 3 and 4. If by some mi schance there is a W for whi ch
rul es 1 and 2 do not appl y, and there exist Nb and Na for whi ch,
frstly, Id. ( W, Nb) and Id. ( W,Na) , and, secondl y, W is i n the form of
Nb and W is in the residue of Na, W woul d have to be pl aced both
in the form and i n the residue of Ni.
But of course such an unfortunate ci rcumstance cannot ari se.
Because, i f W i s i n the resi due of Na and in the form of Nb, then it
di scri mi nates between Nb and Na. But thi s could not be thei r dis
criminant, otherwise it woul d be the case, with regard to thi s l oca
tion, that Na < Nb, whi ch is prohi bited by B < A. Therefore the
di scriminant is smaller than W, and, as before, it i s i mpossi ble that
NblW = NalW; which makes it necessary to suppose their common
equal ity to NilW.
1 5. 1 5. Now we wi l l see whether, wi th these rul es, we do i ndeed
preserve our chances that Ni wi l l sl i p i n between all the Numbers
of B and all the Numbers of A, and therefore between all Nb and
al l Na.
When we appl y rule 1 , we give the val ue F to the ordi nal W. Thi s
certai nl y cannot make Ni become l ess than a Number of B, because,
i f W is the di scri mi nant of Ni and of an Nb, being i n the form of Ni,
it wi l l always be the case that Nb < Ni.
But, given the fact that we put W in its form, don' t we ri sk Ni
becoming larger than a Number of A? For thi s it woul d have to be
the case that W was the di scri mi nant of Ni and of an Na. But then
it woul d al so ul ti mately be the di scri mi nant of the Nb of whi ch W
is the matter ( since we appl y rule 1 ) and of Na. Now, we know that
Nb 5 Na. If thei r di scri mi nant is the matter of Nb, it must be i n the
form of Na. Thi s location of W - W being the di scri mi nant of Ni
and Na - prohi bits us from havi ng the order Na < Ni.
So, in appl ying rul e 1, we can be sure that the location that we fx
for W i n the Number Ni under construction entails neither an unwel
come and frustrating Ni < Nb, nor a fatal Na < Ni . At point W, Ni
stays si tuated ' between' B and A.
The exami nati on of the other rules leads us to the same concl usi on.
Let' s carry out thi s exami nation for rul e 5 ( for rules 2, 3 and 4 the
methods are the same as for rule 1. Let the reader prove this as an
exerci se, wi th the hel p of the note, 9 and above al l of the diagram
bel ow. )
Rul e 5 comes i nto pl ay when rules 1 to 4 are not appl icabl e. The
W under consideration makes no identifcation between any Nb ( or
Na) and Ni if W is located as matter of Nb ( rul e 1 ) , matter of Na
( rul e 2) , form of Nb ( rul e 3 ) or residue of Na ( rule 4) . If, then, it is
the case that I d. ( W, Nb) , or Id. ( W,Na) , it is because W is in the residue
of Nb and/or i n the form of Na. These two hypotheses are compatible
thi s time: the i dentifcations i n question could obtai n, and W could
be both i n the resi due of Nb and i n the form of Na. Rul e 5 then
compel s us to make the gesture of cl osure f( W) M, which deter
mi nes W as matter of the i nterval l ic Number Ni. In the Ni thus
cl osed, W i s located outside the matter. Can thi s choice make Ni less
than some Na, according to the relation R < oM? No, because, i f W
di scri mi nates between thi s Na and Ni, with W in the resi due of Na,
thi s woul d be a case for the appl i cati on of rule 4, whi ch would
excl ude the use of rule 5. And, i n the same way, it cannot be the case
that Ni < Nb according to the relation oM < F, because the l ocation
of the di scri mi nant W in the form of an Nb compel s, for W, the use
of rule 3 rather than rule 5. Rul e 5, appl ied when it is proper to do
so, cannot entai l that Ni < Nb. And, as it cannot entai l Na < Ni
ei ther, it leaves the procedure Ni, at poi nt W, i n the i nterval between
B and A.
So it i s that, at every ordi nal poi nt W, the application of our rules
' l ocal l y' si tuates Ni, in the form of the sub-Numbers Nil, i n an
i nterval l ic posi ti on with regard to B and A. Our step-by-step l abour
i s pursued wi thout Ni surpassi ng any Na, or being surpassed by any
Nb. We conserve our chances al l the way through the construction.
An enl arged di agram shows how Ni proceeds . We have, above, some
Numbers Nb of B, bel ow, some Numbers Na of A, and, i n the mi ddl e,
the process of Ni. The ordi nal s WI to W
s
present, i n order, cases of
the appl icati on of the fve rules. Squares, asteri sks and bars designate
form, matter and resi due. You wi l l recal l that, when a point i s marked
i n an Nb or an Na, it means that, before that ordi nal point, Na ( or
Nb) is identical to Ni ( rel ati on of i dentifcation at an ordi nal
poi nt) .
Ni -
A{
W
V
Z
1/ 1/
/ /
,
.
V
Z
rul e 1 rul e Z rule 3 rul e 4 rul e b
w "
Rb,
Rb2
Rb3
.
.
Ra,
Ra2
Ra3
if fules 1 and Z do not appl y if ful es 1 and 4 do not appl y
The whole subtlety of the enterprise l ies i n mi ni mi si ng the ri sks,
in maki ng sure not to increase the val ue of Ni to poi nt W ( i n particu
l ar, in not giving it val ue F) unti l one is sure that thi s i ncrease wi l l
have no effect with regard to A; and i n not decreasi ng thi s val ue ( the
value R) unless al l effect with regard to B is excl uded. Thus Ni, per
petual l y maxi mi sing the neutral i sati on of the effects of order, sl i ps
in between B and A.
And, when the time for cl osure arrives ( rul e 5) , for a W si tuated
between residue ( Nb) and form ( Na) , we retroactively set the seal on
the tactics, arriving at a Number globally si tuated between B and A,
because it is protected, locally, from any prohi bi ti on agai nst thi s
possi bi l ity.
15. 16. Fundamental theorem, second part: Unicity
We have j ust indicated the strategy - combi ni ng l ocal , neutral i sing
patience with a global deci si on of cl osure - that al l ows the exi stence
to be establ ished, in every case, of at least one Number situated
between two sets of Numbers B and A such that ( i n an abuse
of notation) B < A. In vi rtue of the pri nci pl e of mi ni mal i ty of
ordi nal s, there must exi st at least one such Number of mi ni mal
matter: we wi I I consi der the property ' bei ng the matter of a Number
si tuated between B and A', and the mi ni mal ordi nal for this
property.
It remai ns to be shown that a Number of mi ni mal matter situated
between B and A is uni que, whi ch wi l l permit us to identify the
numerical cut between B and A.
Suppose that there were two: we woul d have the fol l owing
arrangement:
- with N
1
and Nz being of the same matter ( mi ni mal for this
l ocati on) .
Si nce Nl and Nz are of the same matter, N
l
< Nz means that the
di scri mi nant must be i n the residue of Nl and in the form of N1 Take
this di scri mi nant, w. Consi der the sub-Number N
1
/w of N
1
Si nce w
is in the resi due of N h thi s sub-Number belongs to the high set of
N1 : it is therefore l arger than N1 But, since w i s the di scri mi nant of
N I and Nz, and therefore the smal lest ordi nal to di scri mi nate between
them, then N
l
and Nz are identical up to w ( excl usi ve) . This means
that the sub-Number Ndw i s identical to the sub-Number Nz/w. The
di scri mi nant of Ndw and of Nz can onl y be w, which i s outside the
matter of N
1
/w and in the form of N2 Consequently, N
1
/w < Nz.
So fnal l y, we have the arrangement:
Which i s to say that N dw is also situated between B and A. But this
i s i mpossi bl e, given that it i s of lesser matter than that of N J
,
which
i s supposedl y mi ni mal for the location between B and A.
We must reject our i ni ti al hypothesi s: there are not two Numbers
of mi ni mal matter between B and A, there is only one.
The two sets B and A therefore determi ne uni vocal l y one Number
of mi ni mal matter situated between them. This Number wi l l be cal led
the cut of B and A, and we wi l l posi t that N = B/ A, each time that
N can be identi fed as the unique cut of B and of A.
1 5 . 1 7. There i s one very pecul i ar case of the cut: take two Numbers
N 1 and Nz such that N 1 < Nz And take, for B and A, the sets which
have for el ements onl y N I and only Nz, that i s, the singletons (N d
and ( Nz ) . We remai n wi thi n the parameters of the fundamental
theorem, which i s to say that there exi sts a uni que Number N
3
of
mi ni mal matter si tuated between N
l
and Nz We thus rediscover here
the cl assi c condi ti on of density of an order, whi ch we have mentioned
with regard to the rational s: between two Numbers there always
exi sts a thi rd, and thus an infni ty of Numbers . For us, besi des thi s,
there is an addi ti onal determi nati on: between N I and N2 there is
always a unique Number of minimal matter.
We can therefore put forward a pri nci pl e which everything gives
us to expect, and of which the uni city of the cut provides the i nfnitely
strong concept: the order of Numbers i s dense.
But more profound than this is the correlation i n thought between
this numberless density, this coal escence whi ch i nconsists in the
approach to all Number, and the possi bi l ity of counti ng for one the
Number of mi ni mal matter which cuts the fabric wi thout l acuna of
numerical ity at a certain poi nt.
' Cut' here designates the i nci si on of thought i n the inconsistent
fabric of being, that which Number sections from the ground of
Nature. It i s a concept of si ngul arity. Perhaps the concept of si ngul ar
i ty, at l east i n the order of being. For there i s that other si ngul arity
which cuts across being, and which is the event.
1 6
The Numberless Enchantment
of the Place of Number
16. 1 . A review, to begi n with.
1 A Number is an ordi nal - the matter of the Number, M( N) , in
which i s sectioned a part of that ordi nal - the form of the Number,
F( N) . We also consider that part of the ordinal-matter that fal l s
outsi de the secti on, outsi de the form: the resi due of the Number,
R( N) .
2 The location of an ordi nal with regard to Number N is its posi
tion in ( bel ongi ng or non-bel onging to) one of the three 'compo
nents' of Number: form, resi due, matter. There are three locations:
i n the form, i n the residue and outsi de the matter.
3 The di scri mi nant of two Numbers Nt and N2 is the smal lest
ordinal not to be located si mi l arly i n both Numbers. If no
such di scri mi nant exi sts, then the two Numbers are equal ( they
have the same matter, the same form, and therefore the same
resi due) .
4 Dependi ng on the location of the discri mi nant, we can defne an
order-rel ati on ( transi tive and non-refexi ve) between two di fferent
Numbers. We denote thi s through Nt < N2 and by sayi ng that Nt
i s smal l er than Nz This relation i s a total order over the domain
of Numbers i n the sense that, given two di fferent numbers Nt and
N2, it is al ways the case ei ther that Nt < N2 or N2 < Nt .
5 The order-relation is dense: given two Numbers Nt and Nz where
Nt < N2, there always exi sts an N3 whi ch comes in between Nt
and N2 : Nt < N3 < N2 .
NUMBERLESS ENCHANTMENT OF THE PLACE OF NUMBER 1 57
6 Take a Number NI of matter WI and an ordi nal WI smal ler than
WI ( so that WI E Wd. The Number of matter Wb which i s exactly
l i ke NI up to WI excl usive (the form of this Number being con
stituted by al l ordi nal s smal ler than WI that are i n the form of
Nd wil l be cal led a sub-Number of Nb a sub-Number of NI
which is a 'cut' of NI at poi nt WI . We denote thi s sub-Number
Ndw
l
.
7 Amongst the sub-Numbers of Nb some are smal l er than NI ( when
WI is i n the form of Nd, others are l arger than NI (when WI i s i n
the resi due of Nd. The former, gathered together, constitute the
low set of Nb denoted by Lo( Nd. The l atter constitute the high
set of Nb denoted by Hi ( Nd.
8 It can be proved that N I is the cut of its l ow set and its high set
in the fol l owing way: i t i s the Number of mi ni mal matter si tuated,
according to the order of Numbers, between the l ow set and the
high set ( l arger than every Number i n the low set and smal ler
than every Number i n the high set) .
9 More general l y, it can be shown that, given two sets of Numbers
such that all those of the frst set are smal ler than all those of the
second, there exists a uni que Number N of mi ni mal matter situ
ated between these two sets. Taki ng two such sets B and A, we
can say that thi s Number N is the cut of B and A, whi ch i s written
N = BfA. Thus NI = Lo( NdfHi ( Nd. This speci fed cut is cal led
the canonical presentation of N I .
16. 2. We wi l l now take a strol l through the borderless domai n of
Numbers, pointing out some of them, and i n particul ar al l those tra
ditional species: natural whole numbers, negative whole numbers,
ordi nal s, rati onal s, real s. But al so so many others, whi ch fnitude and
the wretchedness of our i nherited practice of Number keeps from us.
How negligi ble are numbers amongst Numbers ! The being of Number
exceeds in every di rection that whi ch we know how to negotiate. Our
strength, however, i s that we possess a way of thi nki ng of thi s excess
of being over thought.
16. 3. Zero
There i s a very distinctive Number, the Number ( O, O) , whose matter
is the void, and whose form, consequentl y, is also the void. Thi s
Number inscri bes as numeri cal gesture the absence of every gesture,
in defaul t of any matter. It is absol ute Zero, the Number wi thout
numerical ity. Of course, its ontological foundation i s the empty set,
the suture to being of every text, the advent of being qua being
to the thi nkabl e. There i s no doubt that it is this void that we are
thi nking here as Number. But thi nking it as Number makes a di ffer
ence. It is not for exampl e the same thi ng, not the same Number, as
it woul d be i f the voi d was onl y i n the posi ti on of matter, or only in
the position of form. The number ( ( 0) , 0) , or ( 1 , 0) , whose form is
void, cannot at all be i dentifed as the Zero of Number. Certai nl y, the
act of secti oni ng it i s equal l y nul l , i t doesn' t extract anything from
its matter, but thi s matter suhsists unaltered, constituting, in the
absence of any act, the real substance of that whi ch thi s gesture never
even started. The only true Zero i s that which subtracts itsel f from
al l numeri cal gesture because it has nothing, no materi al or natural
mul ti pl icity, upon which this gesture could be carried out or not
carried out. Zero is thus outside all appreci ati on, positive or negative,
of the act of numeri cal secti on. It i s, very precisely, neither positive
or negati ve. It subsi sts in itsel f, inaccessi bl e to al l eval uabl e action.
Zero i s being qua bei ng thought as Number, from wi thi n ontology.
16. 4. Si nce we have said, a l ittle metaphori cal l y, that Zero i s neither
positive nor negative, can we not give a precise numerical sense for
these adj ecti ves ? Elementary arithmetic al ready i ntroduces - to the
obscure rel i sh of every schoolchi l d - whole negative numbers such
as
=
4
Consi der for exampl e the Number N, whose matter is the l imit
ordi nal c, and whose form has only the ordi nal as element. Which
i s to say that the form is the si ngleton of 0, and that the number NJ
can be wri tten: ( c, ( O) ) . I f we compare thi s Numher t o Zero, that i s,
to ( 0, 0) , we can cl earl y see that thei r di scri mi nant i s 0, which is in
the form of N, and outsi de the matter of Zero ( any ordi nal whatso
ever, including 0, is outside the matter of Zero, whi ch has no matter) .
The rul es of order indicate to us then that N, is larger than Zero.
It makes sense to say that N, is positive.
Consi der now the Number Nz, whose matter i s al so the l imit
ordi nal c, but whose form i s thi s time the singleton of 1 . This
Number Nz can be wri tten ( c, ( l ) ) . Once agai n, the di scri mi nant of
Nz and Zero is 0. It can be found thi s ti me in the resi due of Nh si nce
the form of Nz does not contai n ( i t onl y contai ns 1 ) , but its matter,
c, does contai n it, c being the l i mi t col lection of all the fni te ordi nal s,
i ncl uding of course. We can see, then, that 0, bei ng outside the
matter of Zero and i n the resi due of Nz, N2, i s smaller than Zero.
So i t makes sense to say that Nz i s negative.
16. 5. Posi ti ve Numbers and negative numbers
Our examples can be general i sed i n the fol l owing fashi on: the dis
cri mi nant between Zero and any other Number whatsoever is always
the empty set O. For Zero is the onl y number whose matter is voi d,
and therefore the onl y Number where 0 i s located outsi de the matter.
For every other Number, 0 i s l ocated i n the form or i n the resi due.
And, si nce 0 is the smallest ordi nal , it i s certai nl y the discri mi nant of
Zero and of every number other than Zero.
The situation i s very simple, then: i f any Number other than Zero
has 0 i n its form, then it is l arger than Zero. If, on the other hand,
o is i n i ts resi due, it is smal ler than Zero, si nce 0 wi l l always be
outside the matter of Zero.
We will thus defne posi ti ve and negative Numbers in the fol l owi ng
way: A Number is positive if 0 is an element of its form. It is negative
if 0 is an element of its residue.
16. 6. Some signifcant consequences of the defni ti on of positive and
negative Numbers:
1 Si nce Zero is wi thout matter, wi thout form and wi thout resi due,
o cannot be an el ement either of the form or of the residue of
Zero. The description i n 16. 3 i s thus transformed into a mathe
matical concept: Zero is neither positive nor negative.
2 Zero is not at al l the smal lest Number. It is l arger than every
negative Number, and negative Numbers consti tute, to al l appear
ances, a l i mitless, inconsi stent domai n. Between the negative
Numbers and the positive Numbers, Zero l ies at the centre of that
which has no periphery.
3 Zero is not defned by extrinsic operati ons, it is not i ntroduced as
the ' frst' term of a successi on, nor as the ' neutral element' of an
operation (an attri bute which it possesses i ncidental l y and second
ari ly) . It is characterised by its numerical being. We have not
strayed from our ontological path, which subordi nates all opera
tional or algebraic consi derations to i mmanent characteri sati on.
4 More general l y speaki ng, the categories ' positive' and ' negative'
have been introduced into the consideration of the order of
Numbers onl y for convenience of exposi ti on. The predicate ' has
o i n its form' or ' has 0 i n its resi due' are whol l y i ntri nsi c. The
examination of the bei ng of a Number al one tel l s us whether it
is positive or negative, wi thout comparing it with any other
Number.
S Positivity does not depend i n the least upon the ' quantity'
of the matter of a Number, or the si ze of its form, but onl y
upon the l ocation of the voi d. The Number ( 2, ( 0) ) i s posi ti ve,
whi lst the Number ( 0, ( 0 0) ) , whose matter is 0 and whose
form takes i n all of this matter apart from 0, i s negati ve. There
is fnite posi ti ve numeri cal i ty, and i nfnite negative numericality,
regardless of whether the question is one of matter or one
of form.
6 If a Number N is posi tive, then, since 0 is in its form and is neces
sari l y mi ni mal , it fol l ows that every sub-Number Nlw of N ( except
for Zero, whi ch is a sub-Number of every Number, the sub
Number N/O) is al so posi ti ve: the elements of the form of Nlw
are actual l y the elements of the form of N up to the ordi nal w,
and, unless w is 0, 0 wi l l be amongst these elements, since N is
posi ti ve. Si mi l arly, every sub-Number of a negative Number N,
apart from 0, i s negative ( it has 0 in its resi due, as N does ) . In
particul ar, the non-nul l elements of the low set and al l the ele
ments of the high set of a positi ve Number are positive; l i kewise,
all the elements of the low set and all the non-nul l elements of
the high set of a negative Number are negati ve.
1 6. 7. Meditation on the negative
The concept of negativity, as proposed by the uni verse of Numbers,
i s every bi t as profound as its apparent paradoxical ity suggests. One
might thi nk at frst that negativity consisted preci sel y i n the incorpo
ration of the voi d i nto the form of Number. Isn't there more positivity
in a form that has not been marked by the stigma of nothingness?
I sn' t the plenitude of the numeri cal secti on better assured i f it expels
from its posi ti ve production that dubi ous index of the multi pl e that
al lows no presentati on?
Number enj oi ns us here to di sabuse ourselves of any remai ni ng
temptation towards an ontol ogy of Presence. If the lack of voi d in
the form of Number seems ' positi ve' , thi s is the case onl y i f we
i denti fy bei ng wi th the pl enitude of the effectively presented. We
are then tempted to i ndex to the negati ve every occurrence of that
which presents nothing, every mark whose mul ti pl e-referent i s sub
tracted. But the truth is entirely otherwi se: it i s precisely under this
mark that being qua being comes to thought. I n which case there
is less ontological dignity in a Number that does not retai n this
mark in its form than i n a Number that does so retain i t. It is from
the poi nt of the voi d that the dignity of being, the superiority of a
Number, can legitimately be measured. Numerical superiority is the
symbol of thi s superi ority with regard to what is at the disposal of
thought.
The ontol ogi cal cl arity ( for a subtractive ontology) of the state
ment ' a Number i s negative i f the mark of the void i s i n its residue'
underlies what might be cal led the ethi cal verdict of Number. I hope
to show one day that what i s Evi l , i n any si tuati on where the void is
attested to ( and such, si ngul arl y, are post-evental si tuati ons) , is the
treating of that testi mony preci sel y as i f it were a resi due of the si tu
ation. What is Evi l i s to take the void, whi ch i s the very bei ng of the
situation, for unformed. The forms of Evi l decl are substance ful l and
l uminous, they expel every mark of the voi d, they rusticate, deport,
chase off, extermi nate those marks. But the verdict of Number tel l s
us: it is in thi s cl ai m to ful l substance, i n thi s persecuti on of the occur
rences of the voi d, that res ides, preci sel y, the negati ve. A contrario,
positivity assembl es and harbours the marki ng of the voi d wi thi n its
forms. And, this being so, it accords thought to being i n an i ntri nsi
cal l y superi or fashi on.
To take the voi d for a resi due i s a negative operati on, a detestabl e
' puri fcation' . Every true pol itics, i n fdel i ty to some popul ar event,
takes on the guardi anshi p of the voi d - of that which is un presented,
not counted, i n the situation - as its highest duty i n thought and in
acti on. Every poem seeks to uncover and to carry to the formal l i mits
of language the latent void of sensi bl e referents. Every sci ence treats
positively the resi due of its own hi story, that whi ch has been l eft
outside of its form, because it knows that preci sel y there dwel l s that
which wi l l refound and reformul ate i ts system of statements. Al l l ove
ultimately establ i shes itsel f i n the j oy of the empty space of the Two
of the sexes which it founds, and from this point of view the romantic
idea of a ful l , fusi onal l ove, under the puri fed sign of the One, is
preci sel y the Evi l of love.
The negative, as its concept is establ i shed by Number, is a punctual
di scord of thought and of bei ng. 'Negative' i s every enterprise of
formation whi ch abandons, fai l s to cheri sh, thi s uni que point upon
whose basi s there can be forms and the unformed, forms and resi
dues; the poi nt where being, i n the gui se of the un presented, assures
us that we do not thi nk i n vai n.
16. 8. The symmetric counterpart of a Number
Not much needs to be done in order to ' negativise' a positive Number:
it suffces to remove 0 from its form. Number teaches us the precarity
of the positive, its a-substanti al character. It is at the mercy of the
transfer of one single point to the resi due. And this point i s the most
transparent of al l , that point that i s not supported by any multi ple
presentation: the mark of the voi d.
Thi s i dea of the transfer of a term from one l ocati on ( here, the
form) to the ' opposite' l ocati on ( here, the resi due) can be general i sed.
Take a Number N and the Number obtained by inverting the form
and the residue of N: The resi due of N is promoted into the form,
whilst al l the terms of its form are demoted into the resi due. Thi s
new Number operates, in the same ordinal-matter, a cut i nverse or
symmetrical to that whi ch defnes N. We wi l l cal l thi s Number the
symmetric counterart of N ( indicating a symmetry whose centre, as
we shal l see, is Zero) . We wi l l denote by -N, and read as ' mi nus N' ,
the symmetric counterpart of N.
A Number and i ts symmetric counterpart can be presented as
fol lows ( usi ng the di agrams introduced i n 1 2. 3) :
F (N) F (N)
*
Number N :
R (N) R (N)
I
I
I
F (-N) F (-N)
*
Number-N :
R (-N) R (-N)
It i s clear i n the di agram that N i s positive (0 i s in its form) and
that -N, its symmetric counterpart, i s negati ve. Evidentl y, thi s wi l l
al ways be the case. Conversely, when N is negative ( 0 i s i n i ts resi due) ,
-N is positive ( 0 i s in i ts form) .
I f we take the symmetric counterpart -N of N, then the symmetric
counterpart -(-N) of -N, we arrive back at N: we have changed the
form into the resi due, and then the resi due into the form. It i s that
old law l earnt in the school room, which spontaneousl y opposes itsel f
both to Hegel and to i ntui ti oni sm: two negati vi si ng operations take
us back to the i ni ti al affrmati on. However, one must sti l l take care,
as al ways, to note that -( -N) i s not necessari l y a positive Number.
If the starting Number N is negative, its symmetric counterpart is
positive, and the symmetric counterpart of its symmetric counterpart
- whi ch is itsel f - is once agai n negative. The sign '-' is not a sign of
negati on but one of symmetry. Whi ch confrms for us that the nega
tive ( unl i ke the symmetrical ) is not an operati onal di mensi on. It is a
structural predicate of the being of Number.
16. 9. A few exampl es.
What i s the symmetric counterpart of the posi ti ve Number ( O, ( O) ) ?
I t i s the Number ( 0, ( 0 - ( 0) ) ) , whose form i s al l of 0 except for O.
It i s obvi ousl y negative.
What is the symmetric counterpart of the negative Number ( 2, ( 1 ) ) ,
whose form is the si ngleton of 1 ? It i s the positive Number ( 2, ( 0) ) ,
whose form i s the si ngleton of ( 0) . In fact, the onl y elements of the
ordi nal 2 are 0 and 1. In the former case, 0 constitutes the resi due,
in the l atter, the form.
Take a positive Number N and its symmetric counterpart -N. To
every Number si tuated ' between' Zero and N we can make corre
spond a Number situated ' between' -N and Zero: we j ust take its
symmetric counterpart. In fact, it is clear that, where it i s the case
that Zero < N, < N, it is also the case that -N < -N, < Zero. Thi s
can be verifed by exami ni ng al l possi bl e cases of i nequal ity between
N, and N ( see 1 3. 1 3) , remembering that -N swaps the form and
residue of N.
There are thus ' as many' Numbers between -N and Zero as there
are between Zero and N, because the function (( NI l -N, is a bi uni
vocal correspondence between the two ' sl ices' of Numbers . But take
care! The correspondence i s not between two sets. The i nterval
between Zero and N is not a consi stent total i ty any more than the
entire domain of Numbers i s. Thi s can easi l y be proved: taki ng, for
example, the Number ( 2, ( 0) ) , we know that all Numbers of the type
( W, ( O) ) , where W i s any ordi nal whatsoever l arger than 2, are smal l er
than ( 2, ( 0) ) . It i s the law that we di scovered in 1 3. 16: i f the form
stays the same and the matter is increased, the Number gets smal l er.
Meanwhi le, al l Numbers ( W, ( O) ) are posi tive, since 0 i s i n thei r form.
So there are ' as many' of these positive Numbers - that i s, those si tu
ated between Zero and ( 2, ( 0) ) - as there are ordi nal s l arger than 2.
But we know for sure that ' al l ordi nal s l arger than 2' is an i nconsi s
tent mul ti pl icity.
Keeping thi s i n mi nd, we can al l ow ourselves to vi sual ise symmetry
in the fol l owi ng way, the axis being that of Numbers taken according
to thei r order:
-N

N
This j ustifes our speaki ng of a symmetry whose centre is Zero.
16. 1 0. The ordinals
We announced a long time ago ( see for example 8. 8) that the ordi nal s,
whi ch constitute the stuff of the bei ng of Numbers, can al so them
selves be represented as Numbers. What do the Numbers that
represent ordi nal s look l i ke?
Let' s consi der the Number (W,W) , whose matter is the ordi nal W
and whose form retai ns all of thi s matter. In other words, thi s is a
case of a maxi mal numerical secti on, or of exhi biting - as certai n
contemporary artists have done - the raw materi al al one as the
' work' . The most interesting thi ng is to compare the Number ( W,W)
with the Number ( W, O) , whose form i s voi d. In both cases, we feel
that the act is somehow nul l . But the two nul l ities are di stinct. The
Number ( W, W) treats the whole of the matter as a form, whereas the
Number ( W,O) does not inscribe any form in the matter. The i mmedi
ate resul t is that ( W,W) , for any W other than 0, i s a positive Number,
whereas ( W,O) is a negative Number ( remembering that 0 is not an
element of 0, and that therefore 0 i s not i n the form of ( W, O) ) . We
di scern a certain posi ti vity in the frst gesture which designates the
matter as form, whereas the second, overwhelmed by the matter, is
unabl e to designate anything whatsoever.
But i f ( W, W) is treated as a posi ti ve production, the assumption
of a matter as form, it remai ns nevertheless a fact that thi s production
repeats the ordi nal-matter. This redoubl i ng of the ordinal (as matter,
then as form) legi ti mates our treating Numbers of the form ( W,W)
as the Numeri cal representatives of the ordi nal s.
We wi l l therefore say the fol l owi ng: An ordinal W is presented as
Number in the form ( W, W); that is, the Number whose matter is W
and whose form is W. Thi s presentation is the ordi nal ' itself', but
thought as Number.
1 6. 1 1 . To be sure that this i s the ordi nal ' itsel f' , we must expl icitly
prove that the order of Numbers respects the order of the ordi nal s,
whi ch i s bel ongi ng. I n other words, that i f it i s the case - ordi nal s
bei ng thought i n thei r own domai n - that WI E W 2, then it i s al so
the case - ordi nal s being thought as Numbers - that ( W"Wd <
( W2 ,W2 ) .
Thi s i s obvi ousl y the case. Because the di scri mi nant of ( W"Wd
and ( W2,W2 ) i s necessari l y the smal lest ordi nal to belong to WI and
not to W2, or to belong to W2 and not to WI ' If WI E W2, this smal lest
ordi nal i s preci sel y W" which belongs to W 2 but cannot belong to
itsel f. Now WI i s outside the matter of ( W"Wd, and it is in the form
W2 of ( W2,W2 ) . So it is i ndeed the case that ( W"Wd < ( W2,W2 ) .
Thus the order of the ordi nal s thought as Numbers, i n the formal
redoubl ing of their materi al being, i s the same as the order of ordi nal s
thought in thei r bei ng, as transitive sets al l of whose elements are
transi ti ve. The Numeri cal representation of the ordi nal s is structur
al ly i somorphi c to the ordi nal s. This being so, there is no reason why
we shoul d not consi der that the ordi nal s ' themselves' are i nscri bed,
identical l y represented, in the order of Numbers.
16. 12. From the fact that an ordi nal i s a Number of the form ( W,W) ,
three consequences can be drawn:
1 Every sub-Number of an ordi nal is an ordi nal . For, if ( W, W) is
an ordi nal , a sub-Number i s of the form ( w"wd, where WI E W.
It is therefore the ordi nal WI '
2 All these sub-Numbers wi l l mani festly be ordi nal s smal l er than
the i niti al ordi nal . It fol l ows that they are all i n the l ow set of
the i ni ti al ordi nal , and that the high set, general l y composed of
sub-Numbers l arger than the Number, i s empty here. Thi s i s a
characteristic property of ordi nal s thought as Numbers. Gener
al l y speaking, a sub-Number of the high set i s a sub-Number
Nlwl such that WI i s i n the residue of N. But, i n the case of
an ordi nal ( and thi s coul d be a defnition of the ordi nal s) , the
residue is empty. The high set of an ordi nal is therefore al so
empty; and, conversely, i f the high set of a Number is empty,
then its resi due i s empty: its form coi nci des with i ts matter; it
is an ordi nal . The canonical presentation of an ordinal wi l l
therefore be of the form Lo( W)/O. But what i s more, as the
low set has for its elements all ordi nal s smal ler than W, it is,
as a set, identical to W ( every ordi nal is the set of all the
ordinals smal ler than it, 1 1. 2) . Fi nal l y, the canoni cal representa
tion - most di stinctive - of an ordinal W thought as Number
is si mpl y W/O.
3 The symmetric counterpart of an ordi nal ( W, W) is obtai ned by
swapping the resi due and the form. Now, the residue is empty.
So it i s the voi d that wi l l be substituted for the 'total ' form
that i s W: the symmetric counterpart of ( W, W) i s the Number
(W,O) . Thought as Number, an ordi nal W al lows of a symmetric
counterpart, so we can freely speak of the Number -W.
It i s clear that every ordinal apart from i s a positive Number, si nce
its form, W, contai ns as an element. The symmetric counterpart
of every ordi nal other than the voi d is therefore a negative Number,
as can be seen di rectly i n writing ( W, O) . It will be found, moreover,
that all the properties of an ordi nal W are i nverted by the passage
to -W. So that now every sub-Number of -W is the symmetric
counterpart -WI of an ordi nal WI smal l er than W; and it i s the low
set of -W that is void, si nce - the form of -W being voi d - every
sub-Number of -W i s l arger than it; and, fnal l y, the high set of -W
i s identical to -W, with the resul t that the canoni cal representation
i s: Ol-W.
We are thereby assured that ordi nal s are Numbers.
l
But what i s
more, grasped in terms of Numerical i ty, the ordi nal s are symmetrici s
able: we have opened up on the other side of Zero ( whi ch is the
ordi nal 0, thought as Number) an i mmense space where wi l l be
i nscri bed those formerly unthi nkabl e entities: natural multi pl icities
submitted to the negative. Numericality i s capable of symmetricising
nature.
1 6. 1 3. Posi ti ve and negati ve whol e numbers
The natural whole numbers, thought i n their being, are none other
than the fni te ordi nal s, which is to say the elements of the frst
l i mi t ordinal . In fact we have al ready given their defnition and di s
cussed thei r operati onal di mensi ons in chapter 1 1 .
Thi s preceding work al ready settles the question, then: thought as
Numbers, natural whole numbers are of the type ( n, n) , where n is a
fnite ordinal. Evidently, they are al l positive. The order of natural
whole numbers qua Numbers coi nci des with the order of natural
whole numbers that we al ready know, the order accordi ng to which
every school boy says that n is l arger than p. For we know that, i f
p e n whi ch i s the ontol ogi cal versi on of traditional order - then
(p, p) < ( n, n) in the order of Numbers. We therefore have the right to
write the Number ( W, W) as W, to i ndicate that an ordi nal ' itsel f' is
being i nscri bed i n the domai n of Numbers. We therefore write a
natural whol e number, thought as Number, as n.
The sub- Numbers of a natural whol e Number are the fnite ordi
nals smaller than it, therefore the natural whol e numbers smal l er than
i t. If n is this Number, these will be natural whol e Numbers
( 0,0) , ( 1 , 1 ) , . . . , ( ( n - l ) , ( n - 1 ) ) , whi ch we coul d also write as
0, 1 , . . . , ( n - 1 ) . Taken together, they form the low set of n. The high
set of n i s empty, and the canoni cal representation of a whole n,
thought as Number, is ( 0, 1 , . . . ,(n - 1 ) )/0. Si nce n' s elements are
preci sel y 0, 1 , . . . ,(n - 1 ) , the l ow set whose elements they are can be
written as n/O. ( NB Thi s is not ci rcul ar, because, considered as a set,
n does not contai n itsel f as an el ement) .
The symmetric counterpart of a natural whol e number is a Number
of the form ( n, O) , where n i s a fnite ordi nal . We wri te it -n, we say
' mi nus n' . We posit that a Number is a whole negative Number if it
is the symmetric counterpart of a natural whole Number, that is, one
which takes the form (n, O) . The sub-Numbers of a negative whole
number -n are all the whole numbers p, where p e n. Taken
together, they form the high set of -n, whose l ow set i s empty. The
canoni cal representati on of a negati ve whol e Number is therefore
ul ti mately written as Of-no
In order to confrm the compl ete identity of the traditional positive
and negative whol e numbers and of the positi ve and negative whole
Numbers, it must obvi ousl y be the case that operations on these
Numbers coi nci de, as order did, with operations on numbers. If for
example we defne an addi ti on NI + Nz on Numbers, then the res ult
of thi s operation in the specifc case of whol e Numbers, hence in the
case of an addi ti on of the type m + n, shoul d be 'the same Number'
as the whole number whi ch, i n the cal cul ati ons of our school days,
corresponded to the addi ti on of these two whole numbers. These
operati onal verifcations wi l l be carried out i n chapter 18.
So far as the inscription wi thi n the Numbers of natural whol e
numbers thought i n their being i s concerned, our task i s complete.
16. 14. Dyadic positive rational numbers
We have al ready spoken of rati onal numbers in relation to Dedeki nd
cuts ( compare 1 5. 5) : a posi ti ve ( or nul l ) rati onal number is a fracti on
or relation % of two natural whol e numbers, which i s to say a pai r
(p,q) of whole numbers. The frst i s cal led the numerator, the second
the denominator. The numerator can be null ( i denti cal to the empty
set) , but it is prohi bited for the denomi nator to be 0 ( we know that
the rel ati on t i s ' undetermi ned' ) .
We have no desi re here t o enter into a rigorous introduction to
these tradi ti onal numbers ( in fact, here we must consider fracti ons as
i rreduci ble, impossi bl e to si mpl i fy) . The intuitive idea of the fraction
wi l l suffce for us.
It is evident that the natural whol e numbers are a subset of rational
positive or nul l numbers; we j ust need to take a rati onal in the form
t to obtain n. In other words: a whole number i s a rati onal of the
type (n, 1 ).
The cl assi cal order of the rational s has the fundamental property
of being a dense order. In other words ( see 1 5. 5) , given two rati onal s
* and : such that * < , however ' near' these two numbers might
be, there always exi sts a third ( and, from there, an i nfnity of them)
which comes between the two i ni ti al numbers: there i s a :' such that
< <
,
q, q, q
z
'
A dyadic rational number is a number of the form J whose
denominator i s a power of 2. Or, in our pai red versi on, a rati onal
number (P,2).
Dyadic rati onal numbers themselves form a dense subset of the
rational s: i f rl and r2 are rati onals such that rl < r2, a dyadic rati onal
can always be intercal ated between them.
The i mportant thi ng for us is that every sequence of augmenti ng
rationals rl < r2 5 . . . < ro < . . . can be ' repl aced' by a sequence of
dyadic rational s dl < . . . < do < . . . : take the dyadic rati onal s si tuated
' between' rl and r2, then r2 and r
3
, etc. We can also say that the dyadic
rationals form a ' basi s' for all the rational s. More speci fcal l y, a non
dyadic rati onal number can be ' approached' as cl osel y as you l i ke
from a dyadic one, because you can al ways l odge a dyadic i n-between
r and r + r
'
, however smal l r
'
might be.
1 6. 1 5. We have, then, the fol l owing statement, perhaps the most
i mportant i n the process of the representation of ( traditional ) numbers
as ( ontological ) Numbers:
Every dyadic rational number can be represented as a Number of
fnite matter, and every Number of fnite matter represents a dyadic
rational number.
1 6. 1 6. How, in general , is a Number of fni te matter presented? I n
t he form ( n, (pt .P2, ' " P, ) ) , where t he whol e numbers PI . P2, etc.
whi ch make up its form are whol e numbers smal l er than n, the matter
of the Number. Si nce we are keeping to rational positive numbers,
we wi l l consi der here only Numbers of positive fnite matter, that is,
Numbers which have i n their form.
The subtle i dea that gui des the ' proj ecti on' of these Numbers of
fnite matter into the dyadic posi ti ve or null rati onal s is the fol l owing.
Let n be the matter of the Number. We take all the elements of this
matter i n order, from to n 1 , which is the l argest whole number
contained in n. In so far as we stay in the l ocation of the frst element,
which is the form, si nce the Number i s positive - we attri bute
the val ue 1 to the whole number i n question. Say that we come across
the frst el ement of n say p not to have the same location as 0,
in other words the smal lest whol e number p i n the matter of the
Number to be i n the resi due. We attri bute to thi s whol e number the
value
e After this, we general l y attribute to whole numbers q

whi ch fol l ow the val ue z if they are in the form, the val ue -z
i f they are i n the resi due.
Fi nal l y, the val ue of the l ast term beyond p ( p bei ng sti l l the frst
term whi ch does not have the same l ocation as 0, i f it exists) , the
val ue attri buted to n 1 , then, wi l l be *, with or without the sign
dependi ng on whether n 1 is i n the resi due or i n the form.
Or, once agai n: a belonging to the resi due wi l l always be affected
by the si gn -. In traversi ng in order al l the el ements q of n, in so far as
one remai ns wi thi n the form, whi ch i s the l ocation of 0, each element
is counted for 1 , for a whol e val ue. As soon as the location changes,
we count the elements for a dyadic rati onal of the form z, where
p is the frst for which the location changes, from now on adding the
sign - whenever thi s location i s the resi due.
Fi nal l y, we associ ate with the i ni ti al Number of fnite matter the
rati onal number obtai ned from the sum ( i n the usual sense) of al l
the values thus attri buted to the elements of n. Thi s rati onal number
is dyadi c, si nce al l the denomi nators i n question are dyadi c, and
si nce - as every schoolchi l d knows - to add fractions, one takes as
denominator the smal lest common mul ti pl e of the denomi nators.
Now, the smal lest common multiple of powers of two i s a power
of two.
16. 1 7. Let' s gi ve an exampl e of the procedure. Take the Number
( 5, ( 0, 1 ,3) ) , whose ( fnite) matter i s the ordi nal 5 and whose form
contai ns 0, 1 and 3. The resi due i s thus composed of 2 and of 4.
Si nce i s i n the form, we give it the val ue 1 .
Since 1 i s al so i n the form, we give it the val ue 1 .
The location changes with 2, which is i n the resi due. We gi ve it
the val ue
=
t
3 h f . . h I
1
1
IS In t e orm; we gI ve It t e va ue 2'-2 , 1
=
2'
4 is in the resi due, we give it the val ue
-
/2.
1
=
-
.
So, in the end, the rati onal number correspondi ng to the Number
( 5, ( 0, 1 ,3) ) wi l l be obtained from the sum:
1 1 1 1 3
1 + 1 - -+ -- -= -
2 2
2
23 23
We can see very wel l that thi s is indeed a dyadi c rational .
16. 1 8. In order better to exhi bi t the construction of thi s correspon
dence, which bears wi tness to an i somorphy, an identity of bei ng,
between positive Numbers of fnite matter and posi ti ve dyadic rati o
nal s, we wi l l formal i se thi ngs a l ittl e. We wi l l then see cl earl y that we
are deal i ng with an i nductive defni ti on, a defni ti on by recurrence.
Take a positive Number of fni te matter. We will defne by recur
rence the fol lowing function (, defned on the elements of the matter
n of the Number:
KU 1 : (( 0) = 1 .
K L 2 : ((p + 1 ) = 1 , i f ((p) = 1 for al l whol e numbers up to and
including p, and i f p + 1 i s i n the form of the Number.
K U 3: ((p + 1 ) = -t i f al l the whol e numbers up to and i ncl ud
ing p are i n the form and p + 1 i s i n the resi due.
K L 4: (( p + 1 ) =
2
!. 1
if the val ue of p is iq or - 2
1
q and p + 1
is in the form.
1 70 ONTOLOGY: DEFI NI TI ON. ORDER. CUTS. TYPES
KL 5: ((p + 1 ) = - if the val ue of p is f or -f and p +
1 is in the resi due.
These rul es wi l l al low us to cal cul ate the rati onal val ue of ( for al l
the el ements of n, the matter of the i ni ti al Number. Usi ng Ra( N) to
denote the dyadi c rati onal that corresponds to N, we then posit
that:
Ra( N) = (( 0) + (( 1 ) + . . . + (( n - 1 )
The sign + i ndicates here the algebraic sum i n the normal sense.
It i s clear that Ra( N) i s a dyadic rati onal .
16. 19. Let' s proceed with the cal culation of another exampl e, the
Number ( 4, ( 0, 1 , 3 ) ) , which is, of course, a positive Number of fnite
matter:
So:
(( 0) = 1 ( by rul e 1 ) .
(( 1 ) = 1 ( by rul e 1 ) .
(( 2) = -. ( by rul e 3; 2 is in the residue) .
2
1 1
3 ) = -- - ( rul e 4,
'
3 i s i n the form) .

2
1 + 1
-
2
2
Ra( ( 4, ( 0, 1 , 3) ) ) = (( 0) + (( 1 ) (( 2) + (( 3) .
1 1
Ra( ( 4, ( 0, 1 , 3) ) ) = 1 + 1 - "+ 2'
Ra( ( 4, (0, 1 , 3))) = , whi ch is a dyadi c rati onal , as we sai d it
woul d be.
2
1 6. 20. Whole ordinal part of a Number
It might appear strange peremptori l y to change the procedure when
we get to the frst whole p that doesn' t have the same location as
o i n the Number of fnite matter under consi deration. Gonshor
real ises thi s: ' The whole idea of a shi ft from ordi nary counti ng to
a bi nary deci mal computation at the frst change in sign may seem
unnatural at frst. However, such phenomena seem i nevi tabl e in a
suffciently rich system. ,
2
Thi s expl anation of Gonshor' s - more of
an apology, real l y - i s a l ittle qui ck.
To fnd the true underlying concept, we shoul d ask what i s actual l y
represented by the p frst consecutive ordi nal s of a Number N of
fnite matter whi ch have the same location as the i ni ti al term O.
Assume once more the posi ti ve case ( 0 l ocated in the form) . If we
partition N at point p ( the frst ordi nal , i n the order of ordi nal s, to
change location) , we obtain the sub-Number NIp all of whose ele
ments have the same location as o. It is cl ear that, si nce thi s l ocati on
is the form, NIp i s the whol e posi ti ve Number p, that i s, the Number
whose matter i s p and whose form is made up of the elements of p.
The function f wi l l attri bute the val ue 1 to al l these elements, and
the sum of the val ues 1 + 1 + . . . wi l l gi ve the ' cl assi c' whol e number
p. Whi ch, we can add i mmediately, i s an algebrai c sum of dyadic
rati onal s of the type or -
2
1
q , where q i s no more than 1 . It fol l ows
that Ra(N) will be the sum of the whole p and a negative dyadic
fraction between -1 and 0 (at least, unless it happens to be a whol e
number) . Fi nal l y, p i s a type of whole part of the positive rati onal
Ra( N) , that i s, the natural whol e number closest to Ra( N) ' from
above' : ( p - 1 ) < Ra( N) < p.
From the point of view of Number, in fact, p i s the l argest sub
Number of N to he an ordinal, si nce ' bei ng an ordinal ' means pre
ci sel y bei ng a Number al l of whose matter i s i n its form. That the
location changes at poi nt p ( p i s i n the resi due) means preci sel y that
NIp + 1 i s no longer an ordi nal ei ther, since p, an el ement of its matter,
is i n the resi due. It i s therefore even more ftting to say that p i s a
'whole part' of N. By whi ch we mean: the largest whol e number p
belonging to the matter of N and such that the sub-Number NIp is
the ordinal p. Or even more simply: the whol e part of N is the l argest
ordi nal to be a sub-Number of N.
Now the procedure becomes clearer: i t works frstly by maki ng
correspond, vi a f, the elements of the whol e part of N and the
whole part ' from above' of the dyadic rational Ra( N) . The 1 val ues
are used to do thi s. And then it is a question of cal culating the
remainder, whi ch is less than 0, but more than -1 , and to do thi s
we use dyadi c fractions of the type j or -j , q i ndi cati ng the
rank of the ordi nal i n question beyond the whole part p. There
is no ' unnatural ' mystery in al l of thi s, but rather a profound
logic.
16. 21 . We can general ise these remarks. Given a posi ti ve Number N
of matter W, we wi l l cal l whole ordinal part of N the l argest ordi nal
WI E W such that the sub-Number Nlwi is the ordi nal WI .
The attentive reader wi l l balk at thi s: how can we speak so freely
of the ' l argest ordi nal ' to sati sfy a property? Doesn't the existence of
l i mit ordi nal s mi l itate agai nst any such cl ai m? Where ordi nal s are
concerned, onl y mi ni mal ity is at work.
The remark i s wel l taken. We wi l l have to reformul ate our defni
ti on, then, and posi t the fol l owi ng: The whole ordinal part of a posi
tive Number is the smallest ordinal located in the residue. Since the
Number is positive, 0 is located i n the form. The smal lest ordi nal
l ocated i n the resi due i s thus indeed the frst ordi nal , i n the ascending
order of the ordinal s of the matter of N, al l of whose elements are
in the form, although it itsel f is in the resi due. These el ements con
stitute the whole ordi nal part of N. Here i s a case where ' the l argest'
transl ates as ' the smal lest' .
I f WI i s the whol e ordi nal part of a posi ti ve Number N then, j ust
as in the above case, N < w" si nce the sub-Number WI is, considered
as an element of the matter of N, in the residue of N, whereas it is
outside its own matter.
It can al so be sai d that the whole ordi nal part of a positive Number
is i n the high set of that Number.
If WI i s a successor ordi nal , once agai n we fnd the ' framing'
)
of
the endpoi nt. Let W2 be the predecessor of WI ; thi s gives WI = S( W2 ) '
Si nce WI is the smallest ordi nal t o be i n the residue, its predecessor
W2 must be in the form. Of course, si nce all the el ements of W2 are
el ements of WI ( transi ti vity of ordinal s ) , and al l the elements of WI
are i n the form, al l the elements of W2 are too; s o N1wz i s the ordi nal
W2 . And, gi ven that thi s ordi nal i s outsi de i ts own matter and in the
form of N, then W2 < N, and so fnal l y W2 < N < S( W2 ) = WI ' This is
the i nterval we are l ooki ng for.
If, on the other hand, WI is a l i mit ordinal, it will certai nl y always
be the case that N < WI , but we woul d search in vai n for the l argest
ordi nal smal ler than N, because on the other side of WI there is no
' predecessor' . N woul d then have a si ngul ar posi ti on: smal ler than a
l i mi t ordi nal , it would be larger than all the ordinals smaller than
this limit ordinal. It woul d come to insert itself in that space we
thought was ' fl led i n' by the ordi nal s that precede the l i mit, the space
' between' a l i mi t ordi nal and the i nfnity of successor ordi nal s of
whi ch it is the l i mi t.
1 6. 22. Let' s gi ve an exampl e. Take the Number N = ( S( O) , S( O) ( 0) ) ,
whose matter i s the successor of 0 and whose form is al l of that
matter except for 0 itsel f, whi ch is the only element of the residue.
The l i mit ordi nal m being the frst ordi nal i n the matter of N to be
in its resi due, is the whol e ordi nal part of N. It is i ndeed the case that
N < , since thei r di scri mi nant is which is in the resi due of N and
outsi de the matter of . What i s more, for every el ement of - that
i s, for every natural whole number n - it i s the case that n < N, si nce
n i s outside the matter of n and i n the form of N. The Number N is
thus at once smal l er than the frst limit ordi nal and l arger than al l
the natural whol e numbers n of which is the l i mi t! Thi s shows to
what extent the domai n of Numbers saturates that of the ordi nal s,
whi ch i t contai ns: there are ' many more' Numbers than there are
ordi nal s.
We can al so say that N i s ' i nfnitely near' to far nearer than even
the most immense of the whole numbers could be. This notion of
' infnite proxi mity' is of a prodi gi ous phi l osophi cal i nterest. It opens
up new spaces for expl orati on in the endl ess ki ngdom of Number. We
shal l undertake these explorations a l ittle later.
16. 23. Sequence and end of the dyadic rationals
We have at our disposal a function Ra( N) whi ch makes a dyadic
rational correspond to every Number of fni te matter. The whol e
numbers are incl uded i n thi s correspondence, because the positive
whole number n thought of as Number will correspond, through the
function Ra, to the sum 1 + 1 + . . . + 1 n times - that is exactl y the
Number n, since, i f a Number i s a natural whol e number, then all of
its sub-Numbers are in its form. It would be better to say that the
function Ra associ ates a dyadic rati onal with every Number of fnite
matter - even i f this Number i s whol e.
To complete the work, and to concl ude that the dyadi c rati onal s
'themsel ves' are represented i n Numbers, we must:

confrm that the order of Numbers of fnite matter is i somorphi c
with the customary order of correspondi ng dyadi c rati onal s, so
that, i f N
l
< N2 in the order of Numbers, then Ra( Nd < Ra( N2 )
in the normal order of rational s; thi s amusi ng mathemati cal exer
cise i s sketched nicely i n the note;
4

prove that all the dyadic rati onal s are obtai ned through the func
tion Ra appl ied to Numbers of fnite matter; thi s comes down to
proving that every positive dyadic rati onal can be put i n the form
of the algebraic sum of a certai n whole number ( its whole part
' from above' ) and dyadic rational s of the form + or -+ ; because,
once thi s i s done, one can reassembl e the Number N, whose value
for Ra is the rati onal thus dismembered; 5

prove that the operational di mensi ons of the rati onal s - addi
ti on, mul ti pl ication, di vi si on, in brief, everythi ng that gi ves them
the al gebrai c structure of a fel d, are i somorphic to the same
operati ons defned for Numbers and appl ied to Numbers of
fnite matter; thi s relates to the exami nati ons made in chapter
1 8, with one obvi ous excepti on: i n order to have negative dyadic
rati onal s, the procedure of symmetricisation woul d be used,
which defnes the general manner of passage to the negative:
i nversi on of swappi ng form and resi due. Of course, we will sti l l
be deal i ng with a Number of fnite matter ( but thi s time with
o in the resi due) .
As far as the ontol ogical si de of thi ngs i s concerned, we have
attai ned our goal . A dyadi c rational , thought i n its bei ng, inscri bed
as Number, has a very si mpl e i ntri nsic defni ti on: its matter is fni te.
As far as bei ng i s concerned, that, thi s cl ari fes however dense the
rati onal s might be, even to the poi nt of an i nfnite swarmi ng between
two consecutive whol e numbers, they nevertheless belong to the
fnite. The numeri cal ontology of the i nfnite begi ns with real
Numbers.
1 6. 24. Real numbers
We know that real numbers provide the model for the geometrical
' conti nuum' : thei r fgure i s that of the poi nts of a l i ne. It i s the
real numbers that have subtended the enti re edifce of anal ysi s,
chef-d'oeuvre and keystone of modern mathemati cal thought,
si nce Newton and Lei bni z.
For a long ti me, the continuum and the functions corresponding
to it were thought either i n terms of geometrical constructions ( Greek
and pre-cl assi cal age) , or in a pri miti ve and pragmatic fashi on ( eigh
teenth and ni neteenth centuries ) . The emergence of a rigorous concept
of reals as entities with whi ch one can cal cul ate took pl ace slowly
during the course of the ni neteenth century, begi nni ng with Cauchy,
and wi th Dedeki nd representing a deci si ve step.
Because it i s the cl osest to that whi ch governs the defnition of
real s in the fel d of Numbers, we will recal l briefy the construction
of real numbers by means of 'cuts' , as i nvented by Dedeki nd.
16. 25. We wi l l begi n wi th dyadi c rati onal s, whi ch we can use here
i n pl ace of rati onal s as such, in view of the remark made in 16. 14.
Take two sets of dyadic rati onal s B and A such that every rational i n
B is smal l er t han every rati onal i n A. We can say both that B has no
internal maxi mum ( for every dyadi c rati onal r1 i n t he set there i s an
r2 i n the set such that rl < r2 ) ; and that A has no i nternal mi ni mum.
Suppose now that the fol l owing rel ati on holds between B and A: there
al ways exists a dyadic in B that is ' as cl ose' as one l i kes to a dyadic
of A. In other words, if r2 is a dyadi c in A and r a dyadi c as small as
one likes, there wi l l always exist a rl in B such that the di fference
between r2 and rl i s less than r.
The situation can be vi sual ised as below, by representing the dyadic
rationals as points on a l i ne:
t e

< t
A
t e A
We can see cl earl y that B ' ri ses' wi thout ever entering i nto A, that
A 'descends' without entering i nto B, and that therefore the two sets
are as close to each other as can be, wi thout ever ' touchi ng' .
Thus did Dedeki nd defne a real number as the poi nt si tuated
exactly ' between' B and A; that i s, the element, created i n this process,
which is si multaneousl y l arger than any element of B and smal ler
than any element of A. We can identify thi s element as the poi nt of
the cut of B and A.
It is characteristic of thi s method that it treats the cut not as a state
of things i n a pre-given uni verse ( which is how we treated i t for
Numbers, see 1 5. 6) , but as a procedure, defni ng a mathemati cal
entity that does not pre-exi st thi s procedure. To begin wi th, there are
only rati onal s. And, if the cut is not a rati onal ( it coul d be, if the
upper limit of B and the l ower l i mit of A coincided) , then it consti
tutes i n itsel f the name, or form of presentati on, of a ' bei ng' whi ch
inexi sts in the fel d of rational s. Therefore the real s are operati onal
productions here; they sign, comi ng forth from non-being, the fctive
point where B and A are touched by the interposi ti on between them
of thi s fcti on. Into that place, where there was nothing but the min
iscule void that separates two sets as close as can be, comes the real ,
which stops up thi s voi d by realising a cut as number.
1 6. 26. Fictions have no place i n the ontological conception of
Number. I f the cl assic real numbers, those whi ch real ise cuts i n the
dyadic rati onal s, i nscri be themselves i n the domai n of Numbers, it is
because they exist and are di sti ngui shed by some property. They
cannot irrupt from inexi stence, i n the form of mere names of a
lacuna. According to an ontological conception of Numbers, every
Number is, none results or is resolved i n the name of an operati on.
We do battle here against a domi nant nomi nal i sm, and we do so in
the feld of number, so commonl y taken for an operati onal fction.
1 6. 27. In fact, our defnition of real numbers as Numbers is quite
l i mpi d:
A Number is a real number if it is either of fnite matter or of
matter 0, and if its form and its residue are infnite.
In what fol l ows we wi l l substanti ate thi s defnition, whi ch repre
sents real numbers ' themsel ves' in the domai n of Numbers.
16. 28. The ' proj ecti on' of thi s defni ti on into the concept of real s as
cuts i s basi cal l y very si mpl e.
If a Number is of fnite matter, then it i s, as we have seen, a dyadic
rati onal .
If a Number i s of matter 0, then al l of i ts sub-Numbers are of
lesser matter than 0, and therefore of fnite matter, si nce 0 is the
smal lest i nfnite ordinal . So all of its sub-Numbers are dyadic ratio
nal s. More speci fcal l y, its l ow set and its high set are sets of dyadic
rati onal s. And, since every Number i s the cut of its low set and its
high set, a Number of matter 0 can be represented as the cut of two
sets of dyadic rati onal s. Or, once agai n, a real number thought as
Number i s a Number whose canonical presentation Lo( N)/i ( N) is
made sol el y from dyadic rational s.
Fi nal l y, i f a Number of matter has an i nfnite form and an
i nfnite resi due, we avoi d i ts l ow set and high set havi ng internal
maxi ma. Because, i f the form of N i s fni te, since it i s composed
of whol e numbers ( the matter being 0) , it admits of a l argest element,
say the whol e number p. The cut of N at point p defnes the sub
Number NIp, which i s obvi ousl y the l argest sub-Number whose
di scri mi nant with N is in the form of N, and therefore the l argest
sub-Number i n the low set of N. And, if the resi due is fnite, there
exi sts a number p such that NIp is the smal lest element of the hi gh
set of N. A cOlltrario, i f both the form and resi due of N are i nfnite
- are sequences of whol e numbers wi thout i nternal maxi ma - then
the low set does not have a maxi mum term, nor the high set a
mi ni mum term.
We thus fnd ourselves precisely i n the conditions of the Dedekind
cut: di sj oi nt ascending and descending sets of dyadic rati onal s with
no maxi mum or mi ni mum. Except that what we characterise as
'reals ' are particular, already existing Numbers, whereas Dedekind
i nstal l s them as a fcti on at the voi d point of a cut. For us, a real
wi l l be that uni que Number of minimum matter situated exactly
between two sets of dyadi c rati onal s whi ch can be shown to be its
low set and its high set, and therefore to be sets of sub-Numbers.
It is parti cul arl y reassuri ng to remark that, in the defni ti on of
real s as Numbers, everythi ng remai ns i mmanent. Dedeki nd cuts des
ignate the fction of a number external to two sequences of rati onal s,
as the poi nt of contact of these sequences . Whereas, on the other
hand, the sets of dyadic rational s that we use are composed of sub
Numbers of a real Number. Thi s i mmanentisation of procedures i s
typical of the ontological approach, that approach which captures
the being of Number. To see if a Number that i s not a dyadic rati onal
is a real number, it suffces to exami ne it accordi ng to its three
components:
-
its matter must be ;
- its form must be i nfnite;
-
its residue must be i nfnite.
Thi s alone al lows us to conclude. Then we can state that the
Number is the cut of two sets of dyadic rati onal s, and that therefore
it is indeed a real number (in the cl assi c sense) . But, all the same, we
have remai ned withi n Number, since dyadic rati onal s are sub- Numbers
of a Number.
The i mmanence of the thi nki ng of being has not faltered for a
moment in thi s approach to the traditional real numbers grasped i n
the space of Numbers. The characteri sati on of a type of pure mul ti pl e
has been substituted for operati onal fcti ons. And real numbers are
no more mysteri ous here than whol e numbers or rati onal s. Thei r sol e
pecul i arity is that they mark the moment where our passage through
Numbers prompts us to envi sage i nfnite matters. From thi s poi nt of
vi ew, the ontological si ngul arity of the real s i n rel ati on to the whol es
and the rati onal s can be summed up in one word: i nfnity. This alone
clarifes, i rrespective of all complexities of constructi on, with an eye
only to that i n which the numerical section operates, the fact that
real Numbers are exempl ari l y modern.
1 6. 29. We now fnd ourselves in possession of a concept of Number
that subsumes as particul ar species the natural whol e numbers, the
whole positives and negatives, the rational s, the real s, and the ordi
nal s. 6 We have overcome the modern resistance to a uni fcation of
the concept of Number ( see 1 . 8) . But, i n the process, we have al ready
seen that this concept al so subsumes other Numbers, that the hi stori
cal deduction from the domai n of Numeri cal ity i s very much l i mited.
Rational s and reals cover the totality of Numbers of fni te matter and
onl y some Numbers of matter . It i s as i f our thi nki ng has so far
only brought to l ight a mi nute i ni ti al segment of that whi ch being
proffers in terms of possi bl e numeri cal access to pure mul ti pl icities.
The future of the thi nki ng of Number i s l i mi tl ess.
16. 30. Infnitesi mal s
We remarked i n 16. 22 that it woul d be possi bl e to fnd a Number N
at once smal ler than 0 and l arger than al l the fnite ordi nal s whose
l i mit is 0. This Number coul d perhaps be said to be ' i nfnitely close'
to 0, and it puts us on the way to a concept of i nfni tesi mal Number.
The idea of i nfnitely small number, freely empl oyed by seventeenth
and eighteenth-century mathemati ci ans, was di smi ssed i n the ni ne
teenth century for its obvi ous i nconsi stenci es. It was repl aced by the
concepts of the l i mi t ( Cauchy) and of the cut ( Dedeki nd) . It reap
peared around thi rty-fve years ago, i n the si ngul arl y artifci al , but
consistent, context of the pure l ogic of model s: Robi nson' s non
standard anal ysi s. 7 In the domai n of Numbers, ' i nfnitely smal l
numbers' or i nfni tesi mal s abound in the most natural fashi on. It is
by means of them that we will complete thi s di mi nutive j ourney
through the enchanted kingdom of Numbers.
1 6. 3 1 . Consi der the Number i = ( 0, ( 0) ) , whose matter i s 0 and
whose form, the si ngl eton of the voi d, has the void as its onl y element.
It i s a positive Number, since 0 i s in its form.
Now thi s positive Number, even i f its matter is the same as that
of real Numbers, is smaller than every positive real Number.
In fact, if a real Number i s positive, 0 i s in its form, as i s the case
for i: 0 does not di scri mi nate between i and a positive real Number.
Al l the whol e Numbers other than 0 being in the resi due of i, the
di scri mi nant of i and a real Number R wi I I be the frst whol e Number
apart from 0 to fgure in the form of R. Such a Number necessarily
exi sts, si nce the defnition of the real s dictates that the form of R
shoul d be i nfnite. And, si nce thi s di scri mi nant is in the residue of i,
i is smal l er than R. Therefore there exi sts a Number i such that
0 < i < R for every real Number. Thi s i i s si tuated ' between' Zero and
all real numbers thought as Numbers. We will say that it is i nfnitesi
mal for the real s.
16. 32. General i sing thi s defni ti on: We say that a set of positive
Numbers, al l of the same matter, tends rati onal l y towards Zero i f,
for every dyadic positive rati onal r, as cl ose to Zero as you l i ke, there
exi sts a Number N, of the set situated between Zero and r. In other
words: for every dyadic rati onal r, there exi sts N, belonging to the
set such that 0 < N, < r. Note that the classic notion of 'tending
towards' i s here rel ati vi sed to dyadic rational s. In the l i mitless domain
of Numbers, we must indicate which scale of measurement is being
employed, because, as we will see, it i s always possible to fnd a sti l l
fner scale.
It is obvious that the set of real positive Numbers tends rational l y
towards Zero. Other sets of Numbers can be found whi ch tend ratio
nal l y towards Zero, for exampl e positive Numbers of the type
( S( c) , ( O, . . . ) ) , whose matter i s S( c) and whose form contai ns at
least O.
We can say, then, that:
z Number is infnitesimal for a set of Numbers that tends rationally
towards Zero if it is:
-
of the same matter as the Numbers of the set;
- positive;
-
smaller than all the Numbers in the set.
So it is that the Number ( c, ( O) ) is i nfnitesi mal for the set of real
Numbers. On the other hand, there i s no i nfni tesi mal for the set
of Numbers ( S( c) , ( O, . . . ) ) , preci sel y because this set contains the
very Number ( S( c) , ( O) ) that is the smal lest positive Number of
matter S( c) .
The l i mi ti ng of the concept of i nfnitesi mal to Numbers of the same
matter as the Numbers of the set that tends towards Zero is necessary
because, i f thi s restriction were not i n place, there would be as many
in{nitesimals as we wished. It woul d suffce to augment the matter:
the Number ( S( c) , ( O) ) i s posi ti ve, and it i s certai nl y smal l er than every
positive Number whose matter is c. In particul ar, i t i s smal ler than
the i nfnitesi mal i ( c, ( O) ) , because the di scri minant is c, which is
outside the matter of i and i n the residue of ( S( c) , ( O) ) . We see to what
extent our concept of the i nfnitesi mal i s rel ati ve: the density of order
over Numbers means that, however ' rel ati vel y' small a posi ti ve
Number might be, there sti l l exists an i nconsi stent mul ti pl ici ty of
Numbers situated between it and Zero.
We can, i f we wish, retai n the cl assi c defni ti on: every posi ti ve
Number smal l er than every positive real is i nfnitesi mal . But then we
wi l l see the i nfnitesi mal s grow and swarm uncontrol l abl y. The
' shores' of Zero contai n ' as many' Numbers as the enti re domai n of
Numbers. Because, at the poi nt where mul ti pl e-being as such incon
sists, the notion of ' as many' loses all meani ng.
16. 33. Cuts of cuts
Take the Number C ( c, ( 0, 1 ) ) , whose matter is c and whose form
is l i mited to the wholes 0 and 1 . Thi s Number i s not real , si nce its
form is fnite. It is positive, si nce 0 is in its form. How can it be
situated amongst the real s, to whi ch its matter belongs ?
A posi ti ve real whi ch does not have 1 in its form is certai nl y
smal ler than C: the di scri mi nant i s 1, whi ch i s i n the resi due of such
a real and in the form of C.
A posi ti ve real which does have 1 in its form i s certai nl y l arger
than C. For al l whol e numbers l arger than 1 are in the residue of C,
whereas some of them are certai nl y i n the form of a real , since this
form is i nfnite. The discri mi nant will be the smal lest whol e l arger
than 1 to be in the form of the real , and, si nce it is in C's resi due, C
wi l l be smal ler.
C therefore i s si tuated precisely between the reals whi ch have 1 in
thei r residue and the real s which have 1 in thei r form. Now these two
cl asses operate a partition into two of the posi ti ve real s, a partition
which i s ordered ( al l the positive real s which have 1 in thei r residue
are smal ler than all the positive real s which have 1 in their form) . We
can, then, perfectly l odge a Number ' between' two di sj oint cl asses of
real s, i n the caesura of a partition of real s. And, si nce the reals are
themselves cuts of rational s, the Number C wi l l be a cut of cuts.
General l y speaki ng, gi ven an organi sed partition into two of a set
of Numbers 'of the same type' , that i s to say, defned by cuts or
canoni cal presentati ons havi ng thi s or that property ( as we saw in
defni ng the real s) , we wi l l cal l a 'cut of cuts' a Number of mi ni mal
matter si tuated in the caesura of the parti ti on, being l arger than al l
those i n the l ower segment and smal l er than al l those in the higher
segment. The Number ( 0, ( 0, 1 ) ) i s a cut of cuts in the numerical type
' positive real Numbers' .
The exi stence of cuts of cuts attests once more to the i nfnite capac
ity of Numbers - as coalescent as they might seem - for cutting at a
point the ul tra-dense fabric of thei r consecuti on.
16. 34. So many other Numbers to vi si t and to descri be! But works
that take delight i n this are begi nni ng to appear. And the phi losopher
i s not defned by curi osity; the j ourney i s not a di sinterested one. The
phi l osopher must, before l eavi ng the ki ngdom convinced that every
number thought of in its being i s a Number, descend back down to
cal cul ati on. Or, rather, to the exi stence of cal cul ati on, because the
phi l osopher i s not a cal culator either. But these numbers, from which
our soul is knitted, the phi l osopher wishes to render over entirely,
even as regards the derivation of thei r operati onal mechani sm, to the
i mmemori al and effectless transparency of Being.
4
Operational Dimensions
I I
1 7
Natural I nterlude
1 7. 1 . The domai n of ordinal s ( and of cardi nal s) hol ds an extreme
charm for thought. A proof by affect - by affection, even - of what
I cl ai m here, is that, on refecti on, thi s charm i s that of Nature itsel f:
an abundant di versity and, at the same ti me, a mute monotony.
Nothing is the same, everythi ng goes to i nfnity, but one hears a fun
damental note, a basso ostinato, signal l ing that these myri ads of
multipl icities and forms, these compl icated mel odi es, prol i ferate
the repose of the identical . If poets' metaphors take as thei r
reference the sky and the tree, the fower and the sea, the pond and
the bird, thi s i s because they woul d speak thi s presence of the Same
that the unl i mited appearances of nature vei l and reveal . In the
same way, the ordi nal s, sti l l si ngul ar i n the i nfnity of thei r i nfnite
number, in the i nconsi stency of their Al l , al so repeat the transitive
stabi l ity and the i nternal homogeneity of natural mul ti pl es, those
multiples that they allow to be thought i n thei r pure bei ng. It i s hard
to tear onesel f away from the intellectual beatitude brought on by
the contempl ation of the ordi nal s, one by one and as a ' set' . I thi nk
of the great Indi an mathematici an Ramanuj an,
I
who hel d each whol e
number to be a personal friend. He was i nvested by thi s poem of
Number, of whi ch the Poem of nature i s the symmetrical counterpart
withi n language. He did not like to construct proofs, but rather, as
a dreamer of the ordi nal site, to draw i n i t with curves of recogni ti on,
which hi s col leagues regarded with some surpri se. Comi ng from afar,
in al l senses of the word, he was not accustomed to our severe modern
distinctions. He saw numbers di rectly for what they are: natural
1 84 OPERATIONAL DIMENSIONS
treasures, where being l avi shes its multiple resource and its fasti di ous
i denti ty i n the same gesture i n whi ch, for the poet, it arranges the
'correspondences' of sensi bi l ity.
1 7. 2. We have at our disposal a concept of Number, and we know
that this concept subsumes our tradi ti onal numbers. Wholes, ratio
nal s, real s, ordi nal s, thought i n thei r multiple-being, are Numbers.
I t must now be shown - a sl ightly less rewardi ng task - that thi s
concept subsumes our traditional numbers not onl y in thei r being,
but also i n their operati ons. As far as we may be from that sensi bi l ity
that i s rul ed by counti ng, it must nevertheless be shown that it is
possi bl e to count with Numbers, and that thi s counting coincides,
for the cl assi cal types of Numbers, with ordinary counting. We must
cover al gebra, addi ti on, mul ti pl icati on, etc. I f we di d not, then who
woul d believe us when, speaki ng from the sol e poi nt of view of being,
we said that these Numbers are numbers ?
1 7. 3. What is meant by ' operati on' , or cal cul ati on, is the consider
ati on of ' obj ects' upon whi ch one no l onger operates one by one, but
at least two by two: the sum of x and y, the divi si on of x by y, etc.
And, as the matter of Number is made of ordi nal s, it is to be expected
that we have to deal with, to thi nk, pai rs of ordi nal s. So we wi l l be
happi l y detained for a few more moments i n the enchanted domai n
of natural mul ti pl es. Thi s whol e i nterl ude is dedicated to some refec
ti ons and proposi ti ons about pai rs of ordi nal s, ordi nal s taken two by
two. And, as we shal l see, these couples are al so total l y natural : we
can connect them back to ' si ngl e' ordi nal s vi a a procedure which in
itsel f hol ds a great charm.
1 7. 4. We wi l l speak of ordered pairs of ordi nal s, whi ch we denote
by (W" W2). ' Ordered' meani ng that one takes into consi deration the
order of the terms in the coupl e - we wi l l thus speak of the frst term,
W" and the second, W2 - which wasn' t the case i n our concept of
the simple pai r, denoted by ( e"e2 ) ( compare 7. 7), which was a pure
' gatheri ng together' of two terms regardless of thei r order. Or, in
other words: if W, and W2 are di fferent, then the ordered pai r
(W"W2) i s not the same thi ng as the ordered pai r (Wz ,W, ). In order
better to di sti ngui sh the si mple pai r from the ordered pai r, we wi l l
cal l the l atter a couple.
We can al so al l ow ' coupl es' of the type (W" W, ). In such cases, W,
occupi es both the frst and the second place.
The concept of ordered pai r, or coupl e, pl ays a deci si ve role in
mathemati cs: it underl i es all thi nking of rel ati ons and of functi ons.
2
NATURAL INTERLUDE 1 85
It can be reduced to a fgure of the pure multi pl e, testi fyi ng to the
fact that relations and functi ons do not depend on any sort of addi
tional being apart from the multi pl e, that there is no ontological di s
tinction between bound obj ects and the bond which bi nds them. But
we wi l l empl oy the concept here i n i ts naiVe sense.
17. 5. We wi l l cal l maximal ordinal of a couple (WhW2), and denote
by Max ( WhW2 ) , either the l arger of the two ordi nal s WI and W2,
if they are di fferent, or, if the coupl e is of the type (W h WI ) the si ngl e
ordi nal WI that fgures i n it. You are remi nded ( see 8. 10) that ordi nal s
are total l y ordered by bel ongi ng: i f WI and W 2 are di fferent, then one
i s necessari l y smaller than the other ( bel ongs to the other) .
Thi s most elementary notion of the maxi mal term of a coupl e wi l l
play a cruci al role i n what fol lows. It i s i mportant to get a frm grasp
of it.
1 7. 6. Take a couple of ordi nal s (WhW2), whi ch we will denote by
Ch and another couple (W
3
,W
4
), which we will denote by C2 We
will defne an order-rel ati on between these couples i n the fol l owi ng
way. We say that CI i s smal ler than C2 and write CI < C2, i f one of
the three fol l owing conditions i s sati sfed:
1 The maxi mal ordi nal of the coupl e C2 is equal to the maxi mal
ordi nal of the coupl e C
1
In other words: i f Max( Cd E Max( C2) ,
it is always the case that C
1
< C2
2 The maxi mal ordi nal of coupl e C
1
is equal to the maxi mal ordi nal
of the coupl e Cz , but the frst term of the coupl e CI i s smal ler
than the frst term of the coupl e C2 In other words, i n a case
where Max( Cd Max( C2 ) , i f WI E W3 , then C
1
< C2
3 The maxi mal ordi nal of the coupl e CI is equal to the maxi mal
ordi nal of the coupl e C2, and the frst term of the coupl e CI is
equal to the frst term of the coupl e C2, but the second term
of C
1
is smal l er than the second term of C2 In other words,
Max( Cd Max( C2 ) and WI W
3
, but W2 E W
4
In thi s case,
CI 5 C2
Evidently, if none of these three condi ti ons are sati sfed, then the
couples C
1
and C2 must be identical : they have the same frst term
and the same second term. A contrario, i f two coupl es of ordi nal s
are di fferent, either one i s smal l er than the other, or the other i s
smal l er than it: the rel ati on is total.
Thi s order fol l ows di rectly from empl oyment of the operator
Max( C) , or, i f this yields onl y an identity, from the comparati ve
exami nati on, in thi s order, frstly of the frst ordi nal of each couple,
then, i f thi s examination too yi el ds onl y an i dentity, of the second
ordi nal of each coupl e. The maxi mum trumps the frst term, and the
frst term the second. 50me exampl es:

(6, 0) is smal ler than (0, 7), because its maxi mum is 6, which is
smal ler than the maximum of the l atter, which is 7;

(0, 0) is smal l er than ( 1 , 0), because, thei r maxi ma being i dentical
( it i s () , the frst term of the former, 0, is smal ler than the frst
term of the latter, 1 ;

(5( 0) , 0) is smal l er than (5( 0) , 5( 0) , because, thei r maxi ma being
equal ( they are both 5( 0) , the successor of () and their frst term
l i kewi se ( it i s 5( 0) in both cases) , the second term of the frst
coupl e, whi ch is 0, i s smal l er than the second term of the other,
which is 5( 0) .
Note that t he coupl e (0,0) i s t he smal lest coupl e of al l , si nce its
maxi mum, its frst term and its second term are all equal to 0, which
i s itself the smal lest ordi nal .
It is al so cl ear that coupl es form an inconsistent multi ple, since,
al ready, the ordi nal s themselves cannot form a set. In speaki ng of
' the' couples, but also of ' the' ordinals, or ' the' Numbers, we must
al ways remember that we cannot attri bute any property to whatever
thi s 'the' designates : there is no question of a thi nkable, or present
abl e, total i ty. In particul ar, i f there exi sts a mi ni mal coupl e for the
order that we are going to defne ( it is the pair (0, 0) , there certainly
i s not a maxi mum couple for thi s order.
1 7. 7. I leave the reader the task of showi ng that the relation between
couples that we have j ust defned i s a genui ne order-rel ation ( and
therefore, essenti al l y, transi ti ve: i f C] < C2 and C2 < C
)
, then
C] < C
)
) .
Far more i nteresting i s the fact that i t i s a well-ordered relation. I
have gi ven the defni ti on of thi s in 6. 4: given any set whatsoever of
terms wel l -ordered by a rel ati on <, there exi sts one ( and one onl y)
el ement of that set that i s minimal for the order-rel ati on, whi ch i s the
smal lest element of that set.
Take any ( non-empty) set E of couples of ordi nal s - that i s, a set
all of whose elements are couples of ordi nal s. Consi der all those
coupl e el ements of E whose maximum is minimal for E. In other
words al l the coupl es C E E such that Max( C) is the smal lest ordinal
to fgure in the el ements of E as maximum of a couple. Thi s i s possible
by vi rtue of the princi pl e of mi ni mal ity that characterises the ordi nal s
( see 8. 10) . Given the property ' being a maxi mal ordi nal in a coupl e
C which belongs to E' , there exi sts a smal lest ordi nal to sati sfy thi s
property. We thus obtai n a subset E' of E, al l of whose elements C
have the same mi ni mal maxi mum. Note that, because of the frst
of the conditions defning the order of couples, all the elements of
E' are smaller than all the elements that remai n, that are i n E-E'
( i f any) .
Now consider, i n E', the set of couples whose frst term i s mi ni mal
for E' . In other words, al l the couples C (Wt .Wz) such that WI is
the smal lest to be found i n all the frst terms of the couples i n E'.
This i s possi bl e for the same reason as before: it suffces to consi der
the property ' bei ng an ordi nal that fgures i n a coupl e i n E' as the
frst term' , and to take the mi ni mal ordi nal for thi s property. We wi l l
thus obtain a set E" of coupl es havi ng the same maxi mum ( because
they are in E' ) and the same frst term ( mini mal for E' ) . Note, in
considering the second of the conditions defni ng the order of coupl es,
that all the elements of E" are smal l er than all the el ements that
remai n, whi ch are in E' _E", whi ch themselves are all smaller than
the elements of E-E'.
There i s a sort of concentric embedding, where the couples of each
i nner ci rcle are smal ler than those of the exterior boundary.
Consi der, fnal l y, in E", the property ' being an ordi nal that fgures
in second position in one of the couples i n E"' . There i s a mi ni mal
ordi nal for thi s property. But, thi s ti me, the set obtai ned consists of
one couple only. Thi s i s because, i n E", the frst term of the couples
is fxed ( it is the minimal frst term for the coupl es i n E' ) . In fxing
the second term (as mi ni mal for thi s place) , one couple i s entirely
determined. But the others are themselves smal ler than all the couples
in E'-E", which are smal l er than the couples i n E-E'. So the mi ni mal
couple obtained i n E" is in fact mi ni mal i n E-QED.
Thi s property of mi ni mal ity for the order of coupl es of ordi nal s
grants us three essenti al freedoms:
1 Given a coupl e C, it is possi bl e to designate the uni que couple
that will come di rectl y after it i n the order we have defned. To
do so, it suffces to consi der, i n a suitabl e set that contai ns C, the
subset of those that are l arger than i t. This subset wi l l have a
mi ni mal element, whi ch is the smal lest one to be l arger than C,
and is thus the ' successor' of C.
2 If a property of couples defnes a set ( the set of couples which
possess that property) , then we can safely speak of the smal lest
coupl e in that set, and therefore of the smal lest coupl e to possess
that property.
3 Gi ven a set of couples, we can speak of the upper bound of that
set, as we can for sets of ordi nal s ( see 12. 16) : it suffces to consi der
the smal lest coupl e that i s l arger than all the couples in the set.
' Wel l -orderedness' al l ows thought to move between interior mi ni mal
i ty and exterior maxi mal ity: the smal lest of a given set, and the frst
( outsi de) to be l arger than al l those i n that set. The trap i s to imagine
that one thereby gains access to internal maxi mal ity: this i s not at all
the case because, for couples as for ordi nal s, that which goes to the
l i mit i s not internal l y maxi mi sabl e.
1 7. 8. We are speaki ng of successi on and l i mit. Here we return, let
us remark, to the di sputations of chapter 9. Di scovering the ki nshi p
between ordi nal s and coupl es of ordi nal s was our latent motive.
1 7. 9. Let' s begi n with an example. What can we say of a couple of
the form (W"S(W2) , where WI i s any ordi nal whatsoever apart from
0, and where S( W2 ) is the successor of any ordi nal W2 ? Everythi ng
depends on the maxi mal ordi nal in the coupl e. Suppose that WI is
maxi mal and thus that S( W2 ) E WI . If we compare the couple to al l
others that al so have WI as thei r maxi mal ordi nal , we see that it i s:
- l arger than al l those where W I onl y comes in second position in
the coupl e ( pri macy of frst posi ti on, condi ti on 2 of the ordering
of couples ) ;
- l arger than al l those whi ch, in second posi ti on, have an ordi nal
smal ler than S( W2 ) ( thi rd condi ti on of order) ; i n particul ar, it i s
l arger than the coupl e (W"W2);
- smal l er than al l those which have, in second position, an ordinal
l arger than S( W2 ) . In particul ar, it is smal l er than the couple
(W"S( S( W2) ) ), supposi ng that S( S( W2 ) ) remai ns smal ler than WI .
thus leaving WI ' S maxi mal status intact. But l et' s assume thi s
hypothesi s.
It seems cl ear from thi s that the coupl e (WhS( W2 ) , given the
assumed hypothesi s, intercalates itself exactly between the couple
(WhW2) and the coupl e (WI , S( S( W2 ) ) ). More specifcal l y, we can say
that it succeeds the frst of the two coupl es.
If, on the other hand, we take the coupl e (W) , L) where L is a l i mi t,
and sti ll suppose that WI i s maximal i n the coupl e, we cannot deter
mine a couple that it succeeds. This coupl e is certai nl y l arger that all
the couples of the form (WhW2) where W2 i s smal l er than L ( thi rd
condition of order) . But L, a l i mi t ordi nal , preci sel y does not succeed
any of the W2 i n questi on. There i s therefore only one possi bi l ity: the
couple (WhL) i s the upper bound ( see N6) of the set of couples
(WhW2), where W2 E L, with WI , of course, being maximal - that
is to say, l arger than L. We can al so say that the coupl e (WhL) i s the
limit of the couples (WhW2) for W2 less than L.
Fi nal l y, take the coupl e (0, S( W2 ) . The Max. of thi s coupl e is
evi dently S( W 2). But it is certai nl y the smallest couple to have thi s
Max. In fact, i ts frst term i s mi ni mal ( i t is 0) , so every coupl e C
where Max( C) = S( W2 ) and where the frst term is not - therefore
every coupl e of thi s sort other than our exampl e - is greater
than it.
Being the smal lest coupl e whose Max. is S( W2 ) , our coupl e must
succeed the ' l argest coupl e' - i f it exi sts - whose Max. i s i mmedi
ately i nferior. Note that these noti ons of ' l arger' and ' i mmediately
i nferior' can be disrupted by the i ntervention of limit ordinal s. Al l
the same, thi s is not the case in our exampl e: si nce the Max. of
our couple is S( Wz ) , an i mmediately i nferi or Max. exi sts: it i s W2.
What woul d be the l argest coupl e whose Max. is W1 ? Obvi ousl y
that coupl e whose frst term i s maxi mal ( condi ti on 2 of order) . But
the frst term of a couple whose Max. is W2 attai ns i ts maxi mum
when it is equal to W2 For, i f it surpasses W2, the Max. changes.
So the coupl e that i mmediately precedes (0, S( W2 ) in the order of
couples is (W2,W2). We can al so say that (0, S( W2) i s the successor
coupl e of (W2,W2).
1 7. 1 0. What we real l y want i s to 'ontologise' couples of ordi nal s, as
we di d for ordi nal s: fnd an i ntri nsic characteri sati on, not bound to
order al one, of successor couples and l i mi t coupl es. The exampl es i n
the previous paragraph wi l l gui de us.
1 7. 1 1 . Let's begin wi th couples contai ni ng 0.
We have remarked that couples of the form (O, WI ), for al l W
I
other
than 0, are the smal lest ones whose Max. is WI
'
Thi s al lows us to
characterise them i mmanentl y:
1 A coupl e of the form (O, S( Wd) is always a successor ( it succeeds
(WhWI ) . Thus the couple (0, 1 ) is a successor ( it succeeds the
mi ni mal coupl e (0, 0) .
2 A coupl e of the form (O, L) is al ways a l i mit: it is the upper bound
of the sequence of couples (WI ,Wz) where WI and Wz pass i nto
the l i mi t ordi nal L. So that the coupl e (0,0) i s the l i mi t of al l the
couples (m, n) where m and n are fni te ordi nal s ( and therefore
natural whol e numbers, see chapter 1 1 ) .
Couples of the form (WhO) depend j ust as di rectl y, as regards their
i ntri nsi c characteri sati on, on the nature of the ordinal WI :
1 A coupl e of the form (S( W I ) ,0) is the smal lest coupl e to have
S( Wd as Max. in frst position. It is l arger than al l those which
have S( Wd as Max. in second position that i s, couples of the
form (W2 , S( Wd) where Wz i s smal ler than S( Wd. In fact it
comes j ust after the l argest of these coupl es, which latter wi l l
evi dently have the l argest possi ble frst term to conserve S( Wd' s
status of maxi mum i n second posi ti on. Thi s l argest frst term i s
WI . the i mmedi ate predecessor of S( Wd. The l argest of the
couples which come before (S(W I ) , 0) i s therefore the couple
(W"S( Wd). We can concl ude: every coupl e of the form (S( WI ) , O)
i s a successor. Thus the coupl e ( 1 ,0) i s a successor ( it succeeds
(0, 1 ) .
2 A coupl e of the form (L,O) is l arger than every coupl e of the form
(W"L) where WI i s less than L. But there is no such couple that
i s l arger than all the rest, because there i s no W I that i s ' closer'
to the l i mit ordinal L than all others ( see 9. 1 8) . The coupl e (L,O)
i s, moreover, smal l er than all the couples of the type (L,WI ) where
WI is not 0. In a sense, it makes a cut between the couples (W" L)
and the coupl es (L, O). Al l the same, amongst the l atter there is a
mi ni mal coupl e, which is the coupl e (L, l ), and whi ch therefore
succeeds (L,O). Here agai n we fnd the stri ki ng dissymmetry, char
acteristic of the ordinal s, between mi ni mal ity ( guaranteed) and
maxi mal ity ( which presupposes successi on) . The couple (L,O) is
the l i mit, or upper bound, of the sequence (W" L) for WI E L,
and it i mmediately precedes the couple (L, l ). It creates an infnite
adherence to its l eft, or ' on thi s si de' of it, and the void of one
si ngle additi onal step to its right, beyond it.
1 7. 12. Let' s now turn to ' homogenous' couples of the type (S hS2) or
(L"L2). Everythi ng wi l l once more depend upon the Max. of these
coupl es:

If 5 1 or LI are the Max. , the problem is tri vi al : (s "sz) is a successor.
Just a moment of refection wi l l show that it comes j ust after the
coupl e consti tuted by 51 (the Max. ) and the predecessor of 52 . As
for (LI , L2), it i s surel y the l i mit of the sequence of couples of the
type (L"WI ), where WI traverses the elements of the ordi nal Lz.

If Sz or L2 are the Max. , thi ngs are not much more di ffcul t. It is
certai n that (5 1 ,52) is a successor: it comes j ust after the couple
consti tuted by the predecessor of 51 and by the Max. 52 . As for
(L"L2), it is assuredly the l i mit of the sequence of coupl es (W"Lz),
where WI traverses the elements of L" from 0 ' up to' LI .
1 7. 1 3. We wi l l fni sh with mi xed coupl es. The method does not
change at al l :

If, in a couple of the type (s, L) or (L,s), it is L which is the Max. ,
these couples are successors: they come j ust after the couples
obtained by replacing 5 with its predecessor.

If 5 is the Max. , the coupl es are l i mi ts of the sequences of couples
of the type (s,WI ) or (W"s), where WI traverses the elements of
the l i mit ordi nal .
1 7. 14. Fi nal l y, we now have a tabl e of i mmanent characteri sati ons
of couples as fol l ows:
Type Max. Exampl e Character
(0,0) Uni que Speci al
(O,s
) S (0, 1 ) Successor
(O,
L) L (0,
0) Li mi t
(s, O) s
( 1 , 0) Successor
(L
,O) L (
0
,0) Li mi t
(
S
"
S2)
S l
(2, 1 ) Successor
(
SI S2)
S1
( 1 ,2) Successor
(L"L2) L
I (0, , 0) Li mi t
(L"L2) L
2 (0
,
0
1 ) Li mi t
(s,
L)
s
(S( o
) ,
o
) Li mi t
(s,
L) L (
l
,
o
) Successor
(L,s) s
(
o
, S(
o
) Li mi t
(L
,s) L (0
, 1 ) Successor
This table has a perfect symmetry, broken only by the i naugural
couple of the void with itsel f, the ontol ogi cal basi s of the whol e
constructi on.
1 7. 1 5. It is entertai ni ng to vi sual i se the begi nning of the sequence of
ordinal coupl es.
We have al ready seen that after the coupl e (0,0) comes the couple
(0, 1 ), then the couple ( 1 ,0). One can quickly see that it is ( 1 , 1 ) that
succeeds (1 ,0), si nce it i s the l argest coupl e whose Max. i s 1 . Comi ng
next is (0,2), whi ch i s, as we have remarked, the smal lest couple
whose Max. i s 2. The readers can exercise themselves by calculating
the rest. If we draw the successi on of coupl es onto a squared back
ground, using the horizontal axis to represent the ordi nal that occu
pi es the frst place and the verti cal to represent that which occupies
the second, we obtai n the fol l owi ng:
<, 1 > <1 , > <Z, > <3, > <4, >
What we see i n this diagram i s that the route through the couples
forms a ki nd of ' chai n' which evidently coul d be proj ected onto an
ordi nal axi s. At any gi ven moment we know how to ' produce' the
nth coupl e, as soon as its predecessor has been determined. It is
tempting to formal ise thi s i ntui ti on by establ i shing a term-by-term
correspondence between ordi nal s and coupl es of ordi nal s, since we
have seen that the ' passage to the l i mit' represents no obstacle to our
doing so: there is a concept of the l i mi t coupl e, structural l y di stinct
from the successor couple. Thi s magni fcent constructi on, which proj
ects couples ( representable on a pl ane or a surface) onto the l i nearity
of their constituents ( si ngle ordinal s ) , is a tri umph of ontol ogy. It
shows that there is no more in the doubl e than in the si ngl e. It l i n
earises the di vergence of twoness.
1 7. 1 6. Our underlying motive here is to show that couples of ordi
nal s behave ' l i ke' ordi nal s themsel ves. The si mplest way is to establ i sh
between couples of ordi nal s and ordi nal s a bi univocal correspon
dence ( see 4. 5 ) . However, it i s dubi ous, absurd even, to speak of a
correspondence or a functi on between two i nconsi stent multi pl i ci ti es.
Neither the ordi nal s nor the ordi nal couples are sets. How can we
j usti fy comparing or l i nking these two untotal isabl e col lecti ons ?
We have given the pri nci pl e for the forci ng of thi s i mpasse i n
chapter 1 0: we must, i f we can, defne the correspondence between
the ordi nal s and the couples via transfni te i nducti on, or recurrence.
The function will only be defned at successi ve level s, without us
having to consider the ' ai l s' between which it operates.
1 7. 1 7. Let (( (W"W2) be the function we wi sh to defne and whi ch,
to every coupl e of ordi nal s, wi l l make correspond bi uni vocal l y an
ordi nal : (( (W"W2) W3
We are frstl y going to root the function ( securel y i n its frst val ue,
whi ch wi l l correspond to the smal lest of the couples, the coupl e (0, 0).
Refer back to chapter 10 for the whol e of thi s procedure.
We posit expl i ci tl y:
KL 1
(( (0, 0) 0.
We wi l l then exami ne the case of successor couples ( compare the
typology of couples in 1 7. 1 4) . Let C2 be a coupl e whi ch succeeds
couple C" which we wi l l denote - i n an extensi on of the notation
adopted for the ordi nal s - by C2 S( Cd. The si mplest way i s to make
correspond to the coupl e, vi a (, a Cl, which is the successor of coupl e
C" the successor ordi nal of the ordi nal whi ch corresponds, vi a (, to
C, : we make the ordi nal s succeed 'in paral l el ' to the successi on of
couples. We thereby respect the basi c i dea of inducti on, or recurrence:
supposi ng the function ( to be defned for the couple C" we defne it
by an explicit rule for the couple C2 which succeeds C, . We therefore
posit:
KU 2 (( C2) (( S( Cd) S( (( Cd) .
let' s move on to l i mit coupl es. We suppose the function ( to be
defned for al l coupl es that precede a l i mi t couple CL. To al l these
coupl es, the function ( makes correspond an ordi nal W (( C) . The
idea i s evidently to take, as val ue of ( for the l i mi t coupl e Cl, the
ordi nal that comes j ust ' after' all of the ordi nal s thus associated, via
(, with the coupl es that precede CL. We know of the existence of
thi s ordinal that comes j ust ' after' a set of ordi nal s ( see N6): it is
the upper bound of that set, denoted by sup. We posit then that
(( Cl) i s the sup. of all the ordi nal s (( C) for the set of C smaller than
CL. So:
K L 3 (( Cl) sup. ( (( C) ) , for C < CL.
The i nductive defniti on of ( i s now compl ete, si nce we have
covered the three cases - the mi ni mum ( D, D)) , successors and l imits
- defni ng ( vi a an expl ici t rul e which makes its value depend upon
the val ues obtai ned ' bel ow' the term in questi on.
1 7. 1 8. A few exampl es.
What, for exampl e, is the val ue of (( (D, 1 ) ) ? We have seen that the
coupl e (D, l ) is the successor, i n the order of couples, of the couple
(D,D). We apply rul e 2: (( (D, l } ) S( (( (D, D} ) ) . But rul e 1 i ndicates that
(( (D, D}) D. Then it must be: (( (D, 1 }) SI D) 1 .
What i s the val ue of (( (D, r} ) ? We have seen that the coupl e (D,r)
comes j ust after the set of al l the couples (m,n), where m and n are
fnite ordi nal s ( the natural whol e numbers ) . Now it is cl ear that (
associ ates a fnite ordinal with each of these coupl es, since a successor
coupl e wi l l be associ ated wi th the successor of the ordi nal that cor
responds to its predecessor, and since one begi ns from D. To couples
of the type (m,1t) wi l l correspond the sequence D, 1 , 2, etc. Conse
quentl y, (( (D, r)) wi l l have as its val ue the upper bound of all the fnite
ordi nal s, that is, the frst i nfnite ( or l i mi t) ordi nal , which is to say ro
Thus (( (D, r} ) ro
These el ementary exampl es demonstrate that we are i ndeed in a
posi tion to cal cul ate ( for any coupl e whatsoever: it is enough to
' progress' the length of the wel l -ordercdness of coupl es. The val ue
for the frst coupl e being fxed, rul es 2 and 3 al l ow us to know the
val ue of ( for a coupl e C on the basi s of the val ues which ( assigns
to the couples which precede it.
1 7. 1 9. That our functi on (, defned i nductively with these three rules,
is defnitely bi uni vocal merits verifcati on, whatever evi dence we may
al ready have on this poi nt.
It must frst of al l be confrmed that ( i s i nj ective, or, in Dedeki nd' s
terms, disti nct ( see 4. 5 ) . In other words that, i f coupl e C
1
i s di fferent
from coupl e C2, then ordi nal (( Cd i s di fferent from ordi nal (( C2 ) . We
can assure oursel ves of thi s by casting our eye over the rules of i nduc
ti on. I f two coupl es are di fferent, they are ordered; say, C
1
< C2 The
value of (( C2 ) depends on the val ue of ( for the couples whi ch precede
it, and it i s di fferent from al l of these val ues. Speci fcal l y, it i s di fferent
from the val ue of ( for CI . whi ch comes before C2 We can therefore
be sure that C
1
.. C2 (( Cd ..(( Cz ) . Function ( is i nj ective.
In fact, we have a stronger property here: the functi on ' proj ects '
the order of coupl es i nto the order of ordi nal s ( techni cal l y, it is a
homomorphi sm from the order of coupl es i nto the order of ordi nal s) ,
such that, i f C
1
< C2, then (Cd E (( C2 ) . For, i f Cz comes after C
I . its
value for ( ( which is either the successor of the value of the coupl e
whi ch precedes i t, or the upper bound of the val ues of ( for al l the
coupl es which precede i t) i n any case surpasses the val ue of ( for C
1

Consequently: C
1
< C2 (( Cd E (( C2 ) .
I t remai ns t o be shown that function ( i s surjective, a modern word
meani ng that every possi bl e value of the function is effecti vel y ful
fl led. In other words that, for every ordi nal W, there exists an ordi nal
couple C for whi ch (C) = W.
Suppose that an ordi nal W exists whose val ue for function ( is not
a coupl e C. Then there exi sts a smal l er such ordi nal ( pri nci ple of
mi ni mal i ty) , say w. Thus al l ordi nal s smal ler than w do correspond,
vi a, (, to a coupl e. We can see then that w must necessari l y al so,
contrary to the hypothesi s, correspond, vi a (, to a coupl e. Because,
i f w i s a successor, whi ch means that w = S( wd and (( C) = N
j
, it must
then be ( rule 2 of the inductive defni ti on of f) that (( S( C) ) = S( wd =
w. And, if w is a l i mit, then, si nce al l the ordi nal s whi ch precede w
correspond, vi a (, to coupl es, w itsel f appears as the upper bound of
al l those ordi nal s, and thus, from rul e 3, i ts val ue for ( wi l l be the
coupl e that comes ' after' al l the couples correspondi ng to ordi nal s
smaller than w.
So, fnal l y, (is i ndeed a bi univocal correspondence between couples
and ordi nal s. Thi s correspondence is, i n addi ti on, an i somorphi sm
between the structure of order of coupl es ( vi a the Max. , the frst term,
and then the second term) and the structure of order of the ordi nal s
( belonging) . Suffce to say at thi s poi nt that the ordi nal coupl es are
a sort of ' doubled' i mage of si ngl e ordi nal s. Taken 'two by two' ,
nature is sti l l si mi l ar to itsel f. Nature i s its own mi rror.
1 7. 20. These wanderings i n nature through the l ooki ng-gl ass of the
double teach us:
- that there exi sts a wel l -orderedness over couples of ordinal s, such
that these couples obey, as do ordi nal s, the principle of
mi ni mal ity;
- that we can speak, as for the ordi nal s, of successor couples and
l i mi t coupl es, and that these attri butes can be uncovered by i mma
nent exami nati on al one of the structure of the couples which
possess them;
- that there exi sts between coupl es and ordi nal s a function ( which
has all the characteristics of a bi uni vocal correspondence, except
that the total ities between whi ch thi s function operates are incon
si stent, so that ( must be defned by transfnite induction;
- that thi s function ( defnes an i somorphi sm between the structure
of order of couples of ordinal s and the structure of order of the
ordi nal s, so that C] < Cz i mpl ies that (( Cd E (( C2 ) .
I n the mi rror of the doubl e, nature perseveres i n al l of its formal
comportments.
Identi cal means woul d al l ow us to establ i sh that triplets of ordi
nal s, of the form (W
J
,WZ,W
3
), have the same properties as coupl es
do, and i n particul ar that they are i n bi univocal correspondence with
the si ngle ordinal s. The same goes for n-tuplets of ordi nal s of the
form (WbWZ , ,Wn). In matter, it i s only the frst step that costs.
Doubled, nature mai ntai ns its order. Redupl icated i n fnite series as
long as you l i ke, nature persi sts in maintai ni ng its frst identity. Stabi l
i ty, homogeneity, order, mi ni mal ity, the ontological hi atus between
successors and l i mits: al l of thi s remai ns when the si mpl icity of the
ordi nal i s multi pl ied withi n the l i mi ts of the fni te. Nature is its own
hal l of mi rrors.
1 7. 2 1 . Mal l arme wrote: ' Nature is there, it wi l l not be added to, .
3
And it is a fact that, if one adds to nature, and even if one adds and
adds, and so on repeatedl y, the domai n of natural multiples attests
unabated to the pregnancy of the Same. Thi s i s what we grasp in
every experience of the natural : that ramifed growth, reproductive
di vi si on, far from suggesti ng to us the Other, reposes in itsel f, in the
eternal seat of its order.
Now, we know that every operati on, every al gebra, i s concerned
with a doubl ing or tri pl i ng of the terms upon which one operates.
We add two numbers to obtai n a thi rd, cal cul ate the smal lest
common di visor of two numbers, arrange i n a fni te sequence the
components of a pol ynomi al . . . All these disci pl i nes of reckoning
and algebra have as their substructure a fnite l i sting of numerical
marks.
If it is true that natural multi pl es, ordi nal s, furni sh the matter of
Number, we can understand why the possibility of operati ons, of
algebra, of reckoni ng, fnds i ts ontol ogical guarantee i n nature' s
capacity to maintain the i denti cal wi thi n di vi si on. Beneath the appar
ent variegation of schemes of reckoning, the vari ety of operati ons
and of algebraic structures, l i es thi s perseverance of natural bei ng,
thi s i mmanent stabi l ity i n fnite seri al ity. An operati on i s never any
thing more than the mode in which our thi nki ng accords with
Mal l arme' s maxi m: i f, wi thout exposi ng oursel ves to the di si ntegra
tion of the Other, we can combi ne two Numbers - ' add' one to the
other - it is because nature, taken as double, added to itsel f, re
attached to itsel f, mai ntai ns the i mmanent form of the mul ti ple-beings
through whi ch it i nconsi sts.
An operation, a counting, an algebra, are onl y marks of our
thought' s being caught i n the mi rror-games whi ch it pleases bei ng
to proffer, under the l aw of the Same to which natural mul ti pl es
di spose it.
1 8
Al gebra of Numbers
1 8. 1 . We must fnal ly come to counting.
Once its bei ng has been fxed, the combi natory capacity of Number
i s a mere consequence. It ari ses from an i nvestigative ingenuity as to
the ways i n which couples or triplets of Numbers can be l i nked. But
the source of these l i nks is hel d completel y withi n the concept by
means of which Number is anchored i n being. Al l that operations
can do i s to deploy - in the numberless domain of Number - the
prodi gal ity of being in its possi bl e connections.
Concomitantly, the di ffculty resi des i n the choice of 'good' defni
tions of the l i nks, so that they shoul d conform to the faci l ities of
cal cul ati on: we wish the operations to be associ ative, for there to be
a neutral element, i nverses, and it woul d hel p i f they were al so com
mutative. We would be even happi er i f operati ons combi ned wel l
behavedly amongst themselves, wi th a di stri butivity of one with
regard to the other. To arrive at these res ults, Number must be scru
tinised and we must careful l y authenticate the links we wish to
defne.
1 8. 2. The substanti al results to be obtained through the i ngenuity of
operati onal defni ti ons are as fol l ows:
1 We can defne a frst operation on Numbers named addition and
denoted by +, whi ch has the properties of a commutative group:
- associ ati vity: Nt + ( N2 + N3 ) = ( Nt + N2 ) + N3 ( one can count
'in any order' , and achi eve the same resul t) ;
ALGEBRA OF NUMBERS 1 99
- a neutral element ( which is Zero) : N, + 0 = N, ;
- i nversion ( which is the symmetric counterpart ) : N, + ( -N, ) = 0;
- commutativity: N, + N2 = N2 + N, .
2 We can defne on Numbers a second operation, named mul ti pl i ca
ti on, and denoted by , which has the fol l owi ng properti es:
- associati vity: N,
.
( N2 N3 ) = ( N,
.
N1) N3 ;
- neutral element ( which is the Number 1 ) : N, l = N, ;
- existence of an inverse i ( N) for every Number di fferent from
Zero: N, i ( Nd 1 ;
- commutativity: N, N2 = N2 N, .
3 Mul ti pl ication is distri butive in relation to additi on: N,
.
( N2 + N
3
)
( N, N2 ) + (N , N3
) .
These three operational consi derati ons woul d lead us t o say that
Numbers form a commutative {eld, i f it were not for one probl em:
Numbers do not even form a set, because they are i nconsistent. How
can something be a feld - which is supposed to be an al gebraical l y
defned entity - i f it cannot be counted as a multi pl e?
Therefore, prudently, we wi l l say onl y thi s: that every set consti
tuted of Numbers whose matter i s less than a given cardi nal i nfnity
(therefore every set constituted from Numbers whose matter i s
bounded by a ' brute' fxed i nfnite quantity) can be gi ven the structure
of a commutative feld.
1
What i s more, as can be proved for the
rationals or the real s, there are other sets of Numbers that are al so
commutative fel ds. Numerical i nconsistency can be 'secti oned' into
innumerable algebraic structures.
These logical caveats aside, the al gebra of Numbers is the richest
concei vabl e: its cal culative capacities equal - for exampl e - those with
which the real Numbers furni sh us ( in particul ar, it can be proved
that every Number has a square root, which i s not the case i f one is
operating, for example, i n the feld of rati onal numbers ) .
1 8. 3. A resul t at once l abori ous ( i n its procedures of veri fcati on) and
of key importance ( for the val idity of our concept of Number) is the
fol l owi ng: operations defned on Numbers coi ncide wi th operations
defned on ' our' numbers, i f the latter are thought i n thei r being as
Numbers. In other words: take two real numbers r, and r2, taken i n
thei r usual algebraic sense. I f the sum of rl and r2, such as we know
i t, is the real number r
3
, then the sum of r, and r2 such as it i s repre
sented ' itsel f' in Numbers ( with Numbers of matter fnite or equal to
0, see 1 6.28) - ' sum' being taken i n the sense of the addition defned
on Numbers - will be preci sel y the representative, wi thi n Number,
of the number r3 . The same will go for mul ti pl icati on, etc. More
200 OPERATIONAL DIMENSIONS
techni cal l y, we can say that the fel d of the real s, as we know it in
classical anal ysi s, i s i somorphi c to the real s thought as subset of
Numbers.
It i s not, therefore, sol el y in thei r being that ' our' usual numbers
can be thought of as si ngul ar types of Numbers, but also i n their
al gebra. Our real Numbers are ul ti matel y indistinguishable from real
numbers. In particul ar, real Numbers constitute a complete ordered
Archi medean fel d, whi ch is the uni vocal customary defnition of real
numbers.
It can be sai d, ul ti mately, that al l the di mensi ons and capacities of
' hi storical ' numbers are retained by thei r presentati ve instance i n the
i nnumerabl e swarm of Numbers. Which confrms:
- that the ontological essence of a number is nothi ng more than
that whi ch our thought apprehends it to be when it is determined
as a type of Number;
- that the operati onal or algebraic properties are onl y the effect of
a correct determi nati on, on the basi s of natural multi pl icities, of
the being of Number.
We therefore fnd the programme of uni fcation of the concept of
Number ( one sol e concept which subsumes the natural whol e
numbers, the negative whol e numbers, the rational s, the reals and the
ordi nal s) to be whol l y real ised, frstly in mul ti pl e-being, and then in
the operati onal di mensi ons .
It is now possi bl e for us to speak freely of, and to submit to cal
cul ati on, enti ti es previ ousl y devoid of any sense, l i ke the sum of an
ordinal and a real number, or the division of a transfnite ordinal by
a rati onal number, or the square-root of the division by three of an
ordi nal , etc. Incredi ble equati ons l i ke:
- whi ch, i n the di spersed and l acunary hi storical theory of numbers,
would have made absol utely no sense - in the uni fed framework of
the concept of Number become perfectly meani ngful algebraic formu
l ae, i ndicating certai n procedures of cal cul ati on and defnite results.
Number thus founds in being the l iteral connection of what,
under the di sparate name of ' numbers' , had defned heterogeneous
domai ns.
ALGEBRA OF NUMBERS 20 I
1 8. 4. The defnition of operations on Numbers is essenti al l y a techni
cal affai r. Whoever wi shes to fol l ow it i n al l i ts detai l is referred to
the I iterature.
2
Nevertheless, its ani mating spi rit al lows a revi si on of
concepts, a fnal passage through the Idea of Number. In particul ar,
the systematic use of transfnite induction hi ghl ights the fact that
Number, thought of i n i ts bei ng, i s essenti al l y an i nfnite multi ple ( the
section of a form from an i nfni te ordi nal-matter) . In the same way,
the recourse to sub-Numbers of a Number in order to construct
operations ' from bel ow' attests to the i mportance of the fact that
every Number can be presented as a cut of its l ow set and its high
set ( see chapter 14) . And agai n, it i s by presenting the result of an
operation as a cut ( see chapter 1 5 ) - that i s, by uti l i sing the funda
mental theorem - that we can handl e i nducti on. Lastl y, the correl a
ti on explored i n chapter 1 7 between couples of ordi nal s and ordi nal s
pl ays a maj or rol e in thi s whol e process - as one might expect ( si nce
an operation connects two Numbers ) . So as not to forego these reca
pitulations i n thought, we wi l l cover the essenti al s of the defnition
of addition.
18. 5. The general idea is as fol l ows: given two Numbers NI and N2,
we can make them correspond to two ordi nal s W I and W 2 si mpl y
by taki ng thei r respective matters, WI = M( Nd and W2 = M( N2 ) . We
know that a certai n ordi nal corresponds to these two ordi nal s vi a
the bi uni vocal function (, whi ch associ ates an ordi nal with every
couple (Wl ,W2) of ordi nal s ( see 1 7. 1 7) . This ordi nal wi l l fx the
'level' of defnition of the additive operati on: we wi l l suppose that
addition i s defned for all couples of Numbers N
3
and N
4
of matter
W3 and W
4
, such that the ordi nal that corresponds vi a (to the coupl e
(W3 ,W
4
) is smaller than the ordi nal associ ated with the coupl e
(WI ,W2). We then propose an expl ici t rule, which wi l l defne the sum
N I N2 on the basi s of sums of the type N3 N
4
, defned at a l ower
ordi nal level .
Now, such sums are given by the sub-Numbers of NI and N2 . A
sub-Number, being a ' partiti on' of a Number for an ordi nal smal l er
than i ts matter, has an i nferi or matter.
We can then pass on to the next stage, which i s the core of the
constructi on. Take the Numbers NI and N2, of matter WI and W2.
Consider a sub-Number N I /W3 of N I , and a sub-Number Nz/w
4
of
N2. Now take the couples (W
3
,W2), or (WhW
4
)'
I say that they are
lower than the coupl e (WI ,W2), by the rules of order of coupl es, and
remembering that W3 and W
4
are respectively smal ler than WI and
W2 ( see 1 7. 6) . Thi s is an excel lent exerci se for the reader, but see
the note.
3
As the function f is an i somorphi sm of the order of couples with
the order of the ordi nal s, we will also have:
f( W
3
,WZ) E f( Wt
.
Wz) ,
and
f( Wt
.
W
4
) E f( W"Wz ) .
Whi ch i s t o say that the ordinal level associated with couples
of Numbers of the type (Ndw3, Nz), or (NhNz/W4) will always be
lower than the ordinal level associated with couples of Numbers
(N
h
N2).
Given thi s fact, i n order i nductively to defne the sum of N, and
Nz, we can suppose defned sums of the type Ndw
3
+ N2, or N, +
Nz/w
4
, which pertai n to a lower ordi nal level . We wi l l thus pass on
to the defni ti on of N, + N2 by formulating a rule which assigns the
val ue of this sum on the basis of the various val ues between N, and
Nz on the one hand, the sub-Numbers of N, and Nz on the other.
The i mmanent concept of sub-Number wi l l serve to underwrite the
i nducti on, which fxes their ordi nal level on the basis of a coupl e
formed of the matters of the two Numbers under considerati on.
Fi nal l y, the strategy wi l l mobi l ise the fundamental theorem of the
cut. We will begin with the l ow set and the high set of the two
Numbers N, and Nz We suppose defned the sums of each of the
two Numbers with the sub-Numbers of the l ow set and of the high
set of the other Number, according to a fxed combi nati on. These
sums can be assumed, because thei r ordinal level is lower. We can
thus obtai n two sets of Numbers, and the sum of N I and Nz wi l l be
the uni que Number defned as cut of these two sets.
1 8. 6. Inductive defnition of the addition of two Numbers
' Level zero' of the i nduction contai ns only the Number ( 0, 0) . It i s the
only one to have as matter. We can thus posi t:
K L 1 ( 0, 0) + ( 0, 0) = ( 0, 0) .
We wi l l now suppose t hat addi ti on is defned for al l levels lower
than an ordi nal W, that is, all levels corresponding to Numbers N3
and N
4
( taken i n that order) such that, their respecti ve matters being
W3 and W
4
, it is the case that f( WJ ,W
4
) E W.
Now take a coupl e of Numbers NI and Nz such that, thei r respec
tive matters being WI and Wz , it i s the case that f( W"Wz ) W. In
other words a coupl e of Numbers belonging to ordi nal level W.
ALGEBRA OF NUMBERS 203
We have remarked that al l the couples of type N, and Nz/w, or
N /w and N2 , where N /w and Nz/w are sub-Numbers of N, and N2,
belong to ordi nal levels i nferi or to those of the coupl e N, and N2,
and therefore i nferior to W.
It follows that we can suppose defned al l the addi ti ons of the type
N, + Nz/w, or Ndw + N2
We must agree on an i mportant written convention here. We wi l l
write N, + Lo( N2 ) for the set of Numbers constituted by al l the results
of the addi ti on of N, with each of the Numbers of the low set of Nz
( the low set of N2 i s constituted, remember, of al l sub-Numbers of
Nz smal l er than N2 ) . I f Lo( N2) i s empty, the Number denoted by
N, + Lo( N2 ) would be undefned ( we wi l l not consi der thi s i n the
calculations) .
In the same way, we write NJ + Hi ( N2 ) for the set of Numbers
constituted by all the results of the addition of N, with each of the
Numbers of the high set of N2 (the high set of N2 being constituted
by all sub-Numbers of N2 l arger than N2 ) . The convention wi l l al ways
be not to bother wri ti ng thi s i f Hi ( N2 ) is empty.
We wi l l adopt the same notati on to designate sets of Numbers whi ch
result from additions i mpl icated i n Lo( N, ) + N2 , or Hi ( N, ) + N2.
Addition wi l l then be defned as fol l ows: on the one hand we take
the set of Numbers constituted by all the Numbers of Lo( Nd + N2,
together with al l the Numbers of N, + Lo( Nz ) ; on the other hand,
the set constituted by al l the Numbers of Hi ( Nd + N2, together wi th
al l the Numbers of NJ + Hi ( N2 ) . In other words, we ' col lect' on one
si de the Numbers whi ch are the sum of N, and N2 and the low sub
Numbers of the other Number, and on the other si de the same sums,
but wi th the hi gh sub-Numbers.
We thus obtai n two sets of Numbers, which we can cal l L and H.
It i s not hard t o prove, by way of a ' i ncremental ' i nduction which
I leave to one side,
4
that L and H are in a si tuati on of a cut: every
Number of L i s smal ler than every Number of H.
We then uti l i se the fundamental theorem ( chapter 1 5 ) . The result
of the addition of N, and N2 wi l l be preci sel y the Number whi ch
makes a cut between these sets, that i s, the uni que Number of mi ni mal
matter situated between the sets:
We posit:
L ( Lo( Nd + N2, N, + Lo( Nz ) )
H ( Hi ( Nd + N2, N, + Hi ( N2 ) )
KU 2 N, + N2 UH, cut of the two sets defned above.
1 8. 7. Addition is commutative
In fact, this cut, which supposedly defnes the sum N2 + Nh operates
on the same sets as the cut whi ch defnes N I + Nl, as one can show
i nductively with no di ffcul ty.
It is true at l evel 0, where there i s onl y the sum, certai nl y commu
tative, + 0.
Suppose that the sums of ordi nal levels i nferi or to (( WI ,W2 ) W
are commutative. Then, in parti cul ar, the sums Lo( Nd + Nz or N, +
Lo( N2 ) are commutati ve. So the set L ( Lo( Nd + N2, NI + Lo( Nz ) ) ,
whi ch serves t o defne N, + Nz , i s composed of the same Numbers
as the set L
'
( Lo( N2 ) + ( Nd, Nz + Lo( Nd) , whi ch serve to defne
Nz + NI . Evidently, the same goes for the set H. And consequently,
Nz + NI , being defned by the same cut as NI + N2, i s equal to it:
addition is commutative.
1 8. 8. The Number 0, which is more precisely the Number ( 0, 0) , is
the neutral element for addition
It i s a questi on of proving that, for every Number N, N + N.
Induction can thi s ti me be appl ied di rectl y to the ordi nal-matter of
the Numbers.
It i s true at l evel 0, si nce rule 1 prescri bes that + 0.
Suppose that thi s i s true for al l the Numbers of l ower matter than
WI ' In other words, for every Number N of matter N such that
W E Wh N + N.
Now take a Number NJ of matter WI ' Let' s exami ne the sum
N I + 0. The sets L and H of the cut which defne the addition
are:
L ( Lo( Nd + 0, Lo( O) + Nd
H ( Hi ( Nd + 0 , Hi ( O) + Nd
But the l ow set and the hi gh set of the Number 0 - that is, ( 0, 0)
are empty ( 0 has no sub-Numbers ) . The conventions adopted in 1 8. 6
prohi bit us from taki ng i nto account the terms Lo( O) + NI and
Hi ( O) + N, . So we actual l y have:
L ( Lo( Nd + 0)
H = ( Hi ( Nd + 0)
But Lo( Nd and Hi ( Nd are composed of sub-Numbers of Nh
and therefore of Numbers of lower matter than WI ' Consequently,
the hypothesi s of i nduction appl ies to all the Numbers of Lo(N d
or of Hi ( Nd: for any such Number, s ay Ndw2' i t i s the case that
Ndw2 + 0 = N1 1w2 '
We can, with a sl ight abuse of notation, write thi s resul t in the
form Lo( Nd + 0 Lo( Nd, Hi ( Nd + 0 Hi ( Nd. So that, ul ti mately,
L and H, whi ch defne by a cut the sum N
1
+ 0, are no other than
Lo( Nd and Hi ( Nd. But the Number defned by a cut between its l ow
set and its high set i s preci sel y the Number N
1
itsel f, and so it i s
indeed the case that N I + 0 = N I '
The i nduction is complete: for a Number N, whatever its matter,
o i s a neutral element for addi ti on.
1 8. 9. Every Number N apart from 0 allows the Number -N as its
inverse for addition: N (-N) = 0
An important poi nt: si nce -N inverts the form and the resi due of N,
the low set of -N is composed of the Numbers -Nlw, where Nlw i s a
Number from the high set of N; and the high set of -N is composed
of the Numbers -Nlw, where Nlw is a Number from the low set of N.
A sub-Number Nlw is in the low set if w i s i n the form, and i t is i n the
high set i f w is i n the resi due. These determi nati ons wi l l be i nverted in
-N. And, since everythi ng that precedes w i n N i s also i nverted ( what
was i n the form i s i n the residue, and what was i n the resi due i s i n the
form) , in addition to the exchange of the low set and the high set, we
wi l l also have an exchange of the signs of posi ti ve and negative.
In an abuse of notation, we could therefore write the high set of
-N as -( Lo( N) ) , and the l ow set of -N as -( Hi ( N) ) .
The resul t ( see the i nductive defni ti on of addi ti on) i s that the two
sets L and H which defne by a cut the sum N + ( -N) are the
following:
L = ( Lo( N) + ( -N) , N + ( -( Hi ( N) ) ) )
H = ( Hi ( N) + ( -N) , N + ( -( Lo( N) ) ) )
So the strategy of the proof consists i n proving that al l the N
umbers of L are negative and al l the Numbers of H posi ti ve. The
result is that 0 is situated between L and H and that, being necessari l y
of mi ni mal matter i n that posi ti on, it is 0 that occupies the posi ti on
of the cut between L and H. Consequentl y, N + ( -N) = O.
MMA I f the sum NI + N2 i s positive, i f NI + N2 0, then
-( Nd < N2, and -( N2 ) < N
1

The lemma is true at ordi nal level 0, because at that level it cannot
possi bl y be the case that N 1 + N2 O.
Suppose that it is true up to ordi nal level W: for every pai r of
Numbers N
3
and N
4
such that {( W3,W
4
) E W, the property in ques
ti on hol ds. I say that it also must hol d for every pair of Numbers N,
and N2 such that {( W"W2) = W.
The sum N, + N2 i s defned by the cut UH. I f thi s cut is
positive, it is because set L contai ns positive Numbers,
s
or else the
cut woul d be negative or null ( see the argument on cuts i n 1 5. 1 1 ) .
As for set H, i t onl y contai ns positive Numbers. Consequently,
there are Numbers in Lo( Nd + N2 or i n N, + Lo( N2 ) that are
positive, and all the Numbers of Hi ( Ni l + N2 or of N, + Hi ( N2)
are so.
Take for example N, lw + N2 as a positive Number of Lo( Nd +
N2 . The pai r of N, lw and N2 is of lower ordi nal level than W, and
the lemma i s therefore supposed to be true of it: si nce the sum N dw
+ N2 i s positi ve, it i s the case that -( N2 ) < N21w, and, since N dw is
i n the l ow set of N" it i s a fortiori the case that -( N2 ) < N, . In exam
i ni ng the other components of sets L and H, the l emma can be
establ i shed in al l general ity.
Now l et' s come back to the sum N + ( -N) . Consi der the set L
whi ch defnes it by a cut, so:
L = ( Lo( N) + ( -N) , N + ( -Hi ( N) ) ) .
Suppose that there are positive Numbers i n L. Take for example
one such Number Nlw + ( -N) , where Nlw i s from the l ow set of N.
I n vi rtue of the l emma, it i s the case that -( -N) < Nlw, so N < Nlw,
whi ch is i mpossi bl e si nce Nlw, being from the low set of N, must be
smal ler than N. If N + ( -N/w) is posi ti ve, N/w being in the high set
of N, it must be the case that N/w < N, which is prohi bi ted, si nce
Ndw bel ongs to the high set. We meet with an i mpasse, and so there
are no posi ti ve Numbers in set L.
Symmetrical deductions woul d demonstrate that there are no nega
tive or nul l Numbers i n set H.
Fi nal l y, the cut LlH which defnes the result of the addition
N + ( -N) operates between a set L of negative Numbers and a set H
of posi ti ve Numbers. The N umber of mi ni mal matter situated between
these two sets i s necessari l y 0, and so N + ( -N) = O.
So we can say that -( N) is the i nverse of N for addition.
1 8. 1 0. Confrmi ng that the additi on of Numbers is associative i s,
as al ways, a ti resome calcul ati on. It i s, it i s . . . To the extent that
we have establ i shed that Numbers, endowed - so to speak - with
addi ti on defned i nductively by the cut:
woul d form ( were it not for the i nconsi stency of thei r ' Al l ' . . . ) an
ordered commutative group, of whi ch the Number Zero ( either ( 0, 0)
or 0, i t' s al l the same) i s the neutral el ement.
To confrm that the ' representatives' in our Numhers of whole
positives and negati ves, rati onal s, rea I s, ordi nal s, are in fact these
numbers themselves, but thought i n their being, we must prove that
addition (in the normal sense) of these numbers coi ncides with addi
ti on of their being as Numbers. For exampl e, i f rl and r2 are numbers
from the real feld, and i f rl + r2 r3 with ' cl assi c' addi ti on, then the
Numbers r J , r2 and r3 defned as Numbers of fnite matter or of matter
as we presented them i n 16. 27, are such that, addition being
defned i nductivel y as above, it i s al ways the case that rl + r2 r3 .
These confrmations of algebraic i somorphy demand no l ittle i nge
nuity, above all when it comes to mul ti pl icati on ( which l abyrinth we
will avoid entering into) .
18. 1 1 . I wi l l content mysel f wi th carryi ng out the verifcation for
natural whol e numbers.
Remember ( from 16. 1 3 ) that a natural whol e number n presented
as Number is of the form ( n, n) , where n is a fni te ordi nal . Recal l
al so ( i bi d. ) that the l ow set of n i s consti tuted by al l the whol e
Numbers lower than n, and that i ts high set is empty.
Take two natural whol e Numbers ( nJ ,nd and ( n2, n2 ) ' Thei r sum i s
formal l y defned by the cut:
But, as Hi ( nd and Hi ( n2 ) are empty, the sums of set H of the cut
are not defned ( convention on the defni ti on of additi on, see 1 8. 6) .
Set H is therefore empty, whi ch amounts t o sayi ng that t he s um i s
si mpl y the upper bound of set L.
Si nce Lo( nd is the set of Numbers l ower than nJ , t he s um Lo( nd
+ n2 i s constituted by al l the sums + n2, 1 + n2, . . . ( nl - 1) + n2'
And, in j ust the same way, nl + Lo( n2 ) is constituted by al l the
sums nl + 0, nl + 1, . . . , nl + ( n2 - 1 ) .
The l argest Number of these sums i s i n al l evi dence the Number
nl + n2
- 1 .
Reasoning by inducti on: suppose that, up to the ordi nal rank
which corresponds to the pai r of Numbers nJ , n2 ( so, i n real ity, the
Numbers ( nJ ,nd and ( n2, n2 ) ) , therefore up to the ordi nal N f( nJ , n2 ) ,
i t i s true that the sum of wholes as Numbers wi l l be the Number
whi ch corresponds to the ( normal ) sum of the wholes. In particul ar,
that it i s true for the pai r nl > ( 112 - 1 ) , which is evi dently of a lower
ordi nal rank than the pair nl >n2 . It i s therefore to be supposed that
( nl > nd + ( ( n2 - 1 ) , ( n2 - 1 ) ) = the Number which corresponds to the
ordi nary addi ti on of the numbers n, and (n2 - 1 ) , that i s, the Number
( n, + ( n2 - 1 ) , n, + ( 112 - 1 ) ) , where the sign + denotes the ordinary
addition of whol e numbers. 6
Now we come to see that the l argest Number in the set L which
defnes n, + n2 i s preci sel y n, + ( 112 - 1 ) . In vi rtue of the hypothesis
of i nducti on, thi s Number is the Number which corresponds to its
being written as an ordi nary addition - the Number which inscribes
the whol e number n, + n2 - 1 .
Now, n , + n2 ( i n the sense of the addition of Numbers) i s the upper
bound of L. And every upper bound i s a Number of the type ( W,W) ,
as we have demonstrated in 1 5. 9. The upper bound of L wi l l therefore
be the smal lest Number of the type ( W, W) to be superi or to the l argest
Number i n L, whi ch is the Number ( n, + n2 - 1 , 11 , + n2 - 1 ) ( where
the signs + and - have thei r traditional meaning, as when dealing
with numbers ) . This Number i s evidently ( n, + n2,n , + n2 ) , because
n, + n2 i s the fnite ordi nal which comes i mmedi ately after the fnite
ordi nal n, + n2 - 1 .
Consequentl y, the sum ( i n the Number sense) of the two whole
Numbers n, and n2 i s the Number that represents the number sum
n, + n2 ( i n the number sense) . The addition of whol e Numbers is
i somorphic to the tradi ti onal addition of whol e numbers.
The treatment of whol e negative numbers poses no great probl em
( an i nteresti ng exerci se) . Thus it is confrmed that the whole positive
and negative Numbers form a commutative group isomorphic to the
additive group of the ring 7 of the algebraic whol e numbers.
The reader will have grasped the essence of operational proce
dures: fnd a ' good' i nductive defni ti on of the l i nks, prove the cl assi c
al gebraic properties ( associ ati vity, commutativity, neutral element,
i nverse, di stri buti vity . . . ) , confrm that what one obtains i s i somor
phi c, for the cl assi cal numbers represented in Numbers, to the
structures which these numbers are endowed with.
However l abori ous these efforts mi ght be, they lead to the desi red
concl usi on: al l the cl assic algebraic structures (the ring 7 of algebraic
whole numbers, the feld a of rational s, the fel d I of real s) , and al l
the ' i nconsistent' al gebras ( addition and mul ti pl ication of ordi nal s)
are i somorphi c to the substructures di scerni ble wi thi n Numbers.
And so it i s that all types of numbers, without exception and
in thei r every di mensi on, are subsumed by the unique concept of
Number.
Concl usi on
1 9
I n Conclusion:
From Number to Trans- Being
19. 1 . Number is neither a trait of the concept, nor an operational
fcti on; neither an empi ri cal given, nor a constitutive or transcenden
tal category; neither a syntax, nor a language game, not even an
abstraction from our idea of order. Number i s a form of Being. More
precisely, the numbers that we mani pul ate are onl y a ti ny deducti on
from the i nfnite profusi on of Bei ng i n Numbers.
Essenti al l y, a Number i s a fragment sectioned from a natural mul
ti pl icity; a mul ti pl icity thought, as ordi nal , in i ts being qua being.
The l inear order of Numbers, l i ke thei r algebra, i s our way of tra
versing or investigating thei r bei ng. Thi s way i s l abori ous and l i mited.
It exhi bits Number in a tight network of l i nks, whose three pri nci pal
categories are successi on, l i mi t and operations. Thi s i s where the i l l u
si on ari ses of a structural or combi natory bei ng of Number. But, i n
real ity, the structures are consequences, for our fni te thought, of that
which i s legible i n Number as pure mul ti pl icity. They depose Number
in a bound presentation whi ch makes us bel ieve that we mani pul ate
it li ke an obj ect. But Number is not an obj ect. Before every bound
presentation, and i n the un-bound eternity of its bei ng, Number i s
avai l able to thought as a formal section of the mul ti pl e.
We mi ght al so say that between Number, whi ch i nscri bes i ts
section i n the unrepresentabl e inconsistency of natural mul ti ples, and
number, which we mani pulate according to structural l i nks, passes
the di fference between Being and beings. Number i s the pl ace of the
bei ng qua bei ng, for the mani pul abl e numerical ity of numbers.
Number ek-sists i n number as the l atency of its bei ng.
21 2 CONCLUSION
1 9. 2. Thi s onl y makes it more remarkable, then, that we can have
some access to Number as such, even if thi s access sti l l i ndicates an
excess: that of being over know ledges, an excess mani fest in the
numberl ess extent of Numbers, compared to what we can know by
structuri ng the presentation of types of numbers. That mathematics
al l ows us at least to designate this excess, to accede to it, confrms
the ontological vocation of that di sci pl i ne. The hi story of mathemat
ics, for the concept of Number as for every other concept, i s precisely
the history, intermi nabl e i n pri nci pl e, of the rel ation between the
i nconsi stency of being as such, and what our thought can make con
sistent of thi s i nconsi stency. Mathematics establ i shes ontology as the
hi storical si tuati on of bei ng. It progresses constantl y within ontico
ontological di fference, bringing to l ight, as the l atency of the struc
tures presented i n the ontological si tuati on, an excessi ve horizon of
i nconsi stency, of whi ch structures are onl y effects for a fnite thought.
It i s this traj ectory which we have reconstructed at one of its points:
that which designates, beyond numbers, the i nconsi stent multi ple
eternity of Numbers .
19. 3. Number is thus rendered over to being, and subtracted from
the humani ty of operations or fgures of order, whi ch nevertheless it
continues to subtend i n thought. The task concerning Number, and
numbers, can onl y be to pursue the depl oyment of their concept
withi n ontico-ontological di fference. Number fal l s wi thi n the excl u
si ve purview of mathematics, at l east so far as the thi nking of number
i s concerned. Our phi l osophi cal proj ect prescribes this exclusi vity,
and designates where Number is gi ven as the resource of being within
the l i mits of a situation, the ontological or mathematical situation.
We must abandon the path of the thi nki ng of Number fol l owed
by Frege or Peano, to say nothi ng of Russell or Wittgenstei n. We
must even radi cal ise, overfow, thi nk up to the poi nt of di ssol uti on,
Dedeki nd' s or Cantor' s enterpri se. There exi sts no deduction of
Number, it i s sol el y a question of a fdel ity to that whi ch, i n its
i nconsi stent excess, i s traced as hi storical consi stency i n the intermi
nabl e movement of mathemati cal refoundations.
The modern i nstance of thi s movement attests to the voi d and the
i nfnite as materi al s for the thi nki ng of Number. Nevertheless, none
of these concepts can be i nferred from experience, nor do they propose
themselves to any i ntui ti on, or submi t to any deducti on, even a tran
scendental one. None of them amounts to the form of an obj ect, or
of obj ectivity. These concepts ari se from a decision, whose written
form is the axi om; a deci si on that reveal s the openi ng of a new epoch
for the thought of being qua bei ng. Being asks nothing more of us,
FROM NUMBER TO TRANS-BEING 21 3
at this poi nt, than that we doggedly pursue the i nscri ption - wi thi n
a revi sed ontological si tuati on - of that whi ch, i n traci ng the i ncon
sistent latency of bei ng, fai thful l y prepares the rupture at a poi nt of
that pl ace where it consi sts for us.
19. 4. It is then possi bl e to mai ntai n that the contemporary ' banal i sa
ti on' of number is outside al l thought. The reign of number, the
portents of whi ch I di scussed at the begi nni ng of thi s book, is i ntran
sitive to the mathematical thought of Number. It i mposes the fal l a
ci ous i dea of a bond between numeri cal ity and val ue, or truth. But
Number, whi ch i s an instance of bei ng as such, can support no val ue,
and has no truth other than that which i s gi ven to it i n mathematical
thought, effectuating its hi storical presentation for us.
I f the reign of number - i n opi ni on pol l s or votes, in nati onal
accounts or in pri vate enterprise, i n the monetary economy, i n the
asubj ectivising eval uation of subj ects - cannot be authori sed by
Number or by the thi nki ng of Number, it is because i t fol l ows from
the si mple law of the si tuati on, whi ch is the law of Capital . This l aw
assures, as does every l aw, the count-for-one of that whi ch i s pre
sented in the situation, it makes our hi storical situation consi st, but
it cannot make any cl ai m to truth: neither to a truth of Number, nor
to a truth which would underlie that which Number designates as
form of being.
In our situati on, that of Capi tal , the reign of number is thus the
rei gn of the unthought sl avery of numeri cal i ty itsel f. Number, whi ch,
so it is cl ai med, underlies everythi ng of val ue, is in actual fact a pro
scription against any thinking of number i tsel f. Number operates
as that obscure poi nt where the si tuati on concentrates its l aw;
obscure through i ts being at once sovereign and subtracted from al l
thought, and even from every i nvestigation that orients itsel f towards
some truth.
The resul t is that al l thought necessari l y depl oys itsel f today in a
retreat with regard to the reign of number, i ncl udi ng every thought
that tri es to make a truth of Number. It i s i n thi s sense that we must
hearken to Mal l arme' s slogan, more perti nent than ever: that of
restrained action.
1
This whole meditation on the concept of Number, because it
restores it to being, necessi tates the inversion of the contemporary
j udgement such as it i s presented under the banner of number. We
must say, against this j udgement, that nothing made into number is
of val ue. Or that everythi ng that traces, in a si tuati on, the passage of
a truth shal l be si gnal led by i ts i ndi fference to numeri cal i ty. Not so
that thi s indi fference can in its turn be made i nto a criteria, because
21 4 CONCLUSION
many proj ects, having no number, have no truth either. But thi s indif
ference is a necessary subj ectivity.
The reverse side of the abundance of capital is the rarity of truth,
in every order where truth can be attested to: science, art, politics
and love.
19. 5. But i f the true i s, on pri nci ple, in our situation, subtracted from
the rei gn of number, whi ch is onl y a law of thi s si tuati on, what is
the origin of thi s process ?
A truth can depend nei ther upon being as such ( thi s is why it does
not signal itsel f through a Number) nor upon the contemporary situ
ation, which i s that of Capital ( thi s i s why it does not si gnal itself
through numbers ) . Its ori gi n i s evental . But the event i s not non-being,
however much it exceeds the resources of situati on-bei ng. The best
way to say it woul d be that the event i s of the order of trans-being:
at once 'held' within the principle of being (an event, l i ke everything
that i s, i s a multi pl e) and in rupture with thi s princi ple ( the event
does not fal l under the law of the count of the si tuati on, so that, not
being counted, it does not consi st ) . Evental trans-being i s at once
mul ti pl e and ' beyond' the One - or, as I have chosen to call it, ultra
One. The possi bi l ity that there can be a truth, in a situation whose
state has whol l y succumbed to numbers, depends upon a fdelity,
subtracted from numbers, to thi s ultra-One.
To thi nk Number, as we have tried to do, restores us, either
through mathemati cs, whi ch i s the hi story of eternity, or through
some faithful and restrai ned scrutiny of what is happening, to a
supernumerary hazard from which a truth originates, always heter
ogenous to Capi tal and therefore to the sl avery of the numerical . It
i s a questi on, at once, of del ivering Number from the tyranny of
numbers, and of rel easi ng some truths from i t. In any case, restrained
action is the pri nci pl e of a remote di sorder: it establ i shes mathemati
cally that order i s but the al l -too-human precarity of a thi nki ng of
the bei ng of Number; it proceeds, effecti vel y and theoreti cal l y, to
the downfal l of numbers, whi ch are the l aw of the order of our
si tuati on:
' Li ke a god, I put in order nei ther one nor the other . . . '
2
Notes
Translator's Preface
1 London: Continuum, 2005.
2 Mi nneapol i s: Uni versity of Mi nnesota Press, 2003.
3 London: Continuum, 2004; in particul ar, ch. 5, 'The Being of
Number', represents an extremely condensed gloss of the present
work.
4 See for example ch. 2n 4.
Chapter 0 Number Must Be Thought
1 [Dedeki nd, R. , Was sind und was sol/en die Zahlen ( Braunschweig:
F. Vieveg, 1 8 8 8 ) ; Engl i sh transl ation The Nature and Meaning of
Numbers, in Beman, W. W. ( ed. , trans . ) , Essays on the Theory of
Numbers ( La Sal l e, IL: Open Court, 1 90 1 ; reprinted NY: Dover
1 963) . Badi ou' s reference i s to the transl ati on by J. Mi l ner, with H.
Sinaceur' s i ntroduction, Les Nombres, que sont-i/s et a quoi servent
ils? ( Pari s: Navarin, 1 979) . Al l references given below are to the
numbered sections of Dedeki nd' s treati se. - trans. )
2 [aei 0 anthropos arithmetizei - ' man i s always counting' - in
Dedekind, ' Numbers' , Preface to the frst edn. - trans. )
3 [ Marx and Engels, Communist Manifesto, transl ated by S. Moore,
with introduction and notes by G. S. Jones ( London: Pengui n,
2002) , p. 222. - trans. )
21 6 NOTES TO PAGES 7- 1 1
Chapter I Greek Number and Modern Number
1 Consi der, for exampl e, the defni ti on of number in Eucl i d' s Elements
( Book VII, defnition 2) : 'ApleI6 EO'lV 'o h: 1OvuBcv OUYEtIEVOV
1Aieo' . We might transl ate i t as fol l ows: ' A number is a multi tude
composed of uni ti es. ' The defnition of number i s secondary, being
dependent upon that of uni ty. But what does defnition 1 , that of
unity, say? Movu EOTLv, Ka8'v EKaOTOV TUN OVTCV EV AEYETOl:
' Uni ty is that by virtue of whi ch each being i s said to be one. ' We
can see i mmedi atel y what ontol ogi cal substructure is presupposed
by the mathematical defnition of number: that the One can be said
of a being i n so far as it is.
2 [ trait: see ch. 2n 3. - trans. ]
3 [ Conway, J. H. , On Numbers and Games ( London Mathe
mati cal Society Monographs no. 6, London: Academic Press, 1 976 ) .
- trans. ]
4 [ Knuth, D. E. , Surreal Numbers ( Readi ng, MA: Addi son-Wesley,
1 974) . - trans. ]
5 [ Gonshor, H. , An Introduction to the Theory of Surreal Numbers
( London Mathematical Society Lecture Note Series, 1 1 0, Cam
bridge: Cambri dge Uni versity Press, 1 986) . - trans. ]
6 [ Since the l anguage of ' whole' and ' natural ' numbers is i nformal
and not al ways appl i ed consi stently, it is worthwhi l e to set out the
usage of the present work, al ong with the formal mathematical
equivalents:

whole numbers: 0, 1 , 2, 3 . . . ( the non-negative integers, Z* ) .

natural whole numbers: 1 , 2, 3 . . . ( the posi ti ve i ntegers, Z+) .

'relative ' whole numbers: . . . -3, -2, -1 , 0, 1 , 2, 3 . . . ( the i nte
gers, Z) . - trans. ]
7 On the di alectic - consti tuti ve of material i st thought - between
algebraic and topological ori entati ons, the reader is referred to my
Theorie du Sujet ( Pari s: Seui l , 1 982) , pp. 231 -49.
8 [ See Bourbaki , N. , Elements de mathematique, Livre 1: Theorie des
ensembles ( Pari s: Hermann, 1 954) ; Engl ish edn Elements of Math
ematics, Vol I: Theory of Sets ( Readi ng, MA: Addi son-Wesley,
1 96 8 ) . - trans. ]
9 The theme of the cut, in its concept and its techni que, is treated i n
chapter 15 of thi s book.
1 0 [ See Dedeki nd, Numbers, 73. - trans. ]
1 1 [ Frege, Gottlob, Die Grundlagen der Arithmetik: Eine logisch
matematische Untersuchung uber den Begriff der Eahl ( Bresl an,
1 8 84) ; The Foundations of Arithmetic, Engl i sh transl ation by J. L.
Austi n (2nd revised edn, Oxford: Blackwel l , 1 974) . References given
NOTES TO PAGES 1 1 - 1 6 2 1 7
below are to the numbered secti ons of Frege' s text. The frst German
edn appeared i n 1 884. - trans. ]
12 [ See Frege, Foundations, 28-29. - trans. ]
13 For a particul arl y bri ef i ntroduction to t he di fferent types of numbers
whi ch are used i n modern anal ysi s, refer for exampl e to J. Dieudon
ne' s Elements d'analyse, I: Fondements de [ 'analyse moderne ( Pari s:
Gauthi er-Vi l l ars, 3rd edn, 1 98 1 ) , chs 1 -4.
14 [unique nombre qui ne peut pas hre un autre: From Mal l arme' s
' Un Coup de des j amai s n' abol i ra I e hasard' , translated in E. H.
Blackmore and A. M. Blackmore, Collected Poems and Other Verse
( Oxford: Oxford Uni versi ty Press, 2006) , pp. 1 6 1-8 1 as 'A di ce
throw at any ti me never wi l l abol i sh chance' ( translati on modi fed) .
- trans. ]
1 5 Natacha Michel proposes the di sti nction between ' frst moderni ty'
and ' second moderni ty' i n L'Instant persuasif du roman ( Pari s: Les
Conferences du Perroquet, 1 987) .
1 6 Dedeki nd, Numbers, 64.
1 7 I give a detai led commentary on the Hegel i an concept of number - a
posi ti ve vi rtue of which is that, according to i t, the i nfni te is the
truth of the pure presence of the fni te - i n medi tati on 1 5 of L'Etre
et l'evenement ( Pari s: Seui l , 1 988) , pp. 1 8 1 -90 [ pp. 1 6 1-70 i n
Ol iver Feltham' s Engl i sh transl ation Being and Event ( London:
Conti nuum, 2005) . - trans. ]
1 8 [ See Frege, Foundations, 84- 86. - trans. ]
1 9 [ Dedeki nd, Numbers, 2. - trans. ]
20 [ Frege, Foundations, 74. - trans. ]
21 [Dedeki nd, Numbers, 73. - trans. ]
22 [ Tout. - trans. ]
23 [ Dedeki nd, Numbers, 66. - trans. ]
Chapter 2 Frege
1 The key text for Frege' s concepti on of number is The Foundations
of Arithmetic [on which see above, ch. I n 1 1 - trans. ] . The funda
mental argument, extremely dense, occupi es paragraphs 55 to 86
( less than thi rty pages i n the ci ted edi tion) . We must sal ute Cl aude
Imbert's excel lent work, i n particular her lengthy i ntroducti on.
[ Badi ou refers to Imbert' s transl ati on Les Fondements de
l'arithmhique ( Pari s: Seui l , 1 969) . - trans. ]
2 [ toute pensee emet un coup de des: Mal larme, ' Coup de des' , p. 1 8 1 .
- trans. ]
21 8 NOTES TO PAGES 1 6-23
3 [ See Frege, Foundations, 46-53. Badiou renders Frege' s Eigen
schaft as trait: Al though Austin has ' property, ' I have used ' trait'
here, so as not to lose the di sti nction present in Badi ou' s text
between trait and propriete. - trans. ]
4 [ In Austi n' s Engl i sh transl ati on, Frege' s term Gleichzablig is ren
dered as ' equal ' , but see Austi n' s note at 67 on possibl e al terna
tives: I fol l ow both Austi n' s advice and Badi ou' s use of equinumericite
in employing equinumerate as the term whi ch avoids at once i mpre
cision and ugly neologism. - trans. ]
5 [ Frege, Foundations, 68 . - trans. ]
6 [ I bi d. , 74. - trans. ]
7 [ Ibi d. , 77. - trans . ]
8 [ Ibi d. - trans. ]
9 [ Ibi d. , 74. - trans. ]
1 1 [ Ibi d. - trans. ]
1 2 The letter ( written i n German) i n whi ch Russel l makes known to
Frege the paradox that woul d take the name of its author is repro
duced in English translation in From Frege to Godel, a col l ection
of texts edited by J. van Heij enoort ( Cambri dge, MA: Harvard
University Press, 4th edn, 1 98 1 ) p. 1 24. Russell concl udes with an
i nformal di stinction between 'col l ection' [ or ' set' , German Menge
trans. ] and 'total i ty' : ' From thi s [ the paradox] , I conclude that under
certai n circumstances a defnable collection [ Mellge] does not form
a total i ty. '
1 3 Zermel o develops hi s set-theoretical axi omati c, i ncl uding the
axi om of separati on, whi ch remedies Russel l ' s paradox, in a 1 908
text written i n German. It can be found i n Engl i sh translation i n
van Heij enoort' s collection, cited i n t he preceding note. It comes
from Investigations in the Foundations of Set Theory, and espe
ci al l y its frst part, ' Fundamental Defnitions and Axioms' , pp.
20 1 -6.
1 4 [ Frege, Foundations, 58. - trans. ]
1 5 The subordination of the exi stenti al quantifer to the uni versal quan
ti fer means that, given a property P, i f every possibl e x possesses
this property then there exists an x which possesses i t. In the predi
cate calcul us: 'x( P( x) ) 3x( P( x) ) . The classical rules and axioms
of predicate cal cul us permit one to deduce this i mpl ication. Cf. for
exampl e E. Mendelson' s manual Introduction to Mathematical
Logic ( NY: Van Nostrand, 1 964) , pp. 70-1 .
1 6 [TO yag alho voEiv taTLv TE Kat ElvUL, from Parmenides' poem.
- trans. ]
NOTES TO PAGES 24-3 0 21 9
Chapter 3 Additional Note on
a Contemporary Usage of Frege
1 Mi l l er' s text appears in Cahiers pour l 'analyse, no 1 ( Pari s: Seui l ,
February 1 966) , pp. 37-49 [ transl ated by Jacquel i ne Rose as
' Suture ( Elements of the Logic of the Si gni fer) ' i n Screen, 1 8: 4
( 1 978 ) , pp. 24-34. - trans. ] . One ought to read al ong with it Y.
Duroux' s arti cl e 'Psychologie et l ogi que' appearing i n the same
i ssue ( pp. 3 1 -6) , whi ch exami nes i n detai l the successor functi on in
Frege.
2 Cf. A. Badiou, ' Marque et manque: A propos du Zero' , in Cahiers
pour l 'analyse, no 1 0 ( Pari s: Seui l , 1 969) , pp. 1 50-73.
3 U'y suis, j 'y suis toujours. From Ri mbaud' s 1 872 poem ' Qu'est-ce
pour nous, mon crur, que les nappes de sang' [ transl ated in Col
lected Poems, ed. and trans. Ol i ver Bernard ( London: Pengui n,
1 986) , pp. 202-3. - trans. ] .
4 [ Mi l ler, ' Suture' , p. 40. - trans. ]
5 [ meconnue. - trans. ]
6 [ Mi l ler, ' Suture' , p. 40. Transl ation modi fed. - trans. ]
8 [ See Frege, Foundations, 26-27. - trans. ]
9 [ Mi l l er, ' Suture' , p. 44. - trans. ]
1 0 [ Ibi d. p. 46. - trans. ]
1 1 [ Ibi d. p. 47. Transl ati on modi fed. - trans. ]
1 2 [ Ibi d. p . 43. - trans. ]
1 3 [ I 'instance de La Lettre. - trans. ]
14 [ I bi d. p. 44. - trans. ]
1 5 On the typology of orientations i n thought, d. Medi tation 2 7 of
L'Ptre et / 'evenement, pp. 3 1 1 -1 5 [pp. 28 1 -5 i n the Engl i sh transla
ti on. - trans. ] .
1 6 [ Mi l l er, ' Suture' , p. 40. Transl ation modifed. - trans. ]
1 7 [ Ibi d. p . 4 1 . - trans. ]
1 8 [ Ibid. p . 47. Translation modi fed. - trans. ]
1 9 [ ' Matrice' , i n Ornicar? 4 ( 1 975) ; transl ated by Dani el G. Col l i ns i n
Lacanian Ink 1 2 ( Fal l , 1 997) : pp. 45-5 1 . - trans. ]
20 [ Mi l ler, ' Suture,' p. 39. - trans. ]
21 [ fourmillement: i f not for i ts unfami l i ari ty, the more di rect
etymologi cal equi valent of the psychi atri c term formication,
designating a pri ckl ing or ti ngl i ng as of ant s crawl i ng over the
220 NOTES TO PAGES 3 0-4 1
ski n, might carry less of an inappropriate sense of dynamic
sel f-organisation than ' swarming' : rather than i mplying any
vital movement, Badi ou' s fourmillement seems to denote 'our'
phenomenological regi stration of the icy 'constel l ati ons' of Number.
- trans. )
22 For exampl e E. Borel , ' La Philosophie mathemati que et I ' i nfni ' ,
Revue du mois, 14 ( 1 91 2) , pp. 21 9-27.
Chapter 4 Dedekind
1 The reference text for Dedeki nd' s doctrine of number is The Nature
and Meaning of Numbers [ see ch. On 1 above - trans. ) The frst
German edi tion was publ i shed in 1 8 8 8 .
2 [ Dedeki nd, Numbers, 1 . - trans. )
3 [ Ibi d. , 2. - trans. )
4 [ I bi d. , 2 1 -25. - trans. )
5 [ Ibi d. , 26-35. - trans. )
6 [ Ibi d. 7 1 . - trans. )
7 [ Ibi d. 73. Dedeki nd' s text has < where Badiou uses f - trans. ]
8 We mi ght say that Frege i s a Lei bni zi an, Peano a Kanti an, and
Cantor a Platonici an.
The greatest logician of our ti mes, Kurt G6del, considered that
the three most i mportant phi l osophers were Plato, Lei bniz and
Husserl - this last, i f one might say so, holding the pl ace of
Kant.
The three great questions posed by mathematics were thus:
1 the reality of the pure i ntel l igible, the being of that which math
ematics thi nks ( Pl ato) ;
2 t he development of a wel l -formed language, t he certitude of
inference, the laws of cal cul ati on ( Lei bni z) ;
3 the constitution of sense, the universal ity of statements ( Kant,
Husserl ) .
9 [ See Dedeki nd, Numbers, Preface to the frst Edn. - trans. )
1 0 [ Dedeki nd, Numbers, 64n. - trans. )
1 1 [ Ibi d. - trans. )
1 2 [ Ibi d. , 66. Dedeki nd' s text has < where Badiou has f, and a, b
rather than S l , S2. - trans. ]
1 3 [ a - al so ' i d' . - trans. )
NOTES TO PAGES 46-5 2 22 1
Chapter 5 Peano
The reference text for Pea no i s a text publ i shed i n Latin in 1 889,
whose Engl i sh ti tl e i s: 'The Pri nci ples of Ari thmetic' . The Engl i sh
translation of thi s text i s found i n J. van Heij enoort ( ed. ) , From
Frege to Gadel, pp. 83-97.
2 [ Ibi d. , p. 85. - trans. )
3 [ Ibi d. - trans. )
4 This passage is taken from a letter from Dedeki nd to Keferstei n,
dating from 1 890. The Engl i sh transl ation can be found in van
Heij enoort ( ed. ) , From Frege to Gadel, pp. 98-1 03.
5 [ Van Heij enoort ( ed. ) , From Frege to Gadel, p. 85. - trans. )
6 [Title of Husserl ' s 1 9 1 1 ' mani festo' ; translated in Q. Lauer ( ed. ) ,
Phenomenology and the Crisis of Philosophy ( New York: Harper,
1 9 1 0) . - trans. )
7 [Van Heij enoort ( ed. ) , From Frege to Gadel, p. 85. - trans. )
8 [ Ibi d. , p. 85. - trans. )
9 [ Ibid. , p. 94. - trans. ]
10 [ Ibid. - trans. )
1 1 [ Ibi d. ( Axiom 6) . - trans. )
12 [froide d'oubli et desuetude, une Constellation: Mal l arme, 'Coup de
des' , p. 1 8 1 . - trans. ]
13 Regarding these questions, one mi ght read the ( purel y historical )
chapter 1 0 of Robi nson, A. Non-Standard Analysis ( Amsterdam:
North-Hol l and, revi sed edn 1 974) . Robi nson recogni ses that
' Skolem' s work on non-standard models of Arithmetic was the great
est si ngle factor i n the creation of Non-Standard Anal ysi s' (p. 278 ) .
For a phil osophical commentary on these developments, cf. A.
Badiou, ' Infni tesimal Subversion' , i n Cahiers pour l'analyse, no 9
( Pari s: Seui l , 1 968 ) pp. 1 1 8-37.
14 [ (a - trans. )
Chapter 6 Cantor:
'Well-Orderedness' and the Ordi nals
1 Cantor's clearest articulation of hi s ordinal conception of numbers
is found i n an 1 899 letter to Dedeki nd. See the Engl i sh translation
of the key passages of this letter i n van Hei j enoort ( ed. ) , From Frege
to Gadel, pp. 1 1 3-1 7. Cantor demonstrates an excepti onal l uci di ty
as to the phi losophical ly crucial di stinction between consistent
222 NOTES TO PAGES 5 2-70
mul ti pl icities and i nconsistent mul tipl iciti es. It i s to hi m, i n fact, that
we owe this termi nol ogy.
2 On thi s poi nt, you are natural l y referred to the work of Alexander
Koyre.
Chapter 7 Transitive Multipl icities
1 [prenez ensemble - wi th the i ntended resonance of ensemble ( ' set' ) .
- trans. ]
2 [ decoupe: a 'carving out' or deducti on. - trans. ]
Chapter 8 Von Neumann Ordi nals
1 John von Neumann gave a defni ti on of ordi nal s i ndependent of the
concept of well-orderedness for the frst time in a 1 923 German
article, entitl ed ' On the i ntroduction of transfni te numbers' . Thi s
article is reproduced in Engl i sh transl ati on in van Hei j enoort ( ed. ) ,
From Frege t o Gadel, pp. 346-54.
The defni ti on of ordi nal s on the basi s of transi ti ve sets seems to
have been taken up agai n in an article in Engl i sh publ i shed in 1 937
by Raphael M. Robi nson, entitled 'The theory of classes, a modifca
ti on of von Neumann' s system' ( journal of Symbolic Logic, no 2,
pp. 29-36) .
2 Throughout thi s book, the ordi nal s, denoted in current l i terature by
the Greek letters, wi l l be denoted by the letters W and w, supple
mented further on with numerical i ndices, W h or W3 , etc. In general ,
W or N desi gnate a variable ordi nal ( any ordi nal whatever) . In
parti cul ar, we empl oy the expression ' for every ordi nal W' . The
notati on with i ndices i s used to designate a particul ar ordi nal , as i n
the expression ' take ordi nal W, whi ch i s the matter of Number N, ' .
The subscripts wi l l be used most often to the l eft of the si gn E to
desi gnate an ordi nal whi ch is an el ement of another, as w, E W
( ordi nal WI is an element of ordinal W) .
3 [ /a Nature ( as opposed to nature) . - trans. ]
4 The Axiom of Foundati on, al so cal l ed the Axiom of Regul ari ty, was
anticipated by Mi ri manoff i n 1 9 1 7, and ful ly clari fed by von
Neumann i n 1 925. To begin with, it was a matter, above all, of
el i mi nati ng what Mi ri manoff cal l ed 'extraordinary sets' , that i s,
those whi ch are elements of themselves or contai n an i nfni te chai n
of the type . . . E an+ , E an E E az E al E E. It was real i sed onl y
NOTES TO PAGES 70-92 223
later that thi s axi om enabl ed a hi erarchical presentation of the uni
verse of sets.
For a hi storical and conceptual commentary on this axiom, cf.
A Fraenkel , Y. Bar-Hi l l el and A. Levy, Foundations of Set Theory
(Amsterdam: North-Hol l and, 2nd edn 1 973) , pp. 86-1 02.
For a phi l osophical commentary, see Medi tation 1 8 of L'Ptre
et / 'evenement, pp. 205-1 1 [ pp. 1 84-90 in the Engl i sh transl ati on.
- trans . ] .
5 A good presentation of the fact that bel onging ( e ) orders the
ordi nal s total l y ( strict order) - in other words that, given two
di fferent ordinals WI and W2, either WI e W2 or W2 e WI -
can be found in Shoenfel d, J. R. , Mathematical Logic ( Readi ng,
MA: Addi son-Wesley, 1 967) , pp. 246-7. This proof is reproduced
and commented upon in L'Ptre et l 'evenement i n the third section
of Meditation 1 2, pp. 1 53-8 [pp. 1 34-9 i n the Engl i sh translation.
- trans. ] .
6 [ decouper. - trans. ]
Chapter 9 Successi on and Li mit. The I nfi nite
1 Badiou, A. , Manifeste pour la philosoph ie, Paris: Seui l , 1 98 9 [ trans
l ated by N. Madarasz as Manifesto for Philosophy ( Al bany, NY:
State Universi ty of New York Press, 1 999) . - trans . ) . The ci rcum
stances and the effects of the phi l osophy's suture to the poem,
beginning wi th Ni etzsche and Hei degger, are described briefy in
chapter VII, entitled 'The Age of Poets' .
2 [ Osi p Mandelstam, from hi s Tristia ( 1 922) . Badi ou quotes Tatiana
Roy's French transl ati on: vers ces prairies infnies 01 Ie temps
s 'arrete. - trans . ] . The i nstant of Presence is beyond all i nsistence,
al l successi on. The 'eternal midday' i s the trans-temporal limit of
time. Here i s the conj oint site of the poem and the sacred.
It is not al ways i n thi s place, i t must be sai d, that Mandel stams' s
poems establ i sh themselves. For in hi s most powerful poetry he
seeks to thi nk the century, and succeeds i n doi ng so.
Chapter 1 0 Recurrence, or I nduction
1 For the demonstration of the val i di ty of defni ti ons by i nduction,
you are referred to K. J. Devl i n' s Fundamentals of Contemporary
Set Theory ( NY: Spri nger-Verlag, 1 980) , pp. 65-70 ( 'The principle
of recursi on' ) .
224 NOTES TO PAGES 1 07
Chapter 1 2 The Concept of Number:
An Evental Nomi nation
1 To repeat, the basic text for the study of the numbers cal l ed ' surre
al s' i s Gonshor' s Introduction to the Theory of Surreal Numbers
[ see ch. 1 n 5 trans . ] . The fact that Gonshor and al l current theorists
of these numbers, which I cal l Numbers, see them as a ' macro
fel d' of the real s results in a presentation quite di fferent from my
own.
The i nital idea of thei r creator, Conway, was to defne ' surreal '
numbers di rectly by means of cuts. A number wi l l be defned as a
pai r of two sets of numbers, conforming to the conditions of the
cut ( every number in the set ' to the right' in the pair i s smaller than
every number in the set 'to the l eft ' ) . The doubl e circularity of this
defnition obviously must be questioned ( number i s defned on the
basis of number, and i nequal i ty between numhers i s mentioned
without havi ng been properly defned) . The operation that serves to
undo thi s circle is obviously transfni te induction, which makes
ordi nal s appear i nevi tabl y on the scene. I n fact, Conway presents
Numbers on the basis of their canonical representation - that i s, i n
my l anguage, thei r ' structural ' character: they are defned on the
basis of their sub-Numbers. D. E. Knuth' s book Surreal Numbers
( Readi ng, MA: Addison-Wesley, 1 974) gives a 'pedagogical ' version
of Conway's presentation i n the form of a dialogue. It seeks to rec
reate the mental ity of a ' researcher' into the matter, but i n fact
becomes quite convol uted, si nce in its exposition the employment
of the ordi nal series i s not made explicit. Besides thi s, it re
establ i shes, to my mi nd to the detri ment of the real 'geni us' of the
invention of Numbers, a creationist and progressive logic ( frst
'creating' zero, then 1 and -1 , etc. )
Gonshor starts from a l iteral ' coding' , whereas, i n my quest for
the concept and its phi losophical deployment, I j oi n a set-theoretical
l i neage. Technical l y speaki ng, Gonshor general i ses the development
i n base 2 of the real numbers. A real number can he presented as
an infnite denumerabl e series of signs 1 and O. Gonshor' s idea is to
consider such series of any ordinal length whatsoever, rather than
l i mi ting them to denumerahle series. He then begins with two signs
+ and -, and calls ' surreal number of l ength W' a series of such signs
i ndexed to the el ements of the ordi nal W. The index ordi nal s affected
by the sign + correspond to the elements of what I call the form of
the Number, and the index ordi nal s affected by the sign -, to the
elements of the resi due. The ordi nal ' l ength' corresponds to what I
cal l the matter of the Number.
As an example: the Number which I write ( 4, ( 0, 3) ) , whose matter
i s 4 and whose form contai ns the elements 0 and 3, i s written by
Gunshor as follows: + - - +.
NOTES TO PAGES 1 07- 1 3 0 225
Now, of course surreal numbers and Numbers are ' the same
thing' . But we might say that Gonshor treats them as i nscriptions,
or marki ngs, after the manner of Frege and of Peano' s ari thmetic.
The inspiration here i s i deographi c. Whereas I approach them from
the poi nt of view of their mul tiple-being, in the Cantorian spirit,
my inspiration being ontological , or Platonic.
In fact even the technical development ends up being quite di f
ferent, although the results can al ways be transl ated from one
version to the other. For example, i t is not i nsi gni fcant that
Gonshor, who, with the signs + and -, is unabl e to denote an
occurence of the voi d, must i nvoke an ' empty seri es' of signs,
where I woul d wri te ( 0,0) . The conceptual advantage of the
ontological approach to Number is that it al lows one to di spense
with al l additional l i teral i sation, with every heterogeneous sign,
in favour of the two fundamental set-theoretical relations of
belonging E and incl usion c. This doubtless expl ai ns why for
Gonshor the theory of surreal numbers is a sort of speci al i st
technique, whereas for me it is a wholly natural extension of the
ontological vocation of set theory to the concept of Number.
2 Gonshor, Introduction, p. 43.
3 [ decoupe. trans. ]
Chapter 1 3 Diference and Order of Numbers
1 Throughout thi s book, I call a relation ( most often one of order)
'total ' when two different basi c terms of the rel ation are al ways
bound by thi s rel ati on. Thus I woul d say that the relation E i s total
in the ordinals or that the relation E i s total in the Numbers.
Someti mes a rel ation is cal l ed ' total ' which is al so refexive,
binding each term to itsel f. This i s the case, for exampl e, with the
relation :: ( less than or equal to) for the natural whole numbers.
Limiting oneself to my defnition, which onl y demands the rela
tion between di fferent terms, and excl udes the relation of sel f with
sel f ( an i rrefexive relation, then) , i s more convenient i n the case we
are deal i ng wi th. Where we speak of an order-relation, we mean to
say that its axioms are those of strict order.
2 For Gonshor, order is easily presented as lexicographical , since surreal
numbers are introduced as series of signs + and
-
( d. ch. 1 2n 1 ) .
3 On this point, d. Mi l l er, J. - c. Libertes, Lettre, Matiere ( Pari s: Con
ference du Perroquet, June 1 985) .
4 [ Paul Celan, from Zeitgehoft ( 1 976) . [ Badi ou' s reference i s M.
Broda' s translation ( Pari s: Cl i vages, 1 985) : chiquenaude / dans
l'abime, dans Ies / carnets de gribouillages / Ie monde se met a bruire,
il n 'en tient / qu'a toi. - trans. ]
226 NOTES TO PAGES 1 3 1 - 1 3 9
Chapter 1 4 The Concept of Sub-Number
1 Category theory is an attempt to reformul ate al l of mathematics
within a structural , non-set-theoretical framework whose starting
point is 'obj ects' , which are 'types of structures' ; and ' arrows' ,
whi ch are transformati ons, or morphi sms, between structures. The
concept of substructure can be understood i n terms of that of sub
obj ect. A ' sub-obj ect' i s i n fact an equi val ence class for certain
arrows. Cf. for exampl e J. L. Bel l ' s book Toposes and Local Set
Theories ( Oxford: Cl arendon Press, 1 988) , i n particular the argu
ments of pp. 49-58.
2 [sectionne. - trans. )
3 [ ' a la mati ere pres' : In mathematics, 'a la x pres' - Engl i sh equi valent
' up to x
'
- i ndicates that abstraction i s to be made from a certain
cl ass of obj ects, which for the purposes of a particul ar statement or
defniti on are to be regarded as a si ngle entity. Thus a certain propo
sition can be said to be true, or a property to be satisfed, ' up to
i somorphi sm' , ' up to rotati on' , 'up to transl ati on' , and so on. In the
present context, the 'cut' between the high and low sets uni quely
defnes a Number, so long as we regard all possible confgurations
of ' matter' as being subsumed under the aspect of the unique minimal
case. In other words, the defni ti on of the cut must be supplemented
by the princi pl e of mini mal i ty. - trans. ]
4 [ encadrement: an i nterval in the mathemati cal sense, as in ' interval
around a real number' . - trans. ]
Chapter 1 5 Cuts: The Fundamental Theorem
1 The problem of the cardi nal ity of the set of parts of an infnite set i s
a central probl em for set theory after Cantor. The ' mi ni mal ' hypoth
esis, whi ch says that this cardinal i ty i s the smallest cardi nal l arger
than that of the i ni ti al set - the cardi nal successor of the one which
measures the quanti ty of that set - i s the famous 'conti nuum hypoth
esi s' , denoted by CH i n the Engl i sh l i terature on the subj ect.
Following P. H. Cohen' s work, we know that the conti nuum
hypothesi s i s undeci dabl e on the basi s of the cl assi cal axi oms of the
theory. It can be affrmed or denied wi thout any contradiction bei ng
i ntroduced.
A particul arly l uci d text on thi s probl em i s K. Godel ' s ' What i s
Cantor' s conti nuum problem? ' . The Engl i sh text has often been
republ i shed si nce its frst appearance i n 1 947; for example in P.
Benacerraf and H. Putman ( eds) , Philosophy of Mathematics ( Cam
bri dge: Cambri dge Uni versi ty Press, 2nd edn, 1 983) pp. 470-86.
NOTES TO PAGES 1 40- 1 65 227
2 [ dans ces parages du vague OU toute rea lite se dissout: Mal l arme,
' Coup de des' , p. 1 8 1 . - trans. ]
4 The concept of the cut, and the way in which it specifes the relation
between punctual i ntervention and the conti nuum of si tuati ons,
traverses al l the procedures of truth. Its occurrence can be remarked
in the poli tics of rupture ( revoutionary pol i tics ) , i n the artistic theme
of novelty or of modernity, in the scientifc theme of crises and
refoundations, or in the amorous fgure of separation. Every fdel ity
i s al so the process of a cut.
S Dedeki nd' s fundamental text on the i dea of the cut, dati ng
from 1 872, is ' Continuity and Irrational Numbers' [ transl ated i n
Dedekind, Numbers, pp. 1 -24. - trans. ] .
6 The exposition in Gonshor, Introduction, begins with the demon
stration of the fundamental theorem. Hi s style i s very di fferent: both
because, as I have already mentioned, Gonshor adopts a line which
is oriented more towards calcul ati on than towards set theory; and
because he i s not content wi th a proof of exi stence, but i ntends to
determine exactly the Number that is cut (what i s cal l ed a 'construc
tive' proof) . This concern for determi nation entai l s the exami nation
of a great many cases.
7 We will see in chapter 1 6 that the upper bound of a set L, being of
the form ( WI , Wd, is an ordinal. This i s a stri ki ng result.
8 The lower bound of a set H is in fact the negative of an ordi nal , a
Number
-
( W) . Cf. ch. 1 6 .
9 For rule 2, the reasoning is exactly symmetrical to that which val i
dates rule 1 . Let us take rule 3: we have Id. ( W, Nb) , and W i s in
the form of Nb. I put W i n the form of Ni. Am I not ri ski ng making
it so that Ni becomes thus as l arge as a Number of A? Take Na to
be this supposed Number. W must be the di scri mi nant of Ni and
of Na, whi ch i s to say that i t i s al so the di scriminant of Na and Nb.
Now W i s in the form of Nb, one must therefore have Na < Nb,
which i s not al l owed.
The same approach can be appl ied for rule 4.
Chapter 1 6 The Numberless Enchantment of the
Pl ace of Number
1 As we have i ndicated in notes 7 and 8 of the preceding chapter, a
very interesting ' topological ' characteristic of the positive and nega
tive ordinals is that every set of Numbers has an ordi nal as its upper
bound, and the negative of an ordinal as its lower bound. This can
228 NOTES TO PAGES 1 65- 1 78
be explai ned easil y enough, si nce every ordi nal is the cut of itself
and the void, and every negati ve of an ordi nal , a cut of the void
and i tsel f.
2 Gonshor, Introduction, p. 32.
3 [ encadrement. - trans. ]
4 The pri nci ple of the i somorphi sm of orders - that i s, of the fact that,
i f N i s a Number of fni te matter and RA( N) r, then Nl < Nz
RA( Nd < RA( Nz ) - is si mpl e enough ( note that < is to the left of
the i mpl icati on in the order-relation in Numbers, to the right of the
rel ati on of ordi nary order i n the rational numbers ) . The resul t is
that, in the decomposition of N i n the form 1 + 1 + . . . - t + etc. ,
what i s added ' at the end' decreases very rapi dl y. Thi s is a quite
si mpl e, normal algebraic cal cul ati on.
5 See Gonshor, Introduction, pp. 30-1 .
6 One might obj ect at this point that our Numbers do not authorise
the representation either of complex numbers or of quaterions,
upon which physics reli es to a considerable extent.
But are complex numbers and quaternions numbers? I thi nk i t
can be reasonably maintai ned that, from the moment we take leave
of all ' l inearity' when we abandon di mensi on 1 , we are deal i ng
wi th constructions based on Numbers rather than with Numbers
per se.
Basical l y, the i nnermost essence of compl ex numbers i s geometri
cal, i t i s the ' complex pl ane' which del i vers the truth of these
' numbers' . Around the complex numbers i s organi sed the profound
l i nk between pure algebra ( the extension of fel ds) and the ontologi
cal scheme of space as topological concept. I am tempted to cal l
complex numbers operators, operators whose function i n thought
i s to articulate algebra and topology. Hence the si mul taneously
combinatori al ( a complex number being a pair of real numbers) and
geometrical character of these ' numbers' . They are in fact numbers
which do not number, but suggest schemes of representation and
i nscription which are al ready, i n effect, something very close to a
conceptual ' physics' .
Moreover, i t seems to me unreasonabl e to speak of ' numbers'
when i t i s not even possible, i n terms of the operational fel d con
si dered, to say that one ' number' i s l arger or smal l er than another.
In short: a fel d of numbers must i n my view be an ordered feld,
whi ch nei ther complex numbers nor quateri ons are.
Final l y, I restrict the concept of Number, i n so far as i t is thought
of as a form of being, to that which can be deployed according to
the intuition of a line. This i s made clear by the decisive part played
i n the defni ti on of the being of Number by that fundamental ' l i ne
of being' constituted by the ordinal s.
7 See Robi nson, Non-Standard Analysis [see ch. Sn 1 3. - trans . ] .
NOTES TO PAGES 1 83-203 229
Chapter 1 7 Natural I nterl ude
1 On the mathematical personality Ramanuj an, see the great number
theorist G. H. Hardy' s autobiographical A Mathematician 's Apology
[ Cambridge: Cambridge University Press, 1 940. - trans. ] .
2 On the set-theoretical reduction of rel ations and functions to the
pure mul ti pl e, and for an ontological di scussion of thi s point, see
L'Ptre et l 'evenement, Appendi x 2, ' A relation, or a functi on, is
solely a pure mul ti pl e' , pp. 483-6 [ pp. 443-7 i n the Engl i sh transl a
tion. - trans . ] .
3 [ La nature a lieu, on n 'y a;outera pas: Mal l arme, ' De la musi que et
des lettres' , i n H. Mondor and G. Jean-Aubry ( eds ) , Oeuvres com
pletes ( Pari s: Pl e"iade, 1 945) pp. 642-57. Transl ated as ' Music and
Literature' , in B. Cook ( trans . ) , Mallarme: Selected Prose Poems,
Essays and Letters ( Bal timore: Johns Hopki ns Press, 1 956) ,
pp. 43-56 ( translation modifed) . - trans . ] .
Chapter 1 8 Algebra of Numbers
1 It is equal l y true that every set of Numbers whose matter is lower
than or equal to a given i nfnite cardi nal i s a commutative feld. In
this regard, Gonshor is right to say that the study of the fel d of
Numbers of fni te matter or equal to ( ' of countable length' [ see
Gonshor, Introduction, p. 1 03. - trans. ] ) woul d be most worth
while. This feld al l ows real Numbers as a subset, but i t al so contai ns
i nfnitesi mal s and cuts of cuts. It woul d be possi bl e to devel op a
wholly origi nal anal ysis here.
2 Gonshor, Introduction, Ch. 3, 'The Basic Operati ons' .
3 Take two ordi nal s WI and W2, where (( (WhW2 ) W. If WI is
maximal in the couple, every couple (WhW4) where W4 E W2 i s
smal l er than the couple (W. ,W2) in the order of couples ( see 1 7. 6) ,
because they have the same Max ( which i s WI ! and the same frst
term ( which is al so WI ! , but the second term of the couple (WhW4)
is smaller than the second term of (WhW2). So, (( (WhW4) E
(( (WhW2 ) , since ( is an isomorphi sm of order between couples of
ordinals and ordinal s.
If i t is Wl that is maxi mal , the same concl usi on fol l ows, si nce the
Max of (WhW4) i s lower.
Similar verifcations can be made for any such case.
4 The induction in question consists of proving simul taneousl y:

that, if N2 < N3, then NI + N2 < NI + N3 ( compati bi l i ty of order
and addi tive structure) ;
230 NOTES TO PAGES 203-2 1 4

that the Numbers of set L are al l smal l er than the Numbers of
set H.
5 To be real l y meti cul ous, we must take i nto account the case where
L does not contai n any positive Numbers, but does contai n O. In
thi s case, 0 i s the internal maxi mum of L. One can take i t as 0
' al one' , or i denti f L wi th the set ( 0) . The reasoni ng is then much
si mpl i fed.
6 The reader might be perturbed by the constant amphi bol i es of nota
tion ( the sign < used i n one pl ace for the order of Numbers, el se
where for that of this or that particul ar type of number, etc. ) In fact,
mathematicians ( who say in such a case ' that there i s no possi bl e
equi voci ty' ) express through such amphiboli es thei r natural ten
dency to identify purely and si mply, and therefore to name i denti
cal l y, rel ations and operati ons whi ch are defned wi th isomorphic
structures. How el se coul d Category Theory have arisen, taki ng as
its ' pri mitives' not mul ti pl i ci ti es, but ' morphi sms' , or arrows, des
ignating 'correspondences' between structural ' obj ects' ?
Chapter 1 9 I n Concl usi on:
From Number to Trans-Bei ng
1 [action restreinte: Mal l arme, 'Acti on restrei nte' , in Oeuvres com
pletes ( see ch. 1 7n 2) , pp. 369-73. ]
2 [ Como urn deus, nao arrumei nem uma coisa new outra: from
A
l varo de Campos aka Fernando Pessoa' s 1 929 poem ' Reticencias' .
See F. Pessoa, edi ted by M. A. D. Gal hoz, Obra Poetica ( Rio de
Janeiro: Agui l ar, 1 960) . Thi s i s a variant of hi s ' Quasi ' ( cf. Vol . I I
of the Edirao Critica ( lmprensa Naci onal - Casa da Moeda, 1 990) ,
p. 2 1 5) , where we read Como um deus, nao arrumei nem a verdade
nem a vida ( ' Li ke a god, I arranged neither truth nor l i fe' ) . Badiou
quotes A. Gui bert's French translation: Tel un dieu, je n 'ai mis de
l'ordre ni dans l 'un ni dans l 'autre. - trans. ]
I ndex
addi ti on, 1 98-9
and i nductive defni ti on, 95, 96,
201 -8
see also cal cul ati on
adherence, 79, 87
al gebra, 1 42, 1 96, 1 97, 1 98-208
Numbers of fni te matter and
dyadi c rati onal s, 1 73-4
as operati onal di mensi on of
Number, 1 1 1
and substructure, 1 32, 1 33
al gebraic ori entati on, 1 0, 1 2, 48
Al l , 14, 87, 88-9, 1 39, 1 44, 1 93
alogos and nomi nati on, 1 06
Al thusser, Loui s, 3
anal yti c geometry, 1 1
Aristotle, 1 09
ari thmetic, Peano' s signs, 49-50,
5 1
ari thmeti c/geometry opposi ti on,
1 0-1 1 , 1 2, 1 4 1
art, 3 , 1 1 1
as truth procedure, 57
see also poetry
Axiom of the Empty Set, 44, 57
Axiom of Foundati on, 70-2
Axiom of Infnity, 57, 82, 94
Axi om of Separati on, 2 1 , 43
axi omati c approach to number, 9,
57
and deci si on, 21 2-1 3
Peano's axi oms, 8, 1 1 , 1 2, 46-5 1
base-el ement of number
( Dedeki nd) , 33, 34
beati tude, i ntel lectual state of, 1 44,
1 8 3
Bei ng and number, 9 , 1 2, 25 , 57-8,
1 1 1 , 2 1 1-1 4
and operati onal properties, 200
and ordi nal s, 82, 83
trans-bei ng, 21 4
zero, 1 57-8
see also exi stence; ontol ogy of
Number
bei ng of the One, 7, 8, 1 3, 55
bel ongi ng, 6 1 -6, 68, 69-70, 71 ,
76, 84, 93
and di fference between numbers,
1 1 6-1 7
and sets of ordi nal s, 1 1 2, 1 1 3-14
bi uni vocal correspondence
coupl es of ordi nal s, 1 93-5, 1 96,
201 -8
232 INDEX
bi uni vocal correspondence ( cont' )
and Dedeki nd, 3 1 , 32-3, 34,
36-7, 38, 39
and Frege, 1 3, 1 6-1 7
Bol zano, Berard, 1 0, 1 3
Borel ,
E
mi l e, 3 0
Bourbaki , N. , 1 0
Bural i - Forti paradox, 55-6
bureaucrati sati on of knowl edge, 2
cal cul ati on
as essence of number, 8, 1 80
and truth, 26, 27
see also addi ti on; al gebra
canoni cal presentati on of the cut,
1 57, 1 65
Cantor, Georg, 1 , 3-4, 8, 1 2, 1 3,
52-8, 69, 73, 21 2
excess of parts over el ements
theorem, 63, 65, 66
transfni te numbers, 1 06
Capi tal , 2 1 3-1 4
cardi nal s, 1 0, 1 2, 1 3, 226n
Frege, 1 1 , 1 7, 3 1
and zero, 34
Carnap, Rudol f, 48
category theory, 1 3 1
Cauchy, Augusti n Loui s, Baron,
1 78
Cel an, Paul ( poem) , 1 30
ci phers, 95
Cohen, P. H. , 1 3 9
commutati ve fel d, 1 99, 204
compl ete i nducti on, 86
compl eti on of Number and the cut,
1 42-3
compl ex numbers, 1 0, 1 1 , 1 2, 228n
and quaterni ons, 228n
concept of number, 9-1 5, 58, 69,
97, 1 0 1-1 4, 1 84, 2 1 1 -1 4
compl ex numbers and
quaterni ons, 228n
Dedeki nd, 34-5
defni ti on of Number, 1 0 1 , 1 02-
1 2, 1 33
defni ti on of operati on on
Numbers, 20 1
di fference and order of numbers,
1 1 5-30, 1 56
Frege, 1 6-23, 55
and mul ti pl e-bei ng, 1 07
Number as form of Bei ng, 2 1 1
sub-number concept, 1 3 1-8
and termi nology, 1 05-7
uni fcati on of number and
operati on, 200
constructi vist orientation of
thought, 27, 30
conti nuous and di screte, di al ectic
of, 1 41
conti nuum hypothesi s, 1 0, 226n
and the cut, 1 4 1 -2
Conway, J. H. , 9, 1 07
counti ng, 9, 55, 58, 1 84
al gebra, 1 98-208
countabl e and defni ti on of
number, 1 09, 1 1 1
and empty sets, 64
and Peano' s axi oms, 5 1
and pol i ti cal val ue, 1-2
and soci ety, 3
coupl es of ordi nal s and operati on,
1 84-96, 201
cul tural representati on, 2-3
cuts and cutti ng, 1 3 9-55
addi ti on, 20 1 -8
construction of a cut, 147-53,
1 75
cut of cuts, 1 79-80, 229n
Dedeki nd' s theory, 1 1 , 1 06, 1 74-
5, 1 76-7, 1 78
and dyadi c rati onal s, 1 74-5,
1 76-7
and fundamental theorem of
ontol ogy of Number, 1 43-55,
1 56, 20 1 , 203
and operati on, 201
pri nci pl e of the cut, 1 43
sub-Numbers, 1 36, 1 38
surreal numbers, 224n
deci si on, 2 1 2-1 3
decomposi ti on and structure, 1 3 1 ,
1 33
Dedeki nd, R. , 1 , 3-4, 8, 1 2, 1 3-14,
3 1 -45, 2 1 2
the cut, 1 1 , 1 06, 1 4 1 , 1 42, 1 74-
5, 1 76-7, 1 78
letter to Keferstei n, 47-8
and Pea no, 46-7, 50
system of t he i nfni te, 32, 35-44,
47, 55, 56, 73, 85-6, 93
deducti on, 47, 2 1 2
defni tion of addi ti on, 96
defni ti on of Number, 1 01 , 1 02-1 2,
1 33, 228n
Del euze, Gi l l es, 49
denomi nators, 1 67, 1 69
densi ty of order
and the cut, 140, 1 4 1-2, 1 42-3,
1 55, 1 56
i nfni tesi mal numbers, 1 79
see also swarmi ng of numbers
Descartes, Rene, Cogito, 3 8-9,
4 1 -2
di fference between matter and form
of Number, 1 1 1 , 1 1 5-30
and order over numbers, 1 1 9-
30, 1 56
di screte magni tude, 1 41
di scri mi nant, 1 1 7, 1 56
di scri mi nati on between numbers,
1 1 5-30
and cutti ng, 148, 1 50-2, 1 54,
1 56, 1 80
and sub-Number concept, 1 35-7
Zero, 1 58-9
di sorder and restrai ned acti on, 21 4
di ssemi nati on, 79, 80-1 , 85
dyadic positive rati onal numbers,
1 67-77
economy, 3
'economy of number' , 48
elements of the el ements of a set,
79-80
empty extensi ons, 1 8-1 9
empty sets, 22, 55, 63-4, 83, 1 1 3,
1 5 7-8
and di fference between numbers,
1 1 6-1 7
INDEX 233
and successi on, 77, 84-5, 95
and von Neumann ordi nal s,
67-8, 69, 71
empty sets axi om, 44, 57
equal s si gn and Peano, 49-50
equi numerate concepts, 1 6-1 7, 1 8,
20, 3 1
Eucl i d, Elements, 7, 36, 1 06
evental si tes, 1 07
evental trans-bei ng, 21 4
events and truth process, 27, 1 55,
2 1 4
Evi l , 1 60-1
excess of parts over el ements, 63,
65, 66, 1 3 9, 140
exi stence
and fundamental theorem of
ontol ogy of Number, 1 46-53
of i nfni te, 37-44, 47-8, 55, 56,
93
of l i mi t ordi nal s, 82
of zero, 22-3, 56-7
see also Bei ng and number;
ontol ogy of Number
extensi on of a concept, 1 6-1 7
Russel l ' s paradox, 20, 2 1
zero, 1 8, 25
fni te
Dedeki nd and Frege, 1 4, 30, 44,
85-6
Dedeki nd' s defni ti on, 36
fni te group theory, 1 3 1
fni te matter see Numbers of fni te
matter
fni te number, 32, 45, 97, 1 66
frst l i mi t ordi nal ( <) , 94-5
' frst moderni ty' of thi nki ng of
number, 1 3
form of Number, 1 02, 1 03-5, 1 08,
1 1 0, 1 56
and the cut, 1 47-53, 1 79, 1 8 0
and di fference, 1 1 5-30
and ordi nal s, 1 1 4
and resi due, 1 1 1 , 1 1 2-1 3, 1 65
and sub-Number concept,
1 33-8
234 INDEX
form of Number ( cont' )
and symmetri cal counterpart,
1 6 1-3, 1 65, 1 74
formal i st approach to number, 8, 9,
1 1 , 1 4
fracti ons and dyadi c rati ona l
numbers, 1 67, 1 69, 1 71
Frege, Gottl ob, 1 , 3-, 8, 1 3-14,
1 6-30, 44, 2 1 2
' cardi nal number' concept, 1 1 ,
1 7, 3 1
and Dedeki nd, 32, 35, 40
Mi l l er' s contemporary usage,
24-30
functi ons, Dedeki nd, 32-3
fundamental theorem of the
ontol ogy of Number, 1 43-55
and operati on, 201 , 203
Gal i leo, 3 6
generi c ori entati on of thought,
27
geometry
anal yti c geometry, 1 1
ari thmetic/geometry opposi ti on,
1 0-1 1 , 12, 1 41
and Greek thought, 1 0
Godel , Kurt, 8 , 1 0, 1 39, 220n
Gonshor, Harry, 9, 1 07, 1 70-1 ,
224-5n, 227n, 229n
Greek numbers, 1 0, 93, 97
cri si s of number, 1 06
Greek thi nkers and number, 7, 1 0,
56, 57
Hegel , G. W. F. , 1 9, 95, 1 41 , 1 62
Hei degger, Marti n, 8 1
hi gh sub-Numbers, 1 35-8, 1 45,
1 57, 1 60, 1 65, 1 66, 1 76,
202
Hi l bert, Davi d, 8
hi story, 2
hi story of mathemati cs, 2 1 2
homogenei ty and mul ti pl i ci ty of
ordi nal s, 68-9, 72, 80
human sci ences, 2
Husserl , Edmund, 220n
i dea of an i dea, Dedeki nd and
Spi noza, 38, 39, 40-1
i deal s theory, 1 32
Imbert, Cl aude, 2 1 7n
i mmensi ty of Numbers, 1 02,
1 07-8, 1 1 2, 1 77-8, 200, 2 1 1 ,
2 1 2
i ncl usi on and transi ti ve sets, 61 -6,
68, 1 1 3-1 4
i nconsi stent mul ti pl i ci ti es, 89, 1 44,
1 93, 1 99, 2 1 2
i nduction
Pea no, 47
reasoni ng by recurrence, 86-92,
95, 96, 1 69
see also i nductive defni ti on;
transfni te i nducti on
i nductive defnition, 89-92, 96-7,
1 69-70
and addi ti on, 95, 96, 201 -8
and coupl es of ordi nal s, 1 93-5,
1 96, 201 -8
i nfni te, 7, 1 3-1 5, 2 1 2
Cantor's i ntroducti on o f 0 53-,
73
Dedeki nd and Frege, 14
Dedeki nd' s defni ti on, 36
Dedeki nd' s system of number,
32, 35-44, 47, 55, 56, 73,
85-6, 93
and natural whole numbers, 93, 97
and Peano' s axi om, 47, 5 1 , 57
successi on and l i mi t, 77-82, 84,
85-6, 94, 95, 1 8 8-96
i nfnitesima l numbers, 1 08, 1 09,
1 78-9, 229n
i nterval l ic Number, 148, 1 52
i ntui ti oni st approach to number,
29, 30, 1 62
i rrati onal numbers
Dedeki nd' s cut, 1 4 1
a s desi gnati on, 1 06
i rrefexi vi ty, 1 2 1 , 1 23-4
i somorphi sm
and al gebra, 208
and dyadi c rati onal numbers,
1 69, 1 73-4
fel d of the rea I s, 200
and wel l -ordered sets, 54-5, 1 96
i teration of number, 29-30
Kant, Immanuel , 35, 1 41
Knuth, D. E. , 9 , 224n
Lacan, Jacques
si gni fer and the l etter, 1 3 0
theory of the subj ect, 24, 25 , 28 ,
29, 40
l ack and function of zero, 26, 28,
34
l anguage
and i nductive defni ti on, 8 9
and Pea no, 46, 48-9
termi nol ogy of Number, 1 05-7
law of Capi tal , 2 1 3, 21 4
Lei bni z, Gottfried Wi l hel m, 26-7,
30, 35
Pri nci pl e of Identi ty, 1 8-1 9, 25
l exi cographi cal order, 1 3 0
l i mi t ordi nal s, 77-8, 78-9, 8 0, 8 1-
2, 84, 85, 2 1 1
coupl es of ordi nal s, 1 8 8-96
and i nductive defni ti on, 90, 92
and maxi mal el ement, 1 1 3
and natural whol e numbers,
93-7
and reasoni ng by recurrence, 87,
95
and whol e ordi nal part of a
Number, 1 72
' l i ngui stic turn' in phi l osophy, 48
logic of the si gni fer, 24-5, 29, 30
logicist approach t o number, 8, 9,
1 1 , 1 6, 1 8-1 9, 22
and ordi nal s, 52
Peano' s axi oms, 46-5 1
see also Frege
love, as truth procedure, 57
l ove and the negati ve, 1 61
l ow sub-Numbers, 1 35-8, 1 57,
1 60, 1 65, 1 66, 1 76, 202
lower bound of a set of Numbers,
145-6
Lyotard, ]. -F. , 48
INDEX 235
Mal l arme, Stephane, 13, 1 6, 1 96,
1 97, 2 1 3
Mandel stam, Osi p, 223n
Marx, Karl , 3
mathemati cs and Number, 2 1 2,
2 1 3
matter o f Number, 1 02, 1 03-5,
1 08, 1 1 0, 1 1 1 , 1 1 3, 1 56
and the cut, 1 47-53
and sub-Number concept,
1 3 3-8
see also Numhers of fni te matter
maxi mal el ements and ordi nal s, 76,
8 1 , 1 1 3
maxi mal ordi na l of a coupl e,
1 8 5-96
medi ci ne, 2
Michel , Natacha, 1 3
Mi l l er, Jacques-Al ai n, 24-30, 3 9-
40, 44, 73
mi ni mal element of ordi nal s, 1 1 3,
1 1 7, 1 8 6-8
mi ni mal matter and cutti ng, 1 45-6,
1 53-5, 1 57, 1 76
mi ni mal i ty see pri nci pl e of
mi ni mal i ty
Mi ri manoff, D. , Axi om of
Foundati on, 222-3n
modern anal ysi s, 1 3, 56, 2 1 2
Moderni ty, 65
Mul ti pl e
and ontol ogy of mathemati cs,
1 3 1-2
and ordi nal s, 1 08, 1 09
processi on from the One, 7
pure mul ti pl e, 8, 44, 58, 65, 83,
1 01 , 1 85
mul ti pl e-bei ng and concept of
number, 1 07
mul ti pl i cati on, 1 99
and i nducti ve defni ti on, 95,
96-7, 207
mul ti pl i ci ti es, 6 1-97
Dedeki nd' s theory, 3 1 5, 56
i nconsistent mul ti pl ici ti es, 89,
1 44, 1 93, 1 99, 2 1 2
236 INDEX
mul ti pl ici ti es ( cont' )
natural mul ti pl i ci ty and ordi nal s,
68-70, 83, 84, 88-9, 1 66, 2 1 1
transi ti ve mul ti pl i ci ti es, 6 1 -6
von Neumann ordi nal s, 67-72
' na'ive' theory of sets, 3 1 -45
natural mul ti pl icity and ordi nal s,
68-70, 83, 84, 88-9, 1 66, 2 1 1
natural whol e numbers, 1 0, 1 2,
93-7, 1 5 7
addi ti on, 207-8
and Dedeki nd' s cut, 1 1
defni ti on, 94, 95
and defni ti on of Numher,
1 08-9, 1 1 0
' Peano' s axi oms' , 8, 1 1 , 1 2,
46-5 1 , 88
positive and negative numhers,
1 66-7
and successi on, 77, 84, 95
Nature and number, 69-70, 83, 92,
1 09, 1 30, 1 83-97
negative numbers, 12, 1 57, 1 58-6 1
addi ti on, 205-6, 208
natural whol e numbers, 1 66-7
and ordi nal s, 1 65
symmetric counterparts, 1 6 1-3,
1 74
Negati vity, 1 60-1
Neopl atoni sm, 7
Neumann, John von see von
Neumann
neutral i sati on and cutti ng, 1 47,
1 48
U as neutral el ement for addi ti on,
204-5
Ni etzsche, F. W. , 49, 65
nomi nal i st ori entati on of thought,
27, 1 75
nomi nati on and si gni fcati on,
1 06-7, 1 08, 1 75
non-refexi ve rel ati ons, 1 2 1 , 1 23-4,
1 56
non-sel f-i denti ty and zero, 25-30
' not i denti cal to itsel f' extensi on
and zero, 1 8-2 1 , 25-30
number
defni ti on of Number, 1 0 1 ,
1 02-1 2, 1 33, 228n
and Nature, 69-70
of Numhers, 1 39, 1 40
Peano' s si gns, 49-5 1
and structure, 1 3 1-8
and thought, 1 0, 88-9, 92
see also concept of number
number theory, 1 1 , 1 32
numbers
and Greek thought, 1 0
i mmensi ty o f Numbers, 1 02,
1 07-8, 1 1 2, 1 77-8, 200, 2 1 1 ,
2 1 2
Numbers o f fnite matter, 229n
addi ti on, 207
1 68-77
numerators, 1 67
numerical i ty of Number, 1 20, 1 21 ,
1 77-8, 2 1 3-1 4
and symmetrici sati on, 1 65-6
obj ects, 2 1 1
category theory and substructure,
226n
Dedeki nd and i nfni te, 38-43
Frege's doctri ne of, 1 9, 20
mathemati cal ontol ogy and
mul ti pl es, 1 3 1-2
and ' not i denti cal to i tsel f'
extensi on, 25-30
One, the, 7-8, 1 0, 1 3-1 5, 55, 85
Dedeki nd and Frege, 1 4, 44
Frege and concept of One, 1 8
and Peano's axi oms, 50
and trans-bei ng, 2 1 4
ontol ogy of mathematics, 1 3 1 -2,
1 33
ontol ogy of Number, 8-9, 58, 1 0 1 -
80, 1 57-8, 21 1-1 4
and al gebra, 200
and the cut, 1 75
and dyadi c rati onal s, 1 74
fundamental theorem of, 1 43-
55, 201 , 203
medi tati on on negati vi ty, 1 60-1
and number as concept, 1 9, 21 -
2, 30, 6 1
and order o f numbers, 1 21 -2
ordi nal s, 68
pai rs of ordi nal s and successi on,
1 89-9 1
and surreal number theory, 224n,
225n
and transi ti ve sets, 65-6
open sets, 142
operati onal di mensi on of Number,
9, 1 0 1 , 1 1 1 , 1 83-208, 2 1 1
al gebra, 1 96, 1 97, 1 98-208
defni ti on of operati ons on
Numbers, 201
fctions and dyadi c rati onal s,
1 75, 1 77
Numbers of fnite matter and
dyadic rati onal s, 1 73-4
and whol e Numbers, 1 66-7
see also cal cul ati on; counti ng;
mul ti pl ication
order of Numbers, 1 2, 1 0 1 , 1 1 1 ,
1 1 9-30, 1 56, 2 1 1
defni ti on, 1 1 9
densi ty of order, 140, 1 4 1 -2,
1 42-3, 1 55, 1 56, 1 79
dyadi c rational numbers, 1 67-8
see also ' wel l -orderedness' and
ordi nal s
ordered pa i rs of ordi nal s, 1 84-96
ordi nal s, 1 0, 1 2, 67-97, 1 56, 1 57
Cantor's theory of wel l
ordered ness, 52-8, 6 1 , 68,
1 88, 1 94, 1 96
coupl es of ordi nal s, 1 84-96, 201
and cuts, 1 39-55
and Dedeki nd, 3 1 , 34, 44
and defnition of Number, 1 02-
5, 1 08-1 2
di fference of numbers, 1 1 5-20,
1 56
and Frege, 1 1
and i nfnity, 44, 85-6, 93, 94
natural whole numbers, 93-7
as Numbers, 1 63-6
INDEX 237
operati onal abi l i ty, 1 84-97
and order of numbers, 1 20-30,
1 56, 1 64
properties of, 86-92, 94-5, 1 1 3,
1 34, 1 70-2
reasoni ng by recurrence, 86-92,
95, 96, 1 69
sets of ordi nal s, 1 1 2-1 4, 1 39
sub-Numbers as, 1 65
and successi on, 73-82, 84-5, 95,
1 8 8-96
and surreal numbers, 1 07
tri pl ets of ordi nal s, 1 96
von Neumann ordi nal s, 67-72
whol e ordi nal part of a Number,
1 70-3
' outsi de matter' posi ti on, 1 20,
1 45
parti cul ar predicates, 8
parti ti oned Number, 1 33-4, 201
parts of ordi nal s, 1 02, 1 03-5, 1 08,
1 09, 1 1 2, 1 1 3-1 4
Pascal , Bl ai se, 32
Peano, Gi useppe, 1 , 3-4, 9, 1 3 ,
46-5 1 , 2 1 2
axi oms, 8, 1 1 , 1 2, 46-5 1 , 8 8
Pessoa, Fernando ( poem) , 2 1 4
phi l osophy and l anguage, 48-9
pl ace of number, 93, 1 56-80
Dedeki nd, 35, 36-7, 39, 42
and successi on, 95
Pl ato, 35, 49, 64, 1 03
' pl atoni si ng' approach to number,
8-9
poem, ' suture' of phi l osophy, 8 1
poetry and the negati ve, 1 6 1
Pol i ti cal Science, 2
pol i ti cs
governed by numbers, 1-2
and the negative, 1 6 1
a s truth procedure, 5 7
revol uti on vs reform, 79
positive Numbers, 1 58-60
addi ti on, 205-8
dyadi c posi ti ve rati onal numbers,
1 67-77
238 INDEX
posi ti ve Numbers ( wnt' )
i nfni tesi mal numbers, 1 78-9,
229n
natural whol e numbers, 1 66-7
and ordi nal s, 1 65
symmetri c counterparts, 1 6 1-3
Number, 1 70-3
Presence, 56, 57
presentati on
and bel ongi ng, 65, 66, 68, 71
and structure, 1 3 1 , 1 33
pri nci pl e of the cut, 143
pri nci pl e of i denti fcati on, 1 43
Pri nci pl e of I denti ty, 1 8-1 9, 25
Pri nci pl e of I ndi sceri bl es, 27, 30
pri nci pl e of mi ni mal ity, 72, 88, 94-
5, 1 1 3, 1 1 7, 1 86-7
and cutti ng, 143, 144, 146, 1 53-5
Pri nci pl e of Non-Contradi cti on, 27
pri nci pl e of recurrence, 1 2, 47
see also reasoni ng by recurrence
Pri nci pl e of Suffci ent Reason, 1 9
procedure o f neutral i sati on, 1 47,
148
progressi on of number, 29-30
properties of ordi nal s, 86-92, 94-
5, 1 1 3, 1 34
Number, 1 70-2
property of succeedi ng, 75-6
proporti ons and ' i rrati onal
numbers' , 1 06
pure mul ti pl e, 8, 44, 58, 65, 83,
1 0 1 , 1 85
quasi -conti nui ty of rati onal s, 1 42
quaterni ons, 1 0, 228n
Ramanuj an, Sri ni vasa, 1 83-4
rati onal numbers, 1 0, 1 2, 1 09,
1 57, 1 99
and Dedeki nd' s cut, 1 4 1-2, 1 55,
1 74-5, 1 76-7
dyadi c posi ti ve rati onal numbers,
1 67-77
and Frege, 1 1
real numbers, 1 0, 1 2, 1 07, 1 09,
1 57, 229n
and al gebra, 1 99
and Dedeki nd's cut, 1 1 , 1 42,
1 74-5, 1 76-7, 1 79-80
defni ti on of, 1 76, 1 99-200
and dyadi c rati onal s, 1 74-7
and Frege, 1 1
i nfni tesi mal numbers, 1 78-9,
229n
posi ti ve real numbers and cut of
cuts, 1 79-80, 229n
recurrence
and Peano' s axi oms, 1 2, 47
reasoni ng by, 86-92, 95, 96, 1 69
see also i nducti ve defni ti on
rei gn of number, 2 1 3, 21 4
rel ati ve whol e numbers, 1 0
repeti ti on and number, 29-30
representati on and i ncl usi on, 65,
66, 68
resi due of the Number, 1 02, 1 04,
1 05, 1 1 1 , 1 1 2-1 3, 1 56
and the cut, 1 45, 147-53, 1 8 0
1 68, 1 69, 1 72
and sub- Number concept, 1 34
and symmetri cal counterpart,
1 6 1-3, 1 65, 1 74
upper bound of Number, 1 45
and zero, 1 59, 1 60-1
restrai ned acti on, 2 1 3, 2 1 4
ri ng theory, 1 32
Robi nson, A. , 50, 1 78
Russel l , Bertrand, 8, 2 1 2
Frege and Russel l ' s paradox, 20,
2 1 , 42, 43
salva lleritate, 26, 27
Sartre, Jean-Paul , 39
sci ence and the negati ve, 1 6 1
sciences and numbers, 2
'second moderni ty' of thi nki ng of
number, 1 3-1 5
section of Number, 1 08-9, 1 1 0,
1 1 1
seri al numerical i ty, 88, 1 2 1 -2
series of numbers, 7
set of concepts, 1 7, 20
set of parts, 9 1
set-theoreti cal approach to number,
8-9, 1 1 , 22
set theory
axi om of foundati on, 70-2
Dedeki nd' s ' na"ive' theory of sets,
3 1-45
set of al l sets, 42-3
and surreal numbers, 224n, 225n
sets of ordi nal s, 1 1 2-1 4, 1 3 9
si gn/si gni fer and the l etter, 1 3 0
si gn/si gni fer and number, 24-5, 29,
30, 49
Gonshor and surreal numbers,
224-5n
nomi nation and si gni fcati on,
1 06-7, 1 08, 1 75
Pea no' s axi oms, 48, 49-5 1
si mi l ar transformati on, Dedeki nd,
32, 34
si ngletons, 64-5, 7 1 , 95
Skol em, T. , 50
soci ety and number, 3
soci ol ogy, 2
Spi noza, Baruch, 3 8-9, 40-1
'square ci rcle' concepts, 1 8, 1 9
square numbers and bi uni vocal
correspondence, 36
square root, 1 99
statistics, 2
structure and substructure, 1 3 1 -8,
2 1 1
subj ect, Lacani an theory, 24, 25,
28, 29, 40
subj ecti vi ty of number
Dedeki nd, 39-40
Mi l l er's chal lenge to Frege,
24-30
sub-number concept, 1 3 1-8, 1 5 7
and cutti ng, 145, 147, 1 48 , 1 54,
1 76-7
and dyadi c rati onal s, 1 76
and natural whol e numbers, 1 66
and operati on, 20 1 , 202
INDEX 239
sub- Numbers as ordi nal s, 1 65
sub- Numbers as posi ti ve, 1 60
sub-obj ects and category theory,
1 3 1
subsets o f ordi nal s, 1 02, 1 03-5
substi tutabi l i ty, 26
substructure, 1 3 1-8, 1 47
succession, 30, 73-82, 84, 85, 95,
1 42, 1 88-96, 2 1 1
successor ordi nal s, 76-8, 79, 8 1 ,
84-5, 94, 95
coupl es of ordi nal s, 1 8 8-96
Dedeki nd, 34, 46-7
and di scri mi nants, 1 1 7-1 8 , 1 1 9
and i nducti ve defni ti on, 8 9-90,
9 1 , 96-7
Peano' s si gns, 49, 50
Number, 1 72
surreal numbers theory, 9, 1 3,
1 07-8
use of ' surreal ' , 1 07, 1 08
swarmi ng of numbers, 30, 1 1 2,
1 39, 1 40, 200
and fundamental theorem of the
ontol ogy of Number, 143-55
symmetric counterpart of a
Number, 1 6 1-3, 1 65-6, 1 74
symmetri ci sati on of addi ti on, 1 2
symmetri ci sati on of mul ti pl i cati on,
1 2
system of number, Dedeki nd, 32-8
termi nol ogy of Number, 1 05-7
thought and number, 1-4, 57-8,
88-9, 92, 2 1 3
the cut, 1 55
and Dedeki nd' s i nfnite systems,
38-4 1 , 47-8, 93
Frege and concept of number,
1 6-23
Greek thi nkers, 7, 1 0, 57
and l anguage, 48-9
topological ori entati on, 1 0, 1 2, 1 42
topol ogy of di fference, 1 20
total rel ati ons, 1 21 , 122-3, 1 56
trans-bei ng, 2 1 4
240 INDEX
transfni te i nducti on, 1 9, 86, 89,
90, 9 1-2
and coupl es of ordi nal s, 1 93,
1 96, 201
and operati on, 20 1-8
and surrea l numbers, 224n
transfni te numbers, 1 06
transformati on, Dedeki nd, 32-3,
34
transi ti ve sets
and i ncl usi on, 6 1 -6, 68, 1 1 3-1 4
and successi on, 74-5
and von Neumann ordi nal s,
67-72
transi ti vi ty, 1 2
di fference and order of numbers,
1 2 1 , 1 25-30, 1 56
tri pl ets of ordi nal s, 1 96
truth
and cal cul ati on, 26, 27
and Number, 2 1 3-1 4
and thought 4 1 , 2 1 3
truth process and event, 27
truth-case of concepts, 1 6, 25
ul tra-One, 2 1 4
unconsci ous, 41
uni ci ty and fundamental theorem
of ontol ogy of Number, 1 45,
1 53-4, 200
uni on of a set, 79-80, 8 1 , 9 1 , 1 02
uni queness and the cut, 1 43,
1 44-6, 1 55, 1 57
uni ty, 7, 8
Peano' s si gns, 49, 5 1
upper bound of ordi nal s, 1 1 3, 1 89,
1 90, 208
of a set of Numbers, 1 44-5
val ue and Number, 2 1 3-14
Venn di agrams, 1 05
voi d, 1 57-8, 2 1 2
and Evi l , 1 60-1
and i nducti ve defni ti on, 89, 91
si ngletons of, 64-5, 71 , 95
and successi on, 77, 84-5, 95
see also empty sets; zero
von Neumann, John, 8, 1 2
von Neumann ordi nal s, 67-72,
73
and successi on, 73, 74-5
' wel l -ordered ness' and ordi nal s,
52-8, 6 1 , 68, 1 88, 1 94, 1 96
whol e numbers see natural whol e
numbers; rel ati ve whol e
numbers
whole ordi nal part of a Number,
1 70-3
Wi ttgenstei n, Ludwi g, 2 1 2
Zermel o, E. F. F. , 8 , 22
Axi om of Separati on, 20-1 , 43
zero, 7-8, 1 3-1 5
and axi om of the empty sets,
44
and being, 25, 56-7
and bel ongi ng and i ncl usi on,
63-5
Dedeki nd and Frege on, 14, 40,
44
di scri mi nant, 1 5 8-9
exi stence of, 22-3, 56-7
Frege and concept of zero,
1 8-1 9, 22, 25-30, 40, 55
and i nfni tesi mal numbers,
1 78-9
l ack and functi on of zero, 26,
28, 34
as neutral el ement for addi ti on,
204-5
as Number, 1 57-8
ordi nal s and wel l -ordered sets,
55
and Peano' s axi oms, 50, 5 1
and posi ti ve and negati ve
Numbers, 1 5 8-6 1
see also empty sets; voi d

Number and Numbers

Diunggah oleh

Informasi Dokumen

Deskripsi Asli:

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Number and Numbers

Diunggah oleh

Hak Cipta:

Format Tersedia

'Breathtaking.Badiou ...announcesanewepochinphilosophy.

the order-rel ati on of E

K L 3: If rules 1 and 2 do not appl y for a given W, but i nstead

e After this, we general l y attribute to whole numbers q

Anda mungkin juga menyukai