Hegel and Deleuze on the Metaphysical Interpretation of the Calculus

Henry Somers-Hall
Manchester Metropolitan University, UK
1. Introduction
In this paper, I want to look at the role of the differential calculus in the development of
two different philosophical systems. I want to show how, prior to the development of a consistent
interpretation of the foundations of the calculus in the nineteenth century by eierstrass,
differential calculus seemed to offer the possibility, and also show the necessity, of a new way of
thinkin! about the fundamental lo!ical principles we use to describe the world. In particular, I
want to look at the way in which He!el drew support for the dialectical method from the
parado"es which seemed to result from these foundational issues. Indeed, whilst by He!el#s time,
the foundations of the calculus had been fre$uently attacked, most notably by %erkeley, He!el
ar!ues that &it is the inability to 'ustify the ob'ect (i.e. the differential coefficient dy)dx* as Notion
which is mainly responsible for these attacks#.
+
,hus for He!el, if we are to understand the
meanin! of the contradictions which seemed inherent in the foundations of the differential
calculus, we need to move to a dialectical understandin! of these foundations. -fter havin!
shown why He!el believes this to be the case, I want to discuss how this metaphysical
interpretation sheds li!ht on a conflict between He!el and the .rench philosopher, /illes 0eleu1e,
about the role of the transcendental in philosophy. In particular, I want to look at the way in which
He!el ar!ues for the need to incorporate the infinite into the finite, while 0eleu1e instead ar!ues
that the calculus re$uires us to move to a transcendental account. ,hus, while He!el ar!ues that
the antinomical nature of the calculus entails the need for a lo!ic capable of incorporatin!
contradiction, 0eleu1e follows Kant in ar!uin! that antinomy pushes us towards a transcendental
+
He!el 2+343, p. 5678.
+
account of the world. .or 0eleu1e, the differential provides the model for an element which is not
sensible, which he uses to !round his transcendental account of the !enesis of the sensible. I will
do this by lookin! at his own use of the calculus, where he also looks back to what he calls a
&barbaric# interpretation of the calculus.
e can see the importance of the calculus for metaphysics in %ertrand 9ussell#s early
adherence to He!elian philosophy. hilst the youn! 9ussell &found comfort for a time#
5
in a form
of He!elianism, this was premised on what appeared at the time to be the fundamentally
contradictory nature of mathematics, and in particular, the theory of infinite numbers, and the
foundations of the calculus. It was only after &eierstrass, soon after the middle of the nineteenth
century, showed how to establish the calculus without infinitesimals, and thus at last made it
secure#
:
that 9ussell was able to apply the method of lo!ical analysis to philosophy. ,his allowed
9ussell to re'ect the synthetic method of He!el in favour of his own formal analytic method.
,hus, for 9ussell, a dialectical understandin!, not only of mathematics, but also of the world in
!eneral, was the only approach open in philosophy until the development of set-theoretical
foundations of mathematics.
7
9ussell#s approach, with its basis in classical lo!ic, is representative
of the model which both He!el and 0eleu1e try and overcome with their metaphysical
interpretations of the calculus. .or He!el, the kind of approach e"emplified by 9ussell would be
classified as &finite thinkin!#, which He!el opposes by attemptin! to brin! the infinite into the
finite. .or 0eleu1e, 9ussell approach would be called &representation#. 0eleu1e ar!ues that if we
are to understand how representation comes into bein!, the account we !ive must use terms which
falls outside of representation. 0eleu1e thinks that the material for such an account can be found
in the calculus. .urthermore, he believes that the calculus !ives us the means to provide an
alternative to He!elian philosophy. ,he dialectical approach to the calculus did not merely allow
5
9ussell 2+36;, p. 5+8.
:
9ussell 2+37;, p. <4:8.
7
.or more on this point, and 9ussell#s early He!elianism, see Monk 2+33<8.
5
He!el to propose a resolution to the problems of the calculus, but also allowed He!el to provide
further support for his own dialectical method. %efore turnin! to He!el#s interpretation, however,
we will first look at how the calculus itself works. e will then look at the lo!ical difficulties
with this approach, particularly as e"pounded by %erkeley. -fter we have an idea of how these
difficulties occur, we will move on to the positive accounts of He!el and 0eleu1e in the second
half of the paper.
2. The Calculus
,he calculus presents one of the !reatest achievements of mathematics. In this first part
of the paper, I want to !ive a brief account of its operations. ,he calculus allows us to characterise
the relations between $uantities whose ratios to one another vary. -lthou!h it can relate any
variables, I will mostly refer to the relation between time and distance travelled by a hypothetical
body. In order to understand the issues resultin! from the calculus# foundations, however, we will
only need to look at a simple case, that is, the application of the calculus to polynomials. ,he
basic approach of the differential calculus was first used by the +7
th
century mathematician
=resme,
6
and came with the reco!nition that the velocity of a body could be represented by a line
on a !raph. If we represent on a !raph the relation between distance travelled and time taken to
travel that distance by an ob'ect, it becomes possible to determine the velocity of the ob'ect by
dividin! the distance travelled by the time, hence velocity is measured in terms such as miles per
hour, metres per second, etc. If the velocity is constant, the relation between distance and time
will be proportional. ,his means that the line representin! the movin! ob'ect will be strai!ht. If,
therefore, we wish to determine the velocity of the ob'ect, we simply need to take a section of the
line, and divide the distance travelled over that time, which will be represented by the chan!e in
the value on the distance a"is over the len!th of the section, by the time, which will be
6
%oyer 2+363, p. 458, althou!h it was -rchimedes who provides the first reco!nised anticipation of the
calculus in the method of e"haustion used to find the area of a circle. -rchimedes# use was concerned with
static fi!ures, rather than rates of chan!e however.
:
represented by the chan!e in the value of the time a"is over the len!th of the section. -s the two
values are directly proportional, a section of any arbitrary len!th will provide the same result.
Movin! on from =resme, if we are dealin! with an ob'ect movin! at a velocity which is not
constant, then this procedure cannot be used, as we were able to determine the velocity at any
point usin! the previous method only because the velocity was the same at every point 2the
avera!e velocity is the same as the velocity at every instant8. Instead, however, we can measure
the velocity at any point of a system with a chan!in! velocity by drawin! a !raph of the function
;
of this chan!e, that is, of the relation of distance to time, and drawin! a line which runs parallel to
a particular point on the curve. ,his produces a vector of the velocity of the system at this
particular moment. ,he difficulty with this approach is that it can only be appro"imate, as we are
attemptin! to draw a line throu!h a point, which in itself can seemin!ly have no direction. ,he
alternative, to draw a line throu!h two points of the curve, is e$ually flawed, as althou!h it !ives
us an accurate line, we are dealin! with a curve, and so the tan!ent we are now drawin! will not
represent the velocity at one particular moment, but the avera!e between the two points. >eibni1#s
solution to this difficulty was to draw a line between the point whose velocity we wish to
measure, and another arbitrary point on the curve, and then to ima!ine the distance between these
two points decreasin! towards 1ero. -s we now have a strai!ht line between these two points, we
can treat the case in the same manner as the case of constant velocity described above, measurin!
the chan!e in values of both a"es alon! a len!th of the line. ,hus mathematically, we end up with
two lines, one representin! the chan!e over the section in terms of distance, and one in terms of
time, neither of which on its own will have any determinate value, as the lines are infinitely short,
but when divided, one by the other, will !ive a vector at the particular point. Since the a"es of the
!raph can represent more than simply time and distance, these values are referred to more
!enerally as dy and dx. .urther, usin! 0escartes# insi!ht that a function can be represented by a
;
- function consists in a domain 2or ran!e8 of numbers, and a rule which associates each of these numbers
with a value of another variable. ,hus, f2"8 ? "
5
associates any real number, positive or ne!ative, with a
value e$ual to "
5
.
7
!raph, and a !raph by a function, we can define this procedure as a function of the ori!inal
formula, dispensin! with the !raphical method alto!ether.
In !ivin! a brief outline of the method of the calculus, I shall follow He!el#s statement
that &the whole method of the differential calculus is complete in the proposition that
dx nx dx
n n +
= , or P
i
fx i x f
=
+ 8 2
#.
<
hilst these statements may appear intimidatin!, they
are in fact simply al!ebraic representations of !eometrical functions. ,he second e$uation simply
represents the !radient function of a curve. /radient, as we saw in the case of strai!ht lines, is
defined in terms of the chan!e in the value on the y a"is divided by the correspondin! chan!e on
the x a"is, thereby !ivin! a ratio. @ow, if we take i to be a len!th alon! the x a"is, assumin! that
8 2x f y =
, that is, that the value of y is dependent on the value of x, the top half of the e$uation
will !ive us the difference between the values of y when
x y =
and
i x y + =
. ,hat is, the top
half of the e$uation will !ive us the correspondin! value of the rise on the y a"is for a run of
len!th, i, on the x a"is, which be!ins at any arbitrary point on that a"is. e thus have a function
which will provide us with the value of the tan!ent at any point on the line. ,he first e$uation
follows from this second e$uation throu!h the substitution of an infinitesimal value for i. ,his
substitution is the e$uivalent of makin! dy and dx infinitesimal values in >eibni1#s work.
-lthou!h showin! that the first e$uation is true for all values of n involves some comple"
mathematics, it is an easy result to prove for any specific function. e will therefore focus on one
particular function, that !iven the e$uation,
5
x y = , then x
dx
dy
5 = .
<
He!el, +343, p. 5<7.
6
,he first step is to rewrite the e$uation, P
i
fx i x f
=
+ 8 2
, replacin! the value i with
dx, the >eibni1ian infinitesimal, and replacin!
8 2x f
with
5
8 2x , as in this case,
5
x y = . e
thus arrive at the functionA
dx
dy
dx
x dx x
=
+
5 5
8 2
e can now multiply out the brackets on the top half of the e$uation to !iveA
dx
dy
dx
x dx xdx x
=
+ +
5 5 5
5
hich in turn reduces toA
dx
dy
dx
dx xdx
=
+
5
5
Bancellin! out dx therefore !ives usA
dx x
dx
dy
+ =5
,hus the differential of
5
x turns out to be dx x + 5 , where, !iven that the value dx is
infinitesimal with respect to x, we can 'ust write x
dx
dy
5 = 2dx retains a value in relation to dy,
which is also infinitesimal, and so is retained on the left side of the e$uation8. ,hus, it is easy to
show that for the value n ? 5, He!el#s first e$uation follows from his second. .or hi!her values, a
similar result can be obtained with more time consumin! applications of such elementary al!ebra.
,he procedure that we have 'ust carried out is called differentiation. If we reverse this process,
however, we have a procedure which allows us to find what is known as the primitive function of
the derivative, i.e. the function from which the derivative is differentiated.
4
,his process is known
4
,he two operations are not strictly inverse, as differentiation removes any constant values from the initial
function, so, for instance, if 8 2x f y = is +
5
+ =x y , the first sta!e of the differentiation will !ive usA
dx
x dx x
dx
dy 8 + 2 + 8 2
5 5
+ + +
= , which is e$uivalent to the function derived from
5
x y = . In inte!ration,
the constant values are therefore normally represented by the value c. ,he value of c can be determined if
;
as inte!ration. .or modern set theory, the interpretation !iven above has !iven way to an
interpretation in terms of the concept of limit, in order to avoid certain difficulties which emer!e
from the parado"ical $ualities of the infinitesimal, which we will come back to below in our
discussion of He!el#s theory of the calculus. In order to escape these difficulties, the modern set
theoretic interpretation of the differential calculus lar!ely i!nores the individual values of dy and
dx, instead !ivin! a value simply to the ratio, dy)dx, as a whole, called a differential. ,his is
accomplished by admittin! the concept of a limit of an infinite series. -n infinite series such as
+)5 C +)7 C+)4 C+)+;D may be said to approach the value, +, and this allows us on the modern
interpretation to e$uate the derivative, not with the final, infinitesimal value of the series, but
instead with the limit of the series. e do not in the case of the series need to worry about the fact
that any finite summation will always be less than one. In a similar way, the differential is
conceived of under the set-theoretic interpretation as precisely e$uallin! the value of the !radient
at any particular point. %y definin! the derivative to be the limit of the ratio, rather than the
values of dy and dx, $uestions about what happens when or whether the ratio actually reaches this
limit are put out of play. ,his is the method of eierstrass, which allowed 9ussell to move away
from an antinomic interpretation of mathematics, and therefore also from his early He!elianism.
e will now turn, however, to He!el#s theory of the calculus.
3. Hegel and the Calculus
,he basic function of the differential calculus is to find the !radient of a curve at a
particular point. ,he difficulty with a curve is that as the !radient varies, we need to find the
!radient of a point. /radient, however, seems to involve a difference. .or >eibni1, at least in his
informal account, the solution was to resort to the notion of the infinitesimal, as an infinitely
any value of the function is known, however.
<
small difference between two points, dy. -s this difference was infinitely small, it could be
discounted for the purposes of calculation, but, as it retained a ma!nitude relative to dx, it could
be used to form a ratio, dy)dx which had a determinate value. .or 0eleu1e, this approach forms
the foundation for his interpretation of the differential calculus, althou!h he takes advanta!e of
later developments which finesse the parado"es resultin! from it. He!el re'ects this >eibni1ian
approach since the idea of ne!lectin! infinitesimal values, whilst it may lead us to the correct
results, does not !ive us the ri!our essential to a proper mathematical proof. &Even if ordinary
common sense in fairness allows such ine"actitude, all !eometricians re'ect this conception. It is
$uite obvious that in the science of mathematics there cannot be any $uestion of such empirical
accuracyF mathematical measurin! by operations of the calculus or by !eometrical constructions
and proofs is alto!ether different from land-surveyin!, from the measurin! of empirical lines,
etc#.
3
In speakin! of >eibni1 in particular, He!el asserts that &it is chiefly this call to ne!lect
which, alon! with a !ain in facility, has !iven the calculus the appearance of ine"actitude and
e"press ine"actness in its method of procedure#.
+G
-lthou!h >eibni1#s method !ives a result which
is out by only an infinitesimal amount 2the !radient is !iven as 2x C dx, rather than the correct
value of 2x8, this error is not con!ruent with the standards of a formal discipline such as
mathematics.
++
,he tar!et of He!el#s criticisms is especially clear in the first postulate of the
calculus !iven in >#HHpital#s Analyse 2+;3;8, which provided the first !eneral introduction to the
methodA
3
He!el, +343, p. 564.
+G
He!el, +343, p. 564.
++
In fact, >eibni1 was aware of these limitations of the infinitesimal method, and in his first published
account of the calculus in +;47, he attempted to provide a proof of the calculus based on finite line
se!ments 2see =U Mathematics ,eam (+3<7, book III p. 6G*8. It turned out, however, that this definition
could not be applied to differentials occurrin! in the inte!ral calculus, and so >eibni1 eventually resorted to
considerin! the differential as a useful fiction. ,his was at odds with the !eneral approach of the >eibni1ian
school as a whole, however, and thinkers such as >#HHpital, who attempted to secure the foundations of the
calculus on the e"istence of infinitesimals.
4
Postulate I. /rant that two $uantities, whose difference is an infinitely small
$uantity, may be taken 2or used8 indifferently for each otherA or 2which is the same
thin!8 that a $uantity, which is increased or decreased only by an infinitely smaller
$uantity, may be considered as remainin! the same.
+5
It is this postulate which allows us to discount the effects of dx in the results of our derivation of
the differential function, but merely by incorporatin! a lack of ri!or into the foundations of the
calculus themselves. >eibni1 did not provide the only account of the foundations of the calculus,
however, and so we shall instead look at @ewton#s interpretation.

hilst >eibni1 was the first to publish his theory of the calculus, @ewton had
independently developed a form of the calculus called the method of flu"ions. @ewton tried to
distin!uish his own method of flu"ions from the calculus, which were practically e$uivalent,
throu!h his attempt to determine a more ri!orous foundation. ,his was the method of prime and
ultimate ratios which @ewton first published in the Principia 2+;4<8. ,he key to this approach
was to incorporate the idea of time into his demonstration. -s he writes in the Quadrature of
CurvesA
I don#t here consider Mathematical Iuantities as composed of parts e"treamly
small, but as generated by a continual motion. >ines are described, and by
describin! are !enerated, not by the apposition of Jarts, but by a continual motion of
points. Surfaces are !enerated by the motion of >ines, Solids by the motion of
Surfaces, -n!les by the 9otation of their >e!s, ,ime by the continual flu", and so
+5
,his $uotation of >#HHpital#s te"t is taken from Struik 2+34;, p. :+78. See =U Mathematics ,eam 2+3<7,
book IK, p. +G8 for a fuller analysis.
3
on in the rest. ,hese Geneses are founded upon @ature, and are every 0ay seen in
the motion of %odies.
+:
@ewton uses this dynamic conception of the !eneration of the curve itself to posit the idea that
the curve could be understood as the movement of a point with a certain velocity. ith this, it is a
small step to see the !radient of the point on the curve as representin! the instantaneous velocity
of the point itself, which we find by notin! that the time it takes the point to move can be reduced
to 1ero. Under this interpretation, the differentials, called flu"ions in @ewton#s system, do not
have to be seen as bein! too small to affect the result of the calculation, but can actually be seen
to vanish at the limit point, when dt?GA
Quantities, and the ratios of uantities, !hich may in any finite time converge
continually to euality, and before the end of time approach nearer to each other
than by any given difference, become ultimately eual"
If you deny it, suppose them to be ultimately une$ual, and let # be their ultimate
difference. ,herefore they cannot approach nearer to e$uality than by that !iven
difference #$ which is contrary to the supposition.
+7
,he ultimate ratio therefore functions much as the idea of a limit does in modern mathematics. If
we take a normal ratio, such as 4);7, if we reduce the values of the denominator and numerator
proportionately, for instance, to 7):5, or 5)+;, we find that the ratio of terms itself remains
unchan!ed. @ewton uses the idea of an ultimate ratio to ar!ue that at the limit case, where the
numerator and denominator are reduced to G)G 2where the lines become a point8, we can still
+:
@ewton 2+3;7, book I, p. +7+8. See =U Mathematics ,eam 2+3<7, book III, p. 5;8, for a fuller discussion
of @ewton#s ar!ument in the Quadrature of Curves.
+7
@ewton 2+3:7, p. 538.
+G
interpret the point as havin! a determinate ratioA L-nd, in like manner, by the ultimate ratio of
evanescent $uantities is to be understood the ratio of $uantities, not before they vanish, nor after,
but with which they vanish.M
+6
It is this conception of the calculus, as !rounded in the idea of an
ultimate ratio, which He!el takes to characterise it most ade$uately, albeit with the elimination of
&those determinations which belon! to the idea of motion and velocity 2from which, mainly, he
took the name, fluxions8 because in them the thou!ht does not appear in its proper abstraction but
as concrete and mi"ed with non-essential forms#.
+;
,o see how @ewton#s understandin! of the
calculus is inte!rated into He!el#s dialectic, as well as why He!el believed it to re$uire a
dialectical foundation, we will now turn to the idea of the ratio as it occurs in He!el#s remarks on
the mathematical infinite in order to trace the !enetic development of He!el#s own understandin!
of the ultimate ratio.
4. The Ratio in Hegels discussion of !uantitati"e Infinity
He!el#s analysis of the differential calculus occurs in his discussion of $uantitative
infinity in the %cience of &ogic. Here, he uses his two different conceptions of infinity, the true
2dialectical8 infinite and the spurious infinite. -s we shall see, in this, the true infinite will share
many of the properties of the more commonly discussed $ualitative true infinite. .or He!el, the
true infinite e"hibits the ne!ation and preservation of both the finite and the infinite within one
whole. ,his overcomes the spurious infinite, which is merely the perpetual repetition of the finite.
ithin the ratio, both these moments appear simultaneously. If we take He!el#s e"ample of the
fraction 5)<, we can represent it in the two followin! waysA
5)< ? G.546<+75D
+6
@ewton 2+3:7, p. :38.
+;
He!el 2+343, p. 5668. See also Eves 2+33G, p. 7GG8 on the redundancy of time in @ewton#s method.
++
=n the ri!ht, we have the decimal e"pansion of the fraction, which in this case forms an infinite
series.
+<
.or He!el, this infinite series cannot be construed as a true infinity, as it 'ust represents a
bare repetition of finite values. &hat the series is meant to e"press remains an ought'to'be and
what it does e"press is burdened with a beyond which does not vanish and differs from what was
meant to be e"pressed.#
+4
hen we assert an e$uality between 5)< and its e"pansion, we have to
note that while continuin! the e"pansion of the series allows the value on the ri!ht of the e$uals
si!n to be made as accurate as re$uired, it will always preserve a difference from the value on the
left. ,his difference is the beyond which would have to be incorporated into the series to make it
a representation of the true infinite, and therefore the source of the ought. .or He!el, the
contradiction implied by the si!n of e$uality combined with the necessary difference between the
two sides of the e$uation emer!es from the fact that what we have on the left of the e$uation is no
lon!er a $uantitative notion, but has passed on into a $ualitative determinateness, as the ratio is
not tied to any specific value, but instead to their relation. ,his ratio cannot be captured by the
e"pansion of the purely $uantitative series of numbers on the ri!ht, no matter how closely it may
approach the fraction. ,he fractional representation, as includin! that which is unobtainable from
the $uantitative determination 2the moment of difference8, therefore becomes He!el#s first
appro"imation of the $ualitative mathematical infinite.
,he fractional understandin! of the calculus is ultimately not an ade$uate representation
of true infinity for He!el, however. hat is important about the fraction 5)< is that it surpasses the
purely $uantitative determination of the series. -s such, the two apparent $uanta of the ratio show
the property of variability, to the e"tent that the values of 5 and < can be replaced with other
values without chan!in! the overall value of the fraction. ,hus, 5)< is e$uivalent to 7)+7 or ;)5+.
+<
He!el notes 2+343, p. 57<8 that fractions such as +)+G, which do not normally lead to an infinite se$uence
can be made to do so by interpretin! them in another number base 2i.e. instead of base +G8. ,hus, for
instance, +)+G, when interpreted in the octal system 2base 48, also leads to an infinite se$uence of
G.G;:+7;:D
+4
He!el 2+343, p. 5748.
+5
In all of these cases, however, we are still dealin! with $uanta for which the relationship is not
essential. ,hus, whilst the unity of the different terms is essential for the fraction itself, it is not
essential for the terms which make up the fraction. Similarly, He!el notes that movin! to an
al!ebraic description of the fraction does not overcome this limitation, as an al!ebraic
formulation such as y)x ? a can e$ually be written in a form which does not contain a ratio, such
as y ? ax, the e$uation of a strai!ht line. Kariability, therefore is shown not to be the definin!
characteristic of the $ualitative mathematical infinite, as even in the case of the al!ebraic variable,
we still have a symbol standin! in for an arbitrary $uantum, that is, we have variability, but
variability still conceived of as ma!nitude. It should be clear now why He!el moves to the
@ewtonian interpretation of the calculus. In @ewton#s ultimate ratio, as we are dealin! with the
ratio of values at a point, neither of the terms in the ratio, dy)dx, can have any meanin! outside of
the ratio itself. &-part from their relation they are pure nullities, but they are intended to be taken
only as moments of the relation, as the determinations of the differential co-efficient dx)dy#.
+3
In
the differential relation, we therefore have a situation whereby both the ratio itself as well as the
terms can only be understood as a totality. .or He!el, this represents the true dialectical relation
of the $uantum and the infinite. In the differential ratio, each of the $uanta can only e"ist in
relation to the other. ,hat is, the dy and dx only have meanin! throu!h their reciprocal relation.
hat is important to reco!nise is that here we have an e"ample of the He!elian synthesis of the
finite and the infinite. ,hat is, the finite determinations, the elements in the ratio, can only e"ist in
relation to one another. /oin! back to He!el#s initial discussion of the ratio, the ratio in its totality
therefore represents the infinite, as in this case, neither the infinite nor the finite can therefore
e"ist without the other. e should note further that this structure resembles the structure of
contradiction found in the doctrine of essence. -s He!el moves to a lo!ic of contradiction in
order to overcome the limitations of the finite understandin!, however, we must now look at why
+3
He!el 2+343, p. 56:8.
+:
He!el believes that this understandin! of the infinite cannot be arrived at purely throu!h an
analytic deduction.
#. $er%eley and the &oundations of the Calculus
He!el writes that &the operations which (mathematics* allows itself to perform in the
differential and inte!ral calculus are in complete contradiction with the finite determinations and
their relations and would therefore have to be 'ustified solely by the Notion#.
5G
e have seen that
He!el re'ected >eibni1#s interpretation as lackin! ri!our. @ewton#s interpretation of prime ratios
in terms of instantaneous velocity also presents problems, however, as velocity is a rate of
chan!e. .or @ewton, these ratios were also described as nascent ratios, or the ratios at the point
where the flu"ions start to be. hat is problematic in @ewton#s model, therefore, is whether such
a thin! as an ultimate ratio can actually be said to e"ist, since it also appears implicitly to rely on
the dual properties of the ratio not yet havin!, but at the same time already havin! a determinate
$uantity. ,he problems of the foundations of the calculus, althou!h known to >eibni1 and
@ewton, were made notorious in %erkeley#s treatise, (he Analyst$ or, a #iscourse Addressed to
an Infidel )athematician !herein it is examined !hether the *b+ect, Principles, and Inferences
of the modern Analysis are more distinctly conceived, or more evidently deduced, than ,eligious
)ysteries and Points of -aith. ,he essence of %erkeley#s criticism lies in the statement thatA
If with a view to demonstratin! any proposition a certain point is supposed, by
virtue of which certain other points are attainedF and such supposed point be itself
afterwards destroyed or re'ected by a contrary suppositionF in that case, all other
points, attained thereby and conse$uent thereupon, must also be destroyed and
5G
He!el 2+343, p. 5678.
+7
re'ected, so as from thence forward to be no more supposed or applied in the
demonstration. ,his is so plain as to need no proof.
5+
,hus, %erkeley attacks @ewton#s notion of the ultimate ratio for both appearin! to be une$ual to
1ero 2as the terms formin! the ratio can be divided by one another8, but also e$ual to 1ero 2in
order for the ratio to be applied to an instant8. /oin! on to $uestion the flu"ions which make up
the elements of both @ewton#s ultimate ratio as well as He!el#s mathematical infinite, %erkeley
asksA &hat are these flu"ionsN ,he velocities of evanescent incrementsN -nd what are these
same evanescent incrementsN ,hey are neither finite $uantities, nor $uantities infinitely small nor
yet nothin!. May we not call them the !hosts of departed $uantitiesN#
55

,he !round of %erkeley#s criti$ue therefore is that the idea of the flu"ion is !rounded in
&direct impossibilities and contradictions#
5:
which result from &the most incomprehensible
metaphysics.#
57
.rom a He!elian perspective, however, %erkeley#s criticisms can be seen as
restin! on a form of picture thinkin! which is based on takin! a purely static view of the !rounds
of the calculus.
56
He!el writesA
-lthou!h the mathematicians of the infinite maintained that these $uantitative
determinations are vanishin! ma!nitudes, i.e. ma!nitudes which are not any
particular $uantum and yet are not nothin! but are still a determinateness relative to
5+
%erkeley 2+335, para!raph +58.
55
%erkeley 2+335, para!raph :68.
5:
%erkeley 2+335, para!raph 48.
57
%erkeley 2+335, para!raph 748.
56
B.f. %erkeley 2+335, para!raph 78A &@ow as our sense is strained and pu11led with the perception of
ob'ects e"tremely minute, even so the ima!ination, which faculty derives from sense, is very much strained
and pu11led to frame clear ideas of the least particles of time, or the least increments !enerated thereinA and
much more so to comprehend the moments, or those increments of the flowin! $uantities in statu nascenti,
in their very first ori!in or be!innin! to e"ist, before they become finite particles.#
+6
an other, it seemed perfectly clear that such an indeterminate state, as it were called,
between bein! and nothin! does not e"ist.
5;
,he fact that there cannot be a state which contains both bein! and nothin!ness does not trouble
He!el, however. &,he unity of bein! and nothin! is, of course, not a stateF D on the contrary, this
mean and unity, the vanishin! or e$ually the becomin! is alone their truth#.
5<
,hus, the true
foundation of the calculus, accordin! to He!el, is to be found in the results obtained in the
dialectic of bein! and nothin! which opens the %cience of &ogic" ,he differential calculus is
therefore seen as bein! !rounded in the fundamental dialectical moment of transition, within
which the two moments of the flu"ion, bein! and nothin!, are to be taken as immanently related.
e thus have a ratio of two terms, both of which are on the brink of vanishin!, but which, when
related to one another, !ive a determinate value. It is the two transitions of the flu"ions, therefore,
when related to one another throu!h the form of the ratio, which !enerate the determinacy of the
calculus. %erkeley#s attempt to show the contradiction of the calculus reduces it to &a diseased
condition e"ternally induced throu!h erroneous thinkin!#,
54
essentially the thou!ht of the finite
understandin! which eschews any consideration of movement. %erkeley#s criticism shows that
the calculus is the e"pression of a contradiction, but accordin! to He!el, without a proper
understandin! of the nature of contradiction itself, this truth separates the differential calculus
from speculative science, rather than bein! the principle of its unity with it. Ultimately, He!el#s
characterisation of the differential calculus will therefore be premised on what is almost an a"iom
of dialectical thinkin!, that &there is nothing !hich is not an intermediate state bet!een being and
nothing#.
53
hen these flu"ions are incorporated into the ultimate ratio itself, we have a structure
which is isomorphic with the structure of both contradiction and the infiniteA
5;
He!el 2+343, p. 5678.
5<
He!el 2+343, p. 5678.
54
He!el 2+343, p. 5678.
53
He!el 2+343, p. +G68.
+;
,he truth is rather that that which has bein! solely in the ratio is not a $uantumF the
nature of $uantum is such that it is supposed to have a completely indifferent
e"istence apart from its ratio, and its difference from another $uantum is not
supposed to concern its own determinationF on the other hand, the $ualitative is
what it is only in its distinction from an other. ,he said infinite ma!nitudes,
therefore, are not merely comparable, but they e"ist only as moments of
comparison.
:G
e thus have the unity of moments which can only e"ist in their difference throu!h this unity.
.urther, we can see that He!el#s approach is not to move to anythin! like a transcendental account
in 'ustifyin! the calculus. Instead, He!el supplements representation by incorporatin! a moment
of movement, in the form of vanishin!, into representation itself. Everythin! therefore happens on
the same ontolo!ical plane for He!el. hat this analysis has attempted to show is how, for He!el,
the differential calculus both re$uires a move to a dialectical understandin! of mathematics, and
also, in its dialectical development, comes to represent the structure of the system as a whole as it
incorporates the movement from bein! to nothin!. hilst He!el#s analysis of the calculus !oes on
to en!a!e with a variety of his contemporaries# interpretations of the calculus, we will now turn to
0eleu1e, havin! showin! the necessity of this dialectical moment in He!el#s account.
:+
.or
0eleu1e also, the calculus represents a foundin! metaphor. In the ne"t section, therefore, we will
move on to e"plore how 0eleu1e#s understandin! of the calculus differs from that of He!el in
order, throu!h the lar!e number of commonalities which they share in their interpretations, to
hi!hli!ht in precisely which ways they differ in their responses to the problems to be found in
representation.
:G
He!el 2+343, p. 5668.
:+
See 0uffy 25GG;8 for a discussion of He!el#s en!a!ements in these contemporary debates.
+<
'. Deleuze and the Calculus
e have an opposition between two conceptions of the calculusF on the one hand, we
have the finite representational schema of the classical interpretation, and on the other, the infinite
representational schema of the He!elian dialectic. In the li!ht of this, it has become relatively
standard in readin! 0eleu1e to treat him as usin! the tools of modern mathematics to cut off the
path to He!elian dialectic by resolvin! the antinomies at the base of the calculus. ,his view is
clearly implicit in 0elanda#s interpretation, and is most clearly e"pressed by Simon 0uffy in (he
&ogic of .xpression/ Quality, Quantity and Intensity in %pino0a, 1egel and #eleu0e, where he
writes that, L0eleu1e D establishes a historical continuity between >eibni1#s differential point of
view of the infinitesimal calculus and the differential calculus of contemporary mathematics
thanks to the a"ioms of non-standard analysis which allow the inclusion of the infinitesimal in its
arithmetisationF a continuity which effectively bypasses the methods of the differential calculus
which He!el uses in the %cience of &ogic to support the development of the dialectical lo!ic.M
:5
I
want to ar!ue, contrary to this view, that 0eleu1e in fact wants to re'ect both positions in order to
develop a theory of the calculus which escapes completely from the dichotomy of the finite and
infinite. 0eleu1e#s aim will be to use the calculus to foster an understandin! of the transcendental
free from all resemblance to the empirical, as it is only once this resemblance has been removed
that the transcendental can be seen as the !rounds of the !eneration of the empirical, rather than
simply as a conditionin! factor. ,hus he will re'ect both finite interpretations, which understand
the differential $uantitatively, as a determinate, if infinitesimal ma!nitude, and the infinite
interpretation of He!el, which only !ets as far as the vanishin! of the $uantum, and therefore
leaves its status as vanished 2from the realm of $uanta at least8 untouched. ,his will allow him
construe the transcendental field as constituted entirely non-ob'ectivally, openin! up the
possibility of a !enerative account of the empirical.
:5
0uffy 25GG;, pp. <7-<68.
+4
0eleu1e#s own en!a!ement with the calculus is intertwined with his criti$ue of He!el.
Historically, 0eleu1e situates his own pro'ect in relation to what he calls a &!eneralised anti-
He!elianism#
::
which pervaded intellectual culture in +3;Gs .rance. In lookin! at 0eleu1e#s own
interpretation of the calculus, I therefore want to show that he also retains a pre-eierstrassian
interpretation of the calculus, and that he uses this interpretation to differentiate his own position
from that of He!el. 0eleu1e#s e"position of the calculus be!ins with the pronouncement that &'ust
as we oppose difference in itself to ne!ativity, so we oppose dx to not--, the symbol of difference
to that of contradiction#.
:7
In spite of this, in his interpretation of the calculus, there are a !reat
many similarities with the He!elian model. =ne of 0eleu1e#s main aims in !ivin! his e"position
of the calculus is to provide a metaphysic which will &take dx seriously#.
:6
.or 0eleu1e, this
approach can only be achieved by providin! an interpretation that will leave it &separated from its
infinitesimal matri"#.
:;
hat is drivin! 0eleu1e in this matter is, in parallel with He!el, an
attempt to provide an interpretation of the calculus which does not !ive the differentials the status
of infinitely small $uanta. Mirrorin! He!el#s ar!ument about the representation of the infinite in
the ratio, 0eleu1e writes, &$uanta as ob'ects of intuition always have particular valuesF and even
when they are united in a fractional relation, each maintains a value independently of the
relation#.
:<
.or this reason, the >eibni1ian interpretation of the calculus, which relies on infinitely
small $uantities is re'ected by 0eleu1e. Similarly, the calculus cannot be conceived of as
operatin! with al!ebraic terms, since in this case &there must always be a particular value char!ed
with representin! the others and with standin! in for them#.
:4
,hus, the e$uation of the circle,
G
5 5 5
= + , y x , is opposed to its differential, ydy C xdx ? G, in which such a substitution of
::
0eleu1e 2+337, p. "i"8.
:7
0eleu1e 2+337, p. +<G8.
:6
0eleu1e 2+337, p. +<G8.
:;
0eleu1e 2+337, p. +<G8.
:<
0eleu1e 2+337, p. +<+8.
:4
0eleu1e 2+337, p. +<+8.
+3
arbitrary values cannot be made. In this e"ample, however, we can already see the difference
between He!el and 0eleu1e becomin! clear. hilst the differential of the circle can be
represented as a ratio, 0eleu1e instead chooses the representation which maintains the
separability of the differentials. His rhetorical device here si!nifies his intention to present an
interpretation of the differential which falls outside of the positions of the infinitesimal and the
ultimate ratio.
.or 0eleu1e, &the symbol, dx appears as simultaneously undetermined, determinable, and
determination#.
:3
,o each of these three moments, we have a principle, which to!ether form the
basis of a transcendental interpretation of the calculus. ,he first sta!e, the undetermined as such,
corresponds to the differentials themselves 2dx, dy8. Here, in what is his most important break
with He!el, 0eleu1e ar!ues that whilst dx is strictly nothin! in relation to x, this is not because the
differential is not in a sense real, but rather because it cannot be captured by either 2Kantian8
intuition or the cate!ories of $uantity. ,his characteristic of fallin! outside of both $uantity and
intuition is what 0eleu1e calls &continuousness#,
7G
presumably to hi!hli!ht its difference from the
discrete nature of $uantitative determinations. In this sense, rather confusin!ly, continuousness is
the &ideal cause of continuity#,
7+
continuity here understood as the kind of continuity discovered
in the sensible realm. Oust as with He!el, 0eleu1e is opposed to thinkin! of the differential ratio in
terms of variability, and continuousness instead represents the transcendental condition for
variability, since it is throu!h the reciprocal determination of differential elements, which
themselves escape determination as $uantities, that determinate $uantities are !enerated by the
differential function. In this sense, the differential, dx, as a symbol of difference, is &completely
undetermined#,
75
that is, as the representation of the &closest noumenon#,
7:
difference, it escapes
:3
0eleu1e 2+337, p. +<+8.
7G
0eleu1e 2+337, p. +<+8.
7+
0eleu1e 2+337, p. +<+8.
75
0eleu1e 2+337, p. +<+8.
7:
0eleu1e 2+337, p. 54;8.
5G
the symbolic order. ,he symbols, dy and dx, and their values of G in respect to y and x, therefore
represent the annihilation of the $uantitative within them in favour of what 0eleu1e calls the sub-
representational, or e"tra-propositional.
It should be clear from what we have said of >eibni1#s interpretation of the calculus that
once the differential is not seen as an infinitesimal $uantity, we can no lon!er conceive of the
process of differentiation as entailin! the infinitesimal appro"imation of the tan!ent of a point.
Instead, 0eleu1e introduces the notion of limit, but not the limit of the modern interpretation,
whereby the differential is the value an infinite series conver!es on. Instead, &the limit must be
conceived not as the limit of a function but as a !enuine cut#.
77
0eleu1e#s point is that whereas dy
and dx are completely undetermined in relation to x and y, they are completely determinable in
relation to one another. e can relate this back to our initial description of >eibni1#s interpretation
of the calculus by notin! that >eibni1 held differentials to be merely $uantitatively distinct from
the values of x and y, as infinitesimally small values. ,his led to the difficulty that on the one
hand they relied on this $uantitative moment to form a ratio and thereby determine the value of x,
but on the other hand, this moment had to be i!nored in order that we could determine the value
of x itself, rather than x 2 dx. .or 0eleu1e, dy and dx are not to be characterised $uantitatively,
and so, when we determine the $uantitative answer, they fall out of the e$uation, as they cannot
be captured by the cate!ories of $uantity. ,hat is to say, the differentials themselves escape
representation, but, throu!h their reciprocal interaction, are able to !enerate determinate
representations. .or He!el, neither dy or dx could be understood outside of the ultimate ratio,
precisely because each could only become determinate in relation to the other. .or 0eleu1e, it is
also true that dy and dx only become determinate in relation to one another, but this does not
imply that the differentials are not determinable prior to their determination. ,hey thus stand
outside of the He!elian notion of the ultimate ratio. ,his determinable level is for 0eleu1e a
77
0eleu1e 2+337, p. +<58.
5+
transcendental condition of the ratio which forms the basis of He!el#s account, what he calls
elsewhere the virtual. ,his brin!s us on to the status of the reciprocally determinable, as dy)dx.
.or 0eleu1e, as for He!el, the differential relation is not to be understood in terms of
$uantities or variable ma!nitudes, but a!ainst He!el, the value of the differential e$uation is to be
understood accordin! to its difference from the primitive function 2the function which is
differentiated8. In the e"ample which we used at the be!innin! of this chapter, the primitive
function was the initial e$uation, y 3 x
5
, which was differentiated to !ive, dy)dx 3 5x. -s we saw,
differentiatin! a function !ives us the tan!ent to that function, enablin! us to determine the
!radient at any point. ,o be more precise, however, we can say that the differential of a function
is itself a function which will !ive us the tan!ent at any point on the primitive curve. ,he
function, dy)dx 3 5x !ives the !radient at any point on the curve, as we can insert any value of x
into the e$uation. ,here is thus a chan!e in $uality between the differential e$uation and the
primitive function. ,he primitive function, which deals with the relations of actual ma!nitudes, is
tied to representation, whereas the differential function, which specifies values in terms of dy and
dx falls outside of representation. ,he movement from the differential to the primitive function is
therefore seen by 0eleu1e to be a movement of !eneration, and solves a difficulty he sees in
standard transcendental accounts. Kant, accordin! to 0eleu1e, merely repeats the structure of the
understandin! at a hi!her level. ,his means that ultimately, his account of the transcendental
cate!ories appears empty 2a criticism similar to one He!el also levels at Kant8. 0eleu1e hopes, by
usin! the difference between the primitive function and the differential, to !ive a model of the
transcendental that does not simply repeat the empirical. ,here is also a third sta!e of 0eleu1e#s
analysis, complete determination, which corresponds to the actual values of dy)dx. ,he
importance of this sta!e is that in specifyin! the tan!ent to various points on the curve, we can
also determine the points, known as sin!ularities, where the nature of the curve chan!es. ,hese
points, where dy)dx becomes null, infinite, or e$ual to 1ero, represent points of transition within
55
the structure of the curve itself, as for instance, when the !radient of a tan!ent is e$ual to 1ero, we
have a local ma"imum or minimum as the direction of the curve chan!es. ,his means, for
0eleu1e that the differential calculus allows us both to specify the !eneral, in the form of the
curve as a whole, and as the sin!ular, in the form of the sin!ularity. 9eciprocal determination
!ives us the characteristic of !enerality, in the form of the line which all of the sin!ularities, as
partakin! in it, to!ether determine.
e have dealt with two conceptions of the foundations of the calculus apart from that of
0eleu1e. ,he first approach is He!el#s, which posits that the calculus is fundamentally dialectical,
and as such can only be understood in terms of infinite thou!ht. Bentral to this interpretation was
the idea that the ultimate ratio, as a vanishin!, fell outside of the procedures which a thinker of
finite representation, such as %erkeley, was able to countenance. ,he differential, as a vanishin!,
could not be understood in the static terms today e"emplified by set theory. ,he second approach
was that of set theory itself, which proceeded by re'ectin! %erkeley#s interpretation of the
foundations of the calculus, and thereby seekin! additional resources within finite representation
itself with which to lend consistency to its operation. ,his led to the theory of limits. .or 0eleu1e,
&it is precisely this alternative between infinite and finite representation which is at issue when we
speak of the &metaphysics# of calculus#.
76
e saw this $uestion of representation played out in
terms of the ontolo!ical status of the differential, with He!el preservin! it, and modern
interpretations removin! it. 0eleu1e attempts instead to finesse this problem by !rantin! the
differential a status outside of both forms of representation. In this sense, the differential is
&neither real nor fictive, (instead* differentials e"press the nature of the problematic as such#.
7;
,he $uestion of the differential therefore bears on the $uestion of the two forms of representationA
&Jerhaps the other alternative collapses as well, that between finite and infinite representation#, as
76
0eleu1e 2+337, p. +<;8.
7;
0eleu1e 2+337, p. +<48.
5:
&what is still missin! is the e"tra-propositional or sub-representative element e"pressed in the
Idea of the differential, precisely in the form of a problem#.
7<
,his sub-representational element
cannot be provided by the move to a modern interpretation, as in order to brin! the differential
into the system of representation, it must conceive of it as the limit of a series, and a limit that
has, Llost its phoronomic character and involves only static considerations.M
74
It is for this reason
that 0eleu1e re'ects the kind of interpretation of the calculus which would simply remove its
metaphysical character.
(. Conclusion
e can summarise the differences found in the interpretations of the differential calculus
by notin! that He!el takes a broadly @ewtonian line in his interpretation of the calculus, whereas
0eleu1e takes his inspiration from the work of >eibni1, albeit with a non-intuitional interpretation
of the differential. .or 0eleu1e, the difficulty of differentials appearin! in the resultant formulae
is resolved throu!h the belief that we are dealin! with two different ontolo!ical planes, for He!el
once a!ain throu!h reco!nisin! that the status of the nascent ratio differs from that of normal
numbers. -side from this difference, we should note that what He!el discovers in the calculus is
&the infinity of &relation##,
73
meanin! that what is important in the calculus is the relation between
the two flu"ions, to such a de!ree that this relation not only defines the determination of them,
but also their e"istence. ,he $uestions of the determination and e"istence of the differential
remain separate on 0eleu1e#s interpretation, as the differential is !iven the characteristic of
&determinability.# ,he centrality of these differences is clear in that He!el first refers to the
differential calculus in his discussion of the notion of vanishin! in the fourth remark to the
dialectic of becomin!, and it is this moment of vanishin! which 0eleu1e uses to characterise
He!el#s philosophy as a whole as one of infinite representation. He!el adds movement to the
7<
0eleu1e 2+337, p. +<48.
74
0eleu1e 2+337, p. +<;8.
73
0eleu1e 2+337, p. :+G, n. 38.
57
dialectic by considerin! finite representation &not as havin! vanished and disappeared but as
vanishin! and on the point of disappearin!#.
6G
,hat is to say, He!el does not renounce finite
representation, but instead calls for it to be reinte!rated into the interiority of the infinite.
&=pposition remains abstract so lon! as it does not e"tend to the infinite, and the infinite remains
abstract every time it is posed outside of finite oppositions#.
6+
In this sense, for 0eleu1e, He!el#s
criticism of those who went before him is that they had not taken difference to the level of
absolute difference, the contradiction. In makin! this final move, He!el !oes from finite, or!anic
representation to what 0eleu1e calls infinite, or!iastic representation. =r!anic representation is
!iven content by participatin! in or!iastic representation, 'ust as both bein! and the differential
are maintained as bein! 'ust on the point of vanishin!.
hereas He!el focuses on the structural elements of the ratio dy)dx, 0eleu1e is instead
interested in the difference in kind between the primitive function and the differential, and the fact
that by inte!ration, the primitive function can be generated from the differential. In reco!nisin!
the difference in kind between them, 0eleu1ian ontolo!y returns to the %er!sonian idea that the
absence of order is in fact the presence of a different kind of order, in 0eleu1e#s terms a
distinction which is played out between sedentary 2representational8 distributions and nomadic
2differential8 distributions. ,he differential provides the tools for this, presentin! a transcendental
lo!ic which is capable of e"plainin! the !enesis of ob'ectivity throu!h a non-ob'ectival
2differential8 transcendental field. Everythin!, however, takes place in the middle for 0eleu1e,
and ne!ation and opposition appear in the world &only insofar as these are cut off from their
virtuality which they actualise, and from the movement of their actualisation#.
65
,he fact that
differential relations have a structure of their own, and, throu!h 0eleu1e#s analysis of the
differential as escapin! representation, a structure which itself escapes representation, means that
6G
0eleu1e 2+337, pp. 7:-78.
6+
0eleu1e 2+337, p. 778.
65
0eleu1e 2+337, p. 5G<8.
56
0eleu1e can attempt to provide an account of the !enesis of determination which does not reduce
the absence of determination to indifference.
5;
$i)liography
%erkeley, /eor!e. +335. #e )otu and the Analyst, trans. Oesseph, 0ou!las M. 0ordrecht. Kluwer.
%oyer, Barl %. +363. (he 1istory of the Calculus and Its Conceptual #evelopment. >ondon.
0over.
%oyer, Barl %. +346. A 1istory of )athematics. Jrinceton, @. O. Jrinceton University Jress.
Ba'ori, .lorian 2ed.8. +3+3. 1istory of the Conceptions of &imits and -luxions in Great 4ritain"
>ondon. =pen Bourt Jub. Bo.
0eleu1e, /illes. +337. #ifference and ,epetition, trans. Jatton, J. @ew Pork. Bolumbia
University Jress.
0uffy, Simon. 5GG;. (he &ogic of .xpression/ Quality, Quantity, and Intensity in %pino0a, 1egel
and #eleu0e. -ldershot. -sh!ate.
Eves, Howard hitley. +33G. An Introduction to the 1istory of )athematics. JhiladelphiaA
Saunders Bolle!e Jub.
He!el, /. . .. +343. 1egel5s %cience of &ogic, trans. -. K. Miller. -tlantic Hi!hlands, @O.
Humanities Jress International.
Monk, 9ay. +33<. &as 9ussell an -nalytic JhilosopherN# in, (he ,ise of Analytic Philosophy, ed.
/lock, H.-O., :6-6G. ="ford. %lackwell.
@ewton, Sir Isaac. +3:7. )athematical Principles of Natural Philosophy and his %ystem of the
6orld, trans. Ba'ori, .lorian. >ondon. University of Balifornia Jress.
@ewton, Sir Isaac. +3;7. (he )athematical 6or7s of Isaac Ne!ton. ed. 0. ,. hiteside, @ew
Pork. Oohnson 9eprint.
=U. History of Mathematics Bourse ,eam. +3<7. 1istory of )athematics. Milton KeynesA =pen
University Jress.
9ussell, %ertrand. +37;. A 1istory of 6estern Philosophy/ And Its Connection !ith Political and
%ocial Circumstances from the .arliest (imes to the Present #ay. >ondon. -llen and Unwin.
9ussell, %ertrand. +36;. Portraits from )emory. >ondon. -llen and Unwin.
Struik, 0. O. 2ed.8. +34;. A %ource 4oo7 in )athematics 8299'8:99. Jrinceton. Jrinceton
University Jress.
5<