Division and Its Relation to Dialectic and Ontology in Plato

Author(s): J. R. Trevaskis
Source: Phronesis, Vol. 12, No. 2 (1967), pp. 118-129
Published by: BRILL
Stable URL: http://www.jstor.org/stable/4181798 .
Accessed: 17/08/2013 18:58
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

BRILL is collaborating with JSTOR to digitize, preserve and extend access to Phronesis.

http://www.jstor.org

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

Divisionandits Relationto Dialecticand

in Plato
Ontology
J. R. TREVASKIS

The

formal divisional exercises which we meet above all in the

Sophist may strike the reader as tedious. Yet it is usually said

that Plato lays great store by Division as a method of philosophy,
one, moreover, to which he gives the title of 'dialectic' and which
reveals the real structure of Ideas.
I wish to discuss how far the method is to be identified with dialectic,
what relation, if any, it bears to Plato's ontology, and what Plato
hopes for from it. I shall be mainly concerned with Phaedrus, Sophist
and Statesman, having discussed the Philebus on a previous occasion
(Phronesis 5,1 [1960], 39-44).
1. How far is Division to be identified with dialectic?
It is usual to see the first announcement of the method in the Phaedrus.'
One thing that may be noted at once is that the Phaedrus example is
not dichotomic, not, at least, in the manner of the divisional schemes
of the Sophist: the method briefly illustrated at Phaedrus 265 d ff. does
not involve the successive rejection of one side of a single column of
dichotomies. The Phaedrus example involves the division of the
concept 'irrationality' into 'human' and 'divine'; and each of these
two subdivisions is pursued, at all events according to the theoretical
summary of 266a, until it yields its quarry.
Socrates declares himself (266b3) enamoured of the method, and a
disciple of its successful practitioners. Furthermore, he says, he calls
A moment later (c8)
such persons, whether rightly or not, &Lx?cXeLxoL.
his respondent remarks (without strict accuracy) that in his opinion
I 265dff., resumed briefly at 277b. D. J. Allan, it is true, has said: "I should be
prepared to maintain that the reference to division and collection in the Phaedrus

presupposes

the fuiller account of these processes in the Sophist..."

(Philosoph1y

XXVIII [1953] 365). But, as will be seen, no fuller account of collection and
divisioin is, in my opinion, given in the Sophist. Both dialogues give a certain
amount of explanation of the method, the Phaedrus here at 265dff., the Sophist
at 218dff. (the illustrative example of the Angler). Neither presupposes the
other.

118

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

Socrates was right to call the method

8LX3XTrXOV.

Thus we have the

position that Plato here in the Phaedrus is prepared to call both the
method and its practitioners 8Lockx'cxo.
What exactly are we to understand him to mean by this adjective
here? Does he, for example, mean that he calls successful practitioners
of the method 'practitioners of dialectic'? and does he therefore equate
the method with 'dialectic'? The answer to these questions seems to be
'no': if Socrates is able to express doubt whether or not he is justified

in calling the method's successful users 8Lt-XOL,

and if Phaedrus is

able to express approval of the method's being called 8LixrXtlXOV (and

hence acknowledgement that the question admits of doubt), the method
cannot itself be equated with 8?xW%rLx'in Plato's mind. Dialectic must
be taken to have an independent meaning for Plato, over and above
the method of Division.
In so far as Socrates's doubt is seriously felt by him, it reflects, then,
his realisation that he may be mistaken about the method's dialectical
nature, that is to say, about its being the best one for his purpose, that
of enquiry into Ideas.2 At 266c8, where the method is called 8&aXewtx6v, what Plato intends would seem to be 'appropriate to philosophical
And if we are looking for a single adjective to cover the
discussion'.3
word as applied both to practitioners and method here, 'philosophical'
is perhaps the best available. At all events, we cannot argue from the
Phaedrus that dialectic and the method of Division are to be equated.
When we turn to the Sophist we find seven dichotomic divisions
concerned with defining the sophist, and a section (253d-e) which
Cormford declares4 to be a r6sum6 of Collection and Division. Perhaps
the first question which we should ask is whether we are meant by
Plato to consider Division as outlined in the Phaedrus and as practised
in the Sophist to be essentially one and the same method. Plato does
not tell us explicitly, but, in view of the similar terminology employed
in the two dialogues, it would be implausible to maintain that the two
methods are to be distinguished.
It looks, therefore, from this brief glance at the Phaedrus and the
Sophist as though Cornford rnight be right when he says: "The method
of Division may be used for two distinct objects: (1) the classification
of all the species falling under a genus in a complete table, or (2) the
2

Cf. Hackforth, Plato's Phaedrus, p. 134, n. 4.

3 Cf. Richard Robinson, Plato's Earlier Dialectic2, pp. 69ff. and Hackforth,

p. 135 init.
4Plato's Theory of Knowledge, p. 264 init.

119

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

definition of a single species only."5 For all that, Comford's object no.
(1) could not be fully justified from the Phaedrus. It is not there
has been laid bare, only
envisaged that the whole articulation of ,uavEcx
sufficient for the purpose of distinguishing the two sorts of love.
Nonetheless there would seem to be nothing in principle to prevent the
method's being used for the division of a genus into all its species; and
Plato probably made few bones about the sort of distinction drawn by
Cornford. The essential thing about the method was that it worked by
dividing concepts.
which enables
rn
At Sophist 253 d 1-3 we hear of a aLOCXSXCLX a'MrG
Is this process the method of
its possessor xm'tocy'v- 8cxLpeSaOac.
Division? I do not think so. And if such a doubt is justified, the passage
lends no support to the identification of the method of Division with
dialectic.
253dl-e2 may be translated as follcws:
Elean: To distinguish between concepts, and not think that one is
another and vice versa, falls under dialectic, shall we not agree?
Theaetetus: Yes.
Elean: Well, a man who can make such distinctions readily perceives
a single Idea extended over the whole field of many instances, though
each instance is quite separate, and many different Ideas embraced
from without by a single Idea; and, again, a single Idea linked in a
unity extending over many whole Ideas, and many Ideas completely
separate and distinct. This is to know how to determine by their classmembership to what extent things can and cannot combine.
Cornford believes that the Elean's longer utterance refers to Collection
and Division, the first half of the main sentence referring to Collection
and the second half to Division, though Cornford is clearly uneasy
about the second half.6 Most readers will probably share Cornford's
uneasiness, I think, particularly over the words xod. no&aq X(OPLq
7r0V'rn

&LG)oatesvOC4

(253d9).

If the words of d8-9 ("and, again, a single Idea linked in a unity

extending over many whole Ideas") refer, as they seem to do, to the
relation of a generic Idea to subordinate Ideas, then the words of d9
("and many Ideas completely separate and distinct") follow oddly if
6 p. 171. The Phaedrus might then be taken to exemplify object (1), and the
Sophist object (2).
6 p. 267. The second half is "less easy to interpret". It "appears to describe"
Division.

120

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

they refer, as Comford would have it, to those same subordinate Ideas.
One would expect, from the whole shape of the sentence, that these
words referred to some different group of Ideas.
I believe that the words do refer to a different group of Ideas and
that we should hesitate before supposing that 253d is a summary of
Collection and Division. Two further reasons may be advanced to
support such hesitation:
1. The Elean claims at 253c6-9 that we have "chanced upon" the
area of the philosopher's ability in our desideration of a skill that shall
empower a man to judge "which concepts go with which and which do
not" (bll-cl).
We have unexpectedly arrived, that is to say, upon a
new scene. The philosopher's ability, then, which the Elean goes on to
describe at dl ff., must surely be something different, not simply the
technique of Division which Theaetetus and the Elean have tried six
times already in the dialogue. As is said at e8-9, we can look for the
philosopher in this area now or later. On Cornford's view we have
explored that area already.
2. How odd it would be for Plato to have employed the method six
times in the dialogue already, and then, and not till then, described it,
without any reference to its previous use. The atmosphere of the
passage, on the contrary, is one in which we are dealing with something de novo.
The argument of 252 e-253 e seems to me to go like this: some terms
or concepts go together so as to produce true7 statements, some do not.
It is like what happens with the letters of the alphabet or with musical
notes: only certain combinations succeed. But in order to decide which
these alphabetical or musical combinations are, a man needs to be
literate, or, in the other case, a trained musician. Similarly, a man who
wishes to demonstrate which concepts are congruent and which are
niot, and whether there are uniting concepts which make combinations
possible and other concepts which cause disjunction in statements in
which concepts are disjoined, will need a particular type of knowledge,
indeed the supreme type of knowledge. In fact, such a man will need to
be a philosopher; for the ability to discriminate concepts falls under
dialectic, which the philosopher alone can practise. The possessor of the
ability to distinguish concepts can recognize the class-concept lying
7 It is important to be clear that wveare concerned with true statements, not
statements that are merely formally legitimate like 'grass is black'. Plato tells
us at 252d6-10 that a statement which won't do is 'change rests'; yet this
statement is of entirely legitimate form. It simply happens to be false.

121

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

behind a group of related individuals, and the generic concept lying

behind a group of related concepts. Again, he can recognize when a
number of concepts can be unified under one generic concept, and when
a number of concepts are in no way related to each other. This dialectical ability is the capacity to understand conceptual relations8
correctly.
One could, I think, paraphrase the argument of the section as follows:
a person who is to say what letters make words and what do not, must
possess the skill denoted by the term 'literacy'; literacy being the basic
underlying knowledge of which spelling, in any particular instance,
will be merely an application. Similarly musicianslhip underlies any
individual composition: "the skill conveying the knowledge which
sounds combine and which do not" (253b2-3) is the 'underlying' skill
of musicianship possessed by the pLouacxo'.In the same way a particular
type of knowledge (b9) underlies the correct combination of concepts
and the correct recognition of manipulatory, as distinct from manipulated, concepts. The knowledge referred to at 253b9 and c4-5 is to
be distinguished from the particular applications described at b 10-c 3.
In order to know which concepts can go with which, and to distinguish
the different functions performed by different concepts, we must know
all about concepts, must have the basic knowledge about concepts
which Plato at 253d2-3 calls 'dialectic'. The literate man, then, will
be able to spell; the musician will be able to compose; the dialectician
will be able to see the correct relationships between concepts.
Before leaving the passage I should like to discuss a little further the
lines 253 d5-9. They fall naturally into two sections, the first ending at
7CCPLSXOX[1VOM (d8). It seems fairly clear that the two parts of this first
section refer respectively to (a) the relation of an Idea to its particular
instances" (b) the relation of a number of related Ideas to their generic
Idea. The two parts of the first section are juxtaposed antithetically:
,uv....

8x

=to?v

and 7ro'...

Utro6xl.

The second section

also, I

think, contains an antithesis, this time between a unified many and

8 Does gxma-xmat e 1 refer to particulars or to
y6vj ? I think it refers to both, and
that the relations which Plato has in mind include those of particulars to one
another and of concepts to one another.
9 Mr. W. G. Runciman appears to be correct in his view (Plato's Later Episteat 253d6 refers to particulars,
mology, p. 62), as against Cornford's, that iroXX&v
not Ideas. Cornford, however, will not have been guilty of a mere grammatical
error. Cf. PTK 189, n. 2 where he shows that he recognizes as neuter a word
he translates as though it were feminine.

122

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

an ununified many: [LUrov...aL'... 7rto?Xv and 7toX?OcXCP'Lp. The first

section spoke of the dialectician's ability to classify related particulars
and concepts; the second, I think, speaks of his ability to distinguish
between those groups of concepts which are related and those which
are not.10
But however that may be, the lines cannot, it seems to me, be
satisfactorily interpreted as a summary of Collection and Division.
The first section, it is true, could represent Collection; but the second
section simply will not naturally fit Division. Perhaps the words 70o
of 253 dl have misled readers so as to expect a
xcxt yzvi aLOCLpELOOCL
description of Division. Yet the immediately following words (dl-2)
could hardly be taken to support that expectation. They support,
rather, the meaning 'to distinguish between concepts' for the words
just quoted.
I conclude therefore that dialectic, as briefly characterized at
253 d 1-9, is not to be equated with the method of Division. The lines,
together with their immediate context, only claim that Ideas are the
touchstone by knowledge of which one may avoid confusion and
achieve precision in philosophic debate."
2. The relation, i/ any, between Division and Plato's ontology
This is a question on which opinions are divided ;12 but it is usually
assumed that Division is essentially concerned with Ideas. According
to Cornford, for instance, "The task of philosophy is regarded in the
Sophist as mainly analytical - the mapping out of the realm of Forms
in all its articulations by Division".'3
Hackforth, in his Additional Note on Collection,'4 has already argued
against "what is, I think, the commonly accepted view, that Collection
is... always Collection of kinds (species)..." I wish to attempt something of the same sort so far as Division is concemed, and to argue
I thus partially agree with A. E. Taylor (Plato, the Sophist and the Statesman,
p. 157) who believes that "the two problems mentioned in the second part of
the sentence are meant to be distinguished from those spoken of in its first half."
The details of Taylor's interpretation, however, I do not find convincing.
II Similarly, the theory of Ideas would, in Plato's opinion, answer the problem
of predication posed at 251 a-c, the beginning of this whole portion of the
dialogue.
12 Cf. Ross, Plato's Theory of Ideas, p. 239 fin.

10

13 PTK
14

183.

Plato's Examination of Pleasure, p. 142.

123

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

against the view that Division is concerned only with species or Ideas.
The first passage one might turn to in such a discussion of Divisioni
is Phaedrus 265-6. Yet no certain light on our problem is, I think, to
be derived from it. Admittedly pavtcxand the types of it distinguished
in Socrates's two speeches may be presumed to be Ideas. But this is of
little help towards determining whether we must always expect to find
Ideas reflected in divisional schemes. It is true that we are counselled
xan' etan a8.TpCV xaTJ &pOpa
h7riuxev (265e 1-2); but we can certainly
place no reliance whatever on Plato's use of the term el8oq, which
varies in meaning from one context to another. It may, but it equally
may not, designate an Idea. Nor, I think, is xcx-r'&pOpx7;recux.v any
more to be relied on. A division of a genus into its constituent species
is doubtless xOCT'&pOpx.But I fancy that Plato would recognize other
'natural' divisions within a class which were not to be identified with
Ideas.'5

Before leaving the Phaedrus, however, I should like to draw attention

to the passage at 269cff. where Socrates says that the orator should
discriminate the different types of soul and fit to them different types

of address. The orator must know fuxn 6ac C87 ?XtL... TOUTVa 8 8
8py??vwv,
,
oL6V8
X?'y&voK t6ax xai rosa scv es8N,yx%ov
(271 d 1-5). Thus +uyx' which we know (e.g. fronmthe Phaedrus Mytlh)

ou&

Plato does not regard as an Idea, is subject to &LaLpeatq

into et8j.
Indeed Hackforth does not blink the fact, but says in his commentary
on p. 151: "The upshot is that we are back again at dialectic"' as the
right ,uCsOoao;
for the orator: he must discern 4u ' as at once a One and
a Many". It seems to me that we have a prima facie case here of Division
applied outside the scheme of Ideas, and that the case needs answering
by anyone who holds that Division is essentially conicerned withi Ideas.
It may be urged, conceivably, that the &a(peaLqof Phaedrus 271 d
is not that of 265eff., but is a mere casual distinction of types of mind.
And even though this would, in my opinion, carry little convictioni, we
may perhaps move on now to the Sophist where there can be no doubt
that we have seven schemes of Division. An important point for us to
recall at the outset is that we have argued that Sophist 253d is not
Cf. n. 24 (p. 128) below.
By which term Hackforth intends the method of Collection and Divisioni:
cf. Plato's Phaedrus, p. 134.
For the irreducibility of Soul and the Ideas cf. G. C. Field, The Philosophy of
Plato, p. 131: '"The Forms are not souls, and Soul is not a Form, and neither is
'made' by or dependent for its existence on the other."
15

16

124

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

concerned merely with Division, but with the whole inter-relationship

or, where appropriate, lack of inter-relationship, between Ideas. Now
it is as plain as anything ever is in Plato, particularly from the words
Ls
136xv

7MOX,E)v,SoV exoarou

&a

zeLyzvOU

X()plq,

7r&VWrf8LCXr?VTLe'V,

together with the fact that the context is concerned with a a.)cexrtx'
E7rLG7tWL, that 253d is concemed with Ideas. So long, therefore, as
253d was thought to be a summary of the method of Division, there
was a strong incentive to suppose that the method of Division was
essentially concerned with Ideas. The equation, too, of the method of
Division with 'dialectic' was a similar incentive, and we have seen
reason to abandon that equation also. We are therefore free to look at
the seven Divisions by themselves and to try to determine from them
whether they are concerned exclusively with Ideas.
A question we are bound to ask ourselves at the beginning is whether
Plato posited an Idea of sophistry. He does not seem to make it clear
one way or the other. Cornfordmight appear to suppose there is such
an Idea when he says at p. 170: "The purpose of the dialogue is to
define 'the Sophist'. Here, at the threshold, we cross the boundary
between the sensible world, to which the Theaetetuswas confined, and
the world of Forms." Yet at p. 173 he is able to say: "Sophistry is the
false counterfeit of philosophy and of statesmanship and has its being
in the world of eidola that is neither real nor totally unreal." This
sounds very unlike an Idea. And, whatever conclusion one comes to
about 'sophistry', what is one to make of the intermediate classes
revealed in the course of the Divisions? We may perhaps set aside for
this purpose the first six Divisions which, in Cornford'sopinion, are
not definitions of 'the Sophist', but "analytical descriptions of easily
recognisable classes to whom the name had been attached" (p. 187).
But the seventh cannot evade being taken for a definition of 'the
Sophist', and, in Cornford's translation, it runs in part as follows:
"The art of contradiction-making, descended from an insincere kind
of conceited mimicry, of the semblance-making breed, derived from
image-making..." (p. 331). Are we to believe, then, that Plato posited
an Idea of 'semblance-production by ignorant mimicry"7 (;u(i>nat
oootO,vqrLxm),

267e 1)? I find it difficult to believe that he did.'8

17

Cf. the final Table of Division, PTK 324.

48cff. where 'self-ignorance' is divided into three, and one of the
resultant concepts, 'imaginary wisdom' (8o,oaocptx, 49d11), is further subdivided, according as it is entertained by the weak or the strong. Can we regard
either of these subdivisions as the name of an Idea?
"8 Cf. Philebus

125

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

Certainly when Plato gives us unequivocal instances of Ideas they

bear no resemblance to classes of this sort, but instead have simple
names like 'just', 'equal', 'man' or 'table'. Yet if 'semblance-production by ignorant mimicry' does not represent an Idea, we are at
once relieved of any need to regard Division as essentially concerned
with Ideas.
When the question of the population of the world of Ideas arises,
recourse is commonly had to Rep. X, 596a, where Socrates is made to
say: "we are accustomed to posit19 a single, Idea for each group of
individuals to which we give the same name". Yet, even taking the
usual view of that passage's meaning, it cannot be relied on to represent Plato's settled opinion: for such an opinion is in effect rejected
at Statesman 262d3-6 where those are adversely criticized who think
that 'non-Greek' "must constitute one real class because they have a
common name 'barbarian' to attach to it".20 Thus the mere existence
of a common name is no guarantee that there exists an Idea with that
same name.
The Statesmian represents itself as a continuation of the Sophist, and
little fresh light on our problem can perhaps be expected from its
dichotomies. But it is well recognized that it is in the Statesman in
particular that we are given advice on right and wrong modes of
employment of Division; and from them we may hope to derive a
certain amount of enlightenment. Not that we can get much help
from 262b: T'O popOq czo e8oq exye'. This gets us no further than
Phaedrus 265el. Nor, I think, areroc TC&V7r('VCvOv a'LOZELOCof 278d any
more helpful: if we are debating whether the "universal alphabet of
things",21 that is to say, the components of definitions by Division, are
invariably to be considered to be Ideas, we must try to determine the
questioni by an inspection of the components of divisional definitions.
We cannot, that is to say, argue from the words Ot& -)5v 7rx&v-v
ar'oLxZexto what they 'must' stand for. Again, the etan,jiepriand
yevj of 285a-b seem as ambiguous as before.
There is one passage in the Statesman, however, wlhich does, I think,
cast light on the problem whether Division always reveals a patterni
of Ideas. This is the passage briefly referred to above where we are
19 A. C. Lloyd suggests
in CQ N.S. II, 1-2 (1952), 106, n. 1 that our hypothesis
is merely provisional. But it would seem unwise to build much on the word

TLOCCOOL.
20
21

J. B. Skemp, Plato's Statesman, p. 132.

Taylor, Plato, the Sophist and the Statesman, pp. 288-289.

126

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

counselled against a Division of the term Man into Greek and Barbarian (262dff.). It would be better and a truer bisection to divide
mankind into male and female.
Three points may be observed from this example:
1. We have a clear case here of Division proceeding below the in/ima
species Man. It is commonly said22 that Division stops at the in/ima

species. Plato's own expression in the Phaedrus is ?LCpr.-TOUw-pi-qrou

(277b7) 'to the limit of division', and in the Sophist &'To[oV 7U&V
(229 d5-6) 'at the limit of divisibility', neither of which expressions seems
to have any necessary connotation of species, whether in/ima or not.
It seems that we have no warrant for attributing the anachronistic
expression in/ima species to the lower limit of Platonic Division, and
that we should not do so.
2. Plato seems to have no objection to Hellenic, Lydian etc. as subdivisions of mankind. His objection is to Hellenic etc. versus the rest,
because in such a split only one side (Hellenic etc.) is a y6vo4 as well
as a p?poq (262e6-263a1). So we have a case here in Division of a subdivision (Hellenic etc.) of which Plato approves, but which does not
reflect an Idea: for it is hardly conceivable that Plato would have
supposed the superficial differences represented by human racial
groups and sub-groups to be based on separate Ideas.
3. The division of human beings into men and women does not
support the view of the correct method of Division held by, for
example, Hackforth (PEP 24-25) in his discussion of Plato's example
of Division taken from musical notes and scales (Philebus 17bff.).
Hackforth objects to Plato's example: "The terms 'high', 'low' and
'level' are not the names of species of sound, which can be further
divided into sub-species. ."23 We might be similarly tempted to object
that the species 'human being', which contains both male and female,
is not divided by Plato into sub-species (e.g. European, Asiatic etc.)
themselves containing both male and female, and further divided into
sub-sub-species (Hellenic etc.) also containing both male and female.
Instead (we might continue) Plato comes to a halt with his arbitrary
division into the two sexes, a division which is not distinctively human.
Yet Plato clearly tells us that he approves such a division. Once again
22 E.g. by Hackforth, Plato's Phaedrus, p. 132, n. 4; PEP 24-26. Cf. Cornford,
PTK 186: "The method of Division exhibits ... indivisible species at the bottom."
23 Hackforth similarly objects at PEP 114 to "a classification of arts or sciences
which are not, in fact, co-ordinate species of a genus, but whose relation is one
of greater or less approximation to truth (&O?xi,Lm)
or precision (&xppertoc)".

127

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

it seems that we should abandon any preconception in favour of

'specific' differentiae. Plato is willing to see different criteria observed
in different divisions, even where the same class is to be divided. This
is what we have found with the class Man, where Plato seems willing
to approve a division either by sex or by race,24 so long as genuine
yevl and not mere ,ukp]are come by.
From this review of Plato's use of the method of Division and of his
comments on its correct use, I think we may conclude, in opposition to
the generally held view, that Division is not essentially concerned with
the splitting of a genus into its constituent species (although of course
it may do this); that it need not stop at the in/ima species; that it is
not essentially concerned with Ideas (though of course it may produce
class-names which are the names of Ideas) and does not aim to produce
an 'ontological map'. The seven divisions of the Sophist, indeed, seem
to have little concern with Ideas.
3. What does Plato hope for Irom the method of Division?
It is commonly claimed that Division aims to reveal a real structure
of Ideas. Hackforth, for example, says (PEP 21): "the right number
[of classes], which the philosopher aims at finding, corresponds to the
Forms really existing... that is why dialectical method is 'quite easy
to indicate, but very far from easy to employ'."
Now that we have seen reason to differ from this view, we must
revise our estimate of Plato's hopes of the method. The method's use,
in the Sophist above all, is to describe by a process of elimiination; or,
again, to elicit different meanings of ambiguous terms, of which we
may take as an example 'love' in the Phaedrus. Of these two uses the
second is by far the more important: the first is largely satirical, and
the formal dichotomies of the Sophist and the Statesman have long been
recognized as only half-serious. But the second use, the eliciting of
different meanings of ambiguous terms, is of extreme importance and
Plato brings it to our notice a number of times: as he says in the
Phaedrus, words like 'just' and 'good' are disputed terms: they mean
different things to different persons; 'love' is similarly a disputed
term which requires definition (263a-d), a process which Plato accommodates to a divisional scheme at 265dff.; or as at Sophist 218c
f 7r6cuzevequally
Such a division by race would, it seems to me, be xocT'&pOpca
with a division by sex. Yet only the latter division would reflect a structure of
Ideas.

24

128

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions

the word 'sophist' is in need of explication so as to make sure that it

means the same thing to both speakers, a process which formally
occupies the seven divisions of the dialogue.
But, over and above this important use, there is a third which is
clearly supereminent. It is referred to in essence at Phaedrus 266b
where Socrates justifies his enthusiasm for the method of Division by
its conferring on him the ability 'to speak and to think'. This benefit
is only fully explained, I think, at Statesman285d, where the Elean is
made to suggest that their chief aim in their present discussion is not
to define the statesman but to improve their general powers of philo7nav'7oc &LocXrcLXWrpOLq yLyVea0oL, d5-6).
sophical discussion (ro5 7tcepL
In this way they will train their minds and make them better able to
apprehend Ideas, which is the end of all their discussions (286a4-7).
The method is thus a practice routine in philosophy.25We need
neither be surprisedat its dullness comparedwith the flights of Plato's
imagination elsewhere, nor expect any direct philosophical results
from it. It proceeds by an oblique, not a direct, path towards Plato's
goal of the Idea.
University of Adelaide
25This conclusion therefore bears some resemblance to Prof. Cherniss's view
which does not see Division as providing an 'ontological map' (cf. ACPA 46:
"The schemata of diaeresis for Plato, then, do not portray the relational
arrangements of the world of ideas but rather are instruments of analysis.")
Cherniss, however, believes that when Plato counsels sZ ,ukpoq&t el8oq 'XrTCo
he intends by cl8oq an Idea (cf. ACPA 252 and 264). It does appear, therefore,
that in Cherniss's view Division is very closely concerned with Ideas.

129

This content downloaded from 161.45.205.103 on Sat, 17 Aug 2013 18:58:34 PM

All use subject to JSTOR Terms and Conditions