You are on page 1of 22

The Divided Line and Plato's 'Theory of Intermediates'

Author(s): John A. Brentlinger

Source: Phronesis, Vol. 8, No. 2 (1963), pp. 146-166
Published by: Brill
Stable URL:
Accessed: 18-05-2018 11:28 UTC

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

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

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

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
The Divided Line and

Plato's 'Theory of Intermediates'


IN THIS essay I shall enter into the vexing question of P

of intermediates," and the relation of this theory to the Sun, Line
and Cave section of Republic VI and VII. My thesis is that in the last
7j years or so scholarly opinion has reached a complete impasse, having
veered from one extreme to another, rather in the fashion of an Hegelian
thesis and antithesis; this conflict of opinion desperately requires a
synthesizing "third," and in the conclusion of the paper I try to supply
this ,third.,"

I Plato did hold a theory of intermediates

In his discussion of Plato's philosophy in Metaphysics A, Aristotle


"Further, besides sensible things and Forms there are the objects of mathematics, which
occupy an intermediate position, in that there are nany alike, while the Fornm itself
is in each case unique. " (987b 14-17)

Aristotle commits himself to these distinctions throughout the Meta-

physics, and makes them the basis of his polemic against the systems of
Plato, Speusippus, and Xenocrates (e.g., Z. 1o28b 17-28, A 1o6ga 33-
36). It is hardly necessary to go into the intricacies of an interpretation
of Aristotle's interpretation of Plato in order to accept these distinctions
as true; the overwhelming amount of evidence in the Metaphysics' and
the almost universal agreement of scholars on the matter, makes the
existence of a Platonic theory of intermediates, along with ideas and
sensibles, highly probable. So far as I know only Cherniss2 has denied
that Plato held a theory of intermediates, and even in this case it is not
perfectly clear whether Cherniss denies that Plato held such a doctrine
or only whether he is denying that Plato clearly set it forth in lecture
form to the members of the Academy. In any case the arguments which

I For the evidence, see Sir David Ross, Plato's Theory of Ideas, (Oxford, 1953) pp. 1 51-53.
Hereafter I will refer to this work as P.T.I.
2 Riddle of the Early Academy (Univ. of Calif., 1942) pp. 7 5-7 8.


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
Cherniss gives are of roughly three types: (i) he points out that Plato
never explicitly sets forth a theory of intermediates, and refers to passages
in the dialogues where Plato asserts that the proper subject matter of
mathematical study is the ideal numbers; (2) he points out passages in
Aristotle where the latter seems to contradict himself in dealing with
details in regard to the intermediates; (3) he shows that Plato's principal
students, Xenocrates, Speusippus, and Aristotle, disagree among them-
selves both concerning the subject itself, i.e., the nature of mathematical
entities, and in their views of Plato's theory insofar as it is different from
theirs. Now the ambiguity in the intention of Cherniss' argument makes
it difficult to evaluate; I would certainly agree that points (2) and
(3) render highly suspect both the simple-minded view of an "oral
tradition," which it is the main purpose of Cherniss' work to criticize,
and a literal acceptance of any isolated remark of Aristotle's about the
"facts" of Plato's philosophy. But (2) and (3) provide no evidence that
Plato never held a theory of intermediates; on the contrary, they indi-
cate that there was such a theory, but that it was either internally
problematic or misunderstood by Plato's followers. Argument (i) is
evidence against the Aristotelian distinctions, but can be, and has been,
obviated by the great majority of scholars; the general view is that Plato
developed the theory of intermediates late in life,' and moreover was
primarily concerned with the ideal numbers, and thus may have failed
to mention the mathematical numbers (e.g., in the Seventh Letter or the
Epinomis) because he felt them to be relatively unimportant. Later in this
paper I will give a somewhat different reply to argument (i).

II The relation of the theory of intermediates to Republic VI and VII

Although the existence of a Platonic theory of intermediates thus

seems unquestionable, the relation of the theory to Plato's writings is
highly problematic. For as mentioned above Plato never once explicitly
says that mathematicians study the objects Aristotle describes, while he
does say, by implication, that intermediates are not the proper objects
of mathematics; for example, he asserts that the mathematician wants
to learn of "the square itself" or "the diameter itself" (Rep. gioD),
and then adds no qualification to these remarks. The question as to
whether Plato held the theory of intermediates in the Sun, Line, and
Cave passages of Republic VI and VII has occupied many scholars, both
I e.g., J. Cook Wilson, "On the Platonist Doctrine of &a643Xyrot &pLuo[," Class. Rev.
XVIII (1904) 247-60, or Ross, P.T.I. pp. g9-6o.


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
because of the obvious importance of that section and also because of
its remarkable ambiguity. In the following discussion of the history of the
interpretation of this passage, I will try to make this anmbiguity as clear
and explicit as possible; for it is because of this ambiguity that two
opposed interpretations have arisen, and by a sort of synthesis of it,
as it seems to me, that a more adequate interpretation can be given.'
Past interpretations of these passages, in respect of the question of
internmediates, fall largely into two groups: one grouj)2 hol(ds that the
subject matter of the mathematical ("dianoietic") sciences is the inte r-
mediates as referred to by Aristotle, which I shall call the "old inter-
pretation;" another group,3 which I shall call "nmodern," holds that the
"dianoietic" sciences, in the Republic at least, have ideas as their subject
matter. Each of these interpretations has a certain plausibility, but each
also faces very great difficulties.
The old intepretation has the virtue of taking the Divided Line
seriously as a literal schenia, and of accounting for Plato's remarks
concerning the proportions which hold between the various levels of the
Line, both in respect of the objects or subject matters an(d the correla-
tive "passive states." (Rep. 5 i D-1E, g34A) It has the considerable
disadvantage, however, of being unable to produce any explicit refer-
ences to a theory of intermediates; as a consequence this group makes
much of statements that refer to a single type of mlathematical entity
in the plural, an(d that consequently merely suggest that Plato was aware
that the mathematical sciences require nmany entities of everv type, i.e.,
the intermediates as Aristotle described them. Adamii, for example, leans
very heavily on Rep. 526 A, where Plato asserts that ever) unit is equal,
invariable an(d indivisible, treating the statement as evidence that Plato
was aware that there were many units. But even if such a statement did
I A clear recognition of the ambiguities of these passages is also present in the discussion
of J. A. Natopoulos, "Mlovenment in the Divide(d Line of Plato's Republic," Hlarvard
Studies in Classical Philology XLVII (1943), pP* 57-83. The conclusion of this essay is
rather similar to his, although I disagree considerably with his overall approach to Plato.
2 This interpretation goes back to Proclus, and was the "general view" up to the time
of Henry Jackson (see note 3). I will take as representatives of this view: Janes Adam,
7he Republic of Plato (Cambridge, 1902), e.g., vol. I1, Appendix I to liook VI; and W. F.
R. Hardie, A Study in Plato (Oxford, 1936), pp. 49-65.
3 This interpretation began with the article of FHenry Jackson (On Plato's Republic VI,
so9)1 sqq.)Jnl. of Phil., Vol. X ( i 882), pp. I 32-go), and has continiued up to the present
day. Its most important recent supporters, and the ones whom I shall discuss hiere, are:
F. M. Cornford, "Mathematics and Dialectic in the Republic," in Alind. Vol 41 (1932),
no. I pp. 37-E2, no. 2 pp. 172-I90); ;Sir l)avid Ross, in P.T.I. Pp. 37-83; R. Robinson,
in Plato's Earlier Dialectic (Oxford, 19X3), see esp. pp. 180-201.


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
show that he was aware that mathematicians deal with many numbers or
figures of the same sort, still Robinson is surely correct in arguing that,
while Plato appears to have observed what we might call the "data" for a
theory of intermediates, he simply never goes ahead and explicitly
presents the theory itself; this fact, along with evidence to the contrary,
says Robinson, make the procedure of Adam] and lIardie rash indeed.
And there are two pieces of very formidable evidence to the contrary.
In the first place, there are a nunmber of passages wvhere Plato explicitly
asserts that the dianoietic sciences take ideas as their subject matter.
One of the most important of these occurs in the discussion explaining
the third level of the Line, where we read that geometers "... make
use of visible forms and talk about them, though they are not thinking
of them but of those things of which they are a likeness, pursuing their
inquiry for the sake of the square as such and the diagonal as such
(tOi To pXy(VoOeV0 TOcu 'VsX TO-royo ?6 YOU; 7OLO'JVOL Xoa tcjLpOU MuTfI
(Rep. ioD) This passage, which unequivocally asserts that the
dianoietic sciences seek to know the forms is reinforced by two
other passages. The following one suggests that dialectic and the dia-
noietic sciences deal with the same object, but only in different ways,
and thus that there are no internmediate objects correlative with dianoia:
"And though it is true that those [i.e. the practitioners of the dianoietic
arts] who contemplate themn are compelled to use their understanding
and not their senses, yet because they do not go back to a beginning in
the study of them but start from assumptions you do not think they
have intelligence about them although the things themselves are intelli-
gibles when apprehended in conjunction with a first principle" (vouv
o0x tCEV 7cepi oc&UTX 8OX0956L , O xcL-ot VOt&V OV-tV L?Ta iXpz
Rep. gi I D). The third passage has essentially the same emphasis
as the first; in the discussion of arithmetic, as a part of the education
of the philosopher, Socrates points out "... that it [arithnmetic]
strongly directs the soul upward and compels it to discourse about
numbers as such..." (...c0 ap6po'pa xvcu 7ot &yst T uV i9XJV XOCL rt
CUsCV TCV CapLOPLv cvOCyX'CL 8ELx&cXCyea6ci... Rep. g 2g D). Adam and
Hardie offer different, but equally unlikely interpretations of these
passages, which needn't detain us here; it is obvious that Plato is
referring to the ideal numbers, and holding that mathematicians
study them. These passages, taken together constitute the first ob-
jection to the old interpretation. The second objection is what I
shall call the "explicit exposition" objection. We have seen Cherniss
apply it against the testimony of Aristotle; the modern critics universally


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
urge it against the interpretation of Adam and Hardie. According to
this objection the apparent novelty which a theory of intermediates
would have, along with its importance, creates at once a strong pre-
sumption that Plato would have set it forth with enmphasis at the first
opportunity following its inception. But where would one find a lore
appropriate place than in the discussion of the Line in Republic VI? And
if Plato actually held the theory at the time of writing the Republic,
why shouldn't he have briefly said as much, even once? These difficulties
have seemed decisive to the modern critics, andl the old interpretation
has fallen into disfavor. I will now briefly discuss somle of the more
important of its more recent successors; for it may be that, bad as things
were yesterday, they are still worse today.
When the question arises, in the interpretation of the Line, whether
the metal state 8&CvoLX is correlative with a particular kind of object,
just as vo6ynt corresponds to e1af, 7tLa-nL to Cz, etc., sLxaxc'a to
Fxove, these critics think in terms of two alternatives: either to
follow the suggestion of the diagram taken abstractly, nanmely that 8Lx-
voto has a peculiar object which differs in being and truth from the
other kinds of objects, and thus to take the difficult, or as it seemils,
impossible position of the older critics; or to assume that, in view of
the difficulties of the former position, the proper objects of 8aLoVOLO
are "quantitative" ideas, "universals" such as "triangularity" or "circu-
larity." But having taken this stand (which indeed seenms the only alter-
native open to theml), our moderns are required to show (i) in what
way the nmathenmatical ideas differ from the other ideas, such that they
are properly the objects of &avoLa, and (2) how atavotLo can be
less clear than, and inferior to, Vo-et, even though its object is of
equal ontological rank with the subject matter of v6aL. We shall
see that no satisfactory interpretation is provided for either of these
Ross takes the position that the system of ideas is divided into two
parts, "... a lower division consisting of Ideas involving number and
space, and a higher division not involving these. When philosophy has
done its work, the Ideas which hitherto were only 8tavoYt& have
become vonyi& by derivation from the unhypothetical first principle;
yet they remain different Ideas from those which were from the start
objects of vo6q." The other ideas are always, for him, objects of voii,
the operation of which is exemplified by the discussion of the "greatest
classes" in the Sophist. Now a little reflection shows that Ross's account
is very problematic. First of all, he does not show satisfactorily why


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
Plato selected a particular group of ideas as the objects of aLc"voLM.
He says that they are lower than the others, but we must ask why they
should be. It is not obvious why they should be lower; Robin, for
example, held that they were higher. The suggestion that they "involve
number and space" does not suffice as an answer; they are like the other
ideas both in being themselves non-spatial, and in the fact that their
instantiations are spatial. Moreover, although Ross suggests that the
'quantitative' forms are peculiar in requiring images for their study,
this point can hardly be accepted either. For the study of moral forms
involves the use of images also, if it is not a purely dialectical study
(see below). But it is also clear that Plato never limits the subject matter
of &L%Vvom to the 'realm' of quantity anyway; this level of inquiry
extends as far as the method is applicable, and Ross himself rightly
asserts that the account of tatvoLcx applies in principle to any science
whatsoever which studies ". . . a particular subject matter without raising
ultimate questions about the status in reality of the subject mnatter and
its relation to other subject matters." Thus Ross's position entails that
there is no real distinction of subject matter between a&xvoLoc and vO6Cta,
for even in the case of moral and political philosophy a treatment
on the level of 3m'&votca is possible.' But this is a problematic position
because at the end of Book V, just prior to the passage under consider-
ation, Plato unqualifiedly asserts the principle that a difference of faculty
requires a difference of subject matter,2 while Ross's attempt to establish
a subject matter peculiar to atocvoLo at least implies that he is aware
that Plato held that principle.
But there is a greater difficulty facing those who take Ross's view,
which we can see by asking in what way &cxvotoc is distinct from
VOY)Mq. In the quotation given above Ross asserts that "When phi-
I Indeed the defect which Socrates finds in the discussion of the four virtues, in Book IV,
is that it fails to establish itself in the light of the Good; it is thus presumably on the level
of hypotheses. We can see that this is so when we consider that it uses the perfect state
and man as images, so that its definitions of justice (in itself) actually define justice as
it is in the state and the individual. Moreover the account of the state proceeds
deductively from assumptions regarding human nature; Robert S. Brumbaugh, in Plato
On the One (Yale University, 196I) pp. 195-97, gives an excellent account of the
dialectical juxtaposition of social theories with which Republic Bk. II is concerned.
2 Robinson (P.E.D. p. 200) holds that since in Book VI Plato never calls 8cxvoca a
faculty (&i.o;) but rather always a "state" (7roct'rx), there need not be an object
correlative with it. But this is to mince words. For LT'L-n, which Plato does call a
faculty (477D), is explicitly distinguished from 8Lavocx (533E). aL&t'V0L must there-
fore either be a different faculty, or be a state produced by a different faculty, than

I 5I

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
losophy has done its work, the Ideas which hitherto vere only &LavoOra'
have become vo-r& .. . " but of course this cannot be precisely true:
the ideas do not change. But our conception of the ideas, for examnple
of the idea of square, cannot change either; for according to Ross Plato
holds that geonmeters study the square, and thus have an a(lequate con-
ception of it. If the geonmeter's conception fell short of a perfect one in
an) way whatsoever, then according to Plato he would be dealing with
an image, which is precisely a partial representation of sonme object.'
But if neither the ideas nor our conceptionis of them change, wlhat (loes
change? Ross does not answer this question, but it appears that the onl)
answer open to him is to say that eitlher our attitudes towar(d our con-
ceptions, or our justifications of them, change. On the level of aLxVOLOC
we were, or should have been, doubtful concerning the truth of some
theorem, but after attaining the unhvpothetical lprinciple we cease
being doubtful and become certain, but vet the conceptions involved
in the theorem remain precisely the same. We do not know anything
more about the subject matter with which the theorenm is concerned;
we are simply now able to slhow that we have no reason to doubt the
theoren. But of course this intepretation conflicts with a central
purpose of the Line, which is to distinguish vo6,Lq froml &a'votc in
terms of clearness. For Plato, the geometer dreaims about being, his
mental state is vague, in which only aspects of being appear. The state
of aLavoux relative to vo`aLm is likened by him to the apprehension
of spatio-temporal objects by means of their shadows, and thus to the
apprehension of what is clear and distinct by means of what is vague and
onl) partially determinate. But what happens to this analogy if we
accept Ross's interpretation? It becomes nmeaningless. v6L;, con-
sidered merely as the apprehension of an object or subject matter is in

See Cratylus 432 A-D. Henry Jackson (see p 148 n 3) and J. L. Stocks (Class. Rev. 5
(191 I), 73-88) both suppose that one can have an inadequate, partial conception of an
idea and still conceive of the idea, or have the idea as the proper object of onle's conception.
Of course thlis is in a way true, as I shall show later; but it can also leadi to confusion,
and does so in their case. The Platonic position is that thought is by nature of something;
but if it is partially of something, then it is also adequate to something (else) that plays
the role of image to the former thing. This view lies behind Plato's entire analysis of
images and imitation; for example opinion would be impossible without this view,

for one svould only be able to say that opinions were oi justice or courage in themselves
wvhile of course in reality opinions are of (in the adequate sense, if they are true opinion)
particulars that are becoming, and of (partially) the ideas, but only because particulars
are images of, inmitations of, ideas. But in any case the "modern critics" I amll dealing
with here are free of this misunderstanding.

1 52

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
no way distinct from &aIvoOeO, according to him; their difference lies
not in the clearness or extent or adequacy of their thought concerning
being, but rather in the fact, important to be sure, that a person pos-
sessing v6o1mq is able to show that his state is indubitable, while a
person in the state of 8&xvomo cannot show this.
Now every one of our 'modern critics' holds a position that is essenti-
ally the same as that of Ross on these points, and thus likewise subject
to these difficulties. Robinson holds that dialectic adds nothing to the
conceptions of the real that are appropriate to &a'cvomo except confir-
mation (P.E.D. p. 176); he even goes further and asserts that a person
with only belief, for example a guardian, is capable of apprehending
the same truths that the philosopher knows indubitably (ibid.). But yet
Plato clearly holds that persons with belief are aware of an idea only
in so far as it can be represented bv a spatio-tenmporal entity; a lover of
beauty on the level of belief takes the form of a "lover of sights and
sounds," who delights in p)articular tones, and shapes and works of art,
but who has not "... the power of thought to behold and to take delight
in the nature of Beauty itself." (Rep. 476A) Again, such people are
called dreamers, and dreaming, "... whether one is awake or asleep,
consists in mistaking a senmblance for the reality it resenmbles." (ibid.
476 B). Furthermore, Robinson's more detailed account of the nature of
intuition of the idea of goodness flatly contradicts his position that
dialectic is only confirmatory; he writes that "Because this knowledge
[of the good itself and the beautiful itself] requires many preliminaries,
and yet is direct when it somes, it may suitably be represented by the
language of vision and of the sights revealed to the religious after a
laborious initiation." (P.E.D. p. 176) But if Robinson is right in the
following statement such a representation is miost unsuitable: "Thus
the intuition that is faintly suggested by the Line is not the sort of
intuition in which conception and confirmation occur simultaneously.
The confirmation, that is, the apprehension that the proposition is true,
may be sudden, ... but it occurs long after the first conceiving of the
proposition." (ibid. pp. I76-7) In the second place, Robinson holds
that L&Ovomx has no proper object or subject matter peculiar to it
(ibid. p. 2oo), and is thus liable to the difficulty brought out above,
concerning the principle that different faculties imply different subject
matters, and to a new one peculiar to himself. For in the face of Plato's
assertion that the subject matter of L&OQvoc stands to that of vO'?aL,
as shadows and the like stand to concrete objects, he writes the follow-
ing: "Between the objects of conviction and conjecture the primary

1 53

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
difference, that the) are the objects of different attitudes, remains the
only difference. Ontologically they are identical; for conviction and
conjecture are different ways of getting at the same thing." (P.E.D.
p. I9i) Against this interpretation it is sufficient just to quote what
Plato savs about "the objects of conviction and conjecture": "Would
you be willing to sav," said I, "that the division [between those objects]
in respect of reality and truth or the opposite is expressed by the pro-
portion: as the objects of opinion are to the objects of knowledge so is
the likeness to that of which it is a likeness?" (Rcp. Si o A) Thus concrete
objects and their shadows are different from one another in ternms of
Plato's primary ontological distinction: appearance and reality.
Cornford (see note 6) admits that Plato intends there to be a
distinction between the objects of 8m&voLm and voa4, but denies that
the distinction amounts to an ontological one (p. 38). He holds that
"The distinction of objects is a matter of expediency in teaching..."
The mathematical forms can be represented by visible images, but of
moral forms there are no visible images; "their likenesses in this world
are invisible properties of souls." And this difference, according to
him, is of great pedagogical importance: ". . . it is harder to see the
difference between the justice of a particular action or character and
Justice itself than to distinguish two apples from the number two, also
represented by other visible pairs. Accordingly, mathematics serves as
the easiest bridge from the sense world to the intelligible and should
precede the study of moral Ideas. " (pp. 3 8-9)
Now there are just about as many problems in Cornford's view as there
are distinct points in it. (X) It is not the case that Plato wishes his guardian
to study mathematics before dialectic because it is easier to distinguish
a 'quantitative' idea from its images than a moral idea from its images;
such a view is neither explicit nor implicit in Plato's introduction to the
mathematical curriculum. In this introduction Plato distinguishes two
classes of images and not two classes of ideas; one class presents its
'object' with a superficial definiteness and adequacy, leaving the mind
satisfied, the other gives a confused or paradoxical impression. The
images of mathematical ideas are said to belong to the latter class. But
Plato does not exclude the moral ideas from this latter class; by impli-
cation (i.e., from the example of the finger) the ideas of natural kinds
would be excluded, but it is by no means obvious how nmoral ideas
would fit into this scheme. (2) While there can be little doubt that the
distinctions made in the Sun, Line and Cave passages are very important
for education, Cornford reverses the direction of causal influence;

I 54

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
these passages are concerned to set forth, as much as is possible, the
nature of the highest knowledge, and its subject matter. But it is this
account, and the distinctions which occur in it, which will determine
the character of education, and not the reverse, as Cornford believes.
(3) There are visible (and generally, sensible) likenesses of the moral
forms as well as of the quantitative ones. Many passages could be cited
to support this; here is one: "Then we must not only compel our poets,
on pain of expulsion, to make their poetry the express image of noble
character; we must also supervise the craftsmen of every kind and forbid
them to leave the stamp of baseness, license, meanness, unseemliness,
on painting or sculpture, or building or any other work of their hands. "
(Rep. 4oIA) Thus we may conclude that Cornford offers us no satis-
factory distinction between the subject matter of &t0voL and vo'iGq,
although he rightly asserts that Plato holds that there is such a dis-
But neither can Cornford adequately distinguish between the methods
associated with these states, and evidently just because he finds no real
distinction between them in respect of subject matter. Criticizing his
position on this point is awkward because there is a very great incon-
sistency in his presentation of 8LavoLo. He sub-divides his discussion,
in terms of M'vor.x as a mental state and 8xL&VLo as a movement of
the mind. Now from the standpoint of the former consideration, it
"is the uncertain state of mind of one whose so-called knowledge
consists only of isolated chains of reasoning depending on an assumption
either not demonstrated or not seen to be indemonstrable." (pp. go-gi)
But from the latter standpoint it is the movement downward following
the ascent to the unhypothetical principle, and is thus absolutely certain
and final in its results. (p. 43ff.) Now it is hardly coherent to correlate
an uncertain state of mind (which in its very description requires
reference to a movement that is also uncertain), with a method that is
preeminently sound. But aside from this problem, what can be said for
the view that the method associated with 8xavorC is nothing else than
the downward movement of dialectic? If this were true then 8tLcvoLa,
so far from being a state that is inferior to vosaLq, would be on a level
with it; they would each simply be a part of the state correlative with
dialectic. It would be tedious and unrewarding to continue this survey.
It will have been fruitful if it suggests to the reader that those who reject
the view of Adam and Hardie have very significant difficulties of their
own. Before I conclude this section, and turn to the task of presenting
my own view, I would like to lay out in a more systematic and detailed

I ss

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
fashion the really appalling amount of evidence against the theory that
in the Republic Plato held that the proper objects of the dianoietic
sciences were ideas.

i. There is one explicit, and several implicit references to an onto-

logical difference between the objects of aL&VOm and vo6CGt. The
explicit reference occurs just after an account of the four tocGLaoex
and their proportions to one another; Plato writes: "We had better
not discuss the corresponding objects, the intelligible world and the

world of appearance, or the twofold division of each of those plrovinces

and the proportion in which the divisions stan(d. We might be involved
in a discussion many times as long as the one we have already had".
(534A). Some implicit references are El I C 3-6, gioB 5, 5i iA 4.. But
the single passage quoted is quite enough taken alone to support the
view that the objects of atLavoLa and voacmt are on diff'erent onto-
logical levels. The rejection of this statenment br interpreters displays
the most amazing myopia; the whole passage in which the quotation
occurs is said to be an afterthought of Plato's (by Robinson, Stocks) in
which, to please himself, he set up neat, but meaningless, mathenmatical
relations between the objects of knowledge and their correspondling
states; others either ignore the passage (Cornford, Cook-Wilson), or like
Ross, try to minimize its importance. But the passage stands nevertheless,
and is as clear and explicit a statement as one could hope for. Moreover
it is doubly significant because it implies a recognition that the (lis-
tinctions between tamvoqa& and voyT& and exLxov?; and ~Cp, etc., had
been inadequately discussed in the earlier passages, and gives a reason for
this omission.

2. The myth of the cave, both in its presentation and in the dis-
cussions of it afterward by Socrates, offers evidence for the doctrine of
intermediates, or at least for an ontological distinction between two
kinds of intelligible objects. The following passage is an imiiportant
example: "When there [when out of the cave and in the intelligible
realm], he was still unable to look at the animals and plants and the
sunlight; he could only see the shadows of things and their reflections
in water, though these, it is true, are works of divine creation and come
from real things, not mere shadows of images, thrown by the light of
the fire, which was itself only an image as conmpared with the Sun."
(532B. Another such passage, although less detailed than this, is at
5 i 6A.) Again these passages, so far as I know, are ignored by the modern
critics; but even those who deny that there is any parallelisnm between

I S6

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
the Line and the Cave passages are bound to admit that here Plato makes
an ontological distinction between two kinds of intelligibles, for the
metaphor in which shadows are contrasted to the concrete things which
cast the shadows is one of the most common devices Plato uses for ex-
pressing distinctions of ontological rank. Just prior to the myth of the
cave, in the discussion of the Line, we read that the shadows which are
objects of exLxXcLac, and the concrete objects which are objects of
7t'Re are to one another as the sphere of becoming is to the sphere
of being. It is only natural to suppose that when Plato chose to write
of "divine shadows" and "divine objects," only a few pages later, he
intended the shadow-object relationship to have an analogous meaning.
Other uses of this metaphor where the significance is clearly ontological
are found at Meno iooA, Republic 365C, Phaedo 69B, and Republic
6o2 Cff.

3. "Would you be willin.g to say, " said I, "that the division [within the sphere of becoming]
in respect of reality and truth or the opposite is expressed by the proportion: as is the
object of opinion to the object of knowledge so is the likeness to that of which it is a
likeness ?" (Rep. 5 i o A).
"Your intepretation is quite sufficient," I said: "and now, answering to these four
sections, assume these four affections occurring in the soul: intellection or reason for
the highest, understanding for the second; assign belief to the third, and to the last
picture-thinking or conjecture, and arrange them in a proportion, considering that they
participate in clearness and precision in the same degree as their objects partake of truth
and reality." (Rep. 5 I i D)
"Are you satisfied then," said I, "as before, to call the first division science, the second
understanding, the third belief and the fourth conjecture of picture-thought - and the
last two collectivelv opinion, and the first two intellection, opinion dealing with gener-
ation, and intellection with being, and this relation being expressed in the proportion:
as being is to becoming, so is intellection to opinion; and as intellection is to opinion,
so is science to belief, and understanding to picture-thinking or conjecture?" (Rep. S3 3 E)

These three passages, taken together, assert without any question of

doubt an ontological distinction between the objects of &LcioVoL and
vo6cra. For in the first sensible images are distinguished from sensible
objects by the same principle which distinguished, in an earlier passage
(Rep. 478A), the sphere of opinion from the sphere of knowledge
(namely appearance and reality, or being and becoming). But the latter
two passages assert that 8L&vota and v6qcrt differ because of a difference
in the truth and reality of their objects (a proper translation, by Shorey,
of the passage, c7o-tep np' o0C 9a'v 0X-qr0tLocc tr&tv (Rep. 5I iE)),
which is in turn proportionate to a difference in the truth and reality
of the objects correlative with belief and conjecture.


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
4. When speaking of the difference between the objects of 8atvoto
and VOYaLM it is important to emphasize the term "ontological." For
of course no one denies that Plato thinks there is some kind of distinction
between their objects. Now besides the implications of the passages
quoted in (2) and (3) above, there are at least two other reasons why
this distinction must be an ontological one, or one in terms of levels
or degrees of being or reality.

a. There is no other kin(d of distinction which seems satisfactory.

Distinctions among the ideas, such as those offered by Ross and Robinson
and Cornford, - distinctions anmong 'lower' an(d 'higher' ideas, between
'quantitative' and 'moral' ideas, between ideas of which there can be
images and ideas which can have no (visible or invisible) images, - have
all been shown to be impossible, although they have an initial plausibility.

b. The principle which Plato formulates at such length at the endI of

Book V, namely that a difference of faculty implies a difference in the
object known, surely requires that the objects of 8&4voto be ontologi-
cally distinct from the objects of voncrcs (I have shown - in note 8
above - that these 'states' are correlative with different 'faculties,'
insofar as this expression auvoy.LL is used technicallv bv Plato). Now
it is undoubtedly true that there are difficulties involved in the application
of this principle. One wonders what constitutes a difference in facultv
or on the other hand a difference in the subject matter. But questions
concerning the value or truth of the principle must surely be (listinguish-
ed from the question, Did Plato think the principle valuable and true?
And again when we wonder what criteria determine when the principle
is applicable we must rephrase the question, for an inquiry such as this,
and ask what Plato thought such criteria were. Now so far as I am able
to discover, at least one necessary criterion, for the determination of a
special faculty for a special subject matter, is the criterion of degrees of
reality; this is the only criterion which I can discover in Book V, 47;C-
480 A, and is the only one operative in places such as Timacus 2 9B, where
Timaeus asserts that Xoyot are like their subject matter, onlv likely if their
subject matter is a likeness, firm and unshakable if their subject matter
is truly real and self-subsistent. The modern critics try to underplay the
principle which correlates a faculty and a subject matter; Robinson writes
that Plato "... was imbued with the Greek belief that like is known
by like," as if this were a part of his heritage that he had not quite
managed to shake off. But it will not do to underplay the importance of

I 58

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
this principle for the Line section; valid or not, it is crucial to Plato's
epistemology and in no merely external way.
I would now like to discuss briefly some indirect evidence for the
view that Plato held that 8&wvoLx had the so-called intermediates for
its special subject matter. This evidence, provided by the systematic
requirements of Plato's philosophy, must be handled rather gingerly;
it is evidence for what Plato should have thought, or could have con-
sistently thought, rather than for what he actually did think. Taken in
itself this kind of evidence is of little use; but it is of importance when
coupled with direct textual evidence for a certain theory. In the case
before us it is helpful, even indispensable, to know that Plato's view of
ideal numbers was such as to require mathematical numbers; in the
light of this knowledge textual indications that Plato was aware of the
implications of his theory become more significant. Moreover, a textual
interpretation which conflicts with the "requirements of theory "becomes
immediately suspect, at least in some degree; this is especially so when
one is dealing with a thinker as great as Plato undoubtedly was. It is in
these ways, then, that systematic considerations can be of use in our
interpretation of the Republic passages.
Cook Wilson (see note 3) has shown, in my opinion conclusively,
that the view of the ideal numbers as &UGiX-TOL &pLOpIOL, and the
theory of intermediates, of which Aristotle speaks, are required by
Plato's theory of ideas from the very beginning. According to his
discussion, which is adequate for my purpose, the ideal numbers and
figures are not quantities at all, but are rather qualities, such as square-
ness and triangularity, or twoness and threeness. Aristotle calls these
entities 0a64t53?TOL "incomparable," or in the case of numbers
"inaddible;" they are so, of course, because as squareness, or as twoness,
they have no determinate properties that are usable in mathematical
studies. One does not inscribe triangularity inside circularity: the ideal
circle has no area, no diameter, no circumference, no center; the ideal
square is not a four-sided entity with an incommensurate diagonal. This
is in a way obvious, and in a way rather shocking; it is obvious when we
consider analogous distinctions, such as are made when we recognize
that the idea of man possesses no arms or legs or desires or thoughts;
it is shocking because we are accustomed to think, as do the modern
critics, that for Plato the theorems of geometry and kindred sciences
are about, in a very simple sense of "about," absolute being. Now to
think of the ideal numbers as &,u3XroL (although of course we do not
know that Plato did, at the time of the Republic), is tantamount to the

I 9

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
realization that the subject matter of the theorems of the mathenmatical
sciences must be somehow internmediate, provided of course that one
rejects Mill's view of mathematics. Pythagoras' theorem is not about,
in a simple sense, the ideal triangle; for the idea of triangularity has no
sides and no hypotenuse. The theorenm must then be about "internmediate"
triangles, which are neither nmerely drawn, nor merely conceptual, nor
which come to be and pass away. For if it were about a drawn triangle
it would be false, or what is the sanme from a nmathenmatical point of view,
nmerely approximative; if it were about a conceptual entity it would not
be objectively true; if it were about entities which come to be and pass
away, and conceptual entities would be an example of these, it woul(d be
sometimes true and sonmetimes simplyn meaningless (neither true nor
false). Although the intermediates are imperfect, and nmerely iinlages,
from the standpoint of the ideas, from the standpoint of "becomling"
they have a sort of perfection, and can rightly be called perfect particulars.
They participate in the ideas, but in a different way than drawn or wooden
triangles participate. The latter "share in" trianguLlarity because they
possess, but only approximately, the (leternminate features of the intcr-
mediate triangles, e.g., three interior angles that are equal to two right
angles, etc. But the intermediate triangles share in triangularity just
insofar as they possess the simple, intuitable quality "triangularity, "taken
apart from their determinate quantitative features. But whatever miiay
be the exact nature of an intermediate or of an idea, and of their re-
lationship, it should be clear that Plato's "systematic app1)roach" requires
these two kinds of entities. Of course whether Plato was aware of this
requirement at the time of the Republic is precisely what nmust be decid-
ed. Some think that Plato's silence on the point indicates that it had not
occurred to him. It is time to turn to this theory - which has been heli
by ever) scholar so far mentioned in this article - viz. that the 'dis-
covery' of intermediates is so inmportant, and such an innovation, that
Plato's failure to mention it constitutes evidence that he had not at-
tained it.
Now as I pointed out at the beginning no one denies that at sonme time
in his life Plato 'taught' the doctrine of 'unaddible' numbers and of
intermediates, to which Aristotle so often refers. But we must ask when
he could have 'taught' this doctrine, since in no place in his writings
does he ever mention it. According to the "explicit discusison" theory
he could only have made the discovery after he wrote his last dialogue,
and this is roughly the positionl taken by Ross. Ross holds that the theory
of intermediates is "beneath the surface" in a nunmber ot dialogues,


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
and that Plato "had long been on the point of formulating it," but that
the theory had not occurred to him prior to the publication of the
Seventh Letter, which was written five years before his death, and which
provided a perfect opportunity for him to bring the theory to light.
Therefore Ross evidently believes that Plato formulated the theory
during the last five years of his life, while he was writing the Laws. Now
a number of objections can be brought against Ross's view. In the first
place, if the Epinomis is genuine, as some people think,' then his theory
must be discarded; for the Epinomis was Plato's last work and in it there
is a reference to the subject matter of mathematics (ggo C 5-8), but no
discussion of internmediates. But it is well known that Plato was still
working on the Laws and the Epinomis when he died. Thus if the Epinomis
is genuine the 'explicit discussion' theory must be given up. (In the
Laws there is no section that could be construed as an 'opportunity' to
discuss the intermediates.) Secondly, the theory begins to seem rather
artificial when it is seen to imply that the theory of intermediates could
only have occurred to Plato in the last five years of his life. That such
an important, and in my opinion, obvious consequence of his early
views could have "escaped his notice" over a period of forty years,
especially since in the last twenty-five years of that period he was almost
constantly occupied, if we may judge from the dialogues, with the
theory of ideas and its relation to logic and epistemology and cosmology;
and then that following this period, after he had turned his attention
to semi-empirical political theory, the entire theory of i tX-noL
ciptLO~to' and of intermediates, should have suddenly occurred to him,
seenms highly improbable. Moreover it is safe to say that Plato's theory
of the generation of the numbers, regardless of the particular form it
assumed, presupposed a distinction between ideal and mathematical
numbers (at least such a distinction is presupposed by Ross's account,
and every other one I am acquainted with); therefore it follows, if we
accept Ross's view, that the theory of the generation of numbers was
also developed while Plato was writing the Laws! But the most obvious
short-coming of the 'explicit discussion' theory is that at Republic s34A,
quoted above, Plato (loes explicitly refer to a division among the
"intelligibles," recognizes that they have been inadequately distinguished,
and gives the reason for omitting a discussion of them. Now I have
shown that this division must be in terms of ontological criteria. The

I For a good discussion of the history of the controversy over the authorship of the
Epinomis, see A. C. Lloyd's introduction to A. E. Taylor's translation (Plato: Philebus
and Epinomis. Thomas Nelson and Sons, Ltd. (i 9 56)).


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
reason for omitting the discussion, i.e., that it would be too long, is
surely an acceptable one; the Republic is basically an ethical and political
dialogue, and its excursions into ontology, although significant, can
hardly be expected to be complete. In the light of this explanation the
'explicit discussion' theory merely takes on the form of a criticism of
Plato's ability or judgment as a writer; the critic thinks the theory of
intermediates is so important that Plato should have said something
about it. (Both Ross and Robinson write that if Plato had held the theory
of intermediates in the Republic, then he should have "presented" it
there.) Such a criticism may be just, although I don't think that it is;
but it is nevertheless just a criticism, and hardly provides evidence,
if set over against the passage at s34A, for a denial that Plato held the
theory of intermediates. It seems highly probable to me, although it
would be out of place to argue my opinion here, that the 'explicit
exposition' theory takes its entire origin fronm a view of the nature of
philosophic writing which is alien both to Plato's own theory, and to
his constant practice. The use of a dialogue presupposes that dramatic
unity is essential to philosophic exposition, and there is thus acceptance
of the possibility that dramatic unity and abstract systematic unity
(whatever that might be) could conflict, to the detriment of the latter.
The Republic attempts to solve a problem that is vital to the way of life
of actual persons, and thus requires enough metaphysics to do that, and
no more. The 'explicit exposition' theory, however, presupposes that a
dialogue is simply a vehicle, although to be sure a very agreeable one,
which Plato uses to set forth his latest thoughts; and of course from this
it certainly follows that, if Plato has not presented a theory, then the
theory has not occurred to him.
Sufficient evidence has been brought against the view of the modern
critics to show that it is untenable. This fact, along with the implausi-
bility of the 'explicit exposition' objection to Adam and Hardie, suggests
that we resurrect their view and attempt to put it on a more stable
foundation. Indeed it appears that either we must do this, or conclude
that the text of the Republic is highly confused and self-contradictory.
For while it asserts that mathematicians seek to know "the square itself
and the diameter itself," it also strongly implies, as I have shown, that
the proper subject matter of mathematics is intermediate between forms
and sensibles. Thus the reader seems to have two alternatives: he may
conclude that the Republic is self-contradictory, which is indeed the
implicit conclusion of both the new and the old interpretations I have
discussed; or he may take the view that somehow, Plato thinks the

I 62

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
dianoietic sciences both do and do not take the ideas as their subject
matter, i.e. that in a sense they do, and in a sense they do not. I have
adopted the latter theory, and in conclusion will offer sonme of the con-
siderations which, if correct, would support such a theory.

III Conclusion

To discover what I believe is the correct interpretation of the Line

passage, in respect of the problem of the intermediates, let us turn
briefly to the philosophical digression of the Seventh Letter. There Plato
writes that for everything that exists, there are three necessary instru-
ments by which it can be known, and as a fourth thing, there is the
knowledge itself (342A ff.) The instruments are inmages, names and
%oyoL. Plato argues that each of these kinds of instruments is inadequate
as a means of expressing or setting forth by its mediation the nature of
something (343A ff.), and that knowledge is an immediate relation
between knower and known (344B), although each and every one of the
instruments must be used in the stages of the process that precedes
knowledge (344B, 342A). Thus the knowing process is mediated, but
the final result is unmediated. It is important to note that here Plato
regards every one of the instruments as essential to the knowing process.
One can only suppose that this is because he thinks that each instrument
is adapted to some one aspect of the nature of things, an aspect that is an
essential but nevertheless partial representation of the whole. At least
this is my suggestion regarding the necessary use of these three. A pro-
visional interpretation of knowledge, in the light of these instruments,
could be that the unmediated insight involves a synthesis of the three
aspects of things, whatever they might be, which the three instruments
were singly capable of setting forth; this synthesis would then amount
to an immediate grasp of the thing as one whole. But there should
also be a state of mind which resulted from the use of any one of the
three instruments apart from the others; now it is my view that this use
of XoyoL is &avcm, in the semitechnical sense of the Republic; of names,
is 7C[GLq; of images, eixocaLaC.
If this is correct the very compressed account in the Seventh Letter can
be set out as follows:

known instruments knower knowledge

B ?oyoL &LXVo L'c
Being G names -- 7cL=TL4 ; L vpY
t images - LLXcaLOC


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
These results can be applied to an analysis of the example Plato gives in
the Letter, although only to illustrate the way the "instruments"
supplement one another, and the way each has a positive and negative
feature, and without any pretensions of accuracy in respect of the actual
nature of a circle. The abstract formula of a circle, e.g., "an enclosed
plane figure, with all points on the perimeter equidistant from the
center, " has as its positive aspect the presentation of circularity as eternal
and unchanging, with determinate structural features which allow it to
be related in a systematic way to all other geometric, mathematical,
and cosmological entities. From this standpoint it is a scientific entity,
and thus these truths are hypothetical. In its negative aspect, the ?u6yo4
treats circularity as "many," or as reproducible and universal, thus
neglecting its irreducible individuality as a single form; it also fails to
take account of its causal efficacy, of the fact that all circles depend
upon it for their being as circles, that it is the 'good' toward which
circles must be understood to 'strive,' and with which they must be
compared and judged. The name, which must be taken to symbolize the
public meaning of circularity, as embodied in language, experience,
and technology, in its positive aspect presents circularity as a goal for
action in the crafts or as a 'meaning' derived from experience and more
or less useful in life; because names, as the Cratylus assures us, are at
least partially conventional, circularity loses here its universality and the
systematic relationships provided by science, and also its irreducible
individuality - that is, in the public realm it is merely the agreed upon,
the 'conventional.' Finally, as presented by images, in their positive
aspect, to the individual in sensible immediacy, circularity possesses
the thorough-going opaqueness and selfsubsistence of the absolutely
unique and individual; it is here an object of personal admiration, of
aesthetic enjoyment, of desire. In its negative aspect the image falls
short in both the ways that are proper to names and X6yot. Now
circularity as such, is evidently a self-subsistent individual, with determi-
nate relations to the other ideas, and with causal relations to space-time
circles and human life; to know it in itself, is evidently to know it at
once in all of its meanings and appearances.
"Then, is it not true, in the same way, that we and the Guardians we are to bring up
will never be fully cultivated until we can recognize the essential Form of temperance,
courage, liberality, high-mindedness, and all other kindred qualities, and also their
opposites, wherever they occur. We must be able to discern the presence of these
Forms themselves and also of their images in anything that contains them, realizing that,
to recognize either, the same skill and practice are required, and that the most insigni-
ficant instance is not beneath our notice. " (Rep. 40 2 D)

I 64

This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
Now to many, no doubt, this interpretation will appear radical, and
even fantastic. But I offer it only provisionally, as something which
seems useful in understanding Plato, and also because it seem-s to me to
allow for a meaningful account of the varied aspects of Plato's thought,
of the fact that he accords imlportance to 'instruments' as diverse as the
myth of Er and the hypotheses of the Parmenides. When Plato introduces
this brief account in the Letter he renmarks that these are distinctions
he has used many times before; one is thus led to suppose that this highly
compressed exposition somehow epitomizes his thought in a way he
regarded as very significant. My interpretation is intended to do justice
to that remark.
When we turn to the Republic, Bk. 6 and nmake use of this interpre-
tation, the following picture presents itself. The inquiry turns to the
nature of the highest knowledge, and the account given takes three
forms, the first corresponding to the namle, the second to the ?o'yo4,
the third to the image. First Goodness is presented in terms of its causal
efficacy, both for knowvledge and being, by means of an experienced
object and its experienced properties, i.e., the sun and daylight and
darkness, in their relations to the eyes and ordinary objects; secondly
it is presented by means of a diagram, which 'categorizes' all types of
knowledge and being, thus giving an abstract scheme of the relations
of Goodness to all knowledge and objects of knowledge; thirdly, there
is the image of the cave, which presents Goodness as a self-subsistent
individual, the final object of desire, enjoyment, and contemplation.
The application of the account in the Letter to these passages should
be obvious; by thus using these three 'instruments' Plato at once shows
us how seriously he takes them, and how flexibly and analogically he
intends them to be understood.
The implications of this interpretation for the theory of intermediates
should also be obvious. In the light of it, assertions to the effect that
mathematicians as mathematicians study "the square itself and the dia-
meter itself" are seen to be true but yet only partially true. I have shown
in what way the mathematician's Xo?yoL both succeed and fail to grasp
an idea as a whole. They deal with the form as a determinately structured
part of a hypothetical system, and thus their enterprise would be meaning-
less if it were not about ideas; but on the other hand, squares and circles,
as these are understood by geometry, are not ideas, and their structural
properties belong to them as representations of, as being other than,
ideas. Thus as long as the nmathematician is concerned with solutions to
particular problems, or proofs of propositions, within a systematic


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to
framework; and failing to question his presuppositions, accepts his
definition of circle (or whatever it may be), and the theorems based on
it, as an adequate and complete account of the idea of circle, i.e.,
circularity; then he may also be rightly said to "dream of being," and to
be actually concerned with an image, an intermediate, although one
which is much closer to the natuLre of circularity than wooden or painted
circles. Thus in a way, for Plato, mathematical knowledge is wholly
adequate, and forms a completely autonomous discipline; in this light
its subject matter is the intermediates, circles, squares, the perfect
solids, and these are treated as being themselves complete and adequate.
But this state is really a "waking dream, " and consists in taking an inmage
to be a reality (Rep. g33C, 476 C); in a sense its subject matter is its
own, but in another sense it is radically dependent upon individual,
self-suLbsistent ideas, of which it has never even dreamed.

University of Massachussetts, Amherst, Mass.


This content downloaded from on Fri, 18 May 2018 11:28:03 UTC
All use subject to