Professional Documents
Culture Documents
In this book, Vasilis Politis argues that Plato’s Forms are essences, not
merely things that have an essence. Politis shows that understanding
Plato’s theory of Forms as a theory of essence presents a serious
challenge to contemporary philosophers who regard essentialism
as little more than an optional item on the philosophical menu.
This approach, he suggests, also constitutes a sharp critique of those
who view Aristotelian essentialism as the only sensible position:
Plato’s essentialism, Politis demonstrates, is a well-argued, rigorous,
and coherent theory, and a viable competitor to that of Aristotle.
This book will appeal to students and scholars with an interest in the
intersection between philosophy and the history of philosophy.
VASILIS POLITIS
Trinity College Dublin
www.cambridge.org
Information on this title: www.cambridge.org/
: ./
© Vasilis Politis
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published
A catalogue record for this publication is available from the British Library.
---- Hardback
Cambridge University Press has no responsibility for the persistence or accuracy
of URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
Acknowledgements page ix
Introduction
Why cannot the ti esti question be answered by example
and exemplar?
Hippias Major
Why cannot essences, or Forms, be perceived by the senses?
Hippias Major. Phaedo. Republic
Why are essences, or Forms, unitary, uniform and
non-composite? Why are they changeless? Eternal?
Are they logically independent of each other?
Phaedo and Republic
The relation between knowledge and enquiry
in the Phaedo
Why are essences, or Forms, distinct from sense-perceptible
things?
Phaedo and Republic V. –
Why are essences, or Forms, the basis of all causation
and explanation?
Phaedo –
What is the role of essences, or Forms, in judgements
about sense-perceptible and physical things?
Republic VII. –
vii
Bibliography
General Index
Index Locorum
My gratitude extends, above all, to John Dillon, The Trinity Plato Centre
(which, as it happens, is located in Trinity College Dublin) and all those
who work there and have every day been providing me with a lived space
that we have been shaping jointly. I had the pleasure and privilege to
experience a similar environment during the winter of in The
Durham Institute of Advanced Study, and to it and its excellent directors,
and Nicholas Saul not least, I am grateful. At the outset of this project
I had two wonderful stays, first at Wuhan University in the summer and
autumn of , made even better by my host there, Hao Chanchi, and
his hospitality, and then at Uppsala University in the spring of , made
just as enjoyable by my host and friend there, Pauliina Remes, and her
hospitality. In both places, I enjoyed many seminars with excellent grad-
uate students. I had the opportunity to present individual talks on one or
the other item in this book in several places in China, Europe and at São
Paulo University and Campinas University in Brazil. I thank my friend,
George Karamanolis, my host at Vienna University, and my friend, Klaus
Corcilius, my host at Tübingen University. I am also grateful to my friend
and host Richard King during my time at Bern University. Several uni-
versities in China provided more extended points of refuge during the past
four years and the writing of this book, most especially Wuhan University,
Renmin University and China University of Political Science and Law.
I am grateful to you all. Many people have contributed generously to this
book, and I fear I will forget some, but I will not forget Matthew Adams,
Keith Begley, Nicolò Benzi, David Berman, Lesley Brown, Friedemann
Buddensiek, Damian Caluori, Laura Candiotto, Luca Castagnoli, Nick
Clairmont, Niall Connolly, Klaus Corcilius, Giulio Di Basilio, Filomena
Di Paola, John Dillon, Nathalie Ek, Paolo Fait, Ge Tianqin, Zuzanna
Gnatek, Margaret Hampson, Vivil Valvik Haraldsen, Verity Harte, Blake
Hestir, Thomas Hodgson, David Horan, Martin Jacobson, Jiao Liming,
George Karamanolis, Kate Kiernan, Richard King, Hermann Körner, Inna
ix
Plato sometimes uses eidos and idea interchangeably, when talking about Forms or about essences
(such as Republic V. a; and, perhaps, Euthyphro d–e); but not always. He sometimes uses idea
when he says that an eidos is, precisely, an idea (Greek term) that is always the same. (See, e.g.,
Parmenides a– and b–c. In passages such as Republic VI. b–, and perhaps even
V. a, it is not clear whether he uses idea for eidos or, rather, in this other way.) When he uses
idea in this way, it would not be right to translate ‘Form’ for idea; we may translate, rather,
‘character’ or ‘quality’. Plato’s point will then be that a Form is, precisely, a character or quality
that is always the same. Remarkably, this shows that Plato does not simply assume that a quality must
be always the same in every one of its instantiations or occurrences.
The inclusion of the Timaeus and Philebus, in addition to Phaedo-Republic-Parmenides, may raise
some eyebrows, since it is generally thought that they are much later. My reason for including them
is that, as I will argue, they are, in regard to the theory of Forms, continuous with the Republic, and
appear to be intended by Plato as so being. Of course, once we include the Timaeus and Philebus, we
ought, ideally, to consider, in regard to Forms, also dialogues generally thought to be later than the
Parmenides but earlier than them; such as, especially, Theaetetus, Sophist, Statesman, as well as
Phaedrus and Symposium wherever we place them in the relative chronology overall and indeed in
relation to each other. It is simply that doing this would take us too far . . . and we would need many
more words than Cambridge University Press admits.
Thus Annas (, ): ‘It is often said that Plato has a “Theory” of Forms and even that it
dominates his entire work. In fact Forms appear rarely and are always discussed non-technically; they
answer to a variety of needs which are never systematically brought together . . .. If we ask “What are
Forms?” we find a variety of answers.’ If the argument of the present study is on the right lines, this
statement is the opposite of the truth.
I recommend Christopher Rowe’s salutary opposition to this tendency: ‘Talk of “versatility” is in
danger of suggesting that we can retreat into interpreting each dialogue on its own (as some scholars
in the last two centuries have attempted to do), and there are too many connections between them,
too many constants, to make that a viable proposition’ (, ). Rowe goes on to spell out the need
to read the dialogues together.
Harte () begins (–) by stating, as basic characteristics of Forms, the following four, in
which the notion of essence does not figure: . Forms are primary beings; . they have causal
responsibility; . they are privileged bearers of certain terms and . they are objects of a knowledge of
a privileged sort. Later Harte characterises Forms as being essences, when she says (in regard of
Phaedo c–d): ‘Socratic questions ask “What is F?” for some range of properties. The Form is
identified as “What is [F]” – that is, as the referent of the answer to this Socratic inquiry’ (). See
Rowe (, ) for a similar characterisation. Rowe is especially clear and explicit that the whole
idea of Forms goes back to the Socratic ti esti question. See Chapter .
Silverman’s The Dialectic of Essence () promises to bring Forms closer to essences, but in fact
Silverman argues against the view that a Form is identical with its essence, or that Forms are essences,
and he considers Forms to be, rather, the ‘bearers’ of essences. I come back to this issue in the
Conclusion of the present study – which the reader is welcome to read in advance.
Chapter is a revised version of Politis (a).
It is a consequence of the thesis defended in the present study that developmentalism, in regard to
Forms and the theory of Forms, is mistaken; and it is mistaken irrespective of whether by
developmentalism we mean the view that Plato’s claims about Forms in such dialogues as Phaedo
and Republic are inconsistent with his claims in dialogues we consider to be earlier, or, on the
contrary, we mean the view that Plato’s claims about Forms in such dialogues as Phaedo and
Republic, though consistent with his earlier claims, introduce new claims that are not continuous
with, or based on, or justifiable on the basis of earlier claims. (For a recent defence of
developmentalism in regard to Forms, see Dancy (). For critical assessments of
developmentalism in general in regard to Plato, see Annas and Rowe (), also Kahn ().)
For the conception of metaphysics that is associated with questions about essences and what things
are like essentially, see the classic paper by Kit Fine ().
Chapter is a revised version of Politis ().
That Plato is not committed to this principle of causation (i.e., that causation works by causes
transmitting their character to their effect) was argued, in regard to dialogues before Phaedo, by
Malcolm (, –, –). However, the principle is now commonly (e.g., by Sedley [])
invoked in the account of Plato’s argument in Phaedo e ff.
The following statement from Oderberg (, ), however incredible, is characteristic: ‘There is
also Platonist essentialism, but I do not count this as real essentialism according to the sort of
position I defend.’
Addendum
I want to note that the thesis I am defending in this study – it says that
every item in Plato’s theory of Forms can, Plato thinks, be derived from
the posing of the ti esti question of certain things or qualities, and from
holding out that this question can be answered adequately and truly – is
intended to be compatible with the view that, for Plato, the knowledge of
(certain) Forms requires a sort of intellectual vision. The thesis I am
defending implies that the knowledge of Forms requires a verbal account
(a logos) of what something is. But it does not imply that such a verbal
account is sufficient for such knowledge. Notably, the thesis I am
defending in this study implies that, if the knowledge of (certain) Forms
requires a sort of intellectual vision, then this vision can be arrived at only
through engaging – discursively and dialectically – with verbal accounts;
Phaedo b says that one who knows is able to give an account, logos. If we suppose the logos
includes an account of what the thing known is, this says that knowing what a thing is is necessary
for knowledge; it does not say that knowing what a thing is sufficient for knowledge.
For this reason, I am sceptical of Rowett’s () recent claim that, after early dialogues in which
the request for definitions is prominent, Plato throws out this request.
Myles Burnyeat and Michael Frede (a; b).
There have been many responses to Geach’s charge that Plato dismissed
out of hand the possibility of definition by example and exemplar – that is,
the possibility of answering the question ‘What is F?’ by pointing to, or in
some other way calling to mind, a particular thing that is conspicuously
F and thus capable of serving as a standard for a thing’s being F. What
critics have missed, remarkably, is that Plato casts a particular character as
an able proponent of a cogent defence of definition by example and
exemplar. When, in the Hippias Major d, Socrates asks Hippias
‘What is beauty?’, he answers by invoking, as standards for a thing’s being
beautiful, and hence as proper answers to the ti esti question, first a
particular girl, then a particular horse and a particular lyre, of exemplary
beauty. He urges that, since such examples and exemplars of beautiful
things are ready to hand, no reasoning, argument or enquiry is necessary
for answering this seemingly momentous question, which is, on the
contrary, ‘something trivial and worth practically nothing’ (e–).
I confess that, since I was struck by this some time ago, I have found it
remarkable that critics, apparently without exception, have missed this
Geach (, –). The literature in response to Geach is large and seemingly never-ending.
Representative examples include: Anderson (); Santas (); Irwin (, –); Vlastos
(; ); Woodruff (; ); Reeve (); Fine (; ); Prior (); Wolfsdorf
().
Balaudé (, ff.) is the only critic I know who recognises that Hippias puts forward the girl as
an example and exemplar of a beautiful thing: see , , . With the exception of Woodruff’s
() judicious and insightful commentary, the literature has not been kind to this dialogue, even
setting aside the issue of authenticity. For a long while critics were agreed that Hippias simply
misunderstands Socrates’ ti esti question and in doing so confuses particulars and universals. This
view is an old one, eloquently formulated by Apelt (, ) more than a century ago, who thinks
Hippias’ confusion is rather between things, or qualities, and essences: ‘Also was ist das Schöne?
Wohlverstanden das Schöne und nicht, was ist schön. Dieser Unterschied dünkt dem Sophisten so
fremd wie böhmische Dörfer . . . Meine [i.e., Socrates’] Frage aber ging auf das Wesen der Schönheit,
auf den Begriff derselben’. Nehamas (b) did much to dispel this approach but then himself
defended an incredible reading of how Hippias understands the ti esti question; and he refused to
admit the crucial point, which is that Hippias defends answering this question by example and
exemplar and that this is the target of Socrates’ argument.
Kahn () is perhaps the main living exponent of the denial of authenticity. For a comprehensive
summary of the history of the question of authenticity, see Ludlam (, ff.). For a recent
defence of authenticity, see Petrucci (). For the question of the relative dating of the Hippias
Major, see Malcolm (); Lee (a, , n. ).
Note the di’ ho (‘because of which’) at a, which refers to the girl as an exemplar and a standard
of beauty; and note further that this di’ ho picks up on the explicative datives in the explanations of
the form ‘It is on account of the F that things that are F are F’ at b–d. I am taking it that these
explanations are intended to mean: ‘It is on account of the standard for a thing’s being F that things
that are F are F.’
Commenting on this passage from the Parmenides, Harte (, ) says: ‘Forms are not needed in
those cases where things are “just as we see them to be”.’ Republic X. a– is sometimes read as
implying that there is a Form for every quality. For why this reading is not necessary, see Harte
(, ), when she says that Smith () ‘proposed that the passage should be construed,
rather, as making the claim that we commonly assume, “[as a rule of procedure,] that the Idea which
corresponds to a group of particulars, each to each, is always one, in which case we call the group of
particulars by the same name as the [Form].” On this construal, the passage does not carry any
implication about the scope of Forms’ (square brackets are Harte’s).
Examples are aplenty, a recent exemplary one being Wedgwood ().
The reader may wonder how Plato can provide both an ‘epistemological’ and a ‘metaphysical’
argument for the same conclusion. I answer this in Chapter , ‘What yokes together mind
and world?’
This is one reason why I cannot agree with Broadie () when she argues that we can, and should,
study the Timaeus on its own and separately from other dialogues of Plato.
Contra Rickless, who appeals to this passage in the Hippias Major, a–b, to argue that: ‘Socrates
moves to the claim that sensible F things are con-F (that is, partake of both the F and the con-F)
from the claim that sensible F things appear to be con-F. Since the Greek word for “appear” in this
context is ambiguous as between a veridical and a non-veridical sense, Socrates’ move from the one
claim to the other strongly suggests that his use of “appear” here is veridical’ (, , n. ,
original emphasis). It is clear that Rickless intends this to be true not only of Socrates in this passage,
. This is what is behind the claim in the Phaedo that essences and
Forms cannot be perceived by the senses
It is familiar that it is in the Phaedo that Plato states, loud and clear, that
what certain qualities are, their essences and Forms, cannot be perceived by
but of Plato generally. For he is arguing that Plato’s (so-called) argument from co-presence of
opposites is not limited to how physical things appear, but is about how physical things are.
I consider this issue in Chapter .
I consider this more closely in Chapter .
My reason for translating ‘that which a thing is’, and not ‘that which is’, for ho esti here is due to the
reference to Socrates’ practice of asking questions and receiving answers, since this is a practice
consisting of, or centrally involving, the asking of ti esti questions.
Vlastos (, ) comments on these lines as follows: ‘For Plato the definition of a concept is “the
account of the essence” (logos tēs ousias) of its Form, Φ’. Therefore, according to Vlastos, that which is
defined is not the Form, but the essence of the Form. Vlastos defends his statement by appeal to this
passage, Phaedo d, in his translation of it: ‘That reality itself, of whose essence we give the account
[hēs logon didomen tou einai] when we ask and answer our questions . . .’. I beg to differ. The genitive
here (tou einai) can, and I think should, be read in such a way that it is precisely the ousia that we give
an account of when we give an account of what it is (tou einai). This is most important, because, on
Vlastos’ translation and commentary, this passage confirms the common view, which is indeed
Vlastos’ view, that Forms are substances that have essences, rather than simply being essences. This
passage in no way supports this view.
See Taylor (, –): ‘It is assumed [in Phaedo] that the technical language of the theory of
Forms is so familiar a thing that Socrates needs to warn the lads not to be misled by it: an odd
representation if the whole theory has been invented by Plato after Socrates’ death.’ See also Rowe
(, ): ‘[Plato] writes, in the Phaedo and the Republic, as if “forms” are already a familiar topic,
so in any case projecting them back into the pre-Republic dialogues’ (he explicates the ‘in any case’
with ‘That is, if they were not there already’, note ). And see esp. (, ), where Rowe says:
‘So far as the Phaedo is concerned, the forms are those things, whatever they are, that the
philosopher seeks to “grasp” through dialectic; the very things that Socrates is seeking to grasp in
the Euthyphro, the Laches, the Meno . . .. That is one reason why Plato can have him, in the Phaedo,
describe “beautiful and good and all such being (ousia)” (D–) as “the things we’re always
chattering about”, and then later refer to them just as “those much-chattered-about (poluthrēleta)
things” (B–)’ (ellipsis in original).
See Chapter .
In Republic VII. b–, Socrates expressly points out that this is not the worry he had in mind.
I think it does contain it; see Chapter , where I consider this passage and the argument in it.
See Chapter .
This is to suppose that the Hippias Major is written before the Phaedo, but in relative proximity to
it. This supposition is defended by Malcolm .
I have argued, in Chapter , that the argument does not extend, and is not intended by Plato to
extend, to all qualities.
It is notable that, in Republic VII. d–a, the issue is the use of visible celestial patterns as
exemplars and standards, paradeigmata, of, precisely, equalities and the adequacy of such sense-
perceptible exemplars and standards of equal (see esp. the isōn at a).
This understanding is summed up succinctly by Verity Harte (, ): ‘Is the claim (PC) that,
for any particular perceptible, having some relevant feature, F, necessarily, that particular
perceptible also has the opposite feature, un-F? Or is it the claim (UC) that, for any perceptible
type, a token of which is F, for some relevant feature, necessarily that type has un-F tokens also?’
I note that, as Harte formulates the claim of the co-presence of opposites, the notion of a standard
(paradeigma) is not part of that argument.
There are three such issues. First, is Plato here criticising the mathematicians for treating of things,
including sense-perceptible things, as unitary, without them asking, with any serious intent,
whether these things genuinely are unitary and partless? (This is Burnyeat’s reading (
[originally, ], –); I am inclined to agree with him.) Secondly, is Plato committing
himself to the view that things that are only one and not also many are necessary for defining
numbers or even are that which numbers are? This question is left open by the fact, which is evident
from the text, that he is committing himself to things that are not both one and many, for defining
numbers. Thirdly, if Plato is committing himself to things that are only one and not also many (and
not only to things that are not both one and many), are these things supposed to be Forms or are
they supposed to be something distinct from both Forms and sense-perceptible things, namely,
numbers so understood? (The latter is what Aristotle thinks Plato, or perhaps rather, certain other
Platonists, supposes numbers to be: neither Forms nor sense-perceptible things, but a third kind of
thing.)
. That Forms cannot be perceived by the senses does not imply
that they can be known only a priori
I have argued that the reason why Plato thinks that the Forms of certain
qualities cannot be perceived by the senses is, purely and simply, that he
thinks that what these qualities are cannot be specified by example and
exemplar. It is, therefore, wrong to suppose, as critics commonly have, that
Plato’s reason for this claim, the sensory imperceptibility of Forms, is that
he thinks that Forms are non-physical things; and only physical things can
be perceived by the senses; and non-physical things can only be known
a priori.
I want to end with what seems a remarkable and significant conse-
quence of this conclusion. For, if this is why the Forms of certain qualities
cannot be perceived by the senses, namely, because what these qualities are
cannot be defined by example and exemplar, then we cannot infer that the
Forms of these qualities can only be known a priori. This is not, of course,
to question that, for Plato, Forms are objects of the intellect and reasoning
(noēta), it is to question that Plato’s distinction and opposition between
I consider this closely in Chapter .
It may seem that he commits himself to this view in a–. But since in these lines he is reporting
what certain people, apparently philosophically inclined mathematicians, would say, it is hard
to tell.
Harte (, –) refers to Plato’s Forms as ‘theoretical entities’; and she says: ‘where there is
no theoretical work for Forms to do, there is no reason to posit them’.
For this last idea, see Chapters and .
Bostock (, ) says: ‘The new method [i.e., the method in Phaedo ff.] is supposed to be
one that does not rely on the senses: it is meant to proceed wholly a priori and without the benefit of
any lessons from experience.’ He responds to Plato: ‘I see no contradiction in the supposition that
there could be a fire that was not hot, or a lump of snow that was not cold. But I imagine that Plato
would not have agreed on this point, and he would have thought that we can tell a priori that fire is
hot and snow cold’ (; original emphasis).
Cf. hēmeteran ousan [hē ousia], ‘being ours’, at e–; i.e., the Forms, or essences, referred to as hē
ousia, ‘essence’ or ‘being’, are characterised as ‘being ours’.
If the reader is sceptical about such chronological suppositions, they do not really matter to my
point; what matters is the different dialectical functions, and different audiences, of
different dialogues.
In the Phaedo, when Socrates is at the end of his life, this familiarity is underscored. In the Republic,
it is intended that Glaucon is well versed in their particular method of argument (see, e.g., e–).
In the Parmenides, even if Socrates is presented as a young man, the theory of Forms is presented as
familiar to all involved in the discussion.
. Plato’s reasoning for the claim that Forms are unitary, uniform
and non-composite
It appears, therefore, that Plato’s reasoning for the claim that Forms are
unitary, uniform and non-composite is this:
P A Form is that which is designated by an adequate (hence general,
unitary and explanatory) and true account of what a thing is.
P If an account of what a thing is is adequate (hence general, unitary
and explanatory) and true, then that which it designates is unitary,
uniform and non-composite. (= The relation between unitary accounts
and unitary things)
Therefore,
A Form is something unitary, uniform and non-composite.
That Plato holds Premise we have already gone some way towards
showing, and to demonstrate this is the aim of this entire study. In regard
to Premise , we shall see that there is reason to suppose that Plato holds it.
We have reason, therefore, to think that this represents Plato’s reasoning
for the claim that Forms are uniform, unitary and non-composite.
It might be objected that the cogency of this argument is rendered
questionable by what might seem to be the following suppositions in it:
first, the word ‘unitary’ is used in the same sense when it is predicated of an
account as when it is predicated of that which is designated by an account;
secondly, the same unity is in question in both; thirdly, the nature of the
unity of that which is designated by an adequate and true account can be
directly derived from the nature of the unity of such an account. I want to
make plain that the argument does not depend on any of
these suppositions.
Of course, we would like more concrete textual basis for Premise .
I believe we have such a basis, in the following passage from Phaedo:
I have argued for this supposition in The Structure of Enquiry in Plato’s Early Dialogues, Part I (Politis
). It is shared by many critics.
In and through Chapters and .
Harte . I note that Forms are introduced shortly after d–a, as the contents and objects of
certain logoi.
This is an important qualification. Plato should not be understood to think that any logos should
immediately be grounded through a Euthyphro dilemma; he should rather be understood to think
that certain logoi need, in the first instance, to be based on a logos tēs ousias – a real definition – and
that such logoi tēs ousias need to be grounded through a Euthyphro dilemma.
Critics do this in a variety of different ways, and these ways would, ideally, have to be distinguished,
but, for present purposes, I shall mention a number of the relevant critics without regard to the
variations among them: Stenzel (, ); Cornford (, –, –); Ryle (;
originally ); Hamlyn (, , ); Prauss (), with a good response by Gadamer
(, –; , ). Bambrough (, ) does not seem to say much, but refers to
Crombie (): ‘Mr. I. M. Crombie has usefully remarked that in these later dialogues, and
especially in the Philebus, there is a more sophisticated awareness than in the earlier dialogues of the
nature and importance of the relations between universals and universals as opposed to relations
between two artificially divided worlds of universals and particulars (An Examination of Plato’s
Doctrines, Vol. II, pp. ff ).’ Moravcsik (, ); and, most recently, M. L. Gill (,
introduction, ).
It may be said that the passage says that Forms appear (phainesthai) many and that this does not
imply that they are many. Correct. But the passage says that Forms manifest themselves
(phantazomena) in communion with one another, with actions, and with bodies, and that, as a
result of this, they appear many. It can hardly be denied that Plato intends this to imply that they do
commune with one another, with actions, and with bodies. But, if the passage says that Forms
commune with one another, it cannot, it seems to me, be read to imply that they are logically
independent of each other. Or, if we want to be extra cautious, it is incumbent on the critic who
thinks that, in Phaedo and Republic, Forms are logically independent of each other, to explain what
it means to say that they commune with one another; and to do so in a way that is consistent with
the fact that when, in the Sophist, he says that Forms commune with one another, he intends this to
mean, or directly imply, that they do stand in logical relations with each other (especially relations
of implication and incompatibility). I know of no such explanation. For this issue, and for the
translation of the passage, Gould (, ; with the Greek transliterated by me in this note) is
instructive: ‘Waterfield, taking the koinōnia as real, implies that the eide are immanent in sensible
particulars; Grube, treating the koinōnia as an appearance, suggests that the eide (while separate
from sensible particulars) are unrelated to one another. Either the koinōnia is real or it is apparent. If
real, then the Forms are related to particulars as they are to one another. If apparent, then they may
be as separate from one another as they are from sensible particulars. To adjudicate this would
require a thorough interpretation of Plato’s middle theory. Without going this far, one can yet argue
for the superiority of Grube’s translation of this passage. The crux of the matter thus is whether the
eide appear everywhere (a) because of their association with actions, bodies, and one another or (b)
as though in association with actions, bodies, and one another. How, then, are we to take the
complex subordinate clause? Let us note that it has a men/de construction. Grube’s translation
projects this more clearly, for it highlights the contrast between hekaston einai and phainesthai
hekaston. To opt for (a) is to overlook the parallel between the einai in the straightforward men
clause and the phainesthai in the more puzzling de, a parallel to which Plato draws our attention by
his use of the men/de structure and the contrasting pair of infinitives.’
This is how Nettleship (, ) understood the two passages: ‘in the Republic . . . each form is
itself related to other forms [N. is evidently referring to the a passage here], and ultimately all the
forms of things are connected together and make one system [N. is evidently referring to the c
passage here]’. Nettleship’s ‘system’ for ‘kosmos’ here is notable and, I find, entirely defensible.
. Plato’s reasoning for the claim that Forms are changeless
If we observe that changelessness is argued to be a feature of essences, and
that an essence is that which is designated by an adequate and true account
of what a thing is, and if we suppose that Forms are, basically, essences,
then we may propose that Plato’s defence of the claim that Forms are
changeless is this:
P A Form is that which is designated by an adequate and true account
of what a thing is.
P If an account of what a thing is is adequate and true, then that which
it designates is changeless.
Therefore,
A Form is something changeless.
The question, therefore, is why and on what reasoning Plato thinks that
Premise is true. There is a simple answer, and there is a complex answer
to this question. The simple answer may seem immediately compelling,
but on reflection there is, by Plato’s own lights, a particular objection to it.
The complex answer builds on the simple answer, but obviates
the objection.
What is the relation, in Plato, between the account of knowledge and the
account of enquiry? Is the account of knowledge independent of the
account of enquiry? Taking up this question is a large task, not least
because, while so much work has been done on Plato’s account of
knowledge, and quite a lot is being done on Plato’s account of
enquiry, I know of only the odd critic who has considered the relation
between the two. It is remarkable that critics have generally treated the
two topics – Plato’s account of knowledge and his account of enquiry – as
if they were separate topics, which suggests they have been tacitly sup-
posing that, for Plato, the account of knowledge is independent of the
account of enquiry.
The following is what is at issue, stated in general terms. To think that
Plato’s account of knowledge is independent of his account of enquiry is to
think that, according to Plato, it is possible to give an account of what
knowledge is that makes no reference to enquiry. To deny this is to
propose that, according to Plato, first, knowledge comes about through
enquiry, either as its end-state or in the course of the enquiry (or in both
In addition to Politis (), I am thinking, most especially, of Hintikka, who writes: ‘there does not
exist, and there cannot exist, a fully self-contained theory of justification independent of theories of
discovery’ (, ). Hintikka takes this view to be Platonic in origin and orientation: ‘Hence it is
seen that Plato had in one important respect the same focus as we: the quest for knowledge rather
than the justification of beliefs. The definition of knowledge was thought of by Plato as a means for
this quest’ (; and see esp. ch. ). One might have expected Fine () to have, at the very least,
raised this question, in a book on a problem about enquiry in the Meno, and in view of her expertise
on Plato’s epistemology. But she does not.
I am thinking of such critics as White (); Ferejohn (); Taylor (); Wedgwood ().
They typically focus on Plato’s account of knowledge and treat of his thinking about enquiry, if at
all, largely in relation to the so-called paradox of enquiry in the Meno. I note, however, that Taylor
argues that, for Plato, knowledge is the product of enquiry: ‘Those themes may be summed up as the
doctrine that the aim of inquiry is to achieve systematic understanding of the intelligible principles of
reality’ (, ); indeed, he says that ‘Given the metaphysical theory of Forms as the basic things
that there are, Forms are the primary objects of inquiry’ (, emphasis added).
My impression is that critics have tended to put more emphasis, by far, on the earlier epistemological
claim. A good example is Bostock (), who says a great deal about the earlier, and hardly
anything about the later, epistemological claim.
Bostock () is a good example of both ways. Notable also is how scholars writing on Aristotle on
science take this view of Plato’s epistemology for granted, which presents Aristotle as the innovator.
Thus Karbowski () rehearses it as being familiar that, for Plato, sense-perception is only a cause
of and a stimulus for the ascent to rational and a priori knowledge and is not part of the justification
that constitutes rational knowledge.
See Rowe (, ch. ), for a good defence of the view that the overall aim of the Phaedo is to argue
that the good person, and especially the person whose goodness is based on philosophy, has reason to
be cheerful and hopeful in the face of death.
koinōnein with dative; also ‘through’, dia with genitive, and ‘with’, meta with genitive, and ‘by’, using
simply the instrumental dative. All these locutions are used, apparently, without distinction.
For the relation between ti esti questions and whether-or-not questions in Plato, see Politis .
See Chapter .
I shall properly consider this passage later, in Chapter .
kath’ hoson dunatai: e, c, c (kata to dunaton). Remarkably, and as Dillon (, ch. ,
–) perceptively remarks, this little, but important, qualification is there in the reference to
‘becoming like God in so far as that is possible’ (homoiōsis theō[i] kata to dunaton; Dillon translates
‘assimilation to God’) in Theaetetus b, and likewise in the similar reference in Republic X. a
(homoiousthai theō[i] eis hoson dunaton athrōpō[i], ‘to become like God as much as is humanly
possible’).
I thoroughly agree with Rowe (, –), when he says: ‘Nowhere in the Phaedo is there any
commitment, on Socrates’ or anyone else’s part, to the idea that the objects of knowledge (the
forms) will be available on demand, as it were, after death, even to the “purified” soul. The myth
makes no provision for it, and in fact doesn’t mention forms at all; and Socrates on his own account
associates access to, or approaching, true reality – the unseen, aïdes – specifically with this life, not
the next.’ And I agree no less, when he says: ‘My conclusion about forms in the Phaedo is that this
dialogue does nothing to alter the picture we seem to get from other parts of the corpus: that access
to forms, and to knowledge, for human beings, is by philosophy – dialectic – alone. There are no
short cuts, even after death. So far as the Phaedo is concerned, the forms are those things, whatever
they are, that the philosopher seeks to “grasp” through dialectic; the very things that Socrates is
seeking to grasp in the Euthyphro, the Laches, the Meno . . .. That is one reason why Plato can have
him, in the Phaedo, describe “beautiful and good and all such being (ousia)” (D–) as “the
things we’re always chattering about”, and then later refer to them just as “those much-chattered-
about (poluthrēleta) things” (B–)’ ().
In the Phaedo (a–c), Plato argues that the Form of the quality equal is
distinct from any pair of sense-perceptible things that are equal. The
conclusion of this argument – which says that Forms are distinct from
sense-perceptible things – is manifestly of great significance. In Phaedo, in
which Socrates’ friends during his last hours are ready to grant him that
there are Forms, and that they cannot be perceived by the senses, it serves
to introduce a basic distinction between ‘two kinds of beings’ (duo eidē tōn
ontōn, a), sense-perceptible beings (aisthēta) and intelligible beings,
objects of reason (noēta); and in Republic (at the end of book V), those
who do not acknowledge that there are imperceptible Forms are set against
philosophers and characterised as ‘lovers of sights’, philotheamones.
The question I want to consider is how, in Phaedo, Plato arrives at the
conclusion that Forms are distinct from sense-perceptible things. I want to
contend that the argument, when properly understood, depends on the
supposition that Forms are essences, essences in the sense of that which is
designated by an adequate and true answer to a ti esti question. Crucial in
the argument is that what certain qualities are cannot be perceived by the
senses; and I shall argue that equally crucial for the argument and its
cogency is the justification of this view, which says that the reason for this
is that what these qualities are cannot be specified by example and
exemplar. At the end of the chapter, I want to compare the Phaedo
argument to the argument at the end of Republic V (e–d).
This is not how the argument is currently understood. For the past
decades, and going back especially to Owen’s paper, a certain
orthodoxy about the argument in these passages has emerged among
critics, according to which Plato’s argument is about the character and
nature of sense-perceptible things, as opposed to the character and nature
of Forms: Sense-perceptible things are subject to co-presence of opposites
(e.g., the same sense-perceptible things that are equal are also unequal);
Forms are not so subject; therefore, Forms are distinct from sense-
Up to the late s, this debate has been carefully summarised by White (). Since then it has
continued apparently unabated: Fine ( [reprinted ], see esp. section IV); Annas (,
–); Bostock (, –); Fine ( [reprinted ], see esp. –); Fine (, e.g.,
). McCabe () says that ‘particulars are thought [by Plato] to suffer from compresent opposites
by their very nature’ (, emphasis added). See also Sedley (a); Rickless (); Hestir ().
Tuozzo () turns up the volume: ‘In favour of this interpretation [i.e., a certain interpretation of
the equals argument in Phaedo] is the fact that Plato incontestably does hold that sensibles suffer the
co-presence of opposites, and that the Forms do not’ ().
I am not claiming originality of this part of my interpretation. This, if I am not mistaken, is how the
argument was read by Ross (, ); Hackforth (, ); Bluck (, ), that is, as claiming
that the same sense-perceptible things appear to the senses to be both equal and unequal, whereas this
does not happen with Forms. It is a good question how, and why, this reading was displaced, in
favour of the reading: the same sense-perceptible things are both equal and unequal, whereas the
Form equal is not both equal and unequal. Gallop (, –), is characteristically careful and
cautious, but in the end appears to think that Plato moves from the many equals appear both equal
and unequal to the many equals are both equal and unequal (see ). Nehamas contributed to
cementing the orthodoxy in the making, when he says (a, –): ‘My claim is that
Hackforth and those who agree with him have been misled by an accidental feature of the
particular example that Plato used in his argument. They thus took it that Plato claims that things
can appear both F and not F in the same context to different observers’ (). It seems to me that
Owen () did not assist in getting the debate off to a good start. He begins by observing that: ‘In
Republic VII (a–a) numbers are classed with such characteristics as light and heavy, large and
small, on the score that our senses can never discover any of them kath’ hauto, in isolation (d):
in perceptible things they are inseparable from their opposites’ (, my underlining). He then
drops, without explanation, this qualification and caveat (i.e., the clause I have underlined) for his
general account of Plato’s argument. I note that Owen’s reading was anticipated by Murphy (,
, n. ). Murphy’s reading was criticised by Hackforth (, ), which is some years before
Owen. Hackforth’s response to Murphy was later criticised by Nehamas (a), as noted above,
who followed Owen, as have so many others since.
I would like to acknowledge that a similar point has been made before: it was made by Kirwan in his
paper, when he urged especial caution before we attribute to Plato, on foot of such passages,
not only apparent, but real co-presence of opposites. But Kirwan’s call for caution has not generally
been heeded, perhaps because he did not consider why, in certain passages (both in the Hippias
Major, as I have noted more than once, and, as we shall see, at the end of Republic V), Plato appears
to move from apparent to real co-presence of opposites. I argue that Plato makes this move in these
passages only on behalf of those he is arguing against, such as Hippias in the Hippias Major and ‘the
lovers of sights’ at the end of Republic V. For they do not believe in a principled distinction between
‘appearing’ and ‘being’, because they think that what a quality is can be defined by example and
exemplar and by how it appears to the senses.
For a thorough criticism of the view that the Phaedo comes equipped with an already available theory
of Forms, see Lee (). Lee likewise argues that ‘there are strong dramatic indications against a
particular metaphysical theory being a necessary presupposition of the argument [i.e., the “equals”
argument we are considering]’ (). This, however, is as far as I am prepared to follow Lee in his
reading of Plato’s argument. From the view that the argument does not depend on an already
available theory of Forms, Lee infers that we must make sense of the argument without referring to
anything beyond the Phaedo and the statement of the argument. This is a dubious inference, and, in
my judgement, what this approach yields, in Lee’s paper, is not a substantial interpretation but
innumerable trees, or bits of trees, without a contour of a wood in sight.
Thus Patterson (, ): ‘Plato’s “middle dialogues,” especially the Phaedo, Republic, Symposium
and Phaedrus, along with the late Timaeus, bring the reader’s attention to a previously unnoticed sort
of thing that he calls “forms” (eide). Forms are entirely imperceptible but grasped in thought; non-
spatial and non-temporal yet fully real; independent of and separate from worldly things yet
“participated in” by them and somehow responsible for their being what they are.’ I suspect that
the view that these dialogues ‘bring the reader’s attention to a previously unnoticed sort of thing’ is
not uncommon in the literature.
Sedley (a) argues that Plato’s argument works only if we read tote . . . tote; and he understands
the tote . . . tote as indicating a distinction in time as such, so that it is the difference in time that
matters to the argument: Forms are changeless, whereas sensibles are changing. I set aside the
possibility of reading tō[i] . . . tō[i] in the feminine, hence as referring to ‘relativity of equality to
different respects’ (Lee , ). Lee (, –) mentions this possibility and the one or two
rare proponents who appear to have noticed it.
I shall consider presently why what the second premise says is, strictly, that this Form does not
appear to be both equal and unequal according to Simmias. This is, we shall see, an important
question. For the view that the argument relies only on Leibniz’s Law, and that the reference to
Simmias is irrelevant, see Dancy (, ), when he says: ‘We are going to advert to Leibniz’s
Law: we need some defeating predicate true of any ordinary equal thing that is not true of the equal
itself. At first sight, it looks as if the defeating predicate is “ – shows itself to Simmias as unequal
under some conceivable circumstances”: this is true of any equal stick but not of the equal. But the
reference to Simmias is irrelevant: Simmias is standing in here for anyone at all. The defeating
predicate is, in effect, “ – shows itself to someone under some conceivable circumstances as unequal”:
any equal stick, stone, or whatever does this; the equal does not.’
Dancy () also connects these passages. He thinks that Socrates’ argument in the Phaedo would
have been ‘less odd’ had he used The Beautiful from the Hippias Major, rather than The Equal
(esp. , but also ).
On the term, ‘the equals themselves’, there is a traditional reading going back to at least to
Heindorf, and endorsed by Archer-Hind (, –), among others, which takes ‘the equals
themselves’ to be referring to the Form of equal, arguing that the plural here as forced by idea that
the Form of equal must be complex or multiple, since equality can only be determined through
comparison. For more on this reading, see Dancy (, , n. ).
Dancy (, –) makes quite a lot of this.
The argument at the end of Republic V (i.e., from b to a) has attracted a huge amount of
attention and analysis during the past forty or fifty years especially. For a good summary of the state
of the art on it, see Lee (b). For a more recent, and different, account of the argument at the
end of Republic V, see Rowe (, ch. ). Rowe reads the argument as constrained by the fact that
it is addressed to the lovers of sounds and sights and constrained in such a way that we cannot
directly infer from it Plato’s own metaphysical or epistemological views.
In Republic b– Plato characterises the view of certain people, or the mindset of many people –
people to whom he refers as ‘the lovers of sounds and sights’ – as being a commitment to thinking
that all there is is ‘colours and shapes and all the things constructed out of such things’ (chroas kai
schēmata kai panta ta ek tōn toioutōn dēmiourgoumena).
See esp. Nehamas (a, ); Bostock (, –). Bostock makes the point in a
characteristically candid way: ‘I conclude that the Republic passage demands interpretation (b)
[i.e., that sensibles are both F and opposite-to-F, for certain qualities, F], and will not tolerate
interpretation (a) [i.e., sensibles only appear both F and opposite-to-F, for certain qualities, F].
Since the Republic is reasonably close in date to the Phaedo this affords a strong presumption that
interpretation (b) is also right for the Phaedo’ ().
In the Phaedo, e–c, Plato argues that essences or Forms are the basis
of all causation and explanation. This, it seems to me, is the right way to
formulate the conclusion of this supremely important argument. The
conclusion is not simply that essences or Forms are causes; for it is also
that causes are, or imply, essences or Forms. Nor is the conclusion that
causes are essences or Forms; for Plato argues that causes, provided that
they are grounded in essences or Forms, may also involve other things.
Above all, as I shall argue, Plato’s argument for this conclusion depends
crucially on the supposition that Forms are, precisely, essences.
It is right to think that Plato’s concept of aitia here should be under-
stood as referring to both causation and explanation. As we shall
see, Plato’s principal concern in this argument is a general understanding
of ‘Why?’ questions and what is required for an answer to a ‘Why?’
question to be adequate: to be, genuinely, explanatory. His
concern, therefore, is with what it is to explain something and with
explanation. At the same time, his concern is with what an answer to a
‘Why?’ question designates, if the answer is adequate and true. His
concern, therefore, is with causes or, as I shall generally prefer to say,
explanantia, and, of course, with those things that they cause and explain,
that is, effects or explananda. Plato is just as much concerned with
causation as with explanation.
The thesis of the present chapter is that Plato’s argument, for the conclu-
sion that Forms are the basis of all causation and explanation, relies on just
two premises. First, the supposition that Forms are, precisely, essences.
And, secondly, the supposition that says that explanations must be
For this point, see Michael Frede (, esp. –, ). Frede offers a general account, both
historically and more generally, of the distinction between causation and explanation and of the
relation between the two. For assessment and criticism of Frede’s account, see Ledbetter ();
Wolfsdorf ().
I follow those critics who think that the transmission theory of causation implies that the cause has
the quality that it transmits to the effect. I am supposing, therefore, that, if Forms are causes in
accordance with the transmission theory of causation, then Forms are self-predicative. Meinwald is
especially clear on this: ‘[Socrates] is thinking that when something acquires a property, that property
is transferred to the participant by coming with a participant’s share of a form that itself has that
property’ (, ; original emphasis). Meinwald is here commenting on the Parmenides, but she
thinks this view goes back to Phaedo-Republic.
See c–, not to be confused with the more general blinding image at e.
I read the scope of the holōs to be peri geneseōs kai phthoras. The point of the holōs is to shift the
discussion from the question whether a particular thing, the soul, is or is not subject to generation
and destruction, to the more general and radical question: Why is any thing that is (or is not) subject
to generation and destruction subject to (or not subject to) generation and destruction? I read the gar
to mean ‘therefore’, rather than ‘for’. The point of the gar is that, since the extended discussion and
argument (going back to b) has failed to establish agreement (on the question whether or not the
soul is subject to generation and destruction), a more general and radical examination is needed (dei).
That the question is, in the first instance at least, about the explanation of things that are subject to
generation and destruction is also indicated by the repeated use of the phrase: ‘why something is, or
comes to be, or ceases to be’ (a–, b–, c). The issue is, in the first instance at any rate,
about the enquiry into natural phenomena (cf. peri phuseōs historian, a).
Socrates does not formulate this aporia in so many words, but he describes his intellectual journey as
a path from youthful confidence in explanations of natural phenomena to a state, as he grew in
maturity, of loss of trust in such explanations and indeed in explanations altogether. It is, I think,
natural and appropriate to describe this as a path from his being concerned entirely with the question
‘What is the explanation of this or that?’ to his being motivated, because of a particular aporia, to
concentrate rather on the question ‘What is an explanation?’ We may note that, in a later dialogue,
the Sophist (b–), Plato uses the very same topos, the contrast between youthful confidence
and mature aporia, to describe a similar path: from being concerned entirely with the question ‘What
things are there?’ to being motivated, because of a particular aporia (expressly so called), to
concentrate also on the question ‘What is being?’.
See Fine (, –); Annas (, –; see also Annas, n. for further critics of a similar
view. My account differs in a number of significant ways. First, I consider how these requirements
ought to be formulated. Secondly, I consider why it is supposed that physicalist explanations do not
satisfy them. Thirdly, I argue that the same charge is levelled at teleological explanations as
traditionally conceived. Fourthly, I consider what is the motivation behind these requirements and
why they are supposed to be plausible. I argue that they are well-motivated but not because of the
principle sometimes referred to as the transmission principle of causation. Finally, I consider how it
is that explanations in conformity with the account that says that explanantia are primarily essences/
Forms is supposed to satisfy them. Here I argue that Fine’s and Annas’ responses are not satisfactory,
unless, as Annas thinks we are forced to do, we add the transmission principle of causation.
See e–b, read together with b–c.
See e–a (also d–e), read together with e–b..
See, e.g., Taylor (, –). Likewise, Annas (, ); Fine (, ), who suppose that
the appeal to teleological explanations is independent of the objection against physicalist
explanations based on their not satisfying the above requirements.
See Larsen (, ).
For convenience, we will often say simply that the same explanandum cannot have different
explanantia (and the same explanans cannot explain different explananda) to mean that the same
explanandum cannot have different and incompatible explanantia (and the same explanans cannot
explain different and incompatible explananda).
By ‘intelligible’ I mean ‘subject to explanation’ (see below for a defence of this). It is worth pointing
out that Plato is not committed to the converse claim: If things satisfy these requirements of
explanation, then their behaviour is intelligible. For this claim would depend on the view that these
(i.e., REQ and REQ) are the only requirements of explanation, and we do not have reason to
think that Plato holds this view.
Greene (, ).
I am not, of course, supposing here that what Plato intends by eikos logos in the Timaeus passage is
‘probabilistic explanation’.
For REQ, compare a–b with b–c. For REQ, compare esp. d–e with
e–b.
Remarkably, David Sedley thinks the schema ‘“It is because of the F that F things are F”’ is ‘treated
as [an example of] self-evident truths’ (, –). I confess I cannot see how it could be so
treated by Plato. Plato intends ‘the F’ here to signify an essence/Form, and he could hardly think it
is self-evident that causes are essences/Forms.
See, e.g., b–: [to kalon, agathon, megan] auto kath’ hauto. Earlier in the dialogue simply F itself
(auto).
That this is indeed what Plato thinks is pointed out by, e.g., Annas (, –); Fine (,
). However, I do not think their accounts are satisfactory to explain why Plato thinks this. Fine
says: ‘Plato’s SA [i.e., the simple schema] disallows this [i.e., explanations by opposites]; something
is, or comes to be, F just in case it is, or comes to be, suitably related to the Form of F, and no Form
(Plato assumes) consists of opposites.’ Annas says: ‘The Form F explains an F instance by its
guaranteed freedom from being the opposite of F.’ But this, surely, is not sufficient to ensure
uniformity. We can see this if we name the Form of F ‘N’ and ask why it should not be possible that
N sometimes causes F in things, sometimes G (opposite to F) in things; or, conversely, why it
should not be possible that things that are F should sometimes be caused to be F by Form N,
sometimes by a different Form, M. As far as I can see, what Fine and Annas say does not address this
crucial question. Nor, in my view, would it be advisable to address it by invoking the transmission
principle of causation, for this would be to beg it. Annas appears to suppose that the question can
only be addressed by invoking this principle, and this is why she concludes that ‘Plato’s demand on
explanation is one which we no longer find compelling; it applies convincingly only to a few simple
cases where the explanandum is a quality that can be transmitted from one thing to another’ ().
I agree with her inference to her conclusion, but not with the supposition behind it, hence not with
her conclusion. What Fine and Annas fail to point out is that the Form F is the essence of the
quality F. This, and this alone, is why, if the Form is the cause, it can only cause F.
Sedley invokes the transmission principle of causation to explain why Plato thinks it is because of
the F that things are F (, –); and he argues that Plato thinks this principle is self-
evidently true ( f.). As if by the invocation of self-evidence, for a supposed principle, we could
dispense with considering why, and with what justification, the philosopher thinks it is indeed
a principle.
See d–, supposing that the hupothesis mentioned here includes the general statement of the
account that explanantia are, primarily, essences, in c–, and not only its particular application
to the present case, in c–, i.e., the case of one thing coming to be two.
Taylor (, ) writes: ‘But the safety of the answer seems to lie in its total lack of information’.
Fine, who raises the question of the safety of the complex schema, answers as follows: ‘Whereas the
SA [i.e., the simple schema] specifies necessary and sufficient conditions, the CA [i.e., the complex
schema] need specify no more than sufficient conditions. Further, the CA allows some explanations
involving opposites. Fever, e.g. is adduced to explain illness; but, as Gallop notes, the opposite of
fever – hypothermia – can also explain illness’ (, ). However, it seems to me that, far from
showing that the complex schema is safe too, what Fine says here suggests that, compared with the
simple schema, the complex schema is not safe at all. With regard to the apparent objection that the
appeal to fever does not satisfy the one requirement of uniformity, we may respond that it is not
fever by itself that is adduced to explain illness but fever appropriately related to the essence of
illness (the same would be true of hypothermia). It is clearly fallacious to argue: fever and
hypothermia (if they are explanantia at all) are opposite explanantia; therefore, fever, if
appropriately related to illness (e.g., if understood as an effect of one kind of illness) and
hypothermia, if appropriately related to illness (e.g., if understood as an effect of a different kind
of illness) are opposite explanantia.
‘Accidentally’ because, as we saw, O is subject to change, and subject to change in particular with
regard to its being F.
This passage does not, of course, say or imply that Simmias or Socrates have an individual essence,
rather it illustrates the distinction between a thing’s being accidentally F and a thing’s being
essentially F, by offering a positive instance of the former and a negative instance of the latter.
I want to consider, in its own right and with a hopefully fresh pair of eyes,
this much-discussed passage from the Republic (VII. a–a). I shall
argue that, in it, Plato sets out a single, extended argument for what I call
Plato’s Principal Hypothesis, in the Republic, for essences and Forms. This
Hypothesis is that:
It is possible for us to make judgements and statements that are about sense-
perceptible and physical things, and that conform to the principle of non-
contradiction, ONLY IF we are capable of individuating certain of those
sense-perceptible things in accord with non-perceptible essences and Forms and
in particular the essence and Form of oneness or unity (to hen).
I argue that the principal question to which this passage is addressed is this:
How is it possible for us to speak of and make judgements about physical
things, if our statements and judgements are to be subject to the principle
of non-contradiction (as this principle is formulated in book IV of
Republic, see esp. b–a; sometimes referred to by critics as the
principle of opposites)? In response to this question, Plato argues that this
is possible, only if there is a kind of thought of which it is true to say that:
first, it is not based in and it is independent of sense-perception; secondly,
its primary objects are sensorily imperceptible essences and Forms and,
thirdly, one of its functions is to individuate physical things according to
essences and Forms and especially the essence and Form of unity. We shall
see that, in the passage, Forms are expressly understood as,
precisely, essences.
I want to note straightaway that this reading of the Republic passage is
compatible with thinking that, in this passage, Plato allows for a mode of
thought and judgement that is based in sense-perception. Our account,
therefore, is mindful of the claim, defended by a number of critics recently,
In this chapter, I use ‘physical things’ and ‘sense-perceptible things’ interchangeably.
This reading goes back to Burnyeat’s paper. It has since been adopted by Lorenz (); Moss
(); Storey ().
I am grateful to Henrik Rojahn for impressing on me the need to make these clarifications.
a–: kai anamnēsas humas ta t’en tois emprosthen hrethenta kai allote ēdē pollakis eirēmena (‘and
reminding you of what was said earlier as well as spoken of on many previous occasions’).
In the Divided Line (VI. b), philosophy is, apparently for the first time in the Republic,
associated with the power of dialectic (hē tou dialegesthai dunamis); an association that Plato picks
up and stays with in book VII. It is worth noting that this association was foreshadowed at V. a, a
passage that begins to prepare for the distinction between the philosopher and the lover of sights and
sounds at the end of book V.
See Republic V. b–, when the lovers of sights and sounds are characterised as espousing only
‘colours and shapes and all the things constructed out of such things’ (chroas kai schēmata kai panta ta
ek tōn toioutōn dēmiourgoumena).
I choose this cumbersome label because, it seems to me, neither ‘phenomenalism’ nor ‘physicalism’
will do on their own. This is because it appears that the lovers of sights and sounds do not properly
distinguish between sense-perceptible things and physical things; on the contrary, it appears to be
distinctive of their view that they move without pause between ‘X sensorily appears Y’ and ‘X is Y’.
See Chapter .
For a good account of the relation of book VII to the book V argument in response to the lovers of
sounds and sights, see Rowe , f.
For a good summary of the different recent accounts of the argument at the end of Republic V, see
Lee b.
See d–: ti oun ean hemin chalepainē[i] houtos, on phamen doxazein all’ ou gignōskein, kai
amphisbētē[i] hōs ouk alēthē legomen; (‘But what if this person should be offended with us, when we
say that he opines but does not know, and should object that we are not speaking truly?’)
It is sometimes argued that what Plato formulates in Republic IV is not a principle of non-
contradiction, because it is formulated entirely in terms of contraries (enantia), not
contradictories. I think the view that says that Plato’s principle is not one of contradiction can be
questioned. It can be argued (Lachance has attempted this argument) that Plato’s formulation
of the principle in Republic IV is associated with his use (in Republic and in dialogues both before
and after) of a variety of phrases of the form: ‘saying things opposite to oneself/to themselves’,
enantia legein heautō[i]/heautois. Here the ‘opposite’ statements referred to by these phrases include
both contrary statements and contradictory statements. That Plato should do this is only to be
expected, since the statement ‘O is both F and con-F (at the same time, in the same respects, etc.)’
implies the statement ‘O is both F and not-F (at the same time, in the same respects, etc.)’. It is a
good question whether Plato also thinks that, conversely, the statement ‘O is both F and not-F (at
the same time, etc.)’ implies the statement ‘O is both F and con-F (at the same time, etc.)’. To
suppose that he does, we would have to suppose that he thinks that the truth of a negative
predication of a property of a subject, of the form, ‘O is not-F’, is grounded in the truth of a
positive predication of a contrary property of the same subject, ‘O is G’ (where G is contrary to F).
I cannot attend to this task here.
I put ‘sometimes’ in CAPITALS to emphasise that there is no implication that always and in all cases
when we feel something, e.g., as hard, we also feel it as soft. I elaborate on this later. For the relevant
contrast, see n. .
I put ‘any thing’ in CAPITALS to emphasise that here there is indeed an implication that this is true
in all cases.
Or ‘be’, kai eiē. The kai eiē is most important, for it implies that the principle is not confined to
doings and sufferings.
.. The ‘finger’ passage and its place in the overall argument
(a–b)
The ‘finger’ passage is the most familiar part of Plato’s argument, when he
explicates the distinction between two uses of sense-perception and does so
through the example of one’s perceiving the three middle fingers of one’s
hand. If the question to which sense-perception is addressed is ‘Is this a
finger?’, then perception does not, at any rate not immediately and
directly, issue in opposite answers, depending on whether it is the finger
in the middle or on the outside, a black or a white finger, a thick or a thin
finger, that one perceives. In such cases, ‘it is not the case that the soul of
the lay person (tōn pollōn) is compelled to enquire of thought “What is a
finger?”’ (d–). On the other hand, if the question to which sense-
perception is addressed is ‘Is this (pointing to the finger) something big or
something small, something thick or something thin, something hard or
something soft?’, then perception does immediately issue in opposite
answers: ‘the sensory perception will report to the soul that it is perceiving
the same thing as being [e.g.] both hard and soft’ (paraggelei [hē aisthēsis] tē[i]
psuchē[i] hōs tauton sklēron te kai malakon aisthanomenē, e–a).
In such cases, the way in which sensory perceptions report their content
For this point, about the force of the hōs hikanōs here, see Mann (, –). Mann
(–) also mentions a number of critics, other than those I have mentioned, who hold the
view that I have found to go back to Burnyeat () and that both Mann and I are arguing against:
Nehamas (; a) and Annas (, , –). In favour of the view that Mann and
I are defending, he refers also to Patterson (, –).
I follow Dancy (, ) here and disagree with Moss (, ), when she says that ‘The
famous finger passage of Republic ( A ff.) makes the same claim [as Phaedo b–c]’.
.. Why unity and number are brought into the argument
(b–c)
At this point unity is brought into the argument. Notably, this is not done
for the purpose of considering whether, as it appears in sense-perception,
unity issues in compresence of opposites: this question is posed only later
(d). Unity is brought in because it is through the use and application
of precisely this concept that thought, when called upon to investigate the
[SOC.] On the other hand, we maintain, don’t we, that sight saw some-
thing (ti) as being both large and small, and not something distinct
(kechōrismenon), but rather, confused (sungechumenon). [GL.] We do.
[SOC.] It is for the sake of clarity on this point (dia de tēn toutou
saphēneian) that thought was compelled to inspect a thing large and also
small, and to inspect them not as confused things (sungechumena) but as
determinate things (dihōrismena), as opposed to what sense-perception did.
[GL.] True. [SOC.] Is it not the case, then, that this is the basis of our
thought’s arriving at the question, ‘What then is the large thing and also the
small thing?’? [GL.] Most certainly. (c–)
[SOC.] And so we called one <kind> that which is intelligible and the
other <kind> that which is visible. [GL.] Perfectly correct. (c–;
my translation)
Why, we must ask, is thought’s first task, in response to the apparent
absurdity of the reports of the content of sensory perceptions, ‘to consider
whether each of the things reported (hekasta tōn eisaggellomenōn) are one or
two (eite hen eite duo)’? Apparently, what the soul considers in considering
this is how the object of perception is to be individuated. But we must ask
why the soul has to consider this, and to what end. The evident answer, in
view of Plato’s argument up to this point, is that the object of perception
must be individuated in a way that is adequate for making a judgement
about it (e.g., ‘This thing is hard’), and avoids the logical absurdity of the
tí oun pot’ esti to mega au kai to smikron (c–). See below for the meaning of this indirect
interrogative phrase.
Unlike Priest (, ): ‘Another example: I walk out of the room; for an instant, I am
symmetrically poised, one foot in, one foot out, my center of gravity lying on the vertical plane
containing the center of gravity of the door. Am I in or not in the room?’ He concludes: ‘I am both
in and not in.’ This passage is a perfect example, it seems to me, of how one may find that a physical
thing can contravene the principle of non-contradiction, if one fails to consider how a thing needs
to be individuated for the principle to apply to it.
.. The meaning of the phrase tí oun pot’ esti to mega au kai to smikron: a
question of how to individuate the object of perception (c–)
This indirect interrogative phrase is usually understood to refer to the
question, Q, ‘What is largeness and smallness?’, and is translated accord-
ingly. However, the phrase can also be understood to refer to the question,
Q, ‘What is the large thing and also the small thing?’ Socrates says that
what they have just been considering prompts (cf. enteuthen pothen, c)
the question: tí oun pot’ esti to mega au kai to smikron. What they have just
been considering is how, without contradiction, one and the same thing can
at the same time be both large and small, thick and thin, hard and soft. This
immediately prompts the question ‘What then is the large thing and also the
small thing?’; in particular, ‘Is it a single thing or two things?’ (see b–).
Suppose the question is Q. And suppose this question is answered, by giving
an account of largeness and smallness, irrespective of whether the account is or
is not in terms of sense-perceptible things or qualities. It is not at all clear how
this assists with the issue Plato is addressing. For, the question he is addressing
is: How, without contradiction, can one and the same sense-perceptible thing
at the same time be both large and small, thick and thin, hard and soft?
It appears that one reason why the interrogative phrase has generally been
understood as referring to Q is that, a little later, Plato will use a seemingly
similar phrase, tí pote estin auto to hen, to refer to the question ‘What is that
thing itself, unity?’ (a) But this overlooks that, in the phrase tí oun pot’ esti
to mega au kai to smikron, the important word, auto (‘itself’), is absent. The
function of this word, auto, in such phrases in Plato is, typically, to refer to the
quality itself, F-ness, as opposed to a particular thing that is F. In the later
passage this word is not only present, it is underscored by the fact that it is said
that what is at issue is ‘unity itself by itself’ (auto kath’ hauto to hen, d–e).
apeira to phēthos, as Glaucon will say when he speaks up for the second lemma; see a–.
See Chapters and of this study; see also Politis , ch. .
It should not be immediately assumed that this supposition is un-Platonic: in Parmenides d–,
Socrates says of some things that ‘they are, just as we see them to be’. Commenting on this passage,
Harte (, ) says: ‘Forms are not needed in those cases where things are “just as we see them
to be”. What cases these are may be for us to discover.’ If these cases do not include those of
immediately perceptible qualities of individual sense modalities, such as the quality red, it is hard to
imagine what they could include.
Contra Pappas (, ): ‘Why does mathematics suddenly enter the present argument? Because
numbers form a special case of opposable properties. They appear in particular things in the same
confusing way that other relative terms do: a may mean, for instance, that my hand is
simultaneously one (hand) and five (fingers).’ Pappas states clearly a view held by many critics.
One thing about which critics are generally agreed in regard to the second
part of the Parmenides, notwithstanding many and profound disagree-
ments, is that a good reading of it should attempt to establish a credible
connection between it and the critical examination of, and dispute about,
Forms, from which it follows on and to which it is in some way a response.
This desideratum, I argue, is met only if we suppose that the second part is
as much about thinking as it is about being. I shall, therefore, examine
whether the second part of the dialogue considers not only being, but also
thinking and its relation to being. I argue that it does, in ways that are
continuous with certain crucially important elements in the first part that
are specifically about thinking and its relation to being. The elements
I have in mind are, first, the claim at b–c, which says that Forms,
and the defining of Forms, are necessary for thought; and, secondly, the
claim at b–c, which says that thought is of something, something
that is, and something that is one. I argue that the earlier claim signifi-
cantly prepares for the later claim and that both claims cry out for further
articulation and defence. I argue that Plato provides this in the second part
of the dialogue.
The critical examination of the theory of Forms, which makes up much
of the first part of the dialogue, does not conclude with the objections
The claim at b–c, that Forms are necessary for thought, has attracted much attention, as has
the claim, at b–c, that thinking is of something that is and is one. Especially since Burnyeat’s
paper, this latter claim has been caught up in the debate over whether Plato’s Forms are mind-
independent – which is not an issue I want to consider here and I shall consider in Chapter .
However, if I am not mistaken, critics have not made the connection between the b–c claim and
the b–c claim; and they have not considered whether, in the second part of the dialogue, Plato
defends the claim, as made in the passage and directly implied in the passage, that Forms are
necessary for thought. M. L. Gill (, ) briefly considers this, or something like it, but she runs
together the question of Plato’s defence of the claim that Forms are necessary for thought with the
question of his defence of the claim that they are necessary for ‘the things themselves’.
. The relation between thinking and being, and its relevance
for the dispute about Forms, in the first part of the dialogue
Following a series of objections against the supposition that there are
Forms, and against the theory of Forms that Socrates introduced at the
Contra Allen (), who, remarkably, has nothing to say about the claim, at b–c, that Forms
are necessary for thought, or about the role of this claim in the overall argument and dialectic in the
dialogue so far.
I shall not consider here whether there are further arguments for Forms in the second part, not based
on Forms being necessary for thought and in general on the relation of thinking and being.
b–c and e–b. The later passage refers back to the earlier passage, and the two
passages need to be read together.
I intend to steer clear, at this point, of the issue whether Plato’s argument here is intended to show
that Forms are mind-independent. Burnyeat () notoriously argues that it is so intended. Allen
(, ) strikes a cautionary note: ‘Nothing in the mere observation that thought is of something
which is and is one over many requires the conclusion that what is thought exists independently of
the thinking of it.’ I take up this issue in the final chapter.
The view that says that, according to Plato, certain words or concepts signify Forms and would be
meaningless if they did not signify Forms has been traditional for the past hundred plus years, both
in its own right and as an account of why, in the Parmenides passage (b–c), Plato claims that
Forms are necessary for thought and speech. It is true that this view is not as common as it used to
be. Recently Crivelli (, –) rejected it out of hand, and Malcolm (, ch. ) attacked it
thirty years ago, as did Gosling (, ) half a century ago. This does not mean it is about to go
away. For a recent statement of the view, see Rickless (, ), who, commenting on Parmenides
b–c, says this: ‘names signify forms (or their natures), and thus would have no meaning if forms
did not exist’. Like other defenders of this view, Rickless appeals to the Cratylus in its defence, as do
Kutschera (, ) and Hestir (, ch. ). Going back at least to Cornford (, –),
critics that defend the semantic view include Ryle ( [originally ], ); Keyt (, );
Teloh (, , ); Bostock (, ); Bestor (, ). Several of these critics defend the
view both in its own right and as an account of why, in the Parmenides passage, Plato claims that
Forms are necessary for thought and speech.
Rickless () argues that the claim here is that the Form of oneness cannot in any way be many;
indeed, that quite generally Plato’s Forms, before the second part of the Parmenides, cannot in any
way be many. Along similar lines, see M. L. Gill: ‘I shall argue that Part II [of the Parmenides] is an
indirect argument demonstrating that to save the theory of forms and philosophy Socrates must
abandon his thesis about the one and admit that it is both one and many (in different ways)’ (,
–, original emphasis). This, it seems to me, cannot be right. When Socrates introduces ‘his
thesis’ early in the dialogue, what he says is that that which one is (ho estin hen) cannot be many
(b–), not simply that the one cannot be many; and, in the second part of the dialogue, he says
the very same about the one that is understood in this way, that is, as‘ the one by itself’ (a) and
‘the one itself’ (b) and ‘the truly one’ (c), and is set against the one in combination with
being (also referred to as ‘the one being’, to hen on, see, e.g., a). What he says is that, while the
one in combination with being is also many, the one by itself (or ‘itself’, or ‘the truly one’, or ‘that
which one is’) is not also many. What this shows is two things. First, the claim is that the Form of
F cannot be both F and contrary-to-F, hence the Form of one cannot be both one and many; it is
not, contra Rickless, that Forms cannot in any way be many (for a similar criticism of Rickless, see
Sayre ). Secondly, and contra Rickless and Gill, there is not an inconsistency or contradiction
between the first and the second part of the dialogue on the question whether the one, itself by itself,
can be both one and many: the first part of the dialogue denies this and so does the second part.
What the second part does is explain, or begin to explain, how the one can, in relation to other
things, hence not itself by itself, be both one and many, an explanation that is taken further in the
Sophist. Remarkably, the claim that Forms are one by themselves but many in relation to physical
bodies, and indeed to other Forms, is made in so many words in Republic V. a–.
See the recent work of Evan Rodriguez (; ). He is working further on this, with a number
of publications forthcoming.
Contra Brisson (). In view of what I have been arguing, it will not be a surprise that I find no
more credible Brisson’s view that the one of the second part of the dialogue signifies, exclusively or
primarily, the unitary cosmos and the all. While the cosmos, and the all, may be one theme of the
second part, it is surely not the only or the primary one. If, as Brisson does, we want to trace
connections between Plato’s dialogue and the poem of Parmenides, a reference that Brisson
remarkably overlooks is to fragment , in which Parmenides marks an unbreakable unity
between thought and being.
It will be evident, to the reader who knows Paul Natorp‘s classic, Platos Ideenlehre (translated
with introduction in Natorp ), that my account of Plato’s account of the relation between
thought and being in the second part of Parmenides is much indebted to Natorp’s chapter on the
Parmenides in that monumental work – and this even setting entirely to one side Natorp’s overall
view (which I do not share; see Chapter ) that, for Plato, thinking is prior to being. Years after my
having worked on Natorp, David Horan, in a paper that he published some time later (),
demonstrated to me that, on this question at any rate – the question of how the thought of the one
is necessary for, and is involved in, all thought – Plotinus is an astute and penetrating interpreter
of Plato.
I do not mean to imply that, if the one is understood in different ways in the two Suppositions, two
ones are in question.
I shall not consider why, or because of what (hidden) supposition, this might be thought to be a
plausible inference. Perhaps this is because of the (hidden) supposition that what a thing is, it is,
purely and simply, in relation to itself.
The claim that the one is not even a being is inferred from the claim that it does not partake of time
in any way (e–). Some critics think this inference is clearly fallacious, and intended as
fallacious. I don’t think it is clearly fallacious, and there is no indication that it is intended
as fallacious.
It is familiar, especially since Dodds’ classic paper, that Plotinus read the ending of the
investigation of the First Supposition as positively introducing the idea of an absolute One that is
beyond thought. It is hard to assess this reading, since demonstrating of something that it is not a
possible object of thought falls short of demonstrating a strict absurdity, or contradiction, of
supposing that there is, or could be, such a thing (see Castagnoli ).
See Meinwald ().
For the view that these two things amount to the same, see Chapters and .
Republic VII. a. At e this one is introduced in and through the question ‘What is this one
itself?’ (tí pote estin auto to hen).
. Conclusion
One remarkable consequence of the Third General Condition of Thought,
as this is articulated and defended in the second part of Plato’s Parmenides,
is this: It is not specific Forms (the Form of justice, of human beings, of
fire, etc.) that provide for the oneness of the object of thought, but, in the
first instance at any rate, the Form of oneness that does this. The same
condition was put forward, without articulation and defence, in the first
It will be evident that the account I have defended of the way in which, for Plato, Forms are
necessary for thought is far removed from the view, defended by, e.g., Bestor (), that Plato’s
account of how Forms are involved in thought is a variation of a Putnam-style externalist account of
the object-directedness of thought. On a Putnam-style externalist account, a thought is of an object,
O, if it is suitably caused by O; and a thinker can be thinking of O without being aware that this is
the object of which she is thinking.
Cratylus b–c, picked up at the end of the dialogue at d–e (I am supposing that, for Plato,
naming is not only a linguistic act but also a mental and intellectual act). Contra several critics (e.g.,
Kutschera ; Rickless ), in the Cratylus Plato does not argue that names signify Forms and
would be meaningless if they did not signify Forms. It is a difficult question what Plato’s account is,
in the Cratylus, of how names name (onomazein), or signify (sēmainein), things, or whether it is even
part of Plato’s aim in this dialogue to provide a general account of how words are related, for their
having a meaning, to the things they signify. However, it ought to be evident that he allows that a
name can signify something, without signifying a Form. Towards the end of the Cratylus, he argues,
carefully and persistently, that a name can be assigned to a thing simply demonstratively and by
pointing (–c). He links up this argument with the account of the function of names and of
naming that he had set out at the beginning of the dialogue (see d–e, where he refers back to
b–c). According to that original account, a name is a tool (organon) whose function is to
distinguish how things are (b–c; a tool by the use of which ‘we distinguish how things are’,
ta pragmata diakrinomen hē[i] echei, b–; ‘a tool for distinguishing the way things are’, organon
diakritikon tēs ousias, b–c). And, according to the argument towards the end of the dialogue, this
function can be accomplished simply demonstratively and by pointing. It is true that, if it is
accomplished in this way, it may not – indeed will not – issue in a true and correct way of
distinguishing how things are, but it will have succeeded in distinguishing how things are in a way
that is sufficient for assigning a name to a thing. For a similar point about the Cratylus, and one
made long before this dialogue became again a focus of serious interest, I recommend Gosling
(, ): ‘But this position [i.e., that ‘all meanings are Forms’] is not open to Socrates. For at
[Cratylus] d– he has first argued that what a noise means is a matter of convention, showing
this by the fact that we readily understand words although they contravene Cratylus’ similarity rule
and so should not be intelligible. Second, he argues that words might well have been ill-devised. But
the danger of an ill-devised word is not that it is senseless, but that it is misleading . . .. Yet if they
[i.e., words] always in virtue of being meaningful isolated Forms which constituted their meaning,
they could not be misleading.’ It may be said that this account of the Cratylus ignores its last two
pages (d–e), where it is argued that without Forms it is not possible ‘to correctly address a
thing’ (proseipein auto orthōs, d). Does it follow that Forms are necessary for naming things,
hence for names to have a meaning? The term proseipein (‘to address’) is closely related to the term
onomazein (‘to name’). We may suppose, therefore, that when Plato says that Forms are necessary
for correctly addressing things, he intends to imply that they are necessary for correctly naming
things. But the ‘correctly’ (orthōs) makes all the difference. Socrates has, in the ten Stephanus pages
coming up to this juncture, been at pains to argue that it is possible to name a thing – and,
therefore, it is possible for a name to have a meaning – without naming it correctly; and the ending
of the dialogue carefully maintains consistency on this point.
See, e.g., the following passages – which should be read seriatim – from Aristotle’s Metaphysics (I am
grateful to Ge Tianqin for having impressed them on me): ‘Being and unity are the same – i.e. one
nature – in the sense that they are followed by one another, just like principle and cause, not as being
indicated by the same formula (though it makes no difference even if we believe them like that –
indeed it helps).’ (Metaphysics Γ. b–, trans. Kirwan, modified) ‘On the other hand, that
the one in some sense signifies the same as being, is clear in virtue of the fact that it follows the
categories in the same number of ways and that it is in none of them (e.g. it is neither in the category
of the what-it-is nor in the category of of-what-quality, but it behaves in the same way as being), and
in virtue of the fact that “one human being” does not add anything else in predication to “human
being” (as being, too, is nothing over and above being a certain something or of a certain quality or of
a certain quantity), and in virtue of the fact that being one is being for each thing.’ (Metaphysics I.
a–, trans. Castelli) See also Metaphysics Λ. a–. I note that the passage from
Metaphysics Γ. (b–) has been thoroughly analysed by Politis and Steinkrüger (). The
passage from Metaphysics Iota is, of course, thoroughly striking for our present purpose.
See Gerson ().
For Plato’s close association of the notion of one’s soul with the notion of oneself, see esp. Protagoras
–, where he moves freely back and forth between existential questions about ‘one’s soul’ to
the same questions about ‘oneself’.
Plato often uses ‘to be present’ (eneinai) for the converse relation of ‘to partake of’ (metechein).
I want to be clear that, when I speak of such things as being ‘non-physical’ things, I do not mean, or
I do not have to be taken to mean, ‘things separate from physical things’, on any account of
separation. All I need be taken to mean is what, a generation or two ago, philosophers – including
analytic philosophers such as Peter Strawson and Donald Davidson – used to mean when they would
say such things as: a human being is a bearer of at once physical and mental (hence non-physical)
properties. I hope such ways of speaking are still intelligible, and respectable enough, even today
when naturalism, both metaphysical and methodological, is the order of the day, and that they are
intelligible and respectable not only in regard to such jejune entities as qualia.
It is striking how, when, in Sophist, Plato formulates the various theories about what there is, he
formulates them as being not only about ‘that which is’ (to on) and ‘the things that are’ (ta onta), but
also about ‘the totality of things’ (to pan) and ‘all things’ (ta panta): e, d, e, b,
b, b, c, d, d, a.
Some critics (such as Broadie ) argue that the fact that the Timaeus adopts a cosmic perspective
shows that it is concerned with cosmology and not with metaphysics. This is, surely, a false contrast
as far as Plato is concerned. We would hardly question that Sophist ff. is concerned with
metaphysical questions, questions about being, what there is and what being is; and we would
hardly suggest what Plato takes himself to be concerned with there are cosmological in
contradistinction to metaphysical questions. And yet, Plato there moves back and forth between
speaking of being and beings (to on, ta onta) and speaking of all things and the totality of things (ta
panta, to pan); see: e, d, e, b, b, b, c, d, d, a.
See esp. b, where the totality of Forms is characterised as a perfect, or perfectly complete
(pantelēs), living thing.
See Dillon (); Sedley (b); Broadie (). See also Sorabji (); Johansen ();
Carone ().
See, e.g., a–b.
I want to leave open whether such a cosmic perspective is already present in Republic, noting its
reference to a kosmos of Forms (at VI. c).
See Politis and Steinkrüger ().
See Chapter , where I also quoted this passage.
I would like to note, and acknowledge in gratitude – he was a mentor and friend – David Evans’
excellent paper, in which he argues, incisively, that those critics who urge that Plato is a realist
fail to take account of the many and important arguments in Plato in which he derives a conclusion
about beings or being from a premise about thinking and knowing. For, as Evans reminds us, such
argument ought not to be available to a realist, unless, that is, he is confused.
Unfortunately, the Slings edition, whose text of these lines is the same as the Burnet edition, has
these lines as d–e. I shall follow the Burnet edition.
The reading of ‘essence’ for ousia here is not crucial to my argument. But we may note that the ‘ho
estin’ at b is naturally understood as ‘that which it is’ (and not: ‘that which is’), and hence as a
way of referring to the essence of a thing or quality.
See also d, outō toinun kai to tēs psuchēs hōde noei, ‘Thus then also in regard of that in virtue of
which the soul thinks’.
It is interesting to note that Natorp, who is famous for arguing that, for Plato, thinking is prior to
being, and who dedicates a -page work (which he published in ) to arguing this, did, some
twenty years later, for the second edition, and not long before he died, recant, and in suitably
elevated, declamatory and near hierophantic style: he argues, in the Metakritischer Anhang, which is
the appendix to the second edition of Platos Ideenlehre, for something like the account, of Plato’s
view of the relation between being and thinking, that I am defending here.
. Addendum
I would like to make, loud and clear, the following clarificatory points,
which serve to make clear what my account in this chapter of Plato’s Sun
Analogy does not imply and what I am not committed to in and through
this account.
First, I am not, in and through this account of Plato’s Sun Analogy,
arguing that Plato’s account of the relation between mind and world is
anti-realist, in the sense of asserting the priority of mind over world.
Secondly, I am not, in and through this account of Plato’s Sun Analogy,
arguing that Plato thinks that the Good is beyond knowledge.
Thirdly, I am not, in and through this account of Plato’s Sun Analogy,
arguing that Plato thinks that the Good is beyond being or
beyond essence.
Fourthly, yes, I admit, the account I have defended of Plato’s Sun
Analogy would benefit from being complemented by our taking up and
investigating the following questions (which I have not posed in this
chapter): Does Plato, in the Sun Analogy, distinguish between the sun
and the light of the sun, and, correspondingly, between the Good and the
power of the Good? If he does make this distinction, what role, if any, does
it play in his argument in the Sun Analogy?
Taking up these questions would require a full interpretation of the Sun
Analogy, which I have not attempted. However, from what I can see, the
crux claim in my account in this chapter of Plato’s Sun Analogy does not
depend on taking a particular, and potentially disputable, stand on
these questions.
I am grateful to Daniel Vazquez for advising me to do something like this.
This is the conclusion of the present study: that Plato’s Forms simply are
essences, not things that have essences. For this is what is established, if it is
established that Plato’s Forms are, basically, that which is designated by an
adequate and true answer to a ti esti question, and if it is established that
the other principal characteristics of Forms – Forms are changeless, uni-
form, not perceptible by the senses, knowable only by reasoning, the basis
of causation and explanation, distinct from sense-perceptible things, nec-
essary for thought and speech, separate from physical things and more –
are based, for their justification, in one way or another, on this basic
identity of Forms.
This contradicts a tradition of interpretation, dominant among many
critics for some time now, according to which, up until the second part of
the Parmenides, when Plato came to think better of it, and before he drove
the point home in a single line in the Sophist (c–), he thought of
Forms as things that have properties, and have properties in just the way in
which any thing has properties. Have we provided an alternative to this
view? Do we not still have to refute the basis for it, which is the view that,
up until the second part of the Parmenides and driven home in the Sophist
Perhaps especially since Vlastos’ paper. As Malcolm (, ), notes, the term ‘self-predication’
was introduced by Vlastos in that paper.
‘But I think you agree that of the things that are, some are spoken of in and by themselves (auta kath’
hauta), while others are always spoken of in relation to others (pros alla)’ (Sophist c–; trans.
Rowe). Remarkably, M. Frede (, , emphasis added) says that the distinction made in this
statement in the Sophist, between that which is kath’ hauto and that which is pros allo, ‘in no way’
occurs here for the first time in Plato: ‘Diese beiden Formen [i.e., to kath’ hauto and to pros allo,
referred to at Sophist d] . . . tauchen übrigens im Sophistes keineswegs zum erstenmal bei Platon
auf. Im Charmides nämlich heißt es (a–) . . ..’ If only Frede had followed up on this
remarkable statement! – it would have saved us so much trouble! This directly implies that
Michael Frede cannot be saying what M. L. Gill thinks he is saying. She thinks, and, as indeed
does this whole tradition, she attributes to Michael Frede the view that the Sophist claim goes back
only to the Parmenides and it contradicts and corrects what Plato thought before the Parmenides
(I am supposing the Charmides is before the Parmenides).
See M. L. Gill: ‘In my view the self-predication assumption goes hand-in-hand with a view of
causation, sometimes called “the transmission theory of causation”’ (, ).
See Chapter .
Gill (, ) appeals to Symposium e–a as a clear case of self-predication; but what that
passage says is that the Form of beauty is not both beautiful and not beautiful, or ugly, in the variety
of ways in which sense-perceptible things are so.
One such passage is Phaedo c–: ‘Consider, then, he said, whether you share my opinion as to
what follows, for I think that, if there is anything beautiful, excepting (plēn, often translated here
with “besides”) the Beautiful itself (auto to kalon), it is beautiful for no other reason than that it
shares in that Beautiful (ekeino to kalon), and I say so with everything. Do you agree to this sort of
cause?’ Harte (, ) comments: ‘Since this passage assumes that the Beautiful itself is
beautiful . . . we have here a pretty clear statement of self-predication in what looks to be a sample
case: the Form, the Beautiful itself, is beautiful.’ I am not sure this passage is such a clear case, if we
take plēn in the sense of ‘excepting’, not in the sense of ‘besides’ or ‘in addition to’. But I shall not
press this point. Harte (, ), when she considers the question of self-predication, cites but this
passage, Phaedo c; and she is careful not to commit herself to the presence in it of self-
predication, for she sums up her examination with: ‘If there are good grounds for supposing that
Forms self-predicate . . .’ (, emphasis added).
It is commonly thought that this view goes back to Michael Frede (). But this is far from clear.
In his paper (–), in which he takes himself to be re-stating his view of Plato’s
account of being in the Sophist (see esp. , –), he appears to dissociate himself from this
understanding of his view. Gill (, ) recognises this; but, incredibly, she refuses to take Frede
at his word, when he dissociates himself from this understanding of his own view, and she says that,
in so dissociating himself, ‘He misrepresents what he has just said.’
Meinwald is careful not to commit herself to any particular account of the relation between a
Form and its nature (‘nature’ is her term for my ‘essence’). All she is prepared to say is that ‘there
must be a connection between forms and natures so close that a nature cannot remain when the
associated form is taken away’ (; also ). Notably, she says that ‘A straightforward way of
ensuring this would be to take natures to be in some way components of forms’ (). This is
notable not least because, if essences/natures are components of Forms, then it is not the case that
Forms are essences. Silverman (), too, denies that a Form is identical with its essence and that
Forms are essences; he speaks of Forms as being ‘bearers’ of essence.
Most clearly at b–.
Aristotle a–,
Metaphysics, , Gorgias,
Gamma , –,
b–, Hippias Major, –, , , , , , –,
Iota , , , , –, –, ch.
a–, –, –, ,
Lambda, c f.,
Lambda c,
a–, d,
Mu, e–, , , ,
Nu, b–d,
Zeta –, d–e,
Zeta , d,
b–, a, ,
Nicomachen Ethics, e–,
a–b,
Heraclitus a–,
B, b,
b,
Parmenides b,
B, , b–,
Plato d–,
Charmides b,
a–, d,
Cratylus, , , , e–c,
b–c, c–,
b–, b,
b–c, Laches, ,
b–c, Meno, , , ,
–c, d ff.,
d–e, Parmenides, , , –, , –, , ,
d–, , , , ,
d–e, e–b,
d, e–a,
Euthyphro, , , , e–a,
d–e, e–a,
d–e, b–,
d–e, , – b–,
d–e, , , ,
e–, a–e, , , –