Knowledge and the Self in Platonic Philosophy

Lloyd P. Gerson

University of Toronto

Perhaps it is an ominous portent, but I would nevertheless like to begin by explaining the title of
my paper. One might have surmised that I am going to talk about self-knowledge, about the
cryptic injunction , or know thyself. Alas, I am not, although I believe that the
topic of self-knowledge in Plato and in ancient philosophy generally is well worth serious study.
In fact, I shall in due course have a bit to say about what I think self-knowledge is in Plato, but
that is not my central subject. Rather, I am going to focus my attention and I hope yours on the
self in Plato and on knowledge. What these two concepts have to do with each other will I hope
become clear. For the moment I shall merely advertise my main conclusion: for Plato the true or
ideal self or person is a knower, an eternal contemplator of eternal truth. This may not in itself be
a particularly contentious claim. But since persons spend very little of their time contemplating,
we must then agree that ordinary persons in their ordinary and not so ordinary lives are not ideal
selves. What will perhaps be held to be contentious is my claim that these less than ideal selves
are images of their ideal, just as the sensible world generally is for Plato an image of its eternal
exemplar. Furthermore, forms of cognition other than or less than knowledge are similarly
images of their exemplar. In short, Plato doesnt merely believe that the everyday world is not
the really real world, but he also believes that everyday selves are not really real selves and that
the relation among these is strictly analogous. If I may, I would like to make one more self-
referential remark about my title and that concerns the ambiguous phrase Platonic philosophy.
To tell the truth, this paper is extracted from a lengthy work in progress about the ideas of the self
and knowledge in the entire tradition that calls itself Platonic. Even more alarmingly, with
respect to these two ideas, that tradition includes not only self-professed Platonists like Plotinus
and Proclus, but also philosophers, like the Stoic Epictetus and Aristotle, who are not usually
called Platonists in any sense. I have no intention of inflicting on you anything approaching a
survey of the nexus of the ideas of the self and knowledge in philosophers after Plato. But I shall
toward the end of my paper try to make good my implicit promise to talk some about Platonism
and not just Plato.
I begin with a search for the source of Platos concept of the self in the early so-called
Socratic dialogues. In particular, I shall try to excavate the concept of the self presupposed in the
so-called Socratic paradoxes. These are paradoxical because they fly in the face of
conventional beliefs regarding human interests. For example, the paradoxical claim that it is
better to suffer than to do evil directly confronts the ordinary and deeply held belief that doing
evil does not harm one at all whereas suffering evil or having evil done to one harms one almost
by definition. The paradox that no one does wrong willingly seems to reject the common belief
that doing wrong is at least sometimes in ones interest and that people normally act willingly in
their own interest. The paradox that a worse man cannot harm a better man just sounds silly.
Surely, this happens all the time. Similarly for the claim that the greatest harm for a wrongdoer
is to go unpunished. The dispute between a Socrates who makes such wild assertions and a
typical Athenian gentleman is not, for example, over whether suffering evil, say, receiving an
unjust blow, is more painful than delivering one. Of course it is. Rather, the underlying dispute
is over whether ones interests are always better served by doing that which produces on balance
less pain. If my only choice is between inflicting pain unjustly and having pain unjustly inflicted
on me, could it ever be in my interest to choose the latter? Socrates thinks not just that it it
sometimes in my interest to choose the latter but that it is always and necessarily so. Clearly,
there is a difference here concerning what constitutes human interest.
 If we state the matter of the dispute in this way, one might aptly reply that different
people have different interests and Socrates is hardly in a position to privelege his own. Perhaps
it is in his interest as he conceives it to suffer rather than to do evil because, say, the shame he
would feel in doing the latter would trouble him more than the pain he would experience in
suffering the former. Someone else, however, may conceive his interest diferently. 1 The point is
superficially a powerful one, but it has an underlying weakness. It presumes that each person is
authoritative in determining his or her own interests, that if, for instance, I hold that my interest is
served better by doing evil than suffering it, then no one can gainsay my claim. Interests are like
matters of taste: each person is their ultimate arbiter. It is fairly obvious, however, that persons
are not always infalible assessors of their own interests. It is even fairly obvious that at times
someone else might make a better judgment regarding your interests than you yourself. But
Socrates position would be uninteresting and unpersuasive if he were merely claiming that the
one who prefers to do rather than to suffer evil might someday reassess his priorities and decide
that, on balance, yes, it is better for him to suffer than to do evil. This simply cannot be
Socratesposition because he holds that even if you go to your death believing that it is better to
do than to suffer evil, you are wrong about your own interests.
Why is Socrates convinced that people habitually and pervesely misconstrue their own
interests? The short answer is that Socrates believes that persons are souls. If Socrates held
simply that psychic interests are more important than bodily interests, it would be open to anyone
to object that they may be more important to him but that does not make them necessarily more
important to anyone else. No, Socrates must be presuming that a persons interests must be
psychic because a soul is what a person is. If a person is a soul then it follows that his or her
body is a possession. Since people literally cannot distinguish themselves from one of their
possessions, namely, their bodies, they mistakenly believe that their interests are the interests of
bodies. When someone says, for example, that it is in his interest to avoid pain, he is assuming
that he is a body, not a soul.
If the soul is the person and the body is a possession, nevertheless the body is an odd sort
of possession. It is especially odd to speak of ones interests over against those of ones body as
if the body were a mere possession. This is so because the body is ambiguous and includes an
agents states such as pain and pleasure. My pleasures and pains are states of me, unlike ordinary
possessions which belong to me but are not me. Granting these points, Socrates risks a reversion
to the previous objection, namely, that he might prefer one sort of state whereas other people
prefer others. In other words, his interlocutors do not have to base their rejection of the
paradoxes on the absurd notion that it makes any sense at all to choose the interest of an
ordinary possession over the interest of oneself. Rather, they can claim that they prefer to be in
whatever state one is in when one does evil to whatever state one is in when suffering evil. And
this would mean that again, the admonition that it is better to suffer than to do evil is nothing
more than an expression of Socrates personal preference.
The Apology is particularly illuminating regarding the underlying problem. Socrates
says, For I go around doing nothing but persuading both young and old among you not to care
for your body or your wealth in preference to or as strongly as for the best possible state of your
soul (30a7-b2; cf. 29e7-d3). This text and others are interesting because they are so obviously
question-begging. Why, we may ask, should anyone be persuaded to alter their interests, say, to
put their soul before their body? The question is an instance of a general question, why should
anyone prefer one thing to another? If the question is about alternative means to an end, it is
quite a good question. At least, one can argue why one thing is more likely to achieve an end
than another. But if the question is about ends themselves, say, the state of ones body and the

1
This is precisely the basis for the objection in the Gorgias that Callicles makes to Socrates refutation of
Polus stubborn assertion that doing evil is better than having evil done to you.
state of ones soul, then it is far from obvious that one can argue, as opposed simply to exhort one
to prefer the latter to the former. 2
If the exhortation to pursue care of the soul to care of the body is, however, an
exhortation to prefer care of oneself to care of ones possessions, then the argumentative
possibilities are only marginally more promising. Assume it is true at least that all parties can
agree that possessions without possessors make no sense. So, if preferring ones possessions over
oneself means preferring ones possessions to the loss of oneself, it is easy to see that this makes
no sense. Nevertheless, if pleasure can be legitimately construed as my interest and not the
interest of my possession, then we are back to the problem of why one should prefer Socrates
ordering of his interests to ones own.
If the person or self is a soul, the Socratic paradoxes can be provided with strong
arguments. Unfortunately, the fact that bodily states are states of the person and only in an odd
sense states of a possession undercuts these arguments for the reason given above. The problem
is evident if ones considers, for example, Socrates absolutist prohibition of wrongdoing in the
Crito: one ought never to do wrong( 49a6-7; cf. Apology 29b6-7,
Gorgias 469b12, 508c, etc.). 3 This prohibition, which I have termed absolutist sums up in a
way Socrates philosophy. It is a prohibition that has not received a perspicuous defense. I
suppose that one might say in Socrates defense that being in a virtuous state is more important
that being in a pleasurable state because somehow the former identfies us more closely than the
latter. But surely absent any supporting evidence this is mere rhetoric. Gregory Vlastos
immortalized the rhetoric of this position long ago when he offered this commentary on Socratic
absolutism. If you have only one more day to live it makes no sense to spend it in any way other
than that which makes you a better person. With all due respect for one of the great Plato
scholars of all times, it makes no sense is mere rhetoric. Indeed, it makes perfect sense to
pursue bodily pleasures right up to the end if you find the state you are in when you experience
these more intrinsically satisfying than any other.
Conceptually, what Socrates need is an argument that more clearly identifies the self
exclusively with the subject of psychic states. An argument for the immortality of the soul
presumably serves this purpose, so long as the immortal soul is the self and is disembodied. If the
permanent or ultimate state of the self is disembodied, then at least there is some reason to hold
that states of the self that require a body are somehow less truly identifying. In the Apology and
Crito Socrates expresses agnosticism about the immortality of the soul. This agnosticism
prevents him from offering a non question-begging argument on behalf of moral absolutism.

2. Phaedo

Let us turn to the dialogue that is filled with ungainly and obscure arguments for
the immortality of the soul, Platos Phaedo. Attempting to prove the immortality of the
soul is a very odd project as evidenced by the fact that no philosopher known to us
outside of the Platonic tradition attempted to do this. Even within that tradition, all the
proofs to be found are variations on those found in the Phaedo, Republic, Phaedrus, and
Laws. Apart from anecdotal reports adduced as evidence of reincarnation and personal
religious revelation, it is very difficult to understand how a proof for immortality is
supposed to work. In a very general sense, we can reasonably interpret the question is
the soul immortal? as roughly similar to the question is piety what is dear to the gods?
or is virtue teachable? questions from the Euthyphro and Meno which ask about a

2
It matters not at all that I might persuade you to like ballet rather than football. Why bother?
3
Cf. my Socrates Absolutist Prohibition of Wrongdoing in Wisdom, Ignorance, and Virtue. Edited by
Mark McPherran. Edmonton: Academic Printing and Publishing Co., 1997, 1-13.
property or of an or Form. And just as in the case of those dialogues it
becomes clear that one cannot know whether something is a of an until one
knows what the is, so we may suppose that one cannot know whether immortality
belongs necessarily to the soul until one knows what the soul is. So much is fairly
straightforward. But there are certain complications.
First, as in the case of the definitions of piety or virtue Socrates principal data for
dialectical analysis are common conceptions or references of the words piety and
virtue. In the case of , the commonest of common conceptions is that soul is
whatever it is that differentiates a living being from a dead one. 4 Whatever that may be
is deeply obscure. Or more precisely, the interest of the interlocutors in the Phaedo in the
immortality of the soul is presumably focused on personal immortality. And the
connection between whatever it is that differentiates a living being from a corpse and
personal immortality is hardly obvious.
Second and relatedly, the sort of evidence that might be adduced to discover
exactly what the soul is and the sort of evidence for personal immortality seem diverse.
What was true in the middle of the 4th century is equally true now, when learned
discussions regarding the definition of death can be seen to have absolutely nothing to do
with philosophical speculation about personal immortality. Indeed, it is not at all clear
why the question of personal immortality should have anything to do whatsoever with the
question of soul as animal life. The Phaedo is deeply involved with evidentiary
strategies that unite the two questions.
Against the background that the soul is simply whatever it is that differentiates a
living thing from its dead remains, a proof for the personal immortality of the soul must
essentially do three things: (1), it must show that the soul is more than a property that
differentiates the living from the dead; it must show that it is an entity on its own, that is,
an entity that can exist when the body does not. (2), it must show that this entity is
identical with the person. (3), it must show that this entity, identical with the person, is
everlasting or indestructible.
It is sufficiently clear that the interlocutors in this dialogue are interested in their
own survival of death or personal immortality but it is far from clear what personal
immortality means. Presumably, a proof that we survive the destruction of our bodies
means that our bodies are not us. Or, stated more accurately, if we are immortal then
nothing that requires a body can be part of who we are essentially. 5
As we saw above, the soul or the person is the subject of pleasures and pains as
well as the subject of thoughts. One might suppose that a proof of the immortality of the
soul is either a proof of the survival of the subject of all the activities and states that a
soul is capable of or it must include an argument that some of these are dispensable for
personal identity. But this is clearly a false dichotomy. For not all that we are capable of
under certain circumstances is necessary for our personal identity under all
circumstances. For example, we might well need a body to feel pleasure but our identity
is surely not obliterated if we are no longer capable of feeling pleasure. It is hardly

4
See Jan Bremmer, Early Greek Concepts of the Soul (Princeton:1983) on common notions of the soul.
5
David Gallop, Plato. Phaedo (Oxford: 1975), 90, rightly observes that if such features as memory or
emotion are required for personal immortality, but are, at least implicitly, excluded from survival by the
philosophical arguments, then personal survival not only goes unproven in the Phaedo but is actually ruled
out.
surprising to learn that the arguments in the Phaedo take thought as the condition sine
qua non of personal identity. What we should find most deserving of close inspection,
however, is how a discarnate subject of thought and an incarnate subject of a broad range
of psychological states are taken by Plato to be related.
Socrates introduces his first proof for the immortality of the soul with an apologia
for philosophy. Here he makes one of his most arresting pronouncements: philosophy is
nothing but the practice for dying and being dead []
(64a5-6, cf. 67d7-10). On the basis of the agnosticism about the after life that Socrates
evinces in the Crito, this pronouncement, or at least the last conjunct of it, is nonsense. 6
Since the Crito and Apology also represent Socrates as devoted to philosophy, we can
either say that the Socrates of the Phaedo is Plato and the Socrates of the Crito is the
historical Socrates or else we can say that Plato changed his view about philosophy
between the time of writing the two dialogues. In either case, philosophy is here
conceived of in a remarkably new manner. It would seem best to take the phrase dying
and being dead to refer both to the metaphorical dying to bodily desires that the true
philosophers pursues and to actual physical being dead when the soul is finally released
from the body Before Socrates has offered a proof for the immortality of the soul, he is
represented here as staking his life on this claim about what philosophy is.
One plausible inference from this claim is that the death of the body is not the
death of the person and incarnate existence is only a prelude to discarnate existence.
Therefore, incarnate existence is not equivalent to the existence of a person. The relation
between the discarnate person and the incarnate version is initially unclear, but it would
be a serious mistake to assume without further argument that the possibility of discarnate
existence for a person entails a particular sort of dualism. It would be a mistake because
the usual devastating arguments against dualism, principally those pointing out the
implausibility of interactionism, do not obviously reach to a refutation of personal
disembodied existence. That is, a proof for personal immortality is not defeated by a
refutation of interactionist dualism if that proof does not depend on such dualism.
There are two fundamental points to be made in this regard, one stemming from
the discussion in the beginning of this paper and one that arises from the metaphysics of
the Phaedo. First, Plato tends to treat of the relation between the soul and the body as
that of a person and his possession. More particularly, as we have seen, this possession is
an instrument of subjectivity and agency such that Plato can say that the same person
who thinks feels pain. This we saw produced a seemingly insurmountable problem for a
claim that one ought to abandon care of the body for care of oneself. Second, the relation
between the discarnate person and his incarnate predecessor (or successor) is quite
naturally seen as analogous to the relation between an immaterial and separate Form and
its sensible instances If this should turn out to be in fact Platos view, then for one thing,
it is not interactionism that potentially threatens Platonic dualism, but the more general
metaphysical issue of how the instances of Forms can be related to the material or stuff
that manifests them.

6
At 64b9 the words suggest that the death philosophers seek is a metaphorical death. But
since the theme pronounced here is obviously the literal survival of the person in physical death, it is best to
take the conjunction dying and being dead to include both metaphorical and literal dying. The ambiguity
between metaphorical and literal dying is consistently reinforced throughout the entire passage up to the
introduction of the theory of Forms at 65d4.
 That philosophers long to die suggests, as Socrates says (64c4-8), that death is
nothing but the separation [] of the soul from the body. Given the phrase
dying and being dead and the above interpretation, the separation is exquisitely
ambiguous. On the one hand, if Socrates stipulates that by death he means
metaphorical separation, this hardly requires an argument. On the other hand, if he
means literal separation, then he has already assumed by definition the truth of (1) above,
namely, that the soul is a separate entity. The ambiguity, however, is not gratuitous. To
the extent that one finds it both possible and desirable to separate oneself metaphorically
from the body, the establishment of (1) is supported. A philosopher is best able to
appreciate the proof of personal immortality because a philosopher is most likely to
appreciate the truth of the premise that the soul is an entity literally separable from the
body. And literal separation is supported by metaphorical separation, which again the
philosopher is most likely to experience. In addition, since the Phaedo is a work of
philosophy addressed to the reader, as the reader does philosophy by studying the
arguments in the dialogue, he is engaged in the act of supporting the truth of the premise
of literal separation by means of metaphorical separation.
The following description of the philosopher who longs for detachment from
bodily concerns is, I would suggest, a constituent of the argument for immortality. It is
not, as is sometimes supposed, one of Platos typical diversionary perorations on
philosophy. The philosophers soul, Plato says, attains truth by reasoning
[] (65b9, c2). And it reasons best when it has the least possible
association with the body(c5-9). The soul of the philosopher utterly disdains the body
and flees from it and seeks rather to be alone by itself [] (c11-d2). I
think that the superficial interpretation of these words is that Plato is urging that bodily
concerns distract the philosopher from doing his work. If you are worried about your
clothes or your food or your job your mind will not be fully engaged in philosophy. This
is undoubtedly true enough so far as it goes and indeed is supported by the long speech at
66b1-67b5 where the distractions of embodiment for the philosopher are adduced. But
the subject of this work is the immortality of the soul. In order to prove this, part of what
must be shown is that the soul can exist alone by itself. This is what the philosopher
habitually shows as does the reader of this philosophical work.
Abruptly and portentiously Plato introduces Forms, Justice, Beauty, Good, and so on (65d4-7).
The philosopher, using his intellect alone by itself and pure tries to hunt for each of the things
that are [i.e., Forms], each alone by itself and pure (66a1-3, cf. 79d5-6). Thus, a parallel of sorts
between soul and Forms is erected. When the soul is alone by itself it is like the Forms.
According to the ambiguity that infects this pass age, the soul is alone by itself either
metaphorically when it thinks without reference to the body or alone by itself literally when the
embodied person dies. That there is critical difference between the two possibilities is evident in
the following passage:

Well now, it really has been shown us that if were ever going
to know anything purely, we must be rid of it [the body], and
must view the objects themselves with the soul by itself; its
then, apparently, that the thing we desire and whose lovers we
claim to be, wisdom, will be ours when we have died, as the
argument indicates, though not while we live. Because, if we
can know nothing purely in the bodys company, then one of two
things must be true: either knowledge is nowhere to be gained, or
 else it is for the dead; since then, but no sooner, will the soul be
alone by itself apart from the body. And therefore while we live,
it would seem that we shall be closest to knowledge in this way
if we consort with the body as little as possible, and do not
commune with it, except in so far as we must, and do no infect
ourselves with its nature, but remain pure from it, until God
himself shall release us; and being thus pure, through separation
from the bodys folly, we shall probably be in like company, and
shall know through our own selves all that is unsullied- and that,
I dare say, is what the truth is; because never will it be
permissible for impure to touch pure (66e4-67b2).

The ambiguity between literal and metphorical separation is evident in this passage. The
soul is alone in the primary sense when the person has died. It is alone metaphorically
when it is not infected with bodily nature. In the former state, knowledge is attainable;
in the latter we are as close to it as we may be when we consort with the body as little as
possible.
One would think that the plain sense of this text is that unqualified knowledge is
not possible for incarnated persons. 7 In any case, let us hypothesize for the moment that
this passage contrasts the knowledge that is possible in the discarnate state with
something else. Even more generally, suppose that discarnate existence enables some
type of cognition that is not available for anyone who inhabits a body. The parallel to the
relation between Forms and sensible images or instances is natural and striking.
Incarnate cognition is to discarnate cognition as sensible instances of Forms are to the
Forms themselves. Furthermore, we might reasonably suppose that the incarnate person
is analogously related to his discarnate self.
The characterization of the philosopher as striving to disengage from bodily
concerns in order to approach knowledge of Forms implicitly contains a moral claim as
well as a psychological and epistemological one. Engaging in philosophy supposedly
prepares one for being dead in way that nothing else does. After all, one would think that
preparing for being dead is about as easy as falling off a log. The cognition that the
philosopher strives for is transforming. One does not have to be a philosopher in order to
be an immortal soul. One has to be a philosopher in order to be a happy immortal soul.
Or at least, philosophy is required for the excellence that is rewarded with happiness.
Here the connection is surely even starker. One who has incarnately sought knowledge
has been transformed into somethng godlike. How exactly does this transformation
work? Whatever the answer to this question, implicit in parallel discussed above is a
threefold distinction: (1) the ordinary person thrown into the world; (2) the ideal striven
for, whether this be the life of philosophy or something else; (3) the discarnate result or
finished product.
If knowledge is going to result in self transformation, then this fact should tell us
something about what Plato regards the self to be. Ideally, a discarnate self engages in
contemplation of Forms. That is, what a self is ideally is a subject of thought completely
taken up with that thought. At 79d6, for example, wisdom [] is the state
[] of a soul that has come to be alone by itself [] with Forms.
It is a mistake to view the transformation that wisdom paradigmatically produces in the
7
Cf. Simmias words at 85c3-4
soul as we might view a set of beliefs assent to which changes a person. For example, if
someone comes to believe the doctrines of a religion or if he comes to believe that he will
die tomorrow, it is natural to speak of such a person as a changed man. But this sort of
change may be relatively superficial. That is, it may be exclusively related to behavior.
Since for Plato, however, the paradigm of cognition is not propositional belief, such a
way of conceiving of what Plato is doing is inadequate. Once we rid our minds of the
view that for Plato wisdom is something like encyclopedic knowledge it will be easier to
view the Forms as something other than the items in the encyclopedia. In any case, the
self-transformation of a person which is evident in the successful philosopher is much
more than behavior modification.
Plato views knowing non-representationally. The state of knowing is a state of
cognitive identification, that is, a state in which the knower becomes what he knows.
Thus, we cannot imagine the ideal self reading about or thinking about the truth where
about is a marker for the gap between thinker and truth. More particularly, the ideal
thinkers intentional objects are the Forms with which it is self-reflexively identical. That
is, knowing is a state ideally in which one is aware of the state one is in. On this basis, it
is not difficult to see that wisdom would be self-transforming. If the ideal is to be
eternally identical with Forms, incarnate cognition is self-transforming because is
approaches the discarnate ideal.
By contrast, incarnate cognition is basically representational. Representations are
images for Plato. So, one might suppose that the sort of self-transformation that one
undergoes incarnately is only an image of the transformation that is the discarnate result
of a successful philosophical life. And that is exactly what we discover. Many examples
are ready to hand. Socrates in the Crito believes that one must absolutely never do wrong
and this leads him to accept death with equanimity. But surely the transformation which
his beliefs have produced in him are not merely behavioral. Socrates refuses to leave
prison because of the man he has become where because really explains something.
The Socratic method of doing philosophy, aimed in part at getting people to understand
that they do not know what they think they do, is supposed to produce self-transformation
at least to some extent when it is successful. Callicles in the Gorgias disdains and rejects
the self-transformation that would occur if he assented to the outrageous claims made by
Socrates. And perhaps most importantly, the claim that virtue is knowledge (Gorgias,
460bd; Protagoras 360bd; Republic 340e, 342ab, 351ac; Theaetetus 176c) is in a way the
ultimate expression of the transforming power of cognition.
Cognitional states transform cognizers. In a way, it is as simple as that. False
beliefs presumably transform in a way opposite to that of their true counterparts.
Intentional wrongdoing harms the wrongdoer necessarily or inevitably because the false
beliefs which underlie that wrongdoing corrupt the believer. Or so I imagine Plato in the
Apology and Crito to reason. But along with the conviction that the soul is immortal
comes a claim regarding an ideal self, eternally imbued with truth. The ideal is the
control for understanding its images. Like all images of Forms, the incarnate self strives
to emulate its model. It is naturally impelled to cognition in general and belief states in
particular. False beliefs are to the soul what poison is to the body. They transform the
self by their content. For example, if one believes that ones body is ones self and not
merely a possession, such a belief will inevitably condemn one to live a non-
philosophical life.
 Since the soul has a life as well as brings life to the body, its beliefs are the
substance of that life. That is surely why Socrates declares at 89d2-3 that there is no
greater evil that can befall anyone than , hatred of argument. To renounce
argument is to sentence oneself to whatever false beliefs happen ones way. The
misologist doesnt renounce beliefs, of course, because beliefs are constitutive of his self;
he renounces his best chance for the acquisition of true beliefs. In his brief
autobiography Socrates expresses the conviction that those who face death with their
souls in a way purified [67c3, ] may have good hope for the
afterlife. This purification is identified with the metaphorical dying at 64a5-6. 8 The
purification appropriate for a thinker is of course cognitive in nature. The specifics of
cognitive purificaton are the false beliefs that are a condition of being born into a corrupt
society. As false beliefs are replaced by true ones, an inferior image of the ideal self is
replaced by a superior one. Superior images are, nevertheless, images. The philosopher
who believes, as does Socrates, that the death of his body does not end his life, is yet in a
state different from the ideal state. For example, the words and thoughts or internal
words with which he expresses to himself his belief in immortality are images. And as a
condition of incarnation, he is incapable of absolute disregard of his body or of
investment of his self as a subject of sensations, emotions, and so on. His incarnate life is
necessarily an image of his true self.
The inferior image, or at least a part of it, is related to the superior as the superior
is to the ideal. The virtue of non-philosophers is a kind of shadowy image
[] of the real thing (69b6-7). This is later described as popular and civic
virtue [] (82a10-b1). 9 The illusion of virtue is
produced by behaving virtuously without being virtuous. This occurs when virtuous
deeds are done exclusively for prudential reasons. It follows that the person who has
become virtuous as opposed to merely acting virtuously is transformed internally, not
externally. But it is still not sufficient to say that the transformation consists in his doing
virtuous deeds for the right reason rather than for the wrong reason. True enough. But
right reason is an empty phrase when it at once seems to suggest an ulterior motive and
no motive at all. That is why Plato understands a transformation of the self as the
transformation in cognition.

The AA (78b4-84b4) has not been treated kindly in the literature. In fact, it is
frequently dismissed as an embarrasment for Plato. Insofar as the argument relies on an
analogy or insofar as it is merely probabilistic, it certainly does not stand on its own as a
demonstration of the immortality of the soul. But this argument, like the others, does not
stand on its own. And, at its core, there is an argument which, far from being
inconsequential, is the origin of a family of arguments for the immateriality of the self.
These arguments are refined and elaborated upon by countless later Platonically-inspired
philosophers. They are still, in my view, worthy of interest.
The basic argument presented by Socrates is this 10:

1. Forms are invisible entities (79a6-7, cf. 65d8-11)

8
With Gallop, Hackforth, and Verdenius I take the reference here to the earlier discussion, not to some
ancient doctrine of purification.
9
Cf. Republic 365c4.
10
See Lloyd Gerson, Platonic Dualism, the Monist 69 (1986), 353, for this analysis.
 2. Whatever is an invisible entity, is invariant (79a9-10)

3. Cognition of Forms is possible (cf. 79d3-4)

4. Cognition of Forms is possible only for something with a

nature similar to Forms (assumed)

5. Therefore, that which cognizes Forms is or is similar to an

invisible entity.

6. Therefore, that which cognizes Forms is or is similar to

something invariant (79d5-6)

Clearly, the crux of the argument is premise 4. The RA has neatly argued that we must
have had cognition of Forms prior to incarnation. This is presumably what gives Cebes
reason to agree with Socrates assertion that the soul is an invisible, not a visible thing
(79b12-14). I take it that by invisible the interlocutors mean immaterial, although
this certainly does not follow necessarily. I assume this because of the inference from
invisible to invariant which is much more likely to be true if invisible means
immaterial. Since Plato certainly believes that the Forms are not merely invisible but
immaterial as well, there seems to be nothing against taking the argument in this way.
Why should we then believe that only an immaterial entity is capable of cognizing an
immaterial entity? Before we attemtp to answer this crucial question, let us begin by noting that
the argument evinces the same ambiguity regarding the possibility of incarnate knowledge that
we saw earlier. The relevant passage goes (79d1-7):

Whenever it [the soul] investigates [] alone by itself, it

departs yonder towards that which is pure and always existent
and immortal and unvarying, and in virtue of its kinship with it,
is always with it,whenever it is alone by itself and whenever it
may do so; then it has ceased from wandering and, when it is
around those objects, it is always constant and unvarying, owing
to its being in contact [] with such objects. And this
state of it is called wisdom, isnt it?

It is unclear from this passage whether it is presumed that it is possible toacquire

knowledge of Forms incarnately, as opposed merely to investigating them. On the one
hand, if the souls immateriality is sufficient for it to be able to cognize Forms, then it is
unclear why such cognition might be thought unavailable for the soul in its incarnate
state. On the other hand, incarnation is not inconsequential to the self, as we have seen.
It is owing to incarnation that cognition of Forms is at the very least exceedingly difficult.
In any case, the souls kinship[] with the Forms is the condition in
virtueof which it is able to have some sortofcognitive contact with them. The question
we are concerned with is why kinship with the immaterial is necessary for cognition of it.
This question amounts to nothing less than the question of what exactly knowledge is
supposed by Plato to be. Since knowledge is not thematized until the Republic and then
the Theaetetus, we cannot expect an entirely satisfactory answer to our question here.
Nevertheless, there is quite a bit that can be gleaned from the Phaedo itself.
First, let us observe a small point from the RA. It is that the ability to give an
account of the objects of knowledge, that which Simmias thinks no one but Socrates can
do, must be distinct from knowledge itself, which all agree everyone has to have had
discarnately. Even the sort of knowledge that is recollection is not equivalent to the
account, even though it be a necessary condition for it. I stress thissmall point because it
tends to indicate that knowledge is a state that does not consist in expressing or thinking
of representations of the knowable. Saying the words Equality is a dyadic relation of
such and such a sort or Justice is the proper operation of the parts of the soul or any
other real or putative accountof a Form is not equivalent to having knowledge of it. For
this reason, even if for no other, knowledge of Forms is not proposition, if this means that
knowledge is expressing or thinking of some representation or description of an
immaterial eternal entity. I take this to mean minimally that knowledge is for Plato non-
representational. But that is a very difficult idea for anyone to understand who comes to
the question of Platos view of knowledge from a background of modern epistemology.
For virtually all accounts of knowledge in at least the analytic tradition of modern
epistemology are representationalist.
Plato does not consider a representationalist account of knowledge seriously until
the Theaetetus where he rejects it, or so I will argue. That he does so in the Theaetetus of
course opens up the possibility that he had already considered and rejected such an
account in the Phaedo. I am sympathetic to this possibility, especially if it makes theAA
a better argument. But whether Plato had by the time of writing the Phaedo already
considered and rejected a representationalist account of knowledge, some of the elements
of a non-representationalist account may be assumed, with only a modicum of charity, to
be present.
A non-representationalist account of knolwedge holds that knowledge is a state
constituted by the knowable, not merely caused by it. The presence of knowledge is
equivalent to the presence of the knowable. The word presence here must of course be
understood literally not metaphorically, although the presence of an immaterial entity is
not the same thing as the presence of a material entity. One excellent reason for holding
that knowledge is non-representational is that knowledge is an infallible state and that I
know but I may be mistaken misconstrues the nature of knowledge altogether. If
knowledge were representational, infallibly could in dprinciple not be preserved because
there would be no way of inferring froma representational state, any objective state of
affairs. There is no sCp (where s is a subject, C any form of cognition, and p the object
cognized) that entails p, so long as C is a representational state. This is so because of the
basic distinction between subject and object or subjectivity and objectivity so long as
representational states are a part of the subjective in relation to the objectioe. I shall just
assume and not argue that there is no entailment relation between any subjective state and
any object ive stateof affairs as these are ordinarily conceived.
If knowledge for Plato is non-representational, why then should we suppose that
only an immaterial entity can know an immaterial entity. A moments reflection should
lead us to object that Plato himself, insofar as he holds that immaterial Forms are present
their instances is committed to denying that an immaterial entity can only be present to an
immaterial entity. As Socrates says later in the Phaedo, something is beautiful because it
participates in Beauty (100c5). Why is this not analogous to: sKB because it participates
in B? Because the beauty in that which participates in the Form of Beauty is, as we have
seen, diminished in reality or defective. Of course, it is materiality that inevitably
produces the defect. If the presence of a Form to a knower were the presence of the Form
defectively, then exactly what the non-representational account of knowledge intended to
insure, namely, infallibility, would be lost. One would know defective beauty not
Beauty. So, it seems we can save Plato from the contradiction of saying that an
immaterial entity can and cannot be present toa material entity, but that still leaves us
with the question of why only an immaterial entity can know an immaterial entity.
Perhaps the question can be answered most perspicuously if we consider a
generalized reductio argument. Let the soul be a material entity and let us ask how an
immaterial entity could constitute it or be present to it. For example, let the knowledge of
the Form of Equality be equivalent to the presence of thisForm to a material entity. of
course, the presence of this Form will not make the material entity literally equal to
anything. Then, the presence of Equality would be like the presence of an element in a
compound, like adding butter to the batter or an attribute, like shaping a lump of batter. I
intend the alternative v ersions of presence to be exhaustive, so we can understand the
attribute in any sort of way or the presence of the element as altering thecompound or
not.
I can think of at least three reasons why such models cannot account for the non-
representational knowledge of immaterial enities. First, and most obviously, a material
element could only be a representation of an immaterial entity and though an attribute can
in some sense be said to be immaterial, an attribute depends for its existence on what it is
an attriute of whereas a Forfm does not. And in that case the attribute could only be a
representation of a Form, not the Form itself. Second, the presence of the element or
attribute is either equivalent to the material and deficient manifestations of a Form or not.
If it is, then again knowledge will be of a deficient instance not of the Form. If it is not,
then it is hard to see how its non-representational presence is supposed to be different
from its absesnce. Third, the mere presence of an element or an attribute is not enough
for knolwedge if, in order to know, one must know that one knows. And this seems a
reasonable condition for knowledge. If knowledge is the presence of an element or an
attribute, then the knowledge that one knows will be the presence of another element or
attribute. But it is far from clear how the presence of another element or attribute is
supposed to be equivalent to the awareness of the presence of the first. Thus, if knowing
a Form is the presence of an element or attribute, then knowing that you know the Form,
there is another Form present and whose presence constitutes knowledge that the first is
present. This certainly seems hopeless.
If will perhaps be objected to the last mentioned reason that in the RA the
interlocutors agree that they have acquired knowledge of Forms but they do not have that
knowledge presently, so they know but do not know that they know. The correct
response to this objection is I think that the interlocutors only believe that they believe
truly that they must have known the Forms. They do not know them, else they could give
an account of them. This does not quite solve the problem. If Simmias and Cebes
believe truly that they have known Forms and that this knowledge can be recovered, then
presumably in some sense they know Forms now. They are in the state that a knower is
in but they do not know that they know. As I have already suggested, for Plato all forms
of cognition other than discarnate are defective or derivative versions of the paradigm
just as sensibles are defective versions of Forms. The state of an incarnate individual
who believes truly that he did know, say, the Form of Equality, but now cannot give an
account of it, is defective with respect to the knowledge of that Form. By derivative I
mean that the account of the knowledge possessed by Simmias and Cebes must defined in
terms of knowledge in the primary sense, that is, the knowledge that the discarnate soul
possesses. So, if they must know Forms in order to recollect, that knowledge would
presumably be accounted for dispositionally. Thus, saying that they know means that if
they were suitably instructed they would activate or realize or be aware of or simply
know Forms.
It will perhaps be objected more generally that if knowledge is what Plato
assumes it is, then knowledge is not possible, not just for incarnate individuals, but for
anyone incarnate or discarnate. This is in fact the position of the Sceptics in Platos
Academy, in my view. I think that Platos only non question-begging response to this
objection is to recur to the RA. Unless we had knowledge of the Forms full-blown
unqualified knowledge it would not be possible for us to make judgments of the
deficiencies of sensibles. But we do make such judgments. Hence, we do possess
knowledge. But we are not aware that we possess it. This is admittedly a strange sort of
argument, made even stranger by the claim that we know but do not know that we know
the Forms against which we judge sensibles to be deficient.
Ultimately, a non-representationalist account of knowledge that claims that
knowing a Form requires that one know that one know that Form must insist that
knowers cannot be material entities. Material entities cannot know that they know. The
intuitive reason for this is that in order to know that you know, the subject of the first
know must be the same as the subject of the second know. Inother words, if I know
that I know, this is not the same thing as my knowing that you know. Indeed, it is
incoherent to say that I know that you know if knowing entails that you know that you
know. I am the only one who knows that I know. In a material entity, the material part
that knows must be different from the material part that knows that (the other part)
knows. Of course, if we take the word identical inn a suitably capacious manner, we
can say that it is the identical machine that knows something and knows that it knows
something. We could also say that when one member of a family is sleeping and one is
awake, the identical family is sleeping and awake. The strict identity that is required for
knowing is not possiblefor material entities.
I do not wish to claim that this argument is explicitly in the AA. I claim only that
the account of knowing presmed in the AA is non-representationalist and that Plato
believes that therefore no material entity could know Forms. Of the three reasons I gave
above for insisting that no material entity can know Forms, I am fairly confident that the
first two underlie the AA. I believe the third is present in the Theaetetus, but I am not
sure it is present here.
There are basically three ways in which the AA can be denied. First, one can try
to show that knowledge is not what Plato says it is. Specifically, one can try to show that
a non-representationalist account of knowledge is an account of nothing real. Second,
one can try to show that material entities can have knowledge as Plato understands.
Third, one can agree with Plato regarding knowledge, but claim that it is not possible for
humans to have it. That is, it would take an immaterial entity to have knowledge, but
since humans are not immaterial entities, they cannot have it. Perhaps gods or angels
have it, but not us. I cannot here undertake a discussion of the pros and cons of each
alternative strategy. I am content if I have shown that the AA raises a serious matter and
that underlying it is a powerful argument.
After agreement is reached on the AA by the interlocuors, a subordinate argument
is introduced (79e8-80a5):

Now look at it in this way, too: when soul and body are present
in the same thing, nature ordains that the one shall serve and be
ruled, whereas the other shall rule and be master. Here again,
which do you think is similar to the divine and which to the
mortal? Dont you think that the divine is naturally adapted for
ruling and leading, whereas the mortal is adapted for being ruled
and for serving?

Cebes readily concurs that the soul resembles the divine and so, presumably, the
immortal. I say this is a subordinate argument because it does not address immateriality,
although this may be inferred from the contrast of the soul to the body. What the
argument does in particular is introduce the relation of the soul and the body with respect
to governance. The rule of the body by thesoul is natural but not always the case.
Indeed, the souls very presence in the body as in a tomb suggests as a condition of
incarnation a state contrary to nature (cf. 82e-83a). I note in passing how typically
Platonic it is to invert the ordinary conception of what is natural.
Socrates presses the analogy of the soul to the immortal and the body to mortal. Since
even parts of the body, like bones and sinews, are practically immortal, how much more likely is
it that the soul itself is immortal (80cd). But now he seeks to differentiate the discarnate souls of
those who have pursued a life of philosophy from the souls of those who have allowed their
soulsto be contaminated with their bodies (80e-81c). the latter, owing to their previous
attachments to their bodies, are doomed to wander about tombs and graves until eventually they
are reincarnated into the type of animal body that suits their previous degenerated characters
(81c-82a). Those who have followed neither philosophy nor lives of wickedness but rather lives
of popular, that is, civic virtue (82a11), will experience a happier fate, first being reincarnted
into tamer animals and then eventually back into human beings (82a10-b8).
Socrates has moved from an argument for the immortality of the soul to an argument that
ones discarnate fate flows from ones incarnate career. The latter is expressed in terms of the
degree of attachment to bodily desires. This argument seems to fly in the face of the
implicationof the RA that incarnation has nothing to do with personal identity. It is not so,
because the consequences of incarnate life do not change what we are ideally, that is to say,
discarnately. Indeed, it is because we are ideally knowers,akin to gods, that an incarnate life
devoted to something other than philosophy results in reincarnation. The failure to become what
we are is what Plato represents as being weighted down with bodily elements (81c10), co-
operating especially in ones imprisonment (82e6-7), and being made corporeal (83d5) and
contaminated by the body (83d10).
That the soul can be corrupted by the circumstances of incarnation in general is readily
understandable and surely this is all that being made corporeal and like metaphors convey.
What it means for the soul to share opinions [] and pleasures with the body (83d7-8)
and ultimately to adopt its false opinions is somewhat less understandable. But let us recall that
Platos dualism does not require him to view the body as having opinions. Bodily pleasures that
seduce are the pleasures of the same person who does or does not desire to submit to the
seduction. Although I shall reserve until the next chapter the discussion of the sort of conflict this
occasions, it is here most relevant to stress that the incarnate self that is subject to seduction is an
imperfect or incomplete creation. This imperfect self who submits or does not submit or submits
occasionally as a necessary consequence of incarnation is an image of its perfect exemplar just as
the equal logs and stones manifest an image of the Form of Equality and just as the opinions this
self holds, true or false, are images of knowledge of the immaterial knowable. In one crucial
respect, however, the empirical self is disanalogous to the other images. Only this image can
become aware that it is an image and only it can assimilate itself to its exemplar. This process of
assimilation is named by Plato philosophy
The Phaedo gives a powerful portrait of immortal persons and their incarnate
images. The soul imprisoned in the body is a person in thrall to those desires of his that
he could not have without a body. The discarnate person is a knower, self-reflexively
contemplating the Forms with which he is, literally, identified. Thus, at one extreme, the
image is mired in all the individuality of embodiment; at the other extreme, the
paradigmatic self has shed all individuality by embracing and becoming the universal.
Between the two extremes is the drama of the life of philosophy that is, a life that consists
in gradually divesting oneself of ones imaginary existence vs. a counterfeit life, that is, a
life mired in images. Immortality ups the ante. For the philosopher, practice for dying
amounts to practice for death or the ideal state of disembodiment. For everyone else,
disembodiment is the experience of unimaginable self-alienation.