NICHOLAS D.

SMITH

PLATO’S ANALOGY OF SOUL AND STATE

(Received and accepted 5 October 1998)

ABSTRACT. In Part I of this paper, I argue that the arguments Plato offers for the tripar-
tition of the soul are founded upon an equivocation, and that each of the valid options by
which Plato might remove the equivocation will not produce a tripartite soul. In Part II, I
argue that Plato is not wholly committed to an analogy of soul and state that would require
either a tripartite state or a tripartite soul for the analogy to hold. It follows that the heart of
the analogy is not to be found in the comparison of the Kallipolis and its three parts to the
soul conceived as tripartite, but rather must be supposed to reside in some other connection
between the ways in which justice characterizes states and souls, and I will suggest what
this other connection consists in.

KEY WORDS: analogy, justice, moral psychology, Plato, political theory, Republic, soul,
utopia

At the beginning of Book II of Plato’s Republic, Glaucon and Adeiman-

tus express dissatisfaction with Socrates’ refutation of Thrasymachus, and
challenge him to offer a more persuasive proof that justice is preferable to
injustice. But at the end of Book I, Socrates compared the conversation
with which Glaucon and Adeimantus are so unimpressed to a feast of
gluttons, where the rude diners attempt to devour each new dish before
properly enjoying the preceding one (354a12–b5). To have conducted the
inquiry properly, Socrates proclaimed at the end of Book I, would have
been to discern first what justice is before turning to the next “dish”
and attempting to discern whether it is preferable to injustice or not
(354b5–c1).
In giving his reply to the renewed Thrasymachean challenge mounted
by Plato’s brothers, then, Socrates ensures that there will be a proper
“feast” by turning first to the question of what justice is, before any attempt
is made to decide whether justice or injustice are preferable. Famously (or
notoriously), to accomplish this goal, Socrates proposes the analogy of
large and small letters (368c7–d7), in which the near-sighted are given the
opportunity to read the same message in large letters, before attempting to
discern it written in small letters. This situation, in turn, is compared to the
attempt Socrates proposes, to “see” justice first in a state, where it will be
“writ large,” before trying to figure out what it is “in smaller letters” in a
soul.

The Journal of Ethics 3: 31–49, 1999.

© 1999 Kluwer Academic Publishers. Printed in the Netherlands.
32 NICHOLAS D. SMITH

Plato’s Socrates and his interlocutors accept from the outset and without
argument that justice in the state and soul are like the same letters inscribed
in a large and a small format, and this assumption has been widely chal-
lenged and questioned by Plato’s critics.1 In this paper, I will neither object
to this assumption nor defend it, but instead will show that the specific way
in which Plato articulates his analogy, and the way in which this analogy is
generally supposed by scholars to work – if it works at all – is irreparably
logically flawed.2 Specifically, I will argue in Part I that the arguments
Plato offers for the tripartition of the soul are founded upon an equivoca-
tion, and that each of the valid options by which Plato might remove the
equivocation will not produce a tripartite soul. The result of this argument,
plainly, is that the way in which Plato actually makes the analogy of state
and soul – by appealing to the similarities between a tripartite state of
rulers, warriors, and craftsmen, in the Kallipolis, to a tripartite soul of
reason, high spirit (or thumos), and appetite – cannot validly be made.
(In the remainder of this paper, I will call this specification of the soul-
state analogy the “3-3 specification”). In the second part of my argument,
however, I will argue that Plato reveals a lack of full commitment to the
3-3 specification of the analogy without thereby calling the analogy itself
into question. If this is so, then it follows that the heart of the analogy is
not to be found in the comparison of the Kallipolis and its three parts to
the soul conceived as tripartite, but rather must be supposed to reside in

1 For examples, according to David Sachs, “A Fallacy in Plato’s Republic,” The Philo-
sophical Review 72 (1963), pp. 141–158 and Gregory Vlastos, “Justice and Happiness
in the Republic,” in Gregory Vlastos (ed), Platonic Studies, second Edition (Princeton:
Princeton University Press, 1981), Plato equivocates on the conception of justice, though
they offer different accounts of the equivocation and come to different conclusions about
how serious such equivocations are to Plato’s overall argument.
2 Other criticisms of Plato’s psychology in the Republic may be found in the literature,
as well. For example, Christopher Bobonich, “Akrasia and Agency in Plato’s Law and
Republic,” Archiv für Geschichte der Philosophie 76 (1994), 3–36 and Terence Irwin,
Plato’s Ethics (Oxford: Oxford University Press, 1995), pp. 217–222, argue that Plato’s
conception of the parts of the soul mistakenly attributes to them a kind of agency which
belongs properly only to the individual of which they are parts. In this paper, I do not
address such concerns directly, but I would note that in so far as my arguments give some
reason to think that Plato was not wholly committed to his partitioning arguments, he
cannot either be regarded as wholly committed to the sorts of errors Bobonich and Irwin
have identified. Yet other problems regarding Plato’s psychological account (and thus its
application to the soul-state analogy) are noted in Mitchell Miller, “Platonic Provocations:
Reflections on the Soul and the Good in the Republic,” in D. J. O’Meara (ed.), Platonic
Investigations (Washington, DC: The Catholic University Press), pp. 163–193, and Lloyd
Gerson, “A Note on Tripartition and Immortality in Plato,” Aperion 21 (1987), pp. 81–96. I
regret that I cannot directly address or incorporate all of their arguments or the issues they
raise in this paper.
 PLATO’S ANALOGY OF SOUL AND STATE 33

some other connection between the ways in which justice characterizes

states and souls, and I will suggest what this other connection consists
in. I will conclude by showing that the failure of Plato’s psychological
arguments that I find in Part I of this paper, which is surely fatal to the
3-3 specification of the analogy, is not fatal to the analogy itself, or even
to the political aspects of Plato’s Republic, even if it would require that
substantial changes be made in Plato’s psychological theory.

I. P LATO ’ S A RGUMENTS FOR P SYCHIC D IVISION

At 434b9–d1, Socrates and Glaucon agree that they have discovered what
justice is in the city:
For the three of these classes to interfere in one another’s business or to exchange with
each other, then, is the worst damage to a city and might rightly be called the worst evil.
Indeed.
And don’t you call injustice the worst evil to one’s city?
Of course.
So this, then, is injustice. Moreover, to say it the other way around, for the wealth-making,
auxiliary, and guardian classes to mind their own business and do their own in the city, is
the opposite – justice – and is what makes the city just.
It is this and nothing else, I agree, he said.

The conception of justice expressed in this passage links it directly to the

tripartite class structure of the Kallipolis, and so it is not surprising that
when Plato’s Socrates goes on to consider what justice is in the soul, we
find him immediately looking for three parts of the soul to correspond to
these three classes in the Kallipolis.
The city was believed to be just when the three natural classes within it each did its own,
and it was temperate and courageous and wise because of other conditions and states of
these classes.
True, he said.
And so, my friend, we will evaluate an individual as rightly called by the same names as
we give in evaluating a city, if he has these same forms in his soul, and the same conditions
in them.
Quite necessarily, he said. (435b4–c3)

The general principle, which Socrates employs in each of the partition-

ing arguments, is announced at 436b8–c1:3
3 I am indebted, in much of the analysis that follows, to suggestions by Fred D. Miller,
Jr., and to subsequent discussions with a former student, Michael Fiebig, whose noble and
thoughtful persistence in trying to defend Plato against my criticisms helped me to sharpen
those criticisms considerably.
34 NICHOLAS D. SMITH

Plainly, the same thing cannot do or undergo opposites in the same respect4 and in relation
to the same thing, at the same time. So, if this should ever happen in these things, we will
realize that they weren’t the same, but many.

Socrates immediately goes on to introduce a paradigm case for the appli-

cation of this principle: the same thing at the same time in the same
part of itself cannot both be at rest and also move (436c5–6). From
the way in which he articulates the principle, and immediately clarifies
his meaning with this example, we might well understand Socrates as
committing himself to nothing more controversial than a principle of
non-contradiction:
(X) The same simple object cannot at the same time and with regard to the same things
perform contradictory actions or undergo contradictory changes.

This interpretation of the principle is especially inviting if we construe

moving and resting as contradictories – that is, where either “resting” just
is “not moving” or “moving” just is “not-resting.” But Socrates actually
states his principle in a way that invites an alternative understanding, by
stipulating that it is not opposite actions or modifications that a thing
cannot do or undergo, where by “opposite,” we do not mean “contra-
dictory.” Pushing and pulling, for example, are not contradictories, but
opposites, and this may be the sort of model that Plato has in mind.5
(O) The same simple object cannot at the same time and with regard to the same things
perform opposing actions or undergo opposing changes.

Evidence that (O) and not (X) should be our interpretation of the parti-
tioning principle may be seen from Socrates’ subsequent list of examples
of the sorts of “opposites” that he has in mind: “assent and dissent, desir-
ing to have something and rejecting it, drawing in and thrusting away”
(437b1–5). This list, however, requires interpretation, as we shall soon see.
The problem comes when Plato seeks to apply his partitioning principle
to the specific cases he gives to divide the soul, for the relevant cases either
do not appear to fall under the above interpretations of the principle – or
if they do, the principle will turn out to do considerably more than Plato
seems to want it to do. To see that this is so, let us begin by reviewing
Plato’s first partitioning argument, which is intended to show that the soul
may be divided into at least two parts: reason and appetite.
4 For this reading of κατ ὰ τ αυτ óν at 436b8, see R. F. Stalley, “Plato’s Argument for
the Division of the Reasoning and Appetitive Elements within the Soul,” Phronesis 20
(1975), pp. 110–128.
5 Fred D. Miller, Jr. argues for this interpretation in Fred D. Miller, Jr., “Plato on The
Parts of The Soul,” in Jan Max van Ophuijsen (ed), Plato and Platonism (Washington, DC:
The Catholic University of America Press, forthcoming).
 PLATO’S ANALOGY OF SOUL AND STATE 35

At 439a1–e1, Plato considers the case of someone who is thirsty, but

refuses to drink. But this would be a case to which we could apply (X)
only if it would be true to say of the person both that she did and also
that she did not wish to drink. As tempting as it might be to say this about
her, it would not be true, however. We often do describe our negative or
aversive desires in this way: “I do not want to drink,” but it is important to
distinguish between cases in which we wish to proclaim a lack of a certain
desire (which would be the contradictory of having the relevant desire) and
cases in which we have a negative desire, or an aversion to something, for
example, where we have very real desire not to drink. Non-sentient entities
lack any desire of any sort; so it is true to say of a thumbtack, for example,
that it does not want to drink. It is most certainly not true, however (assum-
ing my assessment of thumbtacks as non-sentients is correct, of course),
ever to say that a thumbtack wants not to drink. Thumbtacks lack desires,
but never have aversions.
What appears to be happening in Plato’s case of someone who is thirsty
but refuses to drink, then, does not fall under principle (X) because it does
not identify a case in which the person both has and lacks the relevant
desire. Instead, the person seems to have two desires, which appear to
oppose one another. Accordingly, it would appear that (O) is a more likely
rule to apply in this case.6
One of the most notable – and perhaps also most troublesome – aspects
of psychological states is their “intensionality.” The objects to which such
states refer, it is often said, are given “under description.” Accordingly, it
is one thing to say of Oedipus that he sometimes wished to enjoy marital
relations with Jocasta, and quite another to say that he sometimes wished
to enjoy marital relations with his mother. We may fairly suppose, given
Sophocles’ telling of the tale, at least, that the former is true and the latter
is false, about Oedipus.
Once we recognize this feature of desires, however, two distinct possi-
ble interpretations of (O), as it would apply to desires, are open to us,
depending upon what we are willing to count as “the same object.” In one
sense – the extensional sense – the woman whom Oedipus desires is the
same object as Oedipus’ mother. In another sense – the intensional sense
– these are different objects. Accordingly, we can understand (O) to apply
to cases of extensional identity, or to cases of intensional identity. Let us
call these alternatives “Oe” for the extensional case applied to desires and
“Oi” for the intensional case applied to desires:

6 Anthony Kenny appears to understand Plato’s partitioning principle in this way, as

“the principle of non-contrariety” [Anthony J. Kenny, “Mental Health in Plato’s Republic,”
Proceedings of the British Academy 55 (1964), pp. 229–253].
36 NICHOLAS D. SMITH

(Oe) The same part of the soul cannot be responsible for both desiring and also having an
aversion to the same object at the same time, even if the object is taken, in the desire and
the aversion, under different descriptions.

(Oi) The same part of the soul cannot be responsible for both desiring and also
having an aversion to the same object at the same time, taken under the same description.

Now, on the one hand, (Oi) looks like a very plausible principle – after
all, it is difficult to imagine how a person might both desire and also have
an aversion to the same thing at the same time, where the thing desired and
shunned was represented in both the desiring and in the shunning in the
very same way. If such cases are possible at all, it seems plausible to think
that they would only be possible for someone with a wholly bifurcated
desire system, as for example, we might find in someone suffering from
multiple personality disorder.
It is anything but obvious, however, that this is the sort of case Plato has
in mind here. He doesn’t tell us all the details about why his thirsty person
refuses to drink, but he does mention that the case he has in mind is one
where the person has some reason to avoid drinking, while what makes the
person desire to drink is “the result of affections of diseases” (439d1). It
looks as if the case, then, is one in which the person’s desire to drink takes
the intended object (say) as “thirst-quenching,” whereas the same object
is taken by the opposed reason as “unhealthy.” This, then, is more like the
case with Oedipus I mentioned earlier – the attraction and the repulsion
represent the relevant object in entirely different ways. If this phenomenon
falls under any principle, then, it must be (Oe).
But as a partitioning principle, (Oe) is too strong, and if it is this princi-
ple that Plato wishes to employ to divide the soul, his argument threatens
to atomize the soul. Plato wants it to turn out that all examples of thirst,
hunger, and lust – and perhaps also all desires for material acquisition –
will belong to the appetitive part of the soul. All anger, and presumably all
desires for honor, will belong to the high spirit. Calculation and reasoning
belong to the rational part. But cases may readily be found in which exam-
ples will fall under (Oe) that would require Plato to partition the soul into
more than three parts.
Consider our aversion to bodily pain, for example. To which part of the
soul would Plato think this belongs? Parity of reason would suggest that
it should belong to the appetitive part, since the desire for bodily pleasure
seems entirely to belong to this part. If this is so, then how would (Oe)
apply to this case: Jones finds himself caught between a strong desire to
drink the water he sees in front of him, on the one hand, and his aversion
to the pain he anticipates if he were actually to attempt to drink the water,
since he notices that it is boiling hot. He desires to drink, but holds back.
 PLATO’S ANALOGY OF SOUL AND STATE 37

Now, perhaps Plato would count this as a case of reason opposing

appetite, because it is a calculation involving anticipated harm that opposes
the thirst. But surely we can imagine a case in which Plato could not count
the relevant aversion as a feature of reason – for example, where Jones has
sound medical evidence that the boiling water will actually serve to cure
some infection in his mouth, and the burns, though painful, will cause no
permanent damage. In this case, the aversion to the pain would actually
work against Jones’s reason, and surely Plato would have to count such
aversions, therefore, as opposing both Jones’s reason and Jones’s desire
for drink.
Moreover, it makes no sense to suppose that such aversions must belong
to the high spirited part of the soul. Even if we tried to defend Plato’s
account by supposing that the high spirit was the “part” to which they
belonged, we can just as easily add to our account not only that Jones
has good medical reasons to overcome his aversion to the pain the boiling
water would cause, but also that Jones saw that refusing to drink, in these
circumstances, would count as an example of disgraceful cowardice. In
such a case, we would have to understand Jones’s aversion as opposed
to elements of all three of Plato’s psychic parts. Accordingly, then, if we
apply (Oe) to this case, we must find a fourth part of the soul. Nor is there
any hope that the process will stop here, for it seems obvious that in the
right circumstances, literally any of our desires might come into conflict
with any of our aversions, and the same kinds of scenarios – in which
support for either desire or aversion can be found in reason or high spirit
– will always be possible. For example, I am thirsty but to get a drink I
must leave my warm bed and suffer the cold in the room; 7 or, I am hungry,
but there is no food in the house and I can’t bear the thought of fighting
through traffic just to go to a store; or, I find myself sexually attracted
to someone, and at the same time find myself repelled by the scent of
garlic on her breath. If such examples are enough to partition the soul,
to accommodate all the ways I might find myself pushed and pulled at the
same time in regard to the same object-in-the-world – even if my attraction
and my repulsion are not focused on the same aspect of that object, as it
were – then it is difficult to see how the process of partitioning will not go

7 Terry Penner, “Thought and Desire in Plato,” in Gregory Vlastos (ed), Plato II: Ethics,
Politics, Philosophy of Art and Religion (New York: Anchor Books, 1971), pp. 108–111,
113–116 and J.F. Tiles, “The Combat of Passion and Reason,” Philosophy 52 (1977),
p. 322 discuss this case, with differing conclusions. Tiles appears to assume that Plato’s
partitioning principle is what I have called Oe.
38 NICHOLAS D. SMITH

on indefinitely. In any case, it certainly will not end after only Plato’s three
parts have been identified.8

II. J USTICE IN THE S TATE

As we have seen, the text makes it perfectly plain that Plato understood the
analogy of state to soul in a way that presumes what I have called the “3-3
specification” of the analogy. If I am right about the failure of Plato’s argu-
ments for the partitioning of the soul, however, Plato does not have a proper
claim to this specification of the analogy, for he has no good reasons for
dividing the soul into three parts, which is required by the analogy. But the
text (happily) also supplies us with compelling reason for supposing that
Plato did not think that the 3-3 specification of the analogy was required for
the analogy to succeed, since he shows that he is not completely committed
to either of the “threes” in this specification, in circumstances where the
analogy itself is not, thereby, called into question.
Let us consider, first, the tripartite division of the state. One of the
consequences of supposing that the analogy requires the 3-3 specification
is plainly that no state that is not divided up into three parts could satisfy
the analogy. The purpose of the analogy, recall, is to make it easier to
discern what justice is, by first trying to see it “writ large” in a state,
before looking for justice in written in the smaller letters of a soul. If Plato
supposed that this analogy required a 3-3 specification, then,9 we should
not expect to see him “looking for justice” in a state that is not divided
into three parts. And yet, this is exactly what we find him doing – long
before he has introduced the three parts of the Kallipolis – at 371e, when
Socrates and Adeimantus have completed what Glaucon later disparages
as no better than a “city of pigs” (at 372d4–5). Despite Glaucon’s disgust,
however, Socrates seems prepared to “look for justice” in this very basic
city.
8 I have only given an example that would partition Plato’s appetitive part of the soul, if
we apply Oe. But I believe similar examples could be given to partition the other parts, as
well. For example, consider a case where one finds oneself attracted to the honor one could
win in some competition, but honorably repelled by the fact that the only hope of winning
one had would be by using banned performance-enhancing drugs. Or, one finds oneself
attracted to a certain philosophical doctrine’s explanatory potential, while also feeling
doubts about its foundational assumptions. Such cases, I suggest, would partition the high
spirit and the rational parts of the soul, respectively, if we applied Oe.
9 See Daryl H. Rice, A Guide to Plato’s Republic (Oxford: Oxford University Press,
1998), p. 58: “Socrates argues that if his analogy between the city and the soul of the
individual is to be helpful, cities and individuals must be morphologically similar; that is,
they must have the same basic structure, or form.”
 PLATO’S ANALOGY OF SOUL AND STATE 39

Well then, Adeimantus, has our city grown to completeness?

Perhaps.
Where, then, in it, can justice and injustice be found? With which of what we considered
did it come to be?

When Adeimantus admits that he doesn’t know the answer, Socrates

exhorts him not to give up and to consider the question further (372a3–4).
Even after Glaucon has shown his contempt for this city, Socrates pauses
to praise this basic city one last time, before going on to consider the
“fevered” city that Glaucon wishes to consider, by calling the more basic
city the “true city” (372e6) and the “healthy one” (372e7).
Socrates’ willingness to look for justice within this first city – which is
not only expressed but then emphasized by his exhorting the reluctant and
puzzled Adeimantus and then by his remarkably positive final appraisal,
to Glaucon – would be inexplicable if Plato wished to show us that justice
in a city had to be understood in terms of the proper functioning of three
classes within a city. The city that Glaucon calls a “city of pigs,” plainly,
has several different functions to be performed, but all would fall within
what is identified as the class of craftsmen in the Kallipolis. So, either there
are several more than three classes in the basic city, or else there is only
one class, depending upon how one wished to individuate classes. None
the less, Plato’s Socrates is entirely ready to look for what it is in this city
that makes it just.
This basic city is disparaged in much of the scholarly literature. Julia
Annas, for example, speaks about how this city is “allowed to develop, and
becomes corrupt,”10 and claims that this city is fatally flawed because it is
“built up purely on the basis of self-interest.”11 Annas believes that Plato,
in fact, reveals his lack of interest in this city by having Socrates stress
that he is looking for the origins of justice and injustice in this and the
“swollen” city which follows it, and comments,
This suggests that both the first, ‘true’ city and the corrupted city are put forward for us
as models of human nature in association which display both justice and injustice, to be
contrasted with the ideally just city, which displays only justice.12

Annas concludes that, in fact, “Plato has not given the first city a clear
place in the Republic’s moral argument.”13 C. D. C. Reeve sarcastically
dismisses the first city as “the Kallipolis for money-lovers”, and says that
Plato did not mean for us to take it seriously because “it includes nothing
10 Julia Annas, An Introduction to Plato’s Republic and Ethics (Oxford: Oxford
Univeristy Press, 1981), p. 76.
11 Ibid, 78.
12 Ibid.
13 Ibid.
40 NICHOLAS D. SMITH

to counteract the destabilizing effects of unnecessary appetites and the

pleonexia to which they give rise.”14
Such assessments, and the arguments offered for them, however, seem
to me to be quite mistaken, and in any case clearly do not explain why
Plato would represent Socrates as having such a favorable view of this
city, which is never qualified or withdrawn even in retrospect in any of
the remaining books of the Republic. But let us look more closely at the
arguments offered for such negative appraisals by Plato’s commentators.
Is the first city built entirely upon self-interest, as Annas says? Annas cites
369c6–7 for this claim, where Socrates says that cities come into being
because people share by giving and taking, thinking that it is better for
them to do so. It is significant that Socrates makes this claim, however,
before beginning the process of “creating a city, from the beginning, in
words” (369c9–10). Accordingly, this characterization of how cities come
into being is not offered as a description of only how the first city comes
into being, or as a subtle signal of what is wrong with the city that comes
into being with this raison d’être. Rather, there is every reason to believe
that this same basic motivation lies behind the generation of any city what-
ever, including the Kallipolis itself. After all, it is an unmistakable feature
of Plato’s argument – and a central feature of his answer to the challenges
of Glaucon and Adeimantus at the beginning of Book II – that it is in one’s
interest to be just. If the first city is brought into being out of self-interest,
then, and can be supposed to satisfy that self-interest, this is an argument
in favor of the first city’s justice, and not against it.
Reeve, however, argues that the flaw is not the self-interest of the city’s
inhabitants, as such, but the fact that it is purely material self-interest,
which leaves the city wholly unprotected against “the destabilizing effects
of unnecessary appetites and the pleonexia to which they give rise.” But
Reeve imagines this without any help from Plato’s own explicit descrip-
tion of this state, in which the relevant appetites and their satisfactions are
described in ways that highlight their moderation and simplicity:

First, then, let’s see what sort of life our citizens will lead when they’ve been provided for in
the way we have been describing. They’ll produce bread, wine, clothes, and shoes, won’t
they? They’ll build houses, work naked and barefoot in the summer, and wear adequate
clothing and shoes in the winter. For food, they’ll knead and cook the flour and meal
they’ve made from wheat and barley. They’ll put their honest cakes and loaves on reeds
or clean leaves, and reclining on beds strewn with yew and myrtle, they’ll feast with their
children, drink their wine, and, crowned with wreaths, hymn the gods. They’ll enjoy sex

14 C. D. C. Reeve, Philosopher-Kings: The Argument of Plato’s Republic and Ethics

(Princeton: Princeton University Press, 1988), p. 171.
 PLATO’S ANALOGY OF SOUL AND STATE 41

with one another but bear no more children than their resources allow, lest they fall into
either poverty or war. (372a5–c1)15

The citizens of Plato’s first state will eschew even the luxurious excess
of wearing shoes or clothing in the summers, and are careful to do noth-
ing that would strain their very modest resources, and so will avoid both
poverty (which Plato wishes to avoid even in the Kallipolis, as morally
corrosive – see 421d13–422a3) and war. If Plato wished to portray this
city as in danger of “the destabilizing effects of unnecessary appetites and
the pleonexia to which they give rise,” he has done a remarkably poor job
of it!
Annas’s final argument – that Socrates’ urging us to see the origins of
both justice and injustice in this state is a subtle sign of his lack of regard
for this state – is equally flawed. First, let us recall that Socrates stated that
this – the identification of the sources of justice and injustice in a city –
was the goal he was seeking in bringing the city into the discussion in the
first place (at 369a5–7). Second, from the fact that Socrates and his inter-
locutors are in a position to see what injustice in a state would be, it does
not follow that the state they have envisioned is one that is filled, to any
significant degree, with injustice. After all, it is precisely because Socrates
and his interlocutors reach the position where they can see the origins of
both justice and injustice that they are able to conclude the discussion of
the Kallipolis at 434b9–c10: justice in the state is where each element in
the state does its own, and injustice is where the elements meddle and
interfere in each other’s proper functioning. Third, Annas is wrong to say
that, unlike the first state, the Kallipolis “displays only justice.” If this were
true, there would be no need for lawcourts in the Kallipolis, and no need
for the philosopher-rulers to serve as judges for disputes that may arise
among the citizens of the Kallipolis; but Plato tells us that there will be
such needs, at 433e3–8.
What is lost, in views such as the ones we find Annas and Reeve giving,
is the recognition that none of the states that Plato’s Socrates and his
interlocutors conceive of are Forms. Stated so bluntly, this may seem just
obvious, but Annas’s claim that the Kallipolis will display only justice
and no injustice is one that Plato’s metaphysics could never support. Only
Forms are what they are and never can or do seem to be the opposite at the
same time (see 478e7–479b8). Accordingly, we should expect that both

15 To highlight the discrepancy between Reeve’s assessment of the first city, and Plato’s
actual description of it, I have used Reeve’s own translation here [G. M. A. Grube and C.
D. C. Reeve Plato: Republic Indianapolis and Cambridge: MA: Hackett, 1992, p. 47].
42 NICHOLAS D. SMITH

the first city and the Kallipolis will “display” both justice and injustice, for
both are only participants in, and not identical to the Form of Justice.16
This recognition, however, shows us a further thing about how Plato’s
metaphysics must inform his conception of the analogy of soul and state.
In so far as Justice Itself – the Form – is one thing, then we should expect
every one of its images (both in states and in individual people) to make
reference to the same thing. The analogy of justice in the soul and justice
in the state is secured, then, in virtue of the fact that the justice in each
will resemble one another, and they will resemble one another precisely
because both sorts of justice are images of Justice Itself. But it follows from
this that any image of Justice will resemble any other image of Justice, in
the relevant way. But states come in an indefinite number of varieties –
not all are divided into three parts. Some states will be more just than
others; some will have little justice in them to observe at all. Any state
with detectable justice in it, however, should serve to the same degree as
its justice is detectable, for Socrates and his interlocutors – and for Plato’s
readers – to be able to “read the large letters” of justice in the state, in
preparation for an attempt to discern the same message of justice written
in smaller letters within an individual. Plato’s Republic, I am claiming,
actually provides two distinct images of justice in states: in the first and
most basic city, and also in the Kallipolis.
Of course, as Plato tells us later – when he describes how images can
be used intelligently (in Books VI and VII), as they are in arithmetic and
geometry – some images are more useful as propaedeutics and heuristics
than are others. In seeking to achieve a grasp of what justice is, before
turning to the actual challenge itself, which was to show that justice is
always preferable to injustice, Plato’s Socrates must choose the most useful
images of justice. When Glaucon and Adeimantus seem so scornfully
unimpressed by the first city, Socrates acknowledges that a larger and more
luxurious city might provide a better view of what they are looking for:
Alright, I said. I understand. It is not just how a city comes into being that we are inves-
tigating, but rather how a luxurious city comes into being. Perhaps that’s not a bad thing,
either. For by investigating that, we might discern how justice and injustice grow up in a
city. (372e2–6)

Even if the Kallipolis provides a better image of Justice for the purposes
of their investigation, however, as it plainly does, we should not necessarily
16 This also, presumably, accounts for what would otherwise be the very strange quali-
fication Socrates makes, at 444a6, that in identifying the just man and the just city as they
have, they have not told a “complete falsehood.” There would be no falsehood at all, in
what they have said, if they had identified the Form of Justice, rather than merely good
images of it.
 PLATO’S ANALOGY OF SOUL AND STATE 43

suppose that the first city was either unjust or in any absolute way a poor
or deeply flawed image of justice itself. An image may be poorly suited
for an inquiry into the nature of that which the image imitates, for reasons
other than that it is a poor image. Notice that in seeking to discern what
justice is, Plato’s Socrates insists on not looking first at individual human
beings, but at states, for fear that the justice they sought to grasp would
be too difficult to discern in individuals, without the help of the analogy
to states. It plainly does not follow from this, however, that no individ-
ual human beings could ever be good or reasonably accurate images of
justice.
At any rate, once justice is finally located and observed in the Kallipo-
lis, Plato has Socrates repeatedly remind us that it is nothing other than
the very same principle that was used in founding the first, most basic
city, as well as each subsequent development worked upon that city, in
transforming it into the Kallipolis: that each person in the city should do
that and only that task in the city, for which his or her nature best suits
them (see 370a7–b2, 374a4–e8, 432d7–433a6, 443b7–c7). Accordingly, it
turns out that there was justice in the first and most basic city, which now
presumably even Glaucon and Adeimantus are in a position to see: in that
city, no less than in the Kallipolis, each person did that for which they
were best suited by nature, and did not meddle in what others were better
suited to do. Despite proving clearly (in retrospect) to be a good image
of justice after all, however, the first city does not do so in virtue of any
obvious division into three parts. None the less, even if we granted Plato’s
arguments for the partitioning of the soul, we can see now what the analogy
of soul and state would consist in, if we compared Plato’s tripartite soul to
his first and most basic state: justice in each would be in each element or
part “doing its own” and not meddling in the proper business of any other
element or part. We do not have to presuppose the 3-3 specification of the
soul-state analogy in order to preserve the identity of justice, and thus, the
essential basis of the analogy.

III. J USTICE IN THE S OUL

If the argument of the last section is correct, Plato actually offers more
than one model of the city which satisfies the analogy of justice in states
and souls. Plato does not, however, provide any clear alternatives to the
conception of the soul that he offers in Book IV, and so it might seem that
he is as deeply wedded to his conception of the soul as tripartite as he could
be to any theory he offers. But in several places in the Republic – including
immediately before and after giving us his partitioning arguments – Plato
44 NICHOLAS D. SMITH

reveals a certain lack of complete commitment to the consequences of the

arguments he gives for the tripartite soul.
Socrates begins his argument for the partitioning of the soul with a word
of caution:
Well, is it only a small question we have come to, O amazing one, I said, whether the soul
has these three forms in it, or not?
[Glaucon responds] I do not believe it is at all a small question. For perhaps, O Socrates,
the saying is true that “fine things are difficult.”
Apparently, I said, but know well, O Glaucon, that in my opinion, we’ll never comprehend
the matter precisely using the methods we are using in our present arguments, though there
is a longer and more complete way which would lead to this. But perhaps we can proceed
in a way that is comparable in value to our former claims and inquiries.

By having Socrates remind Glaucon of the shortcomings of “the meth-

ods we are using in our present arguments,” and noting that the same
shortcomings can be expected in the arguments that follow, Plato cautions
his readers that his arguments must be taken with a grain of salt. 17 More-
over, immediately after he gives his partitioning arguments, when he is
summarizing his account of what justice is in the soul, in Book IV, Plato
explicitly opens up the possibility that the soul may actually be composed
of more than three parts:
The truth, however, is something more of this sort, it seems, justice is not doing one’s
own with regard to what is external, but with regard to what is within him, what is truly
himself and his own, not allowing any part of his soul to do another’s in himself, or to
meddle with one another, but to arrange well what is really his own and to rule himself. He
puts himself in order and befriends himself, and harmonizes the three things in him like
the three harmonic limits, the low, the high, and the middle, and any others there may be
between them. (443c9–d7)

Later, in Book IX, just when we might expect Plato’s Socrates to remind
us of the tripartite nature of the soul, we find him arguing, instead, that
we must not suppose “that in its truest nature, the soul is this way – as
embroidered in many colors and dissimilarities, and at odds with itself”
(611b1–3). Socrates goes on to say that the condition in which he and
Glaucon have been examining the soul is a condition much like that of the
sea-god Glaucus, damaged and encrusted, as it were, in its association with
the body (611b9–d8), whereas only if we could observe it in its true nature
could we really discern “whether it has many parts or just one, and how and
in what way it truly is” (612a3–5). Only a few pages before comparing the
17 I owe this observation to Mitchell Miller, who makes a similar point in Mitchell
Miller, “Platonic Provocations: Reflections on the Soul and the Good in the Republic,”
in D.J. O’Meara (ed), Platonic Investigations (Washington, DC: The Catholic University
Press of America, 1985), p. 169.
 PLATO’S ANALOGY OF SOUL AND STATE 45

embodied soul to Glaucus (at 603d6–7), Socrates breezily acknowledges

that the soul suffers from “countless oppositions” of the sort he took (in
Book IV) to divide the soul. One wonders if all of these oppositions could
be accounted for (applying Oe) by an appeal to only the three parts of the
soul he identified in Book IV. In regard to his earlier arguments, Socrates
allows only that “we have given a fitting (ε’π ιεικ ω̂ς ) account” of the soul,
as it is in a human life (612a5–6).
Here, too, then, we find Plato doing the opposite of what we would
expect of him if he were committed to the 3-3 specification of the soul-
state analogy, for now he appears prepared to allow that the soul may not
– in its ideal condition, at any rate – be composed of three parts, after all.
Moreover, Socrates allows that if only they were able to observe the soul
in its truest nature, they might then be able to “see justice and injustice and
the other things we have been discussing more clearly” (611c4–5).
W. K. C. Guthrie has argued that Plato believed that in its truest form,
the soul was not a composite after all, but, instead, a simple.18 If so,
because Plato claims that in its truest form, the soul would be a “much
finer thing . . . than we now suppose” (611c4–5), we can infer that it would
be an even better approximation or image of justice as a simple than it
is in its fragmented aspect, “maimed” in its association with the body
(611b10–c1).
In carefully distancing himself in this way from his earlier partitioning
arguments, Plato’s Socrates shows that we should not understand those
earlier arguments as being in any way decisive as to how we are to conceive
of “Plato’s theory of the soul,” if he has such a theory. Instead, we should
remind ourselves what the purpose of the earlier arguments was: to try
to discern the aspect of justice in a soul. In dividing the soul into three
parts, Socrates and Glaucon were better able to see what it was that made
a soul just, by analogy to what it was that made a state just. As such, the
just tripartite soul served the discussion well, as an image of justice. But
again, it does not follow from this that the image we have been given is
the best possible image of justice, either in the soul or in the state. All that
follows is that they are good images of justice – good enough to serve the
propaedeutic and heuristic purposes for which they have been selected.
Given Plato’s characterization of justice in the soul, however – each
part doing its own and not meddling in the functioning of the other parts
– we might well wonder how Plato could account for justice in a non-
partitioned soul, since plainly in such a case there could be no parts to “do
their own.” If we return to the formula that Plato has relied upon all along,
18 W. K. C. Guthrie, “Plato’s View of the Nature and the Soul,” in Gregory Vlastos (ed),
Plato II: Ethics, Politics, Philosophy of Art and Religion (New York: Anchor Books, 1971).
46 NICHOLAS D. SMITH

however, we will see that this worry is no real threat to his conception.
Justice in a composite thing – either soul or state – is all along conceived
as nothing other than the overall harmony that is achieved in the composite
in which each element of the composite “does its own.” States are just,
Plato tells us, when and in so far as each element within the state “does
its own.” In the first and most basic state, and also in the Kallipolis, this
ultimately amounts to the situation in which each individual does what he
or she is naturally best suited to do, and does not meddle in others’ affairs.
Any single individual in the state, therefore, will be civically just in so far
as he or she “does his or her own.” After all, it is not as if the state can
somehow be just in which the citizens themselves act unjustly. Similarly,
just as we might talk about certain impulses and motives being unjust and
others being just, Plato can say that each part of a soul is just when and
only when that part is doing its own, and he can even say that individual
desires and motives are just when and only when they are symptomatic of
the soul or its relevant part “doing its own.” Since the justice of a composite
consists in the harmony resulting from the justice of each of its component
(including, presumably, its indivisible) parts “doing its own,” it follows
that indivisible simples are just when and to the degree that they “do their
own.”
But the “doing one’s own” formula ordinarily applies to activities that
are, as the Greeks would say, πρóς τ ι, or “in relation to something.” For
a carpenter in Plato’s Kallipolis to “do his own” is not just for him to
do carpentry, but also for him not to meddle in others’ work. For one’s
“spirited part” to be just, it must not only provide the right motivations, it
must also not interfere in one’s reasoning or appetites. A single individual
stranded on a desert island, accordingly, would not have the opportunity to
engage in the civic form of justice (even if she might still be psychically
just), since she would have no “other” relative to whom she could “do her
own.” Similarly, it might be supposed, if we conceive the soul as a simple,
we lose any opportunity to characterize it as just, because there are no
longer any others (in this case, parts of the soul) for the soul to be just in
relation to.
But Socrates gives an argument in Book I that neutralizes this concern,
proclaiming that the soul has a function, and that the virtue of a soul would
consist in the soul performing its function well. The function of the soul
is “management, rule, deliberation, and all other such things” (353d4–6),
and to this list Socrates immediately adds “living” (at 353d9). Justice is the
virtue of the soul, and so the just soul will be one that performs these func-
tions well (353d11–e11). The “doing one’s own” formula, accordingly,
may refer to the ways in which one functions in regard to others, but it
 PLATO’S ANALOGY OF SOUL AND STATE 47

can also, we must suppose, apply to the proper functioning of the soul in
regard to the soul’s most basic functions, which include most importantly
the management of one’s own life.

IV. C ONCLUDING R EMARKS

In the first part of this paper, I claimed that Plato’s arguments for the parti-
tioning of the soul are unreconstructable failures. If the analogy of justice
in the state and in the soul required what I have called the 3-3 specification
of that analogy, this failure would infect and ruin much of the argument of
the Republic. In fact, however, I have argued that it does not do so, and I
gave reasons, in the second and third parts of the paper, respectively, why
even Plato himself was not committed in any deep way (in the Republic, at
least) either to the necessity of a tripartite state, for the state to be just, or
to the theory of the tripartite soul, or for this conception of the soul to be a
requirement for the soul to be just. I have not claimed that Plato remained
non-committal about his tripartite analysis of the soul, which shows up
prominently again in the Phaedrus and the Timaeus, though even in these
dialogues, I believe his commitment to the analysis may be doubted. An
argument for this claim, however, must wait for another day. Even if Plato
did come to accept his partitioning arguments without reservation, it would
not count as evidence against my argument in this paper, for I have claimed
only that we should not accept them, that Plato does not actually require
such arguments to succeed in order to defend his soul-state analogy, and
that various things he says show that Plato himself was aware that the parti-
tioning of the state and soul were not required by the soul-state analogy in
the Republic.
At base, my argument is founded upon my claim that Plato’s meta-
physics required that there be a single, unitary, and unequivocal conception
of justice, which Plato would tie to his idea that there was a single pure
Form of Justice. Even this part of my argument is not uncontroversial,
however, for scholars have often talked of there being two conceptions of
justice in Plato’s Republic. The most troubling case of this attribution was
made in the famous paper by David Sachs,19 in which Sachs argued that
Plato’s defense of justice rests on a fallacious equivocation between the
two senses of justice, which Sachs called the “vulgar” and the “Platonic”
conceptions. But even those who have sought to defend Plato against
Sachs20 have allowed that there are two distinct conceptions of justice
19 Sachs, “A Fallacy in Plato’s Republic.”
20 See, for examples, R. Demos, “A Fallacy in Plato’s Republic?,” The Philosophical
Review 73 (1964), pp. 345-398; L. Galis, “The State-Soul Analogy in Plato’s Argument
48 NICHOLAS D. SMITH

– usually identified as the social or civic conception, in which justice is

conceived in terms of the qualities of the performance or abstention from
certain sorts of actions, and the psychic conception, in which justice is
conceived in terms of balance or harmony in the soul. Plato’s defend-
ers have generally and I think rightly argued that Plato could show that
those who were psychically just would also be socially or civically just.
Although I do not deny that individuals whose souls are just will act justly
– after all, Plato explicitly says this at 443c9–444a2 – I do think it is a
mistake to say that there is more than one conception of justice ever at
work in Plato’s Republic.
Justice, for Plato, is ultimately nothing other than the Form of Justice.
Any instantiations of this Form, and images of it, will be just if, and only
to the degree that they approximate or participate in the character of this
Form. But the character of this Form does not change in relation to its
different sorts of participants: whatever their other differences may be, in
so far as these participants are just, they will be just in the same respect,
as it were. Anything that approximates or images the Form of Justice will
resemble anything else that does, and so may be found to be analogous to
anything else that approximates or images the Form of Justice, but the
resemblance will always be on the basis of the same Form and there-
fore in virtue of whatever it is in the two approximations or images that
displays the relevant character associated with this single Form. Accord-
ingly, my argument has been that Plato does not need to be committed to
any specific theory of the soul or of the state, to make an analogy between
the instantiations of justice that may be found within each.
In each and every instance of justice represented to us, in the Republic,
we are told that the justice we are supposed to observe derives from and
is present in virtue of that instance’s proper functioning. It is the proper
functioning of the soul that we are told is justice in the soul, in Book I.
It is the proper functioning of each of the citizens in the first and most
that Justice Pays,” Journal of History of Philosophy 12 (1974), pp. 285–293; R. Hall,
“Plato’s Political Analogy: Fallacy or Analogy?,” Journal of the History of Philosophy 12
(1974), pp. 419–435; R. Sartorius, “Fallacy and Political Radicalism in Plato’s Republic,”
Canadian Journal of Philosophy 4 (1974), pp. 349–363; J. Schiller, “Just Man and Just
Acts in Plato’s Republic,” Journal of the History of Philosophy 6 (1968), pp. 1–14; Vlastos,
“Justice and Happiness in The Republic”; S. Waterlow, “The Good of Other in Plato’s
Republic,” Proceedings of the Aristotelian Society 47 (1973), pp. 19–36; and R. Wein-
gertner, “Vulgar Justice and Just Acts in Plato’s Republic,” Philosophy and Phenomeno-
logical Research 25 (1964/65), pp. 248–252. A few scholars (in my view rightly) do not
make such a concession – see, for example, Julia Annas, “Plato and Common Morality,”
The Classical Quarterly 28 (1978), pp. 437–451; Scot Yader, “Justice in Plato’s Repub-
lic: Constructive Moral Philosophy vs. Moral Apologetics” Binghamton, NY, Society for
Ancient Greek Philosophy, 1995.
 PLATO’S ANALOGY OF SOUL AND STATE 49

basic state that makes these citizens and their state just, in Book II. It is the
proper functioning of each of the citizens in the much expanded state, the
Kallipolis, and each of the classes of this state, that we are told make the
citizens and classes of this state – and the whole state itself – just, in Book
IV. Also in Book IV, we are told that it is the proper functioning of the parts
of the soul that make these parts and the activities they motivate just, and
which makes the whole soul and the person just. In every case, justice as
proper functioning can be understood under the formula “doing well what
the thing is naturally suited to do, and not attempting to do what only some
other thing is better suited to do.” But in all of its instantiations, justice is
proper functioning, for Plato. There may appear to be many distinct charac-
terizations of this, just as Plato recognizes many images of the single Form.
There is a unity that runs through all of the characterizations, however, and
the metaphysical basis for this unity is none other than the Form.21

Department of Philosophy
Michigan State University
East Lansing MI 48824
USA

21 I presented various earlier versions of this paper as one of the Patricia Crawford
lectures at San Diego State University (September, 1998), at the Third Arizona Colloquium
in Ancient Philosophy (February, 1998), and as colloquia presentations at the University
of Toronto (December, 1997) and at Michigan State University (September, 1998). I am
indebted to my hosts in San Diego (especially Angelo Corlett and Mark Wheeler) and
to those in Toronto (especially Brad Inwood and Lloyd Gerson) for their searching ques-
tions and criticisms. I gratefully acknowledge, as well, the assistance I have received from
Mitchell Miller, my commentator at the Third Arizona Colloquium in Ancient Philosophy,
who provided in private correspondence several detailed suggestions, in addition to those
he read at this conference, several of which I have incorporated into the current version
of the paper. I am also grateful to my colleagues at Michigan State University, and to
the members of my Fall 1997 seminar there – Timothy O’Neill in particular – for their
many comments and suggestions. I have also benefited from long discussions of this topic
with Fred D. Miller, Jr. and with Michael Fiebig, both of whom were especially helpful in
leading me to sharpen my criticisms of Plato’s partitioning arguments. Finally, I acknowl-
edge the help I have received from Rachana Kamtekar and Ellen Wagner, who read and
commented extensively on the version of the paper I gave in Arizona. None of the above
can be assumed to agree with my arguments, and all remaining errors in the paper are mine
alone.