Degrees of Separation in the Phaedo

MICHAEL PAKALUK

ABSTRACT
It can be shown that, if we assume Ôsubstance dualismÕ, or the real distinctness
of the soul from the body, then the standard objections to the Cyclical Argument
in the Phaedo fail. So charity would presumably require that we take substance
dualism to be presupposed by that argument.
To do so would not beg any question, since substance dualism is a signiŽ cantly
weaker thesis than the immortality of the soul. Moreover, there is good textual
evidence in favor of this presumption. A closer look at the immediately preced-
ing passage, viz. ÒSocratesÕ DefenseÓ, reveals an argument for a real distinction
between soul and body, not unlike DescartesÕ famous argument, based on the
identiŽ cation of an activity in which the soul can in principle engage on its own,
without assistance from the body.
The argumentative project of the Phaedo, on this reading, becomes: given that
the soul is really distinct, show that it is immortal. And Plato aims to do this in
two stages. The three initial arguments are meant to establish merely the mini-
mal claim that the continued existence of the soul across cycles of reincarnation
is the most plausible view to take, given substance dualism; and it is left to the
Final Argument to argue for something that we might regard as immortality, that
is, the imperishability of the soul, come what may.

The three initial arguments for immortality in the Phaedo – the Cyclical
Argument (CA), the Recollection Argument (RA), and the AfŽ nity
Argument (AA) – are preceded by a passage in which Socrates, as if on
trial, defends himself against the charge that he is being reckless in not
resisting death (63a3-69e5). What is the function of this passage, typically
called ÔSocratesÕ DefenseÕ (SD)? A common view is that the passage is
simply SocratesÕ profession of ÔfaithÕ in life after death; 1 and, in the
remainder of the dialogue, Socrates tries to give plausible reasons for this
ÔfaithÕ – just perhaps succeeding in his Ž nal attempt. Another common
view is that SD is of merely rhetorical or literary signiŽ cance, and that

Accepted November 2002

1
For instance: ÒTo Cebes and Simmias it seems strange that a philosopher, or
indeed any man of intelligence, should be well content to die . . . Socrates replies with
a declaration of faith that after death life will still be under divine protectionÓ,
R. Hackforth, PlatoÕs Phaedo (Cambridge: Cambridge University Press, 1955), 39.

© Koninklijke Brill NV, Leiden, 2003 Phronesis XLVIII/2

Also available online – www.brill.nl
90 MICHAEL PAKALUK

the argumentation of the dialogue begins properly with CA.2 I shall take
a different approach. On my reading, the philosophical substance of the
Phaedo begins properly with SD, which should be understood as offering
an argument for a real distinction between soul and body, roughly analo-
gous to the famous argument of Descartes. The arguments that follow
upon SD, then, naturally enough take this distinction for granted. Yet
when we read the dialogue in this way, we see that its successive argu-
ments are cumulative and constructive: in PlatoÕs intention, the Phaedo is
not a dialogue of Ôfaith seeking reasonsÕ so much as of reason strength-
ening and conŽ rming itself. And SocratesÕ arguments in the end are prob-
lematic, not because they are in some crude way fallacious, but rather
because they articulate and defend, with some power and success, a philo-
sophical dualism which is deeply problematic.
I develop my interpretation by defending four claims: (I) on the
assumption of substance dualism, CA is not evidently unsound; (II) sub-
stance dualism is asserted in SD; (III) SD argues for substance dualism;
(IV) SD was intended by Plato to provide the context of the three initial
arguments and therefore counts as the Ôprimary argumentÕ of the Phaedo.

I. On the assumption of substance dualism, CA is not evidently unsound.

CA is not unfairly presented as follows:

1. Anything that comes to take on an attribute which has an opposite,
previously had that opposite attribute.
2. Being dead and being alive are opposite attributes.
3. When something comes to be alive, it comes to take on the attribute
of being alive.
4. Therefore, anything that comes to be alive previously had the attribute
of being dead.
5. But everything that is dead was previously alive.
6. Therefore, anything that comes to be alive was previously alive.
7. Therefore, living things come from previously living things.
8. Therefore, living things will once again become living things.
9. Nothing comes to take on again, at a later time, an attribute that it
now has, if it perishes in the process.

2
David Gallop says that CA is the Ž rst in Ò. . . the series of arguments that form
the core of the dialogueÓ (Plato: Phaedo, Oxford: Clarendon Press, 1975), 103. C.J.F.
Williams refers to CA as Ò[t]he Ž rst solid bit of argumentation you get in PlatoÕs
PhaedoÓ, in ÒOn DyingÓ, Philosophy 1969 (44) 217-30.
 DEGREES OF SEPARATION IN THE PHAEDO 91

10. Thus, living things do not perish when they come to be dead, and in
this sense they are immortal.
PlatoÕs strategy is to connect this present life of a living thing with a pre-
vious life; that done, he draws the general conclusion that living things
were previously alive; yet, he reasons, they could not have come alive
again, if they did not endure in the interval between their previous life
and their current one; and thus, as regards any living thing, we can have
some conŽ dence that it will continue to endure, in the interval after this
current life and before its next life.
Thus stated, the argument is clearly unsound, because the Ž rst premise
is in need of two familiar qualiŽ cations. That Ôopposites come from oppo-
sitesÕ is true only if: (i) we presume that we are not dealing with a case
of simple generation, where something comes to be F only in coming to
be simpliciter; and (ii) the opposites are ÔcontradictoryÕ, rather than mere
ÔcontraryÕ opposites. But if premise 1. is qualiŽ ed accordingly, premise 2.
needs to be revised: being dead and being alive are not contradictory
opposites, since there are things that are neither dead nor alive. Yet if we
rewrite premise 2., so that it involves opposites that are properly contra-
dictory, e.g. ÔBeing not alive and being alive are opposite attributesÕ, then
premise 5. needs to be changed accordingly, becoming: ÔEverything that
is not alive was previously aliveÕ – which is evidently false.
It would be good to have a diagnosis of why the argument goes wrong,
and for this purpose GallopÕs commentary is particularly useful. Gallop
correctly notes, for instance, that Plato in CA tends to speak as though it
is the soul which comes to be alive, rather than the animal, but – Gallop
objects – this Òinsinuates a view of ÔbirthÕ in which the soulÕs discarnate
existence is already covertly assumed. And since that is precisely what
the argument purports to prove, the very conception of incarnation can be
seen to beg the essential questionÓ (105). Again, Gallop wonders why we
shouldnÕt understand Ôbeing deadÕ (in our premise 2. above) to mean, sim-
ply, Ôceasing to existÕ, in which case, clearly, Ôbeing aliveÕ and Ôbeing
deadÕ could not be treated as opposing predicates, as CA requires, since
one would, in that case, be treating existence as though it were a predi-
cate. Speculating on why Plato might have resisted this identiŽ cation, Gallop
remarks that Òa wedge might be driven between Ôbeing deadÕ and Ôceas-
ing to existÕ by treating SocratesÕ soul as a separate subject, distinct from
Socrates himself, and alternating between incarnate and discarnate states.
But this would be, once again, to assume what has to be provedÓ (106).
Again, objecting to (our) premise 3. above, Gallop remarks that ÒThe
sense of Ôcoming to be aliveÕ required for the argument is not that in
92 MICHAEL PAKALUK

which a living thing comes into being, but that in which a soul Ôbecomes
incarnateÕ in a living body. Yet it cannot do this unless it already exists
before birth or conception. And whether it does so or not is just what is
at issueÓ (110).3
So Gallop maintains that CA goes awry because Plato begs the ques-
tion, surreptitiously supposing that the soul is a distinct subject, indepen-
dent of the body. (Call this view Ôsubstance dualismÕ, for short.4)
Yet because one manÕs begging the question is simply another manÕs
tacit assumption, we might wonder how CA would fare, if we were to add
new premises that made this assumption explicit.
Let us assume:
I. Every living thing has a soul, which is a distinct substance from the
body.
II. Death is the complete separation of a soul from any body. 5 That is,
death and life are states of a soul, which we deŽ ne as follow:
1. To be dead = to be completely separated from any body.
2. To be alive = to be joined to some body.
By these deŽ nitions, being alive and being dead are indeed contradictory
opposites. Furthermore, the generation of a living thing becomes simply
the combination, or coming together, of two elements, body and soul; and
the destruction of a living thing would be their separation.
We may then restate the argument. First we introduce the necessary
qualiŽ cation in the Ž rst premise, making it explicit that the principle involves
contradictory opposites:
1. Anything that comes to take on an attribute which has a contradictory
opposite, previously had that opposite attribute.
We also rewrite the next two premises in accordance with our new
assumptions:

3
He adds: ÒA thing cannot be said to Ôcome to life againÕ in the sense required by
the argument, unless the persistence of an independent subject, Ôthe soulÕ, is already
assumed. Yet this is just what has to be provedÓ (110).
4
We should not take this phrase as implying that, according to Plato, the body of
a living thing, and perhaps even a corpse, is a substance as well as the soul (this Plato
would deny, I believe); or that Plato thought that all existing things could be sorted
into two exclusive and exhaustive kinds, souls and bodies (this too Plato would deny).
5
Compare 64c4-8.
 DEGREES OF SEPARATION IN THE PHAEDO 93

2. To be dead, i.e., a soulÕs being completely separated from any body,

and to be alive, i.e., its being joined to some body, are contradictory
opposite attributes.
3. When a soul comes to be alive, it comes to be joined to a body.
Note that premise 3. now stands. It will do no good to object that Òthe
premise fails to consider the possibility that the soulÕs coming to be alive
is its simply coming into existence, as aliveÓ, since we are construing the
coming into existence of a living thing as a joining, or combination, and
it is appropriate to understand a joining, as such, as the combination of
previously existing parts.6 Note also that our new assumptions force us to
make explicit that CA is about souls and the changes they undergo, not
about living things or animals. GallopÕs intuition was that PlatoÕs talk of
souls Òinsinuates a view of ÔbirthÕ in which the soulÕs discarnate existence
is already covertly assumedÓ. We might say, rather, that the understand-
ing of a living thing as a composite of soul and body implies a view of
ÔbirthÕ as the soulÕs coming to have a relation to a body.
It therefore follows trivially that:
4. Any soul that comes to be alive, i.e. joined to some body, previously
had the attribute of being dead, i.e. completely separated from any
body.
If we simply rewrite the next premise (premise 5.) in accordance with our
deŽ nitions we get:
Every soul that is dead, i.e., completely separated from any body, was previously
alive, i.e., joined to some body.

This is unsupported as it stands, but it is not difŽ cult to provide some sup-
port. Suppose we assume something like ÒThe soul is continually subject
to changeÓ or even ÒNo soul remains unchanged foreverÓ – perhaps
because we have a notion of nature as necessarily involving change, and
we regard the soul as within nature. Then it would follow that any soul
that, as it happens, is completely separated from a body, in fact came to
be so: it was not always so. But any soul that came to be separated from

6
We might imagine, say, a single thing coming into existence out of nothing, and
its becoming steadily differentiated into two components as it springs into existence,
as an instance of two things coming to be joined while coming into existence. But this
speculation, besides involving creation ex nihilo, the coherence of which Plato would
deny, seems rather to be a case of two things coming to be separated, rather than
joined.
94 MICHAEL PAKALUK

any body, by premise 1., was previously joined to some body. So we need
to supplement the argument in some such way as follows:
5.1. Nothing that changes in some respect is ever eternally unchanged in
that respect.
5.2. No soul that comes to be joined to a body is ever eternally in a state
of separation from a body.
5.3. Therefore, every soul that is completely separated from a body but
will be joined to a body was previously joined to a body.
Of course it remains possible that souls do not exist eternally through the
past but come into existence themselves, from being compounded out of
simpler parts. Plato aims to rule out this alternative, of course, initially in
the AfŽ nity Argument and then more deŽ nitively in the Final Argument
of the Phaedo. Yet in the context of the Cyclical Argument he might rea-
sonably maintain, simply, that a conception of nature according to which
souls are successively compounded and decomposed out of simpler parts
introduces unnecessary complications – what agent compounds them, once
their parts are scattered? – and therefore should not be entertained with-
out evidence or supporting reasons.
The rest of the argument proceeds as before, now without any hitches:
6. Therefore, any soul that comes to be alive, i.e. joined to some body,
was previously alive.
7. Therefore, living souls come from previously living souls.
8. Therefore, living souls will once again become living souls.
9. Nothing comes to take on again, at a later time, an attribute that it
now has, if it perishes in the process.
10. Thus, souls that are now alive, i.e. joined to some body, do not per-
ish after they come to be dead, i.e. after they come to be completely
separated from any body.
The argument is not meant of course to be a deduction based on neces-
sary principles and on the meanings of words: obviously, premise 8. involves
something like an inductive inference: from the conclusion that the soul
of a living thing has in the past undergone cycles of reincarnation, we can
be conŽ dent that it will continue to do so in the future. Rather, the argu-
ment aims to sketch the most reasonable picture of the world (diamuy-
ologÇmen 70b6, cp. muyologeÝn, 61e2), given the assumption that the soul
is a distinct substance from the body. 7 Given that assumption – and if we

7
We might expect it to rely on principles that are true typically and Ôfor the most
 DEGREES OF SEPARATION IN THE PHAEDO 95

rule out generation out of nothingness and destruction into nothingness as

incoherent – then that souls undergo cycles of reincarnation is the most
reasonable position to adopt: and if generally such is the place of souls
in nature, then, in the absence of arguments to the contrary, it would be
misguided to fear that oneÕs own soul will perish after its separation from
the body.
So CA is not evidently unsound if we assume substance dualism. However,
someone might object that substance dualism is equivalent to the immor-
tality of the soul, and that, therefore, Plato begs the question after all. Yet
this is not so: substance dualism is clearly a weaker view than immortal-
ism. Consider the following series of increasingly stronger claims:
(i) The soul is capable of existing upon the destruction of the body.
(ii) The soul in fact continues to exist upon the destruction of the body.
(iii) Necessarily, the soul continues to exist upon the destruction of the
body.
Presumably substance dualism is roughly equivalent to (i), and immortalism
involves at least (ii) and maybe (iii), but these are stronger claims than (i).8
Not that it isnÕt easy to con ate substance dualism and immortalism:
even Descartes in his eagerness did just that. Recall that the original, 1641
edition of the Meditations bore the subtitle, Òin which are demonstrated
the existence of God and the immortality of the soulÓ (in qua Dei exis-
tentia et animæ immortalitas demonstrantur). But Descartes quite appro-
priately corrected this in the 1642, second edition, so that it read, simply:
Òin which are demonstrated the existence of God and the distinction
between the human soul and the bodyÓ (in quibus Dei existentia, et animæ
humanæ ˆ corpore distinctio, demonstrantur). Descartes made the change
in reaction to MersenneÕs early criticism, expressed to him in a letter, that
in fact Descartes says nothing about the immortality of the soul in the
Meditations. ÒYou should not be surprised,Ó Descartes writes back in
response, ÒI could not prove that God could not annihilate the soul, but
only that it is by nature entirely distinct from the body, and consequently

partÕ, which should be presumed to have force unless we are given reason to think
otherwise.
8
Or consider a series of temporal, rather than modal, assertions about the soul:
(i)Õ The soul continues to exist only momentarily upon the destruction of the body.
(ii)Õ The soul exists for a signiŽ cant interval upon the destruction of the body.
(iii)Õ The soul exists for an indeŽ nite time upon the destruction of the body.
(i)Õ is perhaps roughly the same as substance dualism, under some understandings, but
(ii)Õ and (iii)Õ are stronger, and immortalism would require at least these.
96 MICHAEL PAKALUK

it is not bound by nature to die with itÓ.9 The difŽ culty also gets stressed
near the end of the second set of Objections, usually attributed to Mer-
senne: Ò. . . you say not one word about the immortality of the human
mind. . . . We now make the additional point that it does not seem to fol-
low from the fact that the mind is distinct from the body that it is incor-
ruptible or immortal. What if its nature were limited by the duration of
the life of the body, and God had endowed it with just so much strength
and existence as to ensure that it came to an end with the death of the
body?Ó 10 In reply, Descartes agrees and refers to his Synopsis, where he
acknowledges that, in fact, a proper demonstration of the immortality of
the soul must go far beyond the mere argument for a real distinction and
depend up on Òan account of the whole of physicsÓ.11 We might appro-
priately understand CA, then, not to be begging any questions, but rather
to be Ž lling in some of the gap Mersenne had noticed.
Note, furthermore, that the precise worry at which CA is directed takes
it for granted that the soul is one thing and the body another, but that the
soul is relatively frail and needs, as it were, to be shielded or protected
by the body. As Cebes says at 70a2-6: Òmen fear that when itÕs been sep-
arated (¤peidŒn Žpallag») from the body, it may no longer exist any-
where, but that on the very day a man dies, it may be destroyed and per-
ish, as soon as itÕs separated (eéyçw Žpallattom¡nh) from the body; and
that as it goes out (¤kbaÛnousa), it may be dispersed like breath or smoke,
go  ying off, and exist no longer (oéd¢n ¦ti) anywhere at all.Ó WhatÕs
imagined here, in fanciful, popular imagery, is momentary separation, then
dispersion. Similar language is used when Socrates paraphrases the con-

9
Letter to Mersenne of 24 December 1640, in J. Cottingham, Robert Stoothoff,
Dugald Murdoch, and Anthony Kenny, eds. The Philosophical Writings of Descartes,
vol. III: Correspondence, (ÔCMSKÕ) (Cambridge: Cambridge University Press, 1991),
163. This is a good statement of the limited character of substance dualism. Descartes
here mentions needing to prove that ÒGod could not annihilate the soulÓ because this
would be required in order to establish the absolute necessity of the soulÕs continued
existence. Of course he also means to insinuate that, since GodÕs interfering to anni-
hilate the soul seems unlikely, then to show a real distinction is to go a long way
towards establishing immortalism.
10
John Cottingham, Robert Stoothoff, Dugald Murdoch, eds., The Philosophical
Writings of Descartes: Vol. II, (ÔCSMÕ) (Cambridge: Cambridge University Press,
1984), 91. Note the similarity between MersenneÕs supposition and CebesÕ objection
that, for all we know, the soul may wear out and die, even if not on account of the
bodyÕs dissolution.
11
CSM 10, 108-109.
 DEGREES OF SEPARATION IN THE PHAEDO 97

cern later, at 80d9-11: Òcan it be that this, which weÕve found to be a thing
of such a kind and nature, should on separation (Žpallattom¡nh) from
the body at once (eéyæw) be blown apart and perish, as most men say?Ó
Someone might object that we should not place too much emphasis on
the precise way in which the worry that motivates CA is presented. That
worry is, after all, said to be one that concerns Ôcommon peopleÕ or Ômost
menÕ: it is feared Ôby people generallyÕ (toÝw ŽnyrÅpoiw, 70a1-2) that the
soul will be dispersed when it leaves the body; it is Ôthe popular fearÕ (tò
tÇn pollÇn, 77b3-4); this is Ôas most people sayÕ (Ëw fasin oß polloÜ
nyrvpoi, 80d10-e1). And it was of course an ancient and common belief
in popular Greek religion that the soul was some shadowy thing distinct
from the body. IsnÕt Plato simply motivating an argument for immortal-
ity by bringing in this popular belief, which he and his interlocutors obvi-
ously do not endorse?
But Greek popular views on the survival of the soul were indistinct and
mixed; Plato might just as easily in the context have made reference to
the Ôbelief of most menÕ that the soul does survive death, in some atten-
uated condition, in Hades. Clearly, what Plato does is to select some
aspect or strand of popular opinion to suit his purposes. But what are those
purposes? We see them at the beginning of SD. After it is mentioned that
Ômost peopleÕ regard philosophers as walking dead men, living impover-
ished lives, Socrates tells his interlocutors to dismiss what most people
think, so that they can speak among themselves (eàpvmen g‹r . . . pròw
²mw aétoæw, xaÛrein eÞpñntew ¤keÛnoiw, 64c1-2). Socrates then introduces
the deŽ nition of death as the separation of the soul from the body (64c4-
8) and gives his Defense. The whole argument of SD is meant to be devel-
oped among friends who share philosophical views that set them apart
from Ômost peopleÕ. Thus, when the views of Ômost peopleÕ are brought
in again, just before CA, this is to introduce a doubt not shared by Socrates
and his circle. (To be precise: it is not shared by them insofar as they rely
on their philosophical conception of the soul. Of course, it is shared by
them insofar as they are open to in uence from common views, on
account of Ôthe child inside us, who has fears of that sortÕ, 77e5.) But
since the purpose of the appeal to what Ômost peopleÕ think at this point
is precisely to raise a doubt Ôfrom the outsideÕ, as it were, then what Ômost
peopleÕ are said to concede – that the soul is separable from the body –
may reasonably be taken to represent the state of the argument at that
point, among Socrates and his friends.12

12
That the Greek mind at the time of Plato might admit some kind of dualism while
98 MICHAEL PAKALUK

So then, the CA is not evidently unsound, if we supply substance dual-

ism. The usual canons of interpretation, then, would suggest that we
should supply substance dualism, if doing so would be at all consistent
with the text. And in fact the text does not hinder us from doing so; rather
it gives us strong considerations in favor.13

II. Substance dualism is asserted in SD.

Yet we can go beyond this, since SD, which immediately precedes CA,
should evidently be construed as afŽ rming substance dualism. Therefore,
without reasons to the contrary, we should regard this afŽ rmation as
remaining in place, throughout the three initial arguments.
That SocratesÕ Defense afŽ rms substance dualism is clear if we con-
sider the nature of its argument. Unfortunately, the usual summaries are
 awed. It is not enough to summarize SocratesÕ view, as the commenta-
tors do, by saying some such thing as: ÒA philosopher will not fear death,
since his entire life has involved the practice of deathÓ.14 For this leaves
entirely unclear what it is to ÔpracticeÕ being dead. Furthermore such a

not admitting immortalism is strikingly although indirectly conŽ rmed in a (much later)
argument attributed to Chrysippus, who grants that death is the separation of the soul
from the body, but who takes this to imply that the soul is not immortal: since only
those things that were in contact can become separated, and since only a corporeal
can be in contact with a corporeal, then the soul as well as the body must be corpo-
real (SVF 2.270). (I owe this point to Sean Kelsey.)
13
That CA begins from conclusions of SD, shared by SocratesÕ interlocutors, is per-
haps suggested as well by CebesÕ remark at 69e6: ÒThe other things you say, Socrates,
I Ž nd excellent; but what you say about the soul is the subject of much disbeliefÓ.
Notice Cebes does not say that he himself doubts anything that Socrates has said about
the soul. Moreover, we need not take d¡ at 70a to be adversative: the sentence could
have the sense, ÒAlthough I personally agree with everything youÕve said, people are
liable to take issue with what youÕve said about the soul.Ó
14
Burnet: ÒThe philosopher will not fear death; for his whole life has been a
rehearsal of deathÓ (63e8 ad loc.), in PlatoÕs Phaedo, (Oxford: Clarendon Press, 1911).
Hackforth: Socrates Òis not so foolish as to complain at the approach of that for which
his whole life has been a preparationÓ (42). Gallop: ÒThe philosopherÕs whole life is
a preparation for death. He should therefore welcome death when it comesÓ (86).
Rowe: Òthe true philosopher will look forward to his death, which is a reward rather
than a punishment for his way of lifeÓ, in Plato: Phaedo, (Cambridge: Cambridge
University Press, 1993), 135. Yet later Rowe writes, somewhat more accurately: ÒIf
[dying and being dead] is the philosopherÕs sole preoccupation, he can reasonably be
expected to be Ôeager for itÕÓ (ibid., 64a8 ad loc.).
 DEGREES OF SEPARATION IN THE PHAEDO 99

summary contains no inference that involves maximization; yet SocratesÕ

Defense is clearly concerned with arguing from a partial to a maximal
condition: the philosopher aims to be as dead as possible while alive;
hence he welcomes the maximal condition of death, when it arrives.
To construe the argument correctly, we would do well to start with
the crucial deŽ nition of dying and death, which Socrates introduces at
64c4-8:
ÔDo we take death to be something?Õ ÔTo be sureÕ, replied Simmias. ÔThen is it
anything other than the separation of the soul from the body? And to be dead is
this: the bodyÕs having come to be separate, alone by itself, apart from the soul,
and the soulÕs being apart from the body, separated off, alone by itself? Is death
anything other than this?Õ ÔNo, thatÕs just what it isÕ, he said.
²goumey‹ ti tòn y‹naton eänai; P‹nu ge, ¦fh êpolabÆn õ SimmÛaw. ‰Ara m¯
llo ti µ t¯n t°w cux°w Žpò toè sÅmatow Žpallag®n; kaÜ eänai toèto tò
teyn‹nai, xvrÜw m¢n Žpò t°w cux°w Žpallag¢n aétò kay aêtò tò sÇma gegon¡nai,
xvrÜw d¢ t¯n cux¯n Žpò toè sÅmatow ŽpallageÝsan aét¯n kay aêt¯n eänai;
”ra m¯ llo ti  õ y‹natow µ toèto;

The deŽ nition is essential to the argument, since Socrates hauls it in at

the very point in his Defense when he wishes to conclude that the philo-
sophical life is the practice of death (67d3-10). Now we need to ask: what,
according to Socrates, is the relationship between what a true philosopher
achieves in living a philosophical life, and what is picked out by this
deŽ nition? To what extent does a true philosopher, in his asceticism and
devotion to contemplation, realize what the deŽ nition describes? We can
distinguish four possibilities, I think, which for clarityÕs sake may be
explained in terms of a simple example:
(1) A true philosopher achieves what the deŽ nition describes only in
intention, in the way that someone who yearns to visit Paris and thinks
about it constantly might reasonably say ÒItÕs almost as if IÕm in Paris
alreadyÓ. But in fact he isnÕt in Paris; and he never gets closer to Paris
simply by yearning to be there. Call this a realization of the deŽ nition
in intention, merely.
(2) A true philosopher achieves some analogue of what the deŽ nition
describes, in the way in which someone with limited resources might
travel to Montreal and say ÒThatÕs as close to being in Paris as IÕll
ever get.Ó He never is in Paris; he is only in that city in a neighbor-
ing francophone country that has a character similar to Paris. Call this
an analogical realization of the deŽ nition.
(3) A true philosopher achieves exactly what the deŽ nition describes
although to a less than maximal degree: as someone who has a
100 MICHAEL PAKALUK

one-day stopover in Paris might say something like ÒI was in Paris

once, but only brie y and would love to go back for a longer visit.Ó
He truly is in Paris and would eagerly be there longer for more of the
same. Call this a strict realization of the deŽ nition, differing from
other such realizations only in degree.
(4) A true philosopher achieves ÔsubstantiallyÕ what the deŽ nition de-
scribes and falls short from a full realization of the deŽ nition in some
accidental or superŽ cial respect: as someone might say that he stayed
long enough in Paris to see all the principal sites but would love to
go back and live there for a year. HeÕs gotten most of what he wants
to get out of Paris, but of course thereÕs always more to see. Call this
a substantial realization of the deŽ nition.
Note that all four cases involve someone who would eagerly welcome a
visit to Paris, and all four involve someone who might, in some sense of
ÔpracticeÕ, be described as having practiced or prepared himself for spend-
ing time in Paris.15
So what kind of separation of the soul from the body does Socrates
maintain that a philosopher practices? We can rule out (1) and (2) straight
away, since that would give SD something of the status of a bad pun. If
no more than that were meant, then, for all his asceticism, a philosopher
would not be one jot closer to the condition of being dead than anyone
else. Furthermore, it would make SD pathetically circular: a philosopher
eagerly anticipates death – and why? Because his life is one of anticipa-
tion of death. Perhaps at best SD could be regarded as an appeal to con-
sistency: a philosopher welcomes death, when it arrives, because he has
welcomed it throughout his life in the past.16 But then SD would lack any
of the character of an apologetic for the philosophical life, which Socrates
clearly intends it to be (63b4-c6, 63e8-64a3, 69d8-e5).

15
Of course someone who visited Paris in one of the stronger senses might also do
so in a weaker sense: a man who once traveled there might spend his years pining
and yearning to return.
16
I do not deny that at some points in SD Socrates appeals to consistency, in the
sense that he points out how absurd and foolish it would be, to vacillate: Òif theyÕve
set themselves at odds with the body at every point, and desire to possess their soul
alone by itself, wouldnÕt it be quite illogical if they were afraid and resentful when
this came aboutÓ (67e6-68a1). But a change in course is a foolish vacillation, rather
than reasonable repentance, only if the course is a good one from the start; yet this is
what SD aims to defend.
 DEGREES OF SEPARATION IN THE PHAEDO 101

Moreover, if Socrates were maintaining that a philosopher separates

soul and body only in senses (1) or (2), then SD would invite and pre-
sumably receive a very different challenge from what it gets. As we saw,
when Socrates Ž nishes his apologia, Simmias and Cebes concede that the
soul is distinct from the body and can be separated from it – what they
are concerned about is whether it stays long in that condition. Their fear
is that, at the moment when the soul of a philosopher comes into posses-
sion of his lifeÕs goal, it gets scattered into oblivion. They do not object,
as they obviousl y would, if Socrates had meant that a philosopher
achieves separation in only a metaphorical sense, that Socrates has still to
show that what the philosopher strives for is at all possible. Their objec-
tion is not that SocratesÕ defense is vitiated, but rather that it is incomplete.
Again, throughout SD Socrates uses phrases that are typical for express-
ing a difference of degree among things similar in kind – which suggests
that he regards the philosopherÕs separation as strict, although limited. The
philosopher Òdiffers from other men in releasing his soul, as far as possi-
ble (÷ti m‹lista ), from its communion with the bodyÓ (65a1); his soul
Òcomes to be alone by itself as far as possible ( ÷ti m‹lista )Ó (65c7); he
approaches Òeach object with his intellect alone as far as possible (÷ti
m‹lista )Ó, (65e6); philosophers are Òclosest (¤ggut‹tv) to knowledge . . .
if we consort with the body as little as possible (÷ti m‹lista )Ó (67a3).
We might imagine, too, that part of PlatoÕs purpose in introducing the
imagery of puriŽ cation into SD, at 65e6-66a7 and 67c5-d1, is to suggest
that separation in this life differs only in degree from death. The process
of removing an impurity is, of course, a process of reaching an extreme
degree of a quality already present. That Òwe can know nothing purely in
the bodyÕs companyÓ (66e4) implies that we do know it already, to some
degree, in the bodyÕs company. In linking the practice of death to the
notion of puriŽ cation, then, Socrates evidently means to suggest that dying
is a matter of reaching an extreme degree of a state already attained 17:
ÒpuriŽ cation,Ó he says, Òturns out to be just whatÕs been mentioned . . .
the parting of the soul from the body as far as possible, and the habitu-
ating of it to assemble and gather itself together, away from every part of
the body, alone by itself, and to live, both in the present life and in the
hereafter (kaÜ ¤n tÒ nèn parñnti kaÜ ¤n tÒ ¦peita), released from the
body, as from fettersÓ (67c5-d1). That Socrates in this way links together

17
Observe that this is exactly where the deŽ nition of death from 64c4-8 gets
recalled.
102 MICHAEL PAKALUK

a philosopherÕs present life with his condition after death would appear to
clinch the case.18
We can conclude, then, that SD afŽ rms that a true philosopher, through
his asceticism and dedication to a life of study, achieves death in the strict
sense and to some degree. But since separation in the strict sense could
not take place at all unless, in principle, it could take place completely,
what Socrates asserts in his Defense is substance dualism. We should take
this afŽ rmation, then, to be in force as Socrates next turns to the CA.

III. SD argues for substance dualism.

Yet does Plato give any argument for substance dualism in the Phaedo,
or does he simply take it for granted? It would help Ž rst to consider what
an argument for substance dualism might look like; we wonÕt have much
chance of recognizing one otherwise. In this regard, it is instructive to
recall the challenge that Descartes once posed to Mersenne: ÒDo you claim
that if we clearly understand one thing apart from another, this is not
sufŽ cient for the recognition that the two things are really distinct? If so,
you must provide a more reliable criterion for a real distinction – and I
am conŽ dent that none can be provided.Ó 19 The intuition behind DescartesÕ
challenge is extremely powerful. Let us distinguish between a Ôfull blownÕ
and a ÔprovisionalÕ argument for a real distinction. Both begin with a con-
ception of a thing which is coherent and putatively complete; and from
this they conclude that that thing which is so conceived can exist sepa-
rately. But a Ôfull blownÕ argument, such as that which Descartes gives in
the Meditations, aims at certitude and is re exive in character, since it
aims to justify that very step of arguing from our conception of a thing,
to the existence of the thing, in the manner conceived. A ÔprovisionalÕ
argument for a real distinction is more modest and aims, not at certainty,
but simply to shift the burden of proof. It too starts from our conception
of a thing, and draws an inference about the reality of the thing. But that
inference is taken to be intuitive and goes unexamined. 20 Presumably it is

18
Yet it should be said that there are passages of SD that indicate that separation
is intended in sense (4). For instance, at 66b6 Socrates says that Òas long as we pos-
sess the body, and our soul is contaminated by such an evil, weÕll surely never ade-
quately gain (kthsÅmeya ßkanÇw) what we desireÓ, which seems to imply that we do
gain it fairly well already in this life.
19
CSM 95.
20
Both sorts of argument may be found in the Meditations. It is not uncommon for
 DEGREES OF SEPARATION IN THE PHAEDO 103

this sort of argument, if anything, that we might expect to Ž nd in the

Phaedo.
But do we Ž nd such an argument? Our suspicion has to turn to the
deŽ nition of death, already cited, with which SD begins. It is curious that
commentators are divided as to whether the mere deŽ nition begs a
question. Gallop insists that it does: ÒFrom this deŽ nition,Ó he objects,
Òconjoined with the admission that there is such a thing as death (c2),
it follows that the soul does exist apart from the body. If it did not,
there would be no such thing as death, in the sense given to the word at
64c5-8. It therefore seems hard to acquit Socrates of prejudging the issue
at this pointÓ (86-87). Hackforth however writes: ÒThe deŽ nition does not,
of course, prejudge the question of the soulÕs survival; all that Socrates
here wants is an admission that we can properly think and speak of soul
ÔapartÕ from body; whether soul continues to exist when thus apart is the
question at issueÓ (44, n. 1). Yet it is telling that Hackforth puts ÔapartÕ
in scare quotes, as if the admission that we can properly speak of soul
apart from the body, without hedging or qualiŽ cation, would itself have
important implications. And then Hackforth adds something not to the
point, because, as we have seen, the issue of the soulÕs independent exis-
tence is distinct from that of its continued existence. The mere deŽ nition
might go some way towards establishing the former, without thereby
establishing the latter.21
I suggest that the reason the deŽ nition provokes controversy is that, if
it is coherent, it does provide prima facie grounds for a real distinction.
But presumably that is what SD is meant to show: it aims to show that
the deŽ nition is coherent, by isolating and characterizing an activity in
which the soul can engage without any evident need of the body.
Consider the following argument:
1. A exists independently from B, and therefore may exist if B ceases to
exist, if we can conceive of A, without thereby conceiving of B.

students, encountering that work for the Ž rst time, to take Descartes to have estab-
lished the distinction between the soul and body in the Second Meditation alone. That
is because the Second Meditation contains an intuitive or ÔprovisionalÕ argument. Yet
the whole apparatus of the ÔCartesian circleÕ, which is meant to justify the inference
from conception to the world, by giving a criterion of the trustworthy operation of our
powers of conception, is of course developed in the successive Meditations and not
concluded until the Sixth.
21
On the other hand, precisely because substance dualism is a weaker position than
immortalism, GallopÕs claim that Socrates has indeed ÔprejudgedÕ the issue is mis-
taken.
104 MICHAEL PAKALUK

2. To conceive of A, without thereby conceiving of B, it sufŽ ces to iden-

tify an activity F attributable to A, which is such that A can F with-
out any assistance from B.
3. A philosopherÕs experience is such that from it we can see that there
is an activity of the soul, thinking, which is such that, in principle,
nothing hinders the soul from engaging in it alone and without any
assistance from the body.
4. Hence we can conceive of the soul, without thereby conceiving of the
body.
5. Hence, the soul exists independently from the body and may exist if
the body ceases to exist.22
I regard an argument of this sort as implicit in SD. The deŽ nition of death
at 64c4-8 has the force of premises 1. and 2., and most of SD which fol-
lows is meant to underwrite 3.
It does so by a series of observations which have something of the char-
acter of a protreptic. A true philosopher, Ž rst of all, cares little about activ-
ities of the soul in which the body plays an essential role, on account of
their involving pleasures derived from nutritive or procreative functions of
the body: eating, drinking, and sexual intercourse (64d1-7). Similarly, he
cares little about aiming to adorn the body and making it attractive (64d9-
e1). Note that these activities are ones that cannot be described without
mention of the body: for instance, for a person to take pleasure in eating
just is for him to derive pleasure from a certain function of the body. So
none of these activities qualiŽ es as one in which the soul can be engaged
without the body, and none of them, therefore, could contribute to a con-
ception of the soul Ôalone by itselfÕ. To the extent that someone engages
in them, then, he is not dead; and when a true philosopher tries to avoid
these activities, he is aiming, at least, not to be not dead.
Socrates next takes his interlocutors to a second stage, pointing out that,
generally speaking, a true philosopher is concerned, rather, with acquir-
ing knowledge and truth. One might think that the use of oneÕs sense
organs is suitable for this, but in fact, Socrates claims, the senses are
deceptive and themselves contain no truth (65b1-7). 23 Thus a true philoso-
pher will care little about simply looking at things, hearing, smelling, and

22
The argument is meant to be roughly analogous to the paraphrase of DescartesÕ
in Margaret D. Wilson, Descartes, (London: Routledge & Kegan Paul, 1978), 189.
23
It is not important for our purposes why he claims this, only that he does so.
Commentators object that, in the RA which follows, Socrates will argue that sense
experience plays an invaluable role in provoking recollection of the Forms; conse-
 DEGREES OF SEPARATION IN THE PHAEDO 105

so on. Of course, all of these activities, too, require essentially the assis-
tance of the body, so in avoiding these, too, the philosopher aims not to
be not dead.
Finally Socrates turns to a particular activity of the soul, thinking or
reasoning: ÒSo isnÕt it in reasoning (logÛzesyai), if anywhere at all, that
any of the things that are become manifest to it?Ó (65c2-3). He considers
this activity both from the point of view of the activity itself (65c5-d1),
and from the point of view of the object of that activity (65d7-e9), and
in each case he argues both that there can be differences in degree of
ÔpurityÕ and that the philosopher strives for the purest degree:
And it reasons best, presumably, when- LogÛzetai d¡ g¡ pou tñte k‹llista,
ever none of these things bothers it, ÷tan aét¯n toætvn mhd¢n paralup»,
neither hearing nor sight nor pain, nor m®te Žko¯ m®te öciw m®te ŽlghdÆn
any pleasure either, but whenever it mhd¡ tiw ²don®, Žll ÷ti m‹lista aét¯
comes to be alone by itself as far as kayƒ aêt¯n gÛgnhtai ¤Çsa xaÛrein tò
possible, disregarding the body, and sÇma, kaÜ kay ÷son dænatai m¯ koin-
whenever, having the least possible com- vnoèsa aétÒ mhd ptom¡nh ôr¡ghtai
munion and contact with it, it strives toè öntow.
for that which is. (65c5-9)
. . . whoever of us is prepared to think ùw ’n m‹lista ²mÇn kaÜ Žkrib¡stata
most fully and minutely of each object parskeu‹shtai aétò §kaston dianohy°-
of his inquiry, in itself, will come clos- nai perÜ oð skopeÝ, oðtow ’n ¤ggætata
est to the knowledge of each? (65e2-4) àoi toè gnÇnai §kaston;

Typically sensations accompany thinking, but these are distracting, and

the soul thinks best when not bothered by them (LogÛzetai d¡ g¡ pou tñte
k‹llista, ÷tan aét¯n toætvn mhd¢n paralup», 65c5-6): we can easily
enough conceive of undistracted thinking and see how desirable it would
be. Furthermore, since the objects of thinking – such as the Beautiful
itself, the Good itself, and so on – are not bodily, we grasp these objects
better to the extent that we do not rely even on imagination (m®te t¯n öcin,
paratiy¡menow ¤n tÒ dianoeÝsyai m®te tinŒ llhn aàsyhsin 65e7-66a1).
Socrates and his interlocutors presuppose that thinking is a kind of
perception of something: it is a Òtouching of the truthÓ (t°w ŽlhyeÛaw

quently it would seem a true philosopher could not afford to dismiss it entirely, even
if this were possible. Yet such an admission makes the senses only instrumentally valu-
able for arriving at the truth, and the use of the senses would play no role in that
activity which, strictly, constitutes our grasping of truth. (We might think that Plato
has Socrates so strongly discountenance sense perception in SD, precisely because
what is at issue there is not truth so much as whether sense perception can be taken
as an end in knowing.)
106 MICHAEL PAKALUK

‘ptetai, 65b7), Òa viewing of the truest nature of thingsÓ (tò Žlhy¡sta-

ton yevreÝtai, e2). Therefore, they reason, if thinking were something that
we accomplished through the body, it would be accomplished through the
organs of perception that the body provides, the familiar Ž ve senses. However,
from the experience of thinking, it turns out that one succeeds better at
thinking, the less use one makes of sense perceptions and images derived
from sense perception. Thinking, therefore, in its purest form, does not
make use of images; hence, it does not make use of sense organs; hence,
it does not make use of the body at all. Thus, thinking is an activity in
which the soul can engage without the assistance of the body.
In sum, then, Socrates maintains, through this appeal to the experience
of a true philosopher, that we can conceive of an undistracted soul think-
ing of the Forms without relying on images. This is the limit, or ÔpuriŽ cationÕ,
of what a philosopher already achieves. To conceive of this is to conceive
of someone Òusing his intellect alone by itself and unsullied, [who under-
takes] the hunt for each of the things that are, each alone by itself and
unsulliedÓ (66a1-3). It is the pure activity of the mind, directed at pure
objects of thought:
Then would that be achieved most ‰Ar oïn ¤keÝnow ’n toèto poi®seien
purely by the man who approached kayarÅtata ÷stiw ÷ti m‹lista aét» t»
each object with his intellect alone as dianoÛa àoi ¤f §kaston, m®te t¯n öcin
far as possible, neither adducing sight paratiy¡menow ¤n tÒ dianoeÝsyai m®te
in his thinking, nor dragging in any other tinŒ llhn aàsyhsin ¤f¡lkvn mhdemÛan
sense to accompany his reasoning; metŒ toè logismoè, Žll aét» kayƒ aêt¯n
rather, using his intellect alone by eÞlikrineÝ t» dianoÛ& xrÅmenow aétò
itself and unsullied, he would under- kay aêtò eÞlikrin¢w §kaston ¤pixeiroÝ
take the hunt for each of the things that yhreæein tÇn öntvn . . .
are, each alone by itself and unsul-
lied . . . (65e6-66a3)

This protreptic supplies premise 3. of the argument above, which then

yields the desired conclusion. The Forms play a crucial role in the argu-
ment, since presumably an activity is not completely speciŽ ed until we
specify the object of that activity. In his Defense, Socrates characterizes
thinking, in effect, as an activity of the soul which has the Forms as its
object. A true philosopher, in having discovered the Forms, has simulta-
neously discovered something about himself, namely, that the soul is the
sort of thing that can engage in this activity of grasping the Forms.24

24
It is because of this connection between knowledge of the Forms and self-knowl-
 DEGREES OF SEPARATION IN THE PHAEDO 107

Consequently, he can conceive of the soul as engaged in this activity,

without thereby bringing in any mention or reference to the body. Thus
he can form a notion of the soul Ôalone by itselfÕ: from which he con-
cludes that the soul, in its activity of thinking, is independent of the body
and may exist even if the body is destroyed.
The basic argument for dualism in SD, then, is a provisional argument
for a real distinction. Yet we may identify in SD two additional and sup-
porting considerations for substance dualism. The Ž rst is at 66a10-67b5,
where Socrates points out that a philosopher who is striving to achieve,
by asceticism, a complete separation of his soul from his body, will
quickly see that he is taking the wrong track: he cannot gain entire pos-
session of what he wants, until the gods liberate his soul from the body
at death. The reason for this, Socrates intimates, is that the human body
and other corporeal things are at war with the activity of thinking.
Thinking is not something encouraged in the corporeal world, supported
by it, or aimed at by it. Rather, a true philosopher will engage in think-
ing in spite of the body, with its cares and distractions, and in spite of the
corporeal world as a whole, with its business, wars, and absence of leisure.
Thinking, then, is something alien to the world, and therefore its complete
fulŽ llment is to be found, if at all, outside the world. Yet this considera-
tion has the force of making it at least initially implausible, that thinking
is in some way dependent on the body, since one might naturally expect
that anything that depended on the body would be at home with the body
and serve its needs: consider, for instance, how we do not need to set our-
selves at odds with the world in order to see with our eyes, or to enjoy
the pleasures of eating.
The second consideration is the deduction of the virtues at 68b2-69d7.
There Socrates famously argues that someone acquires true virtue only to
the extent that he lives a philosophical life and separates his soul from
his body. Yet why does SD conclude with such a deduction? One reason,
of course, is that Socrates has claimed that he is justiŽ ed in not fearing
death, because he has reason to expect that he will be better off, and be
among better companions, after he has died (63b7-8); and if people grow
in virtue in proportion to their success in separating their souls from their

edge that I take Socrates later to assert, in the course of RA, what appears to be over-
stated at Ž rst, viz. that the hypothesis of the Forms, and that of immortality, stand or
fall together: ÒIt is equally necessary that those objects exist, and that our souls existed
before birthÓ, 76e5-6.
108 MICHAEL PAKALUK

bodies, then one might reasonably anticipate that a fully separated soul
will have more virtue than any that is still embodied – thus corroborating
SocratesÕ claim. However, another reason for the deduction is surely that
it suggests that the soul is not naturally joined to the body: that the soul
fares poorly to the extent that it is joined, and that it  ourishes to the
extent that it becomes separated, would seem to indicate that the soul is
independent and distinct, since one might expect that nothing naturally
dependent on another thing would get better by being cut off from it.

IV. SD was intended by Plato to provide the context of the three initial
arguments and therefore counts as the Ôprimary argumentÕ of the Phaedo.
On the interpretation defended here, an interesting structure emerges for
the Phaedo. The three main parts of the dialogue may be taken to corre-
spond to distinctions we drew earlier: SD maintains that the soul is capa-
ble of existing when the body is destroyed; the three initial arguments
maintain that it does in fact do so, through repeated cycles of incarnation
(unless liberated from those cycles through devotion to philosophy); and
the Ž nal argument (FA) maintains that the soul necessarily exists when
the body is destroyed, because by no process of change in nature can the
soul go out of existence.25
The result of SD is provisional in the sense that it takes our ability to
conceive of the soul as independent of the body to be a prima facie rea-
son for regarding it as distinct. This result is, however, powerful, as far
as it goes, since it shifts the burden of proof. But since SD concludes
merely that the soul can exist when the body is destroyed, it provides only
a basis for hope that the soul continues to exist after the demise of the
body, as Socrates stresses in many passages: ÒI expect (¤lpÛzv) to join
the company of good menÓ (63c1); ÒthatÕs why I am not resentful, but
rather am hopeful (eëelpiw) that there is something more in store for those
whoÕve diedÓ (63c4-5); Òa man who has truly spent his life in philoso-
phy . . . is hopeful (eëelpiw ) that, when he has died, he will win very great
beneŽ ts in the other worldÓ (63e9-64a2); Òif thatÕs true . . . thereÕs plenty
of hope (poll¯ ¤lpÛw) for one who arrives where IÕm goingÓ (67b7-8);
Òthey may hope (¤lpÛw ¤stin) to attain what they longed for throughout
lifeÓ (68a1-2).

25
That is, we might say that the dialogue defends, successively, Òsubstance dual-
ismÓ, Òweak immortalismÓ, and Òstrong immortalismÓ. (See the chart at this articleÕs
end.)
 DEGREES OF SEPARATION IN THE PHAEDO 109

It is clear that Plato intended the provisional conclusion of SD to serve

as the context of the three initial arguments. As we saw, SD begins with
a deŽ nition of Ôbeing deadÕ (64c4-8) that employs unusual technical terms
in its deŽ niens: Ôbeing deadÕ indicates the soul alone by itself (aét¯n kay
aêt¯n); the body alone by itself (aétò kay aêtò); and both of them being
in a condition of separation (Žpallag®) from the other. We may there-
fore take any occurrence of these technical terms, in the text that follows,
as involving a reference back to the deŽ nition at 64c4-8. Hence we have
a rough criterion for determining whether a text is meant to depend upon
SD: if it contains this sort of technical language, it should be regarded as
falling under the scope of the discussion which begins with that deŽ nition
and is developed in SD.
However, what we Ž nd is that, as soon as Socrates Ž nishes with his
three initial arguments, this technical language from SD makes a reap-
pearance. In fact the ÒAfŽ nity ArgumentÓ (AA) merges smoothly into an
extended reprise of SD (80c1-84b8). 26 All the chief themes from SD are
brought forward once again: Ò. . . whenever [the soul] studies 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, enters
always into its company, whenever it has come to be alone by itself
(aét¯n kayƒ aêt¯n)Ó (79d1-4); Òsuppose too that it has been gathered together
alone into itself (aét¯ eÞw ¥aut®n), since it always cultivated thisÓ (80e5);
Òdo you think a soul in that condition will separate unsullied, and alone
by itself (aét¯n kayƒ aêt¯n)?Ó (81c1); when the soul is in the body it Òis
forced to view the things that are as if through a prison, rather than alone
by itself (aét¯n diƒ aêt°w)Ó (82e4); philosophy frees such a soul Òby urg-
ing it to collect and gather itself together, and to trust none other but itself,
whenever, alone by itself (aét¯n kayƒ aêt¯n), it thinks of any one of the
things that are, alone by itself (aétò kayƒ aêtò)Ó (83a7-b1). Even the ear-
lier deduction of the virtues is recapitulated at 83e5-7.
So we should regard SD, strictly, not as a passage which merely pre-
cedes the three initial arguments, but rather as that which frames them
and sets them off, providing their context. This conclusion is reinforced
by other references in the three initial arguments back to SD:
(i) Socrates concludes CA with the claim that heÕs shown that Òthe souls
of the dead exist and after death things are better for the good and

26
Surprisingly, none of the major commentaries sets off the passage following AA
and identiŽ es it as ÒReprise of SocratesÕ DefenseÓ, or some such thing.
110 MICHAEL PAKALUK

worse for the dead Ó (my italics, 72e1). Editors tend to strike the ital-
icized phrase, not on the basis of evidence in the manuscripts, but on
the grounds that, as Rowe puts it, the phrase Òmakes no sense in this
context and is clearly an interpolation from 63c6-7Ó, sc. from SD. But
if CA is meant to rely upon and develop SD, the phrase does make
sense: relying on the ÔhopeÕ articulated in SD, and which is implied
by the doctrine of ÔpuriŽ cationÕ (see 69c2-d1), Socrates anticipates a
fuller conclusion, to which he cannot quite help himself here, but
which he will more adequately support by the time he reaches AA
(cf. 80d5-81c).
(ii) AA refers back explicitly to SD at 79c2-8: ÒNow werenÕt we saying
a while ago that whenever the soul uses the body as a means to study
anything, either by seeing or hearing or any other sense – because to
use the body as a means is to study a thing through sense percep-
tion – then it is dragged by the body towards objects that are never
constant, and is confused and dizzy, as if drunk, in virtue of contact
with things of a similar kind?Ó The reference is to 65b1-7. But what
is especially striking is the formula, toèto g‹r ¤stin tò diŒ toè sÅmatow,
tò diƒ aÞsy®sevw skopeÝn ti, which makes it clear Plato wants to
stress the principle that, if it is necessary to mention or refer to the
body in specifying that activity, then that activity is not one that the
soul can engage in Ôalone by itselfÕ.
(iii) Again, AA refers back to SD at 79d9-e1: ÒOnce again, then, in the
light of our earlier and present arguments (¤k tÇn prñsyen kaÜ ¤k tÇn
nèn legom¡nvn), to which kind do you think the soul is more similar
and more akin?Ó Since Socrates is looking for his interlocutor to draw
a conclusion based on what he had just said, at 79c2-8, it is clear
that Socrates means to include SD among the earlier arguments,
viewed as of a piece with the present ones. (It is not to the point that
Socrates speaks of legñmena rather than lñgoi; clearly AA is an argu-
ment, and he uses the same term for SD as AA.)
(iv) The passage at 79e9-80a5 has little cogency if not read with SD as
its backdrop. That passage begins: Ò. . . when soul and body are pre-
sent in the same thingÓ (¤peidŒn ¤n tÒ aétÒ Îsi cux¯ kaÜ sÇma). But
this is an odd Ž gure of speech, apart from the viewpoint of substance
dualism; yet this way of speaking has not been licensed by any par-
ticular text after SD. Furthermore, the passage presumes that Ônature
ordains that [body] shall serve and be ruled, whereas the other shall
rule and be masterÕ, and commentators complain that this premise is
entirely new, unsupported, and at odds with the notion that the soul
 DEGREES OF SEPARATION IN THE PHAEDO 111

is imprisoned in the body. Yet we may take this language to come

directly from SD, where, in his discussion of the false virtues of the
many, Socrates noted that the soulÕs servitude to the pleasures of the
body was declared contrary to the divinely decreed order of things:
see 68e7, 69a2, and 69b7.
Hence SD is properly regarded as providing the context for the three ini-
tial arguments. But then how do those arguments rely upon SD? We have
seen already how CA does: it presupposes substance dualism, which SD
articulates and supports. But RA and AA also depend upon SD. The basic
intuition underlying CA may be summarized as: since the soul is distinct
from the body, and nothing comes from nothing, then we should postu-
late reincarnation, as the most plausible explanation of the origin of body-
soul composites. Similarly, RA may be understood as arguing that, since
the soul is distinct from the body – and indeed since the soul may be
understood as that which is capable of grasping the Forms – then, since
nothing comes from nothing, and since the soulÕs capacity to grasp the
Forms cannot be accounted for by its existence after it is joined to the
body, then this capacity must be accounted for by its existence before its
being joined to the body. 27
Again, AA is aptly understood as strengthening the conclusion of SD,
by proposing that the soul and the body belong to larger classes, each
characterized by attributes which are incompatible with those that belong
to the other. Soul belongs to the class of imperceptible, invariable, and
incomposite things; body belongs to the class of perceptible, variable, and
composite things. Hence we may have additional conŽ dence that soul
and body are distinct, because the classes in which they are appropriately
grouped are distinct and indeed incompatible.

27
In this way we see how a common objection to RA may be answered. Bostock
objects that Plato gives no reason why the capacity to grasp the Forms might not sim-
ply be present in the soul from birth, in the manner of Ôinnate ideasÕ: Ò. . . it is open
to an objector to maintain that we were just born with this knowledge of equality, but
did not have to learn it at any previous time. When Plato himself belatedly notices
this objection, at 76c14-15, he gives a reply which can only be regarded as missing
the pointÓ (102). But it looks rather as if Bostock has missed the point: to hold that
the soul comes into existence with such knowledge would be to postulate an effect
without a cause. How does it come to have those innate ideas? Nothing comes from
nothing. (The doctrine of innate ideas that survives into the modern era presupposes
for its cogency the view that the soul is immediately created by God.)
112 MICHAEL PAKALUK

We may note that, when understood in this way, AA resembles an argu-

ment in Descartes as well. In his Sixth Meditation, Descartes Ž rst draws
to a close his elaborate argument for a real distinction: ÒThe inference to
be drawn from these results is that all the things that we clearly and dis-
tinctly conceive of as different substances (as we do in the case of mind
and body) are in fact substances which are really distinct one from the
other; and this conclusion is drawn in the Sixth Meditation.Ó 28 We have
seen that SD supplies a rough analogue of this argument. But once
Descartes Ž nishes his argument for a real distinction, he gives an addi-
tional argument, intended to strengthen this result: ÒThis conclusion is
conŽ rmed in the same Meditation by the fact that we cannot understand
a body except as begin divisible, while by contrast we cannot conceive of
a mind except as being indivisible . . . and this leads us to recognize that
the natures of mind and body are not only different, but in some way
oppositeÓ.29 That is, he assigns body and mind to different classes, which
have incompatible attributes. 30
All three initial arguments, then, are correctly viewed as taking the sub-
stance dualism of SD as their starting point. But since they argue, not
merely that the soul may exist when the body is destroyed, but also that
it does so exist, they go beyond supplying a basis for ÔhopeÕ, and provide
Ԏ rm evidenceÕ and ÔsufŽ cient evidenceÕ ( ßkanòn tekm®rion, 70d1, cp. 72a6),
of the sort that can ground conviction (72d6). Nonetheless they are induc-
tive and analogical rather than demonstrative, and they reach no conclu-
sion having the character of necessity, and thus they are still open to
reasonable doubts. Note that, on this way of construing the Phaedo, there
is not the slightest reason to take Plato to have regarded any of the three
initial arguments as fallacious or unsound.
But whereas the initial arguments respond to a worry of the common
man, not entirely shared by SocratesÕ interlocutors, the objections of Simmias
and Cebes involve worries that arise from within the new natural philos-
ophy, with which Simmias and Cebes evidently do have sympathy: SimmiasÕ

28
The quotation is from DescartesÕ Synopsis, CSM 9.
29
CSM 9-10.
30
Yet the FA is still needed, because AA says merely that the soul is best likened
to indestructible things, and therefore has the character of an induction or analogical
inference; also, because the whole drift of SD and its reprise is that the soul may have
more or less of the character of corporeality, and so it needs to be made clear that its
having this character is secondary, derivative, and accidental to it.
 DEGREES OF SEPARATION IN THE PHAEDO 113

objection, relying on a view of the soul as a harmony and proportion, aims

to overturn the presumption of dualism carried over from SD; CebesÕ
objection, perhaps broadly indebted to Heracleiteanism, strikes at the absence
of necessity in the earlier arguments, already noted. 31 To respond ade-
quately to these, Plato needs to forge new philosophical tools: dialectic,
to respond to Simmias, and a rudimentary metaphysics, to answer Cebes.
The reply to Cebes of course aims at something like demonstration from
incontrovertible premises, and it would show that necessarily the soul con-
tinues to exist.
The Phaedo, then, is not a dialogue in which Socrates, after an initial
expression of faith, searches without much success for vindicating reasons,
but rather one in which Socrates and his interlocutors, through reasoned
discussion, arrive at increasingly stronger reasons for increasingly stronger
claims, developed in response to increasingly stronger challenges. The dia-
logue is thoroughly constructive in character, with later parts conŽ rming
earlier ones. 32 Its general structure may be represented in a table such as
the one on page 115.
Etienne Gilson concluded his classic study of the history of philosophy,
The Unity of Philosophical Experience by remarking that Òthe recurrence
of similar philosophical attitudes is an intelligible factÓ.33 There are pat-
terns of philosophical thought and argument throughout history, which
have a kind of law, or necessity, that transcends the purposes of the
philosophers who express them. Dualism is one such philosophical atti-
tude, and Plato, of course, was a dualist. It should hardly be surprising,
then, that a kind of inner necessity in dualism governs the way in which
he constructs his great dialogue on the soul, bringing it into correspon-
dence, in interesting ways, with DescartesÕ project. To read the Phaedo
in this manner, selon lÕordre des raisons, is to recognize that the usual

31
I regard the objections of Simmias and Cebes as representing PlatoÕs concern that
the new natural philosophy denies the priority of the soul – a concern which Ž nds full
expression in Laws X. There Plato insists that the soul is older than corporeal beings
and their cause, whereas the new natural philosophy would make the soul derivative
and dependent upon corporeal things. See 889b1-890b1, 891c1-8, 892a4-b8.
32
Note that, at the end of the Final Argument, both Cebes and Simmias afŽ rm that
they now Ž nd all the arguments beyond doubt: Ò. . . for my part IÕve no further objec-
tions, nor can I doubt the arguments at any pointÓ (Cebes, 107a3); Ò. . . nor have I
any further ground for doubt myself, as far as the arguments goÓ (Simmias, a8-9).
33
Etienne Gilson, The Unity of Philosophical Experience, (New York: Charles
ScribnerÕs Sons, 1937), 302.
114 MICHAEL PAKALUK

diagnoses of fallacious argument are much too superŽ cial, as are those
remedies that would have us look instead at incidental dramatic details.
There is no merit in simply overlooking the powerful philosophical view
which the dialogue expresses. 34

Department of Philosophy
Clark University

34
I wish to thank Sean Kelsey, Anthony Price, and Paul Woodruff as well as the
participants in the 25th Texas Workshop in Ancient Philosophy for comments on
earlier drafts of this paper.
 The Argumentative Structure of the Phaedo

Motivating Worry Source of Worry Thesis Defended Kind of Argument Kind of Conviction

I. SocratesÕ Perhaps Socrates SocratesÕ friends Substance dualism Prima facie, Hope
Defense has recklessly and disciples provisional
neglected his
own good in not
resisting death

II. The Three Perhaps the soul The Ôcommon manÕ Weak Inductive, Sound inference
Initial disperses upon immortalism plausible story
Arguments separation from (reincarnation)
the body

III. The Final Perhaps the soul Presocratic Strong Deductive, Certainty, proof
Argument wears out natural philosophy immortalism necessary
eventually (divinity of
the soul)
DEGREES OF SEPARATION IN THE PHAEDO
115