Gregory Vlastos (1983). The Socratic Elenchus. Oxford Studies in

Gregory Vlastos (1983). The Socratic Elenchus. Oxford Studies in

ï
[l
THE SOCRATIC ELENCHUS
å
Ê
I
I
iil
THE SOCRATIC ELENCHUS1
m
GREGORY VLASTOS
Ë
Ë
I
whose ¡ationale he does not investigate. They are constrained by rules he
does not undertâke to justify. In marked contrast to the'Socrates' who
speaks for Plato in the middle dialogues, who refers frequently to the
'melhod' (pétoôoç) he follows (either systematically3 or for some parti@
G¡egory Vlastos
I
19E3.
Reprinted with pe¡mission f¡om
ó.ford
An ea¡lie¡ dr¿ft of this essay was delivered as one of a se¡ies of lectu¡es, ,fhe pbilosophy
served as commentato¡ at that meetiDg, for his exceediDgly âcute a¡d sìtggestive critique of my
paper. I t¡ust he will soon put into print his own, highly original, interpretatioD of thé Soc¡atic
elenchus.
'z The ch¡onological order of those works of Plato which I accept as authetrtic (not mâterially different from that generally recognized in recerìt Platonic scholarship) is as followsì
i¡
Ë
E
S¡adies in Akcient philos-
of Soc¡ates', under the Giflord Trust at the University of St Andrews in the winter Ànd
spd¡g terms of 1981. A ìater draft was presented at a meeting of the Ame¡ican philosophical
Association on 29 Dec. 1982. An âbstract of that drâft was published jD the JouriaÌ of
Philosophy,'7g (1982),711-1,4. I am deeply indebred to my friend Richa¡d Kraur, who
(1) The ea¡lier dialogues (listed
cular purpose in a special contexta), the'Soc¡ates' who speaks for Socrates
in Plato's earlier dialogues never uses this wo¡d5 and never discusses his
method of investigation. He never troubles to say why his way of searching is the way to discover truth or even to say what this way of searching
is. He has no name for it. 'Elenchus' and the cognate verb, elenchein (lo
refute, to examine critically, to censure), he uses to describe,6 not to baptize,
what he does;only in modern times? has'elenchus' become a proper name.
So the'What is F?' question which Socrates pursues elenctically about
other things, he never poses about the elenchus, leaving us only his practice of it as our guide when we try to answer it ourselves. Lacking his
deflnition of it, ours can only be a hypothesis-a guess. And we may guess
wrongly.
In Plato's earlier dialogues2-in all of them, except the Lysis, Euthydemus,
arrd Hippias Major-Socrates' enquiries display a pattem of investigation
opht,I (19$),n-58.
37
âlphabetical oÃet\: Apology, Charmides, Cito,
EuthydÊmu,s, Euîhtph¡o, Gotgía6, Hippias Major, Hippias Mikor, Ion, Laehes, Lysß, MeneÍehus, Proragoras, Rep¡råfic 1.I take the ¿r,r¡.r, EuthydemLß.anð Hippias MdÞr, to be the latest
of these (see the appendix below), falling between the co;g¡dr (which I take to be the only
one of the ea¡lier dialogues to precede this trìo) ând the Merro, which I tåke to mark the point
of t¡ansition fÌom the earlie¡ to the middle dialogues. I group the frrst book of the R¿píbljc
with the ea ier dialogues: Soctatic elenchus (which, ås I argue in the appendix, is dropped in
Lysß, Euthydemus,
Hippias Malo¡) is practised there as vigorously as any.where in the
^trd
corpus,
(2) The middle dialogues (listed in p¡obable chrorological order): C¡atylus, phaedo, Republic 2 10, Symposium, Phaedrus, Parmenides, Theaeretus,
(3) The later dialogues (also in probable chronologicâl otder): Timaew, Crttias, Sophíst,
Po liticus, Phileb u.s, Laws.
I 'Our customa¡y method'0R¿¿ 10,596a5J);'the dialecrical method'(Rep. ?,533c?), which
has beeû explåined (533b2-3) as the only'method which endeavourc in every case to apprehend concefning €ach thing what it really is'.
F
fi
å
&
B
I guessed wrongly twenty-flve years ago in the account of the elenchus
I put into my Iítro duction to the Ptotagoras,s and so have others before or
since. Here is the one in the article 'Dialectic' by Roland Hall in the Encycloped.ia of Philosophy (New York, 1967): 'The Soc¡atic elenchus was
perhaps a refined form of the Zenonian paradoxes, a prolonged crossexamination which refutes the opponent's original thesis by getting him to
draw from it, by means of questions and answels, a consequence that cont¡adicts the thesis.' This comes close, but still makes three mistak$. Obvi
ously wrong is the suggestion that Socrates gets the opponent to draw the
consequence that contradicts the thesis. It is Socrates who draws it; the
opponent has to be calried to it kicking and screaming. Less obviously and
'ûethods' he has followed i¡ wo¡king out the tripartite analysis of the soul (ÿp. 4,
also the desuiption (without use of the \\orð methoàos) i.î P¡¡d, 99d4-100b4 of the
method he is to follo$, i¡ the final argument for the immo¡tality of thp soul.
s Tlþ word method.os, \sed often in dialogues of the middle period and almost as often in
4 The
435d).
Ct
Ë
those of the late¡ period (see s.v. methodos,Lf"or,ard Brandwood,.4 Wo¡d ltld¿x lo PIaîo (L,eedsl
Maney, 1976)), created by Plato (its tust occu¡rence in preserved Greek is in the Pl¡d,, ?9e3,
97b6), is itself an expression of the intensity of its creator's new-found interest iD method. lt is
impo¡ta¡t te¡minological coinage, strangely overlooked by Lewis Campbell i¡ his discussion of
'Plato's Technicalities' (ir få¿ Sophistes a/rd Politicus ofPldto (Oxfo¡d: Oxford Unive$ity Press,
195?), pp. xxivff.), which he locates primarily in the later diâlogues.
ó And this in great profusion. There a¡e dozens of ùses oI the nouD and the verb in Plâto, a
majority of them in the earlier dialogues, as a look ât B¡andwood,.4 Word lñdet to Plaþ,will
$
show.
Þ
E
E
F
Ë
F
å
E
H,
Ë
B
E
ii
? Perhaps no earlie¡ than its use for this purpose by George G¡ote
Plato ûnd the Other
Compøníoks oÍ Socrates,lst edn.3 vols., (London: Murray,1E65) (âü my references to this work
throughout this paper will be to vol. i of the 1st edn.) and by Lewis Campbell, ?¡¿ Sophistes
Politicus of Plato. lr was üsed for the sâme purpose by Henry Sidgwick soon after
('The Sophists', Jounr¿¡ ol Philolo*y, Ns 8 (1A72) ), no doubt under the in-fluence of Grote and
Câmpbell, to whose work he refeß.
3 P¡¿¡ot Protago¡âs, trans. Jowett,rev. M. Ostwald, ed. with introd. by G Vlastos (Indianapolis:
Bobbs-Merrill, 1956). I have revised some of the views I express in that intoduction. lts most
serious e¡ro¡ is its misinte¡p¡etation of the elenchus (oD ìrhich, sep pp. 51-4 below) and,
consequentlf of the p¡ofessioû of ignoranca,
d¿¡d
llll
lrl
i
cREcoRY vlAsros
Ë
more seriously wrong is the assimilation of the elenchus to Zeno's dialectic, from which it differs in a fundampntal respect: Zeno's lefutands are
unasserted counterfactuals:
þ
becomes how Socrates can claim, as
g
has established is the inconsistency
38
#there
are many things, they must be both inflnitely many and
finitely
mâny.
there is rnotion, then the s',viftest cânnot overtake the
Achilles will never catch up with the tortoise.
f
Socrates,
slo\ryest:
on the other hand, as we shall see below, will not
B
g
debate
unasserted premisses-only those asserted categorically by his interlocutor, who is not allowed to answer'contrary to his real opinion'
A third mistake is the notion that the consequence which contradicts
the thesis is drawn from that thesis, that is, deduced from it. The notion is
an invention of Richard Robinson. In his P/alo's Earlier Dialectice Robirtson had maintained that Plato 'habitually thought and wrote as if all
elenchus consists in reducing the thesis to a self-contrâdiction' (28) If
that were true, Soüâtes' procedure would have been as follows:when the
answerer asserts ?, Soüates would derive not-p either directly from p or
else by deriving from p some further premisses which entail not-p-in
either case deducing not-p from p 'without the aid of any extra premiss'
(ibid.). The trouble 'trith this picture is that what it pictures is not in our
texts.loThere are some thirty-nine elenctic arguments by Robinson's count
in Plato's earlier dialogues (ibid.24). Not one of them exhibits this pattern.
The premisses from which Socrates deduces not-p generally do not include
p;and even when they do, there are others in the premiss-set, elicited from
the interlocutor without any reference to p and not deducible from it.
If Socrates thought he proved what, according to Robinson, Plato'habitually wrote and thought' he did, Socrates would have believed he was
producing the strongest possible proof of the falsehood ofp: there can be
no stronger proof of the falsehood of a thesis than to show that it entails
its own negation. What Soüates in fact does in any given elenchus is
convict p not of falsehood but of being à member of an inconsistent
premiss-set;and to do this is not to show thatp is false, but only that either
p is false or that some or all of the premisses are false. The question then
(1st edn. Ithaca, NY 1941;2nd edn. Oxford: Clarendon Press, 1953.) My refe¡ences are to
the latter. Iû spite of this and other mistakes, this is an adrDtâble book, which se¡ved me as a
model of exegesis in my earlier PlatoDic studies. See the t¡rbute to it in my review of Harold
Cherniss, Collected Papers, ed,L.larán, Ameücan Joümal oÍ Philoloty, 89 (1978), 537-43: 538.
r0
As poir¡ted out by Paul Friedländer and Harold Cherniss at the time: fo¡ the references,
ând fo¡ my discussion of the textual evidence, cl my review ot Chemiss cited in the p¡eceding
F
E
Þ
å
É
þ
E
ft
I
39
F
I
È
F
Ë
Ë
fr
F
þ
Þ
Ë
E
F
å
ofp with premisses whose truth he has
not undertaken to establish in that argument: they have entered the argument simply as propositions on which he and the interlocutor are agreed
This is /¡e problem of the Socratic elenchus, and it is spirited away in the
account given by Robinson in 1941 and 1951 and repeated in the Ency'
clopedia article in!967.1'z I shall be returning to this p.roblem in due course.
Let me then suggest a more defensible description:
Soc¡atic elenchus is a search for mo¡al truth by adversary argument
in which a thesis is debated only if asserted as the answerer's own
beliel who is regarded as refuted if and only if the negation of his
thesis is deducedl3 ftom his own beliefs.
Elenchus is first and lasf search. The adversary procedure which is suggested, but not entailed, by the Greek word-which may be used to mean
'refutation', but also 'testing', or still more broadly 'censure, reproach'-is
not an end in itself If it were, Socrates' dialectic as depicted in the earlier
dialogues would be a form of verbal jousting-'eristic'14-which it is not,
be explained ât the start of Sect.Il below.
D And still beiûg ¡epeated:'One of the commonest forms lof elenchus] is to årgue that a
giver statement leads to a sell-cont¡adiction, ií otheÌ words to two statements which ale mutu'
alty cortradictory' (G. B,l<eíeÀ, The Sophßtic Movement (Cambtidger Cambridge University
Press,1981),65, v,ith a footnote citing Robinson as authority).
r3 Tbe nteDded force of the argumeDt is deductive throughout; resort to epaSote is Do eÍcep'
tion, for itr its Socratic use this is not true iDduction: see Vlastos, Int¡od. to Plarot Protagoras,
p. xxix and nn. 18 and 45.
_
la As rnisconceived by G, Ryle in his desc¡iption of'the Soc¡atic Method' (article on Plâto in
P Edx,a¡ds (ed.), Enclclopeàìa of Philosophy (New York Macmillan aDd Free Press; I-ondon:
his Plato's ProSress (Cambridge: Cambridge UnjCollier Mâcmillan, 1967), 317) and also
vefiity P¡ess, 1966), 119, where the elenctic arguments in Plato's eârlier dialogues are reple'
seflted as'speçimens of eristic contests'.The misconception is abetted by a blatant disregârd of
the'say what you believe' reqì]Ilement (to be discussed below), which is igDored even in the
admirable essây by Paul Moraux,'La Joute dialectique d'âprès ?opiqt¿e.t yZ¡ (irr G E. L oweû
(ed.), Atistotle on Diakctic (Oxfordt Oxford Univeßity Press, 1968), 277 ff.| z9É00)-ar.
incomparably more exact discussion ofthe topic, which notes carefully some other points of diÍferencè between Socratic elenchi and Aristotle's 'dialeclicat jousts' Similår disrega¡d of that
requirement accounts for other conflations of Socratic dialectic $rith elistic, begiDDing Ì'ith G
G¡òte: he makes no mention ofit in his discussioî (Plato anà the Othet Companions of Socrules,
531) of'the reat contrast' between Socrates and the o¿¿¡lé eristics iû the Eulhydemus This is
what makes it possible fo¡ him to say thât in the Pro¿a8o¡¿ç Socrates is 'decidedly more Eristic'
than the sophist (ibid. 535): he is usi¡g 'edstic' with cùlPable looseDess to mean'contentious'.
Contentiousness irl argument is indeed one of Socrates' taiÌings (fo¡ which Plato, in retlospect,
gently reproaches him; fltr 167d-168c). But in spite of such personal laPses on the part of its
úumán inst¡ument, elenchus remains in princÞle a method of searching for t¡uth, which e¡ìstic
is not, but only â method (or set of methods a whote bag of tricks) fol winning arguments,
regardless of whethe¡ or not you take what you ale arguing 1o¡ to be t¡ue (cf the excellent
iî
Ë
I
shall be arguing he does claim in
'standa¡d elenchus',ll to have proved that the refutand is false, when all he
tt This term will
'
note.
THE SOCRATIC ELENCHUS
E
40
lilrn
I
GREGoRY
vlAsros
THE socRATIc
Ë
ELENcrtus
41
because its object is always that positive outreach for truth which is
expressed by words for searching (èpeut ôt, ðr.epeut'ôt), inquiring ((7rô,
Ë
Elenchus is used to correct mistakesls-its proper, purely negative, use in
ß
being silenced by the civic authorities, he imagines them saying to him
p
philosophical dialect as conceived in Plato's middle dialoguesr6-but not
to discover, still less prove, the proposition which constitutes the true solution to the problem.l? As to the last example, it does, of course, refet to the
unreconstructed Socrates of the earlier dialogues.ls But note that he does
not elicit from his interlocütors the logical conditions for the dght answet
to a'What is F?' question: he produces them entirely on his own initiative,
tells the interlocutors what they are, and requires them to comply, never
inviting elenctic argument on whether or not they are the right conditions.
Thus when he tells Laches that the definition of'courage' must cover all
of the agreed-upon cases of courageous conduct, he does not ask,'Do you
agree?' but only'Do you understand?'1e And this is generally true. The
interlocutor is never shown as having dissenting vie\rys about the logical
pattem to which a good definition should conform-views which need to
be refuted by elenctic argument before the search can start. He is shown
as all at sea on the topic, too confused to have any opinions at all, needing
instruction on its very rudiments, which Socrates is only too willing to
provide. He offers it encountering not opposition but incomprebension.2o
The logical truths governing deflnition, and the still more abstract ones,
Êpøtôt, ouueparõ), investigating (oror.ât, ðnorotût, orcémopa4
ðrøotcémopøt).Tltis is what philosophy ri for Socrates. When he thinks of
(T1)
,..
you shall no longer etgage in this search ¡ror philosophize . . . (Ap. 29c)
the'nor' is epexegetic. Equivalentl¡ for Socrates to philosophize is
to'examine'-he searches by examining: were he to go to Hades, he says,
\¡/here
&
he would go on
(T2) . .. examining and searching there
as
I have been doing with people
he¡e.
@p.4tb)
F
e
F
What is he searching for? For truth, certainly, but not for every sort
of truth-only for truth in the moral domain. If we lvanted to know
what is the wholesale price of olive oiL on the Piraeus market, Socrates
would not propose that elenctic argument is the way to find out. Nor yet
for, say,
Þ
Þ
p"
What is the right diet for a patient with a fever?
What is the side of
a
ts Men.83b-e,
square whose area is twice that of a given square?
Whât conditions must be sâtisfied by a true answer to a'What is F?'
question?
F
È
There is no reason to suppose that Socrates thinks that truths in the
domain of the prâctical technai oÍ of mathematics or of logic are to be
ascertained by elenctic argument. He never says or implies anything of the
kind. My last two examples are meânt to be provocative. The mathematical one, of course, is f¡om the intenogation of the slave boy it the Meno.
In the'Socrates' of this passage Plato has already taken a giant step-the
doctrine of 'recollection'-in transforming tþe moralist of the earlier diaIogues into the metaphysician of the middle ones. The interrogation is
laid on to support that doctrine-to help Meno 'recollect' it (81e-82a).
Ë
Ë
ü
I
Ë
F
description of it ir KerfeÀ, The Sophi.stic Movement,62l, ard Euthd.272Ã-b, ciled iî r.26
below), while for elenchus the aim of'comirlg to know v]hat is true and what is false' is pa¡aúourt (G.9. 505e; cl Cl¡m.l66o-d aî'd also G/8. 458a, circd in n, 27 below). (For Adstotle's
recognition of 'saying what one believes' in Soc¡âtic diaiectic, see fop. 160"19-22: the answerer
is not just 'maintaining a posítiotr for the sake of argument, bÌrt saying what he belìeves'; though
Sodates is not named, the examples show that the refereoce is to Soqates' arguments with Callicles and Polus ir the Go¡gio,r,)
P¡d. 85c, 101d--e, 107b; Rep. 8, 534b-e.
'ó
L7
The method of discove¡y ií this passage is not elenctic but maieutic (though the midwife
metaphor is not used here, as it is not in any dialogue prio¡ to lhe Theaetetu.s\.me SoÍa|es
of this passagesees the boy getting the ânswe!'not by leaming it from me' (82b), but by'himself
recove¡ing krowledge froú himself' (86d), which is what Socrafes says of his interlocuto$
i^ lhe Theaereu$i they 'have leâmed nothing ftom me but have themselves discoveted for
themselves' the sought-for truth 050d6-7). I agree \¡rith Myles Burnyeat (see his 'Socratic
Midwifery, Platonic Inspiration', Bulletín oÍ the lßtitute ol Classical Sfudier (Uúiversity of
London),24 (1977),7-1,5) that the midwife metaphor is a Pl¿toDic invention: his argumerit
for this thesis I consider conclusive. I would âlso agree with him t¡at midwifery ånd recoÌlection are distìúct úetâphors which should not be conflated. Even so, they have in commorì the
fundamental notion, exp¡essed in eâch of the two texts I have cited, that the true P¡opositions
which are discovèred in the interrogation of the inte¡locÌrto¡ by SocIates do not come froîl
.Srcr¿¡¿r but f¡om the inte¡locutor ('recollected' by the latter in the M¿ro,'b¡ougbt forth' by
himinthe Theaetetus)-a notion which is not expressed or even hiDted ât in any of the eâtlie¡
dialogues.
ts La- 191e-192bi Euthphr. 5ò,6d-e,1la6-b1; Hp. Ma.287 cff.; Men.12a6 ff, (Though this Ìast
passage occurs in a !¡ansitional dia-Iogue, its pÌâce in that dialogue antecedes the introductioû
ofthe theo¡y of ¡ecollection; paralleling closely the specificatioDs which a correct definiens must
meer
the Euthyphro--aL Men.72c6-d1 ,¡tilh Euthpht. 6d9-e6-thú \s,clearly, a faìtblul reproductioD of the deônitional doct¡ine of the earlier dialogues.)
ii
å
Ë
'q 191e11. same questioîiî Men.'12d1.
Thus when Hippias says 'there is no difference' between'Whât is the beâutiful?' and 'What
is beautìJul?', he ìs rlot represented as propounding an erroneous doctrine which cålls for refutation-only as exhibiting pitiful incapacity lo understand the very meaning of those questions
(287d-e).
'1o
rT
I
å
Ë
I
il
ri
F
42
THE SOCRATIC ELENCHUS
Þ
GREGORY VLASTOS
43
E
&
like the principle of non-contradiction,2l ate never treated as elenctic
theses.22 Only moral truths are so treated.
For'moral' Socrates has no special word But neither does he have any
diffrculty indicating that what he is searching for is truth in the moral
Ë
domain:
Ë
arsument is ove! no chance matter but over what ß the way we ought
to live.'fflne same Phrase, in the identical' o¡ds, in Úg 500c341 (Rep l,
3s2d)
(T4) Of ai inquiries, Callicles, this is the noblest-about those things on which
vou ¡epróached mei what sort of man should one be' and what should one
practiie and' up to wbat point, whetr he is young and when he is old (Grg
487e+884)
(T5) For the things we are disputbg are hardly trivjal, but, as o¡e might say, those
which to come to know ii noblest and not to know most base. For their sum
and substance is just this: knowitg, or not knowif.g, who ß happy and who
ß not. (Grg. 472c-d)
lT3l ou¡
'
answers to the questions put to you. In a cooperative endeavour for mutual
enlightenment this is self-explanatory Not so is the second-the'say what
you believe const¡ai¡t:
Ë
(T8) By the god of friendship, Calliclesl Dor't think that you carÌ Play
Ë
Ë
so we may get somewherc.u (Rep. L,346a)
(T10) If you agree with these things, Crito, watch out that you are ûot doing so
cotrtrary to your reat opinion (zapri ódfav). (C . 49cÀ)
'
'
These a¡e the questions Socrates attacks by the elenctic method, and he
treats them as new questions, never investigated before by the dght
method, so that what th€ wise men of the past have or haven't said about
them becomes a mattet of indifference. When he is talking with you he
wants to know your ans\À/er. If you quote some wise man's answer-as
Polemarchus does in Republic L-he will discuss it as your answer, expecting you to defend it as yours. That you do not yourself have high credentials will not trouble him. He may even count it an advantage. As a partner
in the sea¡ch he welcomes
(T6)
(t?)
...
.
of you l happen to meet at any given tim€
anyon", young or old, citizen or foreigner. (-4p 30a)
. . . anyone
Ë
B.
E
w
2t Thus Soc¡ates never feels thât he hâs to a/8¿¿e that when his intellocutors ¡un into contradiction they sutfer logicâl disaster. The principle of Don-contradiction is never so much as
stated (as it iiin the rdid?le diatogues: ReP 4,436e-4374), to say nothiûg of its bei¡g defended
or justifred.
¿ For the view that the conditions ol â successful defrnition ale not themselves subject to
elenctic argument I am indebted di¡ectly to Paul Wood¡uff See his excellent remarks on the
deoendencã of 'defrnition-testing arguments' otr'key premises suPplied by Socrates hiflself'
wliich'govern the form and cont;nt tdefiûition must have to be acceptable . . On lhich mâtter
Socrat; is an autlìo¡ity. . .' (P/aroi HiPPias Major (Oxford: Blackwell, 1982),131-Ð.
Ë
E
E
I
F
g
Why should Socrates object to iffy theses? Hypothetical premisses
had always been Iegitimate not only in disputation, but even in the most
stringent of all forms of argument as yet discovered in Greece: mathematical prool It is standard in Greek geometry, where indirect proofs
(as for example in EucLid 1. 5) employ an unasserted premiss, prefaced
by the word Protagoras had just used: ðorø,'let this be so'. Zeno, whose
dialectic had become classical by this time-Aristotle calls him 'the inventor of dialectic'5-had practised systematically the thing Socrates
forbids: each of his paradoxes investigates the contradictory consequences
of its counterfactual premiss. Why should Socrâtes ban this modality
of philosophical argument? He doesn't say. I suggest he has three
teasons.
First, to test honesty in argument. In eristic, where the prime object is
win,26 one is f¡ee to say anything that will give one a debating advantage. In elenchus, where the prime object is to sea¡ch for truth, one does
to
F
Ë
I
t_
Ë
ß
F
Ë
å
iilill
To Protagoras, who had just said in reply to Socrates' question,'But what
does it matte¡? Let it be so lor us (ëorø i¡ptt), if you wish', Socrates says:
(T11) I won't have this. Fo¡ it isn't this 'if you wish' aIId'if you think so' that I
waût to be retuted, but you and me. I say 'you and me' for I think that the
thesis is best refuted if you take the 'if out of it. (P/r. 331c)
(,4P 29d)
His is the aggressive outreach, the indiscriminate address to all and sundry
of the street evangelist. If you speak Greek and are willing to talk and
reason, you can be Socrates' partner in searching, with the prospect that
truth undisclosed in countless ages might be discovered here and now, on
this spot, in the next forty minutes, between the two of you.
Fo¡ success in this enterPrise two constraints must be observed. The first
is to refrain from speechifying, to give short, spare, direct' unevasive
games
with me and answef whateve¡ comes to youI head, contrary to your real
opid.on (øøpò. td ôøcoøwø) . . .B (Grg. 500b)
(T9) My good man, doD't arswer cottrary to youl real opinion (øpø òo[cw),
E
ã Cf also what he had said to Callicles ea¡lier at 495a, a d also Euthd. 286d,'Dionysodo¡us,
are you sayi¡g this for the sake of tâlking-to say something ouhageous-or do you really
believe that rlo human being is ignorant?'
z For the same requirement in the å¡gument with Thrasymachus, see also 337c: here ?ô
ôoþúv or tà
qanópevot' in àmryíveø$ar, rò rpatvópet'ot' ¿au?O ¡eplaces the more usuâl
ðoKot1,ra, and Socrates' âppa¡ent willingness to waive the le at 340c1-2 is ironical, as is made
clear by Socrates' reiterating the requi¡ement at 350e5, though here again he rcsìgns hiflself,
as apir dller, to Thrasymachus' sâyitrg he will ignore it.
6 Diogenes Laertius 9.25,^nd29.
t6 Ct the description of eristic rop¿¡d:'prowess inverbalcontest and in the rcfutâtion of wha!
ever is said, regãraUess of v,¡hether it is false or tr\te' (Eulhd.272a-b).
ò
44
THE SOCRATIC ELENCTIUS
GREGORY VLASTOS
not have that option. One must say \¡/hat one believes-that is, what one
thinks true-even if it will lose one the debate.2?
2s
Se¡iousness
Second, to test one's seriousness in the püsuit of truth
an
shamming
grave
voice,
a
face,
solemn
put
on
a
can be feigned. One can
earnestness one does not feel. But if one puts oneself on record as saying
what one believes, one has given one's opinion the weight of one's life'
Since people consider thei¡ opinions more expendable than their life,
Socratãs wãnts them to tie their opinions to their life as a pledge that what
they say is what they m€an.
reason comes from that other dimension of the elenchus to
which I have made no allusion so far. It is highlighted in the APology vthere
Soc¡ates' 'search' is, at the same time, a challenge to his fellows to change
their life, to cease caring for money and leputation and not cadng for what
should be for everyone the most precious thing of all-what one is:
Á further
{T121 . . . and if one of vou savs . . . he does care, I will not let him go ¡ror leave
him, but will queition and examine and refute him; and if he seems to me
not io have thi vùtue he says he has,I shall reproach him for undervaluing the things of greatest vaiue aod overvaluing trivial ones (Ap' 29e)
Socrates is not always so inquisitorial and censorious. But those who know
him best realize that the elenchus does have this existential dimensionthat what it examines is not just propositions but lives. Says Nicias, an old
acquaintance of Socrates, to Lysimachus, a ne\ry one:
lT13) I don't think vou realize that he who comes closest to Soc¡ates in discus-
sion, even if Le should staIt discussing something etse, will f,nd himself
unavoidably carried round and rou¡d in argument ultil he falls irito giving
an account of himself-of the way he is living now and the way he has
liyed iû the past. And when he does, Soc¡ates will notiet.him go until he
has done a tiorough job of sifting him. (la 187eó-188a3)
Thus elenchus has a double objective: to discover how every human
being ought Io live qnd to test that single human being thât is doing the
answiring-to find out if lre is living as one ougbt to Live. This is a t\üo-inone operation. Socrates does not provide for two types of elenchus-a
philosophical one, searching for truth about the good life, and a therapeutic
27
Socrates says to Gorgias: 'I am one of those who woutd gtadly be refùted if what I say is
not true', adding that if Gãrgias iloes not share this sentiment' further debate would be pointless t458a-b).
Ñote túe connection Socrates sees between'saying 14hat you beteve' and seriousness in
argument at TB above, which cofltinuesì 'Nor must you thjnk of me as Playing gam-es For yolt
little sense
seË what the argument is all about-and is the¡e aEythìng about which eaen arDan-of
could be more-selious than about this: what is the way v/e ought to live' (G¡8 500b-c;cf T3
abotr).The sume connection of the rule with seriousness in argument is made at Ãep 1,3494,
and is implied it the remârk to Dionysodo s quoted âbove, n. 23'
I
45
one, searching out the answerer's own life in the hope of bringing him to
the truth. There is one elenchus and it must do both jobs, though one or
the other \ryill be to the fore in different phases of it. From this point of
view, too, the'say \ryhat you believe' rule makes sense. How could Socrates
hope to get you to give, sooner or later, an account of yout life, if he does
not require you to state your personal opinion on the questions under
debate?
This will also explain why on some occasions Socrates is willing to waive
the rule. So, for example, when the interlocutor is losing the debate, sees
disaster ahead, and tries to spare his battered ego further mauling by shifti¡g from combatant to bystarder. This happens to Protagoras shortly after
the passage quoted as T11 above. By this time he has lost two atguments.
At the start of the third this exchange ensues:
(T14) s. Do you believe that one who acts unjustly may act temperately irÌ
so
acti¡g?
pRT. Socrates,I would be ashamed to agree to that. But most people would
a8ree.
s. Shall I
address my argument to them, or to you?
pRr, Argue fiist agaiNt that view of the multitude, if you wish,
s. It makes no diffe¡ence to me, provided you do the answering. For
what I chiefly examine is the p¡oposition. But the coNequence may be
that I the questioner and you the answe¡e¡ will also be examined.
(P¡f. 333b8*c9)
When Protagoras was looking for the same kind of shelter earlier on by
hedging his answer with 'if you wish', Socrates had blocked the move, indicating that Protagoras had already taken a stand and would be held to it:
his ego was now on the line, as in elenchus it must, for otherwise Socrates
would be left with a proposition detached from a person willing to predicate his life on it, and this Socrates would refuse, as in fact he did refuse
at the time. Once that is settled, Socrates is willing to make concessions,
zs a pis aller and tnder protest, so that the argument may go on: Protagoras is allowed to save face by handing his part over to that faceless surrogate 'the multitude'. For the same reason Soqates lets the same thing
happen again, and on a bigger scale, later in the dialogue, where Socrates
directs his argument for the impossibility of acrasia to the same notional
. ans',verer,'the multitude', dragging along Protagoras as a make-believe ally
(352eff.). At the end of that debate we see that Socrates takes the consequence to be that Protagoras has been'examined' after all, compelled to
confess that ftis thesis-not just that of the 'multitude'-has been shown
to be 'impossible' (360e).
46
GREGoRY
THE SOCRATIC ELENCHUS
vlAsros
II
arguBecause it is allowed the waywardness of impromptu debate' elenctic
ment may take any number of different routes But through its motley
variations the following pattern, which I shall callze'standard elenchus" is
predominantly preserved:
(1) The interlocutor asserts a thesis which Socrates considers false and
targets for ref utation.
(2) SoJrates secures ag¡eement to further premisses, say 4 and r (each of
which may stand fãr a conjunct of propositions)'3o The agreement is
(3)
ad hoc: Socrales argues from 4 and r, but not to them'
Socrates then argues, and the interlocutor agrees' that q and
not-p.
(4) Theieupon Soüates claims thar not-p
has been proved true,
r entail
p false'
The main alternative to this pattern i.s'indirect' elenchus-so called3l
because here the refutand may be used as a premiss in its own refutation,
hence Socrates is not himself committed to the truth of the whole of the
premiss-set from which he deduces the negation of the thesis All he could
reasonably claim to accomplish by this means is to expose contradiction
within thá interlocutor's premiss-set. To estabüsh the falsehood of the
32
Here, and here only, we conthesis he must türn to standard elenchus
Because it is Socrates'marD instrument of PhilosoPhical reseârch With a si¡gle major
con'
exceDtioû lP¡t. 352d-3584), jt is lhe only torm of argument he uses not merely to exPose
iruãiiiion ìn ¡it oppooenis beliefs. bui to establish sr¡bslântive doctrines of his own' such as
i-fr"l"ü"*¡"g (,h"jirt is nol. tDeant to be exhaustìve): that the just man will oot harm enemies
l,
iirr. i. ¡sslÏ,h"t rlt" ¡ust ruler fltes not in his own interes! but i¡ lhat of his subjects (R¿p
àìei-S¿Zalr'i¡ut ¡usd;, is more profitâble than injustice (Ãep 1, 347e-354â); that to teach meû
2e
them just (Gr'8 4604-<)ithat it is bel(er to sufter w¡ong lhan-to
to suffer desËwed pu-nishmeni than escape it 1(he great argument lo be
àiscusseO Uelow); r¡at the good and th; Pleâsant åre not the sÂme and pleasure should be
r"r tt" íát" of the ãood. not vice versa (Grg 494€-5Q04); rhât in malters o[justice we
""'"""ã
!¡ooiá toito* not',t".ani' but'the man who knóws' (c¡i. 474-484)ì thal the poel versifies
justice, tem'
anà the ¡haosode recites bv a kind of mad¡ess. not by craf! (/o/,): lhat piety and
god'loved
because
pious
âction
is
(P¡¿
329e-333b);that
ìnierentailing
o"r"n". un¿^*itdo-
i"J."
ir ¡ár, faj¡o to make
io.trlii ir'"o¿'t.u"t
".e
i! is godloved (EuthPhLgd-17à\'
it is pious, not pious because
b'I use two'va¡iableg though oìe would iuffrc¿, with a view to the special-case, to be disjust one of the agreed-upon
cussed below, where the interlõcutor hâs the oPtion of weÌshitg on
oremisses,
'-ii i øuo* notinton's
his discrlssion
terminology (Pl¿¡ob Earliet Dialectic,22ff)'i}no¡ing those thirgs in
argument i¡ this paper) {'ith which
of indirect' elenchuslirielevant to my main
I disagee.
¡t S-ee e.g. how indirect
elenchus is used to rough up Polemarchus, discrediting-io.his eyes the
ofjustice as'renderiDg to each what is due'(R€P 1,331e) by shoìvi¡g
definition
ultra-respeËtable
justice
ttrat wtre^n;oilea to ottrer admissions of his it has bizarre conseq'rences (that'in all things
47
front what I called earlier on'the problem of the Soc¡atic elenchus'. Here
there can be no question of taking the aim of the elenchus to be only to
show the interlocutor that lr¿ must consider his thesis false if he chooses
to stick by his further admissions. So personal and contingent an outcome
would not begin to sâtisfy Socrates'drive for universally valid results:
(T15) For I think that we should all be contentiously eager to come to know what
is t¡ue aûd what is false about the things we assert;fol it is a common good
for all that this should be made evident. (Grg 505e)
(T16) Or don't you think that it is a commoû good for practically all mankind
that how each thing is should be made eridenll (Chrm.l66d)
What Socrates must do-and whât, I shall be arguing, he is convinced he
does-is to prove p not just false /or tfte interlocutor,btl lals¿.If he cannot
hope to do this by standard elenchus, he cannot hope to do it at all
My discussion of standard elenchus will focus on points (2) and ( ).This
is how I can best set in historiographic perspective the interPretation of
the elenchus I am defending in this paper.At these two points I am departing sharply from each of the leading lines of past interpretation, represented respectively in the two wo¡ks of nineteenth-century scholarship
which are the landmark studies in the field: Edua¡d Zeller, Phtlosophie
der Griechen,3s and George Grote, Plato and. the Other Compønions of
SoÜates.3a I am going against Zeller at point (2), against Grote at point
(4). Since Grote's Socrates is incomparably more interesting than Zelle¡'s
and, in my view, much closer to the truth, I shall have more to say of the
relation of my Socrates to Grote's than to Zeller's.
The claim I am making at point (2) is that the premisses [4, r] from which
Socrates deduces thât negation of the opponent's thesis are logically unsecured within the argument: no reason has been given to compel agreement
to them. Socrates does, of course, have reasons for q and for r. But he does
not bring them into the argument. He asks the interlocutot if he agrees,
and if he gets agreement he goes on from there. So in elenctic argument
the question of referring to some court of last appeal for settling
philosophical disagreement does not arise. In particular there is no appeal
ir thet use, useful when they are useless' (333d), that the just man is a kind of thief'
(334a-b) ), keeping staDdard elenchus i¡ rese¡ve until it is Deeded to establish the Sosatic thesis
that the just man will not harm enemies (335b-'c). Similarly, Euthyphro's tust deûnitìon is
attacked by indirect elenchus (6e-8a), and standa¡d elenchus is then brought iD to p¡ove the
doct¡i¡e, so fundamentål for Socrates' rational theology, thât pious action is Sodloved ¿¿cdure
is useless
it is pious (9d-11a).
rr The volume entitled Sokrotes und d.ie Sokral¡k¿¡,5th edn, (Leipzig,lgz2i rcv.1963). My
references will be to the Eng. trans. of the 3rd Ge¡man edn,, by O. f Reichel (¡epr. New York:
Russell
& Russell. 1962).
48
GREGoRY
THE SOCRATIC
vlAsros
to what Aristotle takes to seNe this purpose: none, on one hand, to those
self-evident t.ruths which are for Aristotle the foundation of all demonstrative argument;35 and none, on the othêr, to what he calls rà ivðofawhat is woithy of belief because it is believed'by all or by most or by the
wise'-which constitutes for Aristotle the foundation of dialectical argument.36 Socrates spurns both. He never tells the interlocutor that be must
grant 4 or r because they are self-evident truths nor yet because they are
¡he m;st generally accepted opinion on the topic. To self-evidence there is
no upp"uiut ,ll by anyone in Plato's earlier dialogues.3T To common belief,
ELENCHUS
At this point the Soüates of Plato's earlier
49
dialogues is at loggerheads
rvith Xenophon's:
(T19) Whenever Soc¡ates himself argued something out he proceeded ftom the
most gene¡ally accepted opitrions (òn rtu póJ.øm öpd.oyoÐpévuv),
believing that secudty ir a¡gument lies therein. Accordingly, wheneve¡ he
argued he got much g¡eater assent ftom his hea¡ers than anyone I have
eve¡ known. And he said that Homer made Odysseus'the safe speaker'
lod,8.lTLj because he was able to conduct his arguments ftom lyhat is
belíeved by mankind (òrà únt ðorcoortav toq <àúptirø6).1e (Mem. 4.6,1.5)
therãìs, but not by Socrates. It is Polus who apPeals to it in the GorSías'
only to find Socrates rejecting it out of hand. When Polus says
If we were to believe this,' hat would we do Ìvith those doct¡ines of the
lT17) Soqates. don't vou think you've been refuted already when you say things
with which no'one woulà agree? Just ask any of these People here '
better to suffer injustice than to commit it; \¡/hat one should neve¡ harm
one's enemy, never return evil for evil;that happiness is not the reward for
virtue, but virtue itself; that virtue and knowledge are the same so that to
known the good ând fâil to do it is impossible? And what would we do with
the profession of ignorance?4o If Socrates had 'proceeded from the most
generally accepted opinions whenever he argued something out' he could
(Grg.4'13e)
Socrates stands on his previously expressed conviction that the only
opinion which matters in an argument is that of the arguers themselves:
(T18)
If I catnot
p¡oduce one
man-yourself-to witness to-my
asse¡tions,
I
shall have accomplished nothing.. Neitber will you, I
beìieve, if this one man-myself-does not witness for you,lettiDg allthose
believe that
I
other people
go.33
(Gt1 4?2ba)
Demonstfttion (rìn¿ôE¿È.ç) proceeds from premisses which are'Primary' (xp&at)' ie'
à
induce lo¡viction 'th¡àugh themsàlves and not tìroughsome other things ltà þù ó¿ hépuv ò
ót aúrùv lvovrø ritv ,ríor;vl Ïop. l1}"21-b2i cL A Po.64h34fn.' Phvs l9i"+6)
ró The pienrisses of diaiectical reasoning are'those which âre believed by all or by most or
by the wiie aad, of these,by all o¡ by most-or the most distinguished aDd most rePutable'(Top
35
roo29-b23).
37
Normân Gulley appea¡s to hold that there is imp¿icit aPpeal to self-evidencei'while the
initial logical aim oi thËsocratic ¿lencrur is to ¡eveal a cont¡adiction in a respondent's views,
its lurthe'r ajm is to es!âblish as lbe conlradictory of the resPondent's initial thesis â proPosition Dresented âs so obviouslv true that the respondent is driven to âbandon his lhesis' (1'¡¿
Philosoohv of Soc¡ates (Londón: Macmillar,1968),43-4) ff this statement is taken at face valùe'
it is surålyialse: ttr us in the Gol8i¿r in each of soqates' major argumeûts agâinst.Gorgias, Polus,
and Calliâes, the conttadictory ãf tì€ir thesis is a paradox-not at all hkely to strikelhe sponso¡
of the ¡efutand as'obviously true', no matter how it is'p¿eseûted' to him Thus in Socrâtes'
to suffer
second a¡gument against Polus the contrådictory of Polus'- thesis is that it is better
;n¡usrice rña¡r to coñrm¡t it arid better !o submit !o punishment thaû lo escape i(;that Socrates
stiould exDect that his arqumen! would make this proposition look'obviously true to Polus is
iL
tft" f.rstiafter Socrates bas 'provãd' it, its immediate consequences (480a-d)
".i;t.r"""u as out rageo us' (rÌroza; 480e 1). If we revise the ctaim { as we probably should
still st rík e Polús
i¡ the light of other thi¡gí Gu ey says on pp 42-3), taking it to mean that when.the interlocutor is co;fronted with th-e inconsistency oihis thesis with lhe agreed-upon Premisses the lâtter
*ui[ ,t ik" hi. u, aot"'obviously truã' than the forme¡, the claim, thougb mo¡e Plausible, is
stiii unacceptaute: ttre obviousness oI one o¡ more of the premisses is-irrelevant to the logic of
the â¡sume;. Soqates does not argue, 'concede that youf thesis is fâ-lse becâuse the agreed-
ít"ài"r.t u," .ore "obviousli true" than it'; even if those premisses Ìte¡¿ obvious, thei¡
ob\¡ousness would not be a Premiss in the argument.
3i For Socrates' rejection ôf the appeal to commoD opiûion, see also ¿a l84e, Cri 46d-47d'
uptn
Socrates of Plato's earLier dialogues that go dead against'the most generally accepted opinions' of his time and shock its common sense: that it is
$ Since what Xenophon is calling here'the most generally accepted opinions' and'what is
believed by mankind' is the very thing Aristotle would call rd ð'óo€ø, Soüates'philosophical
method, as uDderstood by Xenophol, would be congruentwith the method which Aristotle coít¡¿rts with the Socratic.Hesays that the accomplished dialectician will a¡gue not only (¿)'pelrastically', keeping within the questioner\ tole iû argurnent, content to do ¡lo ñore thafl 'exact an
accounl' Q,óyov )t aBeîv) ftom his interlocutor, without expouûding aDd defendi¡g positive views
of his own, but also (å) 'as one $,ho knows' (óç eló¿ç), going into the answerer's role,'renderi[g aÍ\ accoror' ().óyo' úr¿xo'reç) to 'defend a thesis (pú,átapeu rì¡v Aéotv) f¡om the most generally accepted views' (ðl èt'òoÇotárot') (SE 183b2-6: he is ¡ecalling the distiûction he had
established between ó¡a,tr ¿ Krt Koí ând n¿eaartKoí îtólo¿ in the opening pa¡agraphs of the t¡eatise,
165'3-6). It is at just this poinl that he makes his famous ¡efe¡ence to Soc¡ates:'aDd this is why
[i.e. because he adhered to ¡ole (d)lSocrates put questions and gave no ¡eplies;fo¡ he confessed
tìat he had no knowledge' (SE. 183b?-8). Thus Xenophon's Soüates is at logge¡heads u,ith Aristotle's no less than with Plâto's. Xenophon's Socrates would be doing regularly (ånd does fte"
qtrenrly írthe M¿morcbt¿r¿) wbatAristotle does not represent Soc¡ates as doing at all. (Aristotle
need notbe understood as denying to Socrates the willi¡gness, often attested h Plâto (P¡r.336c,
338c-Å; Grg.449b,462a-b, 504c), to 'reply', no less thaû'âsk'. He is content to ìgnore this in tbe
plesent, âll too brief, allusioD to Socmtes, probably because he had failed to understånd (as have
so many othe$ after him) that Soc¡ates could virdicÂte positive theses of his owrl in adve$ary
argument pitåo¡¿¡ welshing on his confession of ignoranca ar,d. withoul i¡|oking endoxa
a
^s
coult of lâst resort.)
¡0 I am ûot suggesting that there is some obviously true interpretatioD of this extremely perplexing feature of Socrates' thought (which I reserve fo¡ separate discussion in a late¡ pape¡).
ÍEditols notetThis la|er pape¡ appears in this volume as Ch. IL] But it is certainly Socratic,
attested rìot only by Plato a¡d Aristotle but by many other qeditûble witnesses-in fact by all
of ou¡ ancient autho¡ities who address the questioû. Heûce no irìterpætâtion of the elenchus
wiÌI stard if it cantrot be sustained in full consistency with his profession of ignoftnce. I count
it a me¡it of the present inte¡pretation that it is both logically and textually independeDt oI my
particular iûterpretation of the profession of ignomDce,but is none the less perlectly consistent
with the latte¡-indeed, provides the foundation for it.
50
not have argued for those paradoxes,al nor could he have professed to
know nothing-as he, in fact, does not in Xenophon: in Xenophon's
Socrates there is no profession of ignorance, no interdict on harrning the
enemy,az and the identity of virtue with happiness and of knowledge with
vi¡tue are blurred and flattened out-the paradoxes become commonplaces. So at point (2) of my account of the elenchus the conflict between
Þhto's testimony and Xenophon's is ufilegotiable, and the gravest fault
in ZelLer, gteat historian though he was, is that he fudged an issueal
which callei for a flrm decision for one of his two major sources against
the other. He thereby bequeathed to the historians that followed himmost recently Guthriea-an impossible reconciliation of irreconcilable
data.To accept, as Zeller did, Xenophon's description of Soüates' method
of argument inT19 above,a5 is fatal to the elenchus.And so we see in Zeller'
aó
and iow again in Guthrie, the elenchus disappear without a trace It is
not argued out of their account of Soqâtes' philosophical rnethod lt iust
drops out.
4r Thât Socrates uses endoxic plemisses fo¡ âll they a¡e \l'olth should go without saying- But
without some contra-endoxic prámisses how could he hope to get contú-endoxic-corìclusions?
by staridard elenchus for a thesis (cf above' n 32) which goes
Consider R¿p. 1, 335b-c, arguing
-morãüiy
where doing evil to one s eDemies is on a-Par,with doing
against the grain of Greek
gãodto one-'s friends ('belongs to the same characÍer'(rot' aùrcú íiûouç;fop 113"3-4). says Arist-otle, the same ødmi)abte (ineuøvç;113"13) character); ând doing good to.ftieûds ìs a sta¡
ei.làxon (Top.1o4'22).The argument starts wilh e/¡doÌ4-when horses (or dogs) are hârmed
thev are made worse ir¡ resoec-i of equine (or canine) excellence-ând theE moves by analogy
Lo tire o¡emiss that when men are harmed tÅey afe made worse in resPect ol their'human excel'
lence'i-which would be as counter-intuilive for the Greeks as i! would be for us (does stealing
a man's v/allet make him'worse in respect of human excelleíce'?) Equally contra-endoxic is
the next premiss, that'justice ìs hùma; excellence', iJ uÂderstood to mean (as it must for the
o*oo""rìf th" uteomãnt) lhat iustice is not merely ø. but ¡å¿ buman excellence
a'¿'
(Mem 2'
Quite the opp-osite: Íenopñon's Socrates explicitly endo¡ses the tr'rdilioûal.view
'
6.3s).
'By
a]
a curious sort of inadvertence: nowhere irl hrs book is there a sign of his even having
nol;c¿á this conflict, nol yet the conflict between A¡istotle and Xe¡ophol on the same issue
(above, n.39).
'4
K. c. cuthrie,,4 Histotv oÍGreek PhilosoPhv'ä:Th¿ FiÍth centuty L'nli+htenñent,and
UnÚeßity Press, 1969. 1975) Certiûed by
ihe Câmbridge imprint, this is now thé standard ¡èferenêe work througbout the world for all
non-soecialisis and. ofteí enouqh, for specialists as well
t5 2eller quotes it without ãissent iDd takes Socrates' PhilosoPhical melhod tô consist
iû 'de¿lucing'cotrceptions from the common opinions of men'(Sokl¿fes und die Sokrctiker,
iv
THE SOCRATIC
GREGORY VLASTOS
w
PIøto: The Eartiet Pe¡io¿ (Cambridge: Cambridge
127¿).
6 Ño discussion whatever of elenchùs by Zeller, ibid, ch lv ('The Philosophical Method of
Socrates') or Guthrie,,4 Hi.stoly of Greek Philosophy,äi 425-9'^îd iv' P¿ssim The disaPPearance is mo¡e surprising in Guih¡ie since he, unlike Zeller, takes Plato to be 'the chiet, and
XenoDhon onlv a'n auxiliarv sou¡ce of knowledge of Socrates as a PhilosoPher' (üi 350) and
since. moreover, the two m;jor studies of Socrates'úethod of enquiry that had âppea¡ed iD
English during ôuthrie's li[e-ti¡ne-Robinson's in 1941 and 1951, Gu[ey's in 1968 (nine yeârs
befäre the puúlication of vol. iv of Gulhde's ¡l¡rloly) hâd put the elenchus at the cantre of their
interpretation of Socrates,
ELENCHUS
51
III
No.,v for point (4) in my analysis of standard elenchus: the most novel of
my proposals.4? I must begin with the position I had reached in earlier
Socralic studies-the extreme opposite of the view I wish to defend now.
Explicitly in that brief introduction to the Protagoras of 1956 to which
I alluded aboveas and implicitly in an essay'The Paradox of Socrates',
written a¡ound the same time,4e I had maintained that Socrates never
meant to go beyond (3) in his elenctic arguments-that their object
was simply to reveal to his ûìterlocutors muddles and inconsistencies
within themselves, jarring their adherence to some confldent dogma by
bringing to their awareness its collision with other, no less confident,
presumptions of theirs.so This interpretation had a mighty precedent in
the work of that great Victorian student of Greek antiquit¡ whose multi
volume Hßtory of Greece (1851) and three-volume Plato (1865) arc, in
my opinion, still, all in all, the flnest contributions ever made in any language to their respective themes. Unlike Zeller, Grote saw with the utmost
clarity how central was the elenchus to Socratic enquiry as depicted
in Plato's earlier dialogues, hoìv centrâl it had to remain in our picture of
a7
My inte¡pretation ofstandard elelchus, taken as a whole, and applied rigorously, conceived
only rationâl support Soc¡ates offers his úoral doctrires, has no clea¡ precedent i¡ the
scholarly lite¡ature, to my knowledge.Its afnnities are with view6 like those of NoIman Gulley
(The Philosophy ol Soctates,31ff) ard lerence I¡wlrr (P\øto's Motal Theory (Oxford: Oxford
University Press, 1977), 37 Í1.; Plato: Gorgias (Oxford: Oxford University Press, 1979),parsln),
which also recognize that the elenchus has positive, ro less than negative, thrust, aiming to b ng
argltlnentative suppo¡t to Soqates' afn¡mative views. One of my differences with Gulley has
beenûoticed above.Mymain difference with IrwiD arises over his view that'Dot all lofSocrates']
positive doct¡i¡es rely on the elenchus; some rely on the a¡alogy between vktue and craft'
(Plato's Morql Theoty,37).I see no sound reason fo¡ putting this analogy outside of the
elenchus;all of the arguments which draw conclusions frorn that aûalogy are purc elenctic argumeDts.Afurther disagreement a¡bes ove! alleged constrai¡ts whìch, acco¡diDg to IrwiD,Socrates
'no¡mally' imposes or what the interlocútor can or can't say in arguiDg against him ('normally
the i¡terlocutor is rlot allowed this fteedom [to ¡eject counter-examples which refute his
defrnitior\]', Ploto's Moral Theoty,39).me l€xt|ual evidence appears to be that Socrâtes ¿rn¿yr
âllows-indeed requires-the interlocuto¡ to say ånything he believes, if he beÌieves it.
ar n.8.
ae
Queen's Quanerly,64 (1957),49Ç516rcF. in G Vlastos (ed.), Tåe Philosopht oÍ Socrates
(New York: Doubleday, 1971).
50
This is Plato's picture of'the sophist of noble lineage'in Spl¡, 230a-e, whose se¡vice to his
interlocutors is simply theÌapy ('purgatior') by enhaûced self-knowledger their dogmatism is batte¡ed as they âre made awa¡e of conlicts within their own system of beliefs (the identincatior
with Socrates is tightened up in the back-reference to tìe passage at 268b-c'\¡r'e set him doì,vn as
having no knowledge');it should be noted thatnothi¡g is said to make this 'pùgative' art a subdivisioÍ of eristic,pace Lewis Campbe ('fåe Sophistes drd Politicus olP¡¿to',191) and others:
see F. M. Co¡Dford, Plato\ Theory ol Knowledge (Londor: Routledge, 1935), 177-82. No one
would doùbt that this is an authentic, if på¡tial, rep¡esentation of Socratesr this is the Socrâtes
who destroys the conceit ofwisdom (,4p.21b-23b). ButPlato never says this is âll Socrâtes was,
as he wouìd have been,if the account of the elenchus I have given at that time were co¡rect.
as the
q
É
52
GREGoRY vLASTos
F
if we were to put any faith at all in Plato's testimony' and how
valuableacontdbutiofltohumanthoughtitsrelentlesspolemicagainst
Grote
¿ãgÃãìirÀ -oof¿ be, even if it were only the negative instrument
B
THE socRATrc
a
that picture of Socrates in my earlier work was very.much
founits
and
minãrity view, itãid not put me in bad intellectual company'
ili.; it Plato's depictiàn of elenctic argument in his earlier
.dialogues
that his
,""Á"¿ .""nt"t ttow could Plato be telling us, I used to ask myself'
to his interlocutors thât thet theses were
Socrates undertakes to prove
-he
shows Socrates doing is proving the inconsisfalse and his true, if all
inte ,ocutors' theses with other, unargued-for' concessions of
iãtã
n,tt tt"lttt"r was that pictüe trouble-free lt left me with this quesii"ÉZ"ïftit
of
tion: if that were all Socratei had expected of the elenchus-exposure
support
inconsistencies in his interlocutors-where did he flnd positive
his life? If
i;; À;;; strong doctrines of his on whose truth he based no
rational
the elenchus, hii only line of argument, gave those doctrines
question
troubled by that
á.á""¿itg, *ttut ¿idf Grote hãd not been
own positive conSocrates'
ihat
believe
possible
to
it
õ""uor" ñ" found
tracks:
victions and his criticàl assaults on those of others ran on separate
are two
'the negative cross-examination and the affìrmative dogmatism
involve'
or.
to'
lead
not
does
one
the
thought;
of
oi"onnË"t"¿ operations
myself
reconcile
not
could
l
not
I
could
zgzr'
(Plato,i.
o. tJiy trt" otit"t'
himself'
a
dogmatist
was
who
life,
examined
the
to Grote's missionary of
And there were ceÍainly textual grounds too for uneasiness \vith the
pì"tor", ur critics started pointing out.52 So I began to lose my enthusiasm
ioiìi.not it it on" thing tó becomi disgruntled with a picture, quite ânother
that destroys
to liberate oneselîfroà it by discovering the textual evidence
they did
critics
because
my
get
ftom
not
could
i
itr" pi"ture. Îtat
"vidence
for
mysell
it
to
discover
not iave it themselves. I had
E
Ë
I
E.
(droôÉðer'rra¿)
(p)
n
(not-p) To suffer injustice is better than to commit it.
Ë
5t See
eso. Pla¡o.i 23Ç17 and
B
I
E
Ê
Ë
E
ã
a
flock of
To commit injustice is baser (o.ioytov) than to suffer it,
while all the rest can be bundled up in a single gather-all conjunct r, whose
contents need not concern us. Socrates argues, and Polus agrees, that q and
r entail not-p.When this result is reached, Socrates tells Polus, in the words
I just cited, that the Socratic thesis, not-p, has been 'proved' true, that is,
that p has been proved not just inconsistent with 4 and r, as it of course
has, but proved false.sa
Why had Socrates' perfectly clear-cut words in T20 been ignored?
Why had I ignored them myself? Because I had scaled thern down, even
while reading them, discounting them as a careless overstatement. I
would not have done so if I had noticed that it is not only here, in the
last gasps of the debate, that Socrates claims he can prove not-p true:
he makes the very same claim in different words several pages back, at
the staft of the debate. Recall what he had told Polus in T18 above: 'If I
cannot produce one man-yourself-to witness to my assertions, I believe
that I shall have accornplished nothing.. .'. Conceding that 'almost all
men would agree with you. Athenians and foreigners' (472a), he had
declared
E
B
Ë
281f|;and his rema¡ks on the elenchus in his HistorY of Grcece'
ix
Attacking in standard elenctic fashion, he gets Polus to agree to
further premisses, only one of which need be recalled he¡e:
(q)
É
thal what was asserted [by myself]
^uny
To commit injustice is better than to suffe¡ it.
Soc¡ates defends what he takes to be the logical contradictor¡
E
is t¡ue? (Grs.419e)
and I
Here Soqâtes says in so many words that he has done what Grote
he
has
that
says
he
had maintained he never did in an elenctic argument:
;froved' his thesis true. Grote had certainly gone over that senteice in the
times, and so had I, and so had scores of others But it had
ãorgi^
His thesis is
F
The c¡ucial text is
lT20) Has it not been proved
53
not hit anyone between the eyes.s' Let us see ',vhat it means when read in
its own context.
The argument starts half a dozen Stephanus pages back, where Socrates
presses the question: if one were forced to choose between inflicting injustice on another person and suffering it oneself, which would be the better
choice-the better for onesell the chooser? Polus takes the first option.
Socrates
it.51
thought
'--i""*hil"
ELENCHUS
85
àdn. (Londoû:Dent. t906),
(oxlord:
-Eueryman
!;'rilîìr"ãi".ì
tharpest objeition was râised by E R Dodds (Plaro: Gorgias
i;Lr., iô5e), 16 ana n r) s"e also Gullef TÈe P'¡tilosoPhv ol socrates'68ff
o.r"iä uîi"J¡iy "t¿
$ There is no comment on it in Dodds's ot lrwin's commefltarf afld no refetence to it in
Gulley, though it should have been a sta¡ text for all three.
5a
Socrates' other tn'o desc¡iptioris of tle result (Grg. 4'79cA-5,480b2-5) go no further than
poi¡ting out the demonst¡âted iDconsiste[cy between Polus'thesis ând the premisses to which
he has agreed. (The reâder should bear in mirld thât throughout thís pâper I set aside åll questions relatiDg to the logical validity of the argnments by which Soctates undertakes to refute
his oppo¡eûts'theses in specifrc elenchi. For thjs whole aspect ofSoc¡atic dialectic I would refe¡
to G. Santas, Soc/¿¡es (London: Routledge, 1979), with whose detailed analyses of Socratic ålgumeDts
$
I ûnd myself in very substantial agreemelt,)
I
t
I
54
i
GREGoRY
vl,Asros
not comPel ¿1¿, but produce
a multitude offalse witnesses against me,trying to drive me from my property, the truth (é¡<tliç oúoíaç raì øú à14Ûoõç) (Grg' 472b)
(T21) But I, a single man, do not agree; fot you
'
d.o
How do you'compel' yow adversary to affirm what he denies? In an argument your only means of compulsion âre logical. So to compel Polus to
'witness' for not-p Socrøtes would have to Sive Polus a loqícally comPellinq
proof that p ís /a/se. Thus âheady in 4'72b, seven Stephanus pages before
àsserting at T20 above that he has 'proved' his thesis true, Socrates is
announcing that this is exactly what he is going to do. So when he says at
T20 that he has done it, this can't be a slip. Thus, pøce Grote, ex-\4astos,
and who knows how many others, there can be no doubt that this long
argument against Polus, elenchus in its standard form' which in point of
logic has done no more than demonstrate inconsistency within the premissset {p, q, r}, Socrates takes to prove thatP is false, not-p true And in one
of his statements we see him claiming by implication that he can do the
same in all of his elenctic arguments against all comers:
l
I
i
i
I
I
i
I
r
&
I
t
Ê
B
The claim he makes here is perfectly general wheneve¡ he is arguing elencti-
F
essay,'/rø problem of the Socratic elenchus': how is it thât Socrates claims
to have proved a thesis false when, in point of logic, all he has proved in
any given argument is that the thesis is inconsistent with the conjunction
of agreed-upon premisses for which no reason has been given in that argument? Could he be blind to the fact that logic does not warrant the claim?
Let me frame the question in the terms of the metaphor that runs through
the passage: compelling a witness to testify against himsell Suppose the
following were to happen: a witness gives tèstimony p on his own initiative and then, under prodding ftom the Prosecuting attorney, concedes q
and r, whereupon the attorney points out to him that ø and r entail nolp,
and the witness agrees that they do. Has he then been compelled to testify
that p is false? He has not. Confronted with the conflict in his testirnony,
it is still up to him to decide which of the conflicting statements he wants
to retract. So Polus, if he had had his wits about him, might have retorted:
'I see the inconsistency in what I have conceded, and I must do something
to clean up the mess. But I don't have to do it your way. I don't have to
concede that p is false. I have other options. For example, I could decide
THE SOCRATIC ELENCHUS
I
(T22) But I know holY to Produce ote witness to my assertions: the man against
whom I am arytif'g. (Grg.474a5-6)
calty with anyone against any thesis in the domain of morals' Socrates
'knows' how to make his adversary 'witness' to its contradictory: that is
how to prove to him by elenctic argument that his thesis is false.
This brings us srnack up against what I had called, near the start of this
I
I
É
B
Ë
Ë
É
F
F'
å
55
p is true and 4 is false. Nothing you have proved denies me this
alternative.'
And why shouldn't Polus in that crunch decide to throw q instead ofp
to the lions? How strongly he believes inp we have already seen: he thinks
it absüd of Socrates to deny it when almost all the world affr¡ms it.55 For
q, on the other hand, he has no enthusiasm. He may have conceded it only,
as Callicles obse¡ves later,'because he had been ashamed to say ¡tr'hat was
in his mind' (Grs.482eZ). Why shouldn't Polus then jettison ø with the
feeling'good riddance'? He would then have come out of the elenchus
believing that doing injustice is better andnobler lhansuffering it;his latter
state',vould have been worse than the first. Couldn't this always happen?
Wlenever Socrates proved to his interlocutors that the premisses they
had conceded entailed the negation of their thesis, why couldn't they hang
on to their thesis by welshing on one of the conceded premisses? Surely
Socrates would be aware of this ever-present possibility. Why then is he
not worried by it? Because, I submit, he believes that if that wrong choice
were made he would have the resources with which to recoup the loss in
a further elenchus. This, I am suggesting, is his general view: if you disappointed him by denying q instead ofp, he would be confident that he could
start all over again anil find othe¡ premisses inside your belief system to
show you that you haven't got rid of the trouble-that if you keep p, it will
go on making trouble for you, conflicting as much with these other premisses as it did with ø and / before. Can it be shown from the text that
Socrates has this conûdence? I want to argue that it can.
For a start let us observe what happens iî the Gorgiasi the¡e Callicles
enacts the part Polus might have played if he had chosen to retract q
instead of p. Polus was worsted in the argument, says Callicles, only
because he had conceded q, which he should not have done, and would
not have done if he had been less squeamish: if he had had the fortitude
that
to admit that to do injustice is nobler than to suffer it, he would have
escaped unscathed (482d7-eZ).Is Socrates stymied when that happens?
Not on your life. He sheds no tears over the loss of q, He extracts a ner¡/
premiss-set from Callicles and, sure enough, this new set contains all the
premisses Socrates needs to deduce not-p all over again. But what if a
super-Callicles should arise to repudiate those new premisses from which
Socrates derived not-p? Is there evidence that Socrates would not be fazed
Ë
even then, or even by a super-super-Callicles after him? Does our text
attest Socrates' belief that no flesh-and-blood antagonist will ever turn up
without always carrying along, in his own system of belief, a baggage of
Ë
5 Above,T17.
å
56
GREGoRY
vlAsros
premisses from which he can be 'compelled' to 'testify' against P? I want
io argue that it does so in two remarks which, taken in conjunction, yield
clear evidence that Socrates believes this very thing.
The first is one of the things he says to Polus before their argument
begins:
(T23)
F
&
, . . I believe that I and you and the ¡est of mankind believe that committing injustice is worse thâIl suffe rÌgrt.. (Grg 414b)
... don't be astonished that I should say these things lwhich he has
been upholding against Polus]. My love, philosophy, is the one you
must stop from asserting them. It is she, my friend, who asserts these
things you hear fiom me, afld she is much less unstable thar are other
loves. Fo¡ the love of Callicles says now one thing now another, while
philosophy always says the same thing
å
This terminology was suggeste¿l to me by David Gà\thier' Atte¡nativelf we might sPeak
Þ
of 'explicit' and'tâcit' belief .
57
ù. lon 539e: to lon, who had just said that it pe ai¡s to the rhapsode's art to judge dll Passages in Homer, includi¡g those which depict the wo¡k of diffe¡ent craftsmen, Socrâtes replies,
'surely yon do not say 'ialÌ", Ion (oìt ou yc Eiç, ãt "lory ónø'ra) Or âre yoù so torgettul?'
The ref;¡ence is to 538b, whe¡e Ion had conceded that the ¡hapsode's å¡t is ditferent from the
cha¡ioteer's and, further, thât if it is a dilferent art, it is knowledge of difÍe¡ent things. Taking
âdmissioris q alld ¡ to entail not-p, SocrÂtes feels entitled to tell Ion at 539e that he'says' not_
p,in the fac¿ of the îâct that Ion saysp.He does the same thing to CaÌlicles by ProÌepsis at G,'8
495e: anticipatiúg that Callicles will make adrnissions which entail the negation of the identity
of the plea;nt with the good (on which Callictes is 'insisting'; 495bE), SocratesJeels entilled to
dechr; that'Callicles do;s not aglee with lthat identity] when he shall take the right vieù' of
himself .
57
Now consider whai Socrates says to Callicles in part (á) of the little
speech that forms the curtain-raiser to their debâte:
(T24) (ø)
What in the world could Socrates mean by saying that Polus and the multitudes who agree with Polus 'believe' the oPposite of what they assert?
There is one-and, so far as I can seen, only one-way of making sense of
the remark:we must understand Socrates to be using'believe' in that marginal sense of the word in which we may all be said to 'believe' innumet;ble things that have never entered our heads büt are none fheless entaíled'
by what we believe in the common or garden use of the word. Let ¡ne call
the latter 'overt', the former 'covert', beliels6 Thus, if I believe overtly that
Mary is John's sister and that John is Bill's grandfather, I may be said to
believe covertly that Mary is Bill's great-aunt, even if I have never thought
of that fact-indeed, even if I do not have a word fol 'great-aunt' in my
vocabulary Or, to take a less trivial example, if I believe that a given flgüe
is a Euclidean triangle, then I believe covertly the proposition which is so
surprising when we first learn it irì geometry, that the f.gure's interior
angles sum to two right angles. Here then is something Socrates might wish
to express by saying that Polus & Co.'believe' not-p, even while they ins.ist
that p is what they do believe-namely, that they have certain beliefs (of
the ordinary, overt sort) which entail not-p. This gives us a lucid sense for
what Socrates might be saying in our text:he is not declining to take Polus
& Co. at their \ryord when they insist that P is whât they believe; taking
their word for this he is telling them that along with their (overt) belief in
p they have also certain other (overt) beliefs which entail not-P: in this
sense they do (covertly) believe not-p.s?
56
THE SOCRATIC ELENCHUS
F
(å) So you must either refute her saying those very things that I was
saying-that to commitinjustice a¡d to do it with impunity is the g¡eat-
evils-or, if you leave this unrefuted, ther, by the dog, god of
Egypq Call¡cles will disagree with you, Callicles, aîd will díssent ftom
you your whole life long. (Grg. 482a-b)
est of
B
F
R
þ
å
E
Ì
I
E
þ
Ë
&
ß
'What
could Socrates mean by saying that if Callicles cannot refute the
Socratic thesis then, in spite of his scomtul rejection of it, it will remain i¡
him as a source of lifelong internal dissension? How will it remain in him
at all, if he repudiates it absolutely? Surely, in the same way in which Polus
is said in the preceding text to'beLieve'the thesis he repudiates: in virtue
of believing celtain other things which, unbeknown to him, entail that
thesis.Thus Callicles is being told thât ifhe cannot refute the Socratic thesis
(and he is not being encouraged to think he can), he will alwøys-bís
'whole life long'-believe propositions which entail it. Here we have conclusive evidence for what I suggested above: Socrates is convinced that
when he shows his interlocutors the inconsistency of their thesis with
the conjunction of premisses to which they have agreed, Ihey will never
succeed in saving their thesis by retracting conceded premisses: if they try
to save it in this way, they will be bound to failt fail they must, if regardless of rùhich conceded premisses they choose to retract, there will ølways
be others in their belief system which entail the Socratic thesis.
Socrates then is making a tremendous assumption. Stated in fullest generality, it comes to this:
(A)
Anyone who ever has a false moral belief will always have at the
same time trues8 beliefs entailing the negation of that false belief.
That he is counting on the truth of this proposition is implied unambiguously if we assume-as we surely may-that what Socrates is saying at
T24(b) he would also say about any of the theses he refutes in elenctic
E
E
Ë
s At this point the formulation of âssumption A iri the abstract published íî lhe Journal of
Philosophy (c1. above,n 1) is amended by the addition of the wo¡d 'true', This is required to
block an unintended consequence of the unamended formula, drawn to my attention by Richå¡d
K¡aut's inte¡pretatioû of that fo¡mula.
g
t
58
GREGoRY
THE SOCRATIC ELENCHUS
vlAsros
arguments. We could not have derived this result from'123 just by itself.
consistency of his own set of moral beliefs. The self-consistency of his own
position is the only reason he gives Callicles for identifying his own
theses-so presumptuously it would seem-with those of 'philosophy': she
'always says the same'; by implication, so does Socrates too.60 At the conclusion of that speech to Callicles he elevates consistency to a supreme
desideratum in his own search fot truth:
For while this shows Socrates' assurance that'Polus and the rest of
rnankind' who have the false beliefp none the less have t¡ue beliefs entailing its negâtion, it does not sho\ü that Socrates is convinced that they will
always have such beliefs. That this is his conviction becomes entirely clear
in T24(å). But it shows up also, though less saliently, in T22 and its associated texts,T18 and T21. For why should Socrates come to his elenctic arguments confldent that he can produce as'witnesses' for his own thesis those
very persons who deny it, unless he were assuming that if his thesis is true
those who assert its contrâdictory are always harbouring true beliefs which
countermand their denial of his thesis?
If
Ë
Because
A.
Here we come within sight of the solution of'the problem of the
elenchus'. To reach it \rye need only note that from assumption A Socrates
could infer with certainty that any set of moral beliefs which was internally
consistetrt would consist exclusively of /,"¿¿¿ beliefs; for if it contained
even a single false beliel then, given A, it would contain beliefs entailing
the negation of that false betiel We know how highly Socrates prizes the
of
5' That is why I ignored it in my account of standard eletchus at the start of Sect.II above.
h should be ignored ií the analysis of the logical structure of any giveD elenctic argument. A is
not a premiss i¡ t¡e argument, ûor does Soc¡ates ever sùggest that it is The remarks from which
I havé teased it out are obíter dicta,'fbe ilrteÌlocuto¡ would be pe¡fectly justifred if he vere to
igriore them as pule Socratic bluster: he has been giveû to ¡eason why i¿ should think them
t¡ue. That is why A has bee[ brought iri only io explain why Socrutes himselT belíevcs lh^l to
prove the iDconsistency of the thesis with the ag¡eed-upon premisses is Þso /¿r¡o to Plove that,
the thesis is false, no one can af6rm it without geûeratiDg contradiction within his own system
of belief
i
(T25) As for myself, I would rather that my lyre were out of tune, o¡ a chot
I
was training, and that the greater part of mankind should dissent from me
and contradict me, than that I should be out of tune with my own single
self and contradict myseIí (Grg.4a2b1-a.3)
this is what Socrates assumes, why does he not argue for it?
A is a metâ-elenctic statement.5e To support it Soüates would
have to engage in meta-elenctic enquiry And this, as I indicated in the
opening paragraph of this paper, Socrates never does in Plato's earlier dialogues. In every one of them prior to lhe Meno Soffates maintains epistemological irnocence, methodological naïvety. He assumes he has the
right method to search fo¡ moral truth, but never attempts to justify that
âssumption. A fortiori he never attempts to justify the assumption on
which his confidence in the constructive efflcacy of the method is predicated. This is not to say that the assumption is arbitrary. He does have a
reason for it. A proves true in his own experience lt never fails Every time
he tangles with people who defend a thesis he considers false and he looks
for true premisses among their own beliefs from which he can deduce its
negation the needed premisses are in place: they are always where they
should be ifA is true. So he has this purely inductive evidence for the truth
59
x
For years he has been striving for just this, constantly exposing the consistency of his beliefs to elenctic challenge, ready to root out any belief,
however attractive in itsel-f, which if allowed to stand would disturb the
coherence of the system as a whole. So this is where he now finds himself
after all those years of searching: of all the sets of moral beliefs competing for acceptance in elenctic argument, only one has shown up in his
own experience that meets this desideratum-his own. All others, when
tested for consistency, have failed. So he has evidence-as before, inductive evidence6l-for a further assumption,
(B) The set of moral beliefs held by Socrates at any given time
E
is
consistent.
F
m Cf, also GJ'a,490e10-1L Socrates assures Callicles that he, untike CallicÌes, always says'the
same things about the same thìngs'.No one should be misled by his retrospective reúark at the
ß
F
K
å
g
Ë
Ë
Ë
Þ
Þ
dialogue's eDd (527d):'we never think the same about the same thirgs'. As Dodds (Plator
Gorgias) remarks ad loc.,'this reproach applies ofcou¡se to Callicles only, but Soüates politely
includes himselj'. Fo¡ si¡qrlar use of the frfit-person plural-'we' i¡onically substituted fo¡
'yotr'----c1. Euthphr,15c8-9: 'Either we were wrong when we ag¡eed before, or, if we were ¡ight
then, v¿e are wrong nou,' (as the context shows, 'we' in its last occuûence refeIs exclusively to
Euthyph¡o); Cl¿¡ìn. 175bê7:'We have admitted that there is knowledge of knowledge, although
the ârglrrDent said "No" ' (it had been C tias who argued fo¡ 'knowledge of knowledge'; it was
Socrates who produced the'argùment that sard "No"'); ¿¿. 194ci 'Come, Nicias, and,ifyou caû,
¡escue you¡ f¡iends who are stom-tossed by the a¡gument' (or y Lâches hâd been 'stormtossed'; Socrates, sailing ve¡y sûroothly, had rebutted each oI Laches' definitions), The irony in
Gr& 527d is Eanspareût Callicles had been coDvicted of numerous inconsistencies, Socrates of
not even one,
ó¡ The coDsisteûcy of the set ìs being inferred fiom its track record in Socrates' own experience: in aìl of the elenctic arguments in Ìvhich he hâs engâged his set has neve¡ been faulted for
inconsisteDcy-a very chancy infereDce: the results of eÌenctic argument a¡e powe¡fully affected
by the a¡gumentative skill of the contestants; since that of Socrates vastly exceeds that of his
interlocutorE he is ñore effective in spotti[g beliefs of theirs which entail the negatio[ of their
theses thân are theywhen trying to do the same to him;so his ùndefeated record need not show
that his belief-set is consistent;it may or y show that its inconsistencies have defred tbe powei
ofhis adve¡sâries to ferret them out. Socrâtes could ha¡dly have been unaware of this unavoidable hazard in his method.This must contûbute to the sense of its fallibility which,l believe, is
the dght clue to his p¡ofession of ignorance.
ã
60
just been shown, B in conjunction with A entails that Soqates'
belief-set consists exclusively ol true beliefs.
This last move yields the missing piece required for the solution of
'the problem of the elenchus'. The puzzle arises over Socrates' claim at
(4) ú the above analysis of standard elenchus: when he has sho\¡'n that
not-p follows from ø and r, whose truth his ârgument has done absolutely
nothing to support, why should he want to claim that through his argument not-p'has been proved true'? What makes him think it has? The
And,
A
and B, which entail
(C) The set of moral beliefs held by Socrates at any given time
Þ
&
g
n
APPENDIX
tfue,
logue after dialogue, disclaiming åe has knowledge, none the less searches
inàefatigably for moral truth, confident that it is findable, and in the most
unlikely of all places-in the minds of those misguided, confused, wrongheaded people whose souls he seeks to improve. The question'How could
this be true?', which never disturbs Plato's Socrates, never stops disturbing Plato. For years he sees no answer to it.63 Then, one day, he becomes
convinced of something Socrates would have thought fantastic-that eveÌ'y
person's soul had existed long before bfuth, had gone throügh many previous births into different incarnations, and had thus, in some mysterious
The Demise of the Elenchus
Ë
it
the Euthydernus, Zysis, and
H¡ppiq: Major
Ë
It does not seem to have been noticed in the critical lite¡ature that these three dialogues, each of which has been ftequently thought (on the strength of miscellaneous
c¡ite¡ia) to fall late within the earlier dialogues,óó have a common feature which distiDguishes them from all of the othe¡ dialogues iû this group: abandonment of adversary argument as Socrates' method of philosophical investigation. The theses which
Ë
are seriously debated in these dialogues are not contested by the interlocutor;
Socrates himself is both their author and cdtic.
F
å
tiv.
62 Attstotle,Me¡aph.g17"32ff , with coÍìment aa to".
o. no.q ¡¡'srorleb Metâphysics
(Oxford: Oxford U¡iversity Press, 1924).
6 We can pinpoint the iime of bis life at which this question became so i¡sisleût in Plåto's
miDd that heiui it into the ceDt¡e oI his depiction of Socrates' practice of standard elenchùs:
the time whei he w¡ote the Go¡8i¿r, on which I foltow the widely held view that it is one of
the latest of Plato's earlier dialogues (tle best case for this view is in Irwin's commentary
on lhe Gorgias (Plato: Gorgias, 4_€); lor a contrary opinion, see Charles Kahr¡,'Did Plâto writ€
Socratic Diälogùes?', C/asslcal Quarte y,3l (1981). 305 ft) Though (he elenchus ¡s prâctised ir
all of the dialo,-gues which precede the GorSlds (except the Apolo7, arld Menexenß), o ly 11
the Gort¿rs is ¡oc¡ates máde to give utte¡ance to lhar Bock of obiter d¡ct¿ which reveal the
assumptLns on which he predicarcs his confidence that the elenctic method establishes truth
61
in every soul in the form of true covert beliefs. Would not that have struck
Plato as ans\ryedng the question Socrates had never pursued:how could it
have happened that each and every one of his i¡terlocutors did have those
true beliefs Socrates needs to refute all of that person's false beliefs? That
wildest of Plato's metaphysical flights, that ultra-speculative theory that
all learning is recollecting, is understandable as, among other things, an
answer to a problem in Socratic elenchus. Could this be why, when Plato
adopts it, he puts it into Socrates' mouth?65
is
from which it follows that 4 and r are true, since Socrates has agreed to
them. Hence Socrates would feel justifled in making the claim at (4): to
show that a proposition follows from premisses which are true iJ to prove
that proposition true.
Imigine now Plato writing Socratic dialogues with y'rr3 mind full of epistemological worries. Under the influence of a certain Cratylus he had
become convinced that there can be no knowledge of the sensible \ryorld
because it is all in flux,62 and this has left him wondering how there could
be any knowledge of anything at all The Socrates he brings to life in dia-
ELENCHUS
way, acquired knowledge aboul everythíng,6a and thrs knowledge was now
as has
answer is in assumptions
THE SOCRATIC
É
GREGORY VLASTOS
s 'The sout being immortâl and hâving had many btths, and havìng seen eve¡ything both in
this world and i¡ Hades,there is nothing it has not come to know' (Me¿. 81c5-7), It should be
emphasized that there is Dot the slightest evidence of Plato's acceptance of this extraordirary
doctriDe prior to the M¿fto aûd, especially, in vieu, of what was said in the preceding note, no
eviderce of his acceptânce of it in the Go¡gia,r. The eschatological myth with which the Cor8las
coDcludes_is a pu¡ely mo¡al tale, without epistemological content of any so¡t.
6s
This essay is the outcome of discussions of the elenchus in semina$ at Berkeley, To¡onto,
and St And¡ews. An ea¡lie¡ draft was circulated among friends and several of them responded
{'ith comme¡ts. Though I cannot thank them all by rame, I must mention those among them
whose c¡iticisms p¡ompted some specific coÍection: Jülia Annas, Myles Burnyeat, Jim
Dybilcowski, Michael Ferejohn, Alvin Goldman, Charles Kahn, Iar Kidd, Richard Kraut,
Jonathan Lear,Alexander Nehamas, Richard (o¡ty, Jer¡y Santas, Friedrich Solmsen. My grcâtest debt by fa¡ is to Burnyeat fo¡ a discussion which enabled me to cla ly the ârgument i¡ Sect.
lII. But Deither he rior any of the aforenamed should be presumed to agree with any ot the
views I have exp¡essed,
6 Onthe Euthydemus, writes Guthrie (,4
*
ß
¡Iirtory of Grcek Philosoph!, iv.266),'the prevaiting opinior lreviewed in H. Keulen, Untersuchunten z¡l Pl¿¡onb Euthydemus (Kla,ts. Philol.
Sfud., xxvü (wiesbadetrì Hanassowitz, 1971)l is that [it], like the Meno, was written afte¡ the
earlier Socratic dialogues and the P¡o¡a8o¡dr, but before tbe great centrâl group'. On the ¿)s,s,
see esp. the ùsefùl ¡eview of work oD this dìalogue in V Schoplick, De¡ p latonische Dialog Lysis
(Augsburg: Blasaditsch, 1969; diss. Freibu¡g, 1968), supporthg the conclùsipn that the ¿)r¡r is
closely related to the E¡¿thydemus ar'd $obabl! comes belore the Meno but after the Go¡8i¿r.
The câse for the ¡l¡ppi¿s M¿Jol as a transitional dialogue is argued strongly by John Malcolm
('OD the Place of the Hippias Major
fhe Development of Plato's Thought', ,4rcl¡iv für
Geschichte der Philosophíe,so (1968),189ft) and by Paul Woodruff (Plafor Hippias Majo¡,
h
17s-9).
and falsehood.
*
THE SOCRATIC
GREGoRY vLASTos
62
elenctic,refutation
Preveûted bv the eristic clowning of the two sophists f¡om.usitg
a Protreltic disin
dialogue
of
the
uusinesi
tüãËtitut
ãå".
ilffi th";, s;äËt
by Soclates
investigated
theses
the
only
lli."- '^ ì^'i"" cr"iniâs.In this discourse
himself in the didactic stvle
bv
amended
#i'":ådilä:ö;äîi"-áti'"4'
-¿
ask questions
äi:,iËñääËä;ËñÃ,wiere the inte¡locutor is a ves-man'who mav
ãtjections,but never puts up Ãustainedresistance to asocratic
tte interiocuto¡, a teenagàr' is docilitv itseli when ¡e does
it is to anticipate the very
oi
iärìiìuîJ'iãti!ìi,ii! i'ì' o*n 1to "n"'vone's ðurprise.¡'
(290b-d).
A fufhersay in which
rhought
Soc¡areJ'
å round out
ìt L"ïì-f,"ïi,
to $ound his doctrine
"""¿"¿ rl" ,*ãuriti"t or
äiå;;ilåi:ïih
"r"nctic-argument.ís
he presents as unconfor
haDDi¡ess-which
desire
i" pt"póiit*"-,¡e unive$al
and'senseless'
be'ridiculous'
*ould
"
in PrinciPte: to questton tt, he says'
testable
äö#"ti;;;iù-;;ï*
;Ëi:ilffiãil;tue
iàiä"i-sl. dtlrt
i
.ove ii never made in ány preceding dialoguel there evervthing
is contestable.
Lysis
ever..p-ro- forma and
He¡e asain there is no ele¡chus at all against anybody-not
with Hippothales
àncounter
initial
iåi'åäi^Ëù1],ri-r " ¡. nil¡"t uop'Il n the (he
rro thesis) but
has
tefutation
it
noi
Proposed
youtn
;ï"î;h" I;;;;;;;d
!étt
all the
proposes
"
sõcrates
wav
gets
under
; ;il;;:;;*;. \t/hJn the Tnvestigation
is no
rhere
are
refuted
which
theses
uuitte
i.dt"t
ir'Ëilt i'rÏi"l' åìä
poitt
(whose
strong
teenagers
amiable
ää;.\il;-S-ocr;ieslroposes a thesis the
it tbey
i"
and sotd iooks, nor brains) go along;when he tums agaiûst
",'.,a -.nn"r,
in wiih socrates' new move and tas
¿iil'"JÑ
;ìå",iäiläi;Ï;)iããi
ùJiå*"¿iut"lv'rãll
along.67
Híppiøs Major
is lost.on the sophist'
After regaliûg HipPias with fulsome compliments whose i¡ony
Hippias' ludicrouslv
with
sport
makes
socråtes
is
sprung.
iililñ;;is Þ;öìi;tion
while trying to
corpus)
in
the
i"-ã". åti*i,* t,itv are thd goo--f,est definitions
the¡ cashiers
ï"ìi-'i" "*t,ì*ftät is calledior in a definition (cl above, n 20) andThe.definitions
is
¡?'.question
the'\ryhat
to
"hì;T;;;;;i";ãi*",iãuÈ un"""t'
(293dff) 'the uselul'
;id".;'"";td";ìÑ'ã"ãniìo u",ut.n '" ouslv-'the ñtt¡ns'
(297ell)-are all
hearing'
or
pl.ase.
eyesigbt
through
irqi.fil 'thut wúi"h
and are reruted
::ít;iJ;'iT; i;;;;i"{ "*ã"ni"'i"e"no resisiance from Hippias'
who
terrorizes him
'reladve'
his'
oI
Ëubristic
i;'#J^i;;;.-;;-t¡ãiì'ie"t,
were savrng:
il
Plato
as
lt
is
tht"u¡eo"-to thÌash him for his stupidity
is not worth
others
of
^íJ"v"o
rifutation
elenctìc
ihut
,o
,""
come
mv socrates has now
towafds rhe
pfogress
make
to
meet
rhat he must
üii"lrr, ìiirïri'
truth.
"ïïr"ri_"¡ìi"ir-
Thusintheseth¡eeworks,allofwhichmustprecedeÌheMeno'tornoneofthem
socrates
äiìiiLp^ärtï-ãiãprtysicaí' epistemological, a;d methodological novelties'
6?Cf.PaulShorev,l'y¡ia¡Pld¡oSdid(Chicago:UniversityofChicagoPress'19331'490'onLy'
and theû
ìi';."aàuess wit¡ ,"nic¡ i¡iÉrlocutors acìept whatSocrales suggesr\
tte¡i;"o-ur;;
are dashed by his discovery of new objections'
63
It is a reasonable conjecture that it is Plato himself who has
now lost faith in the elenchus and extricates his Socrates ftom it, allowing him to
move out of it quietlf without commeIÌt, without sayitg that he is doing so, aIld a
fortíor¡ witho.u,t explaidng why,
ditches the elenchus.
Euthydemus
BLENCHUS