“TO SAY WHAT IS MOST NECESSARY”locutors fail, the dialogues constantly warn otherwise. Indeed, in the
Theaetetus (191c), Socrates suggests to his young friend that when the
account they have settled upon seems good and true, then they must be
most vigilant, torturing (basanivzw) their discourse so that they may test
it from all sides.^29 The Cratylus suggests that to avoid the grave danger
of self-deception one must continuously turn back again to what has
been said and test it by examining it “forwards and backwards simulta-
neously” (428d).
The Cratylus offers a nice example of the way in which Socratic elen-
chus insists on testing all sides of a hypothesis. In it Socrates attempts
to determine whether names are natural or conventional. Socrates fi rst
posits that names are natural and then argues against that conclusion,
or, we might say, stipulates an assumption in order to uncover its prob-
lems, as is his practice. Then he posits the contrary assumption, that
names are conventional, and proceeds to uncover the problems with
that assumption. However, his treatment of these two alternatives does
not allow names to be some mix of natural and conventional elements
either. His examination backwards and forwards demonstrates the in-
completeness of two opposing yet plausible explanations, suggesting
that one must move beyond either or both to get closer to understand-
ing the nature of the matter.
This is precisely how the refutation of Charmides’ position on
swfrosuvnh works. Socrates’ argument shows Charmides’ account of the
nature of swfrosuvnh to be fl awed primarily in its partiality. To suggest
that swfrosuvnh is calmness is to be obviously incomplete. Socrates does
not need to demonstrate what swfrosuvnh is to refute that partiality.
Indeed, if he is honest in his claims, he does not know. Rather, Socrates
offers another defi nition, equally partial and equally obvious in its par-
tiality, but opposed to the defi nition offered by Charmides, as an invita-
tion for Charmides to see the fl aw in his own position, and the partiality
of such accounts.^30
Socrates never fi nds an unassailable account for any matter he
investigates, and those that come closest are consistently undercut. In
the Parmenides (133b and 135b) a clear distinction is drawn between a
correct logos and an irrefutable logos in such a manner as to suggest that
Socrates is rather uninterested in the latter. The Euthydemus pre sents
the possibility of an irrefutable logos in unfl attering light in the answer
the eponymous character proffers to any question: “Neither and both!”
(300d). In the Phaedo (90c– d), Socrates claims that right or wrong logoi
are extremely rare because most logoi are at times true and at times
untrue. In many dialogues, interlocutors remark that a position that
seemed sound to them but a few minutes ago now seems fl awed. Socrates