7/28/2019 Carried Away in the Euthyphro http://slidepdf.com/reader/full/carried-away-in-the-euthyphro 1/33 z!88z§(T(gN"\0<<J@ØJ@gÉ<"4JÎÓF4@<Ô<BV<JgH@Ê2g@ÂN48äF4<, 6"ÂJÎ 1 ¦<"<J\@<, Ô<BV<JgH2g@Â:4FäF4<, <`F4@<. 4b4-c2 and 6a7-d4 are also designed to provoke the reader to think about what is problematic about 2 appeals to authority. Euthyphro’s definition itself–at least in the way it is understood in the dialogue–can also be thought of as authoritarian: cf. S.M. Cohen, ‘Socrates on the Definition of Piety: Euthyphro 10A- 11B’, Journal of the History of Philosophy 9 (1971), 1-13, at p. 2 (this article is reprinted in G. Vlastos (ed.), Socrates (Garden City, 1971), pp. 158-76). In the present passage Burnet takes B@*gPf:g2" only to indicate ‘accept as a starting point for discussion’ (J. Burnet, Plato’s Euthyphro, Apology of Socrates, and Crito (Oxford, 1924), p. 47); but the contrast here is with ¦B4F6@Bä:g< ... gÆ6"8äH 8X(gJ"4.’ ðïäÝ÷ïìáé at 6a8 also means ‘accept’, not ‘accept as a starting-point’. Carried Away in the Euthyphro Lindsay Judson, Christ Church, Oxford At 9e1-3 Euthyphro offers his second definition of piety: ‘Well, then, I would say that the pious is this, whatever all the gods love; and that its opposite, what all the gods hate, is impious.’ 1 Socrates asks if they are to examine this definition, or merely accept it because someone has stated it–not the first implicit attack in the dialogue on the idea of relying on appeals to authority. The examination begins with the famous question: ‘Is the pious loved by the gods 2 because it is pious, or is it pious because it is loved?’. The argument which begins with this question (10a1-11b5) is much stronger than is often claimed: this argument, and the role of definition in it, are the subjects of this paper. The Euthyphro on definition Although in his characteristic fashion Socrates offers no set-piece discussion of what kind of definition he has in view, we can nonetheless glean a certain amount about this, partly from what he does say and partly from the role which definition plays in the argument with which we are concerned. Two well known remarks are particularly relevant. The first is at 6d9-e6: Do you remember that it was not this that I was bidding you to do–to teach me some one or two of the many pious things–but to teach me that very form by which all pious things are pious? For you said, surely, that it is by one form that the impious things are impious and
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
z!88z §(T(g N"\0< < J@ØJ@ gÉ<"4 JÎ ÓF4@< Ô < BV<JgH @Ê 2g@Â N48äF4<, 6"Â JÎ1
¦<"<J\@<, Ô < BV<JgH 2g@Â :4FäF4<, <`F4@<.
4b4-c2 and 6a7-d4 are also designed to provoke the reader to think about what is problematic about2
appeals to authority. Euthyphro’s definition itself–at least in the way it is understood in the dialogue–can
also be thought of as authoritarian: cf. S.M. Cohen, ‘Socrates on the Definition of Piety: Euthyphro 10A-
11B’, Journal of the History of Philosophy 9 (1971), 1-13, at p. 2 (this article is reprinted in G. Vlastos
(ed.), Socrates (Garden City, 1971), pp. 158-76). In the present passage Burnet takes B@*gPf:g2"
only to indicate ‘accept as a starting point for discussion’ (J. Burnet, Plato’s Euthyphro, Apology of Socrates, and Crito (Oxford, 1924), p. 47); but the contrast here is with ¦B4F6@Bä:g< ... gÆ 6"8äH8X(gJ"4.’ ðïäÝ÷ïìáé at 6a8 also means ‘accept’, not ‘accept as a starting-point’.
Carried Away in the Euthyphro
Lindsay Judson, Christ Church, Oxford
At 9e1-3 Euthyphro offers his second definition of piety: ‘Well, then, I would say that the
pious is this, whatever all the gods love; and that its opposite, what all the gods hate, is impious.’1
Socrates asks if they are to examine this definition, or merely accept it because someone has
stated it–not the first implicit attack in the dialogue on the idea of relying on appeals to
authority. The examination begins with the famous question: ‘Is the pious loved by the gods2
because it is pious, or is it pious because it is loved?’. The argument which begins with this
question (10a1-11b5) is much stronger than is often claimed: this argument, and the role of
definition in it, are the subjects of this paper.
The Euthyphro on definition
Although in his characteristic fashion Socrates offers no set-piece discussion of what kind
of definition he has in view, we can nonetheless glean a certain amount about this, partly from
what he does say and partly from the role which definition plays in the argument with which we
are concerned. Two well known remarks are particularly relevant. The first is at 6d9-e6:
Do you remember that it was not this that I was bidding you to do–to teach me some one or
two of the many pious things–but to teach me that very form by which all pious things are
pious? For you said, surely, that it is by one form that the impious things are impious and
The shifts in focus in the dialogue between pious actions and pious people are a topic I shall explore5
on another occasion.
Not everyone accepts this. For discussion see Terence Penner, ‘The Unity of Virtue’, Philosophical 6
Review 82 (1973), 35-68 (reprinted in H.H. Benson (ed.), Essays on the Philosophy of Socrates (NewYork and Oxford, 1992) and Gregory Vlastos, ‘What Did Socrates Understand by his “What is F ?”
Question?’, in Vlastos, Platonic Studies (2 edition, Princeton 1981).nd
the pious things pious.... So teach me what this form itself is, so that looking to it and using
it as a model, I may say that whatever actions (whether yours or anybody else’s) are of this
sort are pious, and those that are not of this sort are not pious.’3
The expression ‘by which all pious things are pious’ is hardly transparent. Some light is cast on
it by the second remark, at 11a6-b1:
though you were asked, Euthyphro, what the pious is, you are in danger of not wanting to
make clear to me its essence, but only to tell me of some feature that it has .... But what it
is, you have not yet said.4
The sort of definition that Socrates is after is a specification of whatever it is about a pious thing
(or person ) which makes it pious. Though this idea is to a large extent indeterminate, it at least5
excludes the idea of what merely produces or brings about the piety of pious things: Socrates’
reference to the @ÛF\" of piety rules this out. It might then be natural to construe ‘by which all6
pious things are pious’ in terms of identity or of constitution; but there might be other
possibilities too. Participation in a Platonic Form, for instance, might resist analysis in these
ways without resting on a relation of efficient causation either; if so, a more abstractly-specified
For example, Burnet, Plato’s Euthyphro, ..., pp. 47-8; Chris Emlyn-Jones, Plato: Euthyphro (London,13
1991), pp. 78-9; Cohen, ‘Socrates on the Definition of Piety’, p. 3; A. Kim, ‘A Chiastic Contradiction
at Euthyphro 9e1-11b5’, Phronesis 49 (2004), 219-24, at p. 219.
Cohen might seem to hold that both distinctions are grammatical: ‘whereas Socrates’ first distinction14
was between active and passive voices, this second distinction is between two different passive forms’
(‘Socrates on the Definition of Piety’, p. 4). Likewise he claims that ‘Socrates wants to say thatsubstituting the participle for ‘p’ and the inflected passive for ‘q’ [in sentences of the form ‘p because
q’] will yield a truth, whereas substituting the inflected passive for ‘p’ and the participle for ‘q’ will yield
informal procedure. It might be natural to say that he is trying to make a number of quite
sophisticated distinctions, but lacks any well-established technical vocabulary with which to
mark them–and it is certainly right that the obscurity of the passage is at least partly due to the
fact that he lacks any such vocabulary. But this way of putting the point contains an ambiguity
of scope. I do not think that we should endorse it if read as ‘There are a number of quite
sophisticated distinctions such that Plato is trying to make them’: that implies that there are
precisely specifiable distinctions delineated, however obscurely, in the text. Rather we should
take it that Plato’s formulations are aimed at capturing some precisely specifiable
distinctions–which carries no such implication. Of course, we must try to work out, as well as
the text will allow, what distinctions Plato is after. But there can be no guarantee in advance that
Plato–and still less his text–has a fully determinate grasp of any such distinctions; and in the case
of our passage I think that the insistence that there is such a grasp, waiting to be discovered,
would be misplaced.
(a) The NgD`:g<@</NXD@< distinction This is drawn at 10a5-9. A number of recent
commentators think that Socrates is making a grammatical distinction between the passive and
active voices of verbs; and Emlyn-Jones, for instance, recommends translating 10a10 as: ‘Is13
“being loved” one thing and “loving” something else?’ (p. 79). But it is very hard to read the text
this way; moreover, Socrates’ point ought to tie up in some way with the claim that follows, that
X is a thing carried because it is being carried and not vice versa–and that is plainly neither about
grammar nor about linguistic expressions. The grammatical interpretation appears to confuse14
So Terence Irwin, ‘Socrates and Euthyphro: The Argument and its Revival’, in Lindsay Judson and22
Vassilis Karasmanis (eds), Remembering Socrates: Philosophical Essays (Oxford, 2006), p. 59; cf. Mark L. McPherran, The Religion of Socrates (University Park, Pennsylvania, 1996), pp. 43-4. Cohen adopts
a variant of the agent-dependence interpretation (‘Socrates on the Definition of Piety’, pp. 7-8).
subject and being the object of an activity (broadly construed), and that he uses the second
distinction to arrive at two types of thing which we can discern in the case of the occurrence of
an activity:
(1) how things are with X : its being Y -ö-ed.
Obj(2) X being ö-ed by Y –understood as equivalent to the union, as it were, of (2 ), X ’s being
Subjthe object of Y ’s activity of ö-ing and (2 ), Y ’s being the subject of that activity.
On this view Plato uses expressions such as NXDgJ"4 not to pick out X ’s being the object of the
activity alone, but the ‘jointly-owned’ conjunction of this and Y ’s being the subject of the
activity.
(1) and (2) are symmetrically related: necessarily, (1) holds iff (2) holds. Nonetheless we
find Socrates claiming that there is an asymmetrical relationship between them. Thus he says
that X is a Y -carried thing (call this (1) ) because X is being carried by Y ((2) ), and notcarried carried
vice versa. If we do not mind things being a little loaded, metaphysically, this seems quite
plausible. We arrived at (1) only by taking the jointly-owned relation between X and Y andcarried
deriving a single-owner item from it: it thus seems reasonable to think that things are this way
with X because X and Y stand in this relation, and hence because of (2) . Socrates does notcarried
discuss the type of connection he signals by ‘because’ here, and we shall return to this below.
Some commentators take the point of the distinction to be to show the ‘agent-dependence’ of
Subj being a carried thing–in my terms, the dependence of (1) on (2 ) rather than oncarried carried
(2) . While I agree that Socrates will–quite reasonably, as it turns out–rely on this sort of carried 22
agent-dependence in the third section of the argument, when it comes to the particular case of
As I said earlier, of course, I do not in any case think that being carried is (in the relevant sense) agent-34
dependent: see pp. 10-11.
5d8-6d4, 9a1-d5–both passages in which the reader’s doubts about this picture of the gods are35
provoked as much by Socrates’ repeated setting aside the issue of the nature of the gods as by the little
that he says directly about it.
13c6-d3 and 15a5-6. Notice that Euthyphro does not have a unified view of the gods in the dialogue:36
he embraces the Homeric picture of the gods–anthropomorphic, far from perfect, and at war with each
other (for references see the previous note)–then he supposes that they do agree about what is just and
unjust, and finally expresses the more ‘Xenophanean’ view just cited. This (understandable) plethora
of views seems to me to undercut the long-standing debate over Euthyphro’s (un)orthodoxy; it is alsorelevant to Myles Burnyeat’s claim that in the Euthyphro we see radical, rationalist theology (Socrates)
confronting ‘traditional polytheism’ (Euthyphro) (‘The Impiety of Socrates’, Ancient Philosophy 17
(1997), 1-12). I shall take up these issues elsewhere.
Another possible line of thought is that Plato might suppose quite independently that loving X 37
necessarily involved responding to some intrinsically attractive feature of X . (This conception of love
has often been detected in the Lysis; for the view that the point of the Lysis is actually to reject such a
conception, see Catherine Osborne, Eros Unveiled: Plato and the God of Love (Oxford, 1994), pp. 58-
61.) It may be that McPherran has something like this in mind in the passage quoted above; but if so it
is a line of thought which would disassociate the demand for an explanation in terms of something
eliciting the agent’s love from the material on being carried and being seen. A disadvantage of ascribingthis move to Plato here is that it seems to come perilously close to begging the question against someone
who holds that what makes things pious is that the gods love them; for this reason, among others, I prefer
instance.34
But a much better explanation is available, namely that Plato thinks that the claim that the
gods love the pious for a reason is unavoidable –not because Plato thinks that it follows from the
material about carrying and being carried, since it quite clearly does not follow, but on some
other basis. What might this basis be? There could be more than one answer to this, but an
obvious one is that Plato thinks that the ‘no reason’ option is ruled out by the sort of the
considerations which are advanced in the modern Euthyphro dilemma. If the gods love the pious
for no reason, then they are being arbitrary, whimsical; and this, Plato may think, is
unacceptable–indeed, it is just the sort of picture of the gods which was put in doubt earlier on
the dialogue, and which we will find Euthyphro himself committed to rejecting later in the35
dialogue when he endorses the idea of the gods as perfect and lacking nothing. Moreover, if 36
there is no reason for the gods to love the pious, why should we love it?37
the line of thought sketched in the main text above.
This is brought out, for example, in the emphasis throughout the dialogue on a connection between38
piety and*46"4@Fb<0 –an emphasis which is of course fully explicit in the development of Euthyphro’s
third definition, but is also in play earlier. Thus the piety of Euthyphro’s prosecution depends on the
justice or injustice of the killing of the hired hand by his father (4b7-e3; 8b7-9b3); and the discussion of what the gods dispute over, and what they love and hate, at 7e1ff., relies on a connection between being
pious and being *\6"4`H.
If the gods’ having no reason for loving pious things is ruled out, this leaves the two options
which Socrates actually offers to Euthyphro: the gods love what is pious because it is pious, or
they love it for some other reason. Here we come to Plato’s second piece of short-circuiting:
neither Socrates nor Euthyphro is made to argue for the former view or against the latter. This
too may give the impression of philosophical sleight of hand; but once again this impression will
be dispelled if we can suppose that Plato had reason to find the by-passed view hard to swallow.
In the modern Euthyphro dilemma, this option is dealt with by the thought that we are
considering whether God is the source of morality, and for that idea the claims that God loves
moral requirements because of some intrinsic or essential feature and that he does so because of
some accidental feature are equally problematic. But this is not the context of the argument in
the Euthyphro, for here we are considering only the weaker claim that the gods’ love of pious
things is what constitutes the essence of their being pious; and it seems quite compatible with that
claim that the gods should have a reason for loving these pious things. I suggest that the answer
lies in an unquestioned assumption which underpins the whole dialogue, that an action’s piety
or impiety constitutes a powerful reason for doing it or refraining from it, and for commanding
our respect or abhorrence. Suppose that the gods love the pious for some extrinsic reason, and38
that it is the gods’ love which makes things pious. Then we can ask, why should we respect the
pious, as opposed to whatever incidental feature of it that the gods love? It could be, for
example, that the gods love pious acts because such acts make people happy–and that this is not
what make the acts pious, but is simply a feature which pious actions happen to have. It could
be also that what the gods love in this case is genuinely worthy of love. Then it would appear
So what is the connection between the reasons agreement and what precedes it? Cohen thinks that39
the material on carrying, being carried and being a carried thing is Socrates’ explanation of what he
meant by his original question at 10a2-3, ‘Is the pious loved by the gods because it is pious, or is it pious
because it is loved?’ (‘Socrates on the Definition of Piety’, p. 3). But although this material is introduced
by Euthyphro saying that he does not understand the question and Socrates responding that he will speak
‘more clearly’ (F"NXFJgD@< NDVF"4), this material does not clarify that question. I suggest that the
reader is meant to understand that Euthyphro fails to grasp not merely the question but where the
argument is going; and the material on carrying, etc., does clarify that . The äÞ which introduces the
reprise of the original question at 10d1–the reprise which brings us to the reasons agreement–does notmark the start of an inference from what goes before, but rather signals that what goes before now puts
Socrates and Euthyphro in a position to proceed to the next step in the argument.
that what we ought to be concerned with in our attitudes and actions are things which make
people happy, and the piety of some or all of these ought to be of no interest. This problem does
not arise, on the other hand, if the gods love the pious because it is pious, or for itself : then the
idea would be that there is something good about piety itself , and the gods love that. In this case
we would have a reason to be concerned with piety–namely whatever is good about piety itself.39
The third section: consequences
[I] S: But it is precisely because it is loved by the gods that it is a thing loved and
god-loved? E: Of course.
[ii] S: Then the god-loved is not pious, Euthyphro, nor is the pious god-loved, as you say
they are, but the one is different from the other. E: How so, Socrates?
[iii] S: Because we agreed that the pious is loved because of this–because it is pious–but that
it is not pious because it is loved. Is that right? E: Yes.
[iv] S: But on the other hand, in the case of the god-loved, we agreed that it is because it is
being loved by the gods that it, precisely in virtue of this being loved , is god-loved; but that
it is not being loved because it is god-loved. E: You are right.
[v] S: But if they really were the same, Euthyphro my friend–the god-loved and the
pious–then if the pious were loved because of its being pious, the god-loved too would be
loved because of its being god-loved; and if the god-loved were god-loved on account of its
being loved by the gods, then the pious would be pious too on account of its being loved.
[vi] But as things are, you can see that the two things are oppositely placed, as being
altogether different from each other; for the one is such as to be loved because it is loved,
while the other is loved because it is such as to be loved. [vii] Though you were asked,
Euthyphro, what the pious is, you are in danger of not wanting to make clear to me its
So Cohen (‘Socrates on the Definition of Piety’, pp. 8-10), Emlyn-Jones ( Plato: Euthyphro, pp. 7841
and 101), Geach (‘Plato*s Euthyphro*, pp. 376-7), Patzig (‘Logic in the “Euthyphro”’, pp. 298-9), Lynn
E. Rose (‘A Note on the Euthyphro, 10-11’, Phronesis 10 (1965), 149-50); see also Wolfsdorf cited in
n. 45. These commentators differ as to whether the conclusion is non-identity, non-equivalence, or
something else. Thomas D. Paxson (‘Plato’s Euthyphro 10a to 11b’, Phronesis 17 (1972), 171-90, at pp.
185-6), Kim (‘A Chiastic Contradiction’, p. 222), and Panos Dimas (‘Euthyphro’s Thesis Revisited’,
Phronesis 51 (2006), 1-28, at pp. 10-11) claim that the argument involves a substitution of ‘god loved’
or ‘loved by the gods’ for ‘pious’ which they think is indeed–given Euthyphro’s definition– truth-
preserving, although they differ from the other commentators cited in seeing Socrates as using thesubstitution to derive an unacceptable conclusion of one sort or another, rather than merely as arguing
that truth has not been preserved (cf. McPherran, The Religion of Socrates, p. 43, n. 43).
essence, but only to tell me of some feature that it has, when you say that the pious has
received this feature, namely being loved by all the gods; but what it is, you have not yetsaid. (10d9-11b1)40
Does Socrates appeal to substitutivity? In this final section Socrates brings together the
conclusions of the first two sections in an argument to show that Euthyphro’s definition must be
rejected. The passage at 10e10-11a6 (steps [v] and [vi]) is often taken to deploy a substitutivity
argument: if ‘pious’ and ‘god-loved’ are not interchangeable salve veritate, then they do not
denote the same thing. Step [v] (10e10-11a3), might seem to apply this move to the sentences41
Wolfsdorf subscribes to a variant of the substitutivity interpretation, but floats the idea that Plato takes45
the relation signalled by Socrates’ ‘because’s to be extensional (‘ Euthyphro 10a2-11b1’, pp. 64-7) (so
that the ‘because’s in what I call the reasons agreement do not directly signal the gods’ reasons for loving
the pious, but a more generic causal relationship). Wolfsdorf deploys a highly fine-grained ontology for
the relata–‘that which is holy qua holy (that is, holiness)’, ‘that which is god-beloved qua god-beloved
(that is, god-belovedness)’ (p. 2; cf. p. 64), and this seems to be needed if the extensionality in question
is to be plausible. Given this machinery, however, Plato could have gone straight from the reasons
agreement in the second section to his conclusion without any need for either the first or third sections:
from ‘it is loved because [understood extensionally] it is pious, but it is not pious because it is loved’
(10d6-7) it will follow immediately that being loved by the gods (understood in the ‘qua being loved’way) is not the same thing as being pious (qua pious). We have good reason, then, to see the argument
as working in a different way.
substitutivity test (in whatever form) to (S1)(a), and appeals to the truth of (S1)(b) to show that
truth-value is not preserved: the other is an explanation of why he can insist that (S1)(b) is true,
on the basis of (S1)(a). This second ingredient is, of course, the heart of the matter, since it
would appear that Euthyphro’s definition requires him to deny (S1)(b) if he can. I shall suggest
that once we have given the explanation for this, there will be no need to see the text as
containing any appeal to substitutivity.45
‘Because’ We must consider first, however, the nature of the connection(s) which Plato means
Socrates to signify when he says that something is a god-loved thing because it is being loved
by the gods, and not vice versa, and when he says that the gods love the pious because it is pious.
Once again, Socrates offers no explanation of the connection(s), and we must be careful about
how determinate an account we ascribe to him. On the surface, as we shall see, Socrates’
argument involves a matrix of connections of a single (if unspecified) type, which is, apparently,
transitive and, in at least some cases, asymmetrical. These features make it natural, though not
inevitable, to construe the connection as one of dependence; and this construal receives some
confirmation from that the fact that it certainly reveals much about the argument’s points of
difficulty and success. There are many items in this matrix, all but one invoked by the term
‘because’: Socrates and Euthyphro useÓJ4, *4`J4 and*4V interchangeably for this purpose, over
40 times in our passage (10a1-11b4). Only one item is not so invoked–Euthyphro’s definition
As I noted earlier, the definition too should be understood as involving a dependence relation, despite46
Socrates’ use of ‘the same’ at 10e10. David Charles has suggested to me that Plato might be advancing,
or fluctuating between, two different conceptions of definition, and I do not think that these possibilities
can be entirely ruled out. It is unclear, however, what Plato might expect to gain in this context from
introducing an alternative conception, and the argument’s steadfast reliance on dependence makes it
much more plausible that Plato here simply takes definition to introduce a dependence relation than thathis use of ‘the same’ is philosophically scrupulous or careful.
Pace Cohen, who thinks (i) that Socrates equivocates in his use of ‘because’ here (‘Socrates on the47
Definition of Piety’, pp. 5-8), and (ii) that he does so non-fallaciously. According to Cohen, the
‘because’ in ‘Is X a ö-ed thing because it is being ö-ed?’ is a sort of logical ‘because’, but in the converse
question–‘Is it being ö-ed because it is a ö-ed thing?’–the ‘because’ has a quite different sense: it now
indicates the idea of the reason Y has for ö-ing X . But it is bizarre to suppose that, without any signal
or contextual indication, a double question of this form is meant to involve two quite different senses of
‘because’–senses which mean that the alternatives are actually not opposed to each other. (This is a
matter of some delicacy: that Socrates is presented as taking the connection to be of the same type in both
halves of the question is compatible with his not being presented as taking a view about the precise typeof explanatory connection in question.) The case of the 10e6-8 is more difficult, for the reasons given
below.
itself, which enters the matrix at 10eb10-11. I think we should infer that Socrates, as Plato
presents him here, construes definition as itself involving dependence: a thing’s being pious
depends in some way on the essence of piety. How this definitional dependence figures in the46
argument will be explored below.
Trouble begins when we consider the connections invoked in the matrix in more detail. In
the argument’s first section, the only plausible asymmetrical connection which we could identify
between items such as X ’s being Y -carried and X ’s being carried by Y was some sort of
metaphysical or ontological dependence of the former on the latter. When in the same breath
Socrates denies that the latter holds because of the former we would expect that he is speaking
of the same relation. Moreover, the context is such that were he to intend a different one, he
would have to signal this in some way: his question is plainly of the form ‘is it the case that p
because q or vice versa?’, and questions of that form obviously presuppose that no equivocation
on ‘because’ is intended. In the case of the reasons agreement in the second section, however,47
it is hard to avoid taking the connection–signalled once again by ‘because’–to be a matter of the
See p. 16 and n. 36 above. Note that the transparency assumption would not save the argument from54
substitutivity salve veritate: even if she knows that they are one and the same person, it might still be the
case that ‘She wanted to marry him because he was the handsomest man in Britain’ was true while ‘She
wanted to marry him because he was the richest man in Britain’ was not. The same holds in the case of necessary truths, too: the Pythagoreans admired the number 10 because it is the sum of the first four
natural numbers, and not because it is the immediate predecessor of 11.
Socrates first argues from (1) and (2) to (3). As the excursus through (6) reminds us, the
dependence in (2) differs from that in (1) in being a matter of the gods’ reasons for loving P .
This sort of dependence is an intentional matter; so that we cannot rely on a general transitivity
across it to establish the conclusion. What I suggest licences the inference to (3) is Socrates’
conception of definition together with what in the discussion of the second section I called his
assumption of transparency: if the gods love the pious because of its piety, and if the essence of
that piety (specified by its definition) is E , then the gods will grasp both of these facts, and will
correctly regard E , or some aspect of it, as what attracts them about the pious. This seems
sufficient to enable Socrates to infer (3); and with this defence of (3), his next inference, from
(1) and (4) to (5), is in order too. Is the transparency assumption problematic? It would of course
be problematic be on the Homeric conception of the gods advanced by Euthyphro early in the
dialogue, since on that conception the gods can be deceived; but it seems entirely in order on the
sort of ‘Xenophanean’ conception which Socrates has all along accepted and which, as I have
said, Euthyphro himself also turns out to endorse.54
Socrates’s argument involves not only these two inferences, of course, but also the twin
claims that because (4) is true, (3) must be false and that because (2) is true, (5) must be false–i.e.
that the dependencies in question are asymmetrical. Why does Socrates think this? He plainly
needs an argument, since the dependence of A on B does not necessarily exclude the dependence
of B on A; this latter dependence might in principle be of the same sort–as in cases of mutual
Once again, we could recast the point in terms of dependence trees. As David Charles has suggested,55
the argument that D1 and D2 are incompatible with D3 might simply embody the assumptions that thereis one linear order of dependency and that no item can appear in two different positions in that order; but
the problem is that on the general level these assumptions seem quite unfounded.
dependence such as we encounter in arches and card-castles–or of a different one. One might55
be tempted to suppose that (3) was already shown to be false by the considerations advanced in
the first section, where the metaphysical dependence of X being Y -ö-ed on X beingö-ed by Y did
indeed seem asymmetrical. If Socrates thinks this he is mistaken, however, since the purport of
(3) is to some extent indeterminate: the deduction of (3) rests in part on (2), and hence invokes
a distinct form of dependence, as we have seen. In other words, we could concede that the
metaphysical dependence involved in (4) is asymmetrical, but understand (3) as nonetheless true
because the dependence it involves can be construed as a different sort–as a matter of the gods’
reasons for loving P . Thus the asymmetry implied by (4) does not require (3) to be false. If
Socrates thinks that (3)’s truth is ruled out on general grounds–that is, grounds which can be
specified in terms of dependence in general–it is very hard to see what these could be. He does,
however, have a good response available, which focuses on the specific content of (3). The
objection to Socrates which I have sketched says, in effect, that (3) is true, and compatible with
(4), if understood along the lines of:
(3r) The gods love P because of P ’s being a god-loved thing.
But (3r) faces a problem of its own which is familiar to all students of the Euthyphro, since it also
arises in relation to our next question, why Socrates thinks that (2) and (5) cannot both be true:
if Socrates can defend this incompatibility, he can reject (3r). Let us put (3r) to one side for the
moment, and turn to (2) and (5).
(5) relies entirely on what seem (to us) to be metaphysical dependencies–of the sort we
identified in the argument of the first section, and of the sort we inferred to be at work in Socratic
definition. Both of these relations are plausibly asymmetrical, and the asymmetry of the
Compare the following. ‘Obey this command’ is vacuous in the sense that nothing counts as obeying57
or disobeying it. ‘Obey all these commands’, when a member of a set of commands, need not be vacuous
(‘Keep quiet; shut the door; obey all these commands’); but the vacuity reemerges if the set consists of
a number of commands all of the form ‘obey all these commands’.
Couldn’t I love something today only because I loved it yesterday? Possibly–but eventually we shall58
get back to a day when I had some other reason for loving it, or loved it for no reason at all. But couldn’t
the gods love something each day only because they had loved it the day before, so that we never get
back to a day when things are different? Again, possibly–but although in this situation the gods always
have reasons for loving what they love, as with the case discussed in the main text, the gods do not seemto emerge as perfect or unwhimsical, and the fact that they love the pious again offers us no reason to do
so.
under a misapprehension as to the nature of piety, we might well feel able to affirm both (5) and
(2), if the latter were understood in terms of modes of presentation or descriptions under which
the gods love the pious.
Someone might try to counter this account by supposing that each god loves the pious
because all the gods do. This situation does not seem to me to escape the difficulty, since (given
that all the gods have only the reason now alleged for loving the pious) this reason for loving the
pious is as vacuous as the original reason–i.e. it cannot operate as a reason. If it is somehow57
not vacuous in this sense, then we would have to agree that the gods in this situation would have
a reason for loving what they love; but it is one which fails to satisfy the objections which I
sketched on Plato’s behalf as grounds for the reasons agreement–the objections to saying that the
gods have no reason, or only an extrinsic reason, for loving the pious. Though armed with a
reason, the gods in this situation seem to be whimsical in their choices (albeit in a different way)
and imperfect (since they should have better reasons); and the fact that they love the pious offers
us no reason to do so.58
We have, one might say, arrived back at Socrates’ famous question at 10a2-3, ‘Is the pious
loved by the gods because it is pious, or is it pious because it is loved?’ Once the question is
understood in the light of the reasons agreement and Euthyphro’s definition, we can see that
Euthyphro is committed to answering ‘both’–and that this answer is unacceptable. We should
If being loved by the gods were only part of the definition of piety, it would be quite compatible with59
the claim that the gods love pious things because of their piety, as I have understood that claim. If being
pious were, essentially, being loved by the gods and being F , the gods could love the pious because of its essential F -ness. This is the sense in which the argument relies on the idea that Euthyhpro’s definition
is (meant to be) complete.
conclude that the essence of piety must be constituted (at least in part ) by something other than59
being god-loved–something which either is or includes the attractive property which inspires the
gods’ love–and thus that Euthyphro’s definition is incorrect.
Conclusions: the body of the argument
Plato’s commitment to the introduction and deployment of very general structures in the
refutation of Euthyphro’s definition makes it tempting to suppose that the argument is meant to
proceed at an entirely general level–that is, solely on the basis of general features of dependence
relations. As we have seen, however, these general structures cannot provide the resources to
make the argument to go through. We could, of course, suppose that Plato simply failed to see
this; but since considerations specific to the subject matter of the argument (the gods’ love of the
pious) do make the argument at least plausible, we have no reason to ascribe this mistake to
Plato–especially since these considerations are ones we have reason to see Plato as in any case
having in mind in relation to the reasons agreement.
It is a cliche for commentators to say that such-and-such a philosopher of the past has
provided the skeleton of an argument, and that what is needed is to put some flesh on the bones.
Something more like the reverse is true of parts of Socrates’ argument here. As we have seen,
both the technical resources he deploys (especially the notion of definition and the distinctions
drawn in the first section) and the lines of arguments he constructs display crucial and pervasive
indeterminacies, so that the argument as presented is (to change the metaphor) rather lacking in
philosophical musculature. Despite this, we have been able to supply (or in some cases perhaps
detect) enough support for key moves to reveal an argument with an impressively strong overall