I have to say that I am a little stumped by this passage in the Phaedo: (Cebes has just asked Socrates what he should tell Evenus when he asks again about why Socrates has been composing verse)
The ἐπιτάττοι seems hard to explain. The construction is obviously a conditional following ἀφοσιούμενος - a participle functioning as an apodosis. The εἰ must be the protasis, but the use of the present optative would not make much sense - the optative is used only of future unreal conditions (or whatever you know them as: unfulfilled, closed, less real etc): “If it were to rain tomorrow, I would get wet”. But “if after all, perhaps the dream figures were to enjoin me to compose this sort of music/art” is not possible since the ordering occurred in the aforementioned dream now in the past. Indeed Rowe (who has written the most up-to-date commentary on the Phaedo) translates “if it was enjoining me” - i.e. past (unless he is ignorant enough to be saying “was” instead of “were” to mark an english subjunctive which I think unlikely). If it was a present condition I could perhaps understand it - the command is still valid right now as Socrates speaks but that would require an imperfect in the protasis.
This construction returns later with a modified prefix on the verb:
But that is simply not true. The optative can be used in all kinds of conditions,
but the one condition you would most likely not find an optative in is an unreal/contrary-to-fact one.
There you would have the imperfect/aorist/(less often pluperfect) in protasis and impf./aor./(less often plpf.) + ἄν in the apodosis.
The optative itself does not have any tense, and the use of present vs. aorist is merely to convey
aspect.
I think this is just an instance of εἰ ἄρα with optative substituting for indicative after a secondary tense in a clause that is in effect indirect discourse because it represents what Socrates was thinking in the past time. (ἀφοσιούμενος is a present participle, of course, but the primary verb ἐποίησα is aorist.) εἰ ἄρα suggests a doubtful possibility, as you capture in your translation (see Smyth sec. 2796). “I composed these poems . . . and in satisfaction of a sacred obligation, if after all, perhaps the dream figures were enjoining [not “were to enjoin”] me to compose this sort of music/art.” Rowe doesn’t use the English irrealis form “were” with a singular subject because the conditional isn’t contrary to fact: it’s what Smyth calls a “simple condition,” with no implication that the clause was contrary to fact–the possibility is open (though doubtful) that the dream images were in fact enjoining S. to compose poetry.
Guy Cooper makes comments of this sort all the time and I find them somewhat mystifying. For example he states that εἰ with ἄρα “may suggest an interesting possibility that is worthy of consideration” or “mark the bemusing quality of the Oratio Obliqua.”[1] The first part of this statement I understand but Oratio Obliqua is used by Cooper in many contexts where it isn’t at all obvious that there is any speech act involved. Does Oratio Obliqua have an extend meaning which includes text that isn’t a speech act of any sort?
I can see an implicit speech act in ταύτην τὴν μουσικήν μοι ἐπιτάττοι ποιεῖν and that perhaps is what Cooper is driving at. But we are talking about the interpretation of dreams functioning potentially (εἰ ἄρα + optative) as injunctions to some activity ποιεῖν ταύτην τὴν μουσικήν. So the speech act metaphor(?) is buried pretty deep and it doesn’t look like that is the Oratio Obliqua you are referring to.
after thought:
ἀλλ᾽ ἐνυπνίων τινῶν ἀποπειρώμενος τί λέγοι
this wording underlines the notion that the dreams amount to an act of speech τί λέγοι.
Thanks so much everyone for the help. Nate, perhaps my use of (probably British!) terminology was confusing. All my grammar books (Morwood Oxford and Anderson & Taylor BCP) divide conditionals up into two general categories: real and unreal. By “unreal” they mean anything that is less likely to happen, whether it is hypothetical (If X were to happen, Y would happen) or completely closed (If X had happened, Y would have happened). They also treat ἐαν/ἀν/ἠν + subjunctive protasis followed by apodosis in present not as a “general real conditional” like Goodwin and Smythe but as an indefinite temporal clause - ἐαν τις κλεπῃ, κολαζεται, “If anyone (ever) steals, he is punished” could equally be translated “Whenever anyone steals, he is punished”; likewise the past equivalent εἰ + optative (If anyone ever did X…) followed by imperfect is treated as an indefinite in the past. At any rate obviously neither of these “semi-conditionals” are what Plato has in mind here. So in this sense the only use of an optative in the protasis of conditions is, as far as I am aware, in a future unreal or less vivid or whatever you want to call it (If X were to happen, Y would happen). That’s why your brilliant explanations about it being in effect oratio obliqua must surely be right ? The original thought in his mind was something like “I will be purified from guilt if, after all, perhaps the dream figures are enjoining me to compose this sort of art/music” which when being related to us is all put into the optative. Surely, Bartholomew, it is not the ἐπιτάττοι which of itself is the reason why it is in the optative (that the dream figures presumably spoke to him to do it), but the indirect statement is implicit in the εἰ ἄρα? This seems similar to Latin where quod, which usually takes the indicative, may suddenly spring into the subjunctive since it is only the alleged reason of the speaker. Anyways, thanks for all the help again - if after all perhaps it seems to make sense .
P.S. This use of implied indirect statement I guess interestingly means that the optative does in fact work in terms of tense rather than aspect, since in its use in indirect statements it works in this way. So if Plato had used the aorist optative here, then that would have implied that Socrates’ deliberation in the past was about the then past and so would be translated with a pluperfect: “If after all perhaps it had enjoined me to…”.
if Plato had used the aorist optative here, then that would have implied that Socrates’ deliberation in the past was about the then past and so would be translated with a pluperfect: “If after all perhaps it had enjoined me to…”.
I think the difference between present and aorist optative would probably be a difference of aspect, not tense: in English translation, “were enjoining” vs. simply an untensed (in relation to the main verb) “enjoined.” There would not necessarily be any implication of anteriority if the aorist ἐπιτάξοι had been used, although our tendency as English speakers locked into the English tense system might be analyze the injunction as anterior, and to translate the aorist, had Plato used it, as an English pluperfect. But with πολλάκις he seems to be emphasizing the repeated, continuing nature of the dream visions; hence the “present” aorist.
But isn’t the optative used in terms of tense when used in indirect speech (or at least it is with ὁτι) - which is why a future optative exists and a future subjunctive does not? It would clarify things if we found a similar example with a different optative.
That appears to be what Cooper is driving at. It looks as if oratio obliqua is extend to include hypothetical situations in the form of thoughts, not necessarily thoughts about speech acts (which we have in this case), but just hypothetical scenarios, which to my way of thinking is too inclusive a definition of oratio obliqua. I am still confused about this, perhaps someone can clarify what is going on.
Think of it this way: Socrates slips into the optative with the εἰ ἄρα clause because he’s imagining how Kebes will report to Euenos Socrates’ statement or thought process explaining why he (Socrates) composed the poetry. We would put this in the past tense in English: “Tell him I composed the poems . . . as a way to discharge a sacred duty–if maybe the dream vision was repeatedly enjoining me to compose this sort of music.” (I’ve used the singular “dream vision” to avoid confusion between past tense “were” and irrealis “were.”)
Oh dear, I’m afraid I’ve got confused again! Before you suggested it was what Socrates was thinking - in the past about the then present - and that made sense but I don’t totally follow this new theory I’m afraid! For one, isn’t λέγε τοίνυν αὐτῳ (or indeed anything which will happen in the present or future) primary sequence? So you would not expect the optative at all at least for this reason - which you obviously don’t find in the rest of his speech save the two "εἰ ἀρα…"s. There must surely be a different reason, which could not apply since it does not apply to the rest of the speech, why the "εἰ ἀρα…"s suddenly take the optative, and your brilliant previous explanation, that they mark Socrates’ thoughts, in the optative because we are now in virtual historic sequence - (Socrates thought:) if after all… - seems very cogent!
There is another sense of πολλάκις after conditional particles which is frequent in Plato: perhaps. See LSJ III. This old commentary by Wilhelm Wagner (1880) has this gloss with a few references:
and if by any chance you have not given your attention to the subject,
we would remind you that it ought not to be neglected,…
→ W.R.M. Lamb. (Cambridge , 1955)
I’m sorry my explanation was confusing, and I’m not entirely sure it’s right It’s clear that this is not a contrary to fact condition–the εἰ ἄρα πολλάκις clause (thanks, Nate, for explicating πολλάκις) leaves open the possibility that the dream images were in fact telling Socrates to make music. But the optative does call for explanation.
I was thinking that Socrates uses the optative because he’s imagining how Kebes would frame what Kebes will tell Euenos Socrates said or thought. When he speaks to Euenos, Kebes will use a past-tense verb of speaking or thinking. Kebes might say something like: "Socrates said that he composed the poems because . . . "
But maybe the optative just indicates a more tentative “maybe” than an indicative would. (And why isn’t λέγει optative, too, if my first explanation has any validity?)
The MSS. has λέγοι Bekk.; λέγειν Bodl. with λέγει in the margin.
I’m not sure. Cooper’s treatment of εἰ with ἄρα as quoted by Stirling – that it “may suggest an interesting
possibility that is worthy of consideration” – would suggest an optative may be more suited here.
Limiting it further with πολλάκις with the sense of perchance, perhaps left me baffled at the sight
of an indicative in those quotes.
In Smyth 2796, however, nearly all the references have indicative. I don’t understand then what would
the indicative lend to this mere possibility.
I was also wondering about this so I invested my morning in looking into oratio obliqua, first in NT grammars and then Smyth and Cooper. Of the NT grammars A.T. Roberson was helpful by pointing out that NT Koine doesn’t use the optative very often in oratio obliqua. Moving to Cooper[1] the optative and infinitive are both used frequently in oratio obliqua. There is definitely a relationship between optative and oratio obliqua.
“The optative is often used in oratio obliqua after verbs of past tense to represent indicatives in the original. When this occurs the present optative of the O.O. represents a present indicative of the oratio recta O.R., or less commonly, an imperfect indicative of the O.R., an aorist optative of the O.O. represents an aorist indicative of the O.R., a perfect optative of the O.O. represents a perfect indicative of the O.R., the future optative of the O.O. represents a future indicative of the O.R. …” Guy Cooper Attic Greek Syntax 1:54.5.0 v1 p697, 1:53.7.10 v1 p. 658.
some references to oratio obliqua in Guy Cooper Attic Greek Syntax (v. 1-4):1:53.7.10 v1 p. 658, 1:54.5.0 v1 p697, 1:55.4.0 v1 p792, 1.65.1.0 v2 p1038, 2:54.5.0 v3 p2418, 2:55.4.0.i-iV v3 pp2533-2540, 2:65.1.0 v4 p2690.
Thanks, Stirling, for these references. So it seems in this particular passage,
indirect speech is the reason behind the use of the optative.
However, I’ve always been taught to distinguish between a true optative, conveying potentiality (+ἄν)
or (im)possibility, (im)probability and an optative standing in place of an original indicative.
My question is if εἰ ἄρα + original indicative/optative of indirect speech is the common
construction for this mere possibility, is it* equivalent in sense to a potential optative with ἵσως?
Certainly the sense is very similar and perhaps identical. Smyth mentioned this mere possibility as
being undesired, and sometimes full of irony. Perhaps herein lies the difference between the constructions, both of which are very much a Platonic trademark.
Your are right. This is a complicated subject. The potential optative in O.O. is special case. Not only that, but O.O. is frequently in the same mood & tense/aspect as O.R.. There is no obligation to change the mood to infinitive or optative. I’m just swimming in waters over my head trying to get a feel for idioms which Koine doesn’t commonly use. Just read another 10 pages of Cooper on this. 1:54.6.1-11, v1 pp698-710. Not easy reading.
The MSS. has λέγοι Bekk.; λέγειν Bodl. with λέγει in the margin.
One ms. (T, Cod. Venetus app. cl. 4.1, 10th c.) has λέγοι written in by a second hand, and Burnet (1905) adopts this reading against the other mss. (Probably Bekker, too, as a conjecture, judging from your citation.) Doesn’t that suggest they are taking both optatives as indirect discourse? There doesn’t seem to be any other reason for the optative in λέγοι. But the new OCT adopts λέγει.
Curiously, another ms. (W, Cod. Vindobonensis suppl. gr. 7, 11th c.) uniquely reads ἐπιτάττει here and προστάττει in 61a. These look like conjectures because they occur in both places where the other mss., including those from the same family, have optatives. Maybe the copyist puzzled over the optatives just like us.
One other thought: the grammar books are supposed to reflect the corpus of ancient Greek texts, not the other way around. Perhaps there’s a nuance to the optatives that the grammar books just missed. In the end, the meaning is reasonably clear, so maybe it’s best to let go of the problem (ancient texts are full of these) and move on!
I agree. I spend at least 100 hours of reading texts for every hour of reading grammars. Once in a while I take some time off from reading texts to read everything I can find on a topic which I don’t understand well like the optative. The distinctions which are presented as either/or in traditional grammar e.g. pure optative | potential optative sometimes turn out to be more trouble than they are worth. Categories leak, boundaries are fuzzy, not healthy to loose sleep over making everything fit into any particular framework because at some point, all the frameworks break down. I shouldn’t be posting in thread about the optative in Plato. I don’t read Plato.
Just in terms of the manuscripts, Rowe (whose edition I believe is the latest, but obviously not necessarily more accurate) has “λεγει” and has “λεγοι” as a “reading found in one of the secondary manuscripts, or inserted by a late or uncertain hand in a primary manuscript” which sounds like a conjecture to me. "λεγειν " is also down as a “reading found in one or more of the primary sources”.
After our discussion, to me the indirect speech reason for it seems the most likely, since give the sense I’m convinced you would not expect it for any other reason. Furthermore in Classical Greek it’s quite common to find the optative being used in conditionals in indirect speech: “The normal rules for subordinate clauses in indirect speech apply i.e. if the main verb is primary the verb in the protasis remains unchanged. If the main verb is historic, either the mood and tense remain unchanged, or the original tense is retained but the mood becomes optative. The historic tenses of the indicative should not be changed to the optative; and if an original subjunctive is so changed, εἰ must replace ἐαν” - Stephen Anderson, Writing Greek. The only thing that really bugs me is how most of the sentence does not use the optative, yet the εἰ clause does. There must be something unique to the εἰ - unless indicative and optative can just freely mix together; but then that is not entirely satisfactory since the imperative introduces the rest of it and that would be primary sequence. Either one of your ideas - (a) that it represents what Cebes will speak to Evenus, or (b) that this is what he thought in the past about the then present seem to satisfy the problem. The conjecture “λεγοι” and the “ἐπιτάττει” + “προστάττει” seem seem to me also to good reasons for supposing it… Yet I will go and ask someone who is almost certain to know about it and let you know what they say…
P.S. Interesting enough is this WGG § 1498 “One verb may be changed to the optative while another is retained”.
? The original thought in his mind was something like “I will be purified from guilt if, after all, perhaps the dream figures are enjoining me to compose this sort of art/music” which when being related to us is all put into the optative. Surely, Bartholomew, it is not the ἐπιτάττοι which of itself is the reason why it is in the optative (that the dream figures presumably spoke to him to do it), but the indirect statement is implicit in the εἰ ἄρα? This seems similar to Latin where quod, which usually takes the indicative, may suddenly spring into the subjunctive since it is only the alleged reason of the speaker. Anyways, thanks for all the help again - if after all perhaps it seems to make sense 
.