The Relation of the Optative to the Subjunctive and Other Moods.In the chapter on the general view of the moods, no attempt was made to assign to either the subjunctive or the optative a single “fundamental idea” from which all the uses of the mood could be derived, except so far as the idea of futurity was shown to belong essentially to the subjunctive in all its most primitive uses. It would be impossible to include under one fundamental idea all the actual uses of any mood in Greek, except the imperative; for even the indicative is used to express unfulfilled conditions, unaccomplished wishes, and unattained purposes, none of which can be brought under the ideas of “declaration” or “absolute assertion” commonly attributed to this mood. Again, it is not to be expected that the true fundamental idea of any mood should include all its uses in a developed language; for the fortunes of language often depend on causes which are quite independent of the original essence of the forms employed, and which seldom can be referred to invariable laws of thought. The same idea can be expressed in two cognate languages by different moods: as he would have seen is εἶδεν ἄν in Greek and vidisset in Latin, while in Sanskrit it would be expressed by a past augmented future equivalent to the Greek ἔμελλεν ὄψεσθαι (see § 428). Even within the Greek itself, we have if he were wise expressed by εἰ σοφὸς εἴη in Homer and by εἰ σοφὸς ἦν in Attic; and in Homer, both οὐκ ἂν ἔγνως and οὐκ ἂν γνοίης can mean you would not have discerned, while the latter can mean also you would not discern (in the same future sense as in Attic). One doctrine of the original meaning of the Greek subjunctive and optative has gained such general approval of late, that it is entitled to special consideration. This teaches that the fundamental idea of the subjunctive is will, and that of the optative is wish. In the subjunctive, the idea of will appears especially in exhortations and prohibitions and in expressions of purpose. It can also be used to explain the subjunctive in protasis, by understanding ἔλθῃ in ἢν ἔλθῃ to mean originally let him go, suppose him to go (in some case). But before we can decide that will is the fundamental idea of the subjunctive, or even that it is a necessary and essential part of the idea of this mood, we must ask, first, whether it is essential to those uses of the subjunctive which we have a right on other grounds to call the most primitive; and, secondly, whether there is any other idea equally essential and equally primitive, from which the idea of will could have been evolved more simply and naturally than this could have been evolved from the idea of will. The subjunctive nowhere bears more distinct marks of primitive simplicity than when it appears in Homer as a simple future; as in οὐ γάρ πω τοίους ἴδον ἀνέρας οὐδὲ ἴδωμαι, for never yet have I seen such men, nor shall I ever see them, Il. i. 262, and in καί ποτέ τις εἴπῃσιν, and some one will say, Il. vi. 459, followed by ὥς ποτέ τις ἐρέει in vs. 462, referring to the same thing. See other examples in § 284. In this sense it is negatived by οὐ, like an indicative; and it may be modified by κέ or ἄν, like the future indicative in Homer, and thus acquire a potential sense (see § § 285 and 286). It is seldom that any modal form (except a plain indicative) is found so free from associations which might affect its meaning and conceal its original character. It has, moreover, its exact counterpart in Sanskrit in the Vedic subjunctive, which is negatived by na/, the equivalent of οὐ.1 This simple subjunctive has no element of will. It expresses what the speaker regrets as readily as what he is resolved to do. Thus in both the examples above quoted, the subjunctive expresses an act which is decidedly contrary to the speaker's will and wish. This subjunctive and the future indicative run parallel in all their constructions, and the former expresses will only so far as the latter does. The only character that is beyond question in this subjunctive is its reference to future time, and if we were left to this use alone, we should have no hesitation in designating the subjunctive as a form expressing futurity like a future tense. As this use cannot be deduced from the subjunctive as an expression of will, let us see whether the opposite process, the evolution from the simple future meaning of the uses in which will appears, is any easier and does any less violence to the principles of the language. The use of the subjunctive which strikes every one as coming next in simplicity to the Homeric construction just described is seen in exhortations, like ἴωμεν, let us go, and (in its negative form) in prohibitions, like μὴ ἴωμεν, let us not go, μὴ εἴπητε τοῦτο, do not say this. This use of the subjunctive is found also in Sanskrit, and its negative is there generally (though not always) mA, the equivalent of μή. It thus appears that the marked distinction which is seen in the early Greek between ἴωμεν, we shall go, and ἴωμεν, let us go, in both positive and negative forms, was probably inherited from an ancestral language, so that we need not seek for the development of this distinction within the Greek itself. It is obvious that the future element is equally strong in both expressions, while the hortatory subjunctive also expresses will. Now it is much more natural to suppose that a future form expressing exhortation or prohibition originated in a form expressing mere futurity, than that the merely future form originated in the exhortation or prohibition. We cannot derive οὐκ ἴδωμαι, I shall not see, from μὴ ἴδωμαι, let me not see. But it is by no means impossible that, in some language which was a common ancestor of Greek, Latin, and Sanskrit, subjunctive (i.e. originally future) forms came to be used to express both commands and prohibitions; that, when these imperative expressions became distinguished from the subjunctive in its ordinary future sense, they adopted the negative (the ancestor of mâ ' and μή) which was used with similar imperative forms, though this use of the negative might not at first be very rigid; and that thus μή ἴωμεν, in the sense let us not go, became established in early Greek as opposed to οὐκ ἴωμεν, we shall not go. In Sanskrit, however, the use of mA/ in such cases was less fixed, and here na/ (the equivalent of οὐ) is sometimes found with the subjunctive in prohibitions.2 This last is what we should have if in χειρὶ δ᾽ οὐ φαύσεις ποτέ, you shall never touch me, EUR. Med. 1320, we could substitute an Homeric subjunctive (e.g. ψαύσῃς) for the future indicative. The cases of μή with the future in prohibitions given in § 70, like μὴ βουλήσεσθε εἰδέναι, do not wish to know, DEM. xxiii. 117, are too few to be of much weight in the discussion; but they seem to show an abortive tendency to establish the future indicative with μή by the side of the subjunctive in prohibitions. What the future could do in an imperative sense is shown by examples like “πάντως δὲ τοῦτο δράσεις,” “but by all means do this,” AR. Nub. 1352 , and others quoted in § 69; but the natural negative here was οὐ, not μή, as in οὐ ψαύσεις above. If the origin of the interrogative subjunctive in appeals (§ 287) and of its negative μή has been correctly explained in §§ 288 and 291, this is merely an interrogative form of the subjunctive in exhortations and prohibitions, and calls for no special discussion here. The origin of the use of the subjunctive with οὐ μή is still to uncertain to give this construction much weight in determining the essential character of the subjunctive. If the view of this construction which is advocated in this work (see Appendix II.) is accepted, the form is an offshoot of the prohibitory subjunctive. If it is thought to be an original construction, expressing a strong denial or prohibition by its own force, the subjunctive appears in its original future force. Whatever theory we may have of the origin of this subjunctive, the form is interchangeable in use with the future indicative. In dependent sentences, the subjunctive is used in two constructions, —in so-called final clauses, and in conditional sentences. In negative final constructions with μή, the subjunctive was originally prohibitive (§§ 262, 307); in positive clauses with the final particles, it expresses something aimed at, that is, an object of will. But here, as in independent sentences, to derive the more complex from the more simple is far more natural than the reverse. Further, in all final constructions the future indicative may be used in the same sense as the subjunctive; this could hardly be done if the subjunctive contained an essential element of will which is wanting in the future. Again, the subjunctive is very common in final constructions after past tenses, where the optative is the regular form (318); it cannot be supposed that the idea of will is present in such final clauses when they have the subjunctive (as they generally do in Thucydides) and is absent when they have the optative (as is more common in Xenophon). In conditional sentences, although we may explain the subjunctive as originally hortatory, ἢν ἔλθῃ meaning let him come (we will suppose), it is more natural to refer this use to the primitive use of the subjunctive as a simple future, εἴ κεν ἔλθῃ (or εἰ ἔλθῃ), in case he shall come, making a supposition of a future event of which the Homeric ἔλθῃ, he will come, might make a statement (see §§ 11 and 398). We thus avoid the necessity of explaining the indicative and the subjunctive in protasis on different principles. As each of the various tenses of the indicative with εἰ expresses a supposition in the time which it naturally denotes (§ 3, c), so the subjunctive is a natural form to express a future supposition. Thus, as εἰ γενήσεται τοῦτο supposes what γενήσεται τοῦτο states, εἰ γένηται τοῦτο naturally supposes what (in the older language) γένηται τοῦτο, this will happen, states. As the former cannot be explained by the idea of will, it seems unnecessary and illogical to introduce this idea to account for the latter. What has been said of ordinary conditional sentences applies also to relative conditions. The only use of the subjunctive in conditions which cannot be derived from the simple future meaning is that in general suppositions; but the undeveloped state of this construction in Homer and other considerations make it highly probable, if not certain, that this is a use of the subjunctive which grew up within the Greek language itself at a comparatively late period, and that it is not one of the primitive uses of the mood. (See §§ 11, b, 400, 401.) It is certain that no trace of the subjunctive as a mood of will can be seen in its actual use in conditional sentences. Thus ἢν τὴν πόλιν ἕλωσι could always be said as properly by the friends as by the enemies of a city, by the besieged as well as by the besiegers. In Il. iii. 71, ὁππότερός κε νικήσῃ, spoken by Priam, is, as an expression, perfectly neutral as regards the hope or desire of victory. It may be said with truth, that the primitive meaning of a verbal form is apt to be weakened, or even to disappear, in actual use. But is it logical to assume a lost meaning to account for an expression, when the meaning which remains accounts for it satisfactorily without external help? When we find ἢν ἕλωσι τὴν πόλιν actually expressing a mere future supposition, with no idea of will, in all periods of the language, and when we find ἕλωσι meaning they will capture in the earliest period that we know, why should we assume an original idea of will (which was afterwards lost) in ἢν ἕλωσι to account for its actual meaning? The view of the conditional sentence here adopted is confirmed by paratactic conditions like the following: θύσεις δὲ τὴν παῖδ᾽ : ἔνθα τίνας εὐχὰς ἐρεῖς; I. A. 1185, where θύσεις makes a supposition, supposing you shall sacrifice the girl, which would generally be expressed by εἰ θύσεις or ἢν θύσῃς: so ἀδικεῖ τις ἑκών and ἐξήμαρτέ τις ἄκων, both expressing suppositions, DEM. xviii. 274.3 On these grounds we may feel justified in regarding the subjunctive as originally and essentially a form for expressing future time, which the Greek inherited, with its subdivision into an absolute future negatived by οὐ and a hortatory future negatived by μή, and used in independent sentences. The name optative mood (ἔγκλισις εὐκτική), which was invented by grammarians long after the usages of the language were settled, designated the mood by the only use which it then had in independent sentences without ἄν, that of wishing. It is evident that this name in itself is no ground for assuming that wishing was the primitive function, or even an essential function, of the optative, any more than the name of the subjunctive (ἔγκλισις ὑποτακτική) would lead us to assume dependence as an original or necessary characteristic of that mood. We have already mentioned the theory that the optative is the mood of wish, as the complement of that which makes the subjunctive the mood of will. This theory finds no support in the potential use of the optative with or without κέ or ἄν, which is the only independent use of the optative except in wishes and exhortations. Surely ἀπόλοιτο ἄν, he would perish, can never have been developed from ἀπόλοιτο, may be perish, for the former is no more likely to be said by one who wishes the death of a person than by one who fears it, and there is nothing in the addition of ἄν or κέ which can reasonably be supposed to change a form, which in itself expresses wish, to a neutral form or even to one expressing what is feared. The fundamental distinction in negative sentences between μὴ ἀπόλοιτο and οὐκ ἂν ἀπόλοιτο (or οὐκ ἀπόλοιτο) is still more significant. Nor can any support for the theory be found in dependent final constructions or in indirect discourse. No one would see a distinction of will and wish in ἴδῃ and ἴδοι in ἔρχεται ἵνα ἴδῃ τοῦτο and ἦλθεν ἵνα ἴδοι τοῦτο, or in φοβοῦμαι μὴ ἔλθῃ and ἐφοβήθην μὴ ἔλθοι,—not to speak of ἦλθεν ἵνα ἴδοι τοῦτο and ἦλθεν ἵνα ἴδῃ τοῦτο. Still less would any one dream of looking for wish in the optative in εἶπεν ὅτι ἔλθοι, he said that he had come, or in ἤρετο εἴ τις εἴη σοφώτερος. In all these dependent constructions, the optative is only the representative of the subjunctive or indicative when these are, as it were, transferred to the past by depending on a verb of past time; but, if wish were the fundamental idea of the optative, we should hardly expect this to vanish so utterly, since the essential character of the optative would naturally be especially marked where it is used by a fixed principle of the language as a substitute for an indicative or a subjunctive. The only strong argument for the theory that the optative is primarily the mood of wish is found in the optative with εἰ in protasis. It is maintained that a gradual development of this conditional form from the simple optative in a wish can be actually seen in Homer. The strongest and most attractive statement of this argument is given by Lange in his elaborate, but unfortunately unfinished, treatise on the particle εἰ in Homer.4 Delbrück's treatment of the optative in his Syntaktische Forschungen, vol. i., is based on this doctrine. When Lange states (p. 485) that, of 200 examples of εἰ with the optative in Homer, 136 are expressions of wish, the majority seems decisive; although we may even here withhold our judgment until we examine the majority and also see what the minority of 64 have to say. The majority of 136 is made up as follows:— 1. Ordinary wishes with εἰ γάρ, εἴθε (αἲ γάρ, αἴθε), or εἰ, like αἴθ᾽ οὕτως, Εὔμαιε, φίλος Διὶ πατρὶ γένοιτο, Od. xiv. 440; αἲ γὰρ οὕτως εἴη, Il. iv. 189; εἴθ᾽ ὣς ἡβώοιμι, βίη δέ μοι ἔμπεδος εἴη, Il. xi. 670. (Of these there are 38 cases.) 2. Cases in which a wish with εἰ and the optative (like the expressions just quoted) is followed by an apodosis expressing a consequence which would follow the fulfilment of the wish. Thus the last example in 1 appears in Il. vii. 157 with such an apodosis:— εἴθ᾽ ὣς ἡβώοιμι, βίη δέ μοι ἔμπεδος εἴη: τῷ κε τάχ᾽ ἀντήσειε μάχης κορυθαίολος Ἕκτωρ. If we put a comma at the end of the first verse, we have a full conditional sentence. In many cases it is doubtful which punctuation is correct. Lange includes under this head even such sentences as Il. vii. 28, ἀλλ᾽ εἴ μοί τι πίθοιο, τό κεν πολὺ κέρδιον εἴη, and Od. xx. 381. (Of these there are 28 cases.) 3. Ordinary conditional sentences, in which the fusion between the optative with εἰ expressing a wish (i.e. supposing something that is desired) and a following apodosis with κέ or ἄν is said to be complete, as in Il. xiii. 485:— εἰ γὰρ ὁμηλικίη γε γενοίμεθα τῷδ᾽ ἐπὶ θυμῷ, αἶψά κεν ἠὲ φέροιτο μέγα κράτος ἤ κε φεροίμην. (Of these there are 19 cases, against 18 otherwise similar cases in which the optative with εἰ supposes something not desired.) 4. Cases of which the following are examples:— ἤλυθον, εἴ τινά μοι κληηδόνα πατρὸς ἐνίσποις, Od. iv. 317. πάπτηνεν δ᾽ ἀνὰ πύργον Ἀχαιῶν, εἴ τιν᾽ ἴδοιτο ἡγεμόνων, ὅς τίς οἱ ἀρὴν ἑτάροισιν ἀμύναι, Il. xii. 333. Such examples are variously explained, but the protasis generally refers to something that is desired. (Of these there are 43 cases.) 5. Ordinary conditional sentences in which εἰ with the optative expressing a wish follows an apodosis; as in Il. xxii. 20, ἦ σ᾽ ἂν τισαίμην, εἴ μοι δύναμίς γε παρείη. These differ from those in 3 only in the position of the protasis. (Of these there are 8 cases of wishes, against 33 in which no wish is implied, of which last 17 are concessive.) The minority of 64 examples, in which εἰ with the optative does not express a wish, is made up of the 18 dissenting cases under 3, the 33 under 5 which contain no wishes, 5 exceptional cases (as Lange views them) under 4 (2 with doubtful readings), and 8 cases of ὡς εἰ with the optative in similes, like ἴσαν ὡς εἴ τε πυρὶ χθὼν πᾶσα νέμοιτο, Il. ii. 780. It will be seen that the strength of the argument lies in the gradual development of the optative conditional sentence which is supposed to appear in 1, 2, and 3. This is further enforced by reference to cases in which the simple optative in a wish (without any form of εἰ) is followed by an apodosis, like the equivalent optatives with εἰ in 2, thus showing the absence of a conditional force in the latter. See Od. xv. 180:— οὕτω νῦν Ζεὺς θείη, ἐρίγδουπος πόσις Ἥρης: τῷ κέν τοι καὶ κεῖθι θεῷ ὣς εὐχετοῴμην. Since the two clauses are grammatically independent here, it is argued that they must be equally so in the examples in 2. The whole argument is based on the important assumption that the optative with εἰ, εἰ γάρ, etc. in a wish is the same in origin with the simple optative in a wish, so that εἰ γένοιτο τοῦτο and γένοιτο τοῦτο both come to mean may this be done in the same way, by a wishing power inherent in the optative itself; and from this it is argued that εἰ γένοιτο τοῦτο as a protasis is used in a more primitive and natural sense when what is supposed is desired by the speaker than when it is not. Unless we assume this as proved, and reject the opposite alternative which makes the optative with εἰ in a wish a protasis with a suppressed apodosis, we have no right to count the examples in 1 and 2 as evidence that the optative with εἰ denotes a wish by its own nature; for it would be reasoning in a circle to quote these as proof that the optative itself denotes wish, in a discussion which aims at establishing the nature and meaning of the optative in these very expressions. Again, the real nature of the 43 conditions with εἰ and the optative in 4 is in question in this discussion; and it is inadmissible here to assume at the outset that they express wish in themselves and then to use them as evidence that wishing is the original function of the optative. Proof is needed, therefore, that the optatives in 1, 2, and 4 (that is, in 109 of the 136 wishing optatives in Homer) actually express wish by their own force, so that they can properly be used as independent testimony here. Until at least a reasonable presumption in favour of this view is established, we are without evidence that there is any such gradual development of the optative condition as is claimed. We must therefore depend at present on the only cases about which no doubts exist, the complete conditional sentences in 3 and 5, to determine whether the optative with εἰ involves the idea of wish without regard to the nature of its apodosis. If it should be found that the idea of wish preponderates in these optatives, we should have a convincing proof that the same is true of the optatives in 1, 2, and 4, whether these are viewed as protases or as original wishes. A slight inspection of Lange's statistics will show that the question is not to be settled in this simple way. Of the 37 optatives in 3, 19 suppose something that is wished for, while 18 do the opposite. Of the 41 in 5, only 8 suppose desirable things, while 33 do not. Therefore, in the 78 plain cases of εἰ with the optative in conditions in Homer, we find only 27 expressing wishes. If we confine ourselves to the cases in 3, where the protasis precedes, we find as equal a division as is possible (19 : 18), showing very plainly that even here wish has nothing whatever to do with the form of expression. Indeed, if we take εἰ with the optative in protasis by itself, what is there to indicate that it involves a wish? It cannot be doubted that this form is the equivalent of the English if he should go and if we should see him; and who would attempt to find any such idea as wish in these expressions? Unless we are prepared to maintain that if we should be saved expresses the original idea of the English construction better than if we should perish, we must be slow to assert that εἰ σωθεῖμεν gives the spirit of the Greek optative better than εἰ ἀποθάνοιμεν. We must remember also the large class of conditional relative sentences which have the optative. This optative cannot be explained on any different principle from the optative with εἰ, and yet who would profess to find anything like the idea of wish in ὅ τις ῥέζοι, Od. i. 47, ᾧ μὴ εἴη, Od. xi. 490, or in Il. vi. 330, Il. 521, Il. xiii. 344, Il. xiv. 248? I give the first six examples that I meet. It is obvious at once that we must recur to the examples in 1 and 2, and see whether these establish any such strong presumption as will justify us in making wish the fundamental idea of the optative with εἰ, notwithstanding the fact that a large majority of the optatives in protasis in Homer have a contrary meaning. In dealing with the examples in 1 and 2, it will be assumed that εἰ, εἴθε, εἰ γάρ, and αἰ, αἴθε, αἲ γάρ all have the same origin, and involve the same particle εἰ or αἰ which is used in protasis.5 The question in regard to the wishes in 1 amounts to this: is it more probable that the optative here is merely the wishing optative, preceded by a sort of exclamatory particle εἰ,6 so that γένοιτο and εἰ γένοιτο are merely different forms of an exclamation, O may it be done!—or that εἰ γένοιτο in a wish is the same as εἰ γένοιτο in protasis, meaning if it should only be done, deriving its force as a wish from the unconscious suppression of an apodosis like how happy I should be or it would be well? The difficulty of explaining εἰ in an ordinary protasis like εἰ ἦλθεν, if he came, as in any sense exclamatory is a great obstacle in the way of Lange's view; but his alternative is equally hard, to make εἰ in a wish radically different from εἰ in a protasis. In the incomplete state of Lange's work, it is impossible to see how successfully he would have surmounted this difficulty. But, apart from this, we are compelled on his theory to believe that the parallel construction of εἰ γάρ and εἴθε with the past tenses of the indicative in wishes is radically different in principle from that of εἰ etc. with the optative. The former is a later construction; but is it possible that the traditions of so fixed an expression as εἰ with the optative in wishes could have so utterly vanished that, while εἰ γὰρ γένοιτο, may it be done, had no conditional force, εἰ γὰρ ἐγένετο τοῦτο, O that this had been done, was felt as conditional? It is impossible to explain εἰ γὰρ ἐγένετο except as an elliptical protasis, since there is no form of wish like ἐγένετο (alone) corresponding to γένοιτο, may it be done. Even if we could suppose that εἰ γὰρ ἐγένετο was formed ignorantly on the analogy of εἰ γὰρ γένοιτο, it would be incredible that μὴ γένοιτο should not have engendered a corresponding μὴ ἐγένετο. But why is it thought necessary or probable that γένοιτο and εἰ γὰρ γένοιτο should have had the same origin? If we can trust our feelings in the use of our own language, it is beyond doubt that our expressions of wish, like may help come and O if help should (or would) come! are entirely independent constructions, and also that the latter is a condition with its conclusion suppressed. Why should we not accept the same simple distinction in the Greek forms, and admit that the Greek had two ways of expressing a future wish, one by the simple optative, the other by a protasis with its apodosis suppressed? Absolute proof is, of course, impossible in such a case; but it is surely safe to maintain that no such strong presumption is established in favour of identity of construction in γένοιτο and εἰ γένοιτο in wishes, as to make it probable that εἰ γένοιτο in protasis was originally a form of wish, in face of the fact that only a small proportion of Homer's undoubted protases with εἰ and the optative express wishes. But it may be said that the peculiar examples of half-formed conditional sentences in 2 (p. 376) establish the theory of the development of the conditional optative out of a wish. But this connecting link loses its value, when it is seen that it connects merely one construction, in which the wishing force of the optative is at least questionable, with another in which there is no positive evidence of any wishing force at all. If the ordinary theory of the suppression of an apodosis with εἰ γὰρ γένοιτο in a wish is correct, we must suppose that the suppressed apodosis was seldom felt in a definite form of words any more than it is with our O if he would come. But it might sometimes happen that an actual expression of a definite result of the fulfilment of a wish would suit the case better than the uncertain reference to a fulfilment, which the mere clause with if suggests. We have an excellent illustration of this when a wish is repeated as a protasis in almost the same words, and is then followed by an apodosis. See Od. iii. 217-223 (quoted in § 730), where εἰ γάρ σ᾽ ὣς ἐθέλοι is first a simple wish, and then is repeated as εἴ σ᾽ οὕτως ἐθέλοι, with the apodosis τῷ κέν τις, etc. naturally following. The oft-recurring verse εἴθ᾽ ὣς ἡβώοιμι, βίη δέ μοι ἔμπεδος εἴη appears in Il. xi. 670, Il. xxiii. 629, and Od. xiv. 468 (if Bekker is right in omitting vss. 503-506) as a simple wish with no addition; but in Il. vii. 157 it stands as a repetition of the wish contained in vss. 132, Il. 133, αἲ γὰρ ἡβῷμ᾽ ὡς, etc., and is followed by the apodosis τῷ κε τάχ᾽ ἀντήσειε μάχης κορυθαίολος Ἕκτωρ. In the other examples, we have simply the wish O if I were young again, with its vague unexpressed apodosis; but in Il. vii. 157 the result is expressed in the definite form, then would Hector meet his match. See Od. xvii. 496 and xv. 536 (quoted in § 730), in both of which a definite apodosis expressing a result takes the place of the usual suppressed conclusion. A distinction of optatives with εἰ into wishes and suppositions, based on the wishing or non-wishing nature of the verb, is often arbitrary. Thus Lange quotes, among his “paratactic” wishes followed by an apodosis in a distinct sentence (that is, half-developed conditional sentences), Il. xvii. 102:— εἰ δέ που Αἴαντός γε βοὴν ἀγαθοῖο πυθοίμην,
ἄμφω κ᾽ αὖτις ἰόντες ἐπιμνησαίμεθα χάρμης,
while he gives as an ordinary conditional sentence Il. xxiv. 653:— τῶν εἴ τίς σε ἴδοιτο θοὴν διὰ νύκτα μέλαιναν,
αὐτίκ᾽ ἂν ἐξείποι Ἀγαμέμνονι ποιμένι λαῶν.
His ground for distinction is merely that the former expresses a wish, while the latter does not. Even if both sentences were held to be simply conditional (as they probably are), it would still be claimed that the optative is used in a more legitimate and primitive sense in the former than in the latter. But is not the patent fact that there is really no essential distinction between these two optatives with εἰ (taken as conditions) a strong argument against the whole doctrine which derives the optative in protasis from the optative in wishes? As to the 43 examples in 4, in which the optative with εἰ obviously stands without any expressed apodosis, I must refer to the discussion of these in § § 486-493, where they are explained as protases which contain within themselves an implied clause of purpose as the apodosis. Whoever will compare the examples of the optative in § 488 with those of the subjunctive in § 487, or those of the optative in Delbrü ck's Conjunctiv und Optativ, pp. 236-238, with those of the subjunctive in pp. 171-175, will probably be satisfied that the greater part of these optatives represent original subjunctives, which are regularly used in this sense after primary tenses, while the original optatives that occur after primary tenses in this construction are not more frequent than they are in ordinary protasis in Homer (see §§ 499-501). Thus βῆ Πάνδαρον διζήμενος εἴ που ἐφεύροι, he went seeking Pandarus, in case he should find him anywhere (i.e. to find P. if haply he might), Il. v. 167, represents an original form βαίνω Πάνδαρον διζήμενος, ἤν που ἐφεύρω. This is true, whatever theory we hold as to the nature of the condition here. Again, this form is equally adapted to suppositions which are not objects of wish or desire; as in THUC. vi. 100, πρὸς τὴν πόλιν, εἰ ἐπιβοηθοῖεν, ἐχώρουν, they marched towards the city, in case the enemy should rush out (to be ready to meet them if they should rush out). So in Hom. Od. xxii. 381:— πάπτηνεν δ᾽ Ὀδυσεὺς καθ᾽ ἑὸν δόμον, εἴ τις ἔτ᾽ ἀνδρῶν
ζωὸς ὑποκλοπέοιτο ἀλύσκων κῆρα μέλαιναν,
where Ulysses is said to have searched the house, in case any one of the suitors should still be alive and be concealed (i.e. to find any such). This is quite as natural an expression as Il. xii. 333, πάπτηνεν εἴ τιν᾽ ἴδοιτο ἡγεμόνων, where the protasis supposes something desired. The idea of purpose which these sentences imply makes it natural that the supposition should be a desirable one in the majority of cases; but no independent support for the theory we are discussing can be found in them. We come then to the following conclusions. The theory that wish is the fundamental idea of the optative finds no support in conditional sentences with εἰ and the optative in Homer, for among 78 full sentences of this class, only 27 express suppositions which are desired by the speaker. The other optatives with εἰ which are said to express wishes stand without apodosis, and the nature of these expressions is itself in question in this discussion. As the presence of the idea of wish in the optative in ordinary conditions would have been a strong proof that the same idea is inherent in these other optatives, so the conspicuous absence of wish in the former creates a presumption against its existence in the latter; for it appears that, even if the optative with εἰ in wishes does express the wish by its own natural force, this force has not passed over into the ordinary optative in protasis, even in Homer. We have to consider, therefore, whether in spite of this presumption it can be established that the optative is the mood of wish, or that the two forms of optative in wishes (with and without εἰ) are identical in origin and construction. The theory of their identity obliges us to believe that εἰ is a sort of exclamatory particle; whereas the older view, which has the authority of Aristarchus (§ 723), that the optative with εἰ in wishes is a protasis with a suppressed apodosis, avoids this difficulty by making the form of wish the same as that of protasis. The new theory also compels us to explain the past tenses of the indicative with εἰ and the optative with εἰ in wishes on different principles. The cases in 2 (p. 376) of an optative with εἰ in a wish followed by an apodosis in a separate sentence are easily explained by supposing an actual apodosis to be expressed in them, where commonly only a general idea of satisfaction (like καλῶς ἂν ἔχοι) is understood. The cases of εἰ with the optative without an apodosis in 4 are to be explained by the implied apodosis: they are not necessarily expressions of desire, and the optative here generally represents an original subjunctive. As a negative result, we do not find in the Homeric examples as a whole any satisfactory proof that wish is the fundamental idea, or even an essential idea, of the optative. For the original meaning of the optative we must go, not to the developed wish, still less to the developed potential construction with ἄν or to the protasis with εἰ, but rather to certain simpler and less decided expressions, a few of which remain in Homer. In Il. iv. 17-19 we have a full conditional sentence, εἰ δ᾽ αὖ πως τόδε πᾶσι φίλον καὶ ἡδὺ πέλοιτο,
ἦ τοι μὲν οἰκέοιτο πόλις Πριάμοιο ἄνακτος,
αὖτις δ᾽ Ἀργείην Ἑλένην Μενέλαος ἄγοιτο.
This may be translated, and if moreover this should be welcome and pleasing to all, king Priam's city may continue to be a dwelling-place, and Menelaus may take Argive Helen home again. But οἰκέοιτο and ἄγοιτο (without κέ or ἄν) here do not make the usual potential apodosis, nor do they express a wish; and yet a very slight change in the thought would make them either of these. With κέ or ἄν added, the meaning would be Priam's city would continue to be, etc.; without ἄν, in the ordinary language it would be may Priam's city continue to be, etc. The same general result happens to be expressed in other passages in various ways. In Il. iii. 71-75 Paris proposes the duel with Menelaus, and says:— ὁππότερος δέ κε νικήσῃ κρείσσων τε γένηται,
κτήμαθ᾽ ἑλὼν ἐὺ πάντα γυναῖκά τε οἴκαδ᾽ ἀγέσθω
οἱ δ᾽ ἄλλοι φιλότητα καὶ ὅρκια πιστὰ ταμόντες
ναίοιτε Τροίην ἐριβώλακα, τοὶ δὲ νεέσθων
Ἄργος ἐς ἱππόβοτον.
Here ἀγέσθω is used with the same general idea in mind as ἄγοιτο in iv. 19, and ναίοιτε is like οἰκέοιτο. This example would rather lead us to understand both ἄγοιτο and οἰκέοιτο as wishes. But in iii. 255 we have τῳ δέ κε νικήσαντι γυνὴ καὶ κτήμαθ᾽ ἕποιτο, where τῷ νικήσαντι is equivalent to ὁππότερός κε νικήσῃ in 71, and ἕποιτό κε is potential, though expressing the same general idea as ἀγέσθω and ἄγοιτο above. Also, in iii. 256 we have ναίοιμεν (like ναίοιτε in 74) and νέονται (as future). Again, in iii. 138 Iris says to Helen τῷ δέ κε νικήσαντι φίλη κεκλήσῃ ἄκοιτις, where κεκλήσῃ κε is potential, referring to the same result as ἕποιτό κε, ἄγοιτο, and ἀγέσθω. These passages show a use of the optative without κέ which comes very near to that of the optative with κέ, and also to that of the imperative and of the future (with and without κέ). This neutral use of the optative is generally called “concessive.” In other cases, the optative without κέ has a more decided potential force; as in Il. xxiii. 151, νῦν δ᾽ ἐπεὶ οὐ νέομαί γε φίλην ἐς πατρίδα γαῖαν, Πατρόκλῳ ἥρωι κόμην ὀπάσαιμι φέρεσθαι, I would fain send. So in Il. xv. 45, αὐτάρ τοι καὶ κείνῳ ἐγὼ παραμυθησαιμην, I should advise him. In Il. xxi. 274, ἔπειτα δὲ καί τι πάθοιμι may be either then let me suffer anything (i.e. let me perish), or then would I suffer anything: that the latter is the true meaning is made more probable by xix. 321, οὐ μὲν γάρ τι κακώτερον ἄλλο πάθοιμι, for nothing else that is worse could I suffer, where οὐ shows that the optative is potential. On the other hand, in Il. xxiv. 148, μηδέ τις ἄλλος ἅμα Τρώων ἴτω ἀνήρ: κῆρύξ τίς οἱ ἕποιτο γεραίτερος, i.e. let no other of the Trojans go with him; only let an elder herald accompany him (or a herald may accompany him), the general sense and the preceding imperative seem to show that ἕποιτο is hortatory. Compare Il. iii. 407, μηδ᾽ ἔτι σοῖσι πόδεσσιν ὑποστρέψειας Ὄλυμπον, between two pairs of imperatives, where μηδέ shows the nature of the expression. Again, in Il. vi. 164, τεθναίης, ὦ Προῖτ̓, ἢ κάκτανε Βελλεροφόντην, we may doubt whether τεθναίης means you must die or may you die (i.e. die), although the connexion with κάκτανε leads us to the latter interpretation: here also compare Il. iii. 407.The tendency is not very strong in either direction in these passages, as is plain from the difficulty which we sometimes feel in deciding which the direction actually is in a given case.7 But as the potential and the wishing forms are generally clearly distinguished in Homer, we must look upon the few neutral expressions that we find as relics of an earlier stage of the language, in which the optative without κέ or ἄν was freely used in the sense of οἰκέοιτο and ἄγοιτο in Il. iv. 18, Il. 19.Such expressions could not be used in negative sentences, at least after οὐ and μή were established in their regular force, as the use of either negative would at once decide the character of the sentence. In the earlier language ἔλθοιμι and ἴδοιμι, I may go and I may see, probably corresponded to the subjunctives ἔλθω and ἴδω, I shall go and I shall see, as weaker forms for expressing future time. But both moods had inherited another use, by which ἔλθω and ἴδω meant let me go and let me see, while ἔλθοιμι and ἴδοιμι meant may I go and may I see. The reasons given above, for thinking a derivation of the hortatory subjunctive from the simple future expression more probable than the reverse, apply equally to the corresponding uses of the optative. In these neutral optatives, of which Il. iv. 18, Il. 19 gives the most striking examples, we probably come nearest to the primitive use out of which the two most common uses of the independent optative (potential and wishing) were developed. Before the Homeric period these two uses were already established, the potential with its mark of κέ or ἄν and its negative οὐ, and the wishing with no external mark and its negative μή. It is hardly possible that the first potential use of the optative was marked by κέ or ἄν, for we find undoubted potential optatives in Homer without either of these particles (see § 240), and even in Attic poetry such indefinite expressions as οὐκ ἔσθ᾽ ὅστις, οὐκ ἐσθ᾽ ὅπως, etc. have the optative without ἄν (§ 241). Although the early Greek, even in Homer, did not always use κέ or ἄν with the potential optative, there is no evidence that it ever failed to distinguish the wishing optative in negative sentences by the use of μή, while the potential was always negatived by οὐ. The Sanskrit optative, which must have had a common origin with the Greek, appears in its earliest use in the state in which we have supposed the early Greek optative to have been, i.e. used both in a potential sense and in wishes without any particle like κέ or ἄν, and occasionally in a neutral or concessive sense. But while the negative na/ (= οὐ) is always found in the potential use, we have both mA/ (= μή) and na/ in wishes and similar expressions in which the Greek has only μή.8 The same peculiarity has been noticed in the use of negatives with the subjunctive (p. 373). It is probable that at some early period the Greek had two parallel uses of the subjunctive and optative in independent sentences, as follows:— ἔλθω, I shall go (neg. οὐ), or let me go (neg. μή） ἔλθῃς, thou wilt go ( ” ), or go thou ( ” ) ἔλθῃ, he will go ( ” ), or let him go ( ” ) ἔλθοιμι, I may or might go (neg. οὐ), or may I go (neg. μή） ἔλθοις, thou mayest or mightest go ( ” ), or mayest thou go ( ” ) ἔλθοι, he may or might go ( ” ), or may he go ( ” ) Although the Greek which is best known to us did not use the second and third persons of the subjunctive in a hortatory sense, there can be little doubt that such a use existed in the earlier language, as appears from the use in Sanskrit and in Latin, and from the Greek prohibitions with μή. (See § 258.) In an Elean inscription we find two cases of the third person: τὸ δὲ ψάφισμα . . . ἀνατεθᾷ ἐν τὸ ἰαρὸν τῶ Διὸς τῶ Ὀλυμπίω, and (voted) that the decree be set up, etc.; and also ἐπιμέλειαν ποιήαται (subj.) Νικόδρομορ ὁ βωλογράφορ, that N. have charge, etc.9 Both moods alike developed a distinct potential use, which was distinguished from the other by κέ or ἄν; and in Homer we have forms like ἔλθω κε and ἔλθῃ κε parallel with ἔλθοιμί κε and ἔλθοι κε, all negatived by οὐ. The potential subjunctive, however, did not survive the Epic period, while the potential optative became fixed in the language. The future indicative also developed a potential form with κέ or ἄν, which appears to have survived the potential subjunctive, at least in the colloquial language. The English has no form except its vague I may take to express the various shades of meaning denoted by ἑλοῦμαί κε, ἕλωμαι, ἕλωμαί κε, and ἑλοίμην, which once stood between ἑλοῦμαι, I shall take, and ἑλοίμην ἄν, I should take. (See § 399.) The subjunctive, therefore, in its two chief uses in independent sentences, from which all others are derived, was originally accompanied by a weaker future form, the optative, expressing the same idea less distinctly and decidedly. Let us now see how this weaker subjunctive (or future) form enters into the various dependent constructions, that is, into conditional and final sentences and indirect discourse. The only dependent construction in which the optative is an original form, not representing another mood after a past tense, is that of protasis (including the conditional relative clause, but excluding the past generic condition). Here we see the same relation between ἐὰν (or εἰ) ἔλθω and εἰ ἔλθοιμι, if I shall go and if I should go, as between the original ἔλθω, I shall go, and ἔλθοιμι, I may (or might) go, the optative being a less distinct and vivid form for presenting a future supposition, it may be for presenting the same supposition which has already been presented by the subjunctive. The distinction, whatever it may be thought to be, is that which appears in our distinction of shall and should, and there will always be differences of opinion as to the exact nature of this.10 The objections to deriving this form of condition from the optative in wishes have already been considered. On the theory that the protasis is an offshoot of the conditional relative clause (see § 398), we should understand εἰ ἔλθω as meaning originally in case (i.e. in the case in which） I shall go or may go, and εἰ ἔλθοιμι, in case I should go or might go,—should and might being here merely weakened forms of shall and may. (Homeric optatives referring to the present are discussed below.) In the whole class of final sentences, in which the subjunctive and optative are probably the only primitive forms, the optative always represents a dependent subjunctive in the changed relation to its leading verb in which it is placed when this verb is changed from present or future to past time, a change which we represent by our change from may to might or from shall to should; as ἔρχεται ἵνα ἴδῃ τοῦτο, he comes that he may see this, ἦλθεν ἵνα ἴδοι τοῦτο, he came that he might see this, etc. The thought in the dependent clause is in both cases what would be expressed originally by ἵνα ἴδω, adapted to different circumstances; and the original subjunctive (ἵνα ἴδῃ） could always be retained, even after past tenses, and by some writers it was generally retained (§§ 318-321). The change is, in fact, the same which is made in indirect discourse when the leading verb is past, since a past final clause always expresses the past thought of the leading subject (§ 703). This relation to indirect discourse is especially clear when the future indicative is used after primary tenses, with the future optative corresponding to it after past tenses. The optative of indirect discourse has much wider relations, which were greatly extended as the language developed. Here the optative represents not merely the subjunctive but also the indicative in the changed relation in which these are placed by a change of the leading verb from present or future to past time, the tenses of the optative (with some restrictions) representing the corresponding tenses of either subjunctive or indicative at pleasure, the present including also the imperfect. In the development of the language, the want of an optative form to represent the future indicative was felt, and the future optative was added to the verb to supply the need, appearing first in Pindar. In Homer, this use of the optative is imperfectly developed, as the optative with ὅτι or ὡς in a quotation representing a simple indicative is still unknown (§ 671). Still the Homeric language has most of the other constructions of indirect discourse, including the optative in indirect questions representing both the indicative and the subjunctive. This optative in Homer appears (as we should expect) more as the correlative of the subjunctive than as that of the indicative. In indirect discourse, as in final constructions, the optative is not absolutely demanded after past tenses; and in some writers the original indicatives and subjunctives are more common (§ 670). The future optative, as a new form, is always less freely used than the older tenses. In final constructions and in indirect discourse the optative appears as a subjunctive or indicative (as it were) transferred to the past, and it here has many points in common with the Latin imperfect and pluperfect subjunctive. In Homer, moreover, the present optative is regularly used in present unreal conditions and conclusions, and both present and aorist optative with κέ occasionally refer to the past like the imperfect and aorist indicative with κέ or ἄν. These uses, taken in connexion with the secondary terminations of the optative, might lead us to think that the optative was originally a past expression, so that καί νύ κεν ἔνθ᾽ ἀπόλοιτο, and now he would have perished there, Il. v. 311, would represent the regular use of the primitive optative, instead of being (as is commonly thought) a rare exception. Against this view, however, there are many considerations to be urged. 1. The optative is fully established in Homer in wishes and conditions as a future expression, and also in present unreal conditions, the imperfect indicative here being still confined (like the aorist) to the past. In past unreal conditions the optative never appears in protasis, and only rarely in apodosis, the aorist indicative being already established here before Homer. Thus, while οὐκ ἂν γνοίης in Il. v. 85 means you would not have discerned, it would commonly mean, even in Homer, you would not discern (as future), and the common Homeric expression in Il. v. 85 would be οὐκ ἂν ἔγνως. The evidence of the Homeric language, therefore, shows that the present optative is the original form in present unreal conditions and conclusions and in present unattained wishes, but is opposed to the view that the optative was ever regularly past. 2. It is hardly possible that the past unreal conditional preceded in development the ordinary future supposition. Every primitive language must have needed expressions like if he should go he would see this before it ventured upon if he had gone he would have seen this. If now we suppose that οὐκ ἂν γνοίης had originally the sense you would not have discerned, we must assume that the Greek expressed this idea before it could express you would not discern (future), for the language never had any other form to express the latter. We cannot hesitate, therefore, to find in the common future meaning of οὐκ ἂν γνοίης the original force of the expression, and to look upon the occasional reference to the past as a relic of an early attempt to express you would not have discerned by a form already appropriated to another use. 3. The Homeric optative in conditional sentences agrees remarkably with the Sanskrit in both the future and the present use, the Sanskrit optative being used both in future and in unreal present conditions and conclusions, but not in past conditions or conclusions. This seems to show that the Greek inherited the two principal Homeric uses of the optative, (1) in future conditions and wishes, and (2) in present unreal conditions and unattained wishes, while, so far as our evidence goes, the occasional use of the optative in past potential expressions is an extension of its use beyond its hereditary limits made by the early Greek itself. 4. The argument drawn from the past tenses of the Latin subjunctive will not apply to Greek conditional sentences, for here the present and perfect subjunctive in Latin (not the imperfect and pluperfect) correspond to the Greek optative in its most frequent use, and in the older Latin these primary tenses sometimes express present unreal conditions. The most natural view seems to be, that the primitive optative, before it came into the Greek language, was a weak future form, like he may go and may he go, from which on one side came its potential and its future conditional use, and on the other side its use in exhortations and wishes. These uses would naturally all be established before there was any occasion to express either an unreal condition or an unattained wish. The need of a form for present unreal conditions and present unattained wishes would naturally come next, and the present optative was made to include these also, no practical difficulty being caused by having a single form for it would be as both present and future, none being felt in Homer and none being now felt in English. In this state the optative probably came into the Greek, before any attempt was made to extend its use to past unreal conditions. When a form was required for these, the optative may have been used at first, on the analogy of present unreal conditions; but here the serious difficulty of using ἀπόλοιτό κε for he would have perished when it was already familiar in the sense he would perish (hereafter) probably prevented the establishment of this usage. Before our evidence begins, the past tenses of the indicative were firmly established in past unreal conditions, while the optative was here a rare exception, even in apodosis, and was never used in protasis. But no attempt was yet made to dislodge the present optative from present unreal conditions or the corresponding wishes, although the use of ὤφελον or ὤφελλον in Homer shows that a past indicative in a present sense was not absolutely repugnant even to the early usage. But afterwards a new tendency prevailed, and the imperfect indicative took the place of the optative in present unreal conditions, still retaining its older use (with the aorist) in past conditions. The Greek, Sanskrit, and Latin appear to have developed their expressions of past unreal conditions independently. The Sanskrit, which seldom needed such a form, used its past future, as the Greek occasionally used ἔμελλον with the infinitive (see § 428). The optative in past general suppositions only represents the corresponding subjunctive transferred to the past. This is, moreover, not to be treated as a primitive use of the optative, for reasons which apply also to the generic subjunctive (see §§ 11, b, and 17). If the optative, at the time of its origin in some ancestral language, ever actually existed as a past form, as its terminations certainly seem to indicate, no effect has come down to the Greek from this remote origin, except perhaps the use of the optative to represent the subjunctive (and afterwards the indicative) transferred to the past in final constructions and indirect discourse. Even here, its relation to the subjunctive, which is probably all that is primitive in this use, is substantially that of a “remoter future,” as it is in independent sentences and in protasis.