There remains to be considered the aorist ἷξε (second person singular ἷξες, third person plural ἷξον). This verb has given scholars particular difficulty; Wackernagel in 1925, for example, called ἷξον das einzige unerklärte Beispiel des Aoristus mixtus. [1] This is the only “mixed aorist” which is both active and indicative. Some scholars state that ἷξε would be a regular sigmatic aorist form if it were not for the plural ἷξον. [2] Against this view is the fact that the sigmatic aorist ἷξα occurs only as a falsa lectio at H. Apoll. 223 and Quintus Smyrnaeus 12.461. If ἷξε is not a first aorist, is it a secondary future, as Magnien thinks? [3] Hardly, because, as Magnien himself notices, the future from this root is middle (ἵξομαι). Indeed, the middle is regular in all tenses formed from this root except the present ἵκω; ἱκνέομαι, ἴξομαι, ἱκόμην. As Wackernagel says, the active form, as well as the long vowel, of ἵκω most probably means that this was originally a perfect stem, later taken as present and given primary and secondary endings. [4] Commonly an intransitive perfect active is associated with middle forms in other tenses. So ἷξε, as active, can be connected only with ἵκω.

We need not worry about explaining the second person singular ἷξες. This occurs only in the Hymn to Apollo, in narrative addressed to the god. That is, it is simply a translation into direct address of narrative with ἷξε. [5] Compare, for example, Y 31 ἷξον δ’ἐς Πυλίων ἀνδρῶν ἄγυρίν τε καὶ ἕδρας, and H. Apoll. 278: ἷξες δ’ἐς Φλεγύων ἀνδρῶν πόλιν ὑβριστάων. Clearly, this use of the second person is derived from the third person.

If we investigate the relationship of ἷξε ἷξον το ἷκε (*ἷκον does not occur) on the one hand and to ἴκετο ἴκοντο on the other, a curious pattern emerges. In meaning, ἷξε ἷξες ἷξον cannot be distinguished from ἵκετο ἵκεο ἵκοντο: “he came (to),” usually with a personal subject. Compare E 773: ἀλλ’ὅτε δὴ Τροίην ἷξον ποταμώ τε ῥέοντε, and Λ 170; ἀλλ’ ὅτε δὴ Σκαιάς τε πύλας καὶ φηγὸν ἵκοντο. On the other hand, ἷκε(ν) is limited, as Wahrmann saw, [6] to the cadence-formula οὐρανὸν ἷκε (verse-end and before the main caesura – A 317, B 153, B 458, M 338, P 425, T 362, H. Apoll. 442, not in the Odyssey). The subject is noise, light, or smoke; for example, B 458 αἴγλη παμφανόωσα δι’αἰθέρος οὐρανὸν ἷκε. In accordance with Kuryłowicz’s fourth principle of analogy, [7] it is evident that ἷκε, with its restricted usage, is the older form, whereas ἵκετο and ἷξε are the newer forms which have replaced it in the primary function. Of these two, ἵκετο is an ordinary thematic aorist middle. The coexistence of ἵκετο and ἷκε presumably has a metrical cause, namely that ἵκετο could not replace ἷκε in all metrical positions (there are some where it could have, like Z 172, as Leumann points out). [8] We can infer that ἷξον has replaced *ἷκον. The noun πόρος “ford” occurs in Homer virtually only in the formulaic verse Ξ 433 = Φ 1 = Ω 692: ἀλλ’ὅτε δὴ πόρον ἷξον ἐὕρρεῖος ποταμοῖο. The only other occurrence is in the Catalogue, B 592. This might lead us to believe that ἷξον was an old form, in contradiction to what has just been said about ἷξε. But since *ἷκον and ἷξον have the same metrical shape, we may suppose that the original formula contained πόρον ἷκον. When ἷξε replaced ἷκε in the in the primary sense “came”, then ἷξον also replaced *ἷκον in that sense. The plural evidently did not occur in formulae of the sort where ἷκε survived.

Why, then, should ἷκε have been replaced? Evidently because it came to be felt as imperfect and not aorist when the formal opposition between imperfect and aorist was established. [9] Then there was a use for a form marked as aorist and having the same metrical shape as ἷκε. [10] But how was ἷξε created? Meillet suggests a line of approach when he compares ἷξον to a class of Indo-Iranian thematic sigmatic aorists from roots ending in velars. [11] Kuryłowicz has explained the Sanskrit forms (seventh aorist) as thematic aorists which have had an s added (not sigmatic aorists which have been thematized). [12] Following his argument in the Sanskrit case, we can construct a similar explanation of the isolated Greek aorist ἷξε. The starting point is a situation in which certain verbs have aorists with s before endings beginning with a vowel but no s before endings beginning with a consonant. In Greek this situation appeared in the case of verbs which had a radical athematic middle but a sigmatic active, for example: ὧρτο ὧρσε ἔλεκτο: ἔλεξε. These forms, though of different ages, coexisted at least in the epic language. The sigmatic aorists active are causative in sense and a productive formation. The archaic athematic middles may, at least in some cases, have been originally sigmatic, but this does not affect the argument. As Kuryłowicz says of the Indian aorists, the asigmatic aorist, because it has no suffix, is formally dominated by the sigmatic aorist. [13] In other words, the compound morpheme tends to replace the simple morpheme in the same function. [14] Now, in the case of ἵκω, there was not only the imperfect or aorist active ἷκε but also the athematic aorist middle ἷκτο attested in Hesiod Theogony 481: ἔνθα μιν ἷκτο φέρουσα θοὴν διὰ νύκτα μέλαιναν, “Thither carrying him she came swiftly through the black night.” The epithet ἴκμενον may be the participle of this aorist. The opposition in this verb ἷκτο· ἷκε does not show the same distribution of s as in the other verbs mentioned earlier, namely s before a vowel, no s before a consonant: ὧρτο· ὦρσε ἔλεκτο· ἔλεξε Because of the dominance of the sigmatic aorist, this pattern is imposed on ἵκω, giving the new opposition ἷκτο: ἷξε As in the Sanskrit seventh aorist, the process of sigmatization still preserves the thematic vowel. The fundamental morpheme of the “sigmatic” aorist in Greek is the a, which under certain conditions is preceded by s.The principal morpheme a can bring in the accessory morpheme s, for example after a vowel stem where the original s had been lost; but the s without a is not sufficient to characterize a form as “sigmatic aorist,” and therefore does not bring in an a. For ἷκε *ἷκον, we thus get ἷξε ἷξες ἷξον, not ἷξε **ἷξας **ἷξαν. Here is an analogical development which occurred in Sanskrit in the natural language, but in Greek only in an isolated case of the poetic language. The process was available in Greek, but it became operative only when meter made the regular development (ἵκετο) impossible.