Contraction in Homer

Book Nav

378*. Contraction. The question of the use of contracted forms has been already touched upon in connection with the different grammatical categories which it affects: see §§ 56, 81, 105. It will be useful here to recapitulate the results, and to notice one or two attempts which have been made to recover the original usage of Homer in this respect.1

  1. Contraction is most readily admitted between similar sounds, or when the second is of higher vowel pitch, i. e. higher in the scale ο, ω, α, η, ε. Thus we have many instances with the combinations εε, οο, αε, οε; few with εα, αω, αο, still fewer with εω, εο.
  2. In most cases in which contraction is freely admitted we find that the sound which originally separated the vowels was the semi-vοwel ι̯ or y. In case of the loss of σ it is comparatively rare; with ϝ it is probably not Homeric at all (§ 396). Hence (e. g.) although it is common with the combinations εε, εει in most verbs in -εω (§ 55), it is not found in χέω (χέϝ-ω) and is extremely rare in τρέω (τρέσ-ω, see § 29.6). But it is admitted with loss of σι̯, as in the genitive ending -ου from -οσι̯ο (-οι̯ο, -οο), and the verbs in -εω from stems in -εσ, as νεικέω (νεικεσ-ι̯ω).

    a. On these principles we should expect the 2nd singular endings -εαι, -εο, -ηαι, -αο (for -εσαι, etc.) to remain uncontracted; and this view is borne out on the whole by the very careful investigation made by J. van Leeuwen. Omitting the verbs in -αω and -εω we find that there are about 522 occurrences of these endings, and that of these 434 present uncontracted forms; while in 66 instances the contracted syllable comes before a vowel, so that it can be written with elision of -αἱ or -ο.

    Il. 3.138 κεκλήσεʼ ἄκοιτις
                  (for κεκλήσῃ)

    Il. 9.54 ἕπλεʼ ἄριστος
                (for ἔπλευ)

    In the case of -εο this mode of writing finds some support in the MSS. e. g. ψεύδεʼ (Il. 4.404), παύεʼ (Il. 9.260, Od. 1.340), εὔχεʼ (Il. 3.430, Od. 4.752), also ἕπε, read by Aristarchus in Il. 10.146 (ἕπευ MSS.). Against these 500 instances there are only 22 exceptions, 7 in the Iliad and 15 in the Odyssey, some of which can be readily corrected. Thus Il. 4.264 (= 19.139) ὄρσευ πόλεμόνδε should be ὄρσο πτόλεμόνδε (Νauck); in Il. 2.367 γνώσεαι δʼ εἰ omit δὲ (Barnes); in Il. 24.434 for ὅς με κέλῃ read ὅς κέλεαι, and so in Od. 4.812, 5.174. In Od. 18.107 for ἐπαύρῃ read the Act. ἐπαύρῃς (Van L.): as in ll. 1.209 we may retain ἴδῃς (so the MSS.; Ar. ἴδῃ—but the corruption is deeper). The greater frequency of instances in the Odyssey (and in book 24 of the Iliad) is hardly enough to indicate a difference of usage within the Homeric age.

    b. In the corresponding forms of verbs in -αω and -εω there is a concurrence of three vowels, which in our text are always reduced to two syllables, either by contraction, as in αἰδεῖο, μυθεῖαι, νεῖαι, μνάᾳ, or by hyphaeresis (§ 105), as μυθέαι, αἵρεο, ἔκλεο, πώλεαι (Od. 4.811). A single vowel appears in πειρᾷ for πειρά-εαι, ἠρῶ for ἠρά-εο. The meter requires αἰδεῖο, αἵρεο, ἔκλεο, πώλεαι; for πειρᾷ it allows πειράαι (becoming πειράʼ in Il. 24.390, 433, Od. 4.545). The isolated form ὅρηαι (Od. 14.343) for ὁρά-εαι should perhaps be ὁρᾶαι or ὁράᾳ. If the ending is in its original form it belongs to the athematic conjugation (§ 19) ; another example may be found in ὁρῆτο (or ὅρητο), read by Zenodotus in Il. 1.56.

    c. In the future in -εω (for -εσω) contraction is less frequent than in the present of verbs in -εω (-ει̯ω or -εσι̯ω). Forms such as ὀλεῖται, καμεῖται, μαχεῖται, ὀμεῖται, κομιῶ, κτεριῶ, κτεριοῦσι, evidently could not otherwise come into the verse. In

    Il. 17.451 σφῶϊν δʼ ὲν γούνεσσι βαλῶ

    we may read βάλω (Fick),

    Il. 4.161 ἔκ τε καὶ ὀψὲ τελεῖ

    we should take τελεῖ as a present. The remaining exceptions are, κτενεῖ in Il. 15.65, 68 (probably an interpolation), κατακτενεῖ in Il. 23.412, and ἐκφανεῖ in Il. 19.104.

    d. Similarly in the declension of stems in -εσ the ending -εες is rarely contracted. In the phrase φαίνονται (or φαίνεσθαι) ἐναργεῖς (Il. 20.131; Od. 7.201, 16.161) Fick happily reads ἐναργές, to be taken as an adverb. The same remedy is applicable in

    Il. 9.225 δαιτὸς μὲν ἐΐσης οὐκ ἐπιδευεῖς

    and

    ll. 13.622 ἄλλης μὲν λώβης τε καὶ αἴσχεος οὐκ ἐπιδευεῖς

    where the nominative plural is unexplained; read οὐκ ἐπιδευές there is no lack.

    e. The contraction of εο to ευ is rare in the genitive of stems in -εσ (§ 105.3), but frequent in the pronominal genitives ἐμεῦ (μευ), σεῦ, εὗ, τεῦ. Here again, however, we are struck by the number of cases in which we can substitute the forms in -ειο or -εο, with elision of -ο. In our MSS. the elision actually occurs in ἐμεῖʼ (Il. 23.789, Od. 8.462) and σεῖʼ (Il. 6.454, also Hom. H. xxxiv. 19). In

    Il. 17.173 νῦν δέ σευ ὠνοσάμην

    Zenodotus is said to have read νῦν δέ σε, i. e. probably νῦν δέ σεʼ, The full forms in -ειο or -εο occur 121 times, and may be restored without elision 9 times, with elision 56 times. To these we should add the instances in which we may put the form *μεο (6 times) or μεʼ (19 times). There remain altogether about fifty-five exceptions, which are discussed by J. van Leeuwen (Mnemοs. xiii. 215). In the phrase κέκλυτέ μευ, which occurs 19 times, he would read μοι, according to the Homeric construction (§ 143.3). So in the formula κέκλυτε δὴ νῦν μευ, Ἰθακήσιοι (5 times in the Odyssey), where however we are tempted to restore ἐμεῖ’ (cp. Il. 3.97 κέκλυτε νῦν καὶ ἐμεῖο). He suggests putting the dative for the genitive also in

    Od. 10.485 οἵ μευ φθινύθουσι φίλον κῆρ

    Od. 15.467 οἵ μευ πατέρʼ ἀμφεπένοντο

    Od. 16.92 ἦ μάλα μευ καταδάπτετʼ ἀκούοντος φίλον ἤτορ

    In the last passage it is needless to alter the genitive ἀκούοντος (§ 243.3.d), and we may even read in

    Il. 1.453 ἐμοὶ πάρος ἔκλυες εὐξαμένοιο

    Cp. Il. 16.531 ὅττι οἱ ὦκ’ ἤκουσε μέγας θεὸς εὐξαμένοιο. The substitution of the dative seems the most probable correction in various places where Leeuwen proposes other changes.

    Od. 4.746 ἐμεῦ δʼ ἕλετο μέγαν ὅρκον
                     (cp. Il. 22.119 Τρωσὶν δʼ αὖ . . . ὅρκον ἕλωμαι)

    Il. 2.388 ἱδρώσει μέν τευ τελαμὼν ἀμφὶ στήθεσφι

    Il. 22.454 αἲ γὰρ ἀπʼ οὔατος εἴη ἐμεῦ ἔπος
                    (cp. Il. 18.272)

    Also Il. 1.273, 9.377, 16.497, 19.185, 20.464, 24.293 311, 750 & 754; Od. 5.311, 9.20, 13.231, 19.108, 24.257; and perhaps

    Il. 19.137 καί μευ φρένας ἐξέλετο Ζεύς
                    (unless the με of some MSS. is right)

    so Il. 9.377 and Il. 9.335. In

    Od. 19.215 νῦν μὲν δὴ σεῦ, ξεῖνε, ὀΐω πειρήσεσθαι εἰ κτλ.

    Leeuwen restores the accusative σὲ (as in Il. 18.600). In Od. 17.421 (= 19.77) we may perhaps read καὶ ὅτι κεχρημένος ἔλθοι (ὅτῑ as in Il. 20.434 οἶδα δʼ ὅτι σὺ μὲν κτλ.). The remaining exceptions are

    Il. 5.896 ἐκ γὰρ ἐμεῦ γένος ἐσσί

    Il. 23.70 οὐ μέν μευ ζώοντος ἀκήδεις

    ll. 24.429 δέξαι ἐμεῦ πάρα

    and Il. 1.88 οὔ ;τις ἐμεῦ ζῶντος κτλ., where the contraction ζῶντος and the dative plural κοίλῃς before a consonant are also suspicious (Fick, Ilias, p. xvii).

    f. The contraction of οα, οε (from οσ-α, οσ-ε) is doubtful in the nouns in -ω and -ως (§ 105.6), but appears in the forms of the comparative, viz. ἀμείνω, ἀρείω, ἀρείους, κακίους, πλείους, and μείζω (Hesiοd), The uncontracted forms in -οα, -οες do not occur, since the meter slows either -ω, -ους or else the later -οvα, -οvες. But in such a phrase as ἀμείνω δʼ αἴσιμα πάντα (where Hauck reads ἀμείνονα) we may suspect that ἀμείνοα was the original form.

    g. Vowels originally separated by ϝ are so rarely contracted that instances in our text must be regarded with suspicion. Thus ἄκωv (ἀ-ϝέκων) should always be ἀέκων: ἄτη (ἀϝάτη) may be written ἀάτη except in

    Il. 19.88 φρεσὶν ἔμβαλον ἄγριον ἄτην

    where the use of ἄγριον as a feminine is also anomalous, § 119. In Il. 3.100, 6.356, 24.28 (where ἄτης comes at the end of the line) the better reading is ἀρχῆς. κοῖλος may be κόϊλος (cp. Latin cavus), except in Od. 22.385. εἶδον (ἔ-ϝιδον) may be ἔϊδον, except in four places (Il. 11.112, 19.292; Od. 10.194, 11.162). πολέ͜αs (accusative plural of πολύς) is not uncommon, but should probably be πολῦς (§ 100) ; πολέ͜ων occurs once (Il. 16.655). Other instances with nouns in -uς and -ευς are rare (Nauck, Mél. gr.-rom. iii. 219 ; Menrad, p. 60), The feminine in -εῖα is not contracted from -εϝϊα, -εϊα but comes directly from -εϝι̯α. So οἰός, οἰὠv for ὀϝι̯-ός, ὀϝι̯-ῶν (cp. ὄεσσι for ὀι̯-εσσι), and δῖος for δίϝ-ι̯ος. ἕ͜ως and τέ͜ως, which occur several times in our text, are nearly always followed by a particle (μέν, περ, etc.), which has evidently been inserted for the sake of the meter (ἕ͜ως μέν for ἧος, etc.). For ἀλλο͜ειδέ͜α in Od. 13.194 we should doubtless read ἀλλο-ϊδέ͜α (§ 125.2).

    εἴρυσα may be from ἔ-ϝρυσα (but see Schulze in K. Z. xxix. 64), as to ἴαχον, which has been supposed to stand for εἴαχον, from ἐ-ϝίϝαχον, see § 31.1.

    The most important example of contraction notwithstanding ϝ is the word πάϊς (παῖς, παιδός, etc). Other words which present the same difficulty are: ἆσε (Od. 11.61), ἄσατο (Il. 19.95)—in both places Νauck would read ἄασε—ἀθλοφόρος (Il. 9.265, 11.699), ἀθλεύων (Il. 24.734), ἆθλον (Od. 8.160), ἄσαμεν we slept (Od. 16.367), ἐᾷ (Il. 5.256) and other forms of ἐάω (Il. 10.344, 23.77; Od. 21.233), νέα (Od. 9.283), ῥέα (Il. 12.381, 17.461, 20.101 & 263), κρέα (Od. 9.347), χεῖσθαι (Od. 10.518), τιμῆντα (Il. 18.475), τεχνῆσσαι (Od. 7.110), ἥλιος (Od. 8.271), ἑωσφόρος (Il. 23.226), πλέων (Od. 1.184), τεθνεῶτι (Od. 19.331), πεπτεῶτα, -τας (Il. 21.503; Od. 22. 384), βεβῶσα (Od. 20.14),νόου (Il. 24.354), καιρουσσέων (Od. 7.107), the compounds of ἐννέα—ἐννῆμαρ, ἐννέ͜ωρος, ἐννε͜όργυιος—and the proper names Εὐρύκλεια Ἀντίκλεια (-κλέεια Nauck). Some of these may be disposed of by more or less probable emendation; others occur in interpolated passages (e.g. ἥλιος in the Song of Demodocus); others (as πλέων, τεθνεώς) may be explained by the loss of ϝ before ω, ο (§ 393). On the whole they are too few and isolated to be of weight against the general usage of Homer.

    The general result of the enquiry seems to be that the harshness of a synizesis or a contraction is a matter admitting of many degrees. With some combinations of vowels contraction is hardly avoided, with others it is only resorted to in case of necessity. We have already seen that the rules as to lengthening by position (§ 370) are of the same elastic character. And as there is hardly any rule of position that may not be overborne by the desire of bringing certain words into the verse, so there is no contraction that may not be excused by a sufficiently cogent metrical necessity. Thus the synizesis in such words as Ἱστίαια, Αἰγυπτίους, χρυσέοισι stands on the same footing as the neglect οf position with Σκάμανδρος or σκέπαρνον, and again the synizesis in τεμένεα, ἀσινέας, or the contraction in πονεύμενος, ἀμφιβαλεῦμαι is like the shortening of a vowel before προσηύδα, or the purely metrical lengthening of a short vowel (§ 386).

    On the same principles harshness of meter may be tolerated for the sake of a familiar phrase: e. g. the hiatus ἄφθιτα αἰεί in Il. 13.22 (ἄφθιτον ἀεί in Il. 2.46 & 186, 14.238). So when the formula καί μιν φωνήσας ἔπεα κτλ. is used of a goddess (Il. 15.35 & 89) it becomes καί μιν φωνήσασα ἔπε͜α. Again the harsh lengthening in μέροπες ἄνθρωποι (Il. 18.288, at the end of the line) is due to the familiar μερόπων ἀνθρώπων.

  • 1. See especially J. van Leeuwen, Mnemosyne, Nov. Ser. xiii. p. 215, xiv. p. 335: and Menrad, De contractionis et synizeseos usu Homerico (Monachii, 1886).