Contraction, Synizesis, Hyphaeresis

105. Contraction, etc. The loss of ι, υ, and σ between vowels (§ 94) does not generally lead to contraction in the Homeric dialect.

1. The dative singular of stems in -εσ and -υ (genitive -εος) often forms εἰ (for -ε-ϊ), but nearly always before a vowel, so that the εἰ is scanned as a short syllable (§ 380); e.g. τείχει ὕπο Τρώων, ἢ ἔπει ἢ ἔργῳ, etc. No such rule will be found to hold for the dative singular of stems in -ι, as πόλει, ἀγύρει, etc.—either because -ει from -ει-ι became monosyllabic earlier than -εἰ from -εσ-ι or -εϝ-ι; or because, as has been suggested (§ 99), the true form of the dative is πόλῑ, ἀγύρῑ, etc.

   Exceptions, real or apparent, to this rule are

   Il. 6.126 σῷ θάρσει
   (read θάρσεϊ σῷ, cp. Il. 7.153 θάρσεϊ ᾧ)

   Il. 17.647 ἐν δὲ φάει καὶ ὄλεσσον
   (read ἐν φάεϊ)

   Il. 23.515 οὔ τι τάχει γε (read οὐ τάχεΐ γε)

   Il. 23.639 πλήθει
   (read πληθυῖ)

   Also οὔδει, dative of οὖδας (Il. 5.734, 8.385, 14.467, 17.92, 23.719, 24.527), for which read οὔδαι or οὕδᾳ (§ 99).

2. The combinations -εα, -εο, -εω are often scanned as one syllable by synizesis, as θεοί (Il. 1.18), σάκε̅α̅ (Il. 4.113), τεύχε̅α̅ (Il. 7.207, etc.); so with the pronouns ἡμέας, ὑμέας, σφέας.

   In Il. 1.18 ὑμῖν μὲν θεοὶ δοῖεν Ολύμπια δώματʼ ἔχοντες the wοrd θεοί is not certain, since Ολύμπια δώματʼ ἔχοντες the lords if Olympus is used as a substantive, and θεοί is therefore unnecessary (Fick, Ilias p. 75).

3. The genitive singular has -ευς for -ε-ος in a few words
   

   Ἐρέβευς
   θάρσευς
   θέρευς
   θάμβευς

   chiefly ἅπαξ εἰρημένα. It is probably better to write -εος and admit synizesis.

   On -ευ in ἐμεῦ, σεῦ, εὗ, τεῦ see § 378*.

4. Nouns with stems in -εεσ (as κλέος, δέος) and some nouns in -ᾰς are liable to hyphaeresis, or dropping a vowel before another vοwel.
   

   κλέα (for κλέε-α)
   δυσκλέα, ἀκλέα, ἀκλέ-ες
   νηλής, νηλέϊ, νηλέα (neuter singular νηλεές)
   θεουδής, θεουδέα (for θεο-δϝής god-fearing), ὑπερδέα (Il. 17.332)
   γέρᾰ, δέπα, κέρα, κρέα, σφέλα (for γέρα-α, etc.)
   χρέα debts (Hes. Op. 647)

   Cp. δαΐ (for δαϊ-ι), dative of δάϊ-ς; also ἀποαίρεο for ἀποαιρέ-εο (§ 5).

   The forms κλέα (ἀκλέα, δυσκλέα), δέπα, κέρα, σφέλα are only found before hiatus—e g. κλέα only occurs in the phrase κλέα ἀνδρῶν—so that we must either suppose -ᾱ to be shortened by the hiatus, or (better) read κλέεʼ ἀνδρῶν, etc. But γέρα occurs before a consonant

   Il. 2.237 γέρα πεσσέμεν

   and so Il. 9.334, Od. 44.66.

   κρέα occurs in the phrase κρέα ἔδμεναι, and in one or two other places before a vowel; but more frequently it is followed by a consonant, and is to be scanned κρε̆ᾰ or κρε̅α̅ (necessarily so in Od. 9. 347, where it ends the line). Pοssibly the ᾱ is shortened by the analogy of the ordinary neuter plural forms in -ᾰ (Meyer, G. G. p. 348). Or, as is now maintained by Joh. Schmidt (Pluralb. p. 321 ff.), κρέα, γέρα, etc. are stems in -ᾰ, originally distinct from the corresponding stems in -ᾰσ, and are therefore properly singular, but capable of being used in a collective sense. On this view κρέᾰ meant flesh, κρέαα pieces of flesh: cp. μῆρα and μηροί (§ 99*). Schmidt does not admit hyphaeresis in most of these words, holding that it only occurredd when three vowels came together in the oldest Greek: so that (e g.) we may have δέα for δϝέεα (δϝει̯εσ-α), but not κλέα for κλέϝεα.

5. There are also several contracted forms from stems in -εεσ which οffer some difficulty.
   

   ἀκληεῖς (Il. 12.316)
   ἀκλειῶς (Od. 1.241, 14.371)
   ἐϋκλειῶς (Il. 22.110)
   ἐϋκλεῖας (Il. 10.281, Od. 21.331 : al. ἐϋκλῆας)
   ἀγακλῆος (ἀγακλεῖος Hesych.)
   Πατροκλῆος, Πατροκλῆα
   Ἡρακλῆος, Ἡρακλῆα, Ἡρακλῆϊ
   Βαθυκλῆα
   Διοκλῆος, Διοκλῆα
   ζαχρηεῖς, ζαχρειῶν (also ζαχρηῶν Hesych.)
   ἐϋρρεῖος
   δείους (Il. 10.376, 15.4)
   σπείους, σπῆϊ, σπέσσι and σπήεσσι

   But the η or εἰ always occurs where it can be resolved into εε, as Πατροκλεέ-ος, ἐϋρρεέ-ος, ἀκλεέ-ως, etc.; moreover the long final syllable so lost (e. g. in writing ἀκλεέ-ες, δέε-ος, σπέε-ος) is never necessary to the meter. Hence we can hardly doubt that these are the true Homeric forms. So κρειῶν (genitive plural of κρέας) should be κρεά-ων (as in H. Merc. 130), or perhaps κρεέων (see § 107.3) : and ζαχρηεῖς, ζαχρειῶν shοuld be ζαχραέες, ζακραέων. For σπέσσι we can read σπέεσι.

   The vοcative of Πατροκλέης should be written in the uncontracted form Πατρόκλεες in the phrase Πατρόκλεες ἱππεῦ (which ends the line in Il. 16.20, 744, 812, 843), and also whenever it comes before the Bucolic Diaeresis (§ 368), When it stands at the beginning of the line (Il. 16.693, 859) we should perhaps read Πάτροκλος: see § 164.

6. The case forms of nouns in -ως and -ω (genitive -οος) ought generally to be written without contraction.
   

   Nom. ἠώς
   Dat. ἠόϊ
   Acc. ἠόα (see § 368)

   Nom. αἰδώς
   Dat. αἰδόϊ
   Acc. αἰδόα

   Nom. ἱδρώς
   Acc. ἱδρόα (Il. 10.574).

   But the genitive in -οῦς (ἠοῦς, Λητοῦς, etc.) is required by the meter in several places. Naturally the contraction of οο was earlier than that of two unlike sounds, as οι, οα. See L. Meyer, Decl. 23.