393. The chief instances in which loss of an original ϝ can be shown to have taken place in the language of Homer fall under the following rule.
When the original initial ϝ is followed by the vowels ο, ω, or the diphthong oυ, it produces no effect on the meter of Homer.
The following are words to which this rule will apply.[fn]See an article by Leo Meyer, K. Z. xxiii. pp. 49 ff.[/fn]
ὁράω, οὖρος (and οὐρεύς) a watcher; ὄρεσθαι to watch. The original ϝ (Germ. wahr-) will account for the forms ἑώρων and ἐπί-ουρος, but there are no traces in the meter of such forms as ϝοράω, etc.
ὄρος mountain (cp. Βορέας), and ὀρθός upright, which may be from the same root (cp. the Laconian Ἄρτεμις βωρθία). There is only one instance of hiatus (viz. Od. 3.290 ἶσα ὄρεσσιν).
ὄρτυξ (Sanskrit vartakas a quail) appears in the name Ὀρτυγίη, which does not admit ϝ (Od. 5.123).
ὄχος chariot (Lat. veho); ὄχλος (lit. movement, tossing), ὀχλέω to disturb (cp. ὀχλεύς and Latin vectis); ὀχθέω (Latin vehe-mens). A trace of ϝ appears in the form συνεοχμός (Il. 14.465).
οψ, ὄσσα, ὀμφή voice. The traces of ϝ are, one instance of hiatus before ὄπα (Od. 11.421), two of lengthening of a short syllable (Il. 18.222, Od. 12.52), and one or two phrases such as ἀμειβόμεναι ὀπὶ καλῇ, etc.; while there are three undoubtedly adverse places (Il. 11.137, 21.98; Od. 5.61). In the case of ὀμφή the evidence is clear against ϝ; in ὄσσα it is indecisive.
οὐρανός (Sanskrit varuṇas).
οὐλαί coarsely ground barley, connected with the root ϝελ-, meaning to roll, etc. Neither this word nor the derivative οὐλοχύται admits ϝ.
οὐλαμός crowd, press of battle, shows traces of initial ϝ in Il. 20.379 ἐδύσετο οὐλαμὸν ἀνδρῶν and the phrase ἀνὰ οὐλαμὸν ἀνδρῶν (Il. 4.251 & 273, 20.113). It does not occur except in these places.
οὐτάω, ὠτειλή wound; cp. ἄ-ουτος unwounded, and the form γατάλαι in Hesychius.
ὠθέω (ἐώθεον, ἔωσα), root vadh (to beat).
ὦνος price, imperfect ἐωνούμην (Sanskrit vasnas, Latin vēnum).
Other words which may have originally had initial ϝ are, ὅρκος (cp. ἐπί-ορκος), ὀνίνημι (ἐρι-ούνιος), ὀΐγνυμι (ἀνα-οίγεσκον, ἀνέῳγε, etc.), ὀπυίω, ὄκνος, οὐρή, etc. (L. Meyer, l. c.). However this may be, none of them show traces of ϝ in Homer. There remain the forms of the possessive ὅς to which the rule would apply, viz. οὗ, ὅν, ὥ, ὧν, οὕς. Hiatus is found before ὅν in 18 places (before ὅνδε δόμονδε seven times, ὃν κατὰ θυμόν six times, in προτὶ ὅν four times), οὕς twice (Il. 2.832, 11.330), οὗ once (Od. 15.358). On the other hand there are 22 places in which the forms in question do not admit ϝ. The significance of this proportion appears when we know that in the case of the remaining forms of the possessive ὅς the places with hiatus number 50, the adverse instances eight, and that with the forms of the personal pronoun (ἕο, οἷ, etc.) the proportion is 728 to 19. It seems probable, therefore, that in the forms οὗ, ὅν, etc., the ϝ was no longer pronounced, though traces of the former pronunciation remained (as in the case of οὐλαμός and ὄψ). Similarly in English the sound of w is lost before the vowel o in who, whom, whose, but retained in which, what, etc.
The retention of ϝ before the diphthong οι, as in οἷ, οἷο, οἷς, also in οἶκος and οἶνος, may indicate that ο before ι had not its ordinary sound, but one approaching to ε (possibly like French eu). This agrees with the fact that οι and υ were afterwards identical in sound, and that in the modern language both are = ι.
Words with initial υ are not found in Homer with ϝ; but we cannot in this case speak of the loss of ϝ—the combination ϝυ having been originally impossible.
The remaining instances in which loss of ϝ may be assumed in Homeric words are few, and for the most part open to question.
ἕλκω, root valk or vlak (Knös, following Curtius): ϝ is perhaps seen in κατὰ ὦλκα (Il. 13.707, Od. 18.375). This account of the wοrd separates it from Latin sulcus.
ἑλεῖν, ἕλωρ, possibly to be connected with Latin voltur the bird of prey. The instances of hiatus before ἕλωρ are hardly enough to prove ϝ.
ἕλος: from which the name Velia is said by Diοnysius Hal. (Arch. 1.20) to be derived, has no ϝ in Homer (Il. 2.584 & 594, 20.221, Od. 14.474). The ϝ of this wοrd is also wanting in the Cyprian dialect (Deecke and Siegismund, Curt.* Stud. vii. 249).
Ἦλις, Ἠλεῖος is without ϝ in Homer: ϝαλήϊοι is the form found on Elean and Laconian inscriptions.
ἦλος (Latin vallus) rejects ϝ in Il. 11.29 ὲν δέ οἱ ἦλοι; the two other places where it occurs prove nothing.
ἰδίω, ἱδρώς (root svid): the σϝ is lost in Homer.
ἵκω, ἱκvέομαι: the derivation from the root viç is quite uncertain.
ἱστίη (Latin Vesta): the forms ἀν-έστιος, ἐφ-έστιος show that the ϝ is lost in Homer (as also in the Laconian, Locrian, and Boeotian dialects, see § 404).