Page:On translating Homer (1905).djvu/162

 *proached to its Welsh sound, that is, to χλ, it is not wonderful that such a pronunciation as οφρᾰ λλαβωμεν was possible: but it is singular that the ὕδατι χλιαρῷ of Attic is written λιαρῷ in our Homeric text, though the metre needs a double consonant. Such phenomena as χλιαρὸς and λιαρὸς, εἴβω and λειβω, ἴα and μία, εἴμαρμαι and ἔμμορε, αἶα and γαῖα, γέντο for ἕλετο, ἰωκὴ and ἴωξις with διώκω, need to be reconsidered in connection. The εἰς ἅλα ἇλτο of our Homer was perhaps εἰς ἅλα σάλλτο: when λλ was changed into λ, they compensated by circumflexing the vowel. I might add the query, Is it so certain that his θεαων was θeāwōn, and not θeārōn, analogous to Latin dearum? But dropping here everything that has the slightest uncertainty, the mere restoration of the w where it is most necessary, makes a startling addition to the antiquated sound of the Homeric text. The reciters of Homer in Athens must have dropped the w, since it is never written. Nor indeed would Sophocles have introduced in his Trachiniæ, ἁ δέ οἱ φίλα δάμαρ leaving a hiatus most offensive to the Attics, in mere imitation of Homer, if he had been accustomed to hear from the reciters, de woi or de swoi. In other words also, as in οὐλόμενος for ὀλό-*

αλλο with alio, ἁλλ with sali.]
 * [Footnote: μᾶλλον is for μάλιον, and compares φυλλο with folio,