ἐλθοῦσ᾽ Οὔλυμπον δὲ Δία λίσαι, εἴ ποτε δή τι
is the a in Dia long? Why? it should be short.
Short final vowels in Homer are sometimes lengthened before a following word-initial λ-, particularly in the princeps position (the initial long/heavy syllable of the dactylic foot). See West, Greek Metre, pp. 15, 38. The λ was prolonged, closing the preceding syllable, and thus resulting in a long/heavy syllable. Other consonants sometimes having this effect: μ, ν, ρ, σ, ϝ.
thanx,
Be prepared to find many apparent violations of the textbook metrical rules in Homer.
How about the αι in λίσαι? As a diphthong, it is long by nature; however, because the next foot starts with a vowel/diphthong, it could be shortened, according to the rules. Two exceptions within one single foot: I wonder if a more simple compromise could be accepted, i.e. to have the foot start with a short and end in a long syllable.
The αι in λίσαι is shortened there. Here is the verse with ictus bolded:
ἐλθοῦσ᾽ Οὔλυμπον δὲ Δία λίσαι, εἴ ποτε δή τι
I’ll note that if you are reading a verse and emphasizing the ictus in performance, a prolonged short syllable in ictus position does not trip up a line, whether it comes from the consonant or what (Hylander’s list happens to include every consonant that can end a Greek word, and you see -εν and -ος, etc.) or sometimes from lack of correption. The sorts of metrical irregularities that can’t easily be fixed in performance seem to occur less often. Lack of elision would trip up performance, but they write that down.
My list of initial consonants that can sometimes lengthen a preceding short vowel is taken directly from West. The consonants in question are resonants, plus σ. These are the Greek non-plosive consonants. (ϝ, of course, was probably not pronounced when the Homeric poems were committed to writing, but its pesence was preserved in formulaic expressions.) I think the phonetic properties of these consonants – perhaps a slightly longer duration than plosives – accounts for this phenomenon.
West, mostly summarizing Hartel on this, does not engage with the fact that this is a rule that (at least in Homer) mostly affects the ictus position. Notice Hartel’s collection of neuter plurals read long, which I reproduce here:
Ε745 (Θ389) ἐς δ’ ὄχεα φλόγεα ποσὶ βήσετο, λάζετο δ’ ἔγχος
Υ255 πόλλ’ ἐτεά τε καὶ οὐκί· χόλος δέ τε καὶ τὰ κελεύει.
Φ352 τὰ περὶ καλὰ ῥέεθρα ἅλις ποταμοῖο πεφύκει·
Χ64 βαλλόμενα ποτὶ γαίῃ ἐν αἰνῇ δηϊοτῆτι,
Ψ240 εὖ διαγιγνώσκοντες· ἀριφραδέα δὲ τέτυκται·
Ω7 ἠδ’ ὁπόσα τολύπευσε σὺν αὐτῷ καὶ πάθεν ἄλγεα
ξ343 ῥωγαλέα, τὰ καὶ αὐτὸς ἐν ὀφθαλμοῖσιν ὅρηαι·
ι109 ἀλλὰ τά γ’ ἄσπαρτα καὶ ἀνήροτα πάντα φύονται,
ι147 οὔτ’ οὖν κύματα μακρὰ κυλινδόμενα ποτὶ χέρσον
κ353 πορφύρεα καθύπερθ’, ὑπένερθε δὲ λῖθ’ ὑπέβαλλεν·
μ396 ὀπταλέα τε καὶ ὠμά, βοῶν δ’ ὣς γίγνετο φωνή.
ν435 ῥωγαλέα ῥυπόωντα, κακῷ μεμορυγμένα καπνῷ·
ψ225 νῦν δ’, ἐπεὶ ἤδη σήματ’ ἀριφραδέα κατέλεξας
Yes, they are . . _, as West mentions (except Φ352), but restated, we could also say “the -α comes in ictus position.” Notice what happens when the -α does not come in ictus:
Ε358 πολλὰ λισσομένη χρυσάμπυκας ᾔτεεν ἵππους·
Ω755 πολλὰ ῥυστάζεσκεν ἑοῦ περὶ σῆμ’ ἑτάροιο
ν438 πυκνὰ ῥωγαλέην· ἐν δὲ στρόφος ἦεν ἀορτήρ.
In thesis position, where a performer emphasizing the ictus would indeed be tripped up by a short vowel in long position, we only see it in front of λ or ρ. There is the phonological rule in action, perhaps. Though even here, notice that it’s all in the first foot, where there is really no chance of it being a problem for performance or reading.
If this were a primarily phonological phenomenon, one would assume that it would not show the arsis/thesis preference that it clearly does. Instead, the parsimonious explanation is that this is an artifact of performers and readers emphasizing the ictus. West’s claim that “We may probably infer that the tempo of the hexameter was fairly rapid; with a slow and stately delivery, these short syllables in long positions would have been unbearable” is unnecessary.