Accordingly, Sappho’s use of phrases shaped ⏑ – ⏑ ⏑ at the end of her verses could not have been modeled on phrases shaped ⏑ – ⏑ ⏑ – ⏓ at the end of epic hexameter. Nor is my argument damaged by the attestation of

at σ 323. This verse, with its violation of Hermann’s Bridge, is but a once-in-a-thousand event from the standpoint of Homeric hexameter. If anything, its banal context makes it all the less likely a model for Sappho. From the evidence of these phrases shaped ⏑ – ⏑ ⏑, I infer that Sappho’s verse-final phrases shaped ⏑ ⏑ – ⏑ ⏑ and – ⏑ ⏑ are likewise not to be derived from verse-final epic phrases shaped ⏑ ⏑ – ⏑ ⏑ – ⏓ and – ⏑ ⏑ – ⏓ respectively.

Accordingly, we should not expect a disproportionate number of matchings between disyllabic words starting the Sapphic pentameter and disyllabic words starting the Homeric hexameter. Sappho would have had plenty of opportunity to remember disyllabic words from slots other than the first foot of hexameter. Our assumptions about unconscious Homeric influence, however, have led here to a false expectation. The fact is, we find an extremely high percentage of Homeric positional counterparts for not only disyllabic but also trisyllabic and tetrasyllabic words/phrases starting the Sapphic pentameter:

does not originate from the Homeric formula

any more than gl^2d originates from pher^3d. To equate gl^2d with pher^3d minus slots 15 16 is merely to engage in surface description. Similarly, κλέοc ἄφθιτον is not some truncation of κλέοc ἄφθιτον ἔcται. Rather, it is an independent formula built into the end of a gl^2d (or gl^d or gl). By contrast,

is a formula built into the end of a (^)pher^d, while

is adequate as a formula inherited at the end of a gl^2d (or gl^d or gl), but {133|134}

would be inadequate as an ending for a (^)pher^d, since it does not involve dactylic expansion. To rephrase from the standpoint of attested hexameter, such a placement of this expression would violate Hermann’s Bridge.

Transcription:

The qualifying ἑκάεργον in the verse following the one just quoted, however, makes clear that an association with Apollo is retained. Comparable to the Homeric ἀείδοντεc Παιήονα … ἑκάεργον is the Sapphic

The last two epithets are an unmistakable indication that the god Apollo is implicated. From the standpoint of the Iliad, there is deep irony here. Hektor’s death is ensured when Apollo abandons him (Χ 212f):

There is internal irony within the Iliad itself. Achilles, standing over Hektor’s corpse, bids the Achaeans to rejoice with these words (Χ 391f): {137|138}

The name invoked in rejoicing is that of the dead Hektor’s former patron. But the supreme irony is in the epithet conferred on Hektor and Andromache by Sappho (44.34): θεοεικέλοιc ‘god-like’, in line-final position. Already in line 21 of Sappho 44, [ἴ]κελοι θεοι[cι] ‘like unto the gods’ had occurred, again in line-final position. This inverted repetition and the metrical identity of the two phrases (⏑ ⏑ – ⏑ ⏓) suggest that something is afoot. The epic hexameter preserves the Glyconic epithet θεοείκελοc ‘god-like’ in the predictable equivalent position: θεοείκελοc – ⏓#, as in the Hymn to Aphrodite 279:

In the Iliad, there are just two attestations of θεοείκελοc, again in the same position. Both attestations (Α 131, Τ 155) designate the same man: θεοείκελ’ Ἀχιλλεῦ. The man is Achilles, killer of Hektor. And Sappho uses an epithet reserved for Achilles in the Iliad as the epithet for Hektor and Andromache. It is the very last word in a poem celebrating their {138|139} joyous wedding. I cannot help but suspect that Sappho did this deliberately.