5. The Wedding of Hektor and Andromache: Epic Contacts in Sappho 44LP

Had this ill-fated couple ever really lived through a single happy day in their lives? Oh, yes: there was a time, once. In a reverie from the past, Andromache is growing up in the house of her parents. She has seven brothers and she is the only daughter. She is happy. Then, one day, there comes a handsome prince, who takes her away to his city as his own princess—
τὸ πρὶν ἐ π’ εἰρήνηc,
πρὶν ἐλθεῖν υἷαc ʼΑχαιῶν
— Kakridis [1]
          Sappho 44LP
          Κυπρο. [                              ]as·
          κάρυξ ἦλθε θε [          ] ελε[…].θειc
          Ἴδαοc ταδεκα…φ[..].ιc τάχυc ἄγγελοc
         deest unus versus
          τάc τ’ ἄλλαc Ἀcίαc .[.]δε.αν κλέοc ἄφθιτον· {118|119}
5        Ἔκτωρ καὶ cυνέταιρ[ο]ι ἄγοιc’ ἐλικώπιδα
          Θήβαc ἐξ ἰέραc Πλακίαc τ’ ἀ[π’ ἀι]ν<ν>άω
          ἄβραν Ἀνδρομάχαν ἐνὶ ναῦcιν ἐπ’ ἄλμυρον
          πόντον· πόλλα δ’ [ἐλί]γματα χρύcια κἄμματα
          πορφύρ[α] καταύτ[. .]να, ποίκιλ’ ἀθύρματα,
10       ἀργύρα τ’ ἀνάριθμα ποτήρια κἀλέφαιc.
          ὢc εἶπ’· ὀτραλέωc δ’ ἀνόρουcε πάτ[η]ρ φίλοc·
          φάμα δ’ ἦλθε κατὰ πτόλιν εὐρύχορον φίλοιc·
          αὔτικ’ Ἰλίαδαι cατίναι[c] ὐπ’ ἐυτρόχοιc
          ἆγον αἰμιόνοιc, ἐπ[έ]βαινε δὲ παῖc ὄχλοc
15      γυναίκων τ’ ἄμα παρθενίκα[ν] τ..[..]οcφύρων,
          χῶριc δ’ αὖ Περάμοιο θυγ[α]τρεc[
          ἴππ[οιc] δ’ ἄνδρεc ὔπαγον ὐπ’ ἀρ[ματ-
          π[      ]εc ἠίθεοι μεγάλω[c]τι δ[
          δ[      ]. ἀνίοχοι φ[. . . . .].[
20      π̣[    ´]ξα.ο[                              
         desunt aliquot versus
                                         ἴ]κελοι θέοι[c
                                        ]ἄγνον ἀολ[λε-
          ὄ̣ρ̣ματ̣α̣ι̣[                    ]νον ἐc ʼ ́Ιλιο[ν,
          αὖλοc δ’ ἀδυ[μ]έληc [κίθαρίc] τ’ ὀνεμίγνυ[το
25      καὶ φ[ό]φο[c κ]ροτάλ[ων, λιγέ]ωc δ’ἄρα πάρ[θενοι
          ἄειδον μέλοc ἄγν[ον, ἴκα]νε δ’ ἐc αἴθ[ερα
          ἄχω θεcπεcία γελ[
          πάντᾳ δ’ ἦc κὰτ ὄδο[ιc
          κράτηρεc φίαλαί τ’ ὀ[. . .]υεδε[. .] . .εακ[.].[
30      μύρρα καὶ καcία λίβανόc τ’ ὀνεμείχνυτο·
          γύναικεc δ’ ἐλέλυcδον ὄcαι προγενέcτερα[ι,
          πάντεc δ’ ἄνδρεc ἐπήρατον ἴαχον ὄρθιον
          Πάον’ ὀνκαλέοντεc ἐκάβολον εὐλύραν,
          ὔμνην δ’ Ἔκτορα κʼΑνδρομάχαν θεοεικέλο[ιc. {119|120}
The form of the poem is suggestive of its relation to the Epic. Consisting of Glyconics internally expanded by two dactyls, it is a pentameter with the syllable-count of 14 in each verse, vs. the 16 in epic hexameter as I trace it back to pher3d:
⏓ ⏓ – ⏓          gl2d
⏓ ⏓ – – ⏓     pher3d
Note that the elegiac pentameter, a functional correlate of the epic hexameter in elegiac couplets, also shows an inherited syllable-count of 14. [2] Like the pher3d which I reconstruct for epic hexameter, the gl2d of Sapphic pentameter is deployed verse-by-verse (= κατὰ cτίχον). Even the actual genre of Sappho 44LP is narrative, much like that of the Epic. The metrician Hephaistion reports [3] that Book II of the canonical Sapphic corpus consisted entirely of poems composed in this meter. For reasons of traditional form and content, then, the 34 surviving verses of Sappho 44LP are ideal for comparison with the epic hexameter. We are ready to ask how Sappho might have borrowed from epic diction.
From a purely surface metrical analysis, we may say that gl2d equals pher3d minus the last two syllables. If we equate pher3d with the archetype of epic hexameter, we might rephrase as follows: the Sapphic pentameter equals the {120|121} epic hexameter minus the last foot. But this equation cannot work as some sort of automatic conversion-mechanism still current in Sappho’s day, since the first foot of epic hexameter must be – – or – , while the first foot of Sappho’s pentameter is clearly ⏓ ⏓. Even though the equation under discussion cannot work metrically, we nevertheless see a striking number of instances where the phrase that ends the Sapphic verse equals a phrase that ends a Homeric verse minus the last two syllables.
44.3                              … τάχυc ἄγγελοc#
Σ 2, etc.                         … ταχὺc ἄγγελοc ἦλθε#
44.4                              … κλέοc ἄφθιτον#
Ι 413                              … κλέοc ἄφθιτον ἔcται#
44.5                              … ἐλικώπιδα#
Α 98                              … ἑλικώπιδα κούρην#
44.7                              … ἄλμυρον#
δ 511                             … ἁλμυρὸν ὕδωρ#
44.8                              … κἄμματα#
ζ 111                             … εἵματα καλά#
ζ 144                             … εἵματα δοίη#
η 265                             … εἵματα ἔccεν#
44.9                               … ἀθύρματα#
σ 323                              … ἀθύρματα θυμῷ#
44.23                              … ʼ ́Ιλιον#
Δ 416                              … Ἴλιον ἱρήν#
44.31                              … προγενέcτεραι#
β 29                                … προγενέcτεροί εἰcιν#
ω 160                              … προγενέcτεροι ἦcαν#
Β 555                              … προγενέcτεροc ἦεν#
44.32                              … (ἴαχον) ὄρθιον#
H. to Dem. 20                  … (ἰάχηcε δ’ ἄρ’) ὄρθια φωνῇ# {121|122}
The rigidity and pervasiveness of these correspondences are so obvious that we are led to ask ourselves whether they imply wholesale borrowing from epic diction. [4] Had Sappho routinely, and perhaps unimaginatively, copied her formulas from Homeric models? If so, I force myself to imagine that she consciously went through the following kind of mental process in constructing the verse-final phrases which I have just surveyed: (1) select those Homeric formulas which end the hexameter with a disyllabic word; {122|123} and (2) delete this word and rearrange the preceding syntax, if the absence of the last word makes a difference. This sort of second-guessing, of course, underestimates the artistry of Sappho, and my only purpose in attempting it has been to portray how unlikely it is that Sappho should have approached the act of composition in this manner. Notice too what happens in the following instances when we try to convert Sappho’s pentameter endings back into hexameter endings by adding two syllables:
44.7                                … ἐπ’ ἄλμυρον – ⏓#
44.9                                … ἀθύρματα – ⏓#
44.11                              … πάτηρ φίλοc – ⏓#
44.23                              … ἐc Ἴλιον – ⏓#
44.26                              … ἐc αἴθερα – ⏓#
Each of these groups, all scanned – ⏓, would violate Hermann’s Bridge if they were attested in these positions. [5] In the case of the group at 44.23 above, contrast the ending of Λ 196:
                                      … εἰc Ἴλιον ἱρήν#
                                              (also Ο 169, Ω 143)
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
                                       … ἀθύρματα θυμῷ# {123|124}
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.
If we still wish to pursue the idea that Sappho borrowed massively from traditional epic diction, we may alternatively imagine that epic formulas were so ingrained in Sappho’s mind that she could keep inserting them over and over again into her meter, without even being conscious of doing so. If, however, we are witnessing an unconscious process at work in Sappho 44LP, then we can no longer claim that Sapphic combinations like
                                      … τάχυc ἄγγελοc#
                                      … κλέοc ἄφθιτον#
                                      … ἐλικώπιδα#
                                      … προγενέcτεραι#
were inspired by epic verse-final combinations like
                                      … ταχὺc ἄγγελοc ἦλθε#
                                      … κλέοc ἄφθιτον ἔcται#
                                      … ἑλικώπιδα κούρην#
                                      … προγενέcτεροί εἰcιν#
                                      etc. {124|125}
simply because the members of the first set are incomplete formulas from the internal standpoint of epic diction. Cutting a formula short entails the curtailing of function as well as form. If we posit a truncation of these epic combinations in the second set, then we can no longer claim that Sappho was unconsciously inserting epic formulas. If Sappho was competent in adjusting and modifying epic formulas, she had to be conscious of the formal and functional limitations imposed on them by the meter she was using.
Nor is there a way out in arguing that Sappho unconsciously inserted fragments of epic formulas. If it were a simple matter of traditional words or phrases ringing in Sappho’s ear, then we are at a loss to explain the rigid correspondences of their placement in her meter and in Homeric hexameter. For example, when Sappho starts her pentameter with a disyllabic word, her meter allows her to use any word shaped – –, – , –, . Note that words with these shapes are to be found at various positions all over the hexameter. [6] If we imagine a Homeric word with any of these shapes coming to Sappho’s mind, it could theoretically originate from any appropriate part of the hexameter. Even words shaped – –, which she can use only in the first foot, could have originated from at least six different positions in the hexameter: {125|126}
– – – ⏔ – ⏔ – ⏔ – ⏔ – ⏓
– – ⏔ – ⏔ – ⏔ – ⏔ – ⏓
– ⏔ – – – ⏔ – ⏔ – ⏔ – ⏓
– ⏔ – ⏔ – – – ⏔ – ⏔ – ⏓
– ⏔ – ⏔ – ⏔ – – – ⏔ – ⏓
– ⏔ – ⏔ – ⏔ – ⏔ – ⏔ – –
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:
44.1                                #Κυπρο …
δ 83                                #Κύπρον …

44.2                                #κάρυξ (ἦλθε) …
θ 62 = 471                       #κῆρυξ (δ’ ἐγγύθεν ἦλθεν) …

44.3                                #Ἴδαοc …
Ε 20, Η 416                      #ʼΙδαῖοc …

44.5                                #Ἔκτωρ …
Γ 116, etc.                        #Ἔκτωρ …

44.6                                #Θήβαc …
Hymn to Apollo 228           #Θήβηc …

44.8                                #πόντον …
Η 6, etc.                           #πόντον …

44.9                                #πορφύρα … [7]
δ 298, etc.                        #πορφύρε’ …
k 353                               # πορφύρεα …

44.10                               # ἀργύρα … [8]
Ψ 741, etc.                       #ἀργύρεον …

44.12                               #φάμα …
υ 100, 105                        #φήμην … {127|128}

44.13                               #αὔτικ’ …
Α 386, etc.                        #αὐτίκ’ …
α 324, etc.                        #αὐτίκα …

44.14                               #ἆγον …
η 324, u 277                     #ἦγον …

44.16                               #χῶριc δ’ αὖ …
δ 130                               #χωρὶc δ’ αὖθ’ …
ω 278                              #χωρὶc δ’ αὖτε …

44.17                              #ἴπποιc …
Γ 260, etc.                       #ἵππουc …

44.24                              #αὖλοc (…[κίθαρίc] τ’) …
Σ 495                              #αὐλοὶ (φόρμιγγέc τε) …

44.26                              #ἄειδον …
ρ 519                              #ἀείδῃ …

44.27                              #ἄχω θεcπεcία …
Θ 159, etc.                      #ἠχῇ θεcπεcίῃ …

44.28                              #πάντᾳ …
Α 384, etc.                       #πάντῃ …

44.29                              #κράτηρεc …
Ζ 528                              #κρητῆρα …{128|129}
We see the same sort of rigid correspondences between Sapphic phrases at the end of the pentameter and Homeric phrases at the end of hexameter minus two syllables. For example, although Homeric formulas shaped may theoretically occur at four different places in the pentameter, Sappho features them predominantly at the end:
                                      … τάχυc ἄγγελοc#
                                      … κλέοc ἄφθιτον#
                                      … ἐλικώπιδα#
                                      … προγενέcτεραι#
Compare again the Homeric analogues:
                                      … ταχὺc ἄγγελοc ἦλθε#
                                      … κλέοc ἄφθιτον ἔcται#
                                      … ἑλικώπιδα κούρην#
                                      … προγενέcτεροί εἰcιν#
From such evidence, I conclude that the phraseology of a Sapphic pentameter both starts and ends like that of an epic hexameter minus the last two syllables. Notice the parallelism here between phraseology and meter: gl2d equals pher3d minus syllables 15 and 16.
1̄̆ 2̄̆           3̄ 4̆ 5̆ 6̄ 7̆ 8̆ 9̄ 10̆ 11̆ 12̄ 13̆ 14̄̆               gl2d (pentameter)
1̄̆ 2̄̆           3̄ 4̆ 5̆ 6̄ 7̆ 8̆ 9̄ 10̆ 11̆ 12̄ 13̆ 14̆ 15̄ 16̄̆     pher3d
1̄ 2̆ ̅2̆½    3̄ 4̆ 5̆ 6̄ 7̆ 8̆ 9̄ 10̆ 11̆ 12̄ 13̆ 14̆ 15̄ 16̄̆     hexameter {129|130}
The parallelism is even deeper. Phrases in Sapphic pentameter and epic hexameter may correspond not only at the beginning and end but also in the middle. For example, the epithet εὐρύχορον (line 12) is placed at slots 9 10 11 12 of the pentameter; likewise, εὐρύχορον/εὐρυχόρῳ are restricted to slots 9 10 11 12 of the Homeric hexameter (Β 498, Ψ 299, ζ 4, λ 256). There is a semantic parallelism as well. Sapphic εὐρύχορον serves as epithet of πόλιν, while the Homeric εὐρύχορον/εὐρυχόρῳ describe the following πόλειc: Μυκαληccόν, cικυῶνι, ʽΥπερείῃ, ʼΙαωλκῷ.
I have till now withheld elaboration on an important feature which distinguishes the phraseology of Sapphic pentameter from that of hexameter. Pentameter tolerates a short at slot 1 and a long at slot 14:
44.31                              #γŭναίκων …
44.33                              … εὐλύρᾶν#
This sort of flexibility is not shared by hexameter, which barely tolerates a short at 1 and, a long at 14 not at all. In other respects as well, the phraseology of the Sapphic pentameter is more flexible than that of the Homeric hexameter. An expression like πάτηρ φίλοc (line 11) is accommodated at slots 11 12 13 14 of Sapphic pentameter but barred from slots 11 12 13 14 of Homeric hexameter, where it would violate Hermann’s Bridge. Instead, Homeric πατὴρ φίλοc is placed at slots 8 9 10 11 (Χ 408). Similarly, ἐπήρατον (line 32) is accommodated at {130|131} slots 5 6 7 8 of Sapphic pentameter but barred from slots 5 6 7 8 of Homeric hexameter, where it would violate the requirement of a main caesura (either penthemimeral or trochaic). Instead, Homeric ἐπήρατον is placed at slots 8 9 10 11 (Σ 512, Χ 121, ν 103, 347).
In sum, there are rigid correspondences between the Sapphic pentameter and the Homeric hexameter not only in meter and phraseology but also in the placement of this phraseology. Where the correspondences break down, it is because one meter is more or less versatile than the other in accommodating the phraseology. Of course, pentameter is less versatile than hexameter in allowing words or phrases shaped – – … only in slots 1 2. It is also less versatile in barring words or phrases shaped – … from slots 1 2. For example, the expression for ‘he/she said’, banal but essential for narrative, is #ὢc εἶπ’ … in Sappho 44.11 vs. ὣc ἔφατ’ … passim in epic hexameter. On the other hand, pentameter is more versatile than hexameter in its capacity for phrases resulting in short at slot 1 or long at slot 14. So far, every one of these phraseological distinctions in pentameter could have theoretically resulted from archaism. In hexameter, certainly, the substitution of – for – – in slots 1 2 and the substitution of – – for – elsewhere are both innovations. [9] The Sapphic pentameter, as we have just seen, is {131|132} also more versatile than the epic hexameter by virtue of ignoring word-breaks and bridges which are de rigueur in the latter. Because of this flexibility, words or phrases shaped may occur in slots where potential Homeric analogues are excluded. Here too, this phraseological distinction in pentameter theoretically could have resulted from archaism. If indeed the main word-breaks and bridges of epic hexameter were motivated by formulas shaped (^)pherd within a pher3d meter, [10] then there is really no reason to expect the same constraints in Sappho 44LP, which is after all composed in Glyconics, not Pherecratics. That is, there is no reason to expect the same constraints once we rid ourselves of the assumption that the phraseology of the Sapphic pentameter was modeled on that of the epic hexameter.
The time has come to propose what must by now be obvious: traditional expressions shaped , , and – at the end of Sapphic pentameter may have been inherited Glyconic formulas rather than truncated Pherecratic formulas. By the same token, we may assume that the expressions shaped –, –, and – – in the same position are likewise inherited Glyconic formulas. With this theory, we can explain the rigid phraseological correspondences between Sapphic pentameter and epic hexameter as a phenomenon parallel to the rigid metrical {132|133} correspondence between gl2d and pher3d. Thus an expression like
                                      … κλέοc ἄφθιτον# (44.4)
does not originate from the Homeric formula
                                      … κλέοc ἄφθιτον ἔcται# (Ι 413)
any more than gl2d originates from pher3d. To equate gl2d with pher3d 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 gl2d (or gld or gl). By contrast,
                                      … κλέοc ἄφθιτον ἔcται#
is a formula built into the end of a (^)pherd, while
                                      … κλέοc ἔcται||
is a formula built into the end of a plain pher. [11] With this theory, we can explain not only the strict correspondences but also the divergences in the phraseology of Sapphic pentameter and epic hexameter. For example, an expression like
                                      … πάτηρ φίλοc# (44.11)
is adequate as a formula inherited at the end of a gl2d (or gld or gl), but {133|134}
                                      … πατὴρ φίλοc – ⏓#
would be inadequate as an ending for a (^)pherd, 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.
I propose, then, that Sappho had at her disposal a tradition of inherited formulas which were parallel to the inherited meter of her verses. The rigid phraseological correspondences between her pentameter and the epic hexameter are due to parallel inheritance of related formulas from related meters. But the fact is, and it cannot be emphasized enough, that Sapphic pentameter is actually far more archaic in structure than Homeric hexameter. The most important of the archaisms is the preservation of isosyllabism and the Aeolic base. Contrast the extent of evolution in Homeric hexameter, where spondees instead of dactyls are frequent in the first four feet [12] and where a pattern like – in the first foot is so rare as to be considered a curiosity by metricians. History has played havoc with the factor of attestation, which is the fundamental problem. By a fluke of history, structurally primitive material like the pentameter inherited by Sappho is attested at a relatively late phase, while structurally more developed material like the Homeric Epic itself had become a Panhellenic fixed text at so {134|135} early a period that it was already prehistoric from the standpoint of the classical period. A term such as ‘Sapphic pentameter’ merely indicates that the given meter does not happen to be attested before the Sapphic era of Lyric.
All this is not to say that Sappho was unaware of the basic mechanics of epic meter and formula. Nor that she could not selectively make direct use of epic diction. Nor even that she was not under the influence of the Iliad when she composed The Wedding of Hektor and Andromache. If a genre is closely related to another, it has a built-in medium for reacting to this other genre.
Sappho’s poem on The Wedding of Hektor and Andromache ends on a note of celebration. Both men and women are singing (lines 31-32),
33. Πάον’ ὀνκαλέοντεc ἐκάβολον εὐλύραν
34. ὔμνην δ’ Ἔκτορα κ’Ανδρομάχαν θεοεικέλοι[c

“invoking Paon the far-darter, the one with the fine lyre,
and they sang of Hektor and Andromache, the god-like”
The name Πᾱ́ων can be reconstructed as *Payāwōn/*Payāwonos. [13] The Doric form παιᾱ́ν shows contraction of *-āwō- to -ā-; the Attic παιών shows contraction of *-āwō- to -ō-. As for Aeolic Πᾱ́ων, *-ayā- seems to have contracted to {135|136} -ā-. [14] The god *Payāwōn is already attested in the Linear B texts of Crete, on the Knossos tablet V 52:
a-ta-na-po-ti-ni-ja 1[
e-nu-wa-ri-jo 1 pa-ja-wo-ne 1 po-se-da-[o-ne
Athānāi potniāi 1[
Enuwaliōi 1 Payāwonei 1 Poseidāonei . . . . [15]
We see here the name *Payāwōn in the company of divine names which have Homeric reflexes (in the dative): ʼΑθήνη πότνια, ʼΕνυάλιοc, Ποcειδᾱ́ων. The Homeric reflex of *Payāwōn is Παιήων, who still appears with the separate identity of a healing-god in the Theomachiai of the Iliad. [16] Otherwise, the name has become subsumed as an alternative designation of Panhellenic Apollo, much as the Cretan war-god ʼΕνυάλιοc ultimately becomes simply an epithet of Panhellenic Ares in the Iliad (e.g. Ρ 211); in the Theomachiai, however, ʼΕνυάλιοc is still a separate entity from Ἄρηc. [17] Traditional invocations like ἰὴ Παιήων lead to a semantic specialization: Παιήων {136|137} comes to be not so much the god invoked, but the actual invocation of the god: the so-called paean. [18] Hence Α 473:
καλὸν ἀείδοντεc Παιήονα κοῦροι Ἀχαιῶν
“the young warriors of the Achaeans, singing a beautiful paean”
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
Πάον’ ὀνκαλέοντεc ἐκάβολον εὐλύραν
“invoking Paon the far-darter, the one with the fine lyre”
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):
ῥέπε δ’ Ἕκτοροc αἴcιμον ἦμαρ ᾤχετο δ’ εἰc Ἀἱδαο, λίπεν δέ ἑ Φοῖβοc Ἀπόλλων
“Hektor’s day of doom came, and he went off to Hades. Apollo left him”
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}
νῦν δ’ ἄγ’ ἀείδοντεc Παιήονα κοῦροι Ἀχαιῶν νηυcὶν ἔπι γλαφυρῇcι νεώμεθα
“come, young warriors of the Achaeans, let us sing a beautiful paean and return to the hollow ships”
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:
γηθήcειc ὁρόων μάλα γὰρ θεοείκελοc ἔcται
“you will delight at the sight, for he will be very god-like”
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.
Here, then, is an indication that Sappho was intensely aware of epic diction in general and of the Iliad in particular. The metrical and formulaic repertory of her pentameter is cognate with that of the Homeric hexameter, and the consequent structural similarities in the two genres present manifold opportunities for allusion. Using parallel traditional material, the poetess can highlight or shade well-known Panhellenic epic passages. But Sappho’s medium can remain her own. Accordingly, we may analyze the inherited repertory of her medium from the standpoint of its own internal archaisms as well as its own internal innovations.


[ back ] 1. Kakridis 1966:24: “Hat das schwer geprüfte Paar wahrhaftig keinen einzigen fröhlichen Tag jemals erlebt? O ja! Die Phantasie geht in die Zeit zurück, wo Andromache im Elternhaus als einzige Tochter unter sieben Brüdern glücklich aufwuchs und eines Tags ein schöne Prinz kam und sie als Gattin in seine Vaterstadt führte.” The article from which this excerpt was taken (Kakridis 1966) provides an admirable literary appreciation of Sappho 44LP, with useful bibliography. See also Weber 1955:93-100.
[ back ] 2. See pp. 99-101.
[ back ] 3. P. 23 Consbruch.
[ back ] 4. On the overall matter of traditional diction in Sappho, see Page (1955), who has regularly collected phraseological parallels in analyzing the major extant Sapphic fragments.
[ back ] 5. For the basics on Hermann’s Bridge, see again pp. 72f.
[ back ] 6. See O’Neill 1942.
[ back ] 7. On the problem of the long final in πορφύρᾱ, see Page 1955:68f. On the basis of adjectival πορφύρῳ at 98.4LP, which scans – –, I would interpret our form as πορφύριᾱ (also πορφύριῳ), with neuter plural in long -ᾱ instead of the prevalent Greek pattern -ᾰ. There is ample comparative evidence for an original variation ᾰ/ᾱ (from *a2/*ea2) in the neuter plural; the regular form is actually -ā in Indic and Slavic (Germanic has both *-ă and *-ā). Notice that Sappho preserves traces of an original variation ᾰ/ᾱ in the first declension (Hamm 1958:147). For more on * ᾰ/ᾱ in the neuter plural, see Schwyzer 1939:580-582.
[ back ] 8. See again Page 1955:68f; also note 7 above.
[ back ] 9. See pp. 50-56.
[ back ] 10. See pp. 67-102.
[ back ] 11. See pp. 104-109.
[ back ] 12. Maas 1962:59.
[ back ] 13. Cf. Frisk II/1970:460f.
[ back ] 14. Intervocalic *-(y)y- is unstable: cf. ποέω vs. ποιέω. For Homeric Παιήων, see below. In connection with Lesbian Πᾱ́ων, cf. Λέcβιον παιήονα in Archilochos 121W.
[ back ] 15. For a clear and concise discussion, see Ventris and Chadwick 1959:311f.
[ back ] 16. Ε 401, 899f.
[ back ] 17. E.g., Y 69. For the Cretan associations of the name Παιήων (vs. Panhellenic Apollo) see Pagliaro 1951:14, commenting on Hymn to Apollo 516.
[ back ] 18. Cf. Frisk II/1970:460f; also Harvey 1955a: 172f.