DIRK OBBINK

T WO N EW P OEMS BY S APPHO

aus: Zeitschrift für Papyrologie und Epigraphik 189 (2014) 32–49

© Dr. Rudolf Habelt GmbH, Bonn

32

TWO N EW P OEMS BY SA PPHO

The recorded reception of Sappho begins with Herodotus. At 2.135 he documents a song (ἐν μέλεϊ) in
which Sappho criticized her brother Charaxos or his mistress. A trader in Lesbian wines, he conceived
a violent passion for a notorious courtesan, then a slave at Naukratis, sailed to Egypt, ransomed her at a
great price, at which Sappho gave vent to her indignation in a song. Herodotus’ account is re-told, with
variations and corrections by several later authors.1 Charaxos, if we may believe Ovid, took no less offense,
turned back to sea, rejecting all Sappho’s assiduous advice and pious prayers. Grenfell and Hunt, in the
first non-biblical papyrus published by them from Oxyrhynchus, thought they had identified part of a relat-
ed poem, Sa. 5 (P. Oxy. 7), and later made a similar link with what we now call Sa. 15 (P. Oxy. 1231 fr. 1
i 1–12 + fr. 3) – although neither text names Charaxos, nor is it even certain that the ‘brother’ or ‘sister’
now certainly mentioned Sa. 5 are Charaxos or Sappho respectively. The very existence of Charaxos and
his lover in Sappho’s poetry has been doubted by many scholars. The earliest author to mention Charaxos,
after Herodotus, is Posidippus in a third century BC epigram, who describes Sappho’s poetry as showing
both Charaxos and girlfriend (there already called Doricha) in a benign light, notwithstanding an element
of irony, which is as uncertain as it is untrustworthy; then Ovid. Her. 15.17–18 Charaxus / frater. We quite
simply have had no clue, up until now, as to the kind of information, or its source, that could have given rise
to Herodotus’ story in a way that his fifth century Athenian audience might have found credible.
A newly uncovered papyrus2 changes that, offering parts of two new poems by Sappho – one that men-
tions prominently Charaxos and his trading at sea, barely overlapping with P. Oxy. 2289 fr. 5,3 and another
that is an address to Aphrodite employing tropes familiar from Sappho’s love poetry elsewhere, substantial-
ly overlapping with, and supplying more of P. Oxy. 1231 fr. 16 (Sappho fr. 26 Voigt).4 Both poems clearly
come from Sappho’s first book, where they stood in close proximity with Sa. 5 and 15. In what follows I
refer to the first as the Brothers Poem, and the second as the Kypris Poem in shorthand designations for
poems which as yet have no fragment numbers assigned to them in any edition.

The Papyrus
A large fragment (176 × 111 mm) with the better part of the upper portion of a single column from a papy-
rus roll, written along the fibers, containing parts of two poems (20 and 9 lines respectively) in Sapphics
in the Aeolic dialect. Top margin survives to a height of 31 mm, left margin to c.5 mm, right margin to
c.5 mm. No bottom margin is visible. Twenty-two lines are preserved in their entire length. Seven lines
at the bottom lack three to six letters from the beginnings and ends of lines; of the last line there are only
negligible traces. The text is arranged in characteristic Sapphic stanzas consisting of four lines each, the
fourth line (adonaean) being notably shorter than the first three. Column drift to the left at the left margin
(‘Maas’ Law’) is present in both the upper and lower halves of the column.
The papyrus is written in a formal round hand with informal connection of the third century AD. Cor-
rections are both by the main scribe and by a similar contemporary hand, who probably added the accents.
1 Strab. 17.1.33 (p. 808, 16ff. C.); Athen. 13.596b–d (disputing the name of the lover, and adducing Posidip. XVII Gow–
Page = 122 Austin–Bastianini); Ovid. Her. 15.63–70, 117–20; P. Oxy. XV 1800 fr. 1.1–35 (Sa. test. 252 Voigt = Chamael. 29 3T
CPF (vol. I 1* pp. 406–9)); Sud. αι 334 Αἴϲωποϲ, ι 4 Ἰάδμων, and ρ 221 Ῥοδώπιδοϲ ἀνάθημα.
2 P. Sapph. Obbink (Fig. 1), now in a private collection, London. I am grateful to its anonymous owner for access to and
permission to publish the papyrus and its text here. I owe a further debt of gratitude for comments and criticisms to L. Benelli,
J. Hammerstaedt, R. Kassel, and J. Lidov, as well as to S. Burris and J. Fish. For reasons of space, I describe the conservation of
the papyri as part of a separate, forthcoming study. Conventions of reference: Sa. 1, Alc. 1 = fragment 1 of Sappho or Alcaeus
according to the edition of Voigt (unless otherwise indicated).
3 Fig. 2. Ed. pr. Lobel, in E. Lobel, ed., The Oxyrhynchus Papyri part XXI (London 1951) 2–6 at 3, 5 with plate I.
4 Fig. 3. Ed. pr. Hunt, in B. P. Grenfell and A. S. Hunt, eds., The Oxyrhynchus Papyri part X (London 1914) 20–43 at 30–3,
43. Both overlaps were observed by Professor Burris.
 Two New Poems by Sappho 33

For parallels see P. Oxy. III 412, no. 23a in C. Roberts, Greek Literary Hands (Oxford 1956), containing
Julius Africanus’ Kestoi, which was not composed before 227 AD, and having a document of 275–6 on
the back; Dura Parchment 24 in Roberts, Greek Literary Hands no. 21b, containing Tatian, Diatessaron,
datable between 172 and 256 AD.
The handwriting (as well as format and line-spacing) is identical with P. GC. inv. 105 (see note 6 below).
A kollesis is visible running along the right edge of the papyrus, so that it cannot have formed part of the
same sheet as P. GC. inv. 105 frr. 2–3 (containing Sa. 16–17, perhaps 18 and an unknown poem, and Sa. 5),
but is likely to have come from a sheet that stood directly either before or after this sheet. Occasionally, in
places, ink-traces are obscured by spots of adherent material that appears to be light-brown gesso or silt,
specs of which are also to be seen on the back. The top portion of the column was detached horizontally
(perhaps by ancient damage?), but has been reattached in modern times. On the back, there is evidence of
ancient repair along vertical stress-lines that in places have stretched the fibre-structure, with resulting dis-
tortion of alignment of ink-strokes on the front along this vertical band. The roll was apparently damaged
here in ancient times (torn vertically up the middle from the bottom of the column, just to left of center) and
repaired in antiquity with thin strengthening strips of papyrus glued horizontally and vertically.5
The scribe marks punctuation occasionally by space, more often by middle or raised point, sometimes
placed after writing of the text, but more often at the time of writing. An organic diaeresis is written in
line 1 αλλαϊ, which specifies the disyllabic division necessary for the meter, and helps to exclude ἄλλαι,
but does not clarify whether ἀλλ(ὰ) ἄϊ or ἄλλ(α) ἄϊ is meant. The text is written verse for verse (the
adonaean added on a separate line), with final vowels consistently elided before words beginning with a
vowel, occasionally marked with an apostrophe (usually by the second hand). Iota adscript is written in 2
νᾶϊ ϲὺμ πλέαι: it is omitted in 26 if ]ρω is dative and not genitive (see 26n., cf. 17n. on ἀέρρη). It is written,
however, at P. GC. inv. 105 fr. 2 i 21 (Sa. 16.23), if ἔμ’ αὔται has been correctly read and articulated there,
and at fr. 1.2 (Sa. 9.4) ὤ⌈ραι and fr. 2 ii 11 (Sa. 17.3) Ἀτρ⌋[είδα]ι, if these last two are datives and not nomina-
tive plurals; at fr. 2 i 12 (Sa. 16.14) ν⌈οήϲη⌉ι· and fr. 3 ii 12 (Sa. 5.3) ϝ⌈ωι the iota adscript is uncertainly read.
The scribe assimilates consonants (2 ϲὺμ πλέαι, 3 ϲύμπαντέϲ). Corrections have been entered by a second,
different but roughly contemporary hand. Of the 2 corrections, both are certainly or arguably right; the one
variant added (in 14, without cancellation) is also probably to be preferred over the first reading. There is at
least one uncorrected error, apparently in 26 (λυ{ }̣ ϲαντι). Division between stanzas is marked at the left
margin, where preserved, by a paragraphus after the fourth, short line (adonaean), except at end of poem, at
20, where this is replaced by a decorated (i.e. ‘forked’) paragraphus or diple obelismene (functioning as a
coronis), in the shape described by E. G. Turner, Greek Manuscripts of the Ancient World, ed. 2 rev. by P. J.
Parsons (London 1987) 12 n. 60, i.e. with a leg descending diagonally at the left from a horizontal line that
might have otherwise been drawn as a paragraphus, and without any additional space between the lines.
Afterwards there follow nine further verses from the beginning of a hitherto unknown poem.
Accentuation occurs three times, apparently drawn by a second hand in lighter-colored ink, sometimes
crudely. Of the two clear cases of accent, at least one can best be explained as instances where the reader’s
attention was being drawn to accentuation particular to the Lesbian poets.

Content
Metre, language and dialect, and references (in 1 and 8) to Charaxos (known from the ancient tradition as
one of Sappho’s two elder brothers) and (in 18) to Larichos (supposedly Sappho’s younger brother) – togeth-
er with overlaps with the text of at least two previously published papyri (P. Oxy. 1231 fr. 16 and 2289 fr. 5)
and the link with inv. P. GC. inv. 105 – point indubitably to a poem by Sappho from her first book. Ovid
(Her. 15.67–8) says that she ‘advised him extensively’, with good intentions, freely, but with pious speech:
5 For other ancient instances of repair of books see E. Puglia, La cura del libro nel mondo antico. Guasti e restauri del
rotolo di papiro (Napoli 1997); id., Il libro offeso. Insetti carticoli e roditori nelle biblioteche antiche (Napoli 1991); G. Men-
ci, Fabbricazione, uso e restauro antico del papiro: tre note in margine a Plinio, NH XIII 74–82, in Proceedings of the XVIII
International Congress of Papyrology, Athens, vol. 2 (Athens 1988) 497–504.
34 D. Obbink

me quoque, quod monui bene multa fideliter, odit; / hoc mihi libertas, hoc pia lingua dedit. In the Brothers
Poem, a speaker addresses someone, criticizing this person for ‘always chattering about ‘Charaxos’ coming
with a full ship’, saying further that the addressee does not heed what Zeus and the other gods know, instead
of sending her to pray to Queen Hera for a safe return for Charaxos, piloting his boat, to find ‘us’ safe and
sound. ‘Everything else’, s/he continues, ‘let us leave to the gods: fair weather comes of a sudden out of a
great storm. Those who, favoured by Zeus, get a special helping δαίμων to release them from their troubles
become completely happy and blessed.’ The poem closes with well-wishing for Larichos, that he grow up to
be a settled member of the leisured, aristocratic class and so ‘release us from many sufferings’. This poem
is then followed in the papyrus by another, previously unattested poem addressed to ‘Kypris’ on the subject
of love, with some recognisable similarity to the theme and phrasing of Sa. 1 and several other fragments
of the Lesbian poets. Although there are sufficient fragmentary and uncertain readings here to warrant
proceeding with caution, enough of the content can be distilled from the remaining lines to reconstruct a
poem on the subject of unrequited passion. The succession of poems shows a specific sequence, i.e. a love
poem following upon a poem about the Brothers. Given the meter, both poems were presumably from book
one of Sappho, all the poems of which were in the same meter (Sapphic strophe). All the poems of Sappho’s
first book seem to have been about family and cult, on the one hand, together with poems about passion.
Sa. 5 and the Brothers Poem contain many of these elements: cultic addresses to divinities (Nereids, Hera,
Zeus; possibly Aphrodite and Dionysos), a brother’s wanderings, sisterly affection and loyalty, Sappho’s
and others’ relations with him and the community, a hoped-for shift from bad times to good. The Kypris
Poem, as far as we can tell, is almost exclusively concerned with love/Aphrodite, although there may be an
allusion to cult in the address to Kypris. The poem may have also contained biographical details, perhaps
conventional, in documenting the poet’s own feelings for someone and encounter with divinity in the past
or present.

Relation to Sappho fr. 5

The similarity of content and mention of Sappho’s brothers in the Brothers Poem point to a connection
with Sa. 5, where a brother and sister are mentioned but not named. The language of the Brothers poem at
7 ἐξίκεϲθαι τυίδε seems to replicate that of Sa. 5.2 τυίδ’ ἴκεϲθα[ι (cf. Brothers Poem 19n.). Many scholars
have thought that Sa. 5 ends after the fifth stanza at line 20, after the invocation of Kypris at lines 18–20. In
P. GC. inv. 105, this comes at line 29 in its column, the last line of the fragment, which contains the greatest
number of lines of any fragment of P. GC. inv. 105, and the same number of lines as the present fragment.
Whether further lines followed in either column is unknown, since in both fragments the text breaks off
at the bottom before end of column. Based on P. GC. inv. 105 fr. 2 col. ii and analysis of the continuity of
Sa. 16, there could be 0–2 stanzas following Sa. 5 after the fragmentary end of fr. 3 col. ii before the end
of the original column. The present fragment cannot then be the following column – at least one column
must have intervened, since a kollesis or sheet-join is visible along its far-right edge, whereas the preserved
part of Sa. 5 in P. GC. inv. 105 stands at the far right edge of a sheet c.30 cms in length that contained frr.
2–3 and comprises parts of four columns. This seems reasonable, since the intervening material (right,
missing part of fr. 3 ii + intercolumnium + missing column + intercolumnium) added to the width of
P. Sapph. Obbink amounts to c.30 cm, the apparent length of the preceding sheet. It is again possible that
the papyrus preceded P. GC. inv. 105 frr. 2–3 in the roll (although see below n. 6). If it did, it was certainly
followed at least by the Kypris Poem, and then Sa. 15 (also probably about Charaxos) before P. GC. inv. 105
fr. 2 i, which begins in the middle of Sa. 16 and is followed there by Sa. 17, possibly 18 (a sequence that is
also known from the copy of Sappho book 1 in P. Oxy. 1231), and finally by another poem before Sa. 5. If
it did directly follow P. GC. inv. 105 frr. 2–3, at least one column intervened containing at least one poem
plus the beginning of the Brothers poem – both of them together in at least seven stanzas. Since line 1 of
the Brothers Poem cannot be its opening, there must have been at least one and possibly two or even three
stanzas of it beginning in the preceding column. The last would bring it up to the maximum known number
of stanzas in a Sapphic poem from book 1 (Sa. 1 = seven stanzas).
 Two New Poems by Sappho 35

A connection between the Brothers Poem and Sa. 5, now known to begin with an invocation of the
Nereids alone6 (it then mentions a brother, and finally invokes Kypris), is the context of strong personal and
family and social and civic relations, emphasis on social and cultic behavior, civic and cosmic justice, and
the threat of loss of security or safety through seafaring, in which one of or even both brothers Charaxos
and Larichos are engaged. There are also a number of fragments which contain themes of sea-faring (as
in Alcaeus, famous for his use of sea-faring as allegories). In the bare remains of Sa. 20, on a dangerous
sea-journey, there occur elements that concur with ideas in the Brothers Poem, Sa. 5, and Sa. 15: see Sa.
20.4 τ]ύχαι ϲὺν ἔϲλαι, 5 λί]μενοϲ, 6 γ]ᾶϲ μελαίναϲ, 8 ναῦται, 9 μεγάλαιϲ ἀήται[ϲ, 12 πλέοι.[, 13 τὰ φόρτι’
εἰκ[.7 It is at any rate clear that not all of these come from one and the same poem, whether or not any
of them mentioned by name Doricha or Rhodopis or Charaxos, as the Brothers Poem does now the last.
At least one of Sappho’s poems, reflected in a testimonium (Athen. 10.425a = Sa. test. 203 Voigt) almost
certainly named Sappho’s younger brother Larichos, who is now mentioned at Brothers Poem 18, acting as
a boy in the honorary capacity of a wine-pourer for the banquets of the civic elite of Mytilene. We seem
to have a cycle of poems centering on the family’s role in civic, social, and business life and frequently
employing images of sea-faring, viticulture, and wine-trading.

The New Poems and Sappho Book 1

The two new poems instance types that can be seen as alternating in book 1 of Sappho. The book contained
poems about (i) family and biography and/or cult, for adults, and (ii) unrequited passion, for adolescents,
while offering a window into adult experience, since the speaker repeatedly represents herself as having
experienced these things in the past.
At the end of the Brothers Poem, Sappho hopes that Larichos will grow up to become a man (18). ἄνηρ
is a rare word in Sappho (common in Alcaeus). Apart from the τέκτονεϲ ἄνδρεϲ of Sa. 111.3, there is the
god-like groom of Sa. 111.5–7, the god-like man of Sa. 31 who may be in the husband’s role; Menelaus in
both Sappho (Sa. 16) and Alcaeus (Alc. 42), as the good husband Helen left; men who sing a pious hymn
to godlike Hector and Andromache in Sa. 44.32–4; and in Alcaeus various men who are rich, powerful,
distinguished, power-hungry, hard-drinking or even wicked: that is, men of the political and sympotic
class, like the hard-drinking ἄνηρ of Alc. 72. Presumably, Larichos, who might have poured wine in the
prytaneion (Sappho fr. 203a Voigt) as a youth, should grow up to assume the status of those whom he had
the honor of serving, supported by the family wealth – free to live his life as a member of the leading class.
The point is not that Larichos should survive and grow up: he should become an ἄνηρ in all senses.
Presumably this would include aristocratic demeanor, noble marriage, transfer of wealth, and production of
legitimate offspring, all of which could be threatened by Charaxos’ not arriving ‘with a full ship’ (perhaps
after squandering the whole cargo on a courtesan in some port), and failing to come home with a ship laden
with goods, spices and perfume, traded in kind for wine.8
In the Brothers Poem as we have it, Sappho challenges the addressee (and by extension, her audience)
to remember that Charaxos’ success and safety is in the hands of the gods and attainable (if at all) only
through the correct form of prayer in song. Against this is held up universal knowledge of all the gods and
the cosmos by the speaker, and the power of hymnic song, framed in the poem, to help secure Charaxos’
safety, as well as the safety and prosperity of the family or community. Perhaps Herodotus’ readers and
Hellenistic scholars even imagined that, if Charaxos were innocent of any wrong-doing, he would not need
6 P. GC. inv. 105 fr. 3 ii 10: see S. Burris, J. Fish, and D. Obbink, New Fragments of Book 1 of Sappho, ZPE 189, 2014,
1–28 at 11, 23, thus confiming alphabetical arrangement by first-letter (Lobel, Σμ xv); so also now Kypris poem 1 πῶϲ.
7 F. Ferrari, Sappho’s Gift: The Poet and Her Community (Ann Arbor, Michigan 2010) 159 n. 24 compares Sa. 20.4 τ]ύχαι
ϲὺν ἔϲλαι with 15.7 ϲὺν] τύχαι; Sa. 20.5 λί]μενοϲ with 15.7 λίμενοϲ; Sa. 20.8 ναῦται with 15.6 ναυβ]άταιϲ[ι. The connection
between Sa. 5, Sa. 15 and Sa. 20 was noted already by W. Schubart, Bemerkungen zu Sappho, Alkaios und Melinno, Philologus
97 (1948) 311ff., 315 and W. Schadewaldt, Sappho. Welt und Dichtung. Dasein in der Liebe (Darmstadt 1950) 137. For Sa. 20
see also H. J. M. Milne, A Prayer for Charaxus, Aegyptus 13 (1933) 176–8.
8 Cf. Alc. 117b.26–7 who raises the problem; further: A. Aloni, Eteria e tiaso: I gruppi aristocratici di Lesbo tra economia
e ideologia, DArch Ser. 3,1 (1983) 21–35.
36 D. Obbink

Fig. 1
Image copyright © Imaging Papyri Project, Oxford – all rights reserved
 Two New Poems by Sappho 37

her prayers. She looks forward to his thus returning to Lesbos, and imagines and advises how one would
have to pray to the gods to secure his safe passage, and in the course of which she herself does so.
The Kypris Poem that follows in the new papyrus is less well preserved and of uncertain reconstruction
in more than one place. Nonetheless, it substantially augments Sappho fr. 26 Voigt (as previously known
from P. Oxy. 1231 fr. 16), previous reconstructions of which it corrects as it invites new ones. As such, it can
be seen to exhibit some of the motifs and language familiar from Sappho’s love poetry elsewhere in book 1
and in Sapphic metre, while introducing examples of diction and forms of expression previously unattested
in the Lesbian poets.
Diplomatic text Articulated and corrected text
top of column top of column
αλλαϊθρυληϲθαχαραξονελθην ἀ⌈λλ’ ἄϊ⌉ θρύληϲθα Χάραξον ἔλθην
-
ναϊϲυμπλέαι·ταμε ο̣ ι μ
̣ ι̣ ζ υϲ ν⌈ᾶϊ ϲὺ⌉ν πλήαι. τὰ μὲν οἴομαι Ζεῦϲ
οιδεϲυμπαντεϲτεθεοι·ϲε’δ ο̣ υχρῆ οἶ⌈δε ϲύ⌉μπαντέϲ τε θέοι· ϲὲ δ’ ̣ οὐ χρῆ
ταυτανοειϲθαι· τ⌈αῦτα⌉ νόηϲθαι,
__
5 αλλακαιπεμπηνεμεκαικελ⟦η⟧`ε΄ϲθαι ἀλλὰ καὶ πέμπην ἔμε καὶ κέλ⟦η⟧`ε΄ϲθαι
πολλαλ ϲ̣ ϲεϲθα β̣ αϲ ̣ ̣ α
̣ ν ρ̣ αν πόλλα λίϲϲεϲθαι βαϲίληαν ῎Hραν
εξικεϲθαιτυιδεϲααναγοντα ἐξίκεϲθαι τυίδε ϲάαν ἄγοντα
νααχαραξον· νᾶα Χάραξον
__
καμμεπευρηναρ ̣ μ ̣ εαϲ·ταδαλλα κἄμμ’ ἐπεύρην ἀρτ°μεαϲ. τὰ δ’ ἄλλα
10 πανταδαιμονεϲ ι̣ νεπ ̣ ρ̣ οπωμεν πάντα δαιμόνεϲϲιν ἐπιτρόπωμεν·
ευδια ̣ α
̣ ε̣ κμεγαλαναητ ̣ `̣ ΄̣ εὔδιαι γὰρ ἐκ μεγάλαν ἀήταν
αιψαπ ̣ ̣ ̣νται· αἶψα π°λονται.
__
τωνκεβολληταιβαϲιλευϲολυμπω τῶν κε βόλληται βαϲίλευϲ Ὀλύμπω
δάιμονεκπονωνεπαρη`ω΄γονηδη δαίμον’ ἐκ πόνων ἐπάρ{η}`ω΄γον ἤδη
15 ̣ καρεϲπελονται
περτροπηνκηνοι α περτρόπην, κῆνοι μάκαρεϲ πέλονται
καιπολυολβοι· καὶ πολύολβοι·
__
̣ μμεϲαικετανκεφαλ ̣ναερ η
α ̣ κἄμμεϲ, αἴ κε τὰν κεφάλαν ἀέρρη
λαρ χ̣ οϲκαιδηποταν ρ̣ γενηται. Λάριχοϲ καὶ δή ποτ’ ἄνηρ γένηται,
καιμαλεκπολλ⟦η⟧`αν΄βα υ ̣ θ μ̣ ια κ̣ εν καὶ μάλ’ ἐκ πόλλ⟦η⟧`αν΄ βαρυθυμίαν κεν
20 αιψαλύθειμεν · αἶψα λύθειμεν.
__
/
πω κ̣ εδητιϲουθαμε ̣ α ̣ ϲαι ο̣ πῶϲ κε δή τιϲ ου θαμε ̣‒̣ αϲαιτο
κυπριδε ̣ ̣ ̣ ̣ ̣ ’ ̣ ο ̣ ̣ ̣ [̣ ̣ ]̣ φ
̣ ̣ [̣ κυπριδε ̣ ̣ ̣ ̣ ̣ ’ ̣ ο ̣ ̣ ̣ [̣ ̣ ]̣ φ̣ ̣ [̣
̣ ̣ ]̣ θελοιμαλιϲτα ̣ ̣ ̣ ̣ κ̣ α [̣ ̣‒̣ ]̣ θέλοι μάλιϲτα ⏑ ̣ ̣ ̣‒̣ κ̣ α [̣ ‒ ×
̣ ̣ ]̣ ονεχηϲθα ̣‒̣ ]̣ ονεχηϲθα
__
25 ̣ ̣ ]̣ α ̣ λοιϲ μ ̣ ’αλεμ ̣ ̣ ̣ ̣ α ̣ ¨ ̣ ̣ [̣ ̣‒̣ ]̣ α ̣ λοιϲ ̣ μ’ ἀλέματ ̣ ̣ δαΐϲδ[
̣ ̣ ]̣ ρωλυ ϲ̣ αν ̣ ̣ ο̣ ̣ ̣ ̣ ε̣ ̣ [̣ ̣ ̣ ]ρω λυ{ι}ϲαν ̣ ̣ ο̣ ̣‒̣ ̣ ε̣ ̣‒̣ [ ×
‒̣ ⏑
̣ ̣ ]̣ α ̣ α ̣ ̣ [̣ ̣ ]̣ ̣ μ̣ ̣ ̣ ̣ ο̣ [] ε̣ ρηϲ[ ̣‒̣ ]̣ α̣ α ̣ ̣ [̣ ̣ ]̣ ̣ μ̣ ̣ ̣ ̣ ο̣ [] ε̣ ρηϲ[
̣ ̣ ]̣ ν ε̣ ρ [̣ ̣ ]̣ ̣ ̣‒̣ ⏑]̣ νεερ [̣ ̣ ]̣ ×̣
__
c.8 ] ̣ ̣ [̣ c.8 ] ̣ ̣ [̣
. . . . . . . . . . . . . . . . . . . .
38 D. Obbink

Palaeographical Notes (to the diplomatic transcript)

2 init. ante ν: horizontal at letter-top height beginning slightly into margin and connecting to top of upright
of ν first ,̣ tip of pointed upright followed by upper half of serifed upright as of ν (but missing finial on
first upright) or μ, in both cases with loss of middle part in crack second ,̣ top, left and right sides of
round letter (middle and bottom left occluded by adherent material): ε, θ, ο third ,̣ diagonal arcing over
left at top and toward right at bottom, with vestige of low circle underneath, prima facie ο, but no other ο
has connecting stroke at bottom right, α possible 3 ’δ ,̣ apostrophe just before and at the level of the
apex of δ, followed by slight trace of another added again just above the apex of δ (but cf. δ in οιδε earlier
in this line) 6 first ,̣ pointed top of upright second ,̣ top half of upright ̣ ̣ ,̣ top half of upright,
followed by top of diagonal sloping left, then upright with serif left at top, horizontal ink at mid-level con-
necting to lower half of an upright, η suggested last ,̣ top of upright followed by trace at mid-level, then
̣ ̣
top of upright above base of upright 9 , horizontal stroke at top-line with upright descending from right
end, then, after half a letter-space, a short horizontal at top-line with dot at mid-level and short up-curving
stroke ranged beneath it 10 ,̣ curved back of a round letter, missing its top, followed by trace at mid-level
as of mid-stroke of ε or end of down-curving top of ϲ ̣ ,̣ squarish shape as though a backwards ϲ, with
a trace at mid-level between the open parts at the left side (but too small and square to be ο, possibly a wide
but short horizontal stroke, as of ι, with its middle abraded away and now free of any ink; then upright with
serifed foot and horizontal stroke protruding right from top of upright, suggesting γ, but with slight over-
hang of horizontal to left of upright as of τ 11 α ̣ ,̣ after α two uprights, the second with a horizontal
stroke across the top just slightly to the left and a little further to the right: ιγ or π both possible, then trace
at bottom line, as of upright α ,̣ upright with ink curving right in a small circle at the top, but not closed,
ρ possible, but with horizontal ink at mid-level just below, ε suggested, or perhaps a connection stroke from
middle of ρ ̣ `̣ ΄̣ , first ,̣ back of a round letter, surface obliteration, then diagonal sloping left, and above
this a trace of a diagonal above the line, perhaps ω begun then cancelled or changed into α second ,̣
upright (almost touching preceding diagonal at bottom), then a slightly detached short diagonal sloping
right and crossing at bottom with another diagonal sloping right at top, as though a stretched ν (cf. ν at end
of 6) 12 ̣ ̣ ,̣ round letter with horizontal stroke inside at mid-level as of ε θ, then trace in top left quad-
rant and trace in lower right quadrant at base-line, then right half of round letter ε θ ο ϲ (not ω) 15 ,̣ two
uprights, the first curving left and the second right at bottom, suggesting μ missing its middle 17 first ,̣
upright, then trace in lower right quadrant at base-line second ,̣ very top of upright with hook over left
as α δ λ, with flecks of ink below, perhaps bottom of diagonal sloping upwards from left to right in lower
left quadrant third ,̣ upright with slight finial on top and bottom and apparently straight, not rounded
vertical stroke running right from top and slightly curving down at end as of ρ if not closed at top and mid-
dle of upright, or γ 18 first ,̣ bottom half of upright second ,̣ trace of bottom of upright, trace of
horizontal at mid-level on left, connected by specks to upright 19 first ,̣ right side and upper left quadrant
of small bowl in upper half of letter-space, with a short horizontal stroke at the base-line directly under-
neath second ,̣ very short horizontal stroke at top-line in upper left quadrant, short diagonal rising from
mid-level to top-line in upper right quadrant third ,̣ upright with diagonal sloping right from top, then,
after a short space, ghost of upright with ink-trace at base-line 21 first ,̣ top and left side of round letter,
as ε ϲ ̣ ,̣ arc in upper left quadrant with two thin horizontal strokes extending to right from top (the
first perhaps attached to the top of the arc), then thin vertical stroke at mid-level suggesting ι, but compati-
ble with mid-left side of wide ω, followed by arc as of left side of ε θ ο ϲ, not closed at right, ostensibly with
a short thin horizontal connecting from left (if not the remains of thickness of shading of the left side of
round letter), and with a short slightly curved thin stroke at the topline, compatible with top of ϲ or ε ι ,̣
after ι, high horizontal with specks of ink of decreasing size ranged beneath it in a vertical band as τ, less
likely (in this hand) ξ 22 ε ̣ ,̣ after ε, curve in upper left quadrant as of ε θ ο ϲ possibly with horizontal
ink attaching at mid-level, then short horizontal stroke at top-line with slight trace at the line of writing as
of finial on base of upright ̣ ̣ ̣ ’,̣ round letter, followed by bottom of upright, then vague vertical ink,
after which apostrophe? (higher than the others, at about mid-way between the lines) in second
 Two New Poems by Sappho 39

hand ̣ ̣ [̣ , upright with horizontal element protruding left followed by upright bowed inward, both
elements topped by a horizontal stroke (slightly lower than the first) that may be connected, then an upright
as of ι, followed by another upright with horizontal strokes protruding left from top (and perhaps also from
middle) ] ,̣ horizontal stroke at mid-level connecting to upright, prima facie right half of η, but with a
lower horizontal than in η elsewhere φ ̣ [̣ , after φ, upright, as of ι or right half of η, then tip of vertical
or left-leaning diagonal stroke at the top-line (after which surface stripped) 23 ̣ ̣ ̣ ̣ ̣ ,̣ right and left part
of horizontal stroke at top-line with upright descending from the first as of π τ (the horizontal extending
further left from the upright than in many but not all instances of π, thus suggesting τ, while not excluding
π), then two diagonals converging at top with horizontal stroke at mid-level as of α, with foot of second
diagonal extending horizontally at base-line and perhaps curving upwards, then fleck of ink at mid-level,
then two successive flecks of ink at bottom-line, perhaps tips of verticals as of η ν π [̣ , trace of diagonal
sloping upwards right at base-line as of α δ λ 25 ] ,̣ trace at far right just above the line of writing, α β
δ ε λ ϲ χ, then after ϲ, dot on the line close in to ϲ and narrow, as of ι μ ̣ ,̣ trace at base-line close-in to
μ, then traces of a diagonal sloping from upper left to lower right as α δ λ, followed by bottom half of
upright with right and left ends of horizontal stroke at top-line centered above it, τ suggested ̣ ̣ α
̣ ,
before α, round letter as ο followed by short upright or left lobe of ω, after which scattered traces, perhaps
more suitable to remains of a round letter than a square one, then left side of a triangular letter with hori-
zontal ink at base-line as of δ ¨ ̣ ̣ [̣ , diaeresis over space for narrow letter, missing but presumably ι, left
side of round letter not closed at right, ϲ suggested, then diagonals converging at top, with horizontal ink at
base-line, as of δ 26 first ,̣ two traces ranged one above the other, one at base-line and one at mid-level,
upright? second–third ,̣ two uprights connected at top by horizontal ink, τι or π?, followed by high point
above the line fourth ,̣ horizontal ink with curve in lower left quadrant beneath, ϲ or γ or first half of
π? after ο, scattering of traces at mid-level, upright or triangular letter?, then two uprights at mid-level,
the first with horizontal ink protruding right at mid-level, followed by indistinct traces, second at mid-level,
third at upper-level final ̣ ,̣ on upper fibres: upper part of round letter with mid-stroke, not closed at
right, followed by top of upright with diagonal descending to second upright, εν suggested; on lower fibers:
bottom of round letter followed by three successive bottoms of uprights 27 ] ,̣ trace at far right at base-line,
αβλδχ ,̣ short horizontal at top-level turning down into upright, followed by upright with horizontal
connecting tops of the two uprights, π more likely than τι ̣ [̣ , connecting stroke/finial connecting to
bottom of upright curving right at top-level, ϲ?, then top of diagonal in upper-left quadrant as of χ or λ or
α ] ,̣ right side of round letter ̣ ̣ ,̣ three successive tops and bottoms of uprights with feet, followed
by trace of upright at top-level ] ̣ ̣ ,̣ upright with horizontal ink above overhanging upright at right, π?,
then horizontal stroke (joining?) to small bowl, not closed at bottom, as ρ or ο 28 [̣ , trace in far upper
left quadrant at top-level, α λ δ χ π τ ̣
] , end of horizontal finial at top level on top of horizontal as
of ι 29 ] ̣ ̣ [̣ , top of upright, dot on line, top of upright

Translation
Brothers Poem
[…]

but you are always chattering for Charaxos to come

with a full ship. Zeus and all the other gods,
know these things, I think. But it is not necessary
for you to think these things,

5 but especially to send me and command to beseech

Queen Hera over and over again
that Charaxos may arrive, piloting
his ship safe back here,
40 D. Obbink

and find us safe and sound. Let us

10 entrust all other things to the gods:
for out of huge gales fair weather
swiftly ensues.

Of whomever the King of Olympus wishes

a divinity as helper now to turn them
15 from troubles, those men become happy
and richly blessed.

And if Larichos lifts up his head,

and in time becomes an established man,
we would even from our many grievous despondencies
20 be released forthwith.

Kypris Poem
How wouldn’t anyone . . . 1
Kypris … 2
… might especially wish . . . 3
… you have . . . 4

25 … me helplessly divide . . . 5
26 … having destroyed . . . 6

Commentary
Brothers Poem
(P. Sapph. Obbink 1–20; P. Oxy. 2289 fr. 5)
Before line 1 P. Oxy. 2289 fr. 5.1–2 provides remains of two preceding
lines:
. . .
A ̣ ̣ ̣ ̣
]λα[
B ̣ ]̣ ϲέμ [̣
. . .
The continuation of these in lines 3–6 overlaps with Brothers Poem 1–4
(indicated by ⌈ ⌉ in the text above). The fragment was not included by
Voigt in her edition. Hunt notes that A–B ‘appear to be written smaller
than the others’. As such, they might have been notations in an upper mar-
gin before the first lines of a column; however, given the unevenness of the
hand, which appears to be the same as the main text of 2289, and similar
line-spacing, this seems highly uncertain. In addition, the accent over the Fig. 2. Photo courtesy Egypt
ε of ϲεμ would imply that these lines were treated as part of the poetic Exploration Society
text. If so, these would be the first extant lines of the Brothers Poem and
would have stood at the bottom of the preceding column (now lost) in the present roll. In this case A will
have been the third line of a Sapphic stanza (probably missing a long and a short syllable before it), while B
is its continuation the adonaean, both within a letter or so of the beginning of the line. B might be expected
to be part of ϲέμνοϲ: see inc. auct. Lesb. 42.11 (not in Sapphics) and Sa. 5.18 (which cannot be our passage).
However, this is not encouraged by the following trace (not reported by Hunt), a dot on the line which would
exclude ν; in addition, because of its quantity ϲέμν- cannot have stood as the second syllable in the line and
 Two New Poems by Sappho 41

it is not sufficiently far to the right to have been the third. Thus ‒ ̣ ]̣ ϲέ μ [̣ ‒ or ‒ ̣ ]̣ ϲ ἔμ [̣ ‒ or ‒ ̣ ]̣ ϲ’ ἔμ [̣ ‒
should be considered.
1 ἀ⌈λλ’: ]λλά · [̣ P. Oxy. 2289 fr. 5.3 (the raised point actually the first element of the diaeresis over ι,
as suspected already by Lobel). The conjunction ἀλλ(ά) is to be assumed – rather than ἄλλ(α) – together
with a sentence in a preceding stanza in which the speaker says or implies: ‘you should/could (think) x …
but (instead) you are always chattering’ etc. It is uncertain whether the tone is one of reproach or coaxing
or mocking. Cf. Sa. 1 where Aphrodite reproaches Sappho with her own previous words, in a way that
becomes, as here, the prayer of the present poem as spoken by Sappho.
θρύληϲθα: Derogatory, implying either a confused babbling or unharmonious chattering: Theoc. 2.142,
Aristoph. Eq. 348 τὴν νύκτα θρυλῶν καὶ λαλῶν, or to say repeatedly, over and over: Eur. El. 910, often
with connotations of what is gossip or hearsay or commonly said: Demosth. 1.7, Pl. Phd. 65b τὰ τοιαῦτα
οἱ ποιηταὶ ἡμῖν ἀεὶ θρυλοῦϲιν (where it is associated with the song-like power of promulgation of belief
through repetition, or of myth: Isoc. 12.237 τὰ μυθώδη … ἃ πάντεϲ θρυλοῦϲιν) – the point being that the
saying of it is no proof of its truth. Not Homeric, nor in the other lyric poets, this is now its earliest occurrence.
Who is the second person subject being addressed (whether the same or different from the ϲέ in 3)?
While the text may be interpretable without knowing for certain, the possibilities are not unlimited. Assum-
ing Sappho/the poet is speaking, the second person subject of θρύληϲθα might be: (i) a concerned friend/
family member, possibly her mother: Sa. 98.1ff. quotes something her mother (ἀ γάρ με γέννα[τ - - -) said,
apparently authoritatively, in the past; or (ii) Charaxos’ lover (called Rhodopis by Herodotus, but Doricha
by Sappho according to Athenaeus) or another hetaira, here criticized (Herodotus 2.135 attests that Sap-
pho ‘railed violently’ against someone, κατεκερτόμηϲέ μιν, where μιν could refer to either Charaxos or to
the hetaira Rhodopis his lover in a poem concerning Charaxos), (iii) the speaker’s companions or group
(collectively, in the second person singular), or one of them; or (iv) reflective self-address on her own poet-
ic discourse. Another possibility (v) is that θρύληϲθα echoes or even quotes something said by someone
addressing the speaker of the poem as ‘you’, perhaps quoting Sappho’s own words, as Aphrodite does in
Sa. fr. 1; for instances of other voices represented or quoted as addressing Sappho include (indirectly) Sa.
1.15–18 (Aphrodite); (directly) Sa. 1.18–24; Sa. 159 ϲύ τε κἄμοϲ θεράπων Ἔροϲ (addressed to Sappho by
Aphrodite according to Maximus of Tyre or. 18.9 (p. 162 Trapp)). Instances of Sappho speaking, directly
or indirectly, to another: Sa. 1.1–5 addresses, then relates a kind of conversation directly with Aphrodite
(25–8); Sa. 134 ‘I talked <with you?> in a dream, (to?) Kyprogeneia’; Sa. 95.8–13 relates a prayer addressed
directly to Hermes.
(i) seems most likely, since only someone like Sappho’s mother could be expected both to share her
concerns for her brother and to have the authority to direct her to go and pray to Hera. In the case of (ii),
although it is fittingly derogatory, there are problems with Sappho addressing Doricha or Rhodopis at a far-
way place like Naukratis. (iii) has against it that there are no references to a collectivity of companions, nor
references to a chorus or dance; (iv) would entail a serious bifurcation of the poet’s personality into halves
characterized as ‘you’ and ‘me’; (v) introduces the additional complication of quoted dialogue, whereas there
are no indications in the text of change of speaker (presumably necessary after line 1). We at least know of
Sappho’s mother Κλέϊϲ (after whom her daugher was named) from the ancient biographical tradition, and
indeed from Sappho’s own poetry (e.g. P. Oxy. 1800 fr. 1.1–35 = Chamael. 29 3T CPF (vol. I 1* pp. 406–9) =
Sappho test. 252 Voigt; Suda ϲ 107 = Sappho test. 253 Voigt). The reason why her mother herself does not
go to pray to Hera may be old age (Hammerstaedt), or, because the duty was expected to be undertaken by
Sappho going together with her group or companions (Benelli). Sappho may be seen to be remonstrating her
mother in lightly mocking language, perhaps even her own language, as Aphrodite does to Sappho in Sa. 1.
In other words, three different scenarios might be considered:
(i) She responds to another concerned person (her mother?), saying in effect, don’t just keep saying
what you want (i.e. for Charaxos to come), but send me to pray.
(ii) She responds to someone saying to her: ‘[he’s not coming back] but you are always chattering that
Charaxos should come.’ She says: the gods know these things, as you should know. We should pray to them.
42 D. Obbink

(iii) She counters reports of a hetaira who keeps chattering that Charaxos will or has come in the past
(aorist with past time) or should come now – to her ‘with a full ship’, saying to her: don’t tell me what the
gods know and what is wrong for you to think; it’s right for me to still try to save the situation … .
On the whole, the first scenario seems likeliest and in greater conformity with the evidence and Sappho’s
poetry elsewhere. Sappho is engaged in gentle banter with her mother (or another family member) over the
correct attitude to be taken toward her brother’s extended absence, in which she contrasts simple repeated
wishing with the pious ritual act of prayer in accordance with human subordination to the gods’ wishes.
1 Χάραξον: Cf. 9. One of Sappho’s two elder brothers, according to ancient biographers: see e.g. P. Oxy.
1800 fr. 1.1–35 = Chamael. 29 3T CPF (vol. I 1* pp. 406–9) = Sappho test. 252 Voigt; Suda ϲ 107 = Sappho
test. 253 Voigt; Suda αι 334; Tz. Prol. de com. Gr. 2.8 (GCF I p. 26 Kaibel); P. Oxy. 2292 fr. 42(a) = Sappho
test. 213A Voigt (commentary). See also n. 1 above.
ἔλθην: Cf. Aesch. Choeph. 138 ἐλθεῖν δ’ Ὀρέϲτην δεῦρο (discussed below). Does the aorist here mean
‘came’? Or (with one-time aspect) ‘come’? (Cf. Alc. 69.3–4 αἴ κε δυνάμεθ(α) … / εἰϲ πόλιν ἔλθην, ‘in
hopes that we could enter the city’.) The former should refer to a previous, perhaps well-known event; the
latter to one that at least might happen typically and more than once. In the latter case, one approach would
be a non-historical translation of the infinitive so that it represents an original subjunctive or optative in
indirect statement: ‘you are always chattering that Charaxos should come’, or ‘chattering “May Charaxos
come”’ (or ‘chattering for Charaxos to come’, or just ‘chattering: “Charaxos come”’) – just as ἐξίκεϲθαι in
7, since it is unlikely to be causative, would represent a subjunctive or optative as well.
A possible parallel is Aesch. Choeph. 138–9, perhaps even re-working material from the earlier part
of the Brothers Poem (prayer to Zeus?): ἐλθεῖν δ’ Ὀρέϲτην δεῦρο ϲὺν τύχῃ τινὶ / κατεύχομαί ϲοι, καὶ ϲὺ
κλῦθί μου, πάτερ. These lines signal a shift in register, being the only lines in Electra’s prayer that contain
nothing specific to her situation, except Orestes’ name – which is metrically identical to Charaxos here.
πάτερ is here her father, but could in another context be Zeus. Both poets may independently be echoing
the language of prayer for safe return. But given that Electra is praying for her brother’s return, an allusion
looks at least possible, either to the present or another poem about Charaxos by Sappho: some form of
ϲὺν τύχῃ τινί (138) likewise may be more or less similarly phrased at Sa. 20.3 τ]ύχαι ϲὺν ἔϲλαι (as noted
by Milne, n. 7 above, 177) and also very probably at 15.7 (where I suggest that ϲὺν κάλαι] τύχαι may be
restored). The parallel thus tends to confirm that it is Sappho who typically prays ‘may Charaxos come’. In
the Brothers Poem, however, the expression ‘Charaxos come’ in line 1 is not part of a prayer (even if ἔλθην
does represent an original subjunctive), but simply a wish or a statement of what the speaker says has hap-
penbed/hopes to happen, since it is contrasted in the lines that follow by the rebuke about knowledge of the
gods’ ways and recommendation (instead) for pious prayer for Charaxos’ return, a model of which follows.
2 ν⌈ᾶϊ ϲὺ⌉ν πλήαι (correction of the papyrus suggested by M. L. West, cf. Alc. 346.5; articulation
originally suggested by A. Henrichs): ]ᾶιϲυ[ P. Oxy. 2289 fr. 5.4 (where the accent but not the diaeresis,
necessary for the meter, is recorded by the annotator, cf. on 2). For ἐλθεῖν/ἵκεϲθαι ϲὺν νῆι see e.g. Od.
3.61 and 9.173. Unless this refers to Charaxos’ anticipated arriving home ‘with his ship’s crew intact’ (as
Odysseus e.g. notably fails to do) – or whether, for a sea-going trader in Lesbian wines, a ‘full ship’ should
point to arrival back home with his ship packed with luxury goods traded successfully in kind, such as
spices, grain, perfumes, and textiles. Perhaps it could apply to either or both. 7–8 ϲάαν ἄγοντα / νᾶα seems
to voice a parallel (if not identical) concern (except that there the poet goes on in 9 to include concern for
the safety of the family, ‘us’, back at home). One notes the absence of τυίδε here, as is specified below in 7
and at Sa. 5.2, which could, but need not imply that the arrival is to a place other than Lesbos; however, as
a shorthand characterization of what Sappho or someone else has said often in the past, the expression may
be abbreviated, and τυίδε simply may be understood here.
τὰ μέν: ‘Such things’ i.e. referring to the gods’ general knowledge.
οἴομαι: For the parenthetic use (mainly in the first person singular) see LSJ s.v. IV.1 with examples
from Il. and Od., and Aesch. Choeph. 758.
 Two New Poems by Sappho 43

Ζεῦϲ: Rarely named in Sappho (17.9 ∆ί’ ἀντ[ίαον, otherwise only at Sa. 1.2 and 53, in both cases
serving only to designate his descendants genealogically) – although frequently named in Alcaeus,
e.g.129.5 (with Hera and Dionysus) ἀντίαον ∆ία and (in nom., as here) 338.1 ὔει μὲν ὀ Ζεῦϲ. With Hera
in 6 we have two of the three main divinities of the pan-Lesbian sanctuary at Mesa enumerated in Sa.
17 and Alc. 129. For the absence of the third, Dionysus, see on 14.
3 οἶ⌈δε: ̣ ]̣ δ [̣ (‘or: ̣ [̣ ; two dots level with the tops of the letters’) P. Oxy. 2289 fr. 5.5. Professor Bur-
ris observes the parallels at Il. 3.308 Ζεὺϲ μέν που τό γε οἶδε καὶ ἀθάνατοι θεοὶ ἄλλοι (Priam) and Od.
15.523 ἀλλὰ τά γε Ζεὺϲ οἶδεν Ὀλύμπιοϲ, αἴθερι ναίων (Odysseus in disguise), Pi. fr. 94b.33, where these
passages have the sense of ‘maybe’ or ‘God only knows’, as here.
4 τ⌈αῦτα⌉: ]λυ[ P. Oxy. 2289 fr. 5.6 (no further description, but this is clearly the αυ of ταῦτα, visible
even on the darkened surface of the papyrus): presumably referring to τά (line 2), which in its turn refers to
‘the particular things you said’. Correction of the papyrus’ νοεῖϲθαι to νόηϲθαι was suggested by M. L. West.
5 ἀλλὰ καὶ πέμπην ἔμε καὶ κέλ⟦η⟧⸌ε⸍ϲθαι: Literally ‘but also to send me and bid’, presumably under-
standing ἔμε over again with the second infinitive, which is completed by λίϲϲεϲθαι in 6. Obviously πέμπην
cannot be the elaborate lyric metaphor that Pindar constructs with πέμπεται and the Phoenician cargo
at P. 2.67–8. Sappho envisages the addressee as instructing her to go, perhaps with her companions as a
delegation, to pray to Hera for Charaxos’ safe return. Denniston, Greek Particles p. 3, on ἀλλὰ καί, end
of paragraph, suggests that that correlation here would lead to ‘not merely necessary for you to (not) think
these things, but that also you should …’, i.e. that you should do these things, too. (See also Denniston p. 21
on progressive ἀλλά, sometimes with καί.)
6 βαϲίληαν ῎Hραν: Like Zeus in 2, rarely named in Sappho: the only other instance is 17.2 πότνι’
Ἦρα; at 17.9 (in prayer to the Trinity of gods worshipped at Mesa) as the main addressee, she is named
simply as ϲέ, but elaborately styled as Αἰολήιαν [κ]υδαλίμαν θέον / πάντων γενέθλαν in the full cult hymn
at Alc. 129.6–7.
7 ἐξίκεϲθαι: Presumably we need a non-historical translation of the infinitive, representing an original
subjunctive or optative in the putative prayer.
7–8 ἐξίκεϲθαι τυίδε ϲάαν ἄγοντα / νᾶα Χάραξον shows interweaved syntax (ABAB) of the two nouns
and their modifiers. The content of the prescribed prayer strongly resembles Sa. 5.1–2 ἀβλάβη[ν μοι] / τόν
καϲίγνητον δ[ό]τε τυίδ’ ἴκεϲθα[ι ‘Grant that my brother arrive back here unharmed’. The verb of ‘granting’
which is specified in the direct version, disappears in the indirect one. τυίδε, shared between both versions,
is deictic and specific, i.e. ‘to Lesbos’ (perhaps even Mesa, the place of Hera’s worship and of putative per-
formance of the ode). Just as in 2–3, where Charaxos is hoped to come νᾶϊ ϲὺν πλήαι ‘with a full ship’, so
here he is correspondingly wished to ‘arrive driving his ship safe’.
7 ϲάαν: Cf. Alc. 401 B a.1 (on the throwing away of his shield in flight) Ἄλκαοϲ ϲάοϲ. Note that the
framed prayer distinguishes physical safety and well-being from becoming very rich and blessed (lines
14–15).
10 δαιμόνεϲϲιν: here presumably unspecified gods collectively, rather than protective spirits or ‘souls’
of the dead. Cf. 14n.
11 εὔδιαι γὰρ ἐκ: to be preferred over εὐδία παρὲκ in asyndeton. εὔδια(ι) (J. Lidov) is a powerful if
common image or metaphor in Pindar (O. 1.98, P. 5.10, I. 7.38, frr. 52b.52, 109.1; cf. Solon 13.17–24; Hor. C.
1.9.9–12, probably echoing the present poem, as suggested by G. Hutchinson). The train of thought seems
to be: ‘In all other matters, let us turn things over to the gods. For fair weather can come suddenly out of/
apart from a storm (either through natural causes or at the behest of gods). Furthermore, Zeus lets those he
wishes have a special divine helper, to relieve them of their troubles.’
ἐκ μεγάλαν ἀήταν: Cf. Sa. 20.9 μεγάλαιϲ ἀήται[ϲ and Alc. 249.5 ἀνέμ]ω (Page) κατέχην ἀήταϲ – both
passages in the context of sea-faring. Normally ἀήται are strong winds, gales, blasts, thus here = χείμων.
Cf. Sa. 2.10–11 αἰ ⟨δ’⟩ ἄηται / μέλλιχα πνέοιϲιν, 37.2 ἄνεμοι; Alc. 249.11 ὤϲ κ’ ἄν]εμοϲ φέρ[η (Lobel). As
with εὔδιαι earlier, this is a favorite image of Pindar and the lyric poets (Pi. I. 4.9, Bacch. 17.91, Simonides
PMG 595.1, Alcm. 1.5 Dav., Anacr. PMG fr. 379b (with Bergk’s correction), Timoth. Persai 107 Hordern).
44 D. Obbink

13 τῶν κε: I take των to be the relative pronoun (equivalent to the indefinite ὀττίνων, a form which
never occurs), as a genitive of possession depending on δαίμον’ … ἐπάρωγον in 14, and picked up by κῆνοι
in 15. Present general conditions in Sappho all have the form ὄττιϲ or τιϲ, with the use of κε optional (Sa.
5.3 has it, Sa. 16a.30–1 (P. GC. inv. 105 fr. 2 ii 6–7, formerly = Sa. 26.2–3) without; at Alc. 70.8–9, quoted
below, κ’ is uncertainly κ(ε) or κ(αί)). If των equals ὀττίνων, then the κε in the present instance could be
emended to τε, but doesn’t have to be; but if not, the sentence is left in asyndeton. (I see no way to accom-
modate the article τών; according to Hamm, Grammatik 108–9 §192, the relative can use forms of the
demonstrative, but not of the article.) In Sappho sentences beginning without a conjunction are uncommon,
being restricted to questions or commands (including optative of wishes and jussive subjunctive) which
open hymns and prayers. Here the effect is to continue one gnomic statement (in 11–12) on with another
(in 13–16), with a relative pronoun that may have connective force with what precedes, even though it con-
strues as a possessive genitive with δαίμον(α) in what follows.
13–15 βόλληται … περτρόπην: Strikingly similar language in Alc. 70.8–9 κ’ ἄμμε βόλλητ’ Ἄρευϲ
ἐπιτ ύ̣ χε ̣ [̣ ---]/τρόπην.
13 βαϲίλευϲ Ὀλύμπω: In contrast to 2, where Zeus is named, here Zeus’ identity is paraphrased in
terms of cosmic genealogy. Cf. Sa. 27.12 οὐδάμωϲ] (Snell) ὄδοϲ μ[έ]γαν εἰϲ Ὄλ[υμπον with a similar
gnomic flavor. Alcaeus by contrast has Ὀλύμπιοι: Alc. 70.11 τάν τιϲ Ὀλυμπίων, 349b.1 θέων μηδ’ ἔν’
Ὀλυμπίων, perhaps 130b.22 ᾿Oλύμπιοι.
14 δαίμον’: Cf. 10, though perhaps here with connotations of ‘protective spirit’ (see W. Burkert, Greek
Religion, Oxford 1985, 179–81 at 180: ‘Daimon is the veiled countenance of divine activity’). Missing is
the third of the triad of gods at Mesa on Lesbos (with 2 Zeus, 6 Hera), at any rate not mentioned by name
here. Only slightly more defined (but still unspecified) are Alc. 70.11 τάν τιϲ Ὀλυμπίων and 349b.1 θέων
μηδ’ ἔν’ Ὀλυμπίων.
ἐκ πόνων: i.e. such as Charaxos might experience at sea.
ἐπάρ{η}`ω΄γον: Cf. Od. 11.498 etc. – the supralinear correction in m2 apparently correct, rather than the
noun ἐπαρήγoν’ i.e. ἐπαρήγον(α) which the original scribe wrote. The η is not cancelled, which may indi-
cate that this was a variant rather than a correction (cf. 5). It would be proleptic: ‘as a helper’. So also Alc.
34 is a prayer for helper-gods who save at sea; Sa. 95 and Alc. 308 (unless it was a version of the story in
the Homeric Hymn) contain prayers to Hermes, well-known as a helper god and escort. See D. Wachsmuth,
ΠΟΜΠΙΜΟΣ Ο ∆ΑΙΜΩΝ (Berlin 1967). For the combination δαίμων ἐπαρωγόϲ see Eur. Hec. 163–4 ποῖ
δὴ ϲωθῶ; ποῦ τιϲ θεῶν ἢ δαίμων ἐπαρωγόϲ (it is not clear whether this means, disjunctively, ‘a god or a
daimon that helps’ or rather ‘a divinity directly or indirectly to be an ἐπαρωγόϲ’; ἐπαρωγόϲ may apply to
both, as it is often translated, or be restricted to δαίμων in apposition, or as an epithet). It is possible that
Euripides here echoes the present passage, but could rather be simply employing standard phraseology, i.e
the typical language of prayer for safe return. M. L. West suggests emending to ἐπ’ ἄρηον, ‘to turn their
fortune to the better’, but this requires understanding δαίμον’ in a radically different sense from 10.
15 κῆνοι: sc. those mortals assigned a special guardian divinity by Zeus (viz. 13 τῶν).
15–16 μάκαρεϲ πέλονται / καὶ πολύολβοι: The gnomic sequence reaches its climax in a makarismos;
in connection with Charaxos cf. Posidippus Epigr. XVII Gow–Page (= 122 Austin–Bastianini) (Doricha)
line 8 οὔνομα ϲὸν μακαριϲτόν. However, as a gnomic statement in the present context it ought to be poten-
tially applicable to all. Cf. Sa. 133.2 πολύολβον Ἀφροδίταν. The idea is expressed negatively in Sa. 63.5–6
ἔλπιϲ δέ μ’ ἔχει μὴ πεδέχη[ν] / μηδὲν μακάρων and Sa. 27.12–13 (quoted above on 13); cf. (probably also
of divinities) Sa. 3.16 μακα[, 15.1 μάκαι[ρα. In the present case a collapse of the distinctions is anticipated.
17 κἄμμεϲ, αἴ κε: Parenthetical interposition of a conditional clause after introductory κἄμμεϲ is odd
or emphatic. Presumably the personal pronoun κἄμμεϲ (cf. 9) is fronted, because it brings the poem back to
the hic et nunc of the climactic final wish – after consideration of various stances and attitudes toward the
absence of a close associate or family member, and the attendant conditions of their safe return – for what
might optimally happen.
 Two New Poems by Sappho 45

ἀέρρη: Presumably subjunctive (parallel to 18 γένηται connected by καί), with κεφάλαν as the Homeric
‘life’, so that the notion is ‘preserve’ or ‘save’. If 18 ἄνηρ can refer to social status (a member of the peerage,
so to speak, and ἑταῖροϲ), then the notion could be ‘to become a gentleman’, supported by family wealth.
For the absence of iota in the subjunctive see Hamm, Grammatik §249a2. (Alternatively ἀέργη, from the
previously unattested Homeric and Aeolic form ἀέργω (< ἀεργόϲ), see LSJ s.v. ἀργέω.) For ἀέρρη, cf. Sa.
111.3 ἀέρρετε, Il. 10.80 κεφαλὴν ἐπαείραϲ, Eur. Tro. 98–9 κεφαλὴν / ἐπάειρε (F. Ferrari), Soph. OT 22–4
πόλιϲ … ϲαλεύει κἀνακουφίϲαι κάρα / βυθῶν ἔτ’ οὐχ οἵα τε φοινίου ϲάλου (M. L. West).
τὰν κεφάλαν: ‘the head stands for the man himself’ (G. S. Kirk ad Il. 2.258 citing 18.114, 23.94,
24.276, Od. 1.343; cf. LSJ s.v. I.3 with other Homeric passages).
18 Λάριχοϲ: Sappho’s and Charaxos’ younger brother according to ancient testimony: Athen. 10.425a
(Sa. test. 203 Voigt) Ϲαπφώ τε ἡ καλὴ πολλαχοῦ Λάριχον τὸν ἀδελφὸν ἐπαινεῖ ὡϲ οἰνοχοοῦντα ἐν τῶι
πρυτανείωι τοῖϲ Μυτιληναίοιϲ, ‘The lovely Sappho often praises her brother Larichos because he poured
the wine for the Mytilenaeans in the town-hall’. One mention of Larichos may have come in Sa. 141, a
description of a divine wedding with Hermes as wine-pourer (quoted by Athenaeus just below at 10.425d):
‘There a bowl of ambrosia had been mixed, and Hermes took the jug and poured the wine for the gods ….
They [subject uncertain] all held drinking-cups, and they offered libations and prayed for all manner of
blessings on the bridegroom.’ Cf. Sa. 2.16 οἰνοχόαιϲον; Eust. ad Il. 1205.18ff.; schol. T Il. 20.234 ἔθοϲ γὰρ
ἦν, ὡϲ καὶ Ϲαπφώ φηϲι, νέουϲ εὐγενεῖϲ εὐπρεπεῖϲ οἰνοχοεῖν, ‘For it was the custom, as Sappho says, for
handsome young noblemen to pour the wine’. Young Larichos would have been deprived of his patrimony
if Charaxos, while trading Lesbian wine at Naukratis, had squandered the family’s fortune on Rhodopis/
Doricha, or for other reasons failed to return to Lesbos ‘with a full ship’ (2). Although there is no other tes-
timony to this effect with reference to Larichos, it is at least possible that the younger brother’s diminished
prospects are alluded to here (or that this was inferred by ancient readers and scholars). Presumably his
identity was familiarly known to an original audience: Athenaeus loc. cit. above says that Sappho praised
her brother Larichos ‘often’ (πολλαχοῦ).
19–20 μάλ’ ἐκ πόλλ⟦η⟧`αν΄ βαρυθυμίαν κεν / αἶψα λύθειμεν: Similarly Sa. 1.25–6 χαλέπαν
δὲ λῦϲον / ἐκ μερίμναν, where she speaks (as here) about herself and her (or her family’s) sufferings.
λύθειμεν recalls some of the language of Sa. 5.5 ὄϲϲα δὲ πρόϲθ’ ἄμβροτε πάντα λῦϲα[ι (= with a slight
change Sa. 15.5 [ὄϲϲα δὲ πρ]όϲθ’ [ἄμ]βροτε κῆ[να λῦϲαι]?), where it is Charaxos who is concerned, just
as 7 ἐξίκεϲθαι τυίδε above parallels Sa. 5.2 τυίδ’ ἴκεϲθα[ι. Luca Benelli observes the important parallel
between βαρυθυμίαν and the ‘Old Age Poem’ (P. Köln XI 429.2) line 5 βάρυϲ δέ μ’ ὀ [θ]ῦμοϲ (although
the sense is not the same as here, as he notes).

Kypris Poem
(P. Sapph. Obbink 21–9; P. Oxy. 1231 fr. 16)
Line 1 (21) and the first word of 2 (22) are relatively readable. After that we are largely left to guesswork
on uncertain traces or articulation of letters for much of the text apart from a certain word here and there.
Reconstruction proceeds on the hypothesis of (i) an address (or quoted address) to Kypris in 2 (22); (ii) nar-
ration of a past encounter with a lover due to Aphrodite, involving (iii) articulation of the experience of
desire. Details, together with elaboration in a simile or metaphor with images of helplessness and destruc-
tive desire, depend on the soundness of proposed restorations. In addition, there is the possible, if not cer-
tain, overlap at 1 (21)–8 (28) with P. Oxy. X 1231 fr. 16.1–8 (= Sa. 26), in a narrow strip down the centre
of the present column, where there seem to be too many coincidences over too many lines in the vertical
strip to avoid concluding that they are the same poem, and without obvious counter-indications or contra-
dictions (the failed attempt to match lines 2–4 with Etym. Magn. 449.34 not withstanding (see P. GC. inv.
105 fr. 2 ii 6–8 in Burris, Fish, and Obbink 2014, note 6 above, 9, 18): as previously reconstructed, the line
is too long, the alignment wrong, and the quotation actually lacks the essential letters πα[ after μάλιϲτα
46 D. Obbink

present in the papyri), and so which now supplies and

confirms letters here and there (see notes). As such, the
Kypris Poem now supplies more (and corrects previous
reconstructions) of Sa. fr. 26 Voigt. However, given the
uncertainty of traces in many cases, and the tentative
nature of many of the suggestions below, other articu-
lations and alternative schemes of restoration may be
contemplated.
1 (21) πῶϲ κε δή τιϲ ου θαμε ̣‒̣ : ]θαμε [̣ P. Oxy.
1231 fr. 16.1 (θαμέω[ν Hunt: θαμέω[ϲ Diehl1, see
Benelli, Sapphostudien, forthcoming). Reading ωϲ here
after θαμε would require the ω to be rather wide (but
not wider than e.g. the first ω in this line), and having
against it only that the trace interpretable as its right side
extends rather far down vertically (rather than angle-
shaped as elsewhere), and could be taken as ι; reading
ϲ requires assuming that the apparent thin horizontal
stroke (running into a somewhat thicker curving stroke
at mid-level) is the only surviving remnant of the thick-
ness of shading of the vertical back of ϲ – but on the
whole, not an impossible fit. This would then resolve as
πῶϲ κε δή τιϲ οὐ θαμέωϲ. For θαμέωϲ see Sa. ‘Old Age
Fig. 3. Photo courtesy Egypt Exploration Society
Poem’ (P. Köln XI 429.2) line 7 ϲτεναχίϲδω θαμέωϲ,
P. Oxy. 1231 fr. 16.1 (Sa. 26.1) cited above (if different from the present text), and cf. Alc. 358.5 θάμα, 72.6
θάμ’, 306 (13) i col. II.16 πόλλ]α τε καὶ θαμα[. (Although not exampled in Sappho, elided θάμ(α) could be
tried here: e.g. θάμ’ ἐπιλάϲαιτο, but is an even greater difficult fit for πιλ, and the compound is unattested.)
A succession of short single-syllable words is not uncommon in Sappho: e.g. in book 1, Sa. 27.4 [ ̣ ̣ ]̣ κ̣ αὶ
γὰρ δὴ ϲὺ πάιϲ ποτ[ – although the extent and word-order in the present case is baffling, in what is also
the beginning of a poem. Other articulations may be sought, but attempts to extract e.g. πω (ἔ)ϲκε or ποι
(ἔ)ϲκε lead nowhere. For δή see 2n. δή τιϲ, at least, is the usual order for questions (Denniston, Greek
Particles p. 212 (6).(i).(b) (where the speaker has someone in mind, but keeps it to himself). A search of the
TLG (1 December 2013), however, shows an additional oddity with regard to the position of the negative.
κε (or ἂν) δή is common from Homer on, but οὐ, when it appears, always precedes; for an example with τιϲ
see Il. 10.204–5 ὦ φίλοι οὐκ ἂν δή τιϲ ἀνὴρ πεπίθοιθ᾽ ἑῷ αὐτοῦ / θυμῷ; cf. Od. 18.414, 22.132, i.e. this
seems to be normal word-order in questions, especially if neither οὐ nor κε is adjacent to the verb. Thus an
apparently deliberately abnormal word order. Metrically, we could as well have had πωϲ τιϲ οὔ κε δή. The
effect (of postponing οὐ) would be presumably to throw emphasis on οὐ, and perhaps also θαμέωϲ: ‘How
could one not do it frequently?’ Assuming θαμέωϲ, the experiencing of anguish of desire (in an address
to Aphrodite) may, of course, simply mean over and over again on any single occasion; but experiencing it
‘frequently’ or ‘repeatedly’ on more than one occasion (i.e. previous to the present one) is what Sappho may
be seen doing in Sa. 1 and elsewhere in her poems. The present instance may be seen as a working out of
that trope, by reflecting descriptively (as in Sa. 1) on an important previous occasion in which she did so,
as an exemplum and argument for the way in which Aphrodite can surely be expected to help yet again,
while showing the speaker helplessly falling in love yet again, as e.g. in Sa. 1, Ibycus PMG fr. 287.1 αὖτε
and Anacr. PMG fr. 358.1 δηὖτε.
αϲαιτο: Very likely ἄϲαιτο as proposed by Benelli, noting good Lesbian parallels for the verb (from
ἀϲάω, with short alpha, not ἀάω): Sa. 3.7 ἄϲαιο, and the root: Sa. 1.3 ἄϲαιϲι (there associated with, if not
inseparable from, ὀνίαιϲι), referring to ‘heartache and anguish’ (Page); cf. 68a.4 ἄϲαν? (Diehl1); Alc. 39a.11
ἄϲαιϲ. D. A. Campbell (Greek Lyric Poetry, 2nd ed. Bristol 1982, ad Sa. 1.3) notes that they refer to physical
 Two New Poems by Sappho 47

distress in the medical writers (see LSJ s.v. ἀϲάω). Apparently not restricted to females: see also Sa. 103.8
ἄϲαροι; 91 ἄϲαροτέραϲ, cf. ἄϲα (Attic ἄϲη); Cat. 73.5 ut mihi, quem nemo grauius nec acerbius urget /
quam modo qui (quae Birt, Otto Skutsch).
2 (22) κυπριδε: Possible articulations include Κύπρι δε- (Sa. 2.13, 5.18?, 15.9), Κύπρ’ ἰδ- (for the elision
cf. Sa. 1.20 Ψάπφ’), and Κύπριδ’ ἐ- (cf. inc. auct. Lesb. 42.7 -π]λόκω Κύ[π]ριδ[οϲ?, but Sa. test. 200 Voigt
Κύπριν). An address to Kypris (= Aphrodite), a favorite in the poems of book 1 of Sappho, looks likely, with
weak pauses before and after. Thus line 1 begins with τιϲ in an independent clause, then becomes implied
direct discourse addressed to Aphrodite, in advance of the speaker’s – first-person revelation of herself in
line 5 (35) with μ’. In this case the τιϲ-figure in 1 would be a stand-in for ‘Sappho’ as speaker of the poem.
̣ ̣ ̣ ̣ ̣ ’ :̣ After κυπριδε almost everything is variable, and various readings and articulations may be
considered, e.g.: (i) Κύπρι, δέρκεϲθ’ ̣ (Burris: ‘How indeed would anyone not be repeatedly disturbed, Cyp-
ris, to see …’). This allows the meter to continue as in 1231 (see next note). However, it would also require
the θ to have been especially wide (perhaps with its mid-stroke extended like the one in 4 (24) εχηϲθα, or
to have had a letter or two, perhaps α or αι, written after the θ, with the apostrophe above intended as a
replacement in correction); ρ would be missing the top of its upright where the cap should attach to it, and
κ would be almost entirely swallowed up in the overlap of layers in the ancient repair join. It also requires
ἄϲαιτο in 1 to be completed by an epexegetic infinitive (where a participle might have been expected).
(ii) Κύπρι, δέϲποιν’, which suits spacing and traces best: cf. Pi. fr. 122.18 ὦ Κύπρου δέϲποινα, fr. 80 with
A. Henrichs, HSCP 80 (1976) 253–86. For the order cf. E. Med. 632.
ο ̣ ̣ ̣ [̣ : ]τ’τινα[ P. Oxy. 1231 fr. 16.2 (ὄ]ττινα[ϲ γὰρ Hunt). In the present case, the traces could be
construed as ὄττιν[, which would accord with P. Oxy. 1231. ττ will have had to have been written rather
narrowly (cf. τ in 10 -τρόπωμεν), with the horizontal strokes virtually connected (the inward curvature in
the second upright suits particularly neither τ nor π, but this seems to be due to disturbance of the surface).
] φ̣ ̣ [̣ : Here ηφ looks almost inevitable, since the scribe does not elsewhere write ει in ligature (though
he does occasionally extend the mid-stroke of ε beyond its bowl so far as to virtually touch the following
letter, e.g. in 19 εκ). Assuming η, Burris suggests ὄττιν[α δ]ὴ φιλ[είη, ‘whomever one clearly loved/should
love’. The α and δ will have had to have been rather wide, though not wider than elsewhere; however, this
will also require the use of δή (a common particle in the Lesbian poets: eleven times in Sappho including
Sa. ‘Old Age Poem’ (P. Köln. XI 429.2) line 6, four in book 1 alone) twice in two successive lines, perhaps
emphatically: for correlated δή in proximity in poetry see Denniston, Greek Particles pp. 224–5 (12).(i).
3 (23) ̣ ̣ ]̣ θέλοι: Potentially another optative, with subject 1 τιϲ, continuing the sequence above. Thus
perhaps [καὶ] θέλοι, ‘and wished/should wish’ (to do something?), unless the verse began with a relative
clause: (instrumental/causal) ‘with/by which (τῶι) one wished/might wish’, or (object) ‘whom (τὰν) one
wished/might wish’ (to do something). Cf. Sa. 5.3 κε θέλη γένεϲθαι; with μάλιϲτα: Sa. 1.17 κὤττι μοι
μάλιϲτα θέλω γένεϲθαι.
μάλιϲτα ⏑ ̣ ̣ ̣‒̣ κ̣ α [̣ ‒ ×: ]λιϲταπα[ P. Oxy. 1231 fr. 16.3. Although the trace after μάλιϲτα could be con-
strued as τ, enough survives compatible with π (cf. the π in 19 ἐκ πόλλαν) here to confirm the connection
with P. Oxy. 1231. If correct, it would be possible to read -τ’ ἀπα- or -τα πάλιν κάλ[εϲϲαι or κάλ[ειϲθαι,
‘to call back’, ‘to call again’ (κάλ[ην × is ruled out by the avoidance by the Lesbian poets of a bi-syllabic
word in the 9th and 10th positions in Sapphics). For πάλιν see Alc. 393; κάλεϲϲαι: Alc. 368.1. Various other
articulations include: (i) θέλοιμ’ ἄλιϲτα or -τ’ ἀπα- (ἄλ(λ)ιϲτα, neuter plural, occurs in Hom. Hymn to
Hermes 168, when Hermes says to Maia that they should not sit apart, without gifts or prayers; frequently
‘inexorable’ as applied to Hades; West prints it as an emendation at Aesch. Agam. 413, where it means ‘not
reachable/achievable by prayer’); (ii) μάλιϲτ’ ἀπα ̣‒̣ κ̣ α [̣ ‒ ×.
4 (24) ̣‒̣ ]̣ ονεχηϲθα (P. Oxy. 1231 fr. 16.4 has nothing here but blank papyrus, because the adonaean,
at this length, would not have extended far enough): Probably ἔχηϲθα, the present 2nd person singular of ἔχω,
with space for the supplement preceding being uncertain (3–4 letters?) depending on how much progressive
extension of the line-beginning to the left one allows under ‘Maas’ Law’. Possible supplements include e.g.
κόϲμ]ον ἔχηϲθα, ‘you have adornment’ (cf. Sa. 98a.3) or θῦμ]ον ἔχηϲθα (cf. Sa. 86.4 ἔχοιϲα θῦμο[ν, ‘having a
48 D. Obbink

[kindly?] spirit’) or αἶν]ον ἔχηϲθα, ‘you have a reputation for’ or ‘are famed for’; πίκρ]ον and ἄκρ]ον (cf. Sa.
148.2 – but the text is uncertain). A propos of the erotic context would be κέντρ]ον, or κἄρ]ον (i.e. καὶ ἔρον),
‘when you have (a) desire’ (i.e. for another). A different construal for the verb, unattested for the Lesbian poets
(although they do exhibit compounds of ἔχω in περ(ρ)- and πεδ-) is ὄνεχω (i.e. ἀνέχω), e.g. τὴν] ὀνέχηϲθα,
‘you reveal her’. ἔχηϲθα could also have the sense of ‘be able’, ‘have the means to’ (so at Sa. 98b.2). For the
second person singular in Sappho see Sa. 129a ἔμεθεν δ’ ἔχηιϲθα λάθαν, ‘you have forgotten me’, 96.23
-εχη⟨ι⟩ϲθ’ (object uncertain) and (in its other form) Sa. 112.2 ἔχηιϲ δὲ πάρθενον ἂν ἄραο. (Note that ancient
scribes seem to have been uncertain about the inclusion of the iota.) At the same position in the adonaean is
Sa. 27.11 ἔχοιεν (in a wish or prayer, the subject of which is apparently θέοι in the ‘preceding’ line). There
is no punctuation after ἔχηϲθα in the papyrus, possibly (though not certainly) an indication that the syntax
continues into the following verse. If so, a connective will be needed in this gap, or a word that can bear
asyndeton. E. L. Bowie suggests ποῖ]ον, L. A. Holford-Strevens πῶϲ] ὀνέχηϲθα, ‘How do you keep on …’.
5 (25) ̣‒̣ ]̣ α ̣ λοιϲ :̣ At end, only -ι fits, possibly ϲαμ]βάλοιϲι? References to footwear (Anacr. PMG fr.
358.1 ποικιλοϲαμβάλωι), walking, and one’s step or gait (Sa. 16.17 βᾶμα) appear as sensual or erotic images
elsewhere in lyric, where they may also imply motion in the context of dance. Alternatively, ϲάλοιϲι (Ham-
merstaedt), ‘tremblings’, ‘shakings’, i.e. as a physical symptom, cf. 1 (21) above ἀϲαιτο, E. Or. 994 πόντιοι
ϲάλοι, with e.g. νῶν] (E. L. Bowie, cf. Sa. 96.2) or a particle like νῦν] or a relative pronoun at the beginning.
ἀλέματ ̣ :̣ ]ᾱλεμάτ[ P. Oxy. 1231 fr. 16.5. Assuming a match between the two texts, the accent in P. Oxy.
1231 shows that the reading here was either ἀλεμάτων or (more likely) ἀλεμάτωϲ, ‘helplessly’, ‘in vain’
(i.e. not ἀλέματον construed with the preceding μ’; there is insufficient space for ἀλεμάτοιϲι). Cf. Sa. 130.2
where Ἔροϲ is said to be ἀμάχανον ὄρπετον. For ἀλέματοϲ see LSJ s.v. ἠλέματοϲ; Theoc. 15.4 ὦ ἀλεμάτω
(Scaliger : ἀδεμάτω MSS : ]τω⸌ϲ⸍ Antinoë Theoc.); Cat. 50.1 otiosi and 51.13–15 otium; Alc. 70.4.
δαΐϲδ[ (Hammerstaedt): the verb not previously exampled in the Lesbian poets. Possible completions
include 2nd person singular δαΐϲδ[ειϲ (or perhaps imperative); (ii) δαΐϲδ[ον agreeing with a neuter singular
noun ending in 4 (24) in ]ον; (iii) active infinitive δαΐϲδ[ην (perhaps completing 4 (24) ἔχηϲθα in the sense
‘you are able’ or with νῶν] in 5 (25)). But it is difficult to divine exactly what should be ‘divided’ or ‘cleaved’,
whether the object is μ(ε), i.e. of divided, potentially contradictory emotions in the speaker (as e.g. in Sa. 31),
of divided loyalty, or some more elaborate metaphor for physical division or cutting. Even less clear is how
this could be done ἀλεμάτωϲ, ‘in vain’, unless as a description of something Aphrodite is not responsible for,
since a god’s actions could hardly be said to be ‘ineffectual’, but might be done ‘irresponsibly’ or ʻcasuallyʼ.
6 (26) [ ‒̣ ⏑ ̣ ̣ ]ρω λυ{ι}ϲαν ̣ :̣ At Sa. 130.1 Ἔροϲ is said to be λυϲιμέληϲ, thus here perhaps ἰμέ]ρω⟨ι⟩
λύ{ι}ϲαντι, ‘because of desire that loosened’ or ‘destroyed (me? see next note)’. Cf. Cat. 50.16 perdidit. For
ἴμεροϲ see Sa. 95.11, 96.16, 137.3 and cf. 1.26–7. The omission of the iota after ]ρω could be explained by
its writing, by transposition, after λυ. However, a genitive in ]ρω might be expected, since the scribe does
not otherwise omit iota adscript wherever we can tell, either in this papyrus or in P. GC. inv. 105 (see intro.).
ο̣ ̣‒̣ ̣ ε̣ ̣‒̣ [ ×: ]·γονωμ[ P. Oxy. 1231 fr. 16.6 (the raised point to counterindicate αντιγονω?). Thus we
could have [Ἰμέ]ρω λύ{ι}ϲαντι γόνω, ‘because of the destructive (e.g. power?) of your offspring, Desire’,
or rather now that we know that λύϲαντι precedes, it is attractive to see γόν(α), ʻkneesʼ here (suggested by
Professor Tsantsanoglou) as its object, thus forming part of the sequence of physical symptoms delineated
in this and the preceding line (in contrast to the ʻthoughtsʼ of Kypris).
7 (27) ̣‒̣ ]̣ α ̣ α̣ ̣ [̣ ̣ ]̣ ̣ :̣ e.g. παῖ]δα πᾶϲ?
] μ̣ ̣ ̣ ̣ ̣ ̣
ο[: ]αιμ’ουπρ[ P. Oxy. 1231 fr. 16.7, in agreement with the traces here. Then φ]αῖμ’ (Tsantsa-
noglou) or ]αι or ] αἰ μ’ οὐ προ[.
] ε̣ ρηϲ[: The traces visible above the line here are not parts of a suprascript iota after ] ε̣ ρ, but rather
the remains of offsets left when the scrap bearing ] ε̣ ρηϲ[, originally folded over left, was unfolded and
turned back into place during conservation. The piece, now detached, may be ranged up to two or three
letters from προ, or may have continued on directly from προ[ without any break.
28 ̣‒̣ ⏑]̣ νεερ [̣ ̣ ]̣ ×̣ (blank): ]αι (blank) P. Oxy. 1231 fr. 16.9. The adonaean appears not to have contin-
ued in this line after the last preserved trace, where the surface is blank and undisturbed. For the adonaean,
 Two New Poems by Sappho 49

‒̣ ⏑
̣ ]̣ νεερ [̣ ̣ ]̣ ×̣ looks more likely than ̣‒̣ ]̣ νθερ [̣ ̣ ]̣ ×,̣ because of the sinuous flourish at the end of the
horizontal stroke, characteristic of the extended mid-stroke of ε elsewhere, whereas the mid-stroke of θ
elsewhere is straight. One possibility is ἐέρϲ[α]ι, ‘dew’ (Burris), cf. Sa. 73a.9, 96.12 (perhaps slightly short
for the space after the ρ). Or possibly a distracted Homeric epic form such as ϲυνεεργ- / ἀνεεργ- / ϲυνεερμ-
(Benelli): ‒̣ ϲυ]νεέρχ[θα]ι, for example (cf. Sa. 1.28 ϲύμμαχοϲ ἔϲϲο), would suit the traces and spacing.
29 ] ̣ ̣ [̣ : This portion of the line does not survive to overlap in P. Oxy. 1231 (fr. 16.9), although the
latter, several letter-spaces later, has ]ϲέ·θελω[ (presumably ] ϲέ, θέλω[). Traces of six further lines follow
in P. Oxy. 1231 (= Sa. 26.10–15) as given below, but without the left margin preserved there or in the pres-
ent papyrus here, leaving no indication whether they indeed belonged to this poem, or where a new poem
began. In 30 (10) τοῦ]το πάθη[ν was suggested by Hunt ap. P. Oxy. 1231 fr. 16.10. 31 (11)–32 (12) are sup-
plied from Apoll. Dysc. Pron. 1.51.1ff. Schn. (see Voigt on Sa. 26.11–12).
The following composite implements some of the suggestions given above in a continuous text. The over-
lapping relationship with P. Oxy. 1231 fr. 16 is indicated as follows: (i) letters outside of any brackets are
exclusively in P. Sapph. Obbink; (ii) letters inside ⌈ ⌉ are read in P. Sapph. Obbink in certain or compatible
agreement (as indicated) with those of P. Oxy. 1231; (iii) letters in ⌊ ⌋ are supplied from P. Oxy. 1231.
21 (1) πω κ̣ εδητιϲουθαμε ̣ α ̣ ϲαι ο̣ πῶϲ κε δή τιϲ οὐ ⌈θαμέω⌉ϲ ἄϲαιτο,
κυπριδε ̣ ̣ ̣ ̣ ’̣ ̣ ο ̣ ̣ ̣ [̣ ̣ ]̣ φ
̣ ̣ [̣ Κύπρι, δέϲποιν’ ,̣ ὄ⌈ττιν⌉⌊α⌋ [δ]ὴ φιλ[είη
̣ ̣ ]̣ θελοιμαλιϲτα ̣ ̣ ̣ ̣ κ̣ α [̣ καὶ] θέλοι μά⌈λιϲτα πά⌉λιν κάλ[εϲϲαι;
__ ̣ ̣ ]̣ ονεχηϲθα ποῖ]ον ἔχηϲθα

25 (5) ̣ ̣ ]̣ α ̣ λοιϲ μ ̣ ’αλεμ ̣ ̣ ̣ ̣ α ̣ ¨ ̣ ̣ [̣ νῶν] ϲάλοιϲι μ’ ⌈ἀλεμάτ⌉ωϲ δαΐϲδ[ην

̣ ̣ ]̣ ρωλυ ϲ̣ αν ̣ ̣ ο̣ ̣ ̣ ̣ ε̣ ̣ [̣ ἰμέ]ρω⟨ι⟩ λύ{ι}ϲαντι ⌈γόν’ ωμ⌉ε ‒̣ [ ×
̣ ̣ ]̣ α ̣ α ̣ ̣ [̣ ̣ ]̣ ̣ μ ̣ ̣ ̣ ̣ ο̣ [0–3] ε̣ ρηϲ[ ̣‒̣ ]̣ α ̣ α ̣ ̣ [̣ ̣ ]̣ ⌈αιμ’ οὐ πρ⌉ο[0–3] ε̣ ρηϲ[
__ ̣ ̣ ]ν ερ [ ] ̣
̣ ̣ ̣ ̣ ̣ ̣‒̣ ⏑]̣ νεερ [̣ ]̣ ⌊α⌋⌈ι⌉

29 (9) c.8 ] ̣ ̣ ̣[ c.8 ] ̣ ̣ [̣ ̣ ]̣ ⌊ϲέ, θέλω⌋[⏑ ‒ ×

. . . . . . . . . . ‒ ⏑ ‒ × τοῦ]⌊το πάθη⌋[ν ⏑ ‒ ×
‒ ⏑ ‒ × ‒]⌊ ̣ αν, ἔγω δ’ ἐμ’ αὔται⌋
⌊τοῦτο ϲύνοιδα⌋

33 (13) ]⌊ ⌋̣ [ ]̣ ⌊ τ̣ οιϲ⌋[ ̣ ̣ ̣ ]̣ ⌊ ⌋̣
]⌊εναμ⌋[
]⌊ ⌋̣ [ ]̣ ⌊ ⌋̣ [
. . . . . . . . . .

How wouldn’t anyone feel anguish repeatedly, 1

Kypris, Queen, and especially wish to call 2
back, whomever one really loves? 3
What sort of thoughts do you have 4

25 to pierce me idly with shiverings 5

26 because of desire which weakened the knees? 6
27 … not … 7
29 … you, (I) wish … 9
30 … to suffer this … 10
31–2 … but I am myself conscious of this … 11–12

Dirk Obbink, Christ Church Oxford

dirk.obbink@classics.ox.ac.uk