Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

Pindar's "hymn to Cybele" (fr. 80 SM): Meter, form, and syncretism
Lidov, Joel B
Greek, Roman and Byzantine Studies; Summer 1996; 37, 2; ProQuest
pg. 129

Pindar's "Hymn to Cybele" 
(fr. 80 SM): 

Meter, Form, and Syncretism 

Joel B. Lidov 

P 
INDAR REFE RS in several places to the Magna Mater, the 
mother of the gods, but only in the three words of 
fr o 80 SM does he associate her with a proper name: 

O£<J7t]Otv[ClV] K'UPE[A,ClV] IlCl't[€pa]. This has given-th~ fragment 
a special interest as part of the evidence for Pindar's religiosity 
or the history of cult, in particular for syncretism in Thebes; 1 

the recent assumption that the fragment begins a hymn to 
Cybele has lent weight to its apparent significance. A careful 
scrutiny of what is possible for these words will draw attention 
to some unappreciated features of both Pindaric metrics and 
the formal character of lyric invocations. It is unlikely that these 
words begin a hymn, or are even evidence for one, least of all a 
Theban one; taken together, this and other passages from 

I -[D]er Dithyrambos auf die groge Mutter [=fr. 80, see next note] ein be­
deutsames Zeichen fur die beginnende Zersetzung der griechischen Religion 
durch die Einflusse des Auslands": R. Wunsch, -Hymnos," RE 9.1 (1916) 161. 
The biographical aspect implicates this fragment in the controversies sur­
rounding Pyth. 3.77ff, with r and the ancient biographical tradition, and 
brings in as well the Hymn to Pan (fr. 95 SM), Dith. 2 (fr. 70b SM), and other 
passages; all discussions of this fragment have begun with at least some 
assumptions drawn from these. One should be aware that scholars (except 
Slater, below) frequently use 'Cybele' when referring to passages in which 
Pindar has MutllP (MquA,u); this begs some questions and creates con­
fusion. I shall return to the other texts at the end. 

I am indebted to Dirk Obbink, who brought the problems of this fragment 
to my attention, for allowing me the use of the draft of the relevant portion of 
his forthcoming edition of the second part of the De pietate and for his advice 
on all matters relating to the text of Philodemus; and to him and Jacob Stern 
for their comments on the whole. An earlier version was read at the 1993 
meetings of the American Philological Association. 

129 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

130 PINDAR'S "HYMN TO CYBELE" 

Pindar usually discussed in this connection suggest more by the 
evidence they do not supply for Thebes than by what they do. 2 

The source of the fragment is the De pietate of Philodemus, 
who quotes the words in a summary of the opinions of a 
number of poetic authorities about the origins of the gods 
(P.Hercul. 247 fro 6, col. l+N 247 VI left side, lines 17ff; p.19 G): 

TIiv­
[oopoc;] 0' [tIC] KU~E­
[A.llC; ~]11'tPOc; tv 'tC!>' 
OE07t]ow[av] KU~E­
[AaV 1 ~a't[Epa] 

It is preserved only in the copies executed by draughtsmen as 
the successive layers of the carbonized Herculaneum rolls were 
unpeeled. The restoration above is the text established by 
Albert Henrichs. Previous editors, not doubting that the con­
text must have continued after (V 'tw, had drifted farther and 
farther from the remaining traces ~nd spaces. Henrichs re­
examined the disegni and then supplied the cult title. J With this 

2 The following will be cited by author and abbreviations: A. HENRICHS, 
"Toward a New Edition of Philodemus' Treatise On Piety," GRBS 13 (1972: 
hereafter 'Henrichs, GRBS') 67-98, and "Despoina Kybele: Ein Beitrag zur 
religiosen Namenkunde," HSCP 80 (1976: 'Henrichs, HSCP'); E. M. LANE, 
ed., Cybele, Attis and Related Cults: Essays in Memory of M. J. Vermaseren 
(=Religions in the Graeco-Roman World 131 [Leiden 1996]); L. Luigi LEHNUS, 
"Nota al fro 80 di Pindaro," 2PE 10 (1973: 'Lehnus, 2PE') 275ff, and L'lnno a 
Pan di Pindaro (=Testi e documenti per 10 studio deli' antichita 64 [Milan 
1979: 'Lehnus, Pan 'J); A. SCHACHTER, Cults of Boiotia, I: Acheloos to Hera, II: 
Herakles to Poseidon (=BICS Suppl. 38.1-2 [London 1981, 1986]); W. J. 
SLATER, "Pindar's House," GRBS 12 (1971) 141-52; M. J. H. VAN DER WElDEN, 
The Dithyrambs of Pindar: Introduction, Text, and Commentary (Amster­
dam 1991). Pindar is cited after Snell-Maehler. 

3 GRBS 84ff; his supplement is supported by his study of the title in HSCP. 
Henrichs traces the history of the attempts to restore the wording, which 
began with Bergk's assumption that Philodemus was quoting from a dithy­
ramb known then only from a quotation in Strabo (fr. 79 SM), but which is 
now know to be part of Dith. 2 (though neither Snell-Maehler nor Bowra [fro 
77J have removed fro 80 from its old position at the end of the book of 
Dithyrambs, and van der Weiden follows this convention by including it in 
his edition; Turyn placed it with the poems of uncertain genre [fro 148]). For a 
brief history of the Herculaneum documents and an explanation of the most 
recent progress in reconstructing the treatises of Philodemus, see R. Janko, 
"Reconstructing Philodemus' On Poems," in D. Obbink, ed., Philodemus on 
Poetry: Poetic Theory and Practice in Lucretius, Philo demus, and Horace 
(New York 1995) 70-73, and, more fully, Obbink's "Introduction" to his 
edition Philodemus on Piety, Part 1, Critical text with commentary (Oxford 
1996) 24-80. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

JOEL B. LIDOV 131 

text as his basis, Luigi Lehnus argued, in a brief note (ZPE), that 
these words belong to the beginning of a hymn; the name 
would thus appear correctly at the start, an attribute of power 
appears suitably in the initial invocation; and the indirect 
apostrophe in the accusative is a frequent hymnic formula. He 
concluded that Philodemus preferred these words to other 
citations of the Mother because, as the actual opening words, 
they serve as a title to refer to the whole poem. His conclusion 
was embraced by Henrichs, who argues that only Thebes or 
Athens could have been the site for the cult;4 and, of course, 
Lehnus assumes it in his study of Pindar's role in the cult of Pan 
(Pan). The possibility of the hymn is preferred by van der 
Weiden and is given prominence by Maehler in the latest 
Teubner edition of the fra~ments.5 

Before turning to the formal criteria relied on by Lehnus, I 
want to take up the first question that we must ask of any 
poetic fragment: what is its meter? In Pindar's metrics, one 
often has the feeling that anything is possible; nonetheless, some 
things are much more likely, or much more unlikely, than 
others, and if we are being asked to admit these words to the 
prominent position of a hymnal exordium, it would be much 
easier not also to have to mark them as metrically exceptional. 

If the words of the fragment are preceded by two syllables. 
and if the final alpha is long 'by position', they do fit within an 
irregular but still frequent pattern of dactylo-epitrites: -u--uu 
:=::J...C=e_d le, with a missing anceps between dl and e. Obvious­
ly, if there are preceding syllables, these are not the initial 
words. Even with the initial syllables, the fact still remains that 
this particular pattern is not found at the start of a strophe and 
that more than half the time there is a caesura or even period 
end where the missing anceps should be, in this case between 

4 GRRS 98 n.l03; at HSCP 257, he calls Pindar's composition of a hymn to 
Cybele an irrefutable fact. But I do not see that the issue of the original pro­
venance of the quotation is relevant to his general argument that Pindar was 
drawing on the established language of a cult. 

S van der Weiden 22lf; Maehler, ad loc.: -aut ad fro 95 trahendum aut 
initium est hymni <EtC; K\)~£A.'lv > (L. Lehnus)." Wilamowitz had placed the 
fragment in the Hymn to Pan, fro 95: Pindaros (Berlin 1922) 271 n.3, accepted 
in Snell's editions. Even before the new text, Slater (151) had observed that 
this created a difficult-to-explain double reference to her. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

132 PINDAR'S -HYMN TO CYBELE" 

KullEA.o.v and ~.ul'tEpo.. Thus, although these words could be dac­
tylo-epitritic, if they are, they are not likely to be initia1.6 

Among aeolic sequences, the words could form a pherecratic 
colon followed by the first two syllables of an iambic pendant; 
but such sequences are much more likely to form a final ca­
dence than an opening crescendo, and, in any position, most 
such extended pherecratic cola in Pindar have a one-syllable 
basis.! This brings us to what is apparently the simplest of inter­
pretations, the one preferred by Henrichs.8 If the final alpha re­
mains a short syllable, these words could form the beginning of 
a very familiar piece of aeolic verse: an asclepiad, or an even 
longer form of what we would call, in Snell's terminology, a 
glyconic with choriambic expansion: ---uu--uu-u-. Aeolic 
cola of this type are well established in the work of Sappho and 
Alcaeus.9 The familiarity, however, is deceptive. After the Les­
bians, choriambic expansion is extremely rare until Sophocles. 
It is all but absent from the works of Pindar. Given the potential 
confusion between various forms of aeolic with choriambic ex­
pansion and verse forms involving a 'choriambic dimeter', one 
must generalize cautiously. But however a line such as Nem. 6 
ep. 5, ---uu--uu-uu-, is to be analyzed, it represents a 
variation adapted to its context, and is not evidence that a colon 

, For the list of occurrences see Snell's survey of meters, which is printed un­
changed as an appendix to Maehler's Teubner edition: Pindar II: Fragmenta 
(1989) 182. The sequence does appear at the beginning of two periods (Pyth. 1 
str. 2 and Nem. 8 ep. 6; cf. 01. 13 ep. 3, lsthm. 5 str. 6), but I count ten 
instances in which there is period end after d l (01. 11 str.3, 12 str. 4, ep. 6; 
Pyth. 3 str. 4,4 ep. 6; Nem. 8 str. 2; lsthm. 1 ep. 4, 5 str. 2, 6 str. 2, 6). Van der 
Weiden (222) adds the parallels of Nem. 8, which begins ---uu- (=0 with 
substitution in Snell, _d t in Maehler) and of the same sequence within fro 221 
(where there is no responsion to confirm it or clarify the structure; van der 
Weiden's -fro 221 [str. 2]" is misleading). As this pattern is so exceptional, and 
in both instances is followed by an anceps, these do not increase the 
likelihood that fro 80 is an initial dactylo-epitritic sequence. Van der Weiden 
also cites instances of dID, but that pattern is not relevant here. 

7 Pindar uses the full pherecratean by itself as an independent period, even 
an initial one (Pyth. 10 str. 1), but a disyllabic base is not likely to be spondaic 
(Snell's schemes allow this, but see M. L. West, Greek Metre [Oxford 1982] 61; 
exceptions occur: Pyth. 10.25) and an extended pherecratean is not likely to 
have a disyllabic base (cf. OL 1 str. 4 with str. 6). 

8 GRBS 85f with n.55. I am grateful for this opportunity, however delayed, 
to provide a better comment on the meter than the one Professor Henrichs 
acknowledged there. 

9 See E.-M. Voigt's metrical survey in Sappho et Alcaeus: Fragmenta 
(Amsterdam 1971): Sappho C3f-k, Ale. C3c-h. Sappho 103V is an index of 
first lines in a meter with choriambic expansion. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

JOEL B. LIDOV 133 

with choriambic expansion is part of Pindar's normal reper­
toire. tO A better example might be found in fro 106, where the 
rhythm seems to be based on the glyconic; in the fifth line 
apparently we find gf pher c. But as this too exemplifies Pindar's 
normal practice of developing the metrical form of a given ode 
by variation through expansion and compression, it cannot 
justify the use of an asclepiad in the Lesbian fashion as an 
unexceptional form of opening.tt 

Two poems do provide a parallel for this metrical phrase in 
their initial lines. In the case of Of. 5, the parallel is perfect. It 
begins: 

'Y'l'llMV apE'tiiv Kat o'tEcpavwv aw'tov YAUKUV, 
--vv-vv-v-v-=glc cr. 

The opening sequence is metrically identical to fro 80. And in 
Of. 4, each of the first three lines contains one instance of -uu­
-uu-. t2 Both these poems are written for the same patron, 
Psaumis of Camarina, a city in the orbit of Gela and Syracuse. 
But Of. 5 is a troubledjarallel: its status in the Pindaric corpus is 
uncertain, it is marke by an unusual degree of puzzling local 
detail, and more than one unusual quality of its meter has been 
noted.13 Even if we assume Pindar's authorship of Pyth. 5, to­
gether these two parallels still do not suggest tnat this rhythm is 
a normal part of his repertoire, but that he used it for a special 

10 See Snell's alternate analysis in the Teubner edition and West's comments 
(supra n.7: 65, 67); the ode modulates between aeolic and dactylic. Similarly, 
fro 52d str. 8 (of 8) and ep. 3 are, in context, choriambic retardations of a 
continuous sequence of double-short (easily seen in Dale's notation: 
... dd'd'dddds and -d'ddddds). 

II West (supra n.7: 64) cites fro 106.5 as an example of Pindar's methods; cf 
B. Snell, Griechische Metrik 4 (Gottingen 1982) 54-57. Snell's neat schematiza­
tion of the aeolic types and the familiarity of the various asclepiads from 
Horace have perhaps helped to obscure the historical discontinuity in the use 
of choriambic expansion. West notes its rarity in Pindar. Ibyc. 1 ep. 5 (of 5) is 
an exception: -uu--uu-u-, but that appears to be more an ad hoc clausular 
variant than an independent utilization of a pre-existing type. 

12 The first instance is analyzed by Snell as the juncture of two choriambic 
dimeters, the second and third as pherc. 

13 The authenticity remains an unsettled question. See E. von Leutsch, "Ist 
die funfte olympische Ode von Pindar?" Philologus 1 (1846) 11&--27, esp. 124f; 
R. Hamilton, "Olympian Five: A Reconsideration," AlP 93 (1972) 324-29; W. 
Mader, die Psaumis-Oden Pindars (0. 4 und 5): Eine Kommentar (=Comm 
Aenipont 29 [Innsbruck 1990]). The metrical objections center on the combi­
nation of the unusual meter with the "Lesbian sty Ie of composition," i.e. the 
simple development of the three-line strophe, reminiscent of the construction 
of the alcaic stanza, and on the repeated use of the ithyphallic colon at period­
ends. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

134 PINDAR'S ~HYMN TO CYBELE" 

circumstance associated with Camarina. This gives room to 
wonder whether the relationship of the supposed hymn to 
these epinicians may be more than metrical, for the associations 
of Cyoele with the Greek West are strong. The earliest known 
Greek inscription naming Cybele is from western Locris.1-4 And 
what is probably the most notorious reference to the Magna 
Mater in Pindar, Pyth. 3.77ff, occurs in a poem for a Syracusan. 

We are well beyond the bounds where anything can be said 
with certainty, but given the cluster of metrical links (note that 
fro 106 is from a poem to Hieron, and that the expanded 
pherecratean in it occurs in a description of Sicily), it may be 
most economical to conclude that If these words do start a 
Hymn to Cybele, the song was commissioned for a western 
patron. 

With this restriction, then, the metrical argument allows the 
claim made for the fragment. The formal argument speaks more 
strongly against it. The three words are in the accusative, which 
Lehnus calls an indirect apostrophe. An opening in an obli'l,ue 
case is typical of the 'rhapsodic' style rather than the 'cultic .15 

The direct form, the invocation of a god in the vocative, is the 
most ordinary way to begin a prayer; the standard example is 
Sappho 1, nOlKtA,08pov <H:Java:t 'A<PPOOl'ta, and another IS the 
parody of a hymn to Cybele at Ar. Av. 887, O£CJ7tOlVa KU~£A,T}, 
0''tpOU8E, ll11n:p KA,EOKPi.'tOU. All the examples of exordia with 
epIthets of power that Lehnus lists (Pan 276 n.5) as parallels for 
fro 80 are in the vocative case. An accusative beginning depends, 
of course, on a verb that describes the activity of the prayer, as 
in the rhapsodic Hymn to Hermes, which combines a direct 
invocation of the Muse with an indirect invocation of the god 
(4.1): 'Epllftv V Ilv Et, MouO'a, L1l0~ Kat Mauioo~ uiov. The Muse 
need not participate, as in the hymn by Lasus of Hermione: 
L1cXlla'tpa jl£A,1t(O Kopav 1£ KA.UIlEvOl' iiA,0XoV (PMG 702.1). 
Parallels tor Lasus' initial accusative can be found in two 
epigraphical examples that are probably fourth century: PMG 
936, a hymn to Pan from Epidaurus, 

14 M. Guarducci, • Cibele in un epigrafe arcaica di Locri Epizefri," K lio 52 
(1970) 131-38, dating it to no later than the early sixth century and proposing 
Ionia as the source of the cult in Magna Graeca; J. de La Geniere, ~De la 
Phrygie a Locres Epizephyrienne: les chemins de Cybete," ME F RA 97 (1985) 
693-718, argues that figurines found in the Pelopponesus are evidence of an 
earlier cult in peninsular Greece; for a summary of evidence and the subse­
quent discussion, see G. Sfameni Gasparro, ·Per la storia del cuI to di Cibele in 
Occidente: i1 santuario rupestre di Akrai," in Lane 5 1-55. 

IS See W. H. Race. -How Greek Poems Begin," in F. M. Dunn, ed., Begin­
nings in Classical Literature (=YCS 29 [Cambridge 1992]) 20, 28. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

JOEL B. LIDOV 

TIiXva 'tOY Nu~cpay£'tav 
Na'iBCI)v ~£A.ll~' adBro 
XPUO£CI)V xoprov ayaA.~a 

and, as restored, PMC 934, a paean, 

[TIa'iava lCA.U'tO]~l1'tw adoa'tE 
[lCOUpOl Aa'to'i&xv 'ElC]a'tov. 

135 

But Lasus' invocation is unusual among the fifth-century lyric 
forms. Normally, the accusative is postponed by at least one 
major word, most often in the genitive. Thus Bacchylides 
begins an epinician (3.tH): 

aplo'tOlCapnou I.llCEA.(a~ lCpEouoav 
Aa~a'tpa ioo't£cpavov 'tE Koupav 
U~V£\, yA.u1CUBCI)p£ KA.eo'i. 

Closer to our fragment in subject is the scolion PMG 885: 

TIA.ou'tou ~l1't£P' ·OA.u~lav ad&> 
Atl~l1'tpa O't£<Pavll<POPOl~ f:v 6>pat~ 
o£ 't£ na'i At~. 

When Pindar begins with the invocation of a god, as he often 
does, even in the epinicians, he uses the vocative form, but he 
does use the accusative to invoke a city as his theme in Nem. 
10: 

Aavaol> noA.tv ayA.ao8povCl)v 't£ n£v'ttllCov'ta lCOpaV, Xapl't£~, 
"Apy~ "Hpa~ Bro~a 8£Onp£1tE~ U~V£l't£· 

Note again that the first word is in the genitive. There odes in 
which the very first word is in the accusative (e.g. 0/. 10, 13; 
Nem. 7), but their form is not hymnic. The closest parallel to a 
pure accusative hymnic opening in Pindar may be the begin­
ning of his first Hymn, fro 29.1-7, a hymn to Zeus: 

'lo~l1VOV 11 xpuoa)..{h:a'tov MEA.lav 
11 KaB~ov 11 I.nap'tffiv i£pov yEvO~ avBprov 
11 'tCtV 1CUava~nulCa 81lpav 

11 'to nav'toA.~OV o8f:v~ ·HpaKA.£~ 
11 'tCtV AlCl)VUOOU noA.uya8£a 'tt~CtV 

11 ya~ov A.eUlCCI)A.EvOU 'Ap~ovia~ 
u~VtlooJ.lEV; 

These six disjunct accusative phrases appear to form a series of 
discarded foils in an introductory priamel-a parallel that hardly 
suggests that fro 80 is the introduction of a hymn to Cybele. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

136 PINDAR'S KHYMN TO CYBELE" 

If ~EO'1tOlVav KU~EAav Jla'tEpa is quoted from the beginning 
of a hymn to Cybele, these words are most probably not the 
very first words, and the word or words omitted would be 
significant.16 But there is reason to suspect that these words are 
not even part of the beginning phrase of a hymn. A careful look 
at the examples of initial accusatives already cited will show that 
in none do we find the deity invoked with three accusative 
nouns or attributes in a row, as here. If we look at examples of 
vocative invocations (not all necessarily initial) we see that the 
number of nouns or epithets can be quite extended. Three in a 
row occur quite frequently. Four are easily found: Pind. Nem. 
3.1 cn xoww MOloa, Jlau:p aJl£'t£pa, AlOOOJlal), fro 57 
(llroBroval.£ Jl£yaoS£VE<; aplo'to't£XVa xa't£p), PM G 884 (scolion: 
ITaAAu<; TPl'tO,,(EVEl' clvaoo' 'ASTlvu), or the parody from Ar. 
Av. already quoted. Pind. fro 195 works in five: EuapJla't£ 
XpuooXhrov t£pw'ta'tov clyaAJla, eill3a. 17 But in none of the ex­
amples of the indirect invocation, in which the god is named in 
the accusative, does the description of the god-i.e., the list of 
attributes or names typically found in the opening of prayers or 
hymns, and which can be found as part of a vocative series­
extend the invocation to more than two words before some 
other part of the sentence intervenes, something that gives the 
phrase a rhetorical or syntactic shape. Most frequently, a geni­
tive interrupts a potential series of three or more accusatives; 18 

in the scolion PM G 885 the verb; in Bacchyl. 3 an apparent 
series of three actually refers to two different deities. 

Negative arguments based on fragmentary material are always 
suspect, of course, but the pattern here compels attention. We 
have many examples of invocations in both forms, and many 
examples of multiple vocatives, but no lyric examples of more 
than two accusatives. It seems unlikely that this is accidental. It 
makes sense rhetorically. Three of anything create a rhetorical 
unit; the vocative invocations of three words or more consti­
tute a self-contained crescendo, and when that crescendo is 
over, another can begin. The accusatives are by definition sub­
ordinate to a larger structure. But three accusatives, trumpeting 

16 Snell's initial 8£wv would be the obvious choice, but Philoclemus would 
hardly have dropped it from his quotation. 

17 As does the presumably parodic invocation by the cornie poet Chariclicles 
(fr. 1 K): Oe(J1tOlv' 'ElCU'tll 'tptoOt'tl. 'tpittoPqlE.. 1tPUtPOO(1l1t£, 'tpiyAm~ lCllAt\lIL£Va, 
quoted by Henrichs, HSCP 260 n.20. 

18 E.g. Sappho 103.3V (from the index of first lines): ];a 1taiOa KpovlOa 'tav 
i01C[oNtJov [. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

JOEL B. UDOV 137 

the divine name at the start of the poem, would stand out as an 
independent phrase, and thereby confuse the poet's structural 
choice: if he had not wanted to make them part of another 
structure, he could have used the more common vocative. The 
effect is possible, but it does not belong to lyric, which, 
(however long the sentence, and however displaced-from our 
point of view-the word order) builds its crescendos from 
relatively independent sequences of distinct images or ideas. An 
actual hexameter rhapsodic hymn offers the more expansive 
structure that can subordinate such a ringing initial phrase to a 
larger development. There is an artful example at the beginning 
of the Homeric Hymn to Demeter: 

L\lt~l1'tP ' ..;ti1CO~ov aE~vi)v aEW apxo~' aEwnv 
auTTtv ..;~£ SUya'tpa 't<xvuo<pupov ilv 'A i1)CJ)vru~ 
i1 pna9:v, ~ro1CEV ~£ ~apu 1C'tuno~ EU puona ZEU~. 

Here the four accusatives invoking Demeter give way in the 
next line, with a brief summary resumption in the pronoun, to 
the theme of her daughter and Hades (the latter separated by 
the summarizing relative clause with enjambement distributing 
the emphasis), and then to the final crescendo, the involvement 
of ~cxp{)1(tU1tO~ Eupu61tcx ZEl)~. After that, the poet narrates the 
story from its beginning. 

The consequences of treating fro 80 as the beginning of a 
hymn are remarkable. It would display a metrical form associ­
ated with a Sicilian victor, but otherwise quite untypical of Pin­
dar, and it would use this meter to express in lyric a rhetorical 
structure otherwise found only in the hexameter tradition. I 
find this difficult to accept as a probable conclusion. One way 
out of the difficulty, of course, would be a different supple­
ment, with the vocative. 19 I think it is important first to ques­
tion whether we are compelled to seek a solution that saves this 
fragment as an incipit, or even whether we should prefer one if 
it is possible. 

If, as Lehnus proposes (ZPE), these words are cited as a title, 
to stand for the whole poem, they must be the opening words. 
He notes that there were other, non-initial, passages available to 
Philodemus from Pindar (viz., Pyth. 3.78; fro 70b.9=Dith. 2.9, fro 
95.3). Further, the form of the citation here, an introduction of 

19 A vocative would allow the opening [OCcr1t]OlV[' c1] KuJ3c[Aa, crE] ~a"t[cpa], 
which could be continued to produce gl + (')pher, which Pindar uses as 
periods elsewhere. lowe this example to Professor Janko (personal communi­
cation), who provided it only to illustrate the form, not the letter-spaces. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

138 PINDAR'S "HYMN TO CYBELE" 

direct speech by EV tip, is unique in the De pietate, where 
Philodemus generally prefers paraphrase, and introduces direct 
quotation with a verbum dicendi; the use of the words as a title 
would explain the anomaly. But Cybele is named in none of the 
other Pindar passages cited; they refer to the Meter (Megala). 
Philodemus' context requires the name Cybele, in distinction 
to Tithys, Rhea, and Hera-the three mother gods named 
previously. And the problem with this explanation for €V 'tip is 
that it overdetermines the quotation: the words that are quoted 
as the 'title' to identify a whole poem are also the words 
Philodemus needs for his argument at this point. It is otiose to 
suppose that he referring to the poem in its entirety, when he is 
in fact already quoting the relevant portion. But if we un­
derstand €V 'tip to be drawing attention to the actual words of 
the text as evidence (equivalent to 'when he uses the phrase'), 
we can understand that he (or his source) is distinguishing it 
from another possible citation that supplies a different name for 
a mother, for Pindar also uses the phrase OUpav~ (5' £CPPl~E VlV 
leal rata ,.Ul'tTlP in the context of a divine birth (Ot. 7.38).20 
Nothing in Philodemus' citation, therefore, supplies a compel­
ling reason for treating fro 80 as the first words of a hymn to 
Cybele. 

In fact, even if Philodemus is quoting the initial words of a 
song, it does not stand to reason that they must come from a 
poem for Cybele. The corpus of Pindar's surviving work does 
not support even a probable connection between the subject of 
a song and its opening invocation, whether vocative or accusa­
tive. Many of Pindar's poems begin with their theme, as do, of 
course, the great majority of the hymns from antiquity that 
received a less full literary development than his, but it is 
equally part of his style to delay the introduction of the topic. 
Gods frequently appear as part of the postponement. Many of 
the epinicians, hymns to men, begin with hymnic openings to 

20 The line is ironic for emphasis: it is Athena's motherless birth that Gaea 
observes. The personification justifies the treatment of rata as a proper name. 
That name is missing from what we have of Philodemus' list, but is a possible 
reading in the first line of P.Hercul. 1610 fro 3 (=N 1610 III, p.61 G), which 
immediately precedes N 247 VI. This if the first extant line of the passage in 
which the Pindar citation appears (cf Henrichs, GRBS 77ff). In the third line 
the list of poetic sources apparently diverts to those who credited the 
preceding divinities (whoever they were) with giving birth to <Ju!'7tavta; it 
then returns in N 247 VI 6££ to the other candidates for parents of the gods. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

JOEL B. UDOV 139 

gods or divine power before they mention their human 
subject. 21 Fragments 37 (nO'tVUl OeoIlO<P0P£) and 95.1 rn nav) 
are quoted as beginnings, but so is fro 29=Hymn 1, whose 
opening priamel was cited above. 22 The vocative invocation if> 
KU1tpo'\) OE01tOlVIl comes from somewhere in the poem that 
begins (apparently) with the vocative nOAu~evlll VeaVlO£~ (fr. 
128). At Pae. 9 (fr. 52k) we find the vocative 'AlC'tl~ <lEAlO'\) ... if> 
J.1(ltep in lines 1£ and AltllVeuc.o, 'ElCaJ36M in line 38. A dithy­
ramb, fro 75, has the vocative 'OAUJ.11tLOl ... OWL at the ends of 
lines 1 and 2 (cited as the start); the subject is introduced in ac­
cusatives in a prepositional phrase and relative clause in lines 9f: 
E1tl tOV lClOOOOIlTt O£ov, tOV BpOJ.1LOV, tOV 'Epl~oav te ~pOtOl 
lCaA.EOllev. Conversely, if Pindar did write a song for Cybele, 
there is no reason to suppose that it began with an invocation of 
her by name. 

Within a poem, Pindar has a variety of motives for naming a 
divinity; a reference to Cybele (even with an epithet of power) 
need not have come from a hymn dedicated to her. Gods can 
be invoked in the direct speech of a character in a myth (e.g. 01. 
1.75, Nem. 10.76,Isthm. 6.42; cf Bacchyl. 17.52). Prayers or 
descriptions of prayer can occur anywhere in a poem. In the 
epinicians these are usually in the vocative (01. 6.23, O€o1tota 
1tovt0J.1eOov; Pyth. 1.67, Zeu tEMl'), but Pyth. 11 ends with a 
compressed, indirect hymn to Iolaus and the Tyndaridae, each 
first named in the accusative. Pindar often includes descriptions 
of worship: the Hyperboreans in Pyth. 10 and the Delphian 
maidens in the sixth Paean are described as worshipers of 
Apollo. Pyth. 3.77ff-the passage most relevant to a discussion 
of fro 80-is a description of worship: 

aA.A.' E7tEl>l;ao9al j.1EV f:Yrov t9EAm 
Ma'Cpl, 'Cftv Koupal 7tap • Ej.10V 7tp69upov OUV naVl j.1EA.7tOv'Cal 9aj.1a 
OEj.1VaV 9EOV EvVUXl<Xl. 

21 01. 8, 12; Plth. 1 (the lyre), 6 (an indirect invocation), 8, 11; Nem. 7,8,11; 
Isthm. 5. At a . 3 the introductory divine names in the dative postpone the 
theme, in the accusative, to line 2. I omit those addressing a toponymic deity, 
for praise of the place may be part of the theme. 

22 Lehnus (2PE) offers fro 36, ·Aj.1j.1OlV ·OA.uJ.lltOU 5EO"1to'ta. as evidence for the 
divine name in the exordium, but that example begs the question. The words 
are from the scholia to Pyth. 9; the existence of a hymn to Ammon is attested 
by Pausanias; the citation of these words as that hymn's beginning belongs to 
the editors. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

140 PINDAR'S -HYMN TO CYBELE" 

Accusatives can also occur in a narrative. Fr. 30.1ff, the first 
lines of a strophe from somewhere in Hymn 1, to Zeus, pro­
vide a good parallel: 

1tpOnov ~Ev d)~OUAoV 8e~w oupav{av 
xpuoeaunv 'l1t1tOlC; '01ctavou 1tapa 1tay&.v 
Mo~pal ... &:yov. 

Cf Pae. 4.40f, "'tp€ffi 'tOt 1tOAqlOV ~tO~ 'EvvooiDClV 't£ ~Clp{)1C­
'tU1tOV;" Dith. 4.37ff, XPUOOPPCl1ttV ropo£v 'EpIlftV ... leClt 
1tOAtOXOV rAClUlCc.01ttDCl. If Pyth. 3.77ff had been preserved only 
as a fragment, who would have guessed that they came from a 
poem for Hieron? 

I think we can say that in fro 80, just as in the words cited 
above from fr. 30, the weight of the accusative series implies 
that the reference is not trivial (there are not many examples of 
three in a row in any position). But in the works of Pindar, we 
cannot assume that a god who is notably mentioned must have 
either begun the poem or been its primary subject. 

A final parallel-with four accusatives and involving the sub­
ject of the ode-will provide another example of how the 
words might fit within a poem. At the end of the first triad of 
01. 1.22-25, Pindar describes the horse Pherenicus' victory: 

lCPO:ttl OE 1tpOOell€.t~t Oto1tcYcav, III 
1:up<X1COOlOV i.1t7tOxapllav paOtAf\a· Ni~7tt:l oe oi. lCAiOC; 
ev tUaVopl AuSou n€Ao7toc; <l7tOllCiq.· 
'tou ~t"(ao9tvilc; epaooa'to ralaoxoc;. 

The metrical division reinforces the climactic positIOn of the 
phrase in the exordium and then makes it the start of a new 
development, which is itself adapted from hymnic style. 23 This 
suggests that we might print fr. 80 thus: 

Beo7tlow[av] Ku~e[Aa.vJ (II ?) 
~a't[epa] 

This would fit easily into the pattern, e_d1\e (in context, of 
course, IlCl'tEPCl could be elided as well as long by position: e.g. 
j.1Cl'tEp' a. lC'tA.). We would then have a poem of an unknown 

2) The first line of strophe W imitates a Du-Stil invocation: kUPaJCOOlCOV butO­
XaPl1u txxol4u, AaI11t£l or. 'tOl, and is followed in the hymnic style by a rela­
tive pronoun introducing a narrative. Fr. 109.3f offers another striking ex­
ample of a climactic use of the accusative series: O''ta(Jlv altO 1tP<lItiooc; E1tllCotOV 
aVEA.cOV, 1tEVlU~ OOUlp<lV, i.xepltv lCoupmpOcpov. Another motive for their use is 
geographic description: frr. 52b.25, 82, 107.4f. Cf the three genitives of OL 4.7. 
These examples are not meant to be exhaustive. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

JOEL B. LIDOV 141 

type with a significant-but, for us, entirely mysterious-men­
tion of Cybele. A hymn to her is no more probable than 
anything else. 

Whatever the poem was, however, it is unlikely that it was 
written on behalf of the Thebans. Let us look at the passages 
adduced by the proponents of a "mother-goddess" in Thebes 
or in Pindar's life. In addition to this fragment, they are: Pyth. 
3.77ff (already quoted), which connects the MUtllP (OEIlVU but 
not MquACX) with a noctural celebration by 1(OUPCXl, with Pan 
(the "with Pan" is insolubly ambiguous);2<4 two fragments ad­
dressed to Pan as her follower: 95.3, Mcxtpo~ JU:'YUACX~ 01tCXO£, 
and 96, OV tE IlquAcx~ SEOU lCUvcx 1tCXVtoOCX1tOV 1(CXA.£OlCJlV 'OAull-
1tLOl (the testimonia do not reliably associate these with the city); 
Dith. 2.6-9, for the Thebans, envisioning the celebration of the 
rites of Dionysus by the Ou pCXV lOcx l: OEIlVii IlEV 1(CXtUPXEl 
MCX't£Pl 1ttl.p IlEYUA,?- pOIl~l 't,\)1tUVlOV~ and Isthm. 7.3ff, asking 
Theba if she most delights in the time when she exalted XCXA.1(O-
1(po'to'\) 1tUPEOpOV ACXIlU'tEPO~ ... AlOV'\)OOV (I omit references to 
Demeter and Persephone in an unquestionably Sicilian con­
text). 

At issue is not whether there had been a diffusion of religious 
practice from Anatolia to peninsular Greece, but whether 
Pindar's descriptions of a mother-goddess or her worship are 
so specifically Anatolian in character as to justify supposing that 
Pindar would have used an obviously foreign name for her. As 
part of deciding whether any of Pindar's references to the 
Magna Mater refer to Cybele-something taken for granted for 
subsequent centuries-we must also determine whether she is 

24 This passage gave rise to a biographical tradition (kept alive by the tour 
guide who welcomed Pausanias seven hundred years later) that associated 
Pindar with the foundation of a cult for the Magna Mater; moderns have 
embellished it: Wilamowitz explained (supra n.4: 270) that Hieron provided 
the funds for the Theban establishment. The fantasy was exrloded by Slater, 
although Lehnus (Pan 18-43), in a detailed reexamination 0 the biographical 
notices, still argued for some historicity. But Mary Lefkowitz has exposed the 
general methods of ancient biography in her The Lives of the Greek Poets 
(Baltimore 1981), esp. 57-66 on Pindar, and First-Person Fictions: Pin dar's 
Poetic -/- (Oxford 1991) 72-88, 147--60; the matter is summed up by 
Schachter, -The whole case for a 'special relationship' between Pindar and 
Meter rests on the assumption that at Pythian 3.77ff., Pindar is speaking 
propria voce; all later references depend on this interpretation of the passage .... 
It seems, on the whole, to be a Hellenistic invention" (I 166 n.l; more 
forcefully, II 140). Note that the question of whether the 1tp66upov is in Thebes 
or Syracuse (as Slater was arguing) is formally separate from the problem of 
the biography, although they are usually discussed simultaneously. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

142 PINDAR'S -HYMN TO CYBELE" 

separate from the Demeter Thesmophoros worshiped on the 
Theban acropolis or whether Demeter too must be included in 
any possible syncretism.25 Three pieces of evidence clearly link 
the Magna Mater and Pan; the connection would have been 
comprehensible in Sicily but was also possible for Thebes. 26 She 
is associated with the symbols of Anatolian rites only once. in 
Dith. 2-and there no differently from the other gods-in 
which she is accommodating Dionysus, for a Theban audien­
ceY Similarly, Demeter is associated with such rites in the 
context of the recertion of Dionysus to Thebes: the novelty 
may be the point (i XClAKOKPO'tO'U is not taken as a condensed 
mythological reference, it may simply describe war spoils de­
posited at her shrine). Neither Dith. 2 nor Isthm. 7 offers any 
reason to consider either goddess Anatolian independently of 

2, In determining the identity or difference of these divinities for Pindar and 
his audience, it is important to keep separate the early (pre-Euripidean) 
evidence, later evidence, and later comments on earlier evidence (such as 
Strabo's introduction to Dith. 2.8-11), as well as Ionian, western, and penin­
sular provenances, and to note explicitly which names are used. van der 
Weiden's completely uncritical collection of sources (68) cannot justify his 
assertion that the Magna Mater of Dith. 2 is identical to Cybele. For a cata­
logue and succinct appraisal of all the evidence regarding the cults of Demeter 
and Meter (including Cybele) in Thebes, see Schachter I 165-!l8, II 141. The 
three reliefs that might offer evidence for a cult of Cybele are probably later: 
see M. J. Vermaseren, Corpus Cultus Cybelae Attidisque II: Graecia atque 
Insulae (=EPRO 50 [Leiden 1982) 127£. Slater notes (145 n.18) that -Kybele is 
not attested as the name for Pindar's deity by any ancient authority; it is an 
invention of the moderns"; he hesitates to draw any connections among the 
divinities, but sees a syncretism of Demeter and the Magna Mater. The reality 
of the Hymn is taken for granted by Henrichs, who insists on the full 
syncretism and 00 a Thebao locale for the cults (HSCP 256f, with detailed 
objections to Slater's skepticism at n.l0). Lehnus (Pan 13-17), who is willing 
to situate the celebration of Pyth . 3 at Syracuse, separates Demeter and 
Cybele. For the Athenian evidence (the earliest), see M. J. Vermaseren, Cybele 
and Attis: The Myth and the Cult, tr. A. M. H. Lemmers (London 1977) 
32-35; L. Roller, -Reflections on The Mother of the Gods in Attic Tragedy" 
(Lane 305-21), who begins with an overview of the archaeological data. 

2' See Schachter II 139. Lehnus proposes (Pan 17) that the passage in Pyth. 3 
refers to a cult practice any audience in the Greek world could understand in 
local terms. 

27 The juxtaposition with OUpavioal suggests that the Mother here is to be 
identified with the Gaea of the mythic tradition, as at 01. 7.38 (quoted 
earlier). Naiades, Zeus, Ares, and Athena come next, all making noise. Artemis 
drives yoked lions. On the cult of Gaea/Rhea in Thebes, see G. Arrigoni, 
- Alia ricerca dell Mater tebena e dei 7Jeteres di (a proposito della metamor­
phosi di Atalanta ed Ippomene)," Scripta Philoroga 3 (1982) 7-70, esp. 26-30. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

JOEL B. LIDOV 143 

Dionysus.28 The noctural rites ascribed to the Mater in Pyth. 3 
give her as much in common with Demeter and initiation cults 
(the lCoUpat are additionally suggestive)29 as with an Anatolian 
goddess. In sum, the evidence points to an established cult of 
the Great Mother and Pan, and a possibility of assimilating her 
to Demeter;30 for Thebans, either figure can be significantly 
connected to their native son Dionysus. Farther than that we 
should not go. Pin dar was a professional composer of cult and 
celebratory song who worked on commission throughout the 
Greek world. Composing for an audience among whom the 
mother-goddess Cybe1e was recognized (he wrote at. 10-11 
for the western Locrians), he could have learned the correct 
language of her cult. But there is no reason to expect him to 
have used the name associated with a Phrygian origin for a 
Theban audience who appear to have had their own concep­
tions. This may seem a basically skeptical, if not negative, 
conclusion, but if correct it has the positive effect of indicating 

28 The scholarly literature, when it cites these two passages, tends to ignore 
the presence of Dionysus, or treats him as an attribute of Demeter or the 
Magna Mater, even though the emphasis in both places is on him. See e.g. 
Henrich's discussion of both passages (HSCP 256f) and Roller (supra n.23: 
306 n.2, 303 n.26). Lehnus (Pan 14f nn.38f) recognizes the possible influence of 
myth and is more qualified in his treatment of Isthm. 7. B. Moreux (-Demeter 
et Dionysos dans la septieme Isthmique de Pindare," REG 83 [1970] 1-14) 
does not, but still provides a full survey of the possible interpretations of 
XQA.lCOlCpO'tOU; he argues that the presence of the two gods together in Isthm. 7 
is unjustified on theological or cult grounds, and that therefore the connection 
between two divinities must be sought through foreign influence. Nonethe­
less, he concludes (14) that -la cite ... ne pouvait accepter Cybele que si elle 
etait assimilee a Demeter." Noel Robertson (-The Ancient Mother of the 
Gods: A Missing Chapter in the History of Greek Religion," in Lane 239-
304), who argues that the peninsular rites of the Magna Mater are entirely 
independent of the Anatolian, takes Dith. 2 as a description of her festival 
(278). van der Weiden (66, relying on references from Euripides) observes, 
however, that the details of the celebration in Dith. 2 are entirely consistent 
with Bacchic rituals. 

29 Lehnus (Pan 13-16), rejecting the equivalence of Mater and Demeter as­
sumed by Slater, observes that Demeter is only call Meter in relation to Kore; 
he does not see that this passage may be an example of that. See now W. J. 
Slater, -Pindar's Pythian 3: Structure and Purpose," Q VCC N.5. 29 (1988) 56 
n.22, with references to the syncretism of the Sicilian cults. Jacob Stern has 
pointed out to me the particular appropriateness of a reference to a girls' 
initiation rite that forms a contrast to the negative exemplum of Coronis' 
sexual transgression. The participants would be enacting Persephone, hence 
the plural. 

30 On the assimilation of Gaea to Demeter in Thebes, see F. Vian, Les 
origines de Thebes: Cadmos et les Spartes (Paris 1963) 135-38. 



Reproduced with permission of the copyright owner.  Further reproduction prohibited without permission.

144 PINDAR'S "HYMN TO CYBELE" 

that the frank acknowledgement of the worship of foreign gods 
was a significant change that occurred rapidly in the decades 
after Pindar's death.)1 

QUEEN'S COLLEGE AND THE GRADUATE SSCHOOL 

CITY UNIVERSITY OF NEW YORK 

October, 1996 

31 The name Cybele does not occur again in literature before Ar. Av. 887 
(quoted earlier). Roller (supra n.23) argues that the prominence of foreign rites 
in late fifth-century tragedy reflects a new labelling of established forms of 
orgiastic worship as part of a debate about their propriety.