Pindar

What is known. Pindar is generally believed to have been born at Cynoscephalae in Boeotia around 518 BC. His death is placed around 446, not long after Pythian 8.^[1] His earliest surviving poem, Pythian 10, dates from 498, and his last datable one, Pythian 8, from 446.^[2] The Cambridge Companion’s chronology gives “520s/510s to post-446”.^[3] Olympian 9, for a wrestling victory of 466, was written when Pindar was in his fifties.^[4]

He seems to have belonged to a distinguished aristocratic family. But the names of his parents, especially his father, are uncertain. The tradition also names a wife, Megacleia, a son, Daiphantos, and two daughters, Protomache and Eumetis.^[1]

What is only told. The biographical tradition only looks rich. There are five ancient Lives transmitted with the odes, a fragmentary life on papyrus (P.Oxy. 2438), the Suda, and many anecdotes. Yet just one fact about his life is accepted as certain: he was a Theban. Ancient biographies of poets were typically built out of the poems, sometimes refracted through comedy. Familiar story-types attached themselves to famous names.

So the Ambrosian Life tells a story on the authority of Chamaeleon and Ister. Near Mount Helicon, bees built a honeycomb in the sleeping child’s mouth. The story may echo Hesiod’s encounter with the Muses on Helicon. It has a parallel in the tradition about Sophocles.^[5]

A Theban for all Greece. Thebes sided with the Persians in Xerxes’ invasion. Ceos, the island of Simonides and Bacchylides, fought at sea against him.^[2] Pindar nevertheless received commissions from almost every part of the Greek world. So the world-view of his poems is “truly pan-Hellenic”. It complements the Athenian world of the tragedians, with whom he is often compared.^[1] The patrons and occasions are set out on the Victory odes tab.

Language. His poetry is in a largely Doric dialect, with strong Aeolic influence. It allows specifically epic forms, most often in Pythian 4. That is the longest ode. It runs to thirteen triads, where he otherwise never exceeds five.^[6] His native Boeotian has left few traces. Perhaps this is because most of his songs were sung outside Boeotia, to listeners who were not Boeotians.^[7]

Corinna. The Suda says that the Boeotian poet Corinna defeated Pindar five times in competition. In Plutarch she tells the young Pindar to sow by the handful, not by tipping out the whole sack. Aelian reports that he called her a sow. He was perhaps prompted by his own line about the taunt of the “Boeotian pig” (Olympian 6.90). Whether Corinna was his contemporary at all is disputed.^[8]

A classic by the fourth century. By the middle of the fourth century Pindar had something like canonical status. Aristotle quotes him to illustrate general principles. Plato’s Gorgias turns his “custom is the king of all” (fr. 169a.1 S–M) into a philosophical crux. Chamaeleon of Pontus wrote on his life.^[9] How his style struck ancient readers is described on the Style tab.

Prosody. Several rules of Greek prosody have a Pindaric exception.

• Pindar, with Theognis and Solon, shows greater freedom than other early poets in scanning a syllable short before stop + liquid.^[10]
• He admits elision of -αι (-ai) in middle endings. Alone among the poets listed in the handbook, he elides the normally unelidable περί (peri, “around”).^[11]
• A continuant is a consonant whose sound can be held. He occasionally lengthens a final continuant as if it closed the syllable. An example is Nemean 1.51 ἔδραμον ἀθρόοι (edramon athrooi, “they ran in a crowd”), scanned ⏑ ⏑ — — ⏑ —: the -ον of ἔδραμον counts long before the vowel of ἀθρόοι.^[12] West’s text has the two words side by side; Sandys’s prints the line in another order.
• Correption is the shortening of a long vowel or diphthong before another vowel. He allows correption inside a word, as in ἰχνεύων (ichneuōn, “tracking”), where ευ counts short before ω: ἰχ-νεύ-ων, — ⏑ —. He also allows a short syllable before σκ (sk) in σκοτεινόν (skoteinon, “dark”, Nemean 7.61), and before σλ (sl) in the dialect form ἐσλός (eslos, “noble”, for ἐσθλός, esthlos). More often, though, he has ἐσ'λός with the first syllable long.^[13]

What survives. Pindar is unique among the early melic poets. He alone has a large number of complete poems that reach us by direct manuscript transmission.^[1] An epinician is a victory ode, and these poems are the four books of epinicians: the Olympians, Pythians, Nemeans and Isthmians. Of the works of the nine canonical lyric poets, only these came down by direct tradition. Outside the canon, so did the Theognidean corpus.^[2]

Everything else is fragments, chiefly from papyri. There are substantial parts of the book of Paeans from Oxyrhynchus. There are significant fragments of hymns, dithyrambs, partheneia, hyporchemata and enkomia, and very little of the dirges.^[3]

Before Alexandria. How the poems survived before the Alexandrian edition is almost entirely unknown. The Rhodian historian Gorgon tells that Olympian 7 was inscribed in gold letters in the temple of Athena at Lindos. The story may reflect preservation by families and communities.^[4]

In fourth-century Athens the Peripatetic Chamaeleon wrote a monograph On Pindar. Its surviving fragment is on the Corinthian custom of courtesans praying to Aphrodite. It quotes Pindar’s encomium for Xenophon of Corinth (fr. 122, which Chamaeleon calls a skolion), with the opening of Olympian 13. Hadjimichael takes such monographs to have quoted and interpreted the poems, without editing or classifying them. That work was left to Alexandria.^[5]

Seventeen books. The Ambrosian Life says that Pindar wrote seventeen books: hymns, paeans, dithyrambs (2), prosodia (2), partheneia (2), “separated from the partheneia”, hyporchemes (2), encomia, laments and epinicians (4).^[4] Who arranged them is disputed. Two Lives attribute the division into books to Aristophanes of Byzantium. But many scholars doubt that his work went beyond classification.^[6]

Others give the work to Callimachus. Nicholson writes that “Callimachus arranged Pindar’s odes into books”. Barbantani says that he grouped the epinicians by place (fr. 450 Pfeiffer). But Phillips finds “no evidence that Callimachus edited Pindar”. Morrison places Callimachus’ classifying work, such as calling Pythian 2 a “Nemean”, in his catalogue, the Pinakes.^[7]

The order of the epinicians. The books follow the rank of the games. Within each book a rough ranking begins with the four-horse chariot race. Olympian 1, for a single racehorse, was put first by Aristophanes, because it tells of the origin of the games.^[8] Brown describes the order Olympia, Pythia, Nemea, Isthmus as “descending status”. Nicholson holds that the Isthmian book originally came before the Nemean, as the Isthmia ranked higher, and that the two were later reversed.^[9] The sources differ.

Colometry. A colon is a short unit of verse, and colometry is the division of a lyric text into cola. Aristophanes is credited with colometrised texts of Pindar.^[10] Dionysius of Halicarnassus says that “Aristophanes or somebody else” divided Pindar and Simonides into cola. To athetise a line is to mark it as spurious. A scholium reports that Aristophanes athetised a colon on metrical grounds.^[11]

The colometry of the papyri and the Byzantine manuscripts derives from one standard model. But there is “no undisputed evidence” that Aristophanes invented colometry. If Turner’s date is right, a third-century papyrus of Stesichorus already has a colometric layout. Whether the Alexandrian cola reflect the poets’ musical intentions is also disputed (Fleming and Kopff against Parker and Prauscello).^[12] The coronis and paragraphos, which mark poem-end and stanza-end, were systematised in this period.^[13]

Strophic responsion was taught in Roman schools. Statius shows this. He learned from his father by what law the voice of Pindar’s lyre comes back to its start. In Holford-Strevens’s reading, Statius claims to know what Horace had called numbers free of law.^[14]

The selection of the epinicians, the scholia, the titles and the printed editions are on the Scholia and titles tab.

Why the epinicians survived. Up to the mid-third century AD the epinicians were about a quarter of what was read, in papyri as in quotation. Then the other books fade. A codex is a book with pages, as opposed to a roll. Pelliccia associates this selection with the coming of the codex. Unlike festival songs, the epinicians could be dated from victor lists. That made historical scholarship possible, of the kind epitomised by Didymus. It passed into the margins of the texts, and so to Byzantium.^[1] Eustathius gives a reader’s reason: the victory odes are more human, briefer and less obscure.^[2]

Neumann-Hartmann dates the decisive step. In the second century AD a new edition of selected Greek poets began to circulate. From Pindar it took only the victory odes. They were chosen probably for their subjects, and because the Panhellenic games went on. The edition’s lost commentary is the basis of the scholia on them.^[3]

Scholia. Scholia are notes written in the margins of manuscripts. Aristarchus and Didymus wrote commentaries. The medieval scholia descend from such works through repeated redaction. So they preserve them only in part, and through many filters.^[4]

The old scholia on the victory odes keep about sixty-seven fragments of Didymus’ commentary and about as many of Aristarchus’. The match is not chance. Most of what we know of Aristarchus’ views comes through Didymus, who cites him, more often to disagree. Didymus’ commentary was in turn the main basis of the scholia, from the second century AD.^[5] Didymus explained historical allusions from the historians, but he did not edit Pindar’s text.^[6] No discussion of metre by him survives in the scholia.^[7]

Titles and headnotes. Each ode carries a title in the dative, naming the victor, his city and his event. Examples are Ἱέρωνι Συρακοσίωι κέλητι (Hierōni Syrakosiōi kelēti, “for Hieron of Syracuse, with the racehorse”) for Olympian 1 and Μίδαι αὐλητῆι Ἀκραγαντίνωι (Midai aulētēi Akragantinōi, “for Midas of Acragas, the piper”) for Pythian 12. These titles are largely an Alexandrian creation or systematisation, consistent across the books. Where they could, the Alexandrians drew on independent records, such as the Olympic and Pythian victor-lists. The headnotes of the scholia are in effect scholia on the titles. Some of them comment on the title as a text.

The medieval manuscripts have lost the titles of Nemean 8 onwards, and of all the Isthmians. But papyri and scholia preserve several. The book of Isthmians breaks off in Isthmian 9. A papyrus commentary (P.Oxy. 2451) knows Isthmian odes that are lost from it, for a Rhodian boxer and a Megarian runner.^[8]

Editions. The standard text is the Teubner of Snell, revised by Maehler (1987). Its apparatus is austere, and Turyn’s edition (1952) gives more.^[9] Its metrical analyses are the ones West recommends.^[10] The ancient scholarship on papyrus is expected in future volumes of the Commentaria et lexica Graeca in papyris reperta.^[11]

For most of his odes Pindar invents a fresh rhythmic pattern. He keeps each one regular from within by repeating its sections, the “periods”, with great virtuosity.^[1] Mostly he uses the triadic system, going up to five triads, and in Pythian 4 alone to thirteen.^[2] The odes are dactylo-epitrite, aeolic or iambic, with some intermediate mixtures. Detailed analyses of every poem are in Snell’s Teubner edition, revised by Maehler.^[3] Dactylo-epitrite (D/e) is treated here; aeolic and iambic are on the next tab.

Triads, length and classes. A triad is a strophe and antistrophe in responsion, followed by a metrically distinct epode. Brown writes that the odes are “all constructed in triads”. They run from a single triad (Pythian 7) to the thirteen of Pythian 4, nearly 300 lines. The typical ode has about four triads and about 100 lines.^[4] West says only that the three poets “mostly” used the triadic system.^[2] His analysis of Pythian 12 below speaks of its “four strophes”.^[3] The sources differ in emphasis.

Both divide the odes into aeolic, iambic and dactylo-epitrite, with some overlap. Pindar and Bacchylides also combine dactylo-epitrite and aeolic in one poem. They do so in different strophic structures (Bacchylides 3), or even within one strophe (Olympian 13).^[5]

Some odes resist analysis. Olympian 1 is the best known of all, but a precise analysis of it has never been reached. Pindar’s phrase Αἰοληΐδι μολπᾷ (Aiolēïdi molpāi, “in Aeolian song”, 102) encourages its classification as aeolic. That classification is disputed.^[6]

Dactylo-epitrite

D/e is built from d, D, D², e, E and E², joined by link-syllables, which are usually long (the units are set out on the Simonides page).^[7] West’s example is the strophe of Pythian 12. It is given here from the Perseus text, with his analysis:

There may be period-end after μηλοβότου (mēlobotou, 2), but nothing in the four strophes proves it. There is hiatus before (ϝ)άνα ((w)ana) in 3. A short syllable in a final long position is called brevis in longo. The final -να (-na) there, and -τε (-te) at the end of 6, are both brevis in longo. So the period ends at both places.^[3] In line 1 West joins τέω (teō) into one syllable by synecphonesis, the running of two vowels together in one syllable without a change of spelling.

Every link in this strophe is long (in line 3 West marks the α of ἐΰδμᾱτον, eÿdmāton, long). West writes the link in line 7 as ⏓. It is long here (-ᾶν), but short in the fourth strophe, where the line reads οὗτος, ὃ καί τιν’ ἀελπτίᾳ βαλὼν (houtos, ho kai tin’ aelptiāi balōn, “this, which, striking a man unexpectedly”; -πτί-, line 31).

West also marks how stop + liquid is treated. A tie in Ἀκράγαντος (Akragantos, 3) shows the first syllable short before κρ (kr) inside the word. This is the freedom Pindar shares with Theognis and Solon.^[8] Syllable-markers in τέχ'νᾳ (tech’nāi, 6) and διαπ'λέξαισ’ (diap’lexais’, 8) show those syllables long.^[9] Hephaestion quotes Pindar for two combinations of dactylic and iambic lengths, the “iambelegus” and the “Pindaric”.^[10]

Licences in D/e. These are the licences:^[11]

• Resolution can occur in (⏓) ⏕ ⏑ — and, less commonly, in — ⏑ ⏕ — — ⏑ —. Exceptionally, to accommodate proper names, it occurs in — ⏑ ⏕ × D (Isthm. 6.63), and even in — ⏑ ⏑ ⏕ ⏑ ⏑ — (Isthm. 3.63). In about ten cases it recurs in each responding strophe, but more often it is a licence taken once.
• The first biceps of D is contracted only very rarely.
• The catalectic e (— —) occurs in two odes. As in Simonides, it is used to emphasise a word (Pyth. 1.8 ends e — D e∧ ‖). The ithyphallic catalexis of E is not found in Pindar.
• Other abnormal sequences occur occasionally, especially ⏑ ⏑ — at the beginning of a period.

Aeolic

Most odes that are not D/e are aeolic. They are named after the role that certain asymmetrical cola play in them. Above all these are gl and gl¨, tl and tl¨, ph, r, and dod and dod¨ (o o — ⏑ ⏑ —). The sign ¨ marks West’s anaclastic forms. In anaclasis two neighbouring positions change places.^[1]

The aeolic base is the opening pair of positions, written o o. In Pindar its first position can be resolved. So a glyconic may begin — —, — ⏑, ⏖ —, ⏖ ⏑ or ⏑ —. The ⏑ — form is not found in responsion with the others. A double-short form ⏑ ⏑ — ⏑ ⏑ — ⏑ — also occurs, again out of responsion. Pindar also admits resolution later in the colon, and occasionally contraction.^[2]

A small number of the simplest poems can be analysed in these terms straight away. West’s example is the first strophe of Nemean 4. It is given here in the Perseus text, with his labels.^[3] West’s list of symbols gives the cola as follows: tl¨ × — × — ⏑ ⏑ —; dod¨ o o — ⏑ ⏑ —; gl o o — ⏑ ⏑ — ⏑ —; gl¨ o o — × — ⏑ ⏑ —; tl × — ⏑ ⏑ — ⏑ —; r × — ⏑ ⏑ — ×; ia∧ ⏑ — —.^[4]

Notes. In line 3 the two cola are “dovetailed”. The word ἀοιδαί (aoidai, “songs”) runs across the colon-boundary (West’s sign ∫).^[5] West marks long ᾱ (ā) in εὐφροσύνα (euphrosynā, 1), εὐλογία and συνάορος (eulogiā, synāoros, 5) and βαθείας (batheiās, 8), and long ῑ (ī) in ἰατρός (īatros, 2). He puts syllable-markers in κεκ'ριμένων (kek’rimenōn, 1), θύγατ'ρες (thygat’res, 3) and φ'ρενός (ph’renos, 8). In each the stop + liquid closes the preceding syllable, so -κεκ- (-kek-), -γατ- (-gat-) and -σαφ- (-saph-) are long. That is the normal treatment inside a word, and here it is also used across words.^[6]

West’s text has τεύχει (teuchei) in 4 and ἐργμάτων (ergmatōn) in 6, where Sandys prints τέγγει (tengei) and ἑργμάτων (hergmatōn). The metre is the same. The closing iambic metron is already familiar as a tail added to an aeolic colon. In these poets it takes many forms. It is placed before or after an aeolic colon, or in runs of its own.^[3]

Most aeolic poems are harder than this. The opening of the strophe is usually easy to analyse. But having stated an initial “motif”, the poets often develop and embroider it. The sequences that result are bewildering in isolation, but easily derived from what has gone before. What matters is where such a sequence comes from, rather than what to call it. Adding or subtracting a syllable may turn an aeolic colon into one in symmetrical rhythm. In Nemean 6.3–5, for example, a resolved glyconic is followed by a lekythion produced from the glyconic.^[3]

Snell’s edition arranges related sequences in vertical alignment. It notes mutant forms with bracketed sigla such as (gl), which interpret a colon rather than define it.^[7] Itsumi calls this approach “dynamic analysis”. He argues that it should be supplemented by a consistent classification, and he finds two metres, not one, in these odes.^[8]

Iambic

Olympian 2 and fr. 75 are almost wholly iambic. They use the unsyncopated and syncopated metra described on the Simonides page.^[7] A dochmiac unit, ⏑ ⏕ — ⏑ ⏕ (δ), may appear. It is found in Pindar only, and it begins a period. At period-end a trochaic metron — ⏑ — — may also be found.^[7] Olympian 2 opens:

The first period is an iambic metron with a trochaic one at its end. The second begins with that Pindaric dochmius.^[7]

Pindar’s odes are roughly half dactylo-epitrite and half something else (Metres: D/e; Metres: aeolic, iambic). Kiichiro Itsumi’s Pindaric Metre: ‘The Other Half’ (2009) is a full study of that something else. Its subject is the 22 victory odes that are not in dactylo-epitrite.^[1] Readers are sent to it for these odes, and for the vexed metre of Olympian 1.^[2] His terms, his method and his findings are set out below, and so is where he parts company with West.

Scope. Of the 44 epinicians, 23 are in dactylo-epitrite. They can be told apart from the rest without doubt. Only Olympian 13 changes metre in mid-strophe.^[3] Itsumi’s argument rests on parallels and counts. So he bases his statistics on the eighteen non-D/e odes whose stanzas recur at least three times. These are the “eighteen majors”: O. 1, 2, 9, 10, 13 (strophe 1–5), P. 2, 5, 6, 8, 10, 11, N. 2, 3, 4, 6, 7, I. 7, 8. He leaves aside O. 4, 5, 14 and P. 7, which are too short to fix their verse-divisions. He treats them, and the longer fragments, in appendices.^[4] By his count these eighteen odes contain 235 verses.^[5]

Terms. Itsumi follows L. P. E. Parker in his terms. He says verse where West says period. He says phrase rather than colon, since Pindar’s cola are often hard to demarcate.^[6]

His aeolic phrases are built on two six-position units. One is the dodrans, — ⏑ ⏑ — ⏑ —. The other is its mirror image, the reversed dodrans, — ⏑ — ⏑ ⏑ —. A base may stand before them, or not. The base is either a single anceps (his “half-base”) or two positions ○ ○ (“full base”). So the telesillean is × + dodrans, and the glyconic is ○ ○ + dodrans. The heptasyllable × — ⏑ — ⏑ ⏑ — and the wilamowitzianum ○ ○ — ⏑ — ⏑ ⏑ — are the same with the reversed dodrans.^[7]

Lengthened phrases are written with a plus sign for the positions added after the choriamb. So tel+3 is the telesillean with ⏑ — — added.^[8] His “reversed dodrans” is not quite the same as West’s dod¨, ○ ○ — ⏑ ⏑ —.^[9] In Pindar its second position is short in nearly every instance.^[10]

Method. West took over from Snell an approach that Itsumi calls “dynamic analysis”. It follows how each sequence grows out of what came before. It calls a variant a “mutant” of the colon it derives from.^[11] Itsumi grants its merits. But he objects that it cannot show which forms are usual and which odd, and so it can mislead when metre is used to judge the text.^[12] A. M. Dale’s neutral notation describes a verse without naming its cola. It is open to the opposite objection: it gives up classification altogether.^[13]

Itsumi’s own answer is a set of explicit rules for dividing verses into phrases. A boundary goes between two true longs, or before an anceps that stands between two longs, but never before a short. A dactylic run is never divided. A position between two longs is anceps, never a contracted double short. These two principles alone, he estimates, fix perhaps 90% of the boundaries.^[14]

Two metres. Applied consistently, the rules show that the “other half” is not one metre but two. Aeolic is marked by the asymmetrical dodrans and reversed dodrans, and by the absence of link-syllables between phrases. Freer dactylo-epitrite is built, like ordinary D/e, from the symmetrical units e (— ⏑ —) and d (— ⏑ ⏑ —) and their longer forms. But its basic double-short unit is d rather than D. It uses rarer phrases, it often joins units with no link at all, and its verses are shorter.^[15] By his count 101 of the 235 verses, 43%, are freer D/e.^[16]

Three classes. According to how the two metres are combined, the 31 stanza-forms fall into three classes:^[17]

The table condenses his synopsis and lists.^[18] Class III he calls the most characteristically Pindaric, the “consummation of his art”.^[19] Olympian 2 is in Class II.^[20] This reading differs from West’s iambic analysis of it.^[21]

The aeolic base. In Pindar the base takes four forms: — —, — ⏑, ⏑ — and ⏑ ⏑ ⏑. The form — — is rare (six cases), though it is the commonest form in tragedy. Different forms seldom answer each other. The only regular responsion is — — ~ — ⏑, and it is confined to Class I.^[22] Pindar may have been the first to resolve the base, as ⏑ ⏑ ⏑.^[23]

Kiparsky reports Itsumi as showing that Pindar never uses the double-short base ⏑ ⏑,^[24] citing Itsumi’s p. 26. West’s list of Pindar’s glyconic bases includes ⏑ ⏑ —.^[25] Itsumi would admit that form only on a corrupt text. The point is disputed.^[26]

Where he differs from West, Snell and Turyn. Besides the base and Olympian 2, there are three points:

• Divisions. West analyses Nemean 4.1 as tl¨ dod¨. In this he follows Snell and Turyn in dividing before a short. Itsumi reads hepta+2 d.^[27] The two analyses are compared line by line on the Nemean 4 works tab.
• Expansion. Snell’s tables and West’s description lead one to expect much dactylic expansion of aeolic cola in Pindar. Itsumi finds at most six examples, all in freer D/e or amalgamated contexts. He suspects they are not aeolic at all, but variants of D.^[28] Disputed.
• The link-syllable. In his larger Greek Metre (1982) West dismisses the link-syllable of D/e as “a false concept” for the process of composition. That judgement is known here from Itsumi’s and Lourenço’s quotations of it. Itsumi asks whether it is not an a priori argument.^[29] Lourenço shares his doubt.^[30] Disputed.

Reception. Kiparsky cites Itsumi’s earlier work for the unity of the glyconic family, in which glyconic, telesillean and wilamowitzianus answer one another. He also cites it for the refutation of a loose pattern × × × × — ⏑ ⏑ — behind them.^[31] He cites the 2009 book for moving beyond a static list of cola to a “dynamic approach”.^[32] That summary uses the word in the opposite sense to Itsumi’s own. Itsumi calls West’s method dynamic, and argues for keeping a consistent, “static” classification alongside it.^[33]

Ancient and modern verdicts. Horace pictured Pindar as a mountain river in flood (Odes 4.2). Quintilian called him “far the greatest” of the lyric poets.^[1] Dionysius of Halicarnassus made him the model of the “austere” style, and Athenaeus called him “Pindar of the greatest voice”.^[2] For “Longinus”, he and Sophocles burn up everything in their path, though their fire is sometimes quenched.^[3]

Modern judgement has swung widely. For one generation his poems were the model of a Romantic sublimity free of all rules. For another they were the product of chilly rhetorical design.^[4] Wilamowitz found the poet himself lacking in substance (“Er selbst ist kein reicher Geist”). Brown takes this verdict to reflect a failure of interpreters.^[5]

Openings. The odes regularly begin with a rhetorical flourish. Olympian 1 opens with a priamel. A priamel is a list of detached statements, arranged so that the last one receives the emphasis. Here the last is the Olympic games: water is best, gold shines like fire, no star outshines the sun, no contest is greater than Olympia.^[6] Olympian 7 opens with a developed simile. This is a feature of epic, and it is rare in Pindar, who has only five others.^[7]

A shifting voice. Pindar’s focus changes constantly and rapidly. He uses imperatives, and apostrophes to people and places near and far. Pelliccia counts this “constant, rapid change of pragmatic focus” among the hallmarks of Pindar’s manner. So far as the evidence goes, neither Simonides nor Bacchylides shares it.^[8]

Ambiguity adds to the difficulty. In Olympian 9.14 αἰνήσαις (ainēsais) may be an optative (“praise …!”). Or it may be an aorist participle, since Pindar uses the Aeolic participle in -αισ- (-ais-). Pindar poses such problems much more often than Bacchylides.^[9]

In Nemean 4.60–65 it is unclear whether Chiron or Peleus is the subject of ἔκφερεν (ekpheren, “brought to fulfilment”, 4.61). The choice changes the role of Chiron in the myth.^[10] The word order of Pythian 3.8–11 seems designed to make the events look stranger still. No English version can stay faithful to it and still read as English.^[11]

Compression. Isthmian 8.49–55 is a model of Pindar’s compressed style, with its expressive metaphors and its events folded into one another. The race of Pelops in Olympian 1 takes two lines. It is a “bare-bones micronarrative”, whose brevity is itself meaningful.^[12] Bacchylides, by contrast, narrates so that the events tell their own story.^[13]

Catalogue, maxim, myth. Pindar varies the list of the victor’s earlier wins. So in Pythian 9.76–92 a victory at the Theban Iolaia opens into the praise of Iolaus and Heracles.^[14] His gnomic sayings express a world-view shared with his aristocratic patrons. They also carry the transitions between sections. The scholia took the two crows of Olympian 2.86–88 as Bacchylides and Simonides. But the dual is now usually seen as a traditional story-telling device.^[15]

His myths draw on the Epic Cycle, but they seem to avoid details of the Iliad and Odyssey. All but one of the Aeginetan odes tell of the Aeacids. The odes for Hieron favour dark myths of crime and punishment.^[16]

Reading Pindar. For a long time critics sought the unity of each ode in a single governing idea. The result was that Pindar seemed to have nothing to say. Bundy’s Studia Pindarica (1962) recast every element as part of a rhetoric of praise. That correction became reductive in its turn, and a historically informed criticism has since returned.^[17]

Language and rhythm. His Doric carries stronger Aeolic colouring than any other non-Lesbian poet’s. It admits rare Doric forms “when it suits him”.^[18] Most odes have a rhythm made new for the occasion. It is held together by the skilled repetition of sections called periods.^[19] In strophic song the melody overrides the accent. So the difference between Πυθῶνάδε (Pythōnade, “at Pytho”) and Πύθωνάδε (Pythōnade, “at Python”) at Olympian 9.12 was immaterial.^[20]

What an epinician did. A victory ode likens the victor’s achievement to the deeds of heroes. It links a myth to the present through gnomic sayings and through verbal and thematic echoes.^[1] The value of athletic victory was contested. So, Nicholson argues, each ode had two jobs. It had to praise the particular victory, and to establish the worth of athletics as a whole.^[2]

Dates. Precise dates depend on victor lists. These survive in part (the Oxyrhynchus list), or are reflected in the scholia. They cover only Olympic and Pythian victories, and they contain errors. Some odes were also commissioned long after the victory.^[3]

Patrons. His victors came from the whole Greek world:^[4]

• Hieron of Syracuse and Theron of Acragas;
• Arcesilas IV of Cyrene, who won the Pythian chariot race in 462 and the Olympic in 460 (Pythians 4–5);
• the boxer Diagoras of Rhodes (Olympian 7);
• Chromius, a soldier of fortune who became regent of Aetna (Nemeans 1 and 9);
• Psaumis of Camarina (Olympian 4);
• a magistrate of Tenedos (Nemean 11).

Pindar seems to have been “particularly fond” of writing for young Aeginetan athletes. He wrote for Athens too, including Pythian 7.^[5] The scholia assign two odes, for a Rhodian and a Corinthian (Olympians 7 and 13), to the same Olympiad of 464.^[6]

The festivals. The four “sacred contests” awarded only crowns: olive at Olympia, laurel at the Pythia, celery at Nemea and the Isthmus. Pindar knows the hierarchy, with Olympia first. Yet his Olympian odes are neither longer nor grander in diction or structure than his Pythians and Nemeans. Local victories enter the odes in catalogues, and some odes celebrate a victory outside the four games.^[7]

Two odes for Melissus of Thebes share the same metre, which is unique among Pindar’s odes. Isthmian 4 is for a pancration victory, and Isthmian 3 for a later chariot victory at Nemea. The shared metre marks the two victories as one athletic career.^[8]

The “programme”. Brown thinks it likely that the poet received guidance on what to include. So the trainer Melesias is praised in Olympian 8, and the charioteer Nicomachus in Isthmian 2.20–28.^[9] Odes for boy victors in the combat events mostly name the trainer. Charioteers are largely passed over. Pythian 5, where the driver was the victor’s brother-in-law, and Isthmian 2 are the exceptions.^[10] Inherited excellence, φυά (phyā, “nature, inborn quality”), is Pindar’s constant theme.^[11]

Commission and fee. Nicholson calls the odes “typically commissioned by the victor or a family member”.^[1] Brown thinks commission “likely”.^[9] Pindar himself reminds his Muse of her task, “if you agreed, for a fee, to hire out your voice for silver” (Pythian 11.41–2, our translation).^[12] Pelliccia, however, argues that the commission-and-fee model derives from unreliable ancient gossip. In his view the poems’ tone of equality fits poets writing as friends and fellow victors of the men they praised.^[13] The point is disputed.

Chorus or soloist? Choral performance was long assumed from the metres, but the assumption has been challenged. Brown concludes that both modes “are possible”. There might be a chorus at a communal celebration, or a solo singer at a symposium, and perhaps both for the same ode in different settings.^[14]

The controversy was opened by Heath and Lefkowitz (1988), and answered by Burnett and Carey (1989).^[15] Pelliccia thinks Olympian 9 calls its own chorus or kômos into being, wherever it is sung.^[16] Barbantani calls public choral performance the view “widely (though not universally)” held.^[17] The question is disputed.

An ode not by Pindar? Nicholson, following Barrett, treats Olympian 5, for Psaumis, as the work of an unknown poet of the mid-fifth century, preserved in the manuscripts as Pindar’s.^[18] Hornblower leaves its authenticity open.^[19] Its authorship is disputed.