Verse formsStichic metres
The elegiac couplet
The elegiac couplet: a hexameter followed by the shorter line traditionally called a pentameter.
The metrical families one by one: dactylic, iambic, trochaic, aeolic, ionic, anapaestic, dochmiac and the rest.
Elegiac · 1 of 4
The scheme of the couplet
Strictly speaking the elegiac couplet is a strophic, not a stichic metre. A stichic metre repeats the same line over and over, as the epic hexameter does; a strophic one repeats a stanza of several lines. The couplet is a two-line stanza, repeated as often as the poet wishes. Its small compass and wide use in extended compositions make it convenient to treat with the stichic metres, as West does.[1] It is, in effect, a short strophe.[2] In West’s classification of Ionian epodic verse it is the form hexameter ‖ D ¦ D |||.[3] In that formula D stands for the hemiepes, — ⏔ — ⏔ —, where ⏔ is a biceps, a position of two shorts or one long; ¦ marks a caesura, ‖ the end of a period and ||| the end of the stanza.
The two lines
| line | scheme | notes |
|---|---|---|
| hexameter | — ⏔ — ⏔ — ¦ ⏑ ¦ ⏑ — ¦ ⏔ — ⏔ — × ‖ | as in epic, but with less prosodic licence[4] |
| “pentameter” | — ⏔ — ⏔ — ¦ — ⏑ ⏑ — ⏑ ⏑ — ||| | two hemiepes cola, D ¦ D, always divided by a caesura; contraction of the bicipitia only in the first colon[5] |
Here is a whole couplet by Anyte, the hexameter set out in its six feet and the “pentameter” in its two halves:
1 ἵζευ ἅπας ὑπὸ καλὰ δάφνας εὐθαλέα φύλλα
2 ὡραίου τ’ ἄρυσαι νάματος ἁδὺ πόμα
Sit, everyone, beneath the laurel’s lovely, thriving leaves, and from the seasonable stream take a sweet drink.
Anyte, Greek Anthology 9.313.1–2. Line 1, the hexameter, in six feet, with the caesura (dashed bar) after καλὰ (kala). Line 2, the “pentameter”, in two hemiepes with the caesura between them. Red: ζευ (zeu), a diphthong shortened before the vowel of ἅπας (hapas).
synthetic ∿
The second line is traditionally called the pentameter. West keeps the name in inverted commas, since the line is really two hemiepes cola.[4] Each colon is two and a half feet long (Hephaestion’s “penthemimer”, πενθημιμερές, penthēmimeres: five half-feet). So the name is traditional but misleading.[6]
Hephaestion describes the line in just those terms. In our translation of his Greek: “Take the dactylic penthemimer twice and the ἐλεγεῖον is the result. Its second part is always seven syllables, two dactyls and a syllable; in the first, the two feet can change, so that they are both dactyls or both spondees, or the first a dactyl and the second a spondee, or the other way round”.[7] The result is a line of fourteen, thirteen or twelve syllables. Hephaestion lists the four shapes, with a Callimachean example of each:[8]
| syllables | first colon | example[9] |
|---|---|---|
| 14 | — ⏑ ⏑ — ⏑ ⏑ — | παίσατε, τῶν δ’ ἐτέων ἡ δεκὰς οὐκ ὀλίγη |
| 13 | — ⏑ ⏑ — — — | Νήϊδες οἳ Μούσης οὐκ ἐγένοντο φίλοι |
| 13 | — — — ⏑ ⏑ — | ῥέζειν, καὶ στεφέων εὔαδε τῷ Παρίῳ |
| 12 | — — — — — | τίκτεσθαι· βροντᾶν δ’ οὐκ ἐμόν, ἀλλὰ Διός |
Note that only the second colon, doubled, would make an ἐλεγεῖον (elegeion, Hephaestion’s name for the “pentameter”). The first cannot, unless it happens to consist of two dactyls.[10] Hephaestion adds the rule of the caesura in the strongest terms. The ἐλεγεῖον must in every case be cut at one or other of the penthemimers. Otherwise it is faulty, πεπλημμελημένον (peplēmmelēmenon). He cites a line of Callimachus, ἱερά, νῦν δὲ Διοσκουρίδεω γενεή, in which Διοσκουρίδεω runs across the division.[11] West and Hephaestion thus agree on the structure, D ¦ D, and on the invariable caesura. Only the names differ.
The endings
The hexameter of the couplet ends — × like any hexameter. The pentameter ends on the single long of the second hemiepes. Since that is period-end, the position may be filled by a short syllable, called a brevis in longo, “a short in a long”.[12] Elision, the loss of a short final vowel before a word beginning with a vowel,[13] is admitted at the caesura, as in the hexameter. The sense need not pause at the end of a couplet, but it usually does.[4] The first couplet of Theognis’ seal shows most of this:[4]
19 Κύρνε, σοφιζομένωι μὲν ἐμοὶ σφρηγὶς ἐπικείσθω
20 τοῖσδ’ ἔπεσιν· λήσει δ’ οὔποτε κλεπτόμενα
Cyrnus, let a seal be set by me, as I practise my art, on these verses; they will never be stolen unnoticed.
Theognis 19–20: Greek from Bergk 1882 II p. 119. Dashed bars: the caesuras. Red, in 19: the final position, × (here -θω, -thō). Red, in 20: λή- (lē-), a contracted biceps, allowed only in the first half; δ’ (d’), δέ (de) elided at the caesura; -να (-na), a short syllable in the final long position, the brevis in longo.
synthetic ∿
In line 20 the elision of δέ falls at the caesura, and the last position is occupied by the short syllable -να.[4]
Where the hemiepes comes from
The colon — ⏑ ⏑ — ⏑ ⏑ — is not peculiar to the couplet. In Archilochus’ epodes it stands as a separate period after a hexameter or an iambic trimeter. There it admits no contraction and no hiatus, but does admit correption in the first biceps.[14] Hiatus is the meeting of two vowels across word-end with both kept at face value.[15] Correption is the shortening of a final long vowel or diphthong before a word that begins with a vowel, as ζευ (zeu) before ἅπας (hapas) in Anyte’s couplet.[16] The hemiepes also opens the Iliad, μῆνιν ἄειδε θεά (mēnin aeide thea), which D’Angour uses to introduce it. The pentameter is therefore the pairing of two of these familiar cola. The first is allowed the contractions of the hexameter, and the second is kept pure.[4] Euripides’ one elegiac lament shows the same pattern. Its second lines can be set out as D : D, contracted (if at all) only in the first half.[17]
| feature | hexameter of the couplet | pentameter |
|---|---|---|
| contraction | as in the epic hexameter, where the fifth biceps is the one least often contracted (about once in 20 lines in Homer)[18] | first colon only[4] |
| caesura | third foot, as in epic | invariably between the cola |
| end | — × ‖ | — (brevis in longo allowed) ||| |
| prosodic licences | princeps freedoms as in epic, but more seldom exercised than in stichic hexameters | the same; hiatus at the caesura particularly rare[4] |
Notes
- ^ West 1987, p. 23.
- ^ Bartol 2022, p. 221.
- ^ West 1987, p. 31.
- ^ a b c d e f g h West 1987, p. 24.
- ^ West 1987, p. 24; Battezzato 2009, p. 138.
- ^ Bartol 2022, p. 221 n. 1.
- ^ Hephaestion, Ench. 15.14; cf. Bartol 2022, p. 221 n. 1, on the two and a half dactyls of each half.
- ^ Hephaestion, Ench. 15.14–15.
- ^ Callimachus, ap. Heph. 15.15.
- ^ Hephaestion, Ench. 15.14.
- ^ Hephaestion, Ench. 15.15.
- ^ West 1987, pp. 4, 24.
- ^ West 1987, p. 13.
- ^ West 1987, pp. 31–32.
- ^ West 1987, p. 15.
- ^ West 1987, p. 14.
- ^ Lourenço 2011, p. 69, on Andromache 103–16.
- ^ West 1987, pp. 20, 24.
Elegiac · 2 of 4
Usage and poets
The elegiac couplet belongs to the Ionian tradition of the eighth and seventh centuries, alongside epic and iambus.[1] Its origins lie in Ionia, and Tyrtaeus at Sparta took over a form of Ionic with the metre.[2] Archilochus, Tyrtaeus and a host of other archaic and fifth-century poets composed in it. From the middle of the sixth century it was more and more the metre of verse inscriptions.[3] In the sixth century it rose to become the commonest metre of epigram.[4] Kelly judges that the couplet already had a long history behind it when it first appears in Archilochus.[5] In the later centuries it was one of the three stichic metres that came to dominate poetic literature, with the hexameter and the iambic trimeter.[6] And it became the standard metre of epigram.[7]

The archaic and classical poets
| poet | what the sources say |
|---|---|
| Archilochus | used the couplet;[8] in his epodes the same hemiepes colon that makes up the pentameter stands as a separate period after a hexameter or a trimeter, so that the couplet properly belongs among his epodic forms, hexameter ‖ D ¦ D |||[9] |
| Tyrtaeus | named by West with Archilochus as an early user[8] |
| Theognis | West’s specimen of the archaic couplet is Theognis 19–24, the “seal” of the collection, and Theognis 2 illustrates the freedoms of the pentameter’s princeps;[8] de Kreij connects Anacreon’s choice of literary Ionic with the tradition of iambus and elegy and with Theognis’ erotic elegies[10] |
| Simonides | wrote both melic poetry, largely Doric, and elegy, which is basically Ionic; the dialect of his elegies and epigrams is primarily Ionic because the convention of the genre required it[11] |
| Sophocles (as elegist) | Hephaestion reports that Sophocles, in his elegies, thought the name Ἀρχέλαος would not fit into either an ἔπος or an ἐλεγεῖον, and wrote Ἀρχέλεως, “for so it was in measure”[12] |
The last item is a small but telling witness to the couplet’s rigidity. The poet alters the name rather than the metre. In general, special licences are occasionally taken to fit essential proper names into the verse. In verse inscriptions, where the need to include names was especially pressing, the scheme sometimes gives way altogether under the pressure.[13] Hephaestion’s other example is more drastic still: Parthenius, writing an elegiac lament for Archelaïs, made the final line an iambic trimeter instead of a pentameter, so as to get the name in.[12] Critias had done the same in an elegy for Alcibiades. Merging two shorts into one, Hephaestion explains, is rare in epic verse, and so Critias thought the name could not be admitted. He put it in an iambic trimeter in place of the pentameter, and said why:[14]
Ἀλκιβιάδην νέοισιν ὑμνήσας τρόποις·
οὐ γάρ πως ἦν τοὔνομ’ ἐφαρμόζειν ἐλεγείῳ·
… Alcibiades, praising him in new ways; for there was no way to fit the name to an elegiac line.
Critias fr. 4 W, lines 2–3, ap. Heph. Ench. 2.3
synthetic ∿
Bartol reads the substitution as a witty touch of mockery, typical of the metrical experiments of classical elegy.[15]
Dialect
Elegy is an Ionic genre whatever the poet’s origin. Simonides of Ceos, an Ionic island, wrote melic poetry that is largely Doric, but elegies that are basically Ionic.[16] The same generic pull is visible in Anacreon’s literary Ionic.[10] Much later the Hellenistic poet-scholar Callimachus could carry Pindar’s rich lyric language over into non-lyric metres too, the hexameter and the elegiac couplet. Julia Balbilla, at Hadrian’s court, composed four epigrams for the Colossus of Memnon in East Aeolic. She deliberately echoed Sappho’s language while writing in elegiac couplets.[17] The metre, that is, stayed fixed while the dialect became a matter of choice.
Epigram and the long elegiac poem
Hephaestion’s own examples of the pentameter’s four shapes are all from Callimachus,[18] a sign of how far the Hellenistic poet had become the model. Epigram is normally written in elegiacs. Up to the first century AD poets often pick them for longer works too, above all for poems on sub-heroic legend or giving medical recipes.[7] The couplet is the metre of verse inscriptions and literary epigrams.[19] The sixth-century poets of Agathias’ Cycle went back to a far stricter technique, with Nonnian habits in their hexameters. Agathias’ own epigrams, though, are looser than the hexameters of his preface.[20]
More poets. Callinus, Mimnermus, Solon, Xenophanes and the classical elegists, and what Bartol says about how elegy was performed, are on the tab Performance and poets.
Notes
- ^ West 1987, pp. 2–3.
- ^ Battezzato 2009, p. 138.
- ^ West 1987, pp. 23–24.
- ^ Aloni 2009, pp. 169–170.
- ^ Kelly 2022, p. 35.
- ^ West 1987, p. 68.
- ^ a b West 1987, p. 80.
- ^ a b c West 1987, p. 24.
- ^ West 1987, p. 31; D’Angour 2022, ch. 9, §’Archaic Lyric Meters’.
- ^ a b de Kreij 2022, ch. 10, §’Anacreon’.
- ^ de Kreij 2022, ch. 10, §’Introduction’, §’Simonides’.
- ^ a b Hephaestion, Ench. 1.5.
- ^ West 1987, p. 9.
- ^ Hephaestion, Ench. 2.3.
- ^ Bartol 2022, pp. 231–232.
- ^ de Kreij 2022, ch. 10, §’Simonides’.
- ^ de Kreij 2022, ch. 10, §’Late-Classical and Hellenistic Lyric’, §’Imperial Lyric’.
- ^ Hephaestion, Ench. 15.15.
- ^ Bartol 2022, p. 221.
- ^ West 1987, p. 81.
Elegiac · 3 of 4
Rules and later refinements
The archaic rules
The hexameter of the couplet is the epic hexameter, but sparer with prosodic licences.[1] The pentameter has its own constraints. The caesura between the two hemiepes is invariable, and contraction of the bicipitia is limited to the first colon.[1] Within these limits the archaic poets allow themselves the following:
| licence | note | source |
|---|---|---|
| elision at the caesura | admitted, as in the hexameter (Thgn. 20 λήσει δ’ ¦ οὔποτε) | West 1987, p. 24 |
| brevis in longo | the last position of the pentameter is at period-end and may be short (Thgn. 20 κλεπτόμενα) | West 1987, p. 24 |
| synecphonesis | Thgn. 22 Θεο- of Θεόγνιδος as one syllable; 23 (Μεγα)ρέως with -εως as one syllable | West 1987, p. 24 |
| freedoms of the princeps | the pentameter’s princeps has much the same freedoms as the hexameter’s, but both lines of the couplet use them less often than hexameters written line by line | West 1987, p. 24 |
| hiatus at the pentameter caesura | particularly rare | West 1987, p. 24 |
| no sense-pause required at couplet-end | though in practice there usually is one | West 1987, p. 24 |
West’s example of a princeps freedom is the second line of the Theognidea:
2 λήσομαι ἀρχόμενος οὐδ’ ἀποπαυόμενος.
I shall not forget (you) as I begin or as I end.
Theognis 2. Dashed bar: the caesura between the two halves. σεῖο (seio), “you”, the god, is in line 1.
synthetic ∿
Theognis 2, as West quotes it.[1]
Here -μαι is correpted before ἀρχόμενος.[2] The naturally short -νος of ἀρχόμενος stands in the princeps before the caesura, and is treated as long. This is lengthening by a final continuant, which occurs in epic, elegy, Pindar and Bacchylides.[3]
Hephaestion’s rule is simpler and absolute. The ἐλεγεῖον has always to be divided at one of its two penthemimers. If it is not, it is faulty (πεπλημμελημένον). He does not hesitate to call a line of Callimachus faulty on this ground.[4] West’s “separated invariably by caesura” is the same rule in modern dress.[1]
Hellenistic practice
In the Hellenistic period versification in elegiacs is in general rather stricter than in the hexameter poets:[5]
- Hiatus is kept within narrow limits.
- Words that would fit the verse without correption, those shaped (… ⏑ ⏑) — —, are correpted less often.
- Before the feminine caesura of the hexameter the poets normally put a word that ends — ⏑ | by nature, not one that reaches that shape only through correption or elision.
- From the late third century BC the syllable before the pentameter’s caesura tends noticeably to hold a long vowel or diphthong, rather than owing its length to the consonant that begins the next word. The growing preference dates from the second half of the third century BC.[6]
- More and more, the pentameter is kept from ending on an accented syllable. The same trend runs away from oxytone and perispomenon endings towards paroxytones.[7]
The last point is one of the effects of the word accent on metre. From the Hellenistic period onward the accent acquired an element of stress, which began to erode the opposition of long and short in speech.[8] Three of Hephaestion’s four Callimachean pentameters[4] end on an unaccented syllable (ὀλίγη, φίλοι, Παρίῳ). The fourth ends in Διός. The tendency was not yet a rule.

After AD 50
Standards of versification become rather laxer after about AD 50. In the third and fourth centuries hiatus and brevis in longo at the pentameter’s caesura become surprisingly free. West compares this development to the failures of synapheia between cola found elsewhere in verse of the Imperial period.[9] That is, the two hemiepes began to be treated as if each were a period on its own. In the sixth century the poets of Agathias’ Cycle tighten the technique a great deal. Their hexameters follow Nonnus’ tendencies, if less rigorously. Agathias is freer in his own epigrams than in the hexameters of his preface.[10]
| period | pentameter caesura | hiatus and correption | accent |
|---|---|---|---|
| archaic and classical | invariable; elision allowed; hiatus very rare | princeps freedoms as in epic, used sparingly | no effect |
| Hellenistic | invariable; from the late third century BC a long vowel or diphthong preferred before it | hiatus restricted; correption avoided where unnecessary | pentameter increasingly not ended on an accented syllable |
| third to fourth century AD | hiatus and brevis in longo tolerated at the caesura | lax | growing |
| Agathias’ Cycle (sixth century) | strict again | Nonnian tendencies | as in Nonnus, less rigorously |
Notes
Elegiac · 4 of 4
Worked example: Callimachus, Epigram 2.1–2
This is the first couplet of Callimachus’ epigram on his friend Heraclitus of Halicarnassus. Hephaestion draws on Callimachus for every one of his pentameter examples.[1] West sees the Hellenistic hexameter at its most refined in Callimachus.[2] So the couplet can be read against both.
1 εἶπέ τις Ἡράκλειτε τεὸν μόρον, ἐς δέ με δάκρυ
2 ἤγαγεν, ἐμνήσθην δ’ ὁσσάκις ἀμφότεροι
Someone told me of your death, Heraclitus, and brought me to tears; and I remembered how often we two …
Callimachus, Epigram 2.1–2, by feet. Dashed bars: the caesuras.
synthetic ∿

The hexameter
Feet. 1 εἶπέ τις, dactyl. -τις is short because its final ς goes with the following vowel.[3] 2 Ἡρα-, spondee. The syllable ρα- is long before κλ. The metre requires it, and within a word stop + liquid normally makes a long syllable in hexameter verse.[4] 3 -κλειτε τε-, dactyl. 4 -ὸν μόρον, dactyl. -ὸν is a closed syllable before μ-, so it is long.[5] 5 ἐς δέ με, dactyl, with ἐς closed before δ-. 6 δάκρυ: the syllable δακ- is long before κρ, and the free final position follows.
Caesura. Word-end after Ἡράκλειτε falls after the first short of the third foot. This is the feminine caesura, which Callimachus prefers in a high proportion of his lines.[6] The word before it, Ἡράκλειτε, naturally ends — ⏑. That is exactly what Hellenistic practice wants in that place, rather than a word that attains the shape only by correption or elision.[7]
Bucolic caesura. The fourth foot, -ὸν μόρον, is dactylic and ends with word-end. This is the bucolic caesura, which rises from 47% of lines in Homer to 63% in Callimachus.[2] The sense-pause after μόρον falls at it.
Bridges and avoided rhythms. Hermann’s Bridge is respected: the fourth biceps, μό-ρον, lies inside a word (Maas counts a word-end after the fourth trochee about once in 390 lines of Homer and never in Callimachus’ Hymns and Hecale, where every line also has its third-foot caesura).[8] Naeke’s Law does not arise, as the fourth biceps is not contracted.[9] The second foot, Ἡρά-, is a spondee inside a word. So there is no word-end at the end of the second foot without one at the end of the first.[10] There is no effective word-end after the fifth princeps. ἐς is a prepositive and δέ, με are postpositives, and appositives count with the word they lean on.[11] Every one of the Callimachean rules West lists is kept in the line.
The pentameter
First colon. ἤγαγεν | ἐμνήσθην δ’, — ⏑ ⏑ — — —. The first biceps is uncontracted and the second contracted (-νή-), which is permitted only in this half of the line.[12] In Hephaestion’s terms the first penthemimer here has a dactyl followed by a spondee. That makes the line one of his two thirteen-syllable shapes.[13] -γεν is short because the ν is followed by a vowel.
Caesura. The invariable caesura falls after δ’, which is elided before ὁσσάκις. Elision at the caesura is admitted, as in Theognis 20.[14] The syllable before it, -σθην, contains a long vowel, so its length does not depend on the word that follows. That is the preference that becomes perceptible in pentameters after the late third century BC.[7]
Second colon. ὁσσάκις ἀμφότεροι, — ⏑ ⏑ — ⏑ ⏑ —, is the pure hemiepes, two dactyls and a syllable in Hephaestion’s description, with no contraction.[15] -κις is short before ἀ-. The final -οι is long, so there is no brevis in longo to show the period-end; the end of the couplet shows it instead. Pentameters in general shun a short open vowel in that last position. Allen cites Martin’s count of 98 in 1,512.[16]
Accent. The pentameter ends on ἀμφότεροι, whose last syllable is unaccented, in keeping with the growing Hellenistic avoidance of an accented final syllable.[7] The proportions fall from about 18% of couplets in early elegy to 12% in the Alexandrians.[17]
| check | hexameter | pentameter |
|---|---|---|
| scheme | — ⏔ — ⏔ — ⏑ ¦ ⏑ — ⏔ — ⏔ — × | — ⏔ — ⏔ — ¦ — ⏑ ⏑ — ⏑ ⏑ — |
| contraction | foot 2 only | first colon, second biceps |
| caesura | feminine, after a word naturally — ⏑ | invariable, with elision |
| bucolic caesura | yes, with sense-pause | not applicable |
| Hermann’s Bridge | kept | not applicable |
| period-end | — × ‖ | — ||| |
Notes
- ^ Hephaestion, Ench. 15.15.
- ^ a b West 1987, p. 78.
- ^ West 1987, p. 12.
- ^ West 1987, p. 16.
- ^ West 1987, pp. 12–13.
- ^ West 1987, pp. 19, 78.
- ^ a b c West 1987, p. 80.
- ^ West 1987, p. 21; Maas 1962, pp. 60–62.
- ^ West 1987, p. 78; Maas 1962, p. 62.
- ^ West 1987, p. 79.
- ^ West 1987, pp. 9, 21, 79.
- ^ West 1987, p. 24; Battezzato 2009, p. 138.
- ^ Hephaestion, Ench. 15.14–15.
- ^ West 1987, p. 24.
- ^ West 1987, p. 24; Heph. Ench. 15.14.
- ^ Allen 1973, pp. 204–205.
- ^ Allen 1973, p. 266.
