Homer

None of the sources used on this site offers a life of Homer, and this page does not invent one. The metrical literature does say how the poems were made. That bears directly on how they scan.

The ancient lives. Lives of Homer were written in antiquity, but they are entirely fictitious. They rest on inferences from the poems, on guesswork and on plain invention. The bard Phemius has a large part in the Odyssey, for example, so he was made into Homer’s stepfather. The poems give no real clue about the poet himself, and so seven cities could claim him.^[1]

An oral, traditional verse form. For West the Homeric hexameter is an ancient, traditional form in the hands of oral poets who worked fast. In his view they did not improvise. They assembled their lines rapidly in the head while getting ready to recite.^[2] Nagy, following Parry and Lord, places the composing in the performance itself.^[3] Both treat the verse as oral and traditional.

Milman Parry showed in the 1920s and 1930s that the epics were made in the way illiterate singers make their poems. They were built from recurring metrical phrases, the formulae, made to fit one into the next.^[4] Three things follow. The form carried traditional freedoms that later, more fastidious poets avoided. Those freedoms emboldened some singers to push further. And centuries of linguistic change altered the pronunciation of old formulae. This left metrical anomalies, which were accepted as traditional rather than removed.^[2]

Time depth. West reckons with six to eight centuries of development before the Iliad and Odyssey. A conservative poetic language lived on through that time and slowly renewed itself. Meanwhile the spoken language changed around it, in some ways that altered the prosody of words.^[5] The epic language is a blend of Aeolic and Ionic forms from different periods.^[6]

Some phrases scan only if one restores a stage of Greek earlier than the Mycenaean tablets. ἀνδροτῆτα (androtēta), “manhood”, Ἐνυαλίῳ ἀνδρειφόντῃ (Enyaliōi andreiphontēi), “to Enyalius the man-slayer”, and νὺξ ἀμβρότη (nyx ambrotē), “immortal night”, presuppose syllabic r̥, an r that forms a syllable by itself. The Linear B texts have already lost it.^[7] Διὶ μῆτιν ἀτάλαντος (Dii mētin atalantos), “equal to Zeus in counsel”, is explained from a Mycenaean formula, Diweí mḗtim ḥatálantos.^[8] The epithet in ἀσπίδος ἀμφιβρότης (aspidos amphibrotēs), “of the man-covering shield”, may refer to the body-shield of the sixteenth and fifteenth centuries.^[9]

Tichy, working on the Iliad, distinguishes an old layer of at least 250 years’ depth from Homer’s own Ionic hexameters. The old layer is rooted in an Aeolic tradition of oral lyric. She does not take it back to Mycenaean times, formulae and perhaps a few single verses apart.^[10] Nussbaum, in the same volume, identifies Neo-Ionic forms that entered the epic repertory at a relatively recent date.^[11]

Two poets. West treats the Iliad and the Odyssey as the work of different poets. In his judgement the Odyssey poet, whatever his merits, handled words and verse less well than the poet of the Iliad. He borrowed the Iliad poet’s lines and half-lines again and again, and did not always fit them neatly to their new place.^[12] The reuse of Il. 12.447–448 at Od. 9.241–242 is West’s most glaring case.^[13]

Whether one poet made both epics is disputed. A few ancient critics, the chorizontes or “separators”, already gave them to different poets. The introduction to the Loeb Iliad calls the question open. Since separate authorship has not been proved, it keeps to the traditional single Homer.^[14]

Formulae and the metre. Whether formula shaped metre or metre shaped formula is disputed. Nagy holds that formula generated metre, and Kiparsky the reverse.^[15] Parry’s definition of the formula and the arguments are on the Formula and metre tab. The earliest hexameters in writing that Kiparsky cites are the two on Nestor’s Cup, of the eighth century. They are preceded by an iambic line with a choriambic first foot.^[16]

In epic, especially in formulae, the prosody sometimes follows the syllabification of older forms of words than those transmitted. This is notably so of forms containing the digamma, the old sound [w] that Ionic lost. In the majority of cases, words that once had a digamma are scanned as if it were still there.^[17]

Disputed. The origin of the hexameter is disputed. Kiparsky reviews the proposals.^[18] West (1973) suggested two shorter lyric verses welded together: a pherecratean and an expanded reizianum, or a hemiepes and a paroemiac. Nagy (1974) suggested a pherecratean expanded by three dactyls. He restated this in 1990 as a verse assembled from smaller inherited metres.^[19] Berg (1978) suggested an eight-syllable dimeter followed by a pherecratean.

Kiparsky himself derives it from a fused pair of iambic eight-syllable dimeters, by syncopation. West later conceded that the origins can never be known exactly. But he noted that the most widely held theory derives the hexameter from two shorter verses resembling the classical glyconic and pherecratean.^[20]

The Homeric question. The modern debate about how the poems were made began in 1788. That year Villoison published the ancient notes, the scholia, from the margins of the Venetus A, a tenth-century manuscript of the Iliad. In 1795 F. A. Wolf argued from them that the poems were put together long after Homer by a compiler. His starting point was that Homer’s age could not write, which is now known to be false.^[21] Scholars then divided into two camps. The Analysts looked for layers by different hands. The Unitarians stressed the design of the whole. Parry’s discovery of the formula systems changed the terms of the debate.^[22]

When were the poems made? The date is disputed. Two chapters of the Cambridge Companion to Homer put the Iliad about 700 BC, or a little after. The poems know things not found in Greece before the later eighth or early seventh century, such as writing, temples with cult statues and settlement overseas.^[23] M. L. West argued for a later date, in the 670s to 650s BC.^[24] A statistical study of the language by Richard Janko puts the early epics in the order Iliad, Odyssey, Theogony, Works and Days.^[25]

When were they written down? One old view holds that the poems were first written down in sixth-century Athens, for the Panathenaic festival under Peisistratus and his sons. The ancient sources show that a text was made for the festival, but not that it was the first one.^[26] Janko’s order also tells against the view. If the poems were fixed one after another, they were not all fixed together in sixth-century Athens.^[25]

Writing material was not a problem in Homer’s time. The Greek word for “book”, βύβλος (byblos), comes from Byblos, the Phoenician port that supplied papyrus, and parchment was also available. The earliest Greek inscription dates from about 770 BC.^[25] On one view the poet dictated his songs to a writer, as Parry’s South Slavic singers dictated theirs to him.^[27]

By the middle of the sixth century the texts were in general circulation. Until then the Homeridae probably kept them. They were a guild of reciters who claimed descent from Homer.^[28] In the archaic period most of the epic Cycle and the Homeric Hymns were ascribed to Homer as well.^[28]

A well-preserved text. The text has come down in good condition. On the whole it is of very high quality.^[1] Of some 28,000 lines, very few raise any difficulty at the caesura.^[2] So a verse that will not scan is a verse to be explained, not a reason to loosen the rules.

Where the tradition has interfered. The metrical anomalies that do survive show copyists and ancient scholars at work:

• At Il. 4.202 some copyists wrote the place-name Τρίκης (Trikēs) with one kappa to put the metre right: Τρῐ-κης, ⏑ —, against Τρίκ-κης, — —. Yet at Il. 2.729 it has two, and scans as a spondee.^[3]
• At Il. 9.414 scribes wrote ἵκωμι (hikōmi) or ἵκοιμι (hikoimi) for ἵκωμαι (hikōmai), forms that do not exist.^[3]
• ἀδροτῆτα (adrotēta) for ἀνδροτῆτα (androtēta), and ἀβρο- (abro-) for ἀμβρο- (ambro-) at Il. 10.65 and 14.78, are efforts to mend what looked like a fault in the metre.^[4]
• The unmetrical ἕως (heōs), “until” (eighteen cases, plus three of τέως, teōs), stands in almost every witness. Traces of other spellings in papyri and manuscripts are isolated.^[5] The metrician Heliodorus (c. AD 100) read a trochaic first foot. Voss, Hermann and Curtius proposed older forms. Curtius’ ἧος (hēos), τῆος (tēos) were adopted by Nauck, Cauer and others.^[6] Monro’s text, which this site quotes, prints ἧος at Il. 1.193. West 1987 prints the transmitted ἕως and notes “for older ἧος”.^[7]
• Sometimes a syllable is lengthened by a following continuant, a consonant like n or m that can be held (τεν-νιφάδες, ten-niphades; δαμ-μελίην, dam-meliēn). There papyri and sometimes medieval manuscripts write the consonant double.^[8]
• Lengthened ε and ο are written ει and ου in our text. This is a matter of spelling only: the earliest texts will have had plain Ε and Ο.^[9] At Il. 21.362 ζεῖ (zei), “boils”, represents ζέει (zeei), the reading of a few manuscripts.^[10]

The standard text. Even after the poems were fixed, lines were added and changed, as early quotations and papyri show. But the changes were kept in check by a fixed point of reference.^[11] A stable standard text, the vulgate, emerged under the Alexandrian scholars, above all Aristarchus in the second century BC. The number of lines in his edition proved decisive. He left out lines that were very poorly attested. He kept other suspect lines but marked them with a critical sign. This marking is called athetesis, and the practice began with Zenodotus early in the third century BC.^[12]

Aristarchus’ preferred readings in single lines had much less success than his count of lines. His “edition” was probably a good manuscript that he chose and linked to his commentary by critical signs.^[13] Extra lines in the medieval manuscripts, or in the early “wild” papyri, are usually easy to spot.^[13]

Editions named in the sources. Fowler names M. L. West’s Iliad and H. van Thiel’s Odyssey as the best modern editions.^[13] West’s own Teubner edition records conjectures on Il. 9.414, and obelizes (marks as corrupt) the half-line at Il. 18.458.^[14] Richard Payne Knight’s edition of 1820 printed genitives in -οϝο (-owo). West finds it eccentric but often perceptive.^[15]

Allen tells the same story from the side of the sounds. Bentley in 1713 first saw that the lost digamma explained Homer’s anomalies. There are some 2,300 cases of absent elision, 400 of lengthening by position and 160 of absent correption. But Payne Knight’s attempt to restore the letter in the text ran to excess. Allen advises the reader not to attempt it in reciting.^[16] Ahrens in 1843 explained apparent lengthening in Ἰλίου προπάροιθεν (Iliou proparoithen), “before Ilium”, by a genitive in -οο (-oo).^[17]

The ancient metricians. Hephaestion takes his example of the dactylic hexameter, “the so-called ἔπος (epos)”, from the opening of the Iliad. He illustrates the indifferent final syllable from Il. 2.1–2.^[18] So the Homeric text he knew already served as the schoolroom model of the metre. He defines the “common” syllable (κοινή, koinē), one that can count as long or short. His first case is a long vowel or diphthong followed by a vowel. He illustrates it from Od. 1.1–2, Ἄνδρα μοι ἔννεπε (andra moi ennepe), where μοι is short before ἔ-, and πλάγχθη ἐπεί (planchthē epei), where -θη is short before ἐ-.^[19] This is the ancient statement of what West calls correption, and West uses the same two lines.^[20]

This site’s text. Homer is quoted from the Perseus Digital Library edition of D. B. Monro (Iliad), under CC BY-SA 4.0. The reading of a line is noted wherever West’s discussion presupposes a different text.

The scheme. The hexameter has six dactylic feet or metra (— ⏔). The last is catalectic, cut short to — ×. Seen from the side of the language, the line stylises natural phrasing into fixed sequences of light and heavy syllables. It has six feet, the sixth truncated. The fourth foot is usually dactylic, and the fifth overwhelmingly so.^[1] A caesura is a word-end at a fixed place inside the line. West marks three such places with dots. There is always a caesura at one of them at least, normally at one of the two places in the third foot.^[2] In the table ¦ stands for his marks:

A caesura after the third princeps (the third long) is penthemimeral. One after the fourth princeps is hephthemimeral. In the third foot the “feminine” caesura, — ⏑ ¦ ⏑, predominates over the “masculine”, — ¦ ⏔, in the proportion 4 : 3. Formulaic phrases assume it more often. Just occasionally a long word extends across the whole third foot, and the caesura falls after the fourth princeps.

Because of the regular break, West regards the line as two cola: — ⏑ ⏑ — ⏑ ⏑ — ¦ ⏔ — ⏑ ⏑ — ⏑ ⏑ — × ‖ or — ⏑ ⏑ — ⏑ ⏑ — ⏑ ¦ ⏑ — ⏑ ⏑ — ⏑ ⏑ — × ‖. These cola also occur independently in other metres. For West they, not the notional “feet”, are what the line is actually built from. Many of epic’s repeated phrases are shaped to fill one colon or the other.^[2]

Battezzato builds the line the same way. A hemiepes and a paroemiac make the line with the masculine caesura. A feminine hemiepes and an enoplion make the line with the feminine caesura.^[3]

In Homer there is word-end at the hephthemimeral position in 45% of lines (Porter’s count). Only about 1% of lines lack the regular central caesura. These usually break after the second and fourth princeps instead, as Il. 6.3 ἀλλήλων | ἰθυνομένων | χαλκήρεα δοῦρα (allēlōn | ithynomenōn | chalkērea doura), “(as they aimed) their bronze-tipped spears at each other”.^[4]

The same picture appears from the other side. About 99 lines in 100 have a word-end near the middle of the line, in the third foot. About nine lines in ten also have one in the fourth foot, either after its first long or at its end.^[5] Some scholars, but not all, posit a third regular break near the start of the line, after the first foot or after the first long of the second.^[6] With these breaks the line falls into four cola. Repeated words and phrases tend to take exactly these shapes and places. A name with its epithet, for example, often fills the line from the feminine caesura to the end, as ἄναξ ἀνδρῶν Ἀγαμέμνων (anax andrōn Agamemnōn), “lord of men Agamemnon”, does.^[7]

Bridges. A bridge works the other way from a caesura: it marks a place where poets avoid ending a word. Hermann’s Bridge (1805) avoids word-end between the two shorts of the fourth biceps. Prepositives count with the following word, and postpositives with the preceding. Breaches occur about once in 550 lines in Homer.^[8] Maas reckons about once in 390 lines. He puts the rarer caesura after the seventh element at about once in 100 lines, usually to fit a proper name.^[9]

Two further habits of word-placing are agreed by all who have counted. A spondaic fifth foot is hardly ever followed by word-end, with three certain exceptions in Homer. And words shaped — — — — stand at the end of the line in over 95% of their 1,445 Homeric occurrences. Allen explains this pattern by the stress he supposes on heavy final syllables.^[10]

Word-end after an uncontracted fourth biceps, the “bucolic caesura”, is frequent.^[11] It comes in 47% of Homeric lines.^[12] Callimachus later avoids two rhythms that Homer admits. One is word-end after a contracted fourth biceps (Naeke’s Law: Il. 1.2 Ἀχαιοῖς | ἄλγε’, Achaiois | alge’, in the box below). The other is word-end between the shorts of the second foot, or at its end, when there is none at the end of the first (Il. 1.1 μῆνιν ἄειδε | θεά, mēnin aeide | thea).^[13]

Prosodic licence. No other kind of Greek verse takes as many prosodic liberties as Homer’s.^[11] The main heads:

Il. 21.359–360 shows several of these together:^[15]

Digamma. Many words once began with ϝ, the digamma, pronounced [w]. Usually, and above all in formulae, they still scan as though the consonant were there. Yet it was no longer pronounced when the poems were made.^[16] The sound survived in some dialects. In Homer it works mostly inside formulaic phrases.^[17]

Contraction. Contraction is the replacement of the two shorts of the biceps by one long. The biceps is a little longer than the princeps, about 6:5. So a single long in its place puts the rhythm under slight stress. Contraction is least common in the fifth foot, about once in twenty lines in Homer. There it often comes from vowel contraction (Od. 12.7 Ἠῶ, Ēō, one long, for older Ἠόα, Ēoa, two shorts after the long). Word-end after a contracted biceps is frequent only in the first foot. In the fourth foot such a word usually ends in a long vowel (Wernicke’s Law).^[18] West’s specimen, Il. 1.215–222, is scanned on the Dactylic page.^[19]

The first foot is the most loosely regulated part of the line. It has anomalous short syllables in sentence-initial words such as ἐπεὶ δή (epei dē), ἴομεν (iomen) and κλῦθι (klythi). The clausula, the close of the line, is strict, with a marked preference for a dactylic fifth foot.^[20]

Oral composition. Epic was composed in a stichic metre, one repeated line by line: the dactylic hexameter. Its poetic language (Kunstsprache, “art-language”) combined Aeolic and Ionic dialect forms from different periods.^[1] Since Milman Parry and Albert Lord it has been studied as oral poetry. Parry defined the formula as “a group of words which is regularly employed under the same metrical conditions to express a given essential idea”. He based the definition on the Homeric poems alone (Nagy, quoting Parry).^[2]

Fieldwork on South Slavic epic showed that every performance was a distinct composition. A singer who composes as he performs needs a ready-made phrase for every position in the verse.^[3] Epic was accompanied by a stringed instrument. Its delivery is sometimes thought closer to recitative than to song. But Homeric language shows a strong conception of epic as song.^[4]

Formulae and the cola. The description of the metre points the same way. The regular word-break in the third foot divides the line into two cola. West treats these as the true building blocks of the verse. Epic’s recurring phrases are often made to fit one colon or the other. They assume the feminine caesura more often than the masculine.^[5] Formulaic phrases are frequently joined at the caesura. How smoothly the join is made depends on the poet’s skill.^[6] Parry had already observed that the hexameter’s regular word-breaks mark where formulas begin and end.^[7]

Nagy’s example is an epithet and its slot:^[8]

In both lines πολύτροπος (polytropos), ⏑ — ⏑ ⏑, fills the stretch between two junctures. Nagy, after Fränkel, labels them B2 and C2. Here they are the feminine caesura and the bucolic caesura.^[9] The generic epithet διίφιλος (diiphilos), “dear to Zeus”, would fit the same slot. The Hymn to Hermes uses πολύτροπος there for Hermes (lines 13 and 439).

The phrase πολύτροπον ὅς / πολύτροπος ὅν (polytropon hos / polytropos hon) runs across the metrical break at C2. Nagy suggests that such “spilling” over a juncture may be traditional. It may reflect rhythms older than the current norms of the hexameter.^[10]

Which generates which? (disputed)

Nagy: formula before metre. Page’s “law of economy” allows one formula, and one only, for a given idea in a given place in the line. For Nagy this is an effect of traditional diction, not its cause. The regulator is traditional theme rather than metre, and he speaks of a “trend toward thrift”.^[11] His definition leaves metre out. The formula is “a fixed phrase conditioned by the traditional themes of oral poetry”. Over time it is formula that gives rise to metre, not the other way round.^[12]

Seen synchronically, at one moment in time, the metre frames the formula. Seen historically, formula shapes generated colon shapes. The hexameter, put together from smaller inherited metres, can hold two formulas or more.^[13] His evidence is comparative. In the oldest Indic poetry, phraseology is regular even where metre is not.^[14]

Kiparsky: metre before formula. The commonest fixed formulas, such as πολύμητις Ὀδυσσεύς (polymētis Odysseus), “Odysseus of many wiles”, and κορυθαίολος Ἕκτωρ (korythaiolos Hektōr), “Hector of the flashing helmet”, are suited to the hexameter. They could not have become formulaic before such metres existed. It “seems unlikely”, Kiparsky writes, that formulas created metre as Nagy proposed. Fixed formulas presuppose a metre, and flexible formulas and themes are not metrically constrained.^[15]

Flexible formulas move about. The mist poured over the eyes appears as κατὰ δ’ ὀφθαλμῶν κέχυτ’ ἀχλύς (kata d’ ophthalmōn kechyt’ achlys), “and mist was poured over his eyes” (Il. 5.696), and κατ’ ὀφθαλμῶν χέεν ἀχλὺν (kat’ ophthalmōn cheen achlyn), “he poured mist over his eyes” (Il. 20.321). It is placed at the masculine or the feminine caesura, and Archilochus recasts it in iambics. Formulas, Kiparsky concludes, adapt to metrical change, and new ones are made for new metres.^[16]

West’s figures show how a poet managed an awkward word. The vocative Τηλέμαχε (Tēlemache), — ⏑ ⏑ ⏑, occurs thirty-one times in the Odyssey. It is elided before a vowel twenty-seven times, and stands before a double consonant three times. Only once is it used irregularly.^[17] Hainsworth’s “flexibility” of the formula is itself regular.^[18]