Graeco Aryan metre

It is still Greek and Vedic that show the clearest relationship. This may be because they are two of the oldest attested and bear the best witness to an original system that had become deformed by the time the evidence from other branches comes into view. Or it may be because Graeco-Aryan had developed a particular system that never existed in the same form in other parts of the Indo-European area. In any case the best procedure will be first to see what can be established about Graeco-Aryan metre, and then to inquire how far the results can be extended to MIE or PIE.

The governing principles of prosody and versification are essentially identical in Vedic and early Greek. The unit of composition is a verse containing a

62 Meillet (1923), foreshadowed in previous publications. On the development and influence of Meillet's views see Françoise Bader, 'Meillet et la poésie indo-européenne', Cahiers Ferdinand de Saussure 42 (1988), 97-125.

63 R. Jakobson, 'Studies in Comparative Slavic Metrics', Oxford Slavonic Papers 3 (1952), 21-66 = id. (1962-88), iv. 414-63; Watkins (1963), 194-249 = (1994), 349-404.

64 Cf. Schmitt (1967), 307-13; West (1973), 161-87; Durante (1976), 62-5, 70; Gamkrelidze-Ivanov (1995), 738-40; Watkins (1995), 19-21, 54.

determinate number of syllables.65 The boundaries of the verse do not cut into a word or accentual unit; very often the end of the verse coincides with a syntactic pause. For prosodic purposes the verse is a continuum, the words within it being treated as an unbroken stream of sound, divided into syllables without regard to word division or sense pauses. Between verses this continuity is broken.

There is a clear opposition of long and short syllables, a syllable being long if it contains a long vowel or diphthong or if it ends in a consonant. A single consonant between vowels is assigned to the second of the two syllables, but where two or three consonants occur together (other than at the beginning of the verse) they are divided between syllables, making the prior syllable automatically long.66 Where a short final vowel occurs before an initial vowel, the two syllables are generally reduced to one, whether by amalgamation (Vedic) or by elision of the first (Greek).67 Other short vowels may in certain circumstances be lengthened metri gratia. A long final vowel before an initial vowel generally remains but is shortened by correption. The semivowels i/i and u/u, as the second element of a diphthong, are treated as consonants when the diphthong is word-final before an initial vowel.

A verse is given a recognizable identity firstly by the grouping of words within it, so that in a longer verse there is usually a word-break or 'caesura' after a set number of syllables, and secondly by some degree of regulation of the sequence of long and short syllables, particularly in the latter part of the verse. Word accent plays no part.

Much the commonest species of verse in the Rigveda are:68

(i) An eight-syllable line of the form xxxxu-u - II, used in three-or four-line stanzas (called respectively gayatre and anustubh). A variant type, preferred in some early hymns, has the 'trochaic' cadence u u-- II.

65 In many types of Greek verse (such as the Homeric hexameter) and in later Sanskrit verse the number is subject to variation due to the optional substitution of two short syllables for one long or vice versa. But this is not true of the Vedic hymns, nor of Lesbian lyric, which in this as in some other respects evidently represents the older state of affairs.

66 In Greek there was an increasing tendency not to divide plosive + liquid groups ([pr], [kl], etc.), so that a preceding short open syllable remained unlengthened, especially if word-final. But this was clearly a secondary development.

67 Cf. J. Kurylowicz in Cardona et al. (1970), 425 f.

68 For full details see Hermann Oldenberg, Die Hymnen des Rigveda, i (Berlin 1888), 1-162; E. V. Arnold, Vedic Metre (Cambridge 1905); B. A. van Nooten and G. B. Holland, Rig Veda. A Metrically Restored Text (Cambridge, Mass. 1994), vii-xviii and 577-667; on prosody see also Kurylowicz (1973), 42-96. In the metrical schemes that follow, the symbol x denotes 'indifferent quantity', u 'mostly long', and u 'mostly short'. The final position in a verse is shown as long, but it may always be occupied by a short syllable, the full length being made up by the pause at line-end. The dividers I II III represent respectively regular word-end (caesura), verse-end, and strophe-end.

(ii) An eleven-syllable line with a caesura after four or five syllables and the cadence u — II, used in four-line stanzas (tristubh) or to conclude a song in twelve-syllable lines. The second, fourth, and eighth syllables tend to be long, and the two syllables following the caesura short. The typical schemes are (a) xu xu I U U u u u — II, (b) x-u-u I uuu u — II.

(iii) A twelve-syllable line, also used in four-line stanzas (jagati), resembling the eleven-syllable except that it has an extra short syllable in the cadence, u u-u- II instead of u u — II.

The relationship between (iii) and (ii) is matched exactly in Greek verse, where very frequently, besides lines ending in . .. u-u- II, there occur (often to conclude a sequence) others that differ only in ending ... u--II. The latter are called catalectic in relation to the former, and the former acatalectic. This systemic parallel is one important point of correspondence between Greek and Vedic.

A second feature is that in both traditions the regulation of quantities is strictest in the cadence of the verse and least strict at the beginning. Greek verses in historical times are fairly closely regulated throughout, but—as already noted by Westphal—what seems to be a relic of original freedom at the start of the verse appears in the so-called 'Aeolic base' of certain Lesbian metres, where the first two syllables may be indifferently long or short.

The most typical of these metres, one that can be followed through Greek poetry for many centuries, is the eight-syllable glyconic, xx-uu-u-. Apart from its slightly more fixed scheme, this is very much like the Vedic octosyllable, which indeed may appear in a form identical to the glyconic. Out of the thirty octosyllables that make up RV 6. 54, for example, ten are good glyconics, as in stanza 5ab:

There also occur, particularly in early hymns, seven-syllable verses which in some cases appear as a catalectic version of the eight-syllable (x x x x u -- II), in others as a foreshortened or headless ('acephalic') version (x x x u -u - II).69 Both have counterparts in common Greek verses, the pherecratean (x x - u u —) and the telesillean (x - u u - u -) respectively. There is a catalectic form of the latter known as the reizianum (x-uu —).

In the Lesbian poets the glyconic and other such units are used as

Pusa ga anu etu nah, Pusa raksatu arvatah.

components of stanzas of two, three, or four lines. We do not find, as in the Veda, stanzas composed of three or four lines of the same form; at least one will be different. In Sappho fr. 94, for instance, the stanza consists of two glyconics and a third line which is an expanded glyconic, x x - u u -u u - u - II. In fr. 98 it is two glyconics and a third line in which a glyconic is prefixed with - u -. The same principle of an extended third line is seen in the so-called usnih metre of certain early Vedic hymns (e.g. RV 3. 10; 8. 12-13, 15, 18, 23-6; 9. 102-7) in which the third of the three octosyllables is extended by a repetition of the four-syllable cadence: xxxxu-u-l u-u- II.

This iambic tetrasyllabic element has some similarity to the first part of those eleven- and twelve-syllable verses in which the caesura falls after four syllables, since they tend to have the second and fourth syllables long. What follows the caesura then resembles the independent octosyllable or its catalectic heptasyllabic counterpart. Where the caesura comes after five syllables, what precedes it may be identified as the pentasyllabic colon x-u -X I which can also appear, in duplicate, to form a recognized type of verse called dvipada viraj; the rest of the line is equivalent to a headless octosyllable or heptasyllable.

In other words, if we designate the standard octosyllable as G, its catalectic and acephalic versions as GA and AG, and the tetra- and pentasyllabic complements as 4 and 5, we can analyse the eleven-syllable verse as 4 I GA II or 5 I aGa II, the twelve-syllable as 4 I G II or 5 I AG II, and the usnih stanza as G II G II G I 4 III.

Similar formulae can be applied to Greek lyric verse, with G here realized as the glyconic (gl), 4 as the iambic metron x-u - (ia), and 5 as the penthemimer x - u - x (pe). For instance, Alcaeus' fragments 140 and 358-60 Voigt are composed in stanzas made up of gl + gl + ia (G G 4), while in fragments 70 and 117b the first and third lines of the strophe consist of ia + gl (4 G). The combination of ia + telesillean (4 AG) forms the first two lines of the frequently used Alcaic stanza.

The iambic dimeter, x-u-x-u - (2ia), can be regarded as an alternative realization of the G pattern. Again, the Vedic octosyllable may take this form. Examples from the same hymn as cited above (RV 6. 54) are 1b yd anjasanusisati, 3a Pasnas cakram na risyati, 6a PUsann anu pra ga ihi. The iambic trimeter (3ia) then appears as the counterpart of the Vedic twelve-syllable line, particularly in its 5 I AG II form, as it too normally has its caesura after the fifth syllable. When Archilochus (frs. 172-81) and Hipponax (fr. 118) use stanzas composed of trimeters and dimeters in alternation, this is entirely parallel to Vedic strophes of the forms 12 II 8 II 12 II 8 III (called satobrhati), 12 II 12 II 8 III (krti), 12 II 8 II 8 III (pura-usnih), etc.

I have limited myself here to the most salient points of comparison.70 The sum of correspondences, not only in the structure of individual lines but also in their relationship to one another and in the patterns in which they are combined, is sufficient to show the persistence both in the Rigveda and in Greek poetry of forms already established at any rate by the Graeco-Aryan period.

We have, of course, early evidence for another branch of Graeco-Aryan, namely Iranian. Zarathushtra's Gäthäs are in verse, and we should expect their metres to show some relationship to the system reflected in Vedic and Greek. And so they do, inasmuch as they are based on lines with a fixed number of syllables and a caesura, arranged in strophes of three, four, or five lines. On the other hand there is no regulation of quantities. In view of the agreements between Vedic and Greek, we must suppose that such regulation had existed at an earlier stage but lost its significance. Kurylowicz has argued that its role had been taken over by a stress accent on the penultimate syllable of every word and of the verse.71 A prosodic difference from Vedic and Greek is that the meeting of a final with an initial vowel never results in syllabic loss. This may be connected with the fact that, according to the oral tradition as recorded in the Sasanian orthography, all final vowels had come to be pronounced long, as was also the case in Old Persian.

The lines are of the following varieties: 7 I 9 (Y. 28-34, three-line strophes); 4 I 7 (Y. 43-50, five- and four-line strophes); 7 I 7 (Y. 51, three-line strophes), and in a more complex strophe (Y. 53) 5 I 7 twice, followed by 7 I 7 I 5 twice. They all involve a seven-syllable colon. The 4 I 7 and 5 I 7 combinations can be directly compared with the Vedic eleven- and twelve-syllable lines, the four-line strophe of 4 I 7 corresponding to the Indian tristubh. If Kurylowicz's stress theory is right, the original cadence ... u — II has been replaced by ... x x x II. In the Younger Avesta, as for example in the Hymn to Mithra (Yt. 10), the commonest metre is an octosyllable, varied occasionally by lines of ten or twelve syllables.

There have been several attempts to find metre in some of the Phrygian inscriptions, especially the Neo-Phrygian epitaphs from the Roman period, where a number of recurrent and probably traditional formulae appear. Most recently I have pointed out that several of these appear to show metrical patterns resembling the Greek glyconic and pherecratean, in some cases

70 For fuller surveys of Greek lyric cola and their interpretation in terms of the same system see Watkins (1963), 195-210 = (1994), 350-65; West (1973), 165-70; Gasparov (1996), 54-64.

71 J. Kurylowicz, L'accentuation des langues indo-européennes (2nd edn., Paris 1958), 369-80; id., BSL 67 (1972), 47-67; id. (1973), 102-38. On Gathic metre see also Christian Bartholomae, Die Gadäs und heiligen Gebete des altiranischen Volkes (Halle 1879), 1-19; J. Hertel, Beiträge zur Metrik des Awestas und des Rgvedas (Leipzig 1927); J. Gippert, Die Sprache 32 (1986), 257-75.

prefixed by a four-syllable element (— u u — I or x — u — I) or followed by a pentasyllabic one (I x — u--). I have given reasons for thinking that these patterns were not borrowed from Greek models but inherited from the time when the Phrygians' Bronze Age ancestors lived near those of the Greeks.72

The remaining branch of Graeco-Aryan is Armenian. The metres of classical Armenian poetry are derived from Greek. But the fragments of older, pagan poetry are versified on a different, syllable-counting principle. As in Iranian, syllabic quantities are no longer regarded. The poem on the birth of Vahagn begins with four seven-syllable lines; then, after two of nine syllables, there are two more of seven and two more of nine. Other fragments show hepta- and octosyllables, and sequences of 7 I 6 and 6 I 9 verses.73 We must reckon with the possibility that these metres developed under the Iranian influence to which Armenian culture had long been subject. But they may equally represent a native tradition.

0 0

Post a comment