The Center for Hellenic Studies

Comparative Studies in Greek and Indic Meter

  Use the following persistent identifier: http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Nagy.Comparative_Studies_in_Greek_and_Indic_Meter.1974.

1. The Common Heritage of Greek and Indic Meter: A Survey

With the ultimate aim of comparing the metrical contexts of Greek κλέοc ἄφθιτον and Indic śráva(s) ákṣitam, we must start by reviewing the conventions of Greek and Indic meter which have been recognized as cognate by Meillet and his followers. [1] My approach will be to consider first those of the conventions which are directly comparable. From there I will move on to latent comparanda, indicating what seems to have remained the same and what has become different as a result of change. In certain respects, Greek meter might have undergone some changes which are resisted by Indic meter, or vice versa. Moreover, I should note straightaway that formal difference between the two does not necessarily mean that one withstood change while the other did not. There may well be instances where both have undergone change, but in different directions. Then too, change can also be parallel. In short, we have to reckon with the factors of: (1) common archaism, (2) archaism vs. innovation, (3) diverse innovation, (4) common innovation. [2] For the moment, then, I will confine myself to features which are overtly comparable. The descriptions {27|28} in the following list apply to Greek and Indic meter simultaneously:

The rhythm of the meters operates on the principle of opposing long (–) vs. short (⏑) syllables.

A long/short vowel followed by one consonant counts as a long/short syllable, regardless of word-boundary.

A long/short vowel followed by two consonants counts as a long syllable, regardless of wordboundary.

A long vowel at the end of one word may be shortened when it is directly followed by another vowel at the beginning of the next word.

Semivowel/vowel interchange is possible. (For example, prevocalic ι is optionally nonsyllabic in Greek, while prevocalic y is optionally syllabic in Vedic. In both Greek and Vedic versification, such variation between syllabic and nonsyllabic forms is lexically conditioned.) [3]

The meters are isosyllabic, in that there must be the same number of syllables in every instance of a given type of verse. The most common Vedic verse-types consist of 12, 11, 10, or 8 syllables (= dodeca-, hendeca-, deca-, or octosyllables), for each of which there is a Greek equivalent. The dodecasyllables and hendecasyllables are called trimeters and the octosyllables, dimeters. (In Greek, there is a marked erosion of isosyllabism in many meters, {28|29} resulting mainly from the process of substituting a long for two shorts. Isosyllabism persists, however, in such primitive Aeolic meters as the Sapphic hendecasyllable.)

The last syllable of a verse is optionally long or short (⏓).

The last syllable may be deleted in such a way that the syllable which had been next-to-last becomes the new last syllable, likewise optionally long or short (⏓), even if it had originally been short only (⏑) or long only (–). This double process of deleting one last syllable and creating another will be called catalexis. (For example, the Indic hendecasyllable

⏓ – ⏓ – ⏑ ⏑ – – ⏑ – ⏓


equals the dodecasyllable

⏓ – ⏓ – ⏑ ⏑ – – ⏑ – ⏑ ⏓


after catalexis. Or again, the Greek Pherecratic

⏓ ⏓ – ⏑ ⏑ – ⏓


equals the Glyconic

⏓ ⏓ – ⏑ ⏑ – ⏑ ⏓


after catalexis.) I will use the term ‘catalexis’ simply to designate the derivation of one meter from another. In order to specify an operative mechanism whereby an acatalectic meter is actively converted into a catalectic counterpart, I will use the term ‘synchronic catalexis’. [4] {29|30}

The verse is divided into an opening and a closing, [5] which are marked by flexible and rigid rhythms respectively. (For example, consider the Greek and Indic octosyllables shaped

⏓ ⏓ ⏓ ⏓ ⏑ – ⏑ ⏓: [6]


the rhythmically flexible opening ⏓ ⏓ ⏓ ⏓ contrasts with the rigid closing ⏑ – ⏑ ⏓.)

In Greek Lyric, the inherited flexibility of the opening becomes progressively restricted. For example, the octosyllable known as the irregular Glyconic, [7]

⏓ ⏓ ⏓ ⏓ ⏑ – ⏑ ⏓


evolves into the regular Glyconic,

⏓ ⏓ – ⏑ ⏑ – ⏑ ⏓,


with syllables 3 and 4 restricted to – and ⏑ respectively. [8] Sometimes, metrical tendencies in Rig-Vedic openings recur as constants already generalized in the corresponding Greek. [9] For example, the mere statistical frequency of – ⏑ – in syllables 2 3 4 of the Rig-Vedic hendecasyllable {30|31} is matched by the regular placement of – ⏑ – in syllables 2 3 4 of the corresponding Greek hendecasyllable, the iambic trimeter catalectic. [10]

In the Rig-Veda, the rhythmical rigidity of the closing is not yet a constant, as in Greek, but only a strong tendency. For example, the closing ⏑ – ⏑ ⏓ of the standard Rig-Vedic octosyllable

⏓ ⏓ ⏓ ⏓ ⏑ – ⏑ ⏓


is the statistically predominant pattern, but other patterns are also tolerated: – – ⏑ ⏓, ⏑ ⏑ ⏑ –, – ⏑ ⏑ ⏓, ⏑ ⏑ – –, etc. [11] By contrast, the closing ⏑ – ⏑ ⏓ of the Greek Glyconic

⏓ ⏓ – ⏑ ⏑ – ⏑ ⏓


is not just the preferred pattern but actually the rule. In this respect, the Rig-Vedic evidence is more archaic than the Greek. [12]

Some Rig-Vedic metrical tendencies are innovations, such as the gradual elimination of double-short sequences (⏑ ⏑) from most positions in the verse, especially from the closing. [13] Unlike other positions in the verse, however, double-short sequences are exceptionally common at syllables 5 and 6 or 6 and 7 of dodecasyllables and hendecasyllables: [14] {31|32}

1̄̆ 2̄ 3̄̆ 4̄ 5̆ 6̆ 7̄ 8̄ 9̆ 10̄ 11̆ 12̄̆
1̄̆ 2̄ 3̄̆ 4̄ 5̄̆ 6̆ 7̆ 8̄ 9̆ 10̄ 11̆ 12̄̆
1̄̆ 2̄ 3̄̆ 4̄ 5̆ 6̆ 7̄ 8̄ 9̆ 10̄ 11̄̆
1̄̆ 2̄ 3̄̆ 4̄ 5̄̆ 6̆ 7̆ 8̄ 9̆ 10̄ 11̄̆


Notice that this isolated Rig-Vedic instance of a metrical tendency to generalize ⏑ ⏑ is far back from the end of the verse.

Unlike Indic verses, where the sequence ⏑ ⏑ is rare in the closing, certain Greek verses not only preserved the double-short but also regularized it as a metrical constant. Consider the closing of the Greek octosyllable known as the choriambic dimeter:

1̄̆ 2̄̆ 3̄̆ 4̄̆ 5̄ 6̆ 7̆ 8̄̆


The double-short is also a constant in another octosyllable, the Glyconic:

1̄̆ 2̄̆ 3̄ 4̆ 5̆ 6̄ 7̆ 8̄̆


(The long of syllable 5 in choriambic dimeter and the long of syllable 3 in the Glyconic are generalized as a consequence of the double-short in syllables 6 7 and 4 5 respectively, in consequence of an inherited constraint against a triple-short.) [15] By actively retaining the capacity for sequences shaped ⏑ ⏑ in the closing, Greek Lyric preserves an archaic pattern more effectively than does the Rig-Veda. [16] Nevertheless, the {32|33} actual regularization of ⏑ ⏑ into a metrical constant is a Greek innovation. Note too that the opening/closing in the Glyconic is realigned from 1 2 3 4/5 6 7 8 to 1 2/3 4 5 6 7 8, at least in terms of an opposition between flexible and rigid rhythms. [17]

Within the original closing, the last four syllables of a dimeter, an overt double-short sequence could fit in only two positions:

… – ⏑ ⏑ ⏓ (as in choriambic dimeter)
… ⏑ ⏑ – ⏓


The latter is the closing of the so-called Indo-European paroemiac, attested not only in Greek and Indic but also in Slavic. [18] Among inherited Greek dimeters which adhere strictly to the principle of isosyllabism, the only well- known type with regular paroemiac closing is a heptasyllable:

1̄̆ 2̄̆ 3̄ 4̆ 5̆ 6̄ 7̄̆ = Pherecratic


This meter, of course, is the catalectic variant of the Glyconic:

1̄̆ 2̄̆ 3̄ 4̆ 5̆ 6̄ 7̆ 8̄̆


By contrast, Indic preserves the paroemiac closing ⏑ ⏑ – ⏓ not in heptasyllables but in octosyllables:

1̄̆ 2̄̆ 3̄̆ 4̄̆ 5̆ 6̆ 7̄ 8̄̆ {33|34}


On account of the Rig-Vedic tendency to eliminate double-short sequences from the closing, however, the latter type is sporadic at best. [19]

Finally, consider this 9-syllable Greek meter with paroemiac closing:

1̄̆ 2̄ 3̆ 4̆ 5̄ 6̆ 7̆ 8̄ 9̄̆ = Enoplion


The Enoplion, however, is not an inherited dimeter but an internal development within Greek. [20]

I return to the basic problem of metrical primitivism vs. evolution. The point is, Rig-Vedic versification generally tolerates different patterns within a single flexible meter, while Greek Lyric generalizes the corresponding patterns as separate rigid meters. The closing of a Rig-Vedic meter like the Gāyatrī octosyllable tolerates sporadic instances of, say, – ⏑ ⏑ ⏓ besides the predominant pattern ⏑ – ⏑ ⏓. [21] Greek Lyric, on the other hand, restricts – ⏑ ⏑ ⏓ to the closing of choriambic dimeters and ⏑ – ⏑ ⏓, to that of Glyconics. Within the framework of two separate meters, the two divergent patterns can survive on an equal footing. Greek Glyconics are in turn separated from a third meter, the iambic dimeter, by further distinctions in rigidity:

1̄̆ 2̄̆ 3̄ 4̆ 5̆ 6̄ 7̆ 8̄̆ = Glyconic
(flexible at 1 2 8, rigid at 3 4 5 6 7)

1̄̆ 2̄ 3̆ 4̄ 5̄̆ 6̄ 7̆ 8̄̆ = iambic dimeter
(flexible at 1 5 8, rigid at 2 3 4 6 7) {34|35}


To repeat: Greek rigidity in meter is a more advanced phenomenon than Indic flexibility. Notice, however, that even the internal evidence of Greek Lyric reveals traces of a primitive phase parallel to the Rig-Vedic. For example, in the so-called polyschematist style of versification, Glyconics and choriambic dimeters still function as if they were variants of one meter, not as two separate meters. [22]

Extreme rhythmical regularity in the Indic verse is a sign of relatively late composition. [23] In the most archaic phases of Rig-Vedic composition, as I have already emphasized, sporadic rhythmical irregularity is sanctioned even in the closing of the verse. In the opening, moreover, the most archaic versification signals not just sporadic irregularity but absolute freedom in the rhythm, and this freedom is the rule rather than the exception. Besides internal evidence for the archaism of rhythmical freedom in the opening, there is also the external evidence of comparative metrics. Lack of a regular pattern of rhythm in the opening is a regular feature of the Slavic meters cognate with the Vedic, [24] as also of some Greek meters such as the choriambic {35|36} dimeter. [25] The comparative approach, in short, suggests that freedom in the rhythm of the opening is a feature inherited from the archetypal Indo-European poetic language. The Greek poetic evidence is valuable because it attests not only the pristine state of the opening but also the progressions away from this state. [26] These progressions, such as the one from irregular to regular Glyconic

⏓ ⏓ ⏓ ⏓ ⏑ – ⏑ ⏓
⏓ ⏓ – ⏑ ⏑ – ⏑ ⏓


involve the gradual restriction of the original freedom in the opening. The lineal direction is from line-final toward line-initial—a direction which we may describe pictorially as heading from ‘right’ to ‘left’. [27] Using such internal evidence in conjunction with the comparative approach, we may imagine three stages in the evolution of the opening: (1) absolute freedom from regular patterns of rhythm; (2) tendency of some patterns to outnumber others in frequency of occurrence; tendency diminishes from ‘right’ towards ‘left’; (3) regularization of such tendencies. [28]

Footnotes

[ back ] 1. See Meillet 1923, Jakobson 1952, Watkins 1963; for a convenient précis, see Schmitt 1967: 307-313.

[ back ] 2. For an application of these factors in linguistic reconstruction, see Householder and Nagy 1972:778-790 (= 1973:58-70).

[ back ] 3. See Nagy 1970 passim.

[ back ] 4. In Indic meter, catalexis may not be synchronic: see pp. 285-287.

[ back ] 5. I have chosen the term ‘closing’ in preference to ‘cadence’. Although the latter has been in vogue among students of Indic and Greek meter (e.g., Arnold 1905 and Dale 1969 respectively), it is subject to misinterpretation.

[ back ] 6. For specifics on these Greek and Indic octosyllables, see pp. 31, 35f, 37f.

[ back ] 7. Watkins 1963:203-206.

[ back ] 8. See Watkins 1963:206; also pp. 35f below.

[ back ] 9. For the analytical value of this distinction between constant and tendency, see Jakobson 1952 passim.

[ back ] 10. Cf. Watkins 1963:208.

[ back ] 11. For a convenient summary of statistics and other data, see Arnold 1905:152-160.

[ back ] 12. Cf. Watkins 1963 passim.

[ back ] 13. Meillet 1923:45f.

[ back ] 14. Cf. Arnold 1905:183-185.

[ back ] 15. Cf. Watkins 1963:200f.

[ back ] 16. Cf. Meillet 1923:46.

[ back ] 17. See pp. 30f, 37.

[ back ] 18. Jakobson 1952:63.

[ back ] 19. Cf. Watkins 1963:209f.

[ back ] 20. See p. 294.

[ back ] 21. For statistics and other data, see Arnold 1905:152-160.

[ back ] 22. See pp. 42f.

[ back ] 23. Cf. Arnold 1905 passim.

[ back ] 24. Jakobson 1952:63.

[ back ] 25. Watkins 1963:196.

[ back ] 26. Watkins 1963:206.

[ back ] 27. Cf. p. 30.

[ back ] 28. One such tendency will figure prominently in the succeeding discussions, Part II: the selection of long over short in syllable 4 of the Rig-Vedic octosyllable.