Other Indo European metre

We have now reconstructed the outlines of a Graeco-Aryan metrical system, characterized by quantitative prosody and lines of determinate length. There were shorter lines of seven or eight syllables, ending in the cadences ... u — - II or ... u — u— II, and longer lines made by prefixing these with a four- or five-syllable element, X u X u I or X u u u X I. Simple strophes were built, usually from three or four similar lines, but sometimes by alternating lines of different length. This summary account, it should be stressed, does not necessarily encapsulate the whole of Graeco-Aryan metrics, only those details that we are able to reconstruct from the extant evidence.

The next step is to inquire how much of this picture, if any, can be taken back to Level 2 (MIE) or Level 1 (PIE). To what extent can the same features be recognized, firstly in the other ancient European poetic traditions and secondly in the Anatolian?

Taking the European traditions in chronological order of attestation, we begin with the Italic evidence, which consists primarily of the Latin Saturnian metre and other odds and ends of early Latin verse; we ignore, of course, the classical Latin metres borrowed from Greek.74

The basic scheme of the Saturnian may be represented as

73 See L. H. Gray, Revue des Études Arméniennes 6 (1926), 160 f., 164-7.

74 Cf. T. Cole, 'The Saturnian Verse', Yale Classical Studies 21 (1969), 3-73; West (1973), 175-9; Watkins (1995), 126-34; P. M. Freeman, 'Saturnian Verse and Early Latin Poetics', JIES 26 (1998), 61-90.

xxxxix uxlxxxxx xii

—a seven-syllable colon, usually with caesura after the fourth syllable, in syzygy with a six-syllable colon of which the penultimate syllable is stressed (and most often long). There are many irregularities in the number of syllables, most of which can be explained by special rules or put down to faulty transmission, but this description may be taken as the basis for discussion.

In the extant material, the oldest of which dates from the third century bce, the accentual rules of classical Latin apply.75 At an earlier period the stress had fallen on the initial syllable of every word. At that time the caesura after four syllables would have entailed a stress on the fifth, and the verse would have had the form xxxxx xxixxxxx xii

The pattern is strikingly analogous to the type of acatalectic + catalectic pairing typical of Graeco-Aryan quantitative verse:

We have already suggested that in Avestan, with stress having come to be a significant factor in versification, the original cadence . .. u--ii was replaced by ... x x x ii. If a parallel development occurred in proto-Italic, the Saturnian can be satisfactorily explained as the continuation of a prototype of the form AG i AGA ii.

Another verse found in the early Latin material, in the Arval and Saliar hymns and in certain traditional charms, is an octosyllable with medial caesura and penultimate stress: diuom deo supplicate; nouom uetus uinum bibo. This cannot be equated with the ordinary Graeco-Aryan G verse, but it could correspond to the Vedic trochaic gayatri type.

Attempts have been made since the nineteenth century to identify and analyse verse in other Italic dialects such as Oscan, Paelignian, Faliscan, and South Picene. Some of these analyses appear wholly arbitrary; others seem possible but too uncertain to assist the argument, and I pass over them here.76

We move north to the Celtic lands. Here the oldest extant verses are perhaps to be found in a Gaulish curse-tablet from Chamalieres, dating from the first century of our era. The text begins: andedion uediiumi diiiuion risunartiu Mapon Arueriiatin, 'by virtue of the Lower Gods I invoke Mapon

75 Except that words scanning u u u X were presumably still stressed on the first syllable, as in Plautus and Terence.

76 See F. Bücheler, 'Altitalisches Weihgedicht', Rh. Mus. 33 (1878), 271-90; 'Altitalische Grabschrift', ibid. 35 (1880), 494 f.; P. Poccetti, 'Eine Spur des saturnischen Verses im Oskischen', Glotta 61 (1983), 207-17; H. Eichner, Die Sprache 34 (1988/90), 198-206; Watkins (1995), 126-34; P. M. Freeman, JIES 26 (1998), 77-9.

Arveriatis'. This certainly seems to show poetic diction and word order, and it can be arranged into three verses of seven or eight syllables with quantitative cadences in ... u — II or ... u-u- II.77 The similarity with the prototypes reconstructed for Graeco-Aryan is remarkable.

Irish verse does not come into view until half a millennium later, but the high proportion of words that are exclusively poetic is a sign of the antiquity of the tradition behind it.78 In this poetry an initial stress accent has replaced quantity as a formative factor. In the earliest epic and gnomic poetry the commonest type of verse is a heptasyllable of the form xxxxi;< x x II

This closely resembles the heptasyllable that forms the first half of the Latin Saturnian, and we can derive it in just the same way from the AG prototype.

Freely alternating with the heptasyllable, or sometimes as the fourth line of a quatrain, we find a verse that differs from it only in having one more syllable before the caesura. This would correspond to the full G. There is also a longer verse, apparently very archaic and soon abandoned, in which the hepta-syllable is preceded by a protasis of four or (less often) five syllables, marked off by caesura:

(x)xxxxlxxxxlx xxll

This is exactly parallel to the construction of longer verses in Graeco-Aryan by prefixing 4 I or 5 I to G, AG, etc.

If these Irish verses ending in ... I x x x II go back to acatalectic prototypes, we should expect that catalectic ones would appear with an ending ... I x x II, as they do in Latin and (without the consequential caesura) in Avestan. Such verses in fact occur. Indeed a complete series of 'acatalectic' and 'catalectic' measures can be found: 5 I 3 and 5 I 2 (= G: GA), 4 I 3 and 4 I 2 (AG: aGa), 3 I 3 and 3 I 2, 2 I 3 and 2 I 2. When members of these pairs are combined in a stanza, the 'acatalectic' ones precede. Thus Watkins cites stanzas of the forms 4 I 3 II 3 I 2 II 4 I 3 II 3 I 2 III, 3 I 3 II 3 I 2 II 3 I 2 II 3 I 2 III, 2 I 3 II 4 I 2 III, and 2 I 3 II 2 I 3 II 4 I 2 III. This is in accord with the predominantly clausular character of the catalectic type in Vedic and Greek.

77 W. Meid, Zur Lesung und Deutung gallischer Inschriften (Innsbruck 1989), ad fin.; id. (1990), 47 f. For the Chamalieres defixio see Lambert (2003), 152-61. In another major Gaulish text, the lead tablet from Larzac (Lambert, 162-74), G. Olmsted finds metre (JIES 17 (1989), 155 ff.; 19 (1991), 280-2), while Meid (1990), 48 finds 'eine teilweise rhythmisierte Prosa mit einem hohen Anteil an lautfigurativen Elementen'.

78 Thurneysen (1921), 56. In the account that follows I rely on Watkins's seminal paper (as above, n. 63), while noting that it has been subjected to some strong criticism: see E. Campanile, ZCP 37 (1979), 193-7; K. Klar, B. O Hehir, and E. Sweetser, Studia Celtica 18/19 (1983/4), 47-51; McCone (1990), 38-41, 45. Cf. also Kurylowicz (1973), 159-71.

In early British poetry the picture is not quite so sharp, but it fits sweetly enough into the same frame. It has long been a matter of controversy how its metrical practice should be understood, but according to an authoritative modern study syllable-counting gives better results than accentual analysis. The characteristic three-line stanza, the englyn, typically consists of seven-syllable lines; sometimes the first has eight syllables.79

Germanic verse does not contribute very much to the argument. We can certainly speak, on the basis of strong similarities between Norse, Old English, Old Saxon, and Old High German verse-forms, of a common Germanic tradition. There is a standard four-stress line that divides into two halves, linked by obligatory alliteration. The first half tends to be longer than the second, but the number of syllables is very variable. In Old Norse the verse is mainly used in four-line stanzas in what is called 'ancient style' (fornyrQislag). There is also a longer type of line with three stresses in each half, and in Norse and English there is a form of strophic composition in which a four-stress line alternates with a single three-stress colon.

It is not hard to suggest ways in which these measures might have evolved from quantitative prototypes such as we have postulated. The dimensions of the four-stress line would suit an origin from the combination of an eight-or seven-syllable verse (G or AG) with its catalectic counterpart. But here we are only observing that the data are compatible with our theory. We cannot claim that they corroborate it.

When it comes to Slavonic metre, we are dealing with material recorded in much more recent times. But the static nature of the forms as far back as they can be traced, together with their diffusion throughout the Slavonic lands, encourages the assumption that they represent a common heritage of considerable antiquity.80

Here again accentual developments in the languages have affected versification to a marked degree. However, the Serbo-Croat ten-syllable epic line preserves a recognizable quantitative cadence in performance, and more noticeably so in poems recorded in the eighteenth century. The underlying scheme is xxxxI xxuu-xll, with the principal stresses on the first, fifth, and ninth syllables. The ninth tends to be prolonged in recitation, and this feature is also recorded from Moravia and Bulgaria. The ten-syllable line occurs further in Slovakia, Poland, Belarus, and Ukraine.

80 Here and in what follows my information is drawn from Jakobson (as above, n. 63). See also Watkins (1963), 210-12 = (1994), 365-7; Gasparov (1996), 15-35.

The caesura after the fourth syllable suggests analogy with the Vedic, Avestan, Greek, and Irish long lines formed with a four-syllable protasis. What follows the caesura is an exact fit with our AGA prototype. That it is to be assigned to our catalectic category is confirmed by the existence in Russia of an acatalectic equivalent, an eleven-syllable verse, no longer with the caesura, but still with the ninth syllable stressed and prolonged. We may infer that proto-Slavonic had the two types, acatalectic xxxx I xxu u-u - II and catalectic xxxxI xxu u — II.

Another common Slavonic verse, used for epical-historical narratives, is of eight syllables, with stress on the sixth. By the principles applied above, this should correspond to the G prototype, x x x x u - u - II. If we take the Russian form of this verse, where the third syllable is also stressed, and apply the same method, we arrive at xx-uu-u - II, which is exactly the Greek glyconic and a frequent manifestation of the Vedic octosyllable. Outside Russia, however, the stresses in the first half of the verse fall rather on the first and fourth (Bulgaria) or the second and fourth syllables. There is also a caesura after the fifth.

Another kind of octosyllable, associated with laments, is characterized by trochaic rhythm and thus penultimate stress. This suggests the catalectic category, and again there is a Russian acatalectic equivalent with an extra syllable at the end to provide confirmation. There is also, in Russian and Serbo-Croat, a longer form of the line, with an extra four-syllable colon at the beginning—the now familiar protasis element.

The Lithuanian songs make use of cola of four, five, six, and seven syllables in various combinations.81 The penultimate syllable tends to be long and stressed, the distribution of other stresses being irregular. We find strophes consisting of two, three, or four heptasyllables; strophes of two or four lines of 5 I 5, or four of 5 I 4; of 5 I 5 I 7; of four of 4 I 6, or two of 4 I 4; of 4 I 4 I 6. In a folk-tale published by Schleicher a horse utters a series of octosyllables (4 I 4), concluding, as it were catalectically, with a heptasyllable.82 The 4 I 6 combination resembles the Slavonic decasyllable.

The Latvian folk singers use three types of metre. The most frequent is the 4 I 4 octosyllable. Its rhythm is strongly trochaic, as Latvian has an initial stress and monosyllables may not be placed in the fourth and eighth positions (which are normally short). Sometimes the hemistich takes a catalectic form, x x - I instead of x xx u I. Secondly, there is also a 3 I 3 verse with dactylic rhythm: x x u I x x u II, again with catalectic variants. Gasparov suggests that

81 Rhesa (1825), 334-47; P. Trost in Poetics. Poetyka. Poetika, i (Warsaw-The Hague 1961), 119-26; Kuryfowicz (1973), 200-10; Gasparov (1996), 13 f.

82 Noted and quoted by Meillet (1923), 77 f.

this may derive from a longer, twelve-syllable verse broken down into hemi-stichs. Thirdly, there are strophes in which the octosyllable alternates with a six-syllable line without caesura, the most popular arrangement being 6 II 6 II 8 II 6 III.83

Albanian oral poetry, according to a nineteenth-century scholar, was non-strophic and unrhymed, consisting of a series of loose octosyllables, among which seven- or six-syllable lines sometimes appeared. There is also epic verse in 4 I 6 decasyllables similar to those of the neighbouring Serbo-Croat tradition.84

Finally, I should not wish to be reproached for overlooking the Tocharians, who are believed to have migrated to their historical seat in Chinese Turkestan from the west and ultimately from Europe. Their verse appears to be purely syllabic.85 Cola of between three and eight syllables are combined in long lines. The commonest are of twelve syllables (4 I 4 I 4 or 5 I 7), fourteen (7 I 7), and eighteen (7 I 7 I 4). Less common are fifteen (5 I 7 I 3 or 7 I 8 or 8 I 7), seventeen (6 I 6 I 5), and twenty-five (5 I 5 I 8 I 7). Lines of equal or unequal length are grouped in four-line strophes.

These structures might be derived from the same set of prototypes as those discussed above; the prevalence of four-, five-, seven-, and eight-syllable cola, and the occurrence of such conjunctions as 5 I 7 and 8 I 7 among the rest, are suggestive. Watkins notes that the heptasyllable commonly has a caesura after four syllables. But one cannot build anything firm on these foundations.

If the Tocharian, Baltic, and Germanic evidence is inconclusive, the Italic, Celtic, and Slavonic traditions provide positive encouragement to think that the metrical principles extrapolated from Vedic and Greek were not valid only for Graeco-Aryan but, by and large, also for the rest of Europe; in other words, for MIE. It remains to ask whether indications of a similar system can be detected in Anatolia (apart from Phrygian, which does not belong to the Anatolian but to the Graeco-Aryan group). If they can, the inference will be that it can be attributed to PIE.

The usable material is disappointingly slight. We have a small number of extended mythological narratives in Hittite that are clearly poetic in nature, but they are translated or adapted from Hurrian originals, and it is not clear

83 V. J. Zeps in A Festschrift for Morris Halle (New York 1973), 207-11; Gasparov (1996), 11-13.

84 Alberto Stratico, Manuale di letteratura albanese (Milan 1896), 60; J. Kolsti in C. E. Gribble (ed.), Studies presented to Professor Roman Jakobson by his Students (Cambridge, Mass. 1968), 165-7. Examples of octosyllabic songs can be found in Auguste Dozon, Manuel de la langue chkipe ou albanaise (Paris 1878), 85 ff.

85 Emil Sieg and Wilhelm Siegling, Tocharische Sprachreste, i. A (Berlin 1921), x-xi. Cf. C. Watkins in H. Eichner and H. C. Luschutzky (edd.), Compositiones Indogermanicae in memoriam Jochem Schindler (Prague 1999), 601-14, especially 604 f.

whether the Hittite versions are themselves metrical. If they are, the versification seems to be based on a general balance between syntactic cola, as in Akkadian, Ugaritic, and Hebrew verse, and not on any measurement by syllables.86

For syllabic verse we must look elsewhere. Watkins has adduced a plausible example in an Old Hittite funeral song, perhaps as old as the seventeenth century bce, where three lines of 4 I 4 or 5 I 4 are each followed by a five-syllable refrain.87 He has also drawn attention to certain Luwian fragments, quoted in Hittite ritual texts, which certainly look like verse: they show signs of alliteration, rhyme, and poetic word-order. One is what he has sensationally interpreted as the first line of an epic lay about Troy:

ahh-ata-ta älati awenta Wilusati, rendered as 'When they came from steep Wilusa'. This can be seen as a two-colon line, 7 I 7 (or, assuming an elision, 6 I 7), with distraction of epithet and noun so as to give rhyming endings to the hemistichs. A parallel line in another fragment shows the words differently arranged so as to keep a similar length of cola: älati-ta ahha zitis awlta [?Wilusati], 'When the man came from steep [?Wilusa]'. Another piece divides easily into 7 I 7 I 7 I 11 (= 7 I 4), where the first two heptasyllables are the same and the third rhymes with them.88

From the fourth century bce we have a small body of verse inscriptions in Lydian, notable for their consistent use of a kind of rhyme: throughout each poem the last syllable of every line contains the same vowel. In one case the lines are arranged in three-line stanzas. The line is normally of twelve syllables, but sometimes eleven, with a regular caesura before the fifth or sixth syllable from the end. It can be analysed in terms of four trisyllabic 'feet', in each of which the last syllable is accented. An accented syllable may also stand in the first position in the line, less often in the fourth, and occasionally in the seventh.89 It is difficult to relate this pattern in any persuasive way to what we have found elsewhere. But the fact that the metre shows constraints based on the syllable-count and on syllabic weight does bring this Lydian verse into some connection with other Indo-European systems, even if the relationship cannot be more closely defined.

86 Such was the conclusion of H. G. Güterbock, who attempted a metrical transcription of the Song of Ullikummi: JCS 5 (1951), 141-4. Cf. I. McNeill, Anat. St. 13 (1963), 237-42; S. P. B. Durnford, ibid. 21 (1971), 69-75; West (1997), 103; H. C. Melchert in Mir Curad, 483-94.

87 KBo 3. 40 = BoTU 14a 13'-15'; H. Eichner, Denkschriften der Österreichischen Akademie 236 (1993), 100-6; Watkins (1994), 418 f.; (1995), 248. The nine-syllable lines can be reduced to eight syllables if kata arnut is read with crasis or elision, kat'arnut (Eichner, 104).

88 Watkins (1994), 714; (1995), 146 f., 150; for others see Eichner (as n. 87), 112 f.

89 M. L. West, Kadmos 11 (1972), 165-75; 13 (1975), 133-6; H. Eichner, ZVS 99 (1986), 203-19; id., Die Sprache 32 (1986), 7-21; id. (as n. 87), 114-27; cf. Gusmani (1975).

Together with the extremely scanty Hittite and Luwian evidence, this suggests that proto-Anatolian poets may have practised composition in verses or cola of commensurate length, including some of seven or eight syllables, and sometimes arranged them in three- or four-line strophes. We cannot establish whether there were definite quantitative patterns that they favoured, or if they were familar with the catalectic-acatalectic opposition. Proto-Anatolian, of course, is not the same as proto-Indo-European, and if its metre was indeed less finely chiselled than that of MIE, it remains an open question which of the two more faithfully reflects the PIE situation.

0 0

Post a comment