________________
VARIANT ENDINGS -U, -AÜ AND -A IN THE APABHRAMŚA VERSES
25
one from the other within Apabhramśa is not so obvious as has thus far been assumed.
Acknowledgements
I am very grateful to Professor O. von Hinüber for his comments on an earlier version of this article.
NOTES
1 The extended and the long endings appeared to be typical of the nom./acc. sing., with altogether 801 instances of -aü and -aü (no count of the short ending -u is provided); of the loc. sing., with 108 instances of -ae (instead of -e); and of the nom./acc. plur., with 61 instances of -ā (and 3 of -aya, instead of -a). The figures for the extended endings of the instr. -aena (19) and for the gen. -ayaholaho (3) are low in comparison. 2 -aü occurs 608 times after a heavy syllable, 193 times after a light syllable, -ae 84 times after a heavy, 24 times after a light syllable, and -a is found only after heavy syllables. In this connection Alsdorf notes that the exceptions to this rule seem in particular to involve the ordinalia: pahilae, taiyae, etc. S. 25 times after a double consonant, 38 times after a syllable containing a long vowel. 4 Alsdorf adds the suggestion that this extension of the use of the long and extended endings irrespective of the quantity of the preceding syllable might be an indication that Hemacandra's verses belong to a more advanced linguistic stage than the Harivarsapuräna. $ It is as such that Alsdorf's rule is generally quoted. Thus, Baumann notes that in the three Apabhramsa poems investigated by him in word-final position often a long vowel appears. He classifies the instances according to the weight of the preceding syllable: 48 times after a heavy syllable, 48 (sic) times after a light one, and, in addition, 41 times with monosyllabic words (mainly pronouns) (Drei Jaina-Gedichte in Alt-Gujarātī, VI.D.7. (p. 33). See, for instance, also VII.B (p. 48), where Baumann notes that 6 instances of the "extended” ending -ī of adjectives are found after heavy syllables, and the four remaining instances after light syllables. He does not elaborate on the fact that his findings do not agree with Alsdorf's observation concerning the Harivarśapurana, nor, what may after all be more important, does he discuss the possible restriction of the occurrence of the long and extended endings with only a few specific categories of words (see below, $ 8).
See also Bhayani in § 53 of the "Critical Study" preceding his edition of the Sardeśa Rāsaka: he starts with referring to Alsdorf's observation that the enlarged endings would occur mainly after heavy syllables. • Contrast in this respect the rhythmic law proposed by Insler for Pali, which, as may be expected of a rhythmic law, effected in principle all words in that language (“Rhythmic Effects in Pali Morphology", Die Sprache 36 (1995), pp. 70-93). It is interesting to note that the circumstances under which this rhythmic law in Pali took effect are the very opposite to the one assumed by Alsdorf for Apabhramsa: in Pali a final long vowel would be shortened in polysyllabic words if the penultimate syllable was long, but remained in disyllabic words and in polysyllabic words that possessed a short penultimate syllable.
The majority of verses are Dohas, for which see Alsdorf, Apa-Studien, pp. 39–44.