On Verbal Accentuation in the Rigveda.

• Author: Hoenigswald, Henry M.

Klein wishes to refine certain observations we owe to Berthold Delbruck (1888), Hermann Oldenberg (1906), George Dunkel (1979, 1988), Heinrich Hettrich (1988), and others. He summarizes his views conveniently in section 14. Verbs are accented (a) in pada or sentence-initial position, (b) before id, (c) "intonationally," (d) in subordinate clauses, and (e) in clauses following sentence-initial eta (eto). Characteristically, he asks whether accented status of verb forms can be reduced "to one or a few underlying functions" (p. 85). Cases of (a), (b), (d), and (e) are predictable; the fact that here the verb is accented (though of course not the place of the accent within the word) can be told from the environment. K. thinks that two of these associations of accented status with position or co-occurrence - namely, (a) and (b) - have one common meaning, 'this is important' or 'emphasis'. In (c), accented forms contrast with otherwise identical non-accented ones, and accented status is meaningful per se ("the utterance is incomplete"), except that the feature is (freely) "variable, and one can frequently find counterexamples under identical conditions" (p. 87f.). By definition, subordination (d), whether by means of conjunction or pronoun, or simply in bipartite construction, carries with it the same incompleteness. The three eta- (eto-) clauses (e) are said to belong here, too; the author regards them as postposed clauses of purpose (pp. 74-76) which are introduced by a semi-fossilized former imperative.(1) 'Reduction', then, has done its duty; it leaves us with two homonymous but separate meanings: emphasis and incompleteness.

Given the defining circumstances, the main interest resides in (c). Here, Oldenberg distinguished between mere juxtaposition and (within the same pada) 'tension'. In the tension category K. finds no absolute criterion that selects accented over unaccented verbs, and he falls back on what he considers a significant statistical preponderance: only six out of 18 accented cases (= 33%) are not minimal but "show contiguous anaphoric structures which may or may not belong to the same clause" while as many as 17 out of 33 unaccented cases (= 50%) do the same (p. 11):

... the more rigid and isolated the structure, the more likely the accentuation is to occur. Interpolations and adjacent expansions would appear to disfavor verbal accent without, however, obviating it.

Only where the verb stands before rather than after a noun or...