In Russian the normative perfective verbal adverb (PVA) suffix is –(v)ši. However, there exist variants such as –mši, -lši, and -tši/-tči (ex. Inf. est’ “to eat” PVA. evši vs. emši, elši, etči). As illustrated in Avanesov and Bromlej (1989), while the normative variant –vši is used throughout the entire territory of Russia, -tši/-tči is found in the north, -lši in the west-central zone, and –mši in the south. In this paper, the variation of the first consonantal portion of the PVA suffix will be examined both in terms of morphophonemic paradigm and syntagmatic context defined in each geographical zone where a specific type of reflex occurs.
First, I explore the source of the variants and the innovations that produced them. Following Obnorskij (1953), I argue that the variants –mši and -tši/-tči occur from Rusian reflexes such as vzem-ši “having taken” (LCS. *vъzьm-) and šed-ši “having gone” (LCS. *šьd-) respectively. The allomorphs are yielded by stem innovation and metaanalysis of morpheme boundary in the process of generation of the normative variant –vši. In this process, one innovation conditions another, following a specific hierarchy. I consider that the variant –lši resulted from morphophonemic reanalysis of the first consonantal portion of the PVA suffix, conditioned by the identical phonetic value of v and l before an obstruent in the area where –lši occurs.
Subsequently, I investigate what motivates those innovations paying attention to the syntactic function that the PVA assumes in dialects. I agree with Flier (1981) in that the innovations and reanalysis are motivated by approximation between the PVA and Past forms due to their functional similarity as predicates. The morphological approximation is most obvious in the PVA forms in –lši in the west-central zone. In that area, a PVA form with or without an auxiliary verb is used as a perfect predicate, which is the function that the old resultative l-participle used to assume.
To consider the given phenomena in more general perspectives, I attempt to explain what factors in particular patterns of the variation. The query on the distribution of the variants is addressed in terms of phonemic value of each consonantal variant, the hierarchical ranking of relevant features, and the role of systemic context across East Slavic territory. I suggest that the distribution of different phonetic realizations of the first consonantal portion of the suffix (v, m, l, t) is conditioned by different phonological systems found in the relevant dialects: obstruent vs. sonorant languages (against “consonantal vs. vocalic languages” suggested by Andersen (1978)) and tense/lax languages vs. voiced/voiceless languages.
TT The Peter de Ridder Press, pp. 1-12.
Flier, M. S. 1981. “The Morphology of the Russian Past Active Participle”, IJSLP 27, s’Gravenhage, pp.79-106.
Obnorskij, S. P. 1953. Ocherki po Morfologii Russkogo Glagola, Moscow: Nauka.