Russian aspectual prefixes O, OB and OBO : a case study of allomorphy

Abstract

The thesis presents a study of the nontrivial interface of morphology, phonology and semantics found in the distribution of three Russian aspectual prefixes O, OB and OBO. These prefixes can be semantically identical when occur in the forms of the same paradigm (e.g. obo-drat’.INF. vs. ob-deru.1 PERSON.SG.FUT. ‘flay’), but can also carry strikingly different meanings that even yield minimal pairs (e.g. o-sudit’ ‘condemn’ vs. ob-sudit’ ‘discuss’). There are some phonological restrictions on their use but they tolerate a lot of variation in the choice of the prefix. Thus, the behavior of O, OB and OBO does not completely satisfy either of the two crucial criteria of regular allomorphy: first, their distribution is not precisely complementary; secondly, their semantics can be both identical and strikingly different. In order to account for this phenomenon, I apply corpus, experimental and statistical methods and address two major questions: (1) whether these prefixes constitute two separate morphemes, as suggested by the Split Hypothesis, or one morpheme with three allomorphs; and (2) whether these prefixes in Natural Perfectives (Janda 2007b) are pure aspectual markers with no semantic content (Švedova et al. 1980: 583, Zaliznjak & Šmelev 2000: 82) or their semantic contribution is just not perceptible due to conceptual overlap with the meaning of the simplex base (the Overlap Hypothesis proposed in Vey 1952; Schooneveld 1958; Janda & Nesset forthcoming a). First, I show that semantically “empty” and ”non-empty” uses of the prefixes O, OB and OBO exhibit isomorphic relations, which support the Overlap Hypothesis. Second, I demonstrate that the Split Hypothesis (Alexeeva 1978; Andrews 1984; Krongauz 1998) fails to account for the large overlap and variation in the semantic and phonological domains that it proposes. I provide a semantic analysis that shows that the meanings which might seem so unrelated are actually parts of a single semantic network and that all submeanings of this polysemy can be expressed by each of the three prefixes. The impact and statistical significance of various factors in the choice of the prefix are tested against a) corpus data and b) mechanisms of word-production examined in a psycholinguistic experiment with nonce words. On the basis of my results, I argue for the alternative view that treats O, OB and OBO as one morpheme with a non-complementary but at the same time statistically significant distribution of allomorphs. This suggests that the traditional understanding of allomorphy is too narrow and should be revised according to the gradient and complex nature of this linguistic phenomenon.