Separate morphemes or allomorphs? Morphological status of the Russian prefixes o- and ob- in the light of the new data: Corpus and experiment.

This article examines an interesting phenomenon in Russian morphology: the behavior of the prefi xes o- and ob-. Do these prefi xes constitute two distinct morphemes or are they allomorphs of a single morpheme? Previous discussion in scholarly works has been limited to lexical data provided in dictionaries. A number of controversial uses of these prefi xes were discovered and led to the hypothesis that the original single morpheme has split into two distinct morphemes. This article summarizes the results of two empirical tests of this hypothesis: fi rst, against a comprehensive dataset extracted from the Russian national corpus; secondly, via a psycholinguistic experiment. The results of both studies do not fully match the predictions of the split hypothesis. I offer a detailed analysis of the radial network of submeanings expressed by the two prefi xes, an independent account for the subject responses to experimental stimuli, as well as statistically robust analyses of both studies. On this basis I propose an alternative hypothesis of a single morpheme with non-standard allomorphy and argue for a unifi ed account of o- and ob-. The study of this particular problem of Russian morphology encourages us to reevaluate the traditional understanding of allomorphy coined in absolute terms and to revisit a number of forgotten ideas of American structuralists about the inadequacy of the binary (+/-) representation of linguistic units.