The Russian system of stress

Abstract

Russian stress is a puzzle. Numerous investigators have tried their hands at unraveling it, starting around the beginning of the last century. Studies carried out in the context of a general inquiry into the history of Slavic accentuation have predominated, and the purely synchronic focus on Russian stress is largely a product of the last forty years. Reliable monographic descriptions of the patterns of Russian stress now exist (e.g. , Red'kin 1971 , Stricek 1966, Fedjanina 1982), as do historical treatments based on the (re)examination of Old and Middle Russian accented texts (e.g., Kolesov 1972, Xazagerov 1973, Zaliznjak 1985). Despite this progress, with its main emphasis on descriptive refinement and comprehensiveness, the basic puzzle remains unsolved. Why does accentual mobility in declension and conjugation continue to exist and to thrive in certain word classes? Whereas the Common Slavic patrimony, with its phonologically based patterns of accentual alternations, can be looked to as a remote source of the persistence of mobility throughout the history of Russian despite the early loss of tone and length distinctions, what is to explain the pervasive patterns of stress "shifts" in the contemporary language? All sorts of ingenious answers to these questions have been devised by modern investigators, particularly in the wake of transformationalgenerative grammar, but on balance the functional system of Russian stress still awaits explanation. Of course, certain facts are clear. First, Russian stress is part of the morphology and morphophonemics of Russian and is basically not motivated by the phonology (while having phonological consequences, namely vowel reduction).1 Second, though stress is in principle free to occur on any syllable of a word, certain restrictions are known to obtain whereby the morphophonemic or morphological structure of a given word, or its membership in a particular form class or semantic category, narrow the range of possible accent positions. Third, not all accentual phenomena are equally motivated: in quite a number of cases (including whole classes of words) there is no motivation in terms of the contemporary system of Russian stress. Items of this sort have to be learned by speakers individually as "traditional" holdovers from an earlier system. They are, of course, also prime candidates for change. So much is common knowledge. A major advance beyond this stage of