Frequency, Regularity, and the Paradigm: A Perspective from Russian on a Complex Relation

Abstract

The correspondence between irregularity and high frequency is well known (Bybee 1995, Greenberg 1966). What is not always clear is whether the frequency envisaged is based on the lexeme and all its manifestations, including the irregular word form(s), or just the irregular form(s) alone. For example in the case of English went, is it this single word-form that is highly frequent compared to other past tense forms, or is it the lexeme go that is highly frequent, in both its regular and irregular manifestations, compared to other lexemes? Bybee (1985:120) suggests the following: the correlation of irregularity with frequency occurs on two dimensions. The first is the lexical dimension ... where irregularity correlates with frequent lexical entries. The second is within the paradigm. To investigate further we have examined frequencies of noun lexemes, and their word forms, in a one million word Russian corpus (the Uppsala corpus), together with information on regularity. So as to more finely locate any correspondence between frequency and regularity, the types of irregularity we considered range from full suppletion to minor irregularity in stress.