Predicting Non-Deterministic Language Variation

Abstract

Linguists have usually tried to analyze language deterministically. Cases of variation have long been recognized, but until the 1960s these cases were vaguely identified as “free variation”, a term which essentially meant that the behavior was non-deterministic. Labov and other sociolinguists have discovered many complex examples of language variation and have clearly demonstrated that language variation is not at all “free”, that many factors — both linguistic and social — affect the probability of a variant occurring. Moreover, the probabilistic behavior is not only found at the group level, but also in individual speakers. And no matter how many variables are considered, the probabilistic behavior cannot be reduced to deterministic explanation. For a good example of this, see Guy’s discussion of final /t,d/ deletion in English for which “variation is inherent, and cannot be scrubbed out of our linguistic description by ever-finer subdivisions of the data” (Guy 1980:11).