Speech sounds in larger inventories are not (necessarily) less variable

Abstract

This paper reconsiders a classic claim about phonetic variability-that speech sounds in larger phonemic inventories should exhibit less within-category variability in production. Although this hypothesis is intuitive, existing literature provides limited unqualified support for the claim, further complicated by the fact that null results (like those failing to find a difference in variability between languages) often go unpublished. Even so, existing work suggests that factors contributing to extent of variability are multifaceted. While phonological contrast may affect variability patterns, inventory size alone is not a reliable predictor of variability differences. This paper reviews relevant findings in the literature, presents an additional case study, and argues for more nuanced alternatives to account for cross-linguistic differences in extent of phonetic variability.