How do words compete? Quantifying lexical competition with acoustic distance

Research on speech perception and lexical access often uses the activation and competition metaphor to describe the process of spoken word recognition. One way of expressing competition associated with a given word is its phonological neighborhood density, which is a calculation of similarity. The present study uses acoustic distance as an alternative to phonological neighborhood density to measure lexical competition during speech perception. The quantification of a word's lexical competition is given by what is termed its acoustic distinctiveness, which is taken as its average acoustic distance to all other words in the lexicon. A variety of possible abstract acoustic representations for items in the lexicon are analyzed. Statistical modeling shows that acoustic distinctiveness has a similar effect as phonological neighborhood density. Additionally, acoustic distinctiveness consistently increases model fitness more than phonological neighborhood density, regardless of the abstract representation used. Acoustic distinctiveness, however, does not explain all the same things as phonological neighborhood density. Potential theoretical implications of acoustic distinctiveness's usefulness in statistical and psycholinguistic models are discussed.