Abstract:

The rime cognate, which is a set of words containing the same phoneme sequence as the rime of a given word, was proposed as a new lexical competitor set for spoken word recognition. Partial correlation analyses showed that the rime cognate has a much larger contribution to the identification of English spoken words than the neighborhood and the cohort which had been commonly used as lexical competitor sets. The analyses also showed that only parts of the neighborhood and the cohort are reliable as lexical competitor sets in spoken word recognition, and that these parts are included in the rime cognate. This result indicates that the rime cognate integrates the two different competitor sets, the neighborhood, and the cohort. Based on these findings, the rime cognate model of spoken word recognition was proposed.