Representational specificity of lexical form in the perception of spoken words

Abstract

Exemplar and traditional abstractionist models of lexical representation were evaluated by examining flapping in American-English. According to abstractionist models, flaps are considered neutralized surface realizations of the underlying phonemic forms /t/ and /d/, with fully specified /t/ and /d/ forms of words represented in memory. Exemplar models, on the other hand, assume flapped tokens of spoken words have full representational status. An investigation of the representation of flaps in American-English was conducted using the repetition priming paradigm. Participants shadowed spoken words in two blocks of trials. Stimuli in the first block were primes and those in the second block targets. Primes and targets consisted of flapped and carefully articulated bisyllabic words. Time to shadow target words in the second block was measured as a function of prime type. The shadowing time results provide evidence for underlying abstract representations. The implications of these results for the theories of form-based lexical representations will be discussed.