ISCA Archive PSP 2005
ISCA Archive PSP 2005

Acquiring novel words and their past tense forms: evidence from lexical effects on phonetic categorisation

Leanne Sedin, Gareth Gaskell

Here we examined plasticity in speech perception by testing whether familiarisation with a novel word can lead to a 'lexical' bias on phonetic categorisation (cf. Ganong, 1980). We also investigated whether this lexicalisation bias would generalise to an inflected form of the novel item. Three sets of matched /d/- final nonwords were used in the experiment, as follows:

(1) Familiarisation: Participants were exposed to novel items repeatedly in three conditions: (a) with the final /d/ (e.g., WECTADE), (b) without the final /d/ (e.g., WECTAY), or (c) not at all. Exposure tasks involved past tense formation (writing down the past tense of the novel item) and phoneme monitoring. Three item sets were used to ensure that each item was encountered in just one condition by each participant, but that all items were represented in all conditions across participants.

(2) Phonetic Categorisation: Participants made twoalternative forced-choice decisions to /d/-/t/ continua embedded word-finally in the nonwords. The three conditions determined by the item status at familiarisation were: (a) uninflected (e.g., WECTA[d/t] having learnt WECTADE); (b) inflected (WECTA[d/t] having learnt WECTAY); (c) nonword (WECTA[d/t] having never heard it before). Phonetic categorisation took place either immediately (40 participants) or one week following familiarisation (41 participants).

Lexicalisation biases were indicated by significantly more /d/-responses in both the uninflected and inflected conditions than in the nonword condition: participants familiarised with WECTADE or WECTAY both judged the ambiguous phoneme in WECTA[d/t] to be more /d/-like than those not familiarised with either. This effect was more pronounced in the uninflected condition, with significantly more /d/- responses given to WECTA[d/t] by participants exposed to WECTADE than those exposed to WECTAY. These effects were evident both immediately following familiarisation and a week later despite no further exposure to the novel items, and were taken as evidence for the immediate use of episodic representations of the novel words (cf. Gaskell & Dumay, 2003) in categorisation, and their subsequent lexicalisation. The lexicalisation bias in the inflected condition suggested that the past tense forms of newly acquired words were recognised as lexical items despite having never previously been encountered. These results are discussed with reference to current theories of lexicalisation, phonetic categorisation, and the lexical representation of inflectional morphology.

s Ganong, W.F. (1980). Phonetic categorization in auditory word perception. JEP: HPP, 6, 110-125. Gaskell, M.G., & Dumay, N. (2003). Lexical competition and the acquisition of novel words. Cognition, 89, 105-122.