Categorical effects in production of

Previous work suggests that methods used in studying categorical perception for segmental contrasts may be useful in determining the representation of suprasegmental contrasts, particularly those based on intonational characteristics. This experiment investigated phonological representations for intonation contours in English by determining the extent of categorical effects in a task involving imitation of synthetic speech continua in which a fundamental frequency maximum (peak) or minimum (valley) was shifted in 25 ms increments through a target two-syllable sequence with a particular stress pattern (WS or SW). Imitations of both WS and SW peak continua showed categorical effects consistent with two categories of representation, while imitations of valley continua were inconclusive. Results are interpreted with respect to phonological categories of representation that have been proposed for English intonation.