Past research has demonstrated that English-learning infants begin segmenting clauses from speech by six months of age (Nazzi et al., 2000). Clause boundaries are marked by many different acoustic cues, such as, pause, pitch reset, and final vowel lengthening. In the current study, we take a first step towards determining what acoustic information English-learners rely on most heavily to extract clauses from speech. More specifically, we ask whether 6-month-olds can extract clauses from speech when pausal cues to clause boundaries are removed. In addition, we investigate this same issue in Dutch-learners. By testing two different language-learning populations with similar stimuli, we are able to examine the possibility that language experience impacts clause segmentation strategies.
In Experiment 1, the Headturn Preference Procedure was used to familiarize infants to the same string of words produced in two ways: as a prosodically well-formed clause (e.g. Rabbits eat leafy vegetables) and as a prosodically ill-formed clause (... rabbits eat. Leafy vegetables....). In Experiment 2, Dutch learning infants were familiarized to similar word strings produced in two ways: a prosodically well-formed clause (Koude pizza smaakt niet zo goed 'Cold pizza doesn't taste very nice') and a prosodically ill-formed clause (...Koude pizza. Smaakt niet zo goed... 'Cold pizza. Doesn't taste very nice'). In Experiment 1 English-learners preferred to listen to a passage containing this string of words produced as a prosodically well-formed clause over a passage containing this string of words produced as a prosodically ill-formed clause. In Experiment 2, similar results were obtained with Dutch-learners. Thus, we replicated findings showing that English-learning 6- month-olds segment clauses from speech, and we also extended these findings to Dutch-learners.
In Experiments 3 and 4, we used the exact same stimuli and procedure used in Experiments 1 and 2, however the pauses associated with clause boundaries were removed. In this case, English- but not Dutch-learners continued to segment and remember clausal units in their native language better than non-clausal units. This suggests that Dutch-learners may be more reliant on pausal cues than English-learners. In a follow-up study, we are testing 6 month-old English-learners on the Dutch stimuli used in Experiments 2 and 4. If language experience shapes segmentation strategies, then English-learners may impose English-like segmentation strategies on the Dutch stimuli. In other words, in contrast to the Dutch-learning infants, the English-learners may extract the Dutch clauses even when pause cues to clause boundaries are absent.