Investigating the role of input in early word learning: a comparison of Dutch and Chinese infant-directed speech

Child-directed speech (CDS) has distinctive prosody compared to adult-directed speech (ADS) and it is found to facilitate word learning (Ma, Houston & HirshPasek, 2011). However, few studies have investigated the prosody of new words and utterances introducing new words. Additionally, the crosslinguistic differences in CDS haven’t received enough attention. Specifically, tonal languages (e.g., Chinese) employ F0 to differentiate lexical meanings, whereas nontonal languages do not. It can be predicted that the prosodic marking of new words, especially F0, may differ across languages. This project set out to investigate the input in word learning situations by comparing the prosody of Dutch and Chinese CDS. Dutch mothers of 18montholds (N=5) and 24montholds (N=5) and Chinese mothers of 18montholds (N=5) were instructed to read the same picture book consisting of 5 target words (bisyllabic) to their children and to adults. The Dutch target words consisted of three trochaic words and two iambic words, while the Chinese words included four tones and a neutral tone. The target words and the utterances embedding them (N=394) were analysed using Praat. Duration and F0-related prosodic cues were measured. The preliminary results showed that (1) there were significant differences between ADS and CDS in each language in most measures but not in F0 range of utterance, or duration and F0 range of target words, suggesting that even though CDS has distinctive prosody compared to ADS, it’s not the case in all measures during word learning. (2) The variances were unequal between Dutch and Chinese CDS in min F0 of utterance (p=0.001), and all F0 measures of target words (p=0.001), indicating that Chinese CDS showed more variance compared to Dutch. Further analysis includes age effect and the phonetic information of tones. Based on these results, a follow-up study will test children’s word learning when different prosodic cues are manipulated.