Six-month olds detect a novel speech sound contrast more effectively from unfamiliar rhythm

0566073 - PSÚ 2023 ES eng A - Abstrakt
Chládková, Kateřina - Podlipský, V.J. - Paillereau, Nikola - Nudga, Natalia - Šimáčková, Š.
Six-month olds detect a novel speech sound contrast more effectively from unfamiliar rhythm.
WILD 2022. Book of Abstracts. San Sebastian: Basque Center on Cognition, Brain and Language, 2022. s. 114-114.
[Workshop on Infant Language Development (WILD) /5./. 09.06.2022-11.06.2022, San Sebastian]
Institucionální podpora: RVO:68081740
Obor OECD: Psychology (including human - machine relations)
https://www.bcbl.eu/events/wild/en/

Three-month-olds recognize atypical realizations of native-language speech [Kitamura et al., 2006], 10-month-olds prefer engaging with native-accented talkers [Kinzler et al., 2007], while 17-month-olds learn new words selectively better from native talkers [Mattock et al., 2010, Fennel & Byers-Heinlein, 2014]. At the same time though, at about the age of 6 months, infants preferentially listen to non-native accents [Butler et al., 2011, Paillereau et al., 2021]. We ask whether, like words, also phonetic categories are learned better from native talkers, or alternatively whether they are learned better from novel, nonnative accents to which young infants attend preferentially. To discover the segmental categories of their language, infants exploit the sounds' distributional statistics [Maye et al., 2002, Wanrooij et al., 2014]. Here we tested whether infants detect a novel phonetic contrast with differing success from native-accented than from nonnative-accented speech. We recorded utterances spoken with typical and atypical rhythm of the infants' native language (Czech) and delexicalised them by replacing consonantal intervals with [f] and vocalic intervals with one of 406 tokens sampled from a bimodal distribution between [ɛ] and [æ]. Six-and-half-month old Czech infants were first exposed to either the native- or nonnative-accented delexicalized material and subsequently tested with non-alternating trials of [fɛ] and alternating trials of [fɛ]-[fæ] (counterbalanced). Data from 34 infants (collection ongoing) indicate an interaction of familiarization condition and test type, whereby only infants familiarized with the [ɛ]-[æ] contrast in atypical rhythm looked longer to non-alternating than to alternating test trials (atypical fam: non-alt = 8.55s, alt = 6.47s, p = 0.064, typical fam: p = 0.737). In sum, 6-month old infants seem to more effectively exploit distributional information from speech with atypical, non-native, rhythm. This is in line with their previously reported preferences for listening to novel over familiar accents.
Trvalý link: https://hdl.handle.net/11104/0337512