Coverbal Speech Gestures Do Not Impact Preschoolers' Ability to Use Prosodic Information to Constrain Parsing
Résumé
Young children exploit prosodic boundaries in speech to find syntactic boundaries and constrain parsing. For instance, they interpret the word "flies'' as a noun in the sentence "[Do you see the baby flies?]" but as a verb in the sentence "[Do you see?][The baby][flies!]". Recent studies demonstrated that the speech signal also contains coverbal visual information, such as eyebrow movements and head nods, which aligns with prosodic boundaries in speech. However, despite evidence of infants and adults integrating auditory and visual information in language processing, it remains unclear whether coverbal visual information could impact listeners' ability to use prosody to constrain syntactic analysis. To investigate this question, we tested whether children’s ability to use prosodic information to constrain parsing would be affected by the presence of both auditory and visual cues, compared to purely auditory information.
French-speaking 4-to-6-year-olds (N=48) participated in a sentence-completion task (Fig.1), testing their ability to use phrasal prosody to constrain analysis of sentences containing noun/verb homophones (e.g., [la petite][ferme la fenêtre] (the little girl closes the window) versus [la petite ferme][est très jolie] (the little farm is very nice); brackets indicate prosodic boundaries). Although both sentences start with the same three words, they can be disambiguated by the different prosodic boundaries. Crucially, all words following the homophone were masked, such that prosodic cues were the only disambiguating information. Participants heard these three-word sentence onsets either through a video of the speaker producing the sentence (audio-visual modality) or while seeing an image of a radio (audio-only modality; N=24 in each modality) and were asked to complete the sentences. Participants’ sentence completions were classified as either noun completions, if they used the homophone as a noun (e.g., “the little farm… is big”; ferme=noun), or verb completions, if they used it as a verb (e.g., “the little girl closes… the door”; ferme=verb). The results revealed that in both the audio-visual and audio-only modalities, children successfully exploited the sentence’s prosodic structure to assign the appropriate syntactic category to the target word. They provided more noun completions when listening to sentences in the noun-prosody condition (M=.77) than in the verb-prosody condition (M=.22; p<.001). However, there was no significant difference in performance between the audio-only and audio-visual modalities, and no interaction between the effect of prosodic condition and modality (p>.1).
The results show that the presence of coverbal speech gestures signaling phrase boundaries do not impact children’s ability to use prosodic information to constrain syntactic analysis. Children performed equally well whether or not they had visual cues to help them. This suggests that although the speech signal contains multimodal information, such as eyebrow movements and head nods that align with prosodic boundaries, this information does not help preschoolers chunk phrases from the input when robust auditory cues are already present. Our study contributes to the literature on the role of multimodal cues in syntactic analysis, investigating how preschoolers use simultaneous visual and acoustic cues to constrain parsing.
Origine | Fichiers produits par l'(les) auteur(s) |
---|---|
licence |
Copyright (Tous droits réservés)
|