Parental linguistic alignment to their children facilitates language acquisition, with syntactic alignment being particularly relevant for children with ASD

Research output: Contribution to conferenceConference abstract for conferenceResearchpeer-review

  • Riccardo Fusaroli
  • Ethan Weed
  • Deborah Fein, University of Connecticut, United States
  • Letitia Naigles, University of Connecticut, University of Connecticut, United States
Background: Parent language is a crucial factor in language acquisition in children with ASD and TD children alike, but its role as a scaffold for the child’s linguistic production on a turn by turn basis is under-investigated. Here we focus on interactive alignment, that is, the tendency to re-use the child’s lexical (“mommy, give me the giraffe”, “is that a giraffe?”), syntactic (“the giraffe has fallen”, “yes and the elephant has gone down to help her”) and semantic (topic maintenance) choices. Besides supporting common ground and rapport, interactive alignment might facilitate language development by providing a rewarding parental response to children’s productions, as well as contextualized linguistic input via elaboration.

Objectives: We investigate whether i) parental alignment to their children might affect language development beyond more traditional measures of parental input (total and unique words, Mean Length of Utterance, or MLU), and ii) this is modulated by the cognitive and clinical features of the children.

Methods: We analyzed spontaneous speech in 67 parent-child dyads from a longitudinal corpus (6 visits over 2 years), consisting of 30 minutes of play activities. We included 32 children diagnosed with ASD (mean age at recruitment: 32.76m) and 35 linguistically matched TD children (mean age at recruitment: 20.27m). Lexical alignment was based on lemmatized words, syntactic alignment on 2-grams of part-of-speech-tags, and semantic alignment on Word2Vec representations of the corpus (Duran et al 2019). Alignment was calculated as cosine similarity between successive conversational turns (parent following child). To assess the effects of alignment at visit N on language learning at visit N+1, we relied on the best multilevel Poisson (total and unique words) and loglinear (MLU) models from our previous study on the role of parental input (Fusaroli et al 2019), and evaluated by Leave-One-Out-Information-Criteria whether including parental alignment improved these models.

Results: The full findings are presented in the accompanying table. The child’s later MLU, unique and total words were positively affected by lexical and syntactic alignment. Parents' semantic alignment positively affected total amount of words, but negatively the number of unique words. Further, the effects of alignment were modulated by the children’s features, in that children with lower previous production showed stronger effects of parental alignment on their own word use, compared with children with higher scores. Diagnosis provided independent modulation, in that children with ASD showed weaker effects on their own word use of parental lexical alignment, but stronger effects of syntactic alignment, than TD children. Finally, regardless of diagnosis, children exposed to less complex parental input (lower MLU) seemed to be more affected by the levels of parental lexical and syntactic alignment in their future word use, than those exposed to higher MLU.

Conclusions: Parental alignment differentially affected child language development. Parental lexical and syntactic alignment had positive effects on children’s word types, tokens, and MLU; however, child features only modulated word use. Specifically, children with poorer production and less complex linguistic input benefited more from parental alignment of their own speech, and children with ASD in particular benefited from higher syntactic alignment.
Original languageEnglish
Publication year2019
Publication statusPublished - 2019
EventINSAR 2019 - Montreal, Canada
Duration: 1 May 20194 May 2019


ConferenceINSAR 2019
Internet address

See relations at Aarhus University Citationformats

ID: 145108743