Poster #35 - Correlations and Mismatches in Parents' and Their Children's Descriptive Speech

• In Event: Poster Session 20

Abstract

Although a good deal of attention has been paid to the ways that nouns (e.g., Bergelson et al., 2019) and verbs (e.g., Nomikou et al., 2017) have been used in speech to children, less research attention has been directed towards how parents use descriptives in speech and the relation to children’s production. Further, the factors that predict how descriptives are used in speech is understudied. The current project investigates correlations and mismatches between the features parents talk to their children about and the features children subsequently talk about. We analyzed labeling and four types of descriptives: perceptual, conceptual, relational, introspective.

Participants were children (n = 32; 55.09 months; 16 females) and their parents from the CHILDES database (CHILDES: MacWhinney, 2000; Individual Differences Corpus: Gelman et al., 2014). Parents read a wordless picture book to their child. Next, children were asked to read a second wordless picture book, with new pictures, to an alien without knowledge of the pictures.

Figures 1 and 2 show the mean differences between the four descriptive types in children’s and parent’s speech. Parents used significantly less relational and conceptual descriptives than perceptual and introspective descriptives. Children’s speech however, contained significantly fewer relational descriptives, but no other differences.

We next examined the relationship between parents’ and children’s use of descriptives in speech. Importantly, we found a significant correlation (r (21) = 0.582, p = 0.023) between the proportion of relational features in parents’ utterances and the proportion of relational features in children’s utterances. This result fits with previous work showing that hearing an experimenter stating the relationships between objects would elicit a relational mindset in children, signaling that relational attention is malleable in children (Simms & Richland, 2019). Additionally, there was a significant correlation (r (21) = 0.467, p = 0.039) between the proportion of main referents’ labels in parents’ utterances and the proportion of main referents’ labels in children’s utterances. This suggests that parents’ labeling may affect children’s labeling behavior more broadly.

We also asked whether the types of descriptives varied across the kinds of pictures depicted. We found children talked about a higher proportion of perceptual features when talking about animals than foods (t(21)=2.550, p=0.019) and a higher proportion of conceptual features for foods than for animals(t(21)=3.172, p=0.005). However, these differences were not found in parents’ utterances. However, there were significant correlations found between the proportions of perceptual (t(21)=2.162, p=0.043), introspective (t(21)=2.880, p=0.009), taxonomic features (t(21)=2.743, p=0.013) in how parents’ and children’s utterances talked about foods, but not for other kinds of references, suggesting that foods may engender shared descriptives between parents and children.

Altogether the results suggest that children are using multiple types of descriptives in speech and that the types of descriptives children use are related to how parents use descriptives in speech and the types of pictures being viewed. Additional research will seek to examine the specific descriptives in children's and parent’s speech at a more granular level.

Authors

• Erjing Zhang, University of California - Los Angeles
  Presenting Author

• Catherine Sandhofer, University of California - Los Angeles