Poster LCICD LA.pdf (1.42 MB)

Individual differences in early word learning: The effects of category curiosity and density

Download (1.42 MB)
posted on 10.09.2018, 12:56 by Lena AckermannLena Ackermann, Robert Hepach, Nivedita Mani

Children show considerable individual differences in their early vocabularies. While these differences might in part be driven by the input, the child herself should also be considered as a source of variability. How can the interaction between category density and category curiosity help explain individual differences? If a child is curious about particular categories (e.g., animals, vehicles), her curiosity might impact her learning of novel members of these categories. Additionally, we hypothesize curiosity to influence the semantic density of categories in the child’s vocabulary: Children should know more words in the categories they are more interested in.

Here we investigate the influence of category curiosity and category density on the acquisition of new word-object-associations. 30-months-olds (n=40) were, first, presented with 16 familiar objects from two broad (M = 31 members) and two narrow (M = 11 members) categories and heard their corresponding labels while their pupil dilation response was measured as an index of their interest in members of the different categories. Next, they were exposed to novel members from each of the four categories and tested on their learning of the new word-object-associations. In addition, a vocabulary questionnaire and a questionnaire on the child’s interests in different category members were administered.

Growth curve analysis indicate that children are able to learn novel members from both broad and narrow categories, but learning is more robust in the broad categories. This suggests that children are able to leverage their existing semantic knowledge to learn new words, which is in line with previous research (e.g. Borovsky, Ellis, Evans, & Elman, 2016a, 2016b). Pupil dilation analyses however showed that category size is not the only predictor: The best model considers both semantic knowledge and the child's interest in a category as predictors for learning.