Beyond associations: Sensitivity to structure in pre-schoolers' linguistic predictions

Research output: Contribution to journalArticlepeer-review

Abstract / Description of output

One influential view of language acquisition is that children master structural generalizations by making and learning from structure-informed predictions. Previous work has shown that from 3 years of age children can use semantic associations to generate predictions. However, it is unknown whether they can generate predictions by combining these associations with knowledge of linguistic structure. We recorded the eye movements of pre-schoolers while they listened to sentences such as Pingu will ride the horse. Upon hearing ride, children predictively looked at a horse (a strongly associated and plausible patient of ride), and mostly ignored a cowboy (equally strongly associated, but an implausible patient). In a separate experiment, children did not rapidly look at the horse when they heard You can show Pingu … "riding", showing that they do not quickly activate strongly associated patients when there are no structural constraints. Our findings demonstrate that young children's predictions are sensitive to structure, providing support for predictive-learning models of language acquisition.
Original languageEnglish
Pages (from-to)340-351
Number of pages12
Early online date12 Oct 2016
Publication statusPublished - Dec 2016

Keywords / Materials (for Non-textual outputs)

  • association
  • linguistic structure
  • prediction
  • visual-world


Dive into the research topics of 'Beyond associations: Sensitivity to structure in pre-schoolers' linguistic predictions'. Together they form a unique fingerprint.

Cite this