[PS-2.84] Toddlers and pre-schoolers predict meaning but not form

Gambi, C. , Gorrie, F. , Pickering, M. J. & Rabagliati, H.

Department of Psychology, University of Edinburgh

Do children learn via prediction (e.g., Ramscar, Dye, & McCauley, 2013)? We asked when English-learning children start predicting form, an ability which could allow them to learn morpho-phonological generalizations. Although French two-year-olds can use informative gender-marking on determiners to identify the referents of nouns more quickly (e.g., Van Heugten & Shi, 2009), there is currently no evidence that young children use determiners to actively predict the form of nouns.

Two- (N=40), three- (N=47), and four-to-five- (N=40) year-olds, and adults (N=24) listened to instructions such as 'Can you find A car/AN apple?' while looking at pictures of the objects. Using Growth Curve Analysis (Mirman, 2014), we modelled changes in the proportion of looks to the target picture over time. Adults' looks to the target increased from determiner onset (t=9.19), but children's did not (2-year-olds: t=1.82; 3-year-olds, t=0.23; 4-5-year-olds, t=0.36). In contrast, two- (t=3.70), three- (t=5.37), and 4-to-5-year-olds (t=6.70), and adults (t=14.96) looked predictively at the target upon hearing 'Can you find TWO cars/ONE apple?'. In conclusion, children predicted the meaning (semantic number) but not the phonological form of upcoming nouns, providing a potential explanation for why they master the a/an distinction late in production.