Chair: Victor Tamburini
Résumé
Previous research suggests that infants can represent not more than three items at a time. For example, infants are able to represent one, two, or three hidden objects, but fail with four (Feigenson & Carey, 2003), and they also choose the larger quantity when comparing, e.g., 1 vs. 2, and 2vs. 3, but fail when comparing, e.g., 3 vs. 4, and 2 vs. 4 Feigenson et al., 2002), exhibiting the signature limit of object file representations. However, infants can overcome the limit of three by chunking individuals into sets when given perceptual, conceptual, linguistic, or spatial cues (Feigenson & Halberda, 2004, 2008). Infants are limited in the number of items they can chunk (they can form sets of two but not three), and in the number of sets they can track (three sets but not four). However, infants are able to form hierarchically organized ‘superchunks’ (two sets of two sets of two; Rosenberg & Feigenson, 2013).
In our study, we investigate whether infants have a representation of four when they represent two sets of two. We tested 30 10- to 18-month-old children (data collection ongoing). They were presented with two boxes in which the experimenter placed different amounts of crackers. After placing all of the crackers, children were allowed to crawl to one of the boxes and retrieve the crackers within the chosen box. Each child underwent up to four trials with the following combinations: 1 vs. 3, 1 vs. 4, 3 vs. 2x2, and 3 vs. 4. If children represent two sets of two as four items, they should successfully choose the box with two sets of two crackers over the box containing three crackers. We found that children reliably succeeded in the 1 vs. 3 condition (80% choices of 3) and chose at chance in the 1 vs. 4 condition (50% choices of 4). Comparing the 3 vs. 2x2, and 3 vs. 4 conditions, children’s performance was almost identical (around 65% choices of 3), suggesting that two sets of two are not represented as four.
