Infants can use many types of categorical descriptors to individuate objects

• In Event: 2-031 - New advances in the study of object representations in infancy: Formation, structure, and content

Integrative Statement

Prior research indicates that infants can use various types of categorical descriptors (e.g., ontological, linguistic, and functional) to individuate objects: After watching two categorically-distinct objects emerge in alternation from behind a screen, infants detect a violation if the screen is lowered to reveal only one of the objects. Here we examined two questions. First, could infants use another type of categorical descriptors, prior event roles (formerly a “supportee”, formerly a “support”), to individuate objects (Experiments 1-2)? Second, would mere exposure to static arrays of objects from two perceptual categories (cylinders, blocks) be sufficient to induce infants to individuate objects from these categories (Experiments 3-4)?

In Experiment 1, 13-month-olds received two familiarization trials depicting support events. In one, a cylinder (e.g., orange-striped) was placed on top of a star-shaped block; in the other, a rectangular block was placed on top of another cylinder (e.g., blue-starred). In the test trial, infants saw the two cylinders (different-objects condition) or the same cylinder (same-object condition) emerge on either side of a screen, which was then lowered to reveal only one cylinder. Infants in the different-objects condition looked significantly longer at this final display than did infants in the same-object condition, suggesting that they (a) assigned distinct event roles--supportee, support—to the two cylinders during the familiarization trials and (b) used these categorical descriptors to individuate the cylinders during the test trial. Control Experiment 2 provided evidence for this interpretation: When both cylinders played identical roles (e.g., both were supports) in the familiarization trials, infants in the two conditions now looked equally at the final display.

In Experiment 3, 9-month-olds received two familiarization trials depicting static arrays of objects from two perceptual categories. In one trial, three red cylinders with different yellow patterns rested in a row on the apparatus floor; in the other trial, three green blocks with different white patterns rested on the floor. In the test trial, infants saw one of the cylinders and one of the blocks (different-objects condition) or just one of these objects (same-object condition) emerge on either side of a screen, which was then lowered to reveal only one object. Infants in the different-objects condition looked significantly longer than infants in the same-object condition, suggesting that they (a) were induced by the familiarization trials to form two distinct perceptual categories, cylinders and blocks, and (b) used these categorical descriptors to individuate the objects in the test trial. Control Experiment 4 supported this interpretation: When shown mixed arrays in the familiarization trials (two cylinders and one block, two blocks and one cylinder), infants in the two conditions now looked equally at the final display.

Our results thus show that infants can use a wide range of categorical descriptors to individuate objects. As such, they support a new account of early object individuation that focuses on interactions between the object-file and physical-reasoning systems and proposes that young infants succeed at individuating objects whenever the object-file system can assign distinct categorical descriptors—any type of categorical descriptor—to the objects.

Authors

• Yi Lin, University of Illinois Urbana-Champaign
  Presenting Author

• Maayan Stavans, Bar-Ilan University
  Non-Presenting Author

• Renee Baillargeon, University of Illinois at Urbana-Champaign
  Non-Presenting Author