Infants generate goal-based action predictions

摘要

Predicting the actions of others is critical to smooth social interactions. Prior work suggests both understanding and anticipation of goal-directed actions appears early in development. In this study, on-line goal prediction was tested explicitly using an adaptation of paradigm for an eye-tracking task. Twenty 11-month-olds were familiarized to movie clips of a hand reaching to grasp 1 of 2 objects. Then object locations were swapped, and the hand made an incomplete reach between the objects. Here, infants reliably made their first look from the hand to the familiarized goal object, now in a new location. A separate control condition of 20 infants familiarized to the same movements of an unfamiliar claw revealed the opposite pattern: reliable prediction to the familiarized location, rather than the familiarized object. This study suggests by 11-months infants actively use goal analysis to generate on-line predictions of an agent’s next action.