Proceedings of the Annual Meeting of the Cognitive Science Society

Young infants are prolific word learners even though they are facing the challenge of referential uncertainty (Quine, 1960). Many laboratory studies have shown that human infants are skilled at inferring the correct referent of an object from ambiguous contexts (Swingley, 2009). However, little is known regarding how children visually attend to and select the target object among many other objects in view when parents name it during free play interactions. In the current study, we explored the looking pattern of 12-month-old infants using naturalistic first person images with varying degrees of referential ambiguity. Our data suggest that infants’ attention is selective and they tend to only select a small subset of objects to attend to at each learning instance despite the complexity of the data existed in the real world. This work allows us to better understand how perceptual properties of objects in infants’ view influence their visual attention, which is also related to how they select candidate objects to build word-object mappings.