Induction, overhypotheses, and the shape bias: Some arguments and evidence for rational constructivism

Abstract

The authors in this chapter focus on a case study of how object representations in infants interact with early word learning, particularly the nature of the so-called 'shape bias'. A short review of the controversies in this subfield is used to illustrate the two dominant views of cognitive development, which can be roughly classified as nativist or empiricist. Also presented are theoretical arguments and new empirical evidence for a rational constructivist view of cognitive development. The authors' goal in this chapter is to argue for a new approach to the study of cognitive development, one that is strongly committed to both innate concepts and representations, as well as powerful inductive learning mechanisms. In addition to discussing the 'shape bias' and how it relates to object representations, generality of the approach is briefly discussed.