Introduction

Most species are endowed with multiple sensory systems that enable them to rely on a combination of multisensory signals for perception and action (Stein & Meredith, 1993). Moreover, as noted by Maier and Schneirla (1964), the reliance on multisensory inputs for control of behavior becomes the rule rather than the exception as one ascends the vertebrate phylogenetic scale from fish to human. Given these two facts, the obvious question is whether redundant, multisensory event and object specification is advantageous from the standpoint of perception and action. This question has figured more or less prominently in various recent discussions about the role of multisensory control pertaining to perception and information processing in general (Massaro, 1998), brain-behavior relations (Damasio, 1989; Edelman, 1992; Edelman & Tononi, 2000; Ettlinger & Wilson, 1990; Stein & Meredith, 1993), and neural and behavioral development (Edelman, 1992; Lewkowicz & Lickliter, 1994; Thelen & Smith, 1994).

In general, it is reasonable to conclude that the widespread evolutionary emergence of multisensory perception and action in the animal kingdom is likely to reflect the fact that multisensory input and the redundancy that input provides have adaptive value (J. J. Gibson, 1966, 1979; Maier & Schneirla, 1964; Marks, 1978; Piaget, 1952; Werner, 1973). For example, a predator that can spot a prey regardless of whether it can smell it, hear it, or see it is at an adaptive advantage over one that can only smell the prey. Likewise, a prey that can avoid a predator regardless of whether it can smell it, hear it, or see it is also at an adaptive advantage. As we will show, empirical evidence from studies of adult organisms indicates that multisensory redundancy provides a variety of advantages. This chapter reviews the available empirical evidence on the development of intersensory perception in human and animal infants to determine whether the same might be true in early development.