Flaws in the Notion of Rational Imitation: On the Impossible Leap from Computation to Rationality and Normativity

摘要

For over a century, scholars have been intrigued by imitation {e.g., Royce, 1895; Tarde, 1890/1993] across a variety of disciplines including developmental psychology, ethology, neuroscience, cognitive science, sociology, and robotics [e.g., Heyes, 2001]. Imitation has attracted attention because it has been considered a key mechanism in socialization [Aronfreed, 1968; Bandura, 1977] and social understanding [Meltzoff, 2011], Traditionally, research in this area has focused on cataloguing the different forms of imitation, charting the developmental course of imitation, and identifying the processes involved in imitation [Heyes, 2001; Hurley & Chater, 2005a, b; Jones, 2007]. Since its beginning, empirical research on imitation has been accompanied by disagreement about the interpretation and explanation of imitative behaviors (e.g., see the longstanding debate over the status of newborn imitation [Anisfeld, 1996; Guernsey, 1928; Maratos, 1982; Meltzoff & Moore, 1977]). This controversy continues with Paulus' [this issue] analysis of the experimentally produced phenomenon labeled 'rational imitation.' We focus first on Paulus' interesting analysis of the claim that infants imitate rationally, and, second, on the expansion of problematic nativist computational theorizing into the domain of imitation.