Connectionist Models of Development:

The symbolic paradigm of cognitive modeling, championed by Minsky and Papert, Newell and Simon, and other pioneers of the 1950’s and 1960’s, remains very much alive and well today. Yet an alternative paradigm, first championed in the 1950’s by Rosenblatt, in which cognitive processes are viewed as emergent functions of a complex stochastic dynamical system, has continued to have adherents. A cooling of interest in such approaches in the 1960’s did not deter Grossberg (1976, 1978a) or James Anderson (1973), and the approach emerged again full force in the 1980’s. Thereafter, many others begun to explore the implications of this paradigm for development (Elman et al., 1996; McClelland, 1989; Plunkett and Marchman, 1993; Schulz, Mareschal & Schmidt, 1994). A parallel and closely related movement, emerging from the physical sciences, also began to gain adherents during the same decade (Schöner & Kelso, 1988) and began to attract the interest of developmental psychologists in the early 1990s (Thelen & Smith, 1994). The present chapter represents our attempt to underscore the common ground between connectionist and dynamical systems approaches. Central to both is the emergent nature of system-level behavior and changes to such behavior through development.