The Dynamic Field Theory and Embodied Cognitive Dynamics

The goal of this chapter is to explain some of the core concepts of Dynamic Field Theory (DFT) and how this theory provides a formal framework for thinking about embodied cognitive dynamics. The DFT is now 15 years old. In 1993, Gregor Schöner and his colleagues published a proceedings paper presenting a theory of how eye movements are planned, including their neural bases (Kopecz, Engels, & Schöner, 1993). Since that time, DFT has been extended to a range of topics including the planning of reaching movements (Bastian, Riehle, Erlhagen, & Schöner, 1998; Bastian, Schöner, & Riehle, 2003; Erlhagen & Schöner, 2002), the development of motor planning (Thelen & Smith, 1994), the perception of motion (Hock, Schöner, & Giese, 2003; Jancke, Erlhagen, Schöner, & Dinse, 2004), the processes that underlie habituation in infancy (Schöner & Thelen, 2006), the control of autonomous robots (Bicho, Mallet, & Schöner, 2000; Steinhage & Schöner, 1998), the processes that underlie visuo-spatial cognition and spatial language (Lipinski, Spencer, & Samuelson, in press; Spencer, Simmering, Schutte, & Schöner, 2007), the development of visuo-spatial cognition (Simmering, Schutte, & Spencer, 2008), the processes that underlie visual working memory and change detection (Johnson, Spencer, & Schöner, 2006), the fast learning of object labels and other aspects of word learning (Faubel & Schoner, in press; Samuelson, Schutte, & Horst, 2008), and the development of executive function (Buss & Spencer, 2008). Despite the extensive application of DFT, discussion at the conference from whence this edited volume emerged suggested that aspects of DFT are poorly understood. Pockets of researchers are familiar with one model or another, but there is little consensus regarding the broader theoretical concepts that underlie this theory. One barrier to understanding is that there have been relatively few reviews of DFT pitched squarely at the central concepts. The present chapter takes up this charge. We begin with some context, describing the milieu in which the theory emerged and the problems it was initially designed to tackle. In particular, we focus on a central problem in cognitive science: how do cognitive systems interface with real-time sensori-motor systems? That is, how can real-time cognitive dynamics be embodied. We discuss our answer to this question within DFT where we use neural dynamics and stable “peaks” of activation as the basic unit of cognition. Peaks can be continuously coupled to sensori-motor systems in real time. But peaks move far beyond a basic feed-forward picture, allowing for a host of cognitive properties that “come for free” in this theoretical framework. We highlight these properties, and also show how peaks provide critical links to two other central challenges in cognitive and developmental science: the integration of processes over multiple time scales and the origin of behavioral flexibility.