The 2nd annual Interdisciplinary Approach to
Philosophical & Psychological Issues Conference

"The Architecture of the Embodied Mind"
Daniel A. Weiskopf (Invited Speaker)
Georgia State University

Proponents of embodied cognition (EC) claim that a creature’s cognitive capacities are essentially related to or constituted by its bodily features. The EC movement has been gaining significant support in recent years across the cognitive sciences. But its proponents disagree about what bodily features are relevant and what their relevance consists in. Moderate EC claims that the bodily features in question are our cognitive sensorimotor capacities; that is, the perception-action systems of the brain. Moderate EC can be viewed as a modern offshoot of the empiricist tradition. Radical EC claims that cognitive capacities are constituted by extraneural features of the body itself. I aim to undermine the case for EC, focusing first on arguments for moderate EC, and then briefly on those favoring radical EC.

I consider three broad theoretical arguments that purportedly support EC. The first appeals to representational parsimony. The second appeals to the simplicity of the cognitive architecture EC requires. The third is a teleological argument concerning the purpose of thought, language, or other higher cognitive capacities. I claim that none of these arguments are persuasive. And opponents of EC can also make parallel arguments in favor of their own position. The theoretical issues alone are too high-level to resolve the debate.

Next I turn to two kinds of empirical argument in favor of EC. The first focuses on the role of action systems in language comprehension. A large body of research by Glenberg, Kaschak, Zwaan, Pulvermuller, and others allegedly supports the claim that language understanding (and the conceptualized thoughts that are associated with it) essentially involve representing actions. In a slogan: thinking is simulating acting. The evidence for this claim comes from a host of behavioral and neuroscientific studies. I review these studies and show that none of them are conclusive, and some actually argue against modest EC. I sketch an alternative architecture for language comprehension that does not reduce comprehension to action simulation and discuss the sort of evidence that would further empirically distinguish the models.

The second argument focuses on the role of gesture in cognition. Research by Goldin-Meadow and others has shown that spontaneous gestures during speech often convey relevant aspects of content that are not encoded linguistically. They further argue that these gestures play a functional role in cognition, since preventing gesturing impairs some aspects of cognition. Hence they contend that cognition essentially involves embodied action. This is one version of the radical EC thesis. This line of argument is supported by philosophers like Mark Rowlands, who argues that gestures and bodily movements should be considered representational in their own right. I argue that the empirical evidence is equivocal as to whether gestures play a functional role in virtue of their content. Without playing such a role it isn’t clear that we should assign them representational status. I conclude by asking whether there is a notion of cognitive architecture on which bodily states might play an essential but non-representational role.