Mindcraft: a Dynamical Systems Theory of Cognition

Abstract

<p>This dissertation develops a theory of cognition, driven by recent developments in the electrophysiological investigation of the neuronal mechanisms that support adaptive behavior. In the first chapter, I situate the theory in the conceptual landscape of the philosophy of mind, distinguishing componential from systemic dynamical theories of cognition. In the second chapter, I analyze two case studies from electrophysiological cognitive neuroscience, arguing that cognitive neuroscientists are beginning to uncover the dynamical components of cognition. Drawing on the recent literature on mechanisms and scientific explanation, I propose a revised definition of a mechanism that accommodates these dynamical mechanisms, as well as making room for their implementation by physical mechanisms. In the third chapter, I argue that the investigation of a particular class of intelligent behavior begins with the construction of a formal model of the processing problem for that behavior, where this model is distinct from the physical device and the functions performed by the device's components. In the third chapter, I argue that the component dynamical mechanisms of cognitive systems are distinct from though implemented by physical mechanisms. These dynamical mechanisms are described by sets of differential equations, possess a set of organized components and activities that execute the formal models of processing, and are implemented by the physical machinery of the cognitive system, such as the brain. After I argue that these multiple interacting dynamical mechanisms are the components of cognition, defending this componentiality claim against several objections, I define the implementation relation that holds between dynamical and physical mechanisms. I next discuss the grounds for inferring the existence of dynamical mechanisms that are type distinct from physical mechanisms, their implementing substrate. In the fourth chapter, I argue that these dynamical mechanisms are reused: they can execute different formal models and be implemented by different physical substrates. I define this concept of reuse, situating it in the debate on theories of reuse, and illustrate how dynamical mechanisms are reused in cognitive systems.</p>