Rationality, Bias, and Mind: Essays on Epistemology and Cognitive Science

Abstract

In these essays, I offer solutions to a cluster of issues in epistemology, the philosophy of mind, and the philosophy of cognitive science. These issues center around the rational evaluability (and rational evaluation) of mental states found in human beings. Which kinds of mental states can be rationally evaluated? In virtue of what can they be so evaluated? And what role do facts about our cognitive limitations and biases play in rational evaluation? I offer answers to these questions in what follows. The first two chapters concern a particular kind of mental state, the implicit aitude. Implicit aitudes are opaque mental states that we often lack conscious access to, but that influence our behaviors in many ways. In Chapter 1, I argue that implicit aitudes, like beliefs, credences, and intentions (and unlike associations or mere flights of fancy) can be rationally evaluated. I argue this is because they have a capacity to respond to evidence. In Chapter 2, I show that the implicit aitudes are best thought of as belonging to a familiar psychological kind: belief. Our implicit beliefs form a significant portion of our rational outlook, although we often remain alienated from their contents and functions. Coming to terms with these facts is of the utmost importance for both epistemology and the philosophy of mind. In Chapter 3, I turn my aention to the rationality of psychological biases more generally. I consider so-called biases of reasoning, often known as heuristics, and ask about their rational standing. I argue that utilizing heuristic reasoning strategies often represents the best we can do from the point of view of rationality. Heuristic reasoning represents a local rational maximum, producing more accurate conclusions in our reasoning than our aempts to implement more “ideal” reasoning rules. I discuss three components of the human heuristic reasoning strategy (aributive substitution, recognition, and one-reason reasoning), and show how their use can lead to accurate conclusions in information-rich environments.