This dissertation puts forth a series of arguments about the extent to which human cooperative interaction is fundamentally shaped by mindreading; i.e. the capability to reason about the psychological causes (e.g. intentions, beliefs, goals) of behavior.

The introduction to this dissertation discusses the broad philosophical underpinnings that lay the foundations for more specific philosophical issues under discussion in subsequent chapters.

In chapter two, I argue that a thorough interpretation of the relevant empirical evidence suggests that mindreading is fast, effortlessly deployed, and operative sub-personally. For this reason, mindreading is principally well-suited to enable most everyday cooperative interactions. In the appendix, I (in collaboration with Evan Westra ) elaborate on this picture, arguing that the cognitive mechanisms operative in social interactions are, in all relevant respects, similar to those operative in non-interactive situations.

While chapter two and the appendix defend the idea that the cognitive faculties responsible for mindreading are fit to enable cooperative interactions, chapters three and four take this perspective for granted and discusses whether human cooperation is crucially dependent on a form of reciprocal attribution of mental states that is often labeled common knowledge.

In chapter three of this dissertation I address, and reject, the oft defended idea that truly performing an action together with others requires that all parties commonly know their intended goals. I argue that this view is fundamentally mistaken. Successfully acting together with others often requires not knowing these goals.

Chapter four explores reciprocal belief attribution in the context of coordination problems. Humans often coordinate their actions by replicating successful past choices; they reason based on precedent. Philosophers have often claimed that solving coordination problems by relying on precedent presupposes common knowledge that all parties rely on precedent in trying to coordinate their actions. Chapter four points out that this assumption is erroneous: Coordinating behavior on the basis of precedent is broadly incompatible with any higher-order knowledge (or beliefs) about the other agents’ choices.