Cristina Borgoni - The Fragmented Mind

This book is about the hypothesis that the mind is fragmented, or compartmentalized. When it is said that an agents mind is fragmented, it is usually meant that their overall belief state is fragmented. To a first approximation, a belief state can be said to be fragmented if it is divided into several sub-states of the same kind: fragments. Each fragment is consistent and closed under entailment, but the fragments taken together need not make for a consistent and closed overall state. The thesis that the mind is fragmented contrasts with the widespread, if often implicit, assumptioncall it Unitythat the mind is unified, i.e. that an agents overall belief state is consistent and closed under entailment. The motivation for fragmentation comes from a number of places, notably the shortcomings of Unity: the problem of logical omniscience, the problem of inconsistent doxastic states, cases of cognitive dissonance and imperfect information access, and others. In this Introduction, we outline what varieties of fragmentation have in common and what motivates them. We then discuss the relationship between fragmentation and theses about cognitive architecture, introduce two classical theories of fragmentation, and sketch recent developments. Finally, as an overview of the volume, we present some of the open questions about and issues with fragmentation that the contributions to this volume address.

Doxastic states like belief and epistemic states like knowledge are standardly assumed to be inherently rational. Much of epistemic logic, Bayesian accounts of human belief, decision theory, and some influential views about rationality proceed from the underlying view that the mindor at least doxastic statesis unified:

Unity

Agents have a unified representation of the world (at time t)a single state of belief organized by two principles:

1. Consistency: The total set of an agents beliefs (at t) is consistent.

2. Closure: The total set of an agents beliefs (at t) is logically closed. That is, agents believe the logical consequences of their beliefs.

Unity may seem to impose unreasonably strong requirements on doxastic states. In the literature, Unityas we call itis often endorsed with one of the following qualifications.

First, Unity is often thought of as part of a descriptive theory of ideal rational agents, not of real agents. Thus, some authors implicitly or explicitly take their analyses to describe suitably idealized versions of real agents. Idealization affords many theoretical advantages, including simplicity in accounting for logical relations among beliefs. None of this entails that the theory applies directly to real agents. An open question here is of course whether real agents are similar enough to these idealized agents for the theory to have any use in the explanation of real agents doxastic attitudes (see Idealization and Explanatory Power below).