Hilary Kornblith

In some views, knowledge cannot exist except against the background of certain social practices. Thus, in Davidson's view, there are no beliefs, and thus no knowledge, except in creatures that use and interpret language. In other views, such as Brandom's, belief, and thus knowledge, cannot exist except in creatures that have a social practice of giving and asking for reasons. Finally, there are views in which it is possible to have beliefs without social practices, but it is not possible to have justified beliefs, and hence knowledge, without some relevant social practice, such as holding individuals responsible for the beliefs they have. These views are examined in detail. It is argued that knowledge does not require social practices of any sort whatsoever.

In Chapter Four, it is argued that coherence is evidence of truth. A sketch of a theory of approximate truth is developed in terms of the theory of reference outlined in Chapter Three, and this notion is put to work in showing that there is reason to believe that most of our beliefs are at least approximately true. It is then argued that coherence with approximately true beliefs, and thus the beliefs we have, is evidence of truth. ;In Chapter Three, the connection between the theory of knowledge and the theory of reference is explored. The relation between Russell's theory of knowledge and his theory of reference is discussed, and it is argued that Russell was quite correct in insisting that there be a close relationship between these two disciplines. A theory of reference is developed which parallels the causal theory of knowledge presented in Chapter One. This theory is a variant of the causal, or historical theory, presented by Donnellan, Kripke, Putnam and Boyd. ;In Chapter One, an outline of a causal theory of knowledge is presented. It is argued that knowledge is reliably produced true belief, and that a process is reliable just in case it tends to produce true beliefs in relevant actual, as well as counterfactual, situations. ;In Chapter Two, the implications of this account for the debate between foundationalists and coherence theorists is examined. It is argued that standard foundational and coherentist accounts of justification share a common false presupposition, that the proper account of justification must be apsychological. It is argued that by rejecting this presupposition and adopting a causal theory of knowledge, it is possible to develop a theory which contains many of the important grains of truth in both foundationalism and coherentism. ;In this thesis, a causal theory of knowledge is elaborated and its implications examined

We sometimes stop to reflect on our mental states, and such reflection can lead, at times, to changing our minds. It can, as well, lead us to endorse the very attitudes which we previously held. Such reflective endorsement has been called upon to play a wide range of roles in philosophical theorizing. It has been thought to ground a distinction between two fundamentally different kinds of knowledge: reflective knowledge and mere animal knowledge. It has been thought to serve as a ground for rational change of belief. It has been called upon to explain the possibility of freedom of the will. And it has been brought into service to explain the source of normativity. This chapter argues that it can play none of these roles.

Hilary Kornblith presents an account of inductive inference that addresses both its metaphysical and epistemological aspects. He argues that inductive knowledge is possible by virtue of the fit between our innate psychological capacities and the causal structure of the world. Kornblith begins by developing an account of natural kinds that has its origins in John Locke's work on real and nominal essences. In Kornblith's view, a natural kind is a stable cluster of properties that are bound together in nature. The existence of such kinds serves as a natural ground of inductive inference. Kornblith then examines two features of human psychology that explain how knowledge of natural kinds is attained. First, our concepts are structured innately in a way that presupposes the existence of natural kinds. Second, our native inferential tendencies tend to provide us with accurate beliefs about the world when applied to environments that are populated by natural kinds.

Epistemology has changed dramatically since Descartes, but many of the questions epistemologists address today are no different from the questions Descartes addressed. I begin by raising four sets of questions with which Descartes concerned himself, and explain briefly why Descartes regarded these sets of questions as interchangeable. My main purpose, however, is not historical. Rather, I wish to present an outline of a naturalistic approach to these questions. I will not defend naturalistic epistemology. Instead, I hope to explore what a naturalistic approach to some traditional epistemological questions might look like; it is only through a better understanding of the consequences of naturalizing epistemology that we may hope properly to evaluate it. I will argue that questions which Descartes treated as interchangeable will have to be separated by naturalistically minded epistemologists. In the course of evaluating these questions from a naturalistic point of view, I hope to shed some light on the relation between reliability and justification, the dispute between internalist and externalist theories of justification, and the relation between ethics and epistemology.