Evidentialism, Reliabilism, Evidentialist Reliabilism?

Experientialism is compared and contrasted with Evidentialism, Reliabilism, and Evidentialist Reliabilism. The generality problem for Reliabilism is discussed, as well as the issue of how to measure reliability. A probabilistic understanding of reliability is put forward. In particular, reliability is understood in terms of evidential probabilities, not physical probabilities. An extension to credences is explored. Experientialism is non-Evidentialist insofar as it does not take experience to be evidence, and is non-Reliabilist insofar as it appeals to a normatively loaded notion of evidential probability.

The Normative Force of Unjustified Beliefs

The previous chapter argued that our evidence is constituted by those beliefs that are ultima facie justified by our experiences—but what about our unjustified beliefs? Do they have any justificatory power of their own? This chapter discusses a puzzle that arises regarding that question, and argues that we should take to heart the lessons learned from an older puzzle: the puzzle of contrary-to-duty obligations. It is argued that these obligations, as well as what Broome has called “normative requirements,” can be understood in the model of the restrictor view of conditionals.

Probability and Decision Theory

This chapter introduces the mathematics of probability and decision theory. The probability calculus is introduced in both a set-theoretic and a propositional context. Probability is also related to measure theory, and stochastic truth-tables are presented. Problems with conditional probability are examined. Two interpretations of the probability calculus are introduced: physical and normative probabilities. The problem of logical omniscience for normative probabilities is discussed. Dutch Book arguments and accuracy-based arguments for Probabilism (the claim that our credences must satisfy the probability axioms) are examined and rejected. Different interpretations of the “idealization” reply to the problem of logical omniscience are considered, and one of them is tentatively endorsed. The expected utility maximization conception of decision theory is introduced, and representation arguments are considered (and rejected) as another reply to the problem of logical omniscience.

Introduction

This chapter introduces the rest of the book. It is assumed that practical rationality requires theoretical rationality. Three views about basic empirical evidence are introduced: Psychologism, according to which our basic empirical evidence is constituted by our experiences; Factualism, according to which our basic empirical evidence is constituted by those propositions we know through experience; and Experientialism, according to which our basic empirical evidence is constituted by those propositions we are basically justified in believing by our experiences. The rest of the book is dedicated to developing objections to Psychologism and Factualism, and to defending Experientialism.

Conclusion

The main threads of the book are brought together and summarized.

The Problem of Easy Rationality

This chapter discusses a version of Cohen’s problem of easy knowledge, arguing for a novel resolution. A brief history of the problem is presented, and then a systematized version of it is developed. Possible solutions are examined, and denying the existence of ampliative inference is defended as a possible solution. It is explained how this is not as bad as it sounds. The chapter also discusses the relation between this solution and one of the arguments against Factualism presented in chapter 4. It is explained that those arguments against Factualism remain in place even if one rejects the possibility of ampliative inference. Finally, the issue of whether and how the problem of easy rationality applies to Experientialism is discussed.

Excuses, Would-Be Knowledge, and Rationality-Based Decision Theory

This chapter considers a reply to the previous chapter, a reply according to which although we can never have false justified beliefs, we can have excusable (or reasonable) false beliefs. The chapter starts by arguing that any view according to which there cannot be false justified beliefs must face two problems: the demarcation problem and the suspension problem. While the excuses maneuver may help with the demarcation problem, it cannot help with the suspension problem. The chapter then examines two particular instatiations of the excuses maneuver: Williamson’s (briefly) and Lasonen-Aarnio’s (in more detail). Finally, the chapter also examines attempts at defining justification as “would-be knowledge,” and finds them wanting as well.

Knowledge-Based Decision Theory?

This chapter argues against Factualism, holding that the thesis fails in both directions: not everything we know is part of our evidence, and not all evidence is knowledge. Fumerton’s thesis (that rational action requires rational beliefs and, as added here, rational credences) is defended from the charge that it leads to unacceptable dilemmas. It is then argued that full beliefs (and not just credences) have a role to play in Bayesian decision theory. The argument that not all knowledge is evidence appeals to inferential knowledge. The main claim of the chapter is that not all evidence is knowledge. It is argued that it can be rational to act on the basis of false beliefs and therefore (given Fumerton’s thesis) that it is rational to believe those false propositions.

Objective Bayesianism and Knowledge-First Epistemology

This chapter distinguishes between subjective and objective versions of Bayesianism, arguing for the latter variety. The distinction is made in terms of constraints on ur-priors. Lewis’s “Principal Principle” is discussed as one such constraint. The Carnapian program of delineating a unique ur-prior in purely syntactical terms is presented, and rejected for familiar reasons. It is then argued that the failure of the Carnapian program does not entail the failure of Objective Bayesianism more generally. Ur-prior Conditionalization is introduced as a better alternative to Conditionalization. The chapter ends by presenting Factualism, the view that our evidence consists of what we know.

Experientialism

This chapter introduces and defends Experientialism, the view according to which our evidence consists in those beliefs that are ultima facie justified by our experiences. Experientialism is contrasted with Factualism and Psychologism. It is argued that both Psychologism and Factualism undercount how much evidence we have. Against Factualism, it is argued that it violates a “same subject matter” constraint on evidence. Against Factualism, it is argued that it entails the prevalence of epistemic dilemmas. Experientialism is defended as having neither of the problems of the other two views. The objection that Experientialism is compatible with the existence of inconsistent evidence is considered. Finally, the objection that Experientialism has troubles accounting for the defeasibility of empirical justification is also considered.