Fitting Belief
Conor McHugh (University of Southampton)
Aristotelian Society, 10th February 2014
I. Introduction
Two plausible claims:
(a) Belief has a standard of correctness: the belief that p is correct when p is true (or known) and
incorrect when p is false (or not known).
(b) This standard is fundamental to epistemic normativity. For example, it explains why certain
considerations are epistemic reasons for certain beliefs.
First question: What is the nature of this standard of correctness?
Second question: How does it ground epistemic normativity?
Good answers to these questions should
(D1) cohere with plausible claims in the theory of normativity;
(D2) contribute to (or at least be compatible with) an explanation of the dominance of evidence
among reasons for belief.
II-IV. What is correctness for belief?
Prescriptive view: You (ought/) may believe what’s true, and you may not believe what’s false.
- Not clear this view explains why evidence for p is a reason to believe p at all; a fortiori it fails D2.
- Tension with the claim that what you ought and may do are determined by reasons (D1).
Evaluative view: True belief is good (qua belief) and false belief is bad (qua belief).
- Since there are many values, this view seems to fail D2. Maybe it’s constitutive of epistemic
normativity that it only cares about one fundamental value—the value proper to belief. But it’s not
in general true that, if x is an F, whether x is a good F dominates other considerations in rationally
choosing or evaluating x.
- Tension with the view that what’s good is what’s fitting (correct) to value (D1).
- And it’s prima facie odd to suppose that something’s being good makes it correct.
General point: It’s not obvious that either of the above views is well motivated. Why think that
correct belief must be permitted or good belief? Isn’t it enough that it is correct?
V. Correctness as fittingness
I propose that we take correctness for belief to be an instance of the familiar property of
correctness, or fittingness, for attitudes. When you intend, want, admire, regret, or fear something,
your attitude can be fitting or unfitting, correct or incorrect, depending on the features of its object.
Fittingness is a normative property in its own right.
1
VI. Reasons and reasoning
I propose that the standard of correctness for belief can explain why certain considerations are
reasons because what it is to be a reason is explained in terms of correctness/fittingness.
(RR)
For the fact that p to be a reason to believe q is for p to be a premise of a good pattern of
reasoning leading from fitting responses to the belief that q.
(GF)
For it to be the case that {P1, …, Pn} > C is a good pattern of reasoning is for it to be the case
that, normally, when {P1, …, Pn} are fitting, so too is C.
VII. Attractions of the fittingness view
- Fits into a plausible general picture of normativity (D1), and doesn’t arbitrarily take correctness for
belief to be understood in terms of some seemingly distinct property.
- Satisfies (D2).
VIII. Insufficiency of truth-preserving patterns?
(T) Belief that grass is green > Belief that 79 is prime
Reply: In this case one can’t avail of the premise in competent reasoning leading one to add the
conclusion to one’s beliefs. So, either we maintain that the reasoning is good and that the premise is
a reason, albeit a useless one, or we revise the account of good reasoning:
(GFR)
For it to be the case that {P1, …, Pn} > C is a good pattern of reasoning is for it to be the case
that you can come to hold C by competent reasoning from {P1, …, Pn} (where competence is
sensitivity to fittingness-preservation).
IX. Reasons not to believe
(GN1)
For it to be the case that {P1, …, Pn} > C is a good pattern of reasoning is for it to be the case
that, normally, when {P1, …, Pn} are fitting, C is unfitting.
Problem: Undergeneration of reasons not to believe. This might be solved by taking the standard of
fittingness for belief to be knowledge rather than truth.
(GN2)
For it to be the case that {P1, …, Pn} > C is a good pattern of reasoning is for it to be the case
that {P1, …, Pn} > C is not a good pattern.
Problem: Overgeneration of reasons not to believe. But this might not be as bad as it seems; these
reasons will be maximally weak.
C.McHugh@soton.ac.uk
2