Book Review: David Christensen. Putting Logic in its Place: Formal Constraints on Rational Belief
Patrick Maher · Notre Dame Journal of Formal Logic · 2006
Christensen distinguishes two kinds of belief, which he calls binary and graded.Binary belief is qualitative; graded belief is a matter of degree.He argues that the standards of ideal epistemic rationality do not require binary beliefs to be consistent, or closed under logical consequence, but they do require graded beliefs to satisfy the laws of probability.Thus the place of logic-alluded to in the title-is in governing graded, not binary, beliefs.This is not a technical book; instead it gives very readable discussions of the philosophical issues.Christensen often makes a convincing case that previous discussions have been too facile and for that reason I think anyone interested in epistemic rationality should read this relatively short book.On the other hand, Christensen's arguments for his positive views are less successful; I will show that he has failed to establish either that binary beliefs may violate consistency and closure or that graded beliefs must satisfy the laws of probability. Rationality and LogicChristensen begins by trying to clarify the concept of rationality that is central to the book.He says that rational beliefs "are those that arise from good thinking, whether or not that thinking was successful in latching on to the truth" (p.2).He distinguishes two ways in which a belief may be rational: pragmatic and epistemic.Christensen does not give a general account of this distinction but his examples and discussion suggest this: A belief is pragmatically rational if adopting it is a good means to the believer's ends; it is epistemically rational if it is in accord with the evidence (p.4).Christensen says his book is concerned with epistemic rationality, not pragmatic rationality.Since he is concerned with constraints that logic imposes on rational belief, Christensen briefly discusses the nature of logic.He says he is concerned with "formal" logic, but declines to say what that is, except that it is