Rationality and irrationality in evidence and proof: a comment on ‘The structure and logic of proof in trials’ by Professor Tillers

In his short paper,1 Peter Tillers treads a narrow path between rationality and irrationality in legal evidence and proof. He hopes that we can improve our fact-finding practices—make them more rational—but also worries that this is not possible. There is much that I agree with in the paper, though that probably has something to do with the abstract level at which the arguments are pitched—it is often easier to agree with generalities than specifics. However, I want to press Tillers on some points and to suggest that he may sometimes veer too far towards the irrationality side of the path. Tillers’s discussion revolves around a series of contrasts all of which roughly correspond with each other. In terms of existing theories of proof, there is abstract theorizing from first principles, contrasted with traditional doctrinal scholarship that tends to catalogue existing principles of Anglo-American law of proof, principles that may be contingent and culturally variable. This abstract/contingent distinction maps fairly cleanly onto a more general differentiation between the normative and the descriptive. And then, a little less neatly, these bipolarities can be aligned with a distinction between the rational and the irrational. What is obviously tendentious in relation to this last move is the suggestion that actual human reasoning is aligned with irrationality. I do not endorse such a generalization, but it does not seem controversial to suggest that, at least sometimes, normal human reasoning departs from ideals of rationality. It is obvious that Tillers is wary of drawing too sharp a distinction between any of these contrasted perspectives. Most explicitly, he says: ‘I do not believe in the radical separation of descriptive inferential theory and normative inferential theory’. As a generality, I would not want to disagree with that, but I do think it is worth being more explicit here: there are many different ways of blurring the descriptive/normative distinction.2 At the extreme is the sort of strong ‘replacement’ naturalism associated with Quine under which the normative is actually replaced with the descriptive: epistemology becomes a matter of descriptive psychology. In relation to the rationality/irrationality pair, some of L. J. Cohen’s work moves in a similar direction.3 But less extreme ways of connecting the normative to the descriptive exist, as when we take the normative—say, the rules of inductive logic— to be an ideal which we try to approach, even if constraints of time and resources make reaching the ideal impossible. I suspect that Tillers, like me, falls towards the less radical end of this spectrum, but it would be helpful to know more about how he sees the connection between the normative and