Laws of nature have been traditionally thought to express regularities in the systems which they describe, and, via their expression of regularities, to allow us to explain and predict the behavior of these systems. Using the driven simple pendulum as a paradigm, we identify three senses that regularity might have in connection with nonlinear dynamical systems: periodicity, uniqueness, and perturbative stability. Such systems are always regular only in the second of these senses, and that sense is not robust enough to (...) support predictions. We thus illustrate precisely how physical laws in the classical regime of dynamical systems fail to exhibit predictive power. *R. G. Holt gratefully acknowledges the support of the National Center for Physical Acoustics at Oxford, Mississippi, and the Office of Naval Research. (shrink)

Recent experimental work by T.D. Wilson et al. indicates that a consequence of asking subjects to reflect on their attitudes is that they not only reduce the consistency between their attitudes and behavior, but they perform actions which they come to regret. Wilson interprets this work via intra-psychic concepts, and arrives at the conclusion that it is rational to avoid deliberating about a wide range of attitudes and behaviors. This consequence has objectionable implications for philosophical theories of deliberative practical rationality. (...) I respond to this challenge by reinterpreting the experimental results in a way which is not only consistent with a certain theory of deliberative practical rationality but in which the results lend support to that theory. My interpretive focus is on attending closely to the social circumstances of subjects in the experiments[1]. (shrink)

Abstract The thesis that teleological explanations are best understood as causal explanations is defended (contra Valentine). I shift the focus of debate from behavior simpliciter to allegedly rational behavior. Teleological explanation, in the case of rational agents, involves reason?giving; and the reasons agents give for acting must be causative of that action if those agents are to be rational in practice. I argue initially that to abandon the claim that reasons are causes of action is to abandon that which renders (...) many generalisations of systematic human behavior intelligible. I then adduce and defend against two major objections to the secondary thesis that reasons can be causes of action: (1) Experiments in attribution research suggest that agents are characteristically mistaken when reporting on the causes of their behavior; (2) Causal explanation must refer to some law, and reasons cannot enter into such laws as causes of action. In responding to these objections, I note that strict predictability is probably not a reliable indicator of genuine causal explanation; some explanations will afford predictions, others will not. (shrink)