Laws of Nature.

Most philosophers agree that there is a fundamental difference between laws of nature, on the one hand, and accidental regularities, on the other. In this interesting and important book, John Carroll concurs with a number of recent philosophers that the regularity theories of laws that long dominated twentieth-century philosophy of science cannot account for the difference. But unlike many other critics of regularity accounts, he argues that no other reductive account of laws can work in their place.

In Chapter I Carroll develops his important thesis that the nomic concepts (lawhood, causation, chance, the counterfactual conditional, and disposition) are central to our conceptual scheme. He argues that virtually all our concepts--from those of colour and physical object, to value and belief--presuppose nomic concepts; they can only be instantiated if there are some lawful and causal connections between things in the world. For example, a person's mental state qualifies as a belief only if it stands in certain sorts of causal relations to her environment, to her other mental states, and to her behaviour. Although most of Carroll's arguments about the nomic presuppositions of specific concepts are too quick to convince a confirmed nonrealist about causation, they are suggestive and most are quite plausible. They also nicely complement recent arguments that the key aspects of properties are the causal powers or capacities they confer on their instances.

In Chapter 2 Carroll turns to regularity theories of laws. Although philosophers wrangle over the characteristic features of laws, it is frequently thought that laws differ from mere regularities in being confirmed by their instances, possessing genuine explanatory and predictive power, and having some sort of modal nature. This modal aspect of laws shows up in their ability to support counterfactuals and in claims about what they require or preclude (Pauli's exclusion principle requires that two fermions occupy different quantum states; the laws of thermodynamics show the impossibility of perpetual motion machines).

We have no obvious epistemic access to this modal aspect of laws, however, and this stimulated empiricists to search for epistemically accessible necessary and sufficient conditions for a universal generalization to be a law. They responded with so-called regularity theories, according to which laws are nothing over and above contingent regularities, differing from accidental generalizations at most in some special epistemic, pragmatic, or logical trappings (like contaning projectible …