What is a law of nature? Traditionally, philosophical discussion of this question has been dominated by two prominent alternatives; David Lewis's best-systems analysis, according to which a law is a regularity that serves as a theorem in our best axiomatization of the facts about the world, and the Dretske-Armstrong-Tooley analysis, which incorporates universals to distinguish laws from mere accidental generalizations. Marc Lange's first book presents a provocative alternative to this tradition, providing a novel treatment of natural laws that should be of interest to those philosophers concerned with the analysis of lawhood, physical necessity, causation, inductive confirmation, counterfactual analysis, and explanation.

The novelty of Natural Laws in Scientific Practice owes, in large part, to Lange's advancement of a deflationary account of natural law that focuses on the functional role natural laws play in scientific practice. Taking the view that, in scientific practice, laws differ from mere accidental generalizations primarily in the manner in which they support counterfactuals, possess explanatory power, and are confirmed by their positive instances, Lange sets as his goal to “see what can be learned about what natural laws are solely from what they must be in order to function as they do in science” (10). The resulting account can be, most efficiently, divided in two; a formal, descriptive account of lawhood in terms of the notion of counterfactual stability and a normative account of lawhood in terms of what he calls the root commitment—the commitment we make when we believe that a certain statement expresses a law of nature.

Lange's formal account of laws begins with the common intuition that laws govern not simply what actually occurs in the world, but also what would have occurred had circumstances been different. That is, Lange's account is built on the ability of laws to support counterfactuals. Lange rejects the notion that laws are to be identified by their invariance under a certain range of counterfactual possibilities, correctly noting that certain accidental generalizations may remain similarly invariant. In place of the concept of invariance, Lange employs a concept he terms preservation (chap. 2). Briefly, preservation refers to a law's remaining true under all physically possible counterfactual antecedents; where physically possible counterfactual antecedents are those counterfactual possibilities consistent with the laws themselves. To the concept of preservation, Lange adds a concept he calls stability (99–107). Stability concerns sets of statements. A set of statements is stable just in case each member of the set is preserved under all counterfactual possibilities consistent with the members of the set. Now, according to Lange there are trivially stable sets and non-trivially stable sets. As an example of a set possessing trivial stability, consider the set of logical truths. This set is stable because each member of the set, being a logical truth, is trivially preserved under all possibilities consistent with the set of logical truths. On the other hand, Lange demonstrates that there is only one set that possess non-trivial stability—the set of laws (101–03). The uniqueness of the set of laws, Lange argues, accounts for the difference between laws and accidental generalizations; no set of accidental generalizations possesses non-trivial stability and were an accidental generalization added to the set of laws, the set would no longer be stable.

Lange's normative account of natural laws takes as its opposition the view that “we describe no fact by calling ‘All Fs are G’ a law—beyond what we describe by deeming ‘All Fs are G’ true: a certain uniformity of nature” (23). For Lange, there is a root commitment associated with believing a statement to be a law that is absent when only believing a statement to be (possibly accidentally) true; a commitment to believing that the statement has been confirmed inductively (chap. 7). Lange identifies inductive confirmation with what he terms inductive strategies, free moves a scientist makes when attempting to correlate what might otherwise seem to be an arbitrary collection of data and make each piece of data confirmationally relevant to the others. Inductive strategies effectively account for the practices a scientist engages at the very onset of empirical investigation. In undertaking an inductive strategy, the limits of projection for a given hypothesis are fixed by the hypothesis itself along with other hypotheses science is committed to confirming inductively. As such, laws, being inductively confirmed hypotheses, are generalizations that apply to counterfactual states of affairs and not just actual unexamined states of affairs. This serves again to distinguish laws form mere accidental generalizations; but more importantly, it is this unique aspect possessed by natural laws that allows laws to “carry out the special functions that science calls upon them to perform” (9). In particular, a root commitment secures an entitlement to use a hypothesis to license certain inferences and to provide explanations and predictions that range beyond past observation.

The brief sketch just provided fails to do justice to the concert of complexly layered arguments Lange provides in support of his account. These arguments are presented, for the most part, in a manner that should be readily accessible to most readers; the one exception being the rather technically advanced nature of the formal account of laws in terms of counterfactual stability. The exposition, further, displays a keen awareness and penetrating analysis of the philosophical tradition to which the book is addressed. As such, it serves not only to present an alternative conception of natural laws, but offers a fresh perspective on previous accounts. On the negative side, the reader will be left wondering about two startling omissions from Lange's presentation. First, Lange nowhere provides an account of the truth or correctness of counterfactual conditionals. As counterfactuals factor critically in the development of the formal account of natural laws, such an omission leaves the account rather unfinished. Second, one feature of Lange's account of natural laws is that there is a difference between the expressions ‘It is a law that all Fs are G’ and ‘All Fs are G’; the former may be true even if the latter is false. As Lange writes, “[t]he only regularity in which we must believe in believing in some law…is that by applying the law in the course of taking certain kinds of inferential routes to solving certain kinds of problems posed in certain kinds of contexts, we reach conclusions that often enough are accurate for our purposes” (176). This raises questions regarding the relationship to be understood between instrumental value and truth that are, for the most part, left unanswered in Lange's presentation. However, such omissions rather than detracting from the philosophical importance of Lange's work, should serve only to inspire critical response and philosophical discussion of this important contribution to the literature on natural laws.