This review appeared in Book Reviews – Isis, 78 : 3 : 293 (1987), pp. 438-9.

Norman Swartz. The Concept of Physical Law. xi + 220 pp., bibl., index. Cambridge/London/New York: Cambridge University Press, 1985. $29.95.

Norman Swartz begins in the conventional terrain of philosophy of science – the relationship between knowledge and reality. However, whereas most assume a lawlike reality and examine the conditions and status of lawlike knowledge, Swartz focuses on the lawfulness or otherwise of' reality. As he puts it, this is an essay in the metaphysics of science, not its epistemology. The main target of the argument is necessitarianism, the belief that physical laws are the way they are out of fundamental necessity and that they hold this nomological grip upon nature. The notion that things behave in a certain way because they have to according to scientific law awards more status to scientific knowledge than the belief' that things behave that way because that is how they have been observed regularly to behave. This regularist alternative to the necessitarian metaphysic is trenchantly and engagingly argued by Swartz.

The starting point is a strong distinction between physical law and scientific law. Physical laws are many – as many, argues Swartz, as all available descriptions of nature. Scientific laws, on the other hand, are used, and therefore they have to be useful condensations, simplifications, and (thus often) even falsifications of the innumerable physical laws available. Physical laws are singular descriptions of events and states. This weak formulation of physical law is the essence of the regularist view that laws are observed regularities, nothing more. This appears to do away with the problem of free will and determinism at a stroke, as a visiting Martian regularist explains on Swartz's behalf. There is no way of knowing that behavior is necessary, even if we have an apparently complete framework of laws. There is only an observed regularity of behavior. If next time it is different, then we add a new, hitherto unknown law. It is central to this case that all data are "gappy," that there are many laws consistent with the data, that each has its own ways of filling the gaps, and that each law is possible in some world or another. We have no cast-iron way of knowing which one of these worlds we are in; therefore we cannot say which way of filling in the data gaps is the way for our world.

At this point the argument seems to have returned to the familiar debate about the criteria for evaluating which among competing theories or frameworks of interpretation is the most capacious, powerful, and so forth. However, Swartz sweeps this aside as the stuff of pragmatism and science – not unimportant, but beside the point. Scientists may choose their laws for a variety of purposes, but nothing in this justifies the implicitly authoritarian standpoint that the chosen scientific laws somehow impose a necessity upon the behavior of nature thereafter. There is a brief but useful discussion of Nicholas Rescher's somewhat heretical defense of necessitarianism, in which the nomological necessity of laws is located in the uses of those laws by scientists, and not in reality itself. Rescher's point that coherence with other laws is a deciding principle is no redemption because this itself is an imprecise rule: it cannot select the singular law that can be claimed to embody nomological necessity. Of course we can say that all cats have to have vertebra, by the way we use the concept of cat (one is reminded of Imre Lakatos or Leonhard Euler and polyhedra). But this is hardly what necessitarianism in its red-blooded form is about, and by this time the arbitrary authoritarianism of the whole thing is nicely exposed.

It is a pity that despite the engaging style and the crispness of the argument, the book rather stops in midair, with the slightly unsatisfactory feeling that all that has been disposed of are a few dusty academic theories. But there are some immense practical implications in the realm of social psychology of science, and perhaps its politics, which could at least have been given an airing. Perhaps scientists, al least in their modern context, can only develop knowledge by imposing a false necessity on nature and granting a false metaphysical power to their current cognitive commitments? Could they continue to concentrate faithfully on the job in a context of radical agnostic pluralism of the sort implied by Swartz? Is this innate necessity inspired by the drive for domination and manipulation of nature that is so much a part of modern knowledge and its institutional contexts? These are of course big questions, and anything but a brief exposure might have confused a tightly argued philosophical thesis. Perhaps they are for the next book.