This Critical Notice appeared in Philosophy and Phenomenological Research, Vol. 57, No. 4 (Dec., 1997), pp. 971-973


Laws of Nature

Cambridge: Cambridge University Press, 1994

Pp. ix + 200.

NORMAN SWARTZ  Simon Fraser University

Nature it would seem, judged on the evidence of this book, is the province solely of physics. What few examples of law there are in this book, are all drawn from physics (or, better, pseudo-physics – more on this in a moment). One will look in vain for any discussion of laws of chemistry and molecular biology, let alone of psychology, sociology, etc. This focus on physics, alone, while resolute, is quite unacknowledged and hence is undefended.

     Even with its exclusive focus on laws of physics, the raw (preanalytic) data – what Carnap used to call "the explicanda" – are remarkably scant. There are two reasons for this.

     Early in chapter one, Carroll reveals (pp. 22-23) his acquaintance with the distinction that I and others have made, viz. between the actual laws of Nature (mostly unknown) and the laws (nearly all false) that scientists use in doing science, in explaining, predicting, controlling, testing, probing, etc., the world. But Carroll dismisses the import of this distinction: "In our daily inquiries, we take ourselves to be seeking, and sometimes finding, truth. It would be surprising if scientists – our most revered investigators – sought less. So, to the extent that laws are one object of scientific discovery, it is natural to think that laws must be true" (p. 22). Which laws: those of nature itself or those of scientific practice? Carroll's answer is strange, but it does accurately foreshadow the rest of the book: "… for expository reasons, I frequently rely on simple and familiar generalizations from the history of science (or even simpler, wholly fictitious examples) that are no longer (or perhaps never were) believed to be true" (pp. 22-23).

     There is a second reason (p. 153) for his eschewing the actual findings and practice of science: "I am a philosopher and not a scientist."[Note 1] And so Carroll invents (i.e. constructs perfectly fictitious) laws – usually variations on a theme of X-particles having the property spin up or spin down in a 'Y-field'. It is these fictitious laws – always, incidentally, devoid of reference to 'theoretical entities' – that then form the test-cases, examples, refutations of analyses, etc., in much of the book.[Note 2]

     These naive fabrications, isolated from scientific theory, are claimed to have – for Carroll anyway – a distinctive lawlike 'feel' (my characterization) such that if they were true, they would be laws. This is reconstructionism with a vengeance, and – I would argue – a latent apriorism: it simply ignores all the subtleties that actual scientists bring to their work.

     How cogent a philosophical methodology is this reliance on invented examples? I am reminded of the story (probably apocryphal) of the man who set out to do ethics using as his touchstone such 'obviously certain' ethical principles as "No man should ever wed his brother's widow". I remain dubious of the value of an exercise that would analyze the concept of a law of nature using, almost exclusively, fictitious examples. If one must – time and again – invoke fictitious examples whose lawfulness is self-evident to the author, I take that to be powerful reason to believe: (1) that science not only does not provide (many) laws of nature, science – more strongly – does not need laws of nature; and (2) that there probably are no more than a handful of laws of nature (of the sort Carroll seeks), i.e. that he is pursuing a will-o'-the-wisp.

     Nearly half the book is devoted to Carroll's arguing what laws of nature are not. He rejects necessitarian accounts (pp. 23-25); he rejects naive regularity accounts (as being both too weak [pp. 31-36] and too strong [pp. 36-40]); he rejects inductivist accounts (pp. 41-45); he rejects systems accounts (pp. 45-55); and he rejects supervenience accounts (pp. 57-85).

     We come, then, to chapter four, Carroll's positive theory of laws of nature. Earlier, in chapter one, Carroll explained that he proposes neither to define, nor to reduce, nor to offer a supervenience account of lawfulness. Instead he proposes to insist upon and explore the nomic aspects of lawfulness. Nomicity marks a vague family of interrelated (but not quite interdefinable) concepts, most especially (read "centrally") that of the counterfactual conditional, but also including such notions as lawhood, causation, lawful sufficiency, production, nonaccidentality, explanation, disposition, etc. (p. 7).

     With the web of the nomic drawn so wide, it is hardly surprising that Carroll (repeatedly) claims: "if there were no laws … in an important sense, there wouldn't be any things" (e.g. p. 9).

     Most writers who invoke the concept of the nomic tie the concept to a kind of nonlogical necessity. But Carroll eschews necessitarianism. Indeed – so far as I can tell – he offers no metaphysical account of nomicity whatever, contenting himself with (1) exploring the relationships between various concepts in the nomic-family; and (2) arguing that – contra Hume – there can be, and more strongly are, grounds in sensory experience for justified beliefs in the nomic.

     This latter thesis is the capstone of the book. Seeking a grounds in perception for a warrant in the belief of nomicity has for decades been 'the Holy Grail' of the anti-Humeans. Like a number of earlier philosophers, Carroll tries his hand at grabbing this elusive prize.

     After some remarkably good preliminaries, discussing both analogies and disanalogies with the familiar problems of phenomenalism and inference-to-the-best-explanation, Carroll states his goal:

… most knowledge involving lawhood arises either from the testimony of others or via a deductive inference from some prior knowledge already involving lawhood. … What worries the Humeans is how knowledge of a proposition that it is a law arises from scratch. (p. 113)

I will have to précis the preliminaries of Carroll's case study (p. 114). As usual, it involves the fictitious X-particle subjected to a Y-field. The experimenter, Jones, repeats the experiment many times, with a variety of X-particles, varying other factors, e.g. the number of particles, their immediate history, etc. In all cases she finds that the examined X-particles have spin up in a Y-field. On the basis of enumerative induction, the variety of the surrounding factors, simplicity, (predictive) strength, and consistency with everything else she knows, Jones believes L1 [i.e. that X-particles subjected to a Y-field have spin up]. Now I quote:

At some point in her experiments, perhaps concurrently with her coming to believe L1, Jones forms some counterfactual beliefs. … [E.g.] Jones believes that the next X-particle in the sample – particle cwould have spin up if it were subject to a Y-field. … Jones's counterfactual belief about particle c together with her other beliefs make plausible other counterfactual beliefs. … As a result, Jones forms the belief that L1 is a law of nature. … Assuming that L1 really is a law and that the situation is otherwise normal, it is plausible to think that L1 is a law. … [In this example,] Jones gains the belief that L1 is a law from scratch. (pp. 114-16).

With the exception of the penultimate quoted sentence "Assuming that L1 really is a law …", I find little or nothing to object to in this passage. As a description of how some of science works (viz. the minor part that is generalized from sensory experience), there is not much to criticize. But that is the very difficulty. How, exactly, is any of this supposed to show the falsity of Humeanism? Even died-in-the-wool Humeans (myself for example) are unlikely to take exception to the fact that all of us do advance counterfactuals sometimes on the basis of observed regularities. Was that ever an issue? I can't recall it ever having been in dispute. I find it difficult to discern just what problem Carroll is supposing himself to be solving, especially when the antecedent of the final conditional statement of the argument reads "Assuming that L1 really is a law …". At the end of this book, I find myself as little enlightened as to what Carroll supposes a law to be as I was at the outset.

     Chapter five deals with causation. Carroll's thesis is: "With regard to our total conceptual apparatus, causation is at the center of the center" (p. 118). I wish Carroll would explain just who it is that is being alluded to by the pronominal adjective "our" in "our total conceptual apparatus": the bulk of humanity (including children)? [recall Piaget's work with children's concepts of causality]; the bulk of adult humanity?; scientists?; physicists?; 20th-century physicists?; late-20th-century physicists?; late-20th-century quantum-mechanical physicists?; or …? It makes a difference. It is by no means clear, indeed it has been vigorously challenged by scores of philosophers, that there is anything whatever that can be called "the causal relation" (p. 123, my emphasis). Readers will have to decide for themselves how much such an approach illuminates the concept(s) of causality, even when restricted to physics. (Do not look for, or expect, any discussion of such well-known accounts as those of Mackie or Salmon in this chapter.)

     Some parts of this book are done very well, e.g. the discussions of inference and phenomenalism. And many parts (e.g. that dealing with nomicity-without-necessity) give evidence of the promise of a profitable research program. What will be needed in pursuing the work begun here will be considerably more emphasis on the empirical – on the real practice of science and the propositions that scientists actually use in their work – and less on the a priori and the fictitious.


  1. On one occasion when he tries to invoke an actual concept, viz. that of solidity (p. 9), he gets the physics wrong, suggesting obliquely that the solidity of material objects is to be accounted for by the solidity of the constituent particles. No mention is made of strong inter-electronic forces between those constituents.[ Resume ]
  2. Does "Y-field" refer to a theoretical entity? The question is never raised.[ Resume ]

Return/transfer to Norman Swartz's Philosophical Notes

Return/transfer to Norman Swartz's Home Page