American Scientist

Human Nature and the Limits of Science. John Dupré. x + 201 pp. Clarendon Press, 2001. $24.95.

How should we study human beings scientifically? You might think that in principle we could just use the methods we have developed to study the rest of nature. But John Dupré thinks that those methods, which he regards as committed to reductionism and mechanism, will seriously distort our understanding of ourselves if we employ them in the human sciences—but not because we stand above or apart from the world of nature. Rather, the failure of the mechanistic picture to apply to people reflects a general failure of the universe to conform to the unified, reductionistic model of its workings that many scientists have bought into. Moreover, alongside the disunity of nature we find the disunity of science: According to Dupré, there is no such thing as the scientific method. He should go a little further and argue that "science" is in fact just an honorific term that we reserve for forms of inquiry that establish especially reliable bodies of knowledge. It does not name a method, and the epistemic virtues that all sciences share are also exhibited by other intellectual activities.

In Human Nature and the Limits of Science, Dupré advocates seeing the universe, not as a law-governed machine in which the behavior of the very large depends on the behavior of the very small, but as being made up of a wide variety of phenomena related in diverse and complicated ways; at every level of the hierarchy are distinctive causal powers that do not depend on the behavior of lower-level entities. This position is familiar to readers of Dupré's earlier work; not all the details are provided here, but enough argument is given to permit someone new to Dupré to get a sense of the care and virtuosity with which he has worked out his view. He puts this overall picture to compelling use in the philosophically most interesting chapter of the book, which presents a fascinating and attractive take on the venerable problem of free will.

This book presents a metaphysics and a philosophy of science, along with critiques of evolutionary psychology and rational choice theory (an economistic perspective that conceives of decision-making as "estimating which action will generate the largest expected excess of benefits over costs to the agent"). You might think that treating all these topics satisfactorily would require a very long book indeed, but Human Nature and the Limits of Science is very short. Brevity is almost never a vice in academic writing, but this book is an exception, because it presents a blend of global, large-scale theories and local, small-scale theories without coming up with middle-range theories telling us how the two are connected. In consequence, the reader may end up, like me, agreeing with nearly everything in the book but unsure quite what to make of it all.

Dupré attacks both rational choice theory and evolutionary psychology, in part because they exhibit the shortcomings of what he calls scientific imperialism ("the tendency to push a good scientific idea far beyond the domain in which it was originally introduced, and often far beyond the domain in which it can provide much illumination"). He is also worried about a possible union of the two in which evolutionary psychology would be used to explain where we get the preferences that rational choice theory just takes as given for its purposes—a synthesis he believes would lead to both intellectual and political disaster. This view would have been strengthened by a discussion of the assumption, made by some rational choice theorists, that people share the same preferences and differ only with respect to their opportunities. However, Dupré makes no serious attempt to show how the union might work; in particular, he does not deal with the question of whether an evolutionary picture of humanity implies not just that we have preferences but that we are the relentlessly maximizing creatures that rational choice theory requires us to be.

The criticisms of rational choice here are not new (as Dupré acknowledges); they draw on existing critiques within philosophy and the social sciences. Dupré's moral is that rational choice theory errs because of its methodological individualism—what Dupré thinks is "the erroneous assumption that social phenomena are no more than the aggregates of individual behaviour." But it is quite possible to endorse Dupré's reservations about rational choice—and his view that a correct general theory of human behavior requires acknowledging the importance of norms—while at the same time regarding methodological individualism as trivially true. Dupré correctly points out that we are in large part the product of our cultures. But this does not threaten methodological individualism unless it can be shown that culture does more than shape beliefs and constrain actions. What is needed is a general claim about the metaphysics of humans to connect the global theory with the local one.

The problem arises the other way around, as it were, in the discussion of evolutionary psychology. Dupré argues that the brain cannot contain symbolic representations and that the diversity of human behavior causes problems for any view of the mind as "modular"—that is, as containing special-purpose computers with proprietary information about the world. Suppose these claims are correct: The first undermines nearly all, and the second large parts of, contemporary cognitive science. Here we have local claims that do seem to have wider significance, but Dupré does not address the wider issues, so we are left wondering whether he believes that any cognitive psychology can work at all.

All in all, it is better to read this book for the excellence of its particular arguments rather than to look for one overarching argument, since I do not think that the philosophy of the human sciences here either follows from or supports the overall metaphysics. The book has abundant merits, but its components do not cohere. Perhaps that is appropriate in a philosopher who stresses how messy and piecemeal everything is.-Dominic Murphy, Philosophy, California Institute of Technology