Philip Kitcher's book begins with a familiar historical overview. In the 1940s and 50s a confident, optimistic vision of science was widely shared by philosophers and historians of science. The goal of science was to discover the truth about nature, and over the centuries science had advanced steadily towards that goal; science discerned the real kinds of things of which the world was composed and the causal relations between them; the methods of science were rational and its deliverances objective; and so on. Only where science failed in some of these respects was there any need to provide external, that is social, political, or individual, explanations of scientific belief. In the late 50s, and especially subsequent to Kuhn's classic, The Structure of Scientific Revolutions, all this started to change. Historians increasingly insisted that the development of science must be treated just like any other cultural process, which meant embedding the narrative of the growth of scientific knowledge fully in the social and political context in which it occurred. This suggested that the view of science as deriving from a uniquely rational process could no longer be sustained. And, notoriously, Kuhn argued that science could not be seen as cumulative across the most dramatic changes in scientific theory. Though philosophers have never given up entirely on the old optimistic picture ("Legend" as Kitcher refers to it), its influence has steadily waned. For various reasons philosophers have become increasingly concerned over whether one could believe scientific claims to be literally true. And influenced by Duhem's thesis, revived by Quine, that scientific theory must always be underdetermined by empirical evidence, they became more sympathetic to the possibility that scientific belief must be explained, at least in substantial part, by much more than the rational objective processes envisioned by Legend. These philosophical doubts existed in uneasy tension with more extreme tendencies toward thoroughgoing relativism or skepticism, and with the movement in the sociology of science to see the whole concern with truth and falsity as an irrelevant diversion.
Kitcher's aim in The Advancement of Science is no less than the restoration of the central claims of Legend. He argues that science does indeed discover the truth about nature, that it discerns the real kinds and their causal relations, that it does so progressively and cumulatively, and that it progresses through methods that are rational and ultimately objective. Of course, Kitcher does not simply restate the doctrines of, say, Rudolf Carnap. The importance of his book is that it attempts to reassert these basic themes in a way responsive to the subsequent intellectual developments that have cast such doubt on them. First, Kitcher identifies many of the problems that arose for Legend's view of science as resulting from a too simple conception of what science consisted in. Instead of describing scientists simply in terms of what they believed, Kitcher introduces the notion of a scientific practice, which includes a scientist's professional language; the questions she identifies as significant; the statements she accepts; the patterns of explanation, or schemata, she accepts; the examples of, and criteria for, credible informants; the paradigms of experimentation and observation, including the instruments that she considers reliable; and the exemplars of good and faulty scientific reasoning (p.74). Of particular note is the idea of schemata for explanation, which Kitcher uses to describe the growing explanatory resources of science. The various developing practices of individual scientists feed into what is the real locus of scientific progress, an advancing consensus practice. A further major respect in which Kitcher diverges from the classical tradition is in emphasizing that scientific practices develop in response to encounters both with nature and with colleagues.
If there is one point on which substantial consensus has been reached in the post-Legend era, it is that claims about the progressiveness of science need to be sensitive to the real history of science. The most impressive aspect of Kitcher's book is the seriousness with which he responds to this demand. The central focus of his argument is the detailed treatment of three key episodes in the history of science, Galileo's defence of heliocentrism, Lavoisier's refutation of phlogiston chemistry, and in greatest detail, Darwin's argument for evolution by natural selection. Historians in a Kuhnian tradition often take these crucial scientific transitions to be paradigms of non-rational change. The conception, and even perception, of the world, it is often argued, were so deeply embedded in the earlier views that their rejection could in principle derive only from extra-rational grounds. But Kitcher argues forcefully that such claims are misguided. Galileo, Lavoisier, and Darwin triumphed in the end because they had the better arguments. And except for a few scientists who were either too strongly committed to the old views for personal or external reasons, or just plain stupid, the scientific community eventually came to see that they had the better arguments. The reason they had the better arguments, moreover, was that what they claimed was true--or at least approximately so. I shall not attempt to predict how historians will respond to the details of Kitcher's accounts. But there is no doubt that the historical narratives are detailed and plausible--far removed from the "rational reconstructions" characteristic of the Legend era. And there is also no doubt that this is the kind of work that is needed if the historical arguments against rationality and progress in science are to be rebutted.
The achievements so far discussed will surely establish The Advancement of Science as an essential text for anyone concerned with the fundamental issues it addresses. Let me now turn to some reservations I have about Kitcher's project. For various reasons the correct account of the major scientific advances just discussed is of great significance. Most philosophers, myself included, believe that Galileo, Lavoisier and Darwin advanced our understanding of nature, but it is not always easy to articulate this belief. The significance of these episodes is that they constitute advance if anything does: they are the crown jewels of scientific history. So that while they are natural targets for the skeptics and cavillers who deny that science ever progresses, it is much less clear what positive conclusions follow from Kitcher's defence of their rationality and progressivity. Though Kitcher remarks on occasion that the work is preliminary, and insists that much more historical work of the kind he exemplifies is needed, I think it is clear that he takes himself to have provided a preliminary argument that science is rational and progressive. But it seems to me that the demonstrationss that a few of the most celebrated episodes in scientific history do, indeed, constitute major advances in our understanding of nature, important though it is, falls far short of establishing this point. For the argument does little to identify what constitutes these episodes as science, and therefore does little to tell us where, or how often, we should expect to find such rational progressive projects occurring. (It is clear from Kitcher's earlier work on sociobiology that he does not think that anything that counts sociologically as science is rational and progressive.)
I imagine that Kitcher takes the long final chapter of his book to provide the beginnings of an answer. I shall approach this chapter by way of my deepest disagreement with Kitcher. Much of the truth Kitcher sees science as providing is in the form of a pervasive underlying mathematical structure to nature. Thus after outlining the explanatory schema for biological phenomena that Darwin introduced, Kitcher leads us through a series of subsequent, putatively progressive developments of that schema. But whereas the schemata that Kitcher offers for Darwin's contribution can readily be filled out with selective explanations for the presence of organismic traits, the general schema he offers for the mathematical models of contemporary population genetics provides us with a kind of explanation that could only be applied in the most degenerately simple cases. ("The fitness of the allelic combination a11a11...an1 an1 is w1 ...1 , the fitness of the allelic combination a12a11a21a21 ...an1an1 is w21...1, ... {continued through all allelic combinations}" (pp.46-7) strikes me as not merely the sort of thing we will never know but the sort of thing about which we have not the slightest reason to suppose there is in general any fact of the matter.)
Kitcher's weakness for the mathematical model-building aspects of science--something which is hardly sustained by his defence of those parts of scientific history he addresses--is displayed most strongly in his final chapter. Here he addresses the question of what social organization of science will be most conducive to the rational deployment of scientific resources. The argument is carried on almost entirely in terms of extremely abstract and idealized formal models of scientific work. At the risk of sounding like a Luddite, I am convinced that if there are things worth saying here, they can be said in English, with the same clarity and accessibility that is characteristic of the rest of Kitcher's book. While I entirely agree with Kitcher that here are matters of public policy to which philosophy of science should be able to contribute, it seems to me that far too much of public policy is already controlled by so-called experts hiding behind highly technical and impenetrable jargon. Kitcher remarks in his introduction that "philosophical reflections stand in relation to the complex practice of science much as economic theory does to the complicated and messy world of transactions of work, money, and goods" (p.10). As a philosopher of science who sees economic theory as one of the more dismal empirical failures in the history of science, I find this remark highly discouraging. Fortunately, only the final chapter of Kitcher's book seems to me to be blighted by this vision.
As will be apparent, I am more sympathetic to some parts of the project of The Advancement of Science than I am to others. However, I should stress in conclusion that the book is full of insightful and important ideas and analysis, and is written with the clarity and force of argument that readers of Kitcher's earlier works will expect. There can be no doubt that it constitutes a significant contribution to fundamental issues in the philosophy of science.