Beyond Positivism and Relativism.
This useful and stimulating book is a collection of thirteen recent essays written over the last decade by a distinguished philosopher of science. Taken collectively, the overall point of the essays is to undermine the central doctrines of the most influential epistemologies of science--postivism and relativism. According to Larry Laudan, these views have not only failed, they have failed for similar reasons, and they can and should be replaced by an epistemology which takes science as a rational enterprise capable of progress. The book is divided into five parts: the first part is an introductory essay which acts also as a stand-alone article, Part Two criticizes the case for "the undetermination of theory by data"; Part Three defends, from relativist critiques, the notion of scientific progress and method. Part Four offers some new thoughts on Laudan's own theory of such matters, which he calls "normative naturalism". Laudan closes with a section on the difference between science, pseudo- science and non-science. Of the essays in the book, two are entirely new and the remainder are generally greatly expanded or revised versions of papers which were previously scattered throughout the literature. Thus, in bringing together the essays in one volume, the book, besides its obvious interest to the specialist, will be useful to those instructors looking for an accessible foil to Kuhn and the positivists to include on their syllabus.
In his new introductory essay "The Sins of the Fathers ... Positivist Origins of Postpositivist Relativisms", Laudan argues for one of the main themes of the book, namely, that the Kuhnian relativism which replaced postivism shares some of its central tenets with the earlier theory. Among these commonalities are the identification of theories with languages, a commitment to some form of the undetermination thesis, and an embracing of what Laudan calls "the thesis of methodological subjectivity". Methodology, scientific or otherwise, has to do with norms and prescriptions. As Laudan reminds us, the positivists were centrally committed, via a sharply drawn fact/value distinction, to non-cognitivism about norms, and hence about methodology itself. For the positivist, there could be no fact of the matter about whether one methodological standard was better than another. All such standards could only be conventions or, as Carnap had it, "proposals". Laudan argues that it was an implicit commitment to this underlying metaphysic of values which led Kuhn and Feyerabend into relativism: for if methodology was a non-factual matter, then obviously methodological disputes were incapable of being brought to rational closure--there was no fact of the matter as to whether one paradigm's standards for problem solving were to be preferred over another (p. 16). Hence, according to Laudan, "Kuhn's relativism about methodological matters emerges directly out of his acceptance of a sharply drawn fact/value distinction, coupled with a conviction that value differences are not adjudicable by facts" (p. 16).
Laudan's supposition that Kuhn was a closet non-cognitivist is appealing: it certainly explains, both philosophically and historically, why Kuhn would infer from the fact that there are disagreements between scientists on methodological issues, that such disputes do not admit of rational closure. But despite its plausibility, I am not convinced that Kuhn is best described as a tacit non-cognitivist about methodological issues. For to say that Kuhn is a non-cognitivist about methodological standards is to say that he is committed to holding that statements by individual scientists of the form "theory x is better than theory y with respect to s" are literally without truth-value. But this seems at odds with understanding Kuhn as a relativist, according to which Kuhn is saying that such statements can only be assessed as true or false relative to, or from the standpoint of, a paradigm. On this view, the reason conflicting methodological evaluations are incommensurable is not that they are expressions of attitudes, but that there is no paradigm-neutral standpoint from which to adjudicate, or even understand, the dispute. As Kuhn notes, it is because of this fact that a transition from one scientific theory to another will not be one step at a time, but all at once; a conversion experience, in other words (The Structure of Scientific Revolutions, 2nd ed., Chicago, IL: University of Chicago Press, 1970, p. 150). I take this to imply that a transition from endorsing one theory to endorsing another is subjective or emotive in regard to the explanation for its occurrence, not because all endorsements of the form "x is better than y with respect to m" are non-factual. According to this interpretation, cognitive judgments about theory evaluation must always be relative; according to the non-cognitivist interpretation, they would not be judgments at all.
As noted, Laudan's purpose in these essays is not simply critical; he is also concerned to demonstrate that a rational view of science can be developed which does not fall prey to the positivist assumptions which undermined relativism. In particular, Laudan is concerned in Part Four of the book to argue for what he has called his "normative naturalism" about scientific methodology. Normative Naturalism is a meta-methodology: its goal is to explain how methodological rules can themselves be justified. According to Laudan, rules of scientific methodology are neither conventions nor self-justifying intuitions; instead a methodological rule is a "hypothetical imperative", in the sense that they are means to desired ends. Thus, a methodological rule such as Popper's "avoid ad hoc hypotheses", according to Laudan, is best represented as "If your cognitive goal is to have falsifiable theories, avoid ad hoc hypotheses". Hence, rules are "normative" since they tell us what (given certain ends) we ought to do. Yet at the same time, Laudan argues that such rules are open to empirical verification, since their acceptability will rest in part, at least, on the empirical question of whether (in the case of the above example) avoiding ad hoc hypotheses actually does lead to falsifiable theories. However, and as Laudan acknowledges, we not only want our methodological means/ends rules to be justified by the empirical evidence, we want them to be aimed at the right or proper goals. Thus, in evaluating whether rule M is justified, we not only need to assess whether the means in question reliably produce the ends, we must evaluate the ends themselves. In general, we need a way of picking out "good" ends from "bad". As Laudan candidly admits, such an axiological investigation is extremely difficult (p. 140). One of those difficulties involves the justification of the axiological component of a methodological rule. Laudan is quite clear that the project of assessing methodological rules is one which is not different in kind from scientific inquiry; they, are, as he says, "a part of empirical knowledge" (p. 133). If so, then presumably, in assessing whether one cognitive aim (simplicity, falsifiability, coherence, explanation of previous theories, etc.) is better than another, I will be abiding by one or another "means/ends" methodological rule. Thus, since any such rule has its aim, I shall always be presupposing some cognitive aim in assessing any cognitive aim. This would suggest that there is no objective standpoint from which to assess methodologies, and casts doubt on Laudan's attempt to provide an objective answer to what he recognizes as the relativist's legitimate questioning of the justification of those methodologies (p. 166).
Let me state again that, the above quibbles aside, this is an extremely useful and excellent book. Laudan's writing style, as usual, is crisp and, as a consequence, will be accessible to a wider audience than the typical book written on these subjects. These papers are chock-full of arguments, criticisms and ideas that will interest anyone working in the general area of philosophy of science. Finally, in reading the essays therein, one gradually forms an impression that they are indeed all contributing to the shape of a particular philosophical view, and yet it is a view which continually resists neat labeling as realist, relativist, positivist, or even naturalist. Laudan's work goes beyond the usual "isms"--a feature which more philosophers need to imitate.
Department of Philosophy and Religion University of Mississippi University, MS 38655 USA