Despite decades of controversy, functionalism continues to be a lively research tradition in the social sciences. Furthermore, social scientists who reject Parsonian functionalism and its variants frequently make liberal use of functional explanations. Yet, neither social scientists nor philosophers have reached any consensus on when and where functional accounts are legitimate. Critics doubt that functionalism is really explanatory or testable; advocates are unfortunately vague on their claims and the evidence for them. In what follows I try to clarify what functionalism entails, discuss some ways in which functional hypotheses can be confirmed, and evaluate criticisms. While I think much functionalism is bad science, I hope to show that the problem is not one of principle but of practice.

Social scientists use the term “functionalism” in diverse ways. At times the term refers to a specific theoretical movement in sociology and anthropology—the movement identified with Malinowski and Parsons.

The relationship between the objects of mathematics and physics has been a recurrent source of philosophical debate. Rationalist philosophers can minimize the distance between mathematical and physical domains by appealing to transcendental categories, but then are left with the problem of where to locate those categories ontologically. Empiricists can locate their objects in the material realm, but then have difficulty explaining certain peculiar “transcendental” features of mathematics like the timelessness of its objects and the unfalsifiability of (at least some of) its truths. During the past twenty years, the relationship between mathematics and physics has come to seem particularly problematic, in part because of a strong interest in “naturalized epistemology” among American philosophers. The tendency to construe epistemological relations in causal and materialist terms seems to enforce a sharp distinction between mathematical and physical entities, and makes the former seem at best uncomfortably inaccessible and at worst irrelevant.

Of the many controversial claims in Thomas Kuhn’s The Structure of Scientific Revolutions perhaps none are more troublesome than those made in his account of the role of a paradigm in perception. For if we take it that “a paradigm is prerequisite to perception itself (Kuhn 1970, p. 113) and that “two groups of scientists see different things when they look from the same point in the same direction” (Kuhn 1970, p. 150) then we seem to be burdened with all the familiar problems about whether, and how, experimental evidence could possibly serve to promote the acceptance of a paradigm. And, indeed, these problems have been amply discussed in the last several decades. Dudley Shapere, for instance, has questioned Kuhn’s account of how science comes to move between incommensurable paradigms. (Shapere 1964) Still, Kuhn’s discussion of perception has an empirical foundation in psychological research which gives it a certain resilience in the face of surface problems with the consequences of his account.

Professor Grosholz packs a lot into her interesting and suggestive paper “Formal Unities and Real Individuals” (Grosholz 1990b). In the limited space available I can comment briefly on its several parts, or direct more substantive comments at a single issue. I will opt for the latter; specifically, I want to address her critique of mathematical structuralism, as found especially in the writings of Michael Resnik.

I begin with a brief, hence necessarily caricatured, summary of Resnik’s influential view. According to structuralism, the subject matter of a mathematical theory is a given pattern, or structure, and the objects of the theory are intrinsically unstructured points, or positions, within that pattern. Mathematical objects thus have no identity, and no intrinsic features, outside of the patterns in which they occur. Hence, they cannot be given in isolation but only in their role within an antecedently given pattern, and are distinguishable from one another only in virtue of the relations they bear to one another in the pattern (see, e.g., Resnik (1981)).

In Word and Object, W.V. Quine made thinkable the idea that speech and cognition bear a burden of semantic indeterminacy. On Quine’s account, the upshot of semantic indeterminacy is that meaning and mentalism resist successful naturalization, and thus fail the test of scientific respectibility. For Quine, semantic indeterminacy is a fatal shortcoming.

Recent attempts to naturalize meaning in our thought and our talk (e.g. Dretske 1981, Fodor 1987), belonging to a tradition that has thrived in reaction to Quine, have sought to nail down semantic properties in a scientifically respectable fashion, avoiding the scourge of indeterminacy. However, both Quine and the new redeemers of mentalism may have missed the boat. Three case studies offered here will suggest that a thriving theory of visual perception still makes room for some rather striking indeterminacies.

Our deepest commitments about history are reflected in how we break it down into periods. (Cf. Galison 1988) By drawing a break at a certain point we emphasize the novelty and importance of a new development. It is also how we contain and dismiss certain work as no longer relevant. Thus, in the history of physics we break the story with Newton, both to emphasize his roles in bringing previous developments to a close and in initiating new lines of work, and also to suggest that the ongoing practice of physics thereafter can appropriately in large measure ignore what preceeds Newton. Periodizing history is essential to understanding it, including when we periodize the work of a given writer. In philosophy, anyone who did not see a gulf between Kant’s early work and his critical philosophy or between Wittgenstein’s Tractatus and his Philosophical Investigations would be missing something enormously important.

Does recent work in the cognitive sciences have any implications for theories or methods employed within the philosophy of science itself? The answer to this question depends first on one’s conception of the philosophy of science and then on the nature of work being done in the various different fields comprising the cognitive sciences. For example, one might think of the philosophy of science as being an autonomous discipline that is both logically and epistemologically prior to any empirical inquiry. If the cognitive sciences are empirical sciences, then research in the cognitive sciences could not have any significant implications for the philosophy of science. And that would be the end of the matter. Logical Empiricism is now typically understood as having exemplified this point of view.

More specifically, Logical Empiricism took it for granted (i) that scientific knowledge should be understood as ideally having the structure of a formal logical calculus, and (ii) that the empirical warrant for scientific claims is given by directly observed data together with formal rules which determine the weight of the evidence for or against the particular claims in question.

Bas van Fraassen (1980) maintained that acceptance of a scientific theory does not involve belief that the theory is true, though it does involve belief that the theory is empirically adequate. This provoked a large literature attempting to refute van Fraassen’s position. For the most part, the critics attempted to show that van Fraassen’s position violates principles of inductive inference, or of rationality.2 Such criticisms do not question the possibility of conforming to van Fraassen’s conception of science; they merely argue that there is good reason not to do so.

But some of van Fraassen’s critics have taken a different tack. Simon Blackburn (1984, p. 223) has remarked, and Sam Mitchell (1988) and Paul Horwich (forthcoming) have argued, that acceptance of a theory, as understood by van Fraassen, is the same as believing the theory to be true. If this is right, then what van Fraassen is advocating is a logically contradictory position, and thus one to which it is impossible to conform.

This paper is abstracted from a forthcoming book which defends a particular answer to the question of just what it is that shifts when a paradigm shifts. The claim is that what shifts are habits of mind. And in particular the claim is that the most striking cases of paradigm shift will characteristically turn on a shift in some single, uniquely critical, habit of mind: the barrier. An account of a radical discovery — discovery that prompts the Kuhnian symptoms of incommensurability, so that intuitions that seem irresistible to some seem perverse to their rivals — then characteristically turns on how some individual got past the barrier (escaped the critical habit of mind), while at least for a while others could not do so.

If one compares various skeptical arguments about our perceptual beliefs with arguments against scientific realism one immediately notices important similarities. Skeptical arguments about perceptual beliefs are often based on the premise that all of our perceptual beliefs could be wrong. Our experience is consistent with many different states of affairs; some familiar examples are hallucination, an evil demon, and brains in a vat. Thus it is claimed we have no reason to believe that the perceptual beliefs we normally form are true and alternative beliefs are false.

One of the most powerful arguments against scientific realism is based on the claim that our data underdetermines our theory. There exist empirically equivalent theories, which means that there is no observation that could lead us to accept one and reject the other. Each member of a class of empirically equivalent theories will be equally supported by the evidence.

Recent work by historians and philosophers has called attention to the complexity of modern experimental physics and its intricate relation to interpretive and explanatory theory (Franklin 1987, Hacking 1983, Galison 1987 and 1989, Pickering 1984 and 1987). One point of general agreement has been that the contexts of discovery and justification often significantly overlap and do not uniformly correlate with experiment and theory. Instead, intended applications, instrumentation, experiment, phenomenological description, explicative models, and explanatory theory all act both as incentives and as constraints during the production of scientific knowledge. The present paper investigates how one node of this network operates in the domain of solid state physics and chemistry. In particular, a case study analysis of how models function in transition metal oxide research provides ample support for several conclusions.

First, it must be emphasized that the term “theory” can have misleading connotations in this domain if it is allowed to imply a single and foundational level of explanatory generality.

Contrary to what the name suggests, the recent work of Howard Margolis (1987) and Ronald Giere (1988) demonstrates that the “cognitive turn” in the philosophy of science is not simply the application of cognitive science to the study of science. For one thing, neither one is what Jerry Fodor (1981) has called a “methodological solipsist,” that is, someone who wants to account for thought processes without presupposing an account of the world of which those thoughts are about. Margolis, in fact, comes perilously close to Fodor’s anti-cognitivist foe, J. J. Gibson (1979), whose “ecological” perspective requires that an organism’s thought processes be specified in terms of structures in the environment, “affordances,” capable of satisfying the organism’s desire to know. The Margolian focus on overcoming “barriers” to alternative “habits of mind,” as the psychic basis of Kuhn’s paradigm shifts, only serves to highlight this broadly “functional” side of thinking that typically has no place in cognitive science.

In a recent paper (1989), Ian Hacking has extended his discussion of entity realism, a discussion begun six years ago in the final chapter of Representing and Intervening (1983). This extension allows us to examine for the first time the whole of one impressive attempt to rescue scientific realism from the ever more subtle skepticism of post-positivist thinking (Laudan 1984; Fine 1986). Hacking’s approach complements that of Nancy Cartwright. Like Cartwright, he implies that a full-blown realism about scientific theories and entities can survive neither the facts of history nor the arguments of opponents. Hacking is a realist only about some unobservable entities, those he thinks durable enough to survive most alterations in theory. This is the science that Kuhn said survived even paradigm change.

Electrons are Hacking’s example of a real unobservable entity. Their behavior is so consistent and so well-known that they are used as tools: they are manipulated in the engineering of scientific experiments designed to extend the bounds of the known.

I regret to inform you that Allan Franklin is unable to be here because of the consequences of his collision with a truck in Boulder, Colorado several weeks ago. The three of us have decided to proceed with the symposium in his honor, even though it is now missing its fulcrum. The original point of the symposium was to have an informed discussion of two versions of atomic parity-violation experiments, versions that embody opposed philosophical conceptions of the experiments. The first conception is embodied in Andy Pickering’s account in his Constructing Quarks, an account that is explicitly criticized by Allan Franklin’s more recent discussion in his Experiment, right or wrong (Pickering 1984; Franklin 1990). The symposium would have brought this confrontation into focus, with Franklin’s presentation of his critique followed by Pickering’s rejoinder at this symposium, both of them in sufficient command of the detailed history of the atomic parity-violation experiments to allow for the possibility of a useful exchange of differences of opinion.

While most philosophers of science agree that science is, perhaps in many ways, a highly successful enterprise, there is no consensus about the best way of defining and measuring this success. Philosophers are also divided in their views about two further issues: What does the success of a scientific theory indicate? What is the best way of explaining this success?

After some remarks about institutional and pragmatic measures of success, this paper concentrates on rival ways of defining cognitive success. Four realist measures of epistemic credit are discussed: probability, confirmation (corroboration), expected verisimilitude, and probable verisimilitude. Laudan’s non-realist concept of the problem-solving effectiveness of a theory is compared to Hempel’s notion of systematic power. I argue that such truth-independent concepts alone are insufficient to characterize scientific advance. But if they are used as truth-dependent epistemic utilities, they serve as fallible indicators of the truth or truthlikeness of a theory.

In recent years a unified strategy in dealing with constructivism has been emerging in the writings of historians and philosophers of science. In my own experience, the strategy is exemplified in the long critiques of all or parts of my book, Constructing Quarks (CQ), set out by Paul Roth, Peter Galison and Allan Franklin. These critiques have two common features. First, the substance of constructivist claims is more or less ignored, in favour a fictional version that simply asserts the opposite of what the critic wants to affirm, which is, second, that the evolution of science should be grasped in terms of some relatively simple and unsituated concept of ‘reason’ (or ‘logic’ or ‘persuasive argument’).1 Allan Franklin’s discussion of the history of parity-violation experiments in atomic and high-energy physics exemplifies both of these features.2 Concerning the first, the position he attributes to CQ is summarised as a pure negative: Pickering, he says, ‘obviously doubts that science is a reasonable enterprise based on valid experimental or observational evidence’ (165).3

In 1986-1987, a number of ecologists were involved in a dispute over design of wildlife reserves and species losses resulting from deforestation. The battle was played out largely in the pages of the Ecological Society of America Bulletin. The most focused aspect of the controversy began in August 1986 when P. C. Kangas gave a paper at the meetings of the Fourth International Congress of Ecology, held in Syracuse, New York.

Using data on trees in Costa Rica and the “objective approach” of species-area curves, Kangas argued that extinction rates due to deforestation in the tropics are actually much lower than most other researchers have alleged (Kangas 1986, p. 194). Ecologists in the audience disagreed with Kangas’ claim that extinction rates were indeed lower, complained that his questionable results would be used to justify more rapid and widespread deforestation, and claimed to have “sharply trounced his [Kangas’] arguments” (Noss 1986, pp. 278-279).

Does Allan Franklin’s study of atomic parity-violation experiments provide convincing evidence against social constructivism? According to Franklin (1990a, p. 2), “when questions of theory choice, confirmation, or refutation are raised they are answered on the basis of valid experimental evidence… [and] there are good reasons for belief in the validity of that evidence.” Franklin asserts that social constructivists take the opposite position: “They would say that it is not the experimental results, but rather the social and/or cognitive interests of the scientists, that must be used in the explanation.” Having set up the contrasting positions, he then asks the reader, “which of us is telling the more plausible story?” (Franklin 1990b, p. 163). Consistent with his evidence model, he proposes to discriminate between the two opposed positions by consulting historical evidence. He describes the results of two different sets of experiments on atomic parity-violations, and assesses the extent to which the experimental data match predictions based on the Weinberg-Salam unified theory of electroweak interactions.

In the opening chapters of Progress and Its Problems, Laudan presents a taxonomy of scientific accomplishments that has become very well-known among philosophers of science (Laudan 1977). I wish to point out some important omissions in this taxonomy and to recommend an alternative scheme. I believe that the distinctions I have drawn among scientific tasks are philosophically more interesting than Laudan’s. But I am not prepared to defend this opinion in detail. My chief claim is that the new taxonomy is demonstrably closer to being exhaustive. So far as I know, it is exhaustive, although I would not be greatly surprised to discover oversights. A major benefit of Laudan’s scheme was that it called attention to conceptual (as opposed to empirical) activities in science which had frequently been left out of account in generalizations about the nature of scientific work.