In this paper I will propose and defend a view of rationality which bears on three important issues in the philosophy of the social sciences. First, there is the issue of the nature of the principle of rationality in the social sciences. When the principle of rationality is formulated as an interesting empirical generalization, it seems false; but when the principle of rationality is formulated in a way which is clearly defensible and true, it seems trivially tautologous ([A],pp. 172-178). Secondly, there is the issue of the criteria for a rational action. When the criteria for a rational action are made relative to a culture, there is an implausibly large number of standards for judging rationality; but when the criteria for a rational action are not made relative to a culture, there is an implausibly small number of rational actions ([5], pp. 94-111).

My aim in this necessarily compact paper is to sketch a philosophical defense of the kind of position assumed by the later Kroeber, when he wrote:

It has become commonplace to claim that a scientific revolution has taken place in linguistics as a result of Noam Chomsky's contributions to the theories of syntax, linguistic metatheory and the philosophy of mind. Although the label ‘revolutionary’ was applied to these developments almost from the start (cf., [19], p. 375), the cliche has also been encouraged by the appearance of Thomas Kuhn's The Structure of Scientific Revolutions in the early sixties at about the same time that the theory of transformational-generative grammar (hereafter abbreviated TG) was consolidating its position as the dominant school of linguistics and attracting the attention of the wider intellectual and scientific community. Because of this coincidence of currency, it was perhaps inevitable that many of those who were aware of the developments which had been taking place in linguistics would find themselves reading Chomsky's revolution into the pages of Kuhn's book.

The philosophical literature is heavy with discussions of the logical structure of various proposed schema of functional explanation (most of which are modifications of Hempel's schema [7]) but it is very light on detailed analysis of the way in which these schema actually work or the way in which they apply to the instances in which functional explanations are used by scientists. These papers use statements by scientists ascribing particular functions to particular traits, along with intuitions about the way in which ‘function’ is used in ordinary language, to ascertain the referents of the term and thus to determine the class of uses against which any schema should be tested; but such examples only show whether the schema under discussion refer to the same things as those the scientists’ refer to; they do not show whether the logical structure of the schema really captures the logical structure underlying the scientists’ use.

A scientific explanation is prima facie micro-reductive if the explanans refers to the behavior or properties of the micro-parts of something mentioned in the explanandum. Entity x is appropriately regarded as a micro-part of y whenever the following individually necessary and jointly sufficient conditions are satisfied:

1. (i) x is spatio-temporally contained in y

2. (ii) for any other part of y, either

   1. a) x is spatio-temporally discrete from it, or

   2. b) x and it are assigned to different “levels”, such that parts from only one level are used in any given micro-explanation,

3. (iii) x and y. share at least one quantitative, additive property such that

   1. a) the amount of this property attributed to x is less than the amount attributed to y, and

   2. b) the former amount arithmetically contributes to the latter.

The issue of the admissibility into scientific discourse of teleological terms such as ‘function', ‘goal-direction, ‘purpose’ and ‘goal’ has not received widespread attention in recent years. In a provocative book, On Purposeful Systems, Ackoff and Emery claim that “the lack of more recent debate seems…no more than scholarly collusion in agreeing not to raise embarrassing questions” ([11], p. 15). Most of those who have attended to teleological problems have attempted to show that these teleological terms are eliminable in favor of non-teleological terms. ‘Function’ has been analyzed in this manner by Nissen [12], Nagel [11], Canfield [6], Hempel [8] et al. ‘Goal-direction’ and ‘purpose’ have been given similar treatment by Braithwaite [4], Nagel [11], Sommerhoff [21][22], inter alia. Notable exceptions to these reductionist analyses occur in the works of C. Taylor [23] and Wright [27].

Much has been written on the 1905 paper in which Einstein founded the special theory of relativity, his justly famous “Zur Elektrodynamik bewegter Körper” [4]. In it he innovatively defined the simultaneity of events at two distant points by imagining a light signal to be sent from one point to the other and back. In effect he defined the midpoint of the time interval between the signal's emission and return at one point to be simultaneous with its reflection at the other. And a large fraction of the writing about this paper (e.g., [19], [20], [3], [12]) has been concerned to show that, without denying any physical fact, he could have chosen any other point in the interval and thereby given an equally satisfactory alternative definition.

In the present paper I describe a general formal semantic scheme for the interpretation of quantum mechanics (QM) , and on the basis of this scheme I examine the modal interpretation of QM — both the Copenhagen and the anti-Copenhagen variants — proposed by van Fraassen [19, 20, 21], This is intended to be a fragment of a larger work [12] which additionally investigates a number of closely related interpretations, including ones proposed by Bub and Demopoulos [1, 2, 3, 4], Fine [5, 6], and Krips [16, 17].

Formal semantically speaking (see, e.g., Thomason [18]), a logic L may be characterized as an ordered pair <SYN,SEM>, where SYN is the underlying syntax (language) of L, and SEM is the semantics of L, which consists of a class of semantic assignments on SYN.

In his recent book Space, Time, and Spacetime, Lawrence Sklar has argued for a fundamental re-orientation in the philosophical study of the logical and epistemic status of the geometry of physical space. A very large part of the recent literature in this area, he notes (p. 109) has been concerned with the alleged bearing of the continuity of space on the question of whether geometry is the object of empirical knowledge, conventional stipulation, or utter ignorance. That continuity has been thought crucial is due, of course, largely to the enormous influence of the work of Adolf Gruhbaum (e.g., [9]). Sklar, however, argues (pp. 109-112) that the issue of continuity is in fact irrelevant to the logical and epistemic status of geometry, and that its pursuit leads one away from the fundamental questions involved. (See also Michael Friedman for a definitive critique of Grünbaum.)

Since there is a philosophy of science, it is not surprising that philosophers make use of case studies from the history of science in their work. However, the uses to which such case studies can be put depend on important, but not always acknowledged, ideas about how to understand science. According to one view, philosophy provides a theoretically precise account of science,and case studies are the raw material which, when reinterpreted according to a theoretical account, illustrate that account. Thus, just as formal logic is sometimes said to reduce ordinary language to a fixed and, from one standpoint, more convenient form, so the philosophy of science is often thought to perform a similar function for science. But other philosophers, with different ideas about how to understand science, think that case studies from the history of science can provide evidence for philosophy's claims about (scientific) rationality.

Galileo occupies a unique position in the philosophy of science. He has a significant place both in the history of the philosophy of science and in the history of science, that is, both in the theory of scientific practice and in the practice of which philosophy of science is the theory. Moreover, virtually every philosopher of science has felt or feels the need to come to grips with Galileo, in the sense that he either derives his theories from his analysis of Galileo, or he tests the theories he has otherwise formulated by applying them to the case of Galileo. Finally, it has turned out to be possible for philosophers of science of almost any persuasion to use Galileo for their purposes, in the sense of finding evidence from Galileo to support their theories. It is this last fact to which I wish to call attention.

Feyerabend, Kuhn, and Lakatos have held that no objective reasons can be given for entertaining or working on a particular new theory, paradigm, or research program. I shall examine one of the-main cases they appeal to and argue that, although Copernicus's theory was not superior to the Ptolemaic theory according to any of the usual criteria for comparing theories, it did have features which provided the early Copernicans with good reasons for entertaining it and trying to develop it further.

Let us begin with Feyerabend. He has been a vigorous advocate of a “principle of proliferation” according to which scientists should always be considering alternatives to the theories currently accepted. Indeed, Feyerabend holds that scientists should strive to “invent and elaborate theories which are inconsistent with the accepted point of view, even if the latter should happen to be highly confirmed and generally accepted” ([4] , p.26).

There is something methodologically amiss in Cartesian science. Apart from some recent commentators ([1], [12], [13]), the source of Descartes’ problems is usually located in a Rationalist preference for the evidentiary force of “reason” rather than “experience” in the construction and testing of scientific hypotheses. In other words, Descartes’ metaphysics and theory of knowledge are said to imply a scientific methodology which precludes, or greatly diminishes, the decisive role of empirical evidence in the development of a viable science of nature ([2]; [14], p. 384-5; [15]. P. 15; [16]).

I wish to agree in general with the recognition of some kind of miscarriage in Descartes’ methodology, but I locate the source of his scientific weakness elsewhere. I argue that there is no unusual or especially significant opposition between reason and experience in Cartesian methodology - nor do Descartes’ problems derive primarily from his. metaphysics.

In adults and even in small children, visual perception is framed against a richly developed and relatively stable cognitive background. This background affects vision itself. Experimental evidence in cognitive psychology shows that the effects of this background on visual perception are logically immediate and not the products of reasoning or of deliberation. Nor are the effects of this cognitive background discovered solely in the patterns of use which the perceiver makes of visually received information. In this paper three of the principle aspects of epistemological treatments of visual perception are assessed in view of the cognitive structuring of the visual process. First, traditional distinctions among the sensory, phenomenal, and cognitive aspects of the visual process prove to be less clear than formerly they were thought to be. Second, the concept of perceptual strategies is argued to be reasonable and fruitful in the epistemological analysis of the visual process.

In his book, Russell and Moore: The Analytic Heritage, A. J. Ayer takes issue with Russell's defense of scientific realism and his rejection of naive realism. ([1], pp. 126f.) Russell's position is succinctly stated in An Inquiry into Meaning and Truth ([3], p. 13) as follows:

There is an increasing recognition by friends of personal probability that the standard systems of personal probability do not provide a fully adequate basis for the theories of scientific inference and rational decision making. This recognition has methodological and formal components. On the methodological side, Jeffrey [8] and Spielman [16], [17] have suggested that personal probabilities should be interpreted as judgments about the credibility of propositions, i.e., as appraisals of the degrees of confidence that are warranted by the information available to the appraiser. The idea behind this interpretation is that there are criteria for appraising degrees of confidence as rational, plausible, silly or unwarranted. They have not been canonized, with the exception of the traditional informal fallacies of insufficient evidence, and if they were, would be highly vague and subject to revision as knowledge advances.

The project of formulating an account of scientific inference in terms of concepts drawn from probability theory, and based on the programmatic belief that inductive logic (that is, the theory of the principles of inductive reasoning) is the same as probability logic, is one usually associated in our own day with the names of such twentieth-century philosophers of science as John Maynard Keynes, Hans Reichenbach and Rudolf Carnap ([13], [27], [3]).