Advocates and opponents of nuclear power agree that cold water reactors will not explode like atom boaba. Such a logical poaaibility is incompatible with our knowledge of the design of such plants and the physical principles regulating their operation.

No such knowledge rules out the possibility of a core meltdown. Estimates of the rate at which such meltdowns will occur range from one in a million years (so that the probability of a reactor running one year without a meltdown occurring is 0.999999) to one in a thousand years (so that the probability of no meltdown in one year of operation is 0.9990005). (These probabilities are computed on the assumption that the production of meltdowns is a poisson process.)

Some philosophers think the difference between the hypothesis about atomic explosions and the one about core meltdowns to be a matter of degree.

Many philosophers used to believe that the objectivity and rationality of science lay in its reliance on “brute facts”, on what was “given” in sense-experience without any interpretation (and therefore any presuppositions) whatever. For those philosophers, granted their sharp distinction between the pure given and what is built on or added thereto, a distinction between “objectivity” and “rationality” was comparatively easy to draw (though I know of no place where it was explicitly so drawn). For a person to be “objective” in a given inquiry would be for that person to base his inquiry (either his deductions or inductions from the given, or his justification of his theories, depending on the particular version of the view we are considering) solely on the unvarnished facts, the pure given. And for a proposition (conclusion, hypothesis) to be “objectively based” would be for it to be based solely on (deduced or induced solely from, justified in terms of) the pure, uninterpreted given.

There is a common theme in the papers by Levi (1985) and Shapere (1985). Both are much concerned with the nature and role of background knowledge in science. Yet the approaches to this common problem represented in these two papers are so different as to be almost incommensurable. This divergence, I think, is not solely the fault of the authors, both of whom are distinguished and experienced philosophers of science. It reflects widespread features of the philosophy of science as it is now practiced. I will begin by saying what I think these features are and by suggesting an alternative conception of our enterprise as philosophers of science. Then I will comment on some specific points in each of these papers from the standpoint of this alternative conception. Finally, I will give my own account of the role of background knowledge in a specific scientific case.

The papers by Isaac Levi and by Dudley Shapere, despite their many differences, share a common theme. This is the idea that science in a certain sense builds upon itself, that some of its assertions, having become firmly established, play important roles in the further development of science. Established science not only guides practical action, it also constrains possible conjectures, guides the construction of more extended and related theories, and even guides the construction of their eventual successors (though this last needs very careful handling). This basic idea is surely correct; but I have many disagreements with the particular (and different) ways that the basic idea is developed by Isaac Levi and Dudley Shapere. I shall therefore first, outline very roughly the sort of development of this basic idea that I would advocate.

For many of us nurtured in idealist ways inclining us to unwavering acceptance of Kantian principles, 1951 was a bad year. We read in Reichenbach's The Rise of Saientifio Philosophy of the “disintegration of the synthetic a priori”. Kant had taught us that there are very general principles—each one connected to a primal category of thought—that are necessary in the formation of mathematics and physics and are expressible in nonanalytic propositions. However, since Kant's death in 1804 both mathematics and physics have developed revolutionary traits: non-euclidian geometries, new developments in symbolic logic, relativity physics, and finally, quantum mechanics, sealed the negative fate of Kant's high principles. Henceforth we must accept that there are no non-empty claims about reality that can be counted as necessary or indispensable to mathematics and science.

One way of putting our Interpretive problem, as Professor Butts, does in his very interesting paper, is to characterize the relation(s) between Kant's transcendental and metaphysical principles, on the one hand, and between these principles and the presumptive laws of empirical science, on the other. Another way of putting the problem is to distinguish between Kant's merely illustrative use of the science of his day, his appropriation, generalization, and redeployment of contemporary, often chemical, concepts, his commitment to certain features of classical physics, in particular, as constitutive of any adequate conception of experience, and his identification of the deep presuppositions of any science worthy of the name.

In fact, I do not think that there is any general and precise way in which transcendental and metaphysical principles and the laws of empirical science are related. At best there are analogies.

The central purpose of the symposium in which this paper was originally presented was to elicit from anthropologists those problems in their discipline which, from the perspective of a practitioner, might merit philosophical attention. Setting aside the difficulties that must arise when an anthropologist, unfamiliar with the scholarly tradition of philosophy, attempts to identify philosophically interesting issues, any anthropologist is confronted with a further problem. Anthropology is not a coherent discipline in the same fashion as many other academic departments. Nor is any anthropologist really “expert” in more than one of the major disciplines (I would caution wariness of those who claim more) and most only within a particular strain of the relevant discipline.

Being an archaeologist, I feel constrained to treat that discipline.

To begin with, an apology for ray subtitle, which consciously but unforgivably alludes to Hollywood's current fondness for numbered sequels to successful films. My intention was to signal that this paper, like a film such as Halloween II, will make further use of old materials, partly to bring the story up to date, partly to re-work material of proven box-office value, partly to answer those critics who, perhaps, deserve an answer.

Synoptically speaking the paper returns to the issue of whether anthropology aims to understand or to explain. No claim was made in Part I (Jarvie 1970) that the two are mutually exclusive, only that understanding is a vague idea which derives much of its appeal from poor anthropological information about science and its aims. Part I concentrated mainly on criticising Peter Winch's ethnography of science and of Azande witchcraft.

It is frequently argued that the phenomenological or qualitative features of our sensations will never be satisfactorily reduced by a purely materialistic neuroscience. They constitute, it is often said, a permanent barrier to the reductive aspirations of physlcalism. In what follows I provide a critical examination of several of the most popular of these arguments. I also explore the transformation and expansion of human subjective consciousness that a successful psycho-physical reduction could sustain. The conclusion is that, far from being Immune to neurophysiological explanation and reduction, the subjective qualities of our inner lives will enjoy a dramatic unfolding as the representational and computational strategies of the brain are revealed.

Churchland (1985) is much struck with the possibility that we might learn to make introspective reports in “the conceptual framework of a matured neuroscience”. He thinks that this would yield “a quantum leap in self-apprehension”. And he apparently thinks that the possibility of this strengthens the case for reductionism. I will not question the claim that we might learn to make in a quasi-introspective way the sorts of reports Churchland envisages, about dopamine levels in the limbic system, spiking frequencies in neural pathways, and so forth; nor will I question that this might have a beneficial effect on our aesthetic appreciation, our emotional self-understanding, and the like. I will question whether the awareness Churchland envisages would be genuine introspective awareness, although in some sense it would involve it.

Karl Marx has written that “modern Industry … makes science a productive force distinct from labor and presses it into the service of capital” (1867, p. 361) Moreover, according to Marx, “The ideas of the ruling class are In every epoch the ruling ideas”, which are “nothing more than the ideal expression of the dominant material relationships” (1846, p. 64). Part of what Marx means by ‘ideal expression’ is revealed when he argues that ruling ideas have a legitimating function: a ruling class must “represent its Interest as the conmon interest”, “give its ideas the form of universality, ana represent them as the only rational, universally valid ones.” (1846, pp. 65-66).

The idea that scientific inquiry is objective is unquestionably among the ruling ideas of our epoch, and it represents science as serving not the Interests of a particular class, but a purely general interest in the understanding of nature.

Suppose that some event F occurs and later an event E either occurs or does not. I am going to talk about the extent to which F tends to cause E, or the causal propensity of E provided by F, or the propensity or tendency of F to cause E, denoted by Q(E:F). For example, F might be the event that the captain of a firing squad shouts “Fire” and E might be that the shootee's life is_ Ended (F for Fire, E for Ended). The negation of E is denoted by E, which in this example means that the shootee does not drop dead.

Probability kinematics is the theory of how subjective probabilities change with time, in response to certain constraints (accepted by the subject in response to his experience or incoming information). Transformations — rules for updating one's probabilities — which have been described and to some extent accepted in the literature, include (in increasing order of generality): conditionalization, Jeffrey conditionalization, and INFOMIN (relative information minimizing). The present paper investigates the general problem. By an argument purely based on symmetry considerations it is demonstrated that conditionalization and Jeffrey conditionalization are the unique admissible rules for the cases to which they apply. The general problem is thereby reduced to a sequence of two operations, the second a Jeffrey conditionalization and the first a determination of posterior odds on some partition. In Section 2 we give a new deduction of INFOMIN from a simple postulate; in Section 3 a rival rule (MTP) derived from an analogy to quantum mechanics. The Appendix presents some preliminary comparisons.

As everyone knows, in almost all modern cosmological models the origin of the Universe is placed between 10 and 20 billion years ago (van den Bergh 1984). The only exception I am aware of is the steady state theory, which has recently been given new life by the inflationary universe model (see Gott 1982, and Schramm's conference presentation), and which requires the universe to be infinitely old. No modern cosmological theory holds the Universe's age to be a mere 6,000 years, the traditional value in the West for centuries (Ussher 1660; Whiston 1696, 1717; Haber 1959). It is universally thought that radioactive dating and astronomical observation have definitely ruled out such a short age for the Universe. It is universally thought that it is impossible to construct a falsifiable theory which is consistent with the thousands of observations indicating an age of billions of years, but which holds that the Universe is only a few thousand years old.