Bas van Fraassen

doc: city-a.nov Nov 24, 1995

[1] A challenge to Duhem

[2] Scientific inquiry and its presuppositions

[2.1] The argument for proliferation

[2.2] Presuppositions

[2.3] Questions and parameters of inquiry

[2.4] Empirical hypotheses forget their origin

[2.5] Exploration of rival empirical hypotheses

[3] Example: Presumptive materialism

[3.1] What is matter?

[3.2] Two moves for materialists

[3.3] The first move: materialism as scientific hypothesis

[3.4] The second move: whatever it takes

[3.5] Materialism as false consciousness

[4] Naturalism, materialism, scientism: where does science end?

[4.1] Satisfaction with objectifying inquiry

[4.2] Naturalism and scientism

[4.3] Conclusion: the relation of metaphysics to science

Bas C. van Fraassen

As activity, science has become a large-scale cultural phenomenon. As product, it is drawn on by industry, agriculture, and medicine, thus affecting not only the scene of its activity but all the rest of the world as well. Western philosophy has always harboured a tradition which regards scientific inquiry as a paradigm for rational inquiry in general. Yet almost every philosopher in that tradition has pointed to limits of this paradigm and its scope.

Every philosophy provides a different lens through which to view this object of common admiration. In this essay I shall reflect on two views of science which are at first glance inimical to each other. The first is Pierre Duhem's, who saw science as neutral on all issues of metaphysics, theology, and religion. The second is exemplified by Paul Feyerabend, who called for alternate research programs guided by rival metaphysics, and argued that such rivalry has always been a driving force in science. I will argue that Duhem is right, in the main, though our picture of science must be leavened by the insights of the contrary point of view. This will not be an archaeological inquiry into those thinkers' thought; I will make it mainly an independent reflection on these same issues.

[1] A challenge to Duhem

Pierre Duhem walked a tightrope. Today we accept much of his critique of certain views of science as correct. But in his own day he was accused of allowing his religious and metaphysical views to enter his physics in an improper way. Duhem contended that science describes but does not explain -- at least in the sense in which metaphysicians want explanation. This leaves room for metaphysicians (and therefore, theologians) to add their explanations to science, whether as foundation or as supplement. Is this view of science not simply the philosophical dodge of someone who tries to make the world safe for religion? In response Duhem articulated his view of science as truly neutral on all metaphysical questions. The correct or proper way to pursue physical theory requires this neutrality. Why so? Without this neutrality, he says, the disagreements that run riot in metaphysics will ingress into physics. As a result, physics would not be an activity we can all work at together, regardless of our metaphysical views. Science is to be a common ground, where all can join in a common enterprise, regardless of creed, worldview, religion, or metaphysical beliefs.

On this fundamental point there is no real agreement today. Some do indeed hold that neutrality in this sense is part of the objectivity that characterizes scientific inquiry. Others point to the far-reaching assumptions with which famous scientists have approached their subject. Some, perhaps especially philosophers in the scientific realist camp, frankly approve of this, and take some metaphysical assumptions to be beneficial or even indispensable to science. Others, such as Marxist, feminist, social constructivist, and religious critics attempt to expose the bias which crept into scientific theories due to the scientists' unacknowledged background opinions. Duhem could rest easy with such views if they are understood as asserting defects in extant science. But not all are thus; indeed, some advocate that science be pursued under the aegis of certain worldviews or of presumed knowledge come by extra-scientifically.

The most weighty challenge posed for Duhem's view of science is, I think, this: if metaphysical assumptions can make a difference to science, who is to say that science will not be more successful if it is less neutral? Is this not also an empirical question?

Philosophical positions, and specifically those embodied in metaphysical systems, normally identify themselves through their theses. Those theses can certainly have empirical implications, even when they are not first and foremost empirical hypotheses. What happens, exactly, if a scientist begins with or draws on such a position? I have two answers. My first answer (in part [2]) is this: if assumptions have some non-metaphysical, empirical content, then once admitted into science the implied empirical hypotheses will, as it were, forget their origins. They can play a role, but not one that will offer comfort to the original motives or extra-scientific orientations that brought them in.

However, there is more to a metaphysical thesis than its empirical implications. Will the remainder, the metaphysics advanced as 'underpinning' for science, not play a much more significant role? My answer to this (in part [3]) is that in fact, metaphysical assumptions do not in themselves -- apart from any empirical implications -- make a difference to science. I cannot pretend to prove this very general contention, but I shall try to support and illustrate it with the central example of science's 'presumptive materialism'. There is indeed a way in which materialism can play a role in shaping science, as I shall try to show: not through the content of its theses, entering as assumptions, but through attitudes for which these theses are mere codes. That is, I shall ascribe the true effect of metaphysics on science as a case of false consciousness.

[2] Scientific inquiry and its presuppositions

Just how can a metaphysical standpoint contribute empirical hypotheses to science? To examine this, we need to look at the senses in which science is and is not presuppositionless. At any given historical stage, science includes a large body of putative knowledge about nature, that is, the currently accepted theories and facts classified as evidence. This body is quite obviously not neutral with respect to empirical hypotheses. The willingnes to stand on the shoulders of our past is the very basis prerequisite for progress; but it may hinder progress as well. As Feyerabend and Kuhn have shown so convincingly, progress in science depends crucially on the exploration of rival empirical hypotheses. Would it not therefore be a very good thing if there could be more and larger scale rivalries? Such diversity could come from science being guided in different research communities by various systems of metaphysics. Feyerabend advocated such proliferation, for exactly that reason.

[2.1] The argument for proliferation

If we agree, are we not led to a further conclusion? For then, exactly to the extent that these large scale rivalries are in fact absent, science is not neutral after all between the metaphysical systems under whose guidance empirically differing sciences could be developed. In this way, the admission of non-neutrality with respect to empirical hypotheses seemingly entails non-neutrality with respect to metaphysics.

This argument insinuates a good deal more than it establishes. It would be a very interesting argument if some or all of the following suppositions were true. Behind these lies some deeper supposition, about how metaphysical positions contribute empirical hypotheses to science, which I shall take up below. That apart, three suppositions are subliminally at work here:

(For example, hypotheses with a common origin would be related in some way which is both due thereto, and of importance within science.)

It seems to me that all of these are mistaken.

[2.2] Presuppositions

The idea that science is, or even can be, presuppositionless is rightly dismissed today as incorrect. But there is a distinction to be made. In one way, it cannot be: every empirical inquiry has its presuppositions. The investigator approaches his subject with specific questions or with a specific way of asking questions and so the very design of the inquiry involves some preliminary idea of that subject matter. In another way, however, science must be presuppositionless if it is to be science at all. No rational inquiry may 'beg the question' by presupposing the results of that inquiry.

Every inquiry has presuppositions, exactly because it is defined in part by specific questions or a specific way of asking questions. These questions, and to that extent the subject matter of the inquiry, are delimited in part by a class of 'relevant' features -- the parameters in which the subject and its problems are to be stated. Whether or not those features alone are relevant to other concerns we might have, it is they which discipline and focus the inquiry. This is a crucial point which Kant saw as the revolution that initiated modern scientific method. He took it to be one necessary condition of fruitful science: that the inquirer decide on and strictly determine as a preliminary which questions he will put to nature.

But does this procedure, perhaps a crucial prerequisite for scientific progress as Kant maintains, not open science up to the charge of biasing all its inquiries from the outset?

If an inquiry turns out to have presupposed any part of its results from the beginning, then it was flawed -- isn't that so? Such a flaw is in principle detectable by anyone, including by those who engage in the inquiry. Detection of such a presupposition, however, does not automatically invalidate the results. Quite the contrary -- the obligation not to beg the question means simply that when the presupposition comes to light, it must be turned into an acknowledged assumption. As example take the aberration in starlight, to allow for which astronomers have to point their telescopes away from the star to be observed, at an angle which varies with the seasons. From this it was concluded that the aether is motionless with respect to the stars, and is not e. g. dragged along by the Earth. This conclusion followed in a context in which certain presuppositions were in force. In retrospect we re-interpret this finding: one conclusion properly stated is that if light be a wave travelling through a material medium, the aether, then ... etc. With the wave theory of light triumphant perhaps there seemed to be no point in rendering that "if" explicit. Nevertheless, this acknowledgment of previously tacit assumptions, once laid bare, is required for science to be scientific.

Prior probabilities provide a more difficult example of the ubiquity of initial presuppositions. Historically it is clear that they play a role in theory choice and research design. Perhaps the opposition within science and orthodox statistics against the role of prior probabilities derives exactly from the sense that they are to some extent a presupposition of results of the inquiry. If I investigate hypotheses A and B with the prior opinion that A is ten times as likely as B, am I not going someways toward the assumption that the outcome is to be that A is true? If those prior probabilities enter experimental design, is that not a flaw of the same sort, even if not to the same degree, as a design that ensures the outcome in certain respects? Exactly the same point is to be made here as above: the disclosure of such partial presuppositions does not disqualify the research, but once disclosed or challenged they are to be fully acknowledged in the announced research results. There is a flaw only if the researcher begs his own question. Prior probabilities, however, are not theses but propositional attitudes. They cannot be removed by listing them in the antecedent, though some "explicitating" response is possible.

The simple idea that presuppositions in science are examples of flaws in design is therefore not tenable. The unearthing, bringing to light, laying bare of presuppositions is an unendliche Aufgabe. The hidden assumptions form a potential infinite; for every one we uncover, there will be (in some platonic sense) another. To this aspect of our finitude we must be resigned: consciousness is not transparent to itself.

The inevitability of presuppositions and the consequent need for proliferation in theoretical research are no objection to Duhem's concept of the neutrality of science. The question is, however, whether this gives purchase to religious or metaphysical interests, which may insinuate themselves by contributing empirical presuppositions to science. Again, as a purely historical or psychological point about the origin of hypotheses, Duhem cannot be understood to have denied this. The Duhemian point must be that if this happens, it does not matter to science -- that these origins provide neither credentials to hypotheses nor guidance to inquiry. In this he would be right, I think. To make that point we must look a bit further into that way in which science is not presuppositionless.

[2.3] Questions and parameters of inquiry

The passage I cited from Kant is famous. It marks the historical recognition of the sense in which science is not presuppositionless. But it does not make sufficiently explicit exactly how Galileo and his peers set about this new way to study nature. Scientific inquiry is an objectifying procedure. This is in the main a 'European' rather than 'Anglo-Saxon' way of putting it. What it denotes is a specific way in which presuppositions enter into scientific inquiry, a way in which scientific inquiry is delimited in a certain way beforehand.

A science, or a research project within a science, has a domain. It is tempting to think of its initial delimiting as specifying a set of things; e.g. "this will be a study of frogs". But that is not accurate, since the central subject matter will inevitably have to be connected with other things. Even an anatomical study of frogs will inquire into, for example, the rate of coagulation of frog blood when exposed to air, and the air is not part of a frog. The more accurate way of delimiting such a domain is in terms of the quantities which will be allowed to figure in the description of the phenomena to be studied. Under quantities I include here properties and relations; "quantities" is the more inclusive term, since we can think of properties as two-valued quantities, and so forth.

So the first step in setting up the scientific inquiry is to select the 'relevant' quantities. All questions to be broached must be questions formulable in terms of these quantities. The next step will be to describe the phenomena solely in terms of these quantities. This description must be systematic; the result will be a 'data model', which can be studied to discern patterns in the phenomena. The further step of framing and investigating hypotheses, which will stand or fall depending on their fit (in various ways) to these data models, will be equally constrained in terms of admissible quantities (though these may include "theoretical" quantities not used in the construction of the data models).

The paradigm example of this procedure is Galileo's list of primary qualities for physics: physical description was to proceed solely in terms of these qualities, no others to be admitted. Of course, Galileo's proposal was extreme: he wished to lay down the list of relevant quantities for the scientific description of nature once and for all. All quantities not on the list were to be the subject of reductive or eliminating analyses. We do not have to think of the objectifying procedure either as delimited once and for all or as necessarily linked to a reductionist thesis. Such a thesis may appear as a completeness claim for a certain theory, produced as science-as-product by the inquiry. In the case of theories of frogs, it will be less likely to come along than in the case of fundamental physics. In Galileo's case, the claim was definitely premature, as it was again in the case of Descartes' physics which admitted only kinematic quantities. A more recent instructive example is the methodological innovation attempted by behaviorism in psychology. The attempt to set up a well-delimited and well defined scientific inquiry is one thing; the shortlived completeness claims associated with it are another.

Such a clearly delimited inquiry may face radical failure, namely if questions formulated in its own terms cannot be answered in its own terms. The result is a "crisis in the foundations" of the particular science. The science in question must then augment or revise either its list of relevant parameters or its range of questions to be addressed. If for instance it is asked how current values of a certain quantity depend functionally on its preceding values, and/or current or preceding values of other quantities, there may be no answer. Then the modifications introduced to escape this crisis may indeed revise or extend the list of admissible 'relevant' quantities. This describes exactly the transition from Cartesian to Newtonian physics. It is not the only way: alternatively one might admit statistical modes of description that still utilize only the 'old' quantities. There is a recurrent trade-off in the history of science between determinism and 'hidden variables'.

Could revisions in the set of descriptive parameters shift science into the metaphysical, or occult? It could, in the sense that the new parameters were earlier on excluded exactly on the basis that they were 'occult'. This is illustrated very well by the Cartesian reaction to Newton's mass and force as being a reintroduction of the 'occult' qualities which had been banished by the mechanical philosophy. It is equally illustrated in the 'orthodox' Newtonians' attitude in the 18th century to aether and subtle fluid theories, as precluded by Newton's vera causa rule.

But we must be careful not to generalize too shallowly on this point. First of all, even if occultism or alchemy had inspired the introduction of those new parameters, their role in scientific inquiry once introduced robbed them of all connections with those 'disciplines' whence they 'came'. But secondly, science has a 'shell' structure which imposes strict discipline on how the list of relevant parameters can be revised. Each scientific project lives in an environment created by larger projects, past and present, on which it draws, and with which it needs to stay in harmony. The set of parameters is revisable; it is not fixed in stone. But an alteration in the set of 'relevant' parameters 'revolutionizes' the project, engenders an heir-project so to speak. The revolution may be more or less radical depending on how high up in the 'shell' structure the alteration or augmentation of parameters occurs. So the study of frogs, and indeed of animal physiology generally, was definitely revolutionized when Volta did his experiments in which static electricity made frog-legs move. But notice that the new parameter in the description of the physiological phenomena belongs to a higher shell, physics. Physics itself and everything in between physics and physiology felt the effects of this revolutionary move. Only a narrow frog specialist would refer to it solely as a scientific revolution in the theory of frogs.

The crucial point -- about the revolutionary character of alteration in the set of relevant parameters, due to the 'shell' structure of science -- is the difference between possibility in principle and historical feasibility. Imagine a judge in an American court sentencing a thief to have his hand cut off. It is true that our values, laws, and beliefs can be revised so as to incorporate those operative in whatever other culture you care to mention, in principle -- but that does not mean that such an alteration is now feasible.

[2.4] Empirical hypotheses forget their origin

I will use a single example to illustrate two points which are closely related to this delicate interplay of assumption and disclosure. The first point is that an empirical hypothesis introduced into science for reasons, and perhaps with putative credentials, coming from outside that science, will so to speak forget its origins -- as I said above -- so that its fortunes within science will soon have little or nothing to do with those origins. What this means is that within science the suggested empirical hypothesis will have none of the significance it had in its extra-scientific context, nor need there be any relation to other hypotheses with the same origin.

Examples of this sort of thing must necessarily be fanciful. Imagine the introduction of the following hypothesis of a paranormal phenomenon: horses did not evolve from lower forms of life, they appeared on earth quite separately at some definite time in the past. Suppose that some community of biologists admits this hypothesis as basis for research, but is in other respects like our biological research community in every way. That means: the set of questions that defines their inquiry would still be the same. Therefore the origin of the hypothesis, possibly in contexts outside scientific inquiry, would not play any role. Only the description of the paranormal phenomenon within the parameters of biology and its parent sciences (physics, chemistry) would be 'relevant'.

It would become crucial for them to theorize about the hypothetical introduction of the horse. When does the first horse exist in the world? Did it grow from a baby horse to a mature horse? What animal does it mate with? Did many 'first' horses come into existence at the same time? In addition, in either case, there needs to be an inquiry into the survival rate of these new babies, and whether many of them are created so as to counteract attrition by exposure etc. or whether they always come to be in conditions that allow them to survive till they have offspring.

These questions within biology and paleontology lead to questions for physics. What exists just prior to the existence of the first horse, in the location where the first horse first appears, whether as baby or as mature specimen? This latter question is itself circumscribed by the questions that define the background of biological inquiry: the 'what' refers to physical conditions, the configuration of the matter at that location. Was energy, mass, momentum conserved in this transition that marks the appearance of the first horse? What, if any, of the matter located there immediately beforehand, is part of that first horse?

The second point, which may at first look tangential to this but is not, concerns the curious notion one sometimes sees of "science stoppers". For example, it is sometimes suggested that paranormal hypotheses, such as the existence of ghosts or hypotheses of special creation of different animal species, stop science in its tracks if admitted. But there cannot be such a thing as a true 'science stopper' at all.

Some of the questions raised in our example may be answered in unorthodox ways, but the inquiry does not, cannot come to a stop. Suppose the theory developed is that the first instance of every natural kind just springs into being, at its moment of entry into history, always on top of some mountain. Then all the conservation laws I mentioned are violated, very many times, so physics has to be carefully but radically revised so as to allow for these phenomena. Moreover, answers have to be sought to what happened to the air in that place: exactly what is the distribution of additional momenta imparted to the air molecules so as to move them out of the way? Can this be described deterministically or only in terms of irreducible probabilities? Alternatively, of course, the answer hypothesized may be that the organic and inorganic matter present changed into a baby of the new species. The exact process of change would then become the topic of inquiry. The research program for such a biology seriously pursued is vast. If the community of biologists in question did not proceed in this way, but treated the hypothesis as a "science stopper", then its pursuit would not be scientific.

To support this conclusion, let me take up two putative objections. The first is that someone might proceed quite differently from the above, so as to preserve most of extant science while maintaining the hypothesis, by restricting prior theory to "normal" as opposed to "exceptional" occurrences. The intuitive idea would be that the paranormal phenomena occur very rarely or perhaps even only once in the distant past, while science only formulates regularities that hold 'in general'. There are two reasons why this is not possible without disengaging oneself from science. The first is that the theory cannot be maintained unmodified however small or large the "area of exception" is chosen. If for example the initial creation is of many horses (to prevent rapid extinction), then the general predator-prey balance is upset, with consequences for the evolution of other species. If instead the initial creation is of few, then there must be a compensating hypothesis, such as of rapid mutation, to offset the damaging effects of inbreeding. The hypothesis might be more insulated from counterevidence the larger in time and space we mark out the area/epoch of exception, and the farther it is placed in the past. But that does not remove the problem in principle. In addition, quite apart from the questionability of manoeuvres designed to eliminate possibilities of refutation, they weaken science by curtailing its treatment of other phenomena in that epoch/area.

The second reason is more important. When the theory is thus restricted, we change it from an assertion of form A to one of form A, unless X. Now X cannot be something like A miracle occurs; it must be formulated in terms of the parameters demarcating the domain of the theory. But X must also be genuinely informative, otherwise the 'restriction' becomes something like A, unless not A. This would be less of a problem for someone who held that in the whole history of the universe, only the appearance of horses was such an exceptional, paranormal occurrence. It becomes a serious problem if this manoeuvre is suggested as a general way to proceed in science, i.e. to formulate theories for 'normal' developments with 'exceptional areas/epochs' in which science is silent. Now, describing the areas of exception in terms of the relevant parameters is as challenging a task as any outlined above. Moreover, if the motivation is a belief in agencies properly outside the domain of all science, that belief itself may imply that the task is impossible. But then the 'restriction' would leave us only with a 'science' asserting that certain regularities always hold except when they are violated. That is an empty assertion. The proposed policy would therefore certainly stop science for those who adopt it; but that is why I say that adopting it involves disengaging oneself from science altogether.

The second putative objection is that the inspiration which suggested this hypothesis may turn out to make a great difference to science, whether for benefit or harm. The reason would be that scientists not strongly motivated by beliefs contrary to evolution will not pursue biological theories that would require such radical revisions in physics and chemistry. Therefore, the credit or blame for the effects on science of this new direction in research belong in that case to the extra-scientific inspiration.

Note well: the putative reason offered is a psychological hypothesis about the private life of scientists who explore hypotheses contrary to current accepted scientific opinion. Such a thesis may always be able to take refuge somewhere between the subjects' behaviour and their rhetoric. But on the face of it at least, as we will see below, the actual history of science does not seem to support the hypothesis.

[2.5] Exploration of rival empirical hypotheses

Science is very clearly not neutral with respect to empirical hypotheses. Equally clearly, rival empirical hypotheses are explored in science, and should be. Duhem could not possibly have meant to deny that, given how eloquently he wrote on these very matters. So his reason for advocating neutrality with respect to metaphysics cannot simply be the need for agreement and unity. Yet the one danger he mentions explicitly is that

Can these really be the words of a physicist so engaged in polemics and controversy with the scientific orthodoxy of his day?

Some reflection on Duhem and Feyerabend's own examples will bring us to a more palatable view of the matter. General agreement on a large body of background theories and data can go hand in hand with exploration of possibilities that would overturn them -- indeed, that is exactly how it is in science!

To see this, consider how the call for exploration of rival hypotheses fits in with the point of developing science in the first place. Unlike pure philosophy, science is not a purely academic exercise. Science fails to attain its goal if it does not produce some such thing as the 'body of science', 'accepted scientific opinion', science-as-product to function as guide to practical life. The end is to come up with theories that meet science's criteria of success, and having thus succeeded can be drawn upon for the many needs for which we value the success of science. This requires a certain unity and agreement within science -- the science that science is building is one -- which, a little paradoxically, derives its value exactly from its emergence out of disagreement and disharmony.

Everyone is free to design his own inquiry; a scientific community will do so in its own way, and hope to be vindicated in its choices. Kuhn and Feyerabend have amassed impressive historical evidence for the crucial role in scientific progress of alternatives to accepted theories, developed while the evidence was still overwhelmingly against the alternative. On the other hand, the economics of scarcity that govern all human life applies here as well, and random pursuit of alternative theories would be as unproductive as experimentation not guided by any theory. Anyone basing advocacy of specific departures in scientific research on Feyerabend's sorts of grounds must, it seems to me, also submit to Feyerabend's harsh response to the "But then anything goes!" objection:

Feyerabend's own examples show that even 'mainstream' science is the constant scene of exploration of rival hypotheses and replete with controversies over alternative theoretical approaches. Moreover, there is a quite distinct and often lively area of marginal 'foundational' research in which those engaged in 'normal' science participate along the way. Feyerabend complained at great length of the dominance of the Copenhagen school of thought in quantum mechanics. But not only was he able to point to alternatives under investigation by such eminent physicists as de Broglie and Bohm, we have since seen ample energy devoted to testing the predicted violations of Bell's inequalities, developing the many-worlds interpretation beloved of cosmologists, and the dynamic reduction program (GRW model), to mention a few. Each of these passes Feyerabend's criterion quite handsomely -- in contrast to such examples as Velikovsky or Dianetics, which do not belong even to the margins of science at this point, by that criterion.

This is well illustrated with aspects of science which Duhem himself brought to light in the history of science. Specifically, Newton's development of his theory of the heavens, without preserving consistency with Kepler's laws of planetary motion or Galileo's law of fall, was a clear case of proliferating rival hypotheses in the teeth of the evidence. Thus, the value of proliferation in science has a form compatible with Duhem's claim of neutrality for science; indeed, actual science already recognizes that value and gives it its due.

The advocacy of proliferation would reach absurdity if carried to its logical extreme. Even if we were immortal and had unlimited resources, we could not have an alternative research project going on for every definable set of parameters and questions, and every set of prior probabilities for the answers. However, the mere thought of this impossible fancy forces us to consider a distinction with respect to introduction of hypotheses. If there were a research project going for every set of hypotheses and every allocation of probabilities thereto, then would science as a whole imply nothing at all? Would accepted scientific opinion consist only of tautologies?

Clearly that does not follow, since we can distinguish (1) formulating a hypothesis and designing research to tease out or test its implications, and (2) ceasing to accept theories that imply its falsity. Indeed, Feyerabend's most dramatic examples of the value of exploring new hypotheses which are in conflict with the evidence, sub specie accepted scientific opinion, are of exactly the former sort. It follows therefore, contrary to what Feyerabend seems to imply, that if science is not neutral on some hypothesis, the creation of a rival research project to explore it does not engender neutrality. There is a need for proliferation in theoretical research, but not a need to suspend acceptance of current science while we proliferate.

The cases examined by Feyerabend, in which exploration of rival hypotheses and indeed whole cosmologies turned out to be of crucial and indispensable help in unseating unsatisfactory older theories, illustrate this very well. They do not turn on that ('suspension' or 'non-acceptance') sense of exploration! In each case we see scientists exploring hypotheses which are admittedly and often blatantly counter to the accepted evidence as well as accepted scientific opinion. Since Feyerabend himself illustrates this abundantly, it is surprising that he does not see the corollary: inconsistency with accepted science and lack of positive prior probability does not prevent exploration of rival empirical hypotheses within science. On the contrary, many great advances in science provide counterexamples to such a view. There is a second corollary: the remedy for a perceived paucity of such exploration need in general not come from scientists who do have positive probabilities for alternative hypotheses or who reject the accepted theories in extant science. Who knows what motives are needed? Intellectual playfulness, fascination with the long shot or with the role of rebel, who knows?

The view we must associate with Duhem is realized if there is agreement on what is to be accepted combined with free exploration of alternative theories and experimental inquiry suggested by those alternatives (to the extent that our scarcities permit). In that case there can be great diversity, but there will not be the sorry spectacle of a science that never comes to any conclusion because physical theories "reputed to be satisfactory by the sectarians of one metaphysical school will be rejected by the partisans of another".

[3] Presumptive materialism

The more important question confronting Duhem's concept of science is therefore whether science starting from certain metaphysical beliefs might be -- because of the way the world is, not because of the historical context of inquiry -- more successful than science which is neutral with respect to metaphysics.

One popular plea for metaphysics rests on the idea that for the good of science, scientists must start with provisional realism -- there are unobservable causes for all observable phenomena or some such thesis -- and presumptive materialism -- matter is all there is, so those causes are all material mechanisms of some sort.

This thesis certainly sounds debatable: surely it must be either so or not so .... But appearances are deceiving. Genuine debatability presumes understanding of what the suggestion actually means. A much more basic presupposition is at stake: do provisional realism and presumptive materialism in this sense make any difference to science at all? By "in this sense" I mean, in the sense of assumptions, theses, factual claims about what there is and what there is not, apart from the implied empirical consequences which we discussed in the preceding part. Let us look specifically at presumptive materialism.

[3.1] What is matter?

Does the thesis that matter is all there is rule out at least some kinds of theories, so that they are not even candidates for scientific exploration? I will argue that this is quite illusory. There may however be a certain attitude, orientation, or stance, associated with this thesis, which does affect science as well as practical and intellectual life generally, and for which this thesis functions as code. If that is so then materialism may be a prime example of false consciousness in philosophy. For in that case materialists may take themselves to be maintaining a theory while they are in reality merely expressing attitudes, in ways which lend themselves to such expression only under conditions of confusion and unclarity.

This is not simply an indictment of materialism. I offer this diagnosis in part as explanation of something which materialism has in common with other hardy perennials of philosophy. That is: there is besides the theses on which the day's materialists take their stand, and which vary with time, also such a thing as the 'spirit of materialism' which never dies. False consciousness can be avoided in two ways: (1) the philosopher may lack that spirit and be genuinely concerned solely with certain definite views on what there is, or (2) the philosopher may have that spirit and not confuse its expression with any particular view of what the world is like. The latter, however, has to my knowledge never yet been instantiated.

To begin, it certainly sounds as if matter is all there is (briefly, the Thesis) is a substantive factual claim. Does it not rule out Descartes' mind-body dualism, Aquinas' souls, spirits, entelechies, cosmic purpose? Well, it genuinely rules those out only if each can be made clear enough so that the denial has some genuine content; and even then it may not. Now, if the Thesis is the important part of materialism, then materialists will not rest till the distinction between matter and what is not matter has been made so clear that the Thesis is clearly a factual claim, which can clearly be either true or false. But if they do not, then by modus tollens I conclude ....

What would count as something that is not material? Descartes said that matter is extended and mind is not; mind thinks. But if that is not a stipulative definition, it is certainly wrong. Else we would have to say that Hertz' massive point particles, if they exist, are not material. Equally we would have to say that Hartry Field denies materialism when he claims that space-time points are real, concrete individuals. And finally, we would have to conclude that Alvin Plantinga either does not think or is not incarnate, both of which are clearly false, to an extraordinary degree, given his subtlety and size.

It may be unfair to take Descartes as our whipping boy. But more recent putative statements of materialism do not, it seems to me, fare much better. Typically they start from some version of received scientific opinion, perhaps with some anxiety about being up to date. They will not say that elementary particles are all there is, since they know that there are trees, persons, and rocks as well. But they will say that everything is composed solely of elementary particles. If we take this seriously we shall, I wager, once more land in an untenable historical parochialism. When Newton introduced forces in addition to bodies, did he deny the Thesis? Forces are not composed of particles. When Huyghens' waves-in-the-aether theory defeated Newton's particle theory of light, was that a set-back for materialism? Surely not, although the aether was a continuous medium, not particulate. When a recent article in a physics journal bore the title "Particles do not exist", was that a denial of materialism? The author's argument was that particle number is not relativistically invariant, so that how many particles there are is as relative as left/right, up/down, rest/motion. If that is so, does materialism bite the dust? Surely not. But if materialism were really, purely and simply, some such thesis as that everything is composed of elementary particles, I could not so readily say "Surely not"!

[3.2] Two moves for materialists

Soi-disant materialists have certainly taken cognizance of this difficulty, that their most important terms seem to lack content. In response, they have opted for one of two moves. Some have attempted to formulate very specific theses relating to the putative subject matter of psychology, argued that these are empirical, and offered the results as a specific version of materialism. (By "empirical" I mean here "having some empirical implications"; it does not preclude other implications concerning the unobservable.) Others have nailed the Thesis down by nailing it to a specific science, by means of a completeness claim for that science. It seems to me -- and I will try to show -- that neither move leaves us with materialism as an identifiable substantive thesis.

[3.3] The first move: materialism as scientific hypothesis

U.T. Place argued that it is tenable to say that certain events and processes traditionally classified as mental (for example, sensation) are identical with events and processes in the brain. That this is indeed so he labelled as materialism, and argued that it is in fact a scientific hypothesis. In response to Smart he agreed further that the conditions required for the assertability of such a hypothesis -- conditions under which alone such an identity statement can be true -- are subject to philosophical debate rather than empirical testing. But once such conditions are specified, the remaining question is empirical. For the described 'mental' events and processes have a certain complexity, which brain events and processes may or may not have. The name "materialism" is also given to this or closely similar claims about the psychological e.g. by David Armstrong.

There are three preliminary questions to be raised. First of all, not every replacement for what I have called the Thesis can be accepted as the 'real' materialism -- can this one? Since the main question before us is what exactly the materialist's main thesis could be, we should perhaps accept any seriously offered contender. But if we could identify certain familiar psychological events and processes with physiological ones in some not too weak sense, we would hardly be finished with the traditional concerns of materialism. That a person has a purpose, for example, does not consist in any specific type of occurrent event or process; nor that her sins are forgiven, that she is in a state of grace, or that she is precious beyond rubies. And these are only examples about persons; what else may there not be between heaven and earth never dreamt of in materialist philosophy? I don't want to be fanciful, but merely establishing that sensations are brainstates seems hardly more than a drop in the bucket for the materialist. The virtue of such a ringing Thesis as "Matter is All" was to settle the hash of all such stuff once and for all.

Second preliminary question: does the description of the 'mental' or the psychological in terms of which the replacement thesis is formulated, do justice to its intended concern? Armstrong was rather more conscious than Place of the second preliminary question when he was debating Malcolm, a Wittgensteinian. Today he would also have to contend with putative failures of functionalism, arguments that no computational theory of consciousness could even in principle be successful, and demonstrations that truth conditions for belief attributions must have historical and social parameters outside the believer.

But leave these debates aside. Third question: supposing the empirical claim is false, or is scientifically investigated and found wanting, will there or will there not be a fall-back position to call 'the real materialism after all'? It would be a poor game if after much scientific strife, the loser could say "that's not it at all, that is not what I meant at all." Well, what if we accept Place's or Armstrong's formulation, and their empirical claims are found wanting? Suppose, for example, that no neurological process can be identified which can even in principle predict human decisions reached simultaneously or at the exact end of that process. The next empirical question would be what probabilities can be assigned to the (neutrally described) actions being decided upon, conditional on the states of the central nervous system. If these probabilities cannot even in principle be made as near zero and one as we like, is that the end of materialism?

Think of the exact parallel: no quantum state will predict the exact time of radio-active decay. Is that the end of materialism? It is not; and neither would materialism come to an end if what humans do could be related only probabilistically to their brainstates. A favorite belief of the materialists would have to be relinquished, but they would all know how to retrench. For the spirit of materialism is never exhausted in piece-meal empirical claims.

[3.4] The second move: whatever it takes

If you press a materialist, you quickly find that the most important constraint on the meaning of the Thesis is that it should be compatible with science, whatever science comes up with. This is contrary to what some of them say. If, they say, certain phenomena could not be explained purely in terms of material factors, then the scientific thing to do would be to give up materialism. But, holding the Thesis, they make the bold conjecture that this will never happen. That what would never happen?

If that question cannot be answered with a precise and independent account of what material factors are, there is still one option. That is to nail a completeness claim to science, or to a specific science such as physics. The instructive example here is J.J.C. Smart, who begins his essay "Materialism" with an offer to explain what he means:

He quickly discusses some older and more recent postulations in actual physics, which make that 'theory' look substantive. But of course the parenthetical qualification makes that discussion completely irrelevant!

Smart may believe, or think that he believes, the 'theory' here formulated; but if he does, he certainly does not know what he believes. For of course he has no more idea than you or I of what physics will postulate in the future. It is a truly courageous faith, that believes in an 'I know not what' -- isn't it?

Indeed, in believing this, Smart cannot be certain that he believes anything at all. Suppose science goes on forever, and every theory is eventually succeeded by a better one. That has certainly been the case so far, and always some accepted successor has implied that the previously postulated entities (known, after all, only by description) do not exist. If that is also how it will continue, world without end, then Smart's so-called theory -- as formulated above -- entails that there is nothing. Let's not be too quick to celebrate this demonstration of clear empirical content (about what the future of physics will not be like). Most likely Smart did not notice this implication and would have preferred to rephrase if he had.

In a clear indication that he is at least subliminally aware of the problem, Smart quickly adds some extra content. Not content with his initial formulation once he realizes that it is compatible with emergent properties, holism, and the irreducibility of biology to physics, he says

We should read this as an amendment of the above definition of materialism, for the 'theory' formulated above does not fit this bill. We must wonder how Smart knows that it is not adequate. Is he perhaps telling us that either physics will forever eschew emergent properties, or else materialism is false? Since quantum physics provides, at this point, a clear example of holism, should we conclude that materialism has already come to an end?

Of course not. Faced with the consequences of the stance that materialism should be whatever it takes to be a completeness claim for physics, Smart started backpedalling. Everything that is "repugnant" to him (to use his phrase) may be incorporated in future physics. So he adds, in effect, that physics will be false if that happens. But faced with that consequence, no materialist will stick by him if he sticks by that. They'll point out, quite rightly, that he was of a 'classical' mind, and like so often happens with the older generation in physics itself, quite unable to assimilate new visions of the structure of the material world.

[3.5] Materialism as false consciousness

So is it all just a matter of scientific reactionaries with their self-trivializing theses dressed up as uncompromising metaphysical constraints on science? No, it is not. For all this effort to codify materialism bespeaks something much more important: the spirit of materialism. Materialism is a hardy philosophical tradition, which appears differently substantiated in each philosophical era. Each instantiation has its empirical as well as its non-empirical claims, which interpret for that era, in its own terms, the invariant attitudes and convictions which I call here the 'spirit of materialism'.

How shall we identify what is really involved in materialism? Our great clue is the apparent ability of materialists to revise the content of their main thesis, as science changes. If we took literally the claim of a materialist that his position is simply belief in the claim that all is matter, as currently construed, we would be faced with an insoluble mystery. For how would such a materialist know how to retrench when his favorite scientific hypotheses fail? How did the 18th century materialists know that gravity, or forces in general, were material? How did they know in the 19th century that the electro-magnetic field was material, and persisted in this conviction after the aether had been sent packing?

Of course it is possible to measure certain quantities. But that cannot provide the criterion needed. Just think again of the transition from Cartesian to Newtonian physics. Newton identified forces as the causes of changes in states of motion. Accordingly, if you measure the direction and rate of change of momentum, you obtain a description of that cause in terms of its effects. (The recipe for measuring force direction and magnitude is exactly to measure those effects.) But it could be added consistently that these causes are immaterial, spiritual -- even mental, if Mind does not need to be someone's mind. If instead the forces are said to be material just like the extended bodies so classified before, the materialist must seemingly have some rather mysterious type of knowledge: a knowledge-that the newly introduced entities have the je ne sais quoi which makes for materiality.

But what is it then, in this metaphysical position, that guides the change in content, which it would be pedantic to signal with a change in name? If the "physicalist" or "naturalist" part of this philosophical position is not merely the desire or commitment to have metaphysics guided by physics -- i. e. something that cannot be captured in any thesis or factual belief -- then what is it? This knowledge of how to retrench cannot derive from the substantive belief currently identified as the view that all is physical. So what does it derive from? Whatever the answer is, that, and not the explicit thesis, is the real answer to what materialism is.

Hence I propose the following diagnosis of materialism: it is not identifiable with a theory about what there is, but only with an attitude or cluster of attitudes. These attitudes include strong deference to science in matters of opinion about what there is, and the inclination to accept (approximative) completeness claims for science as actually constituted at any given time. Given this diagnosis, the apparent knowledge of what is and what is not material among newly hypothesized entities is mere appearance. The ability to adjust the content of the thesis that all is matter again and again is then explained instead by a knowing-how to retrench which derives from invariant attitudes. This does not reflect badly on materialism; on the contrary, it gives materialism its due. But it does imply that only the confusion of theses held with attitudes expressed, which yields false consciousness, can account for the conviction that science requires presumptive materialism.

I mean this as a diagnosis of materialism, not a refutation. Its incarnation at any moment will be some position distinguished by certain empirical consequences, and these will either stand or fall as science evolves. But whether they stand or fall, materialism as general philosophical position, as historical tradition in philosophy, will survive. Given this, however, there can -- for that very reason -- be no question of regarding materialism as an assumption at the foundations of science. There is no 'presumptive materialism' which constrains scientific theories to consistency with certain determinate factual theses. For even materialism itself is not so constrained, and it survives by changing so as to accommodate the new sciences.

[4] Naturalism, materialism, scientism: where does science end?

Duhem conceived of science as neutral on all issues of metaphysics and religion. It seems to me that, properly understood, science is thus automatically. No special effort is needed to make science be this way, or to prevent it from becoming otherwise. But this holds only for the process of scientific inquiry and its products, scientific findings and theories. The practice of science may, as many other good and much needed work, be attended by collateral damage to other spheres of life. As paradigm of rational inquiry, for example, it may be extrapolated from what that literally means to a paramount guide to life. There may be psychological reasons for such effects of scientific education. But if science does take on this additional status, it can only be through mediation by certain attendant attitudes and stances which are in no way inherent in science itself.

[4.1] Satisfaction with objectifying inquiry

Scientific inquiry is the paradigm of 'objectifying' inquiry, in the sense I explained. It is possible to be satisfied -- or think that one is satisfied -- with such inquiry as sole cognitive approach to the conditions of our existence. This is an attitude, which may be expressed rhetorically as "What else is there to be found out, except what can be found out in this way? What more is there to understand, once we have all the information needed to provide the missing pieces in our theoretical puzzles?" Given this attitude, all wonder ends once scientific curiosity has been satisfied.

Several writers have taken this attitude to be the characteristic of secular life. On that premise, there is at least a logical possibility of being at once not secular and yet a committed (lay or professional) participant in the scientific enterprise. For after all, that attitude is not part of science itself, but circumscribes a role it can have in our existence.

However, this sort of satisfaction with science, which preempts wonder when scientific curiosity is satisfied, singles out at best one subclass of the secular. It does not charaterize any who seek meaning or significance outside objectifying inquiry. Such a quest may focus on art, poetry, aesthetic experience of any sort, nature apprehended as the Romantics taught, or on understanding through emotion, empathy, meditation, kinesthesis .... The varieties of religious experience too are so diverse that each of the above shades into them without any sharp break. Moreover, with the recognition of non-theistic religions, spiritual disciplines such as Zen and yoga, and the fervor of certain kinds of social ideals, we must almost speak of secular religions, so blurry is the borderline. This is not to say that there is no difference, but that we should probably be content only to chart certain important forms of secular orientation. It may be a mistake to try and draw an exact boundary that could be recognized as such from both sides.

[4.2] Naturalism and scientism

"Naturalism" is today a much more popular term among philosophers than "materialism". Epistemology has been naturalized by Quine, philosophy of science by Ronald Giere; Putnam has spoken to deaf ears with his "Why reason cannot be naturalized". Even to diagnose what it is to naturalize something is far from easy. To identify what naturalism is, apart from something praiseworthy, I have found nigh-impossible. Perhaps it is just the materialism I have been discussing here, under a different name. Alvin Plantinga recently offered a critique of naturalism, but says only that it entails that God does not exist nor any other supernatural being. Obviously this cannot be taken as definition, on pain of the circularity involved in characterizing "natural" in terms of "supernatural". I will not stop to analyze the texts of recent and current soi-disant naturalists and naturalizers to the extent that I have here scrutinized "Australian materialism". Yet I venture to assert: we see here too a position that only purports to be a factual thesis. Most likely it can not be identified with any factual thesis at all, but derives all its strength and support from attitudes that engender affinity with certain theses at each historical stage.

If this is correct, no such argument as Plantinga has recently given against naturalism can succeed. For this argument hinges on reflections on the probability of certain events on a supposition which includes philosophical naturalism. That makes sense only if naturalism is an identifiable thesis, a factual assertion. Even if the argument is internally correct, therefore, it is no more effective against naturalism than points about quantum field theory are against materialism -- they affect only a some temporary tradeable asset of the position.

Another term that comes to mind here is "scientism". The term was originally devised for philosophical attempts to cast science itself as a metaphysical system or as ideology, but also came to mean something like "science-idolatry". It should be sharply distinguished from positivism, even in its original form, though Comte (and later Ayer) certainly did not disentangle the two. Part of scientism is that very satisfaction with science (or more generally, objectifying inquiry) alone, such that wonder is preempted and no questions remain at all once scientific curiosity is satisfied. If this is combined with faith in the possibility of, or a felt need for, traditional philosophical system building then of course science must include any viable metaphysics. Absent such faith or need, the attitude can only be combined with positivism of one sort or another. With this stance, no questions remain at all once science has answered its questions. For those having this stance, there is certainly nothing to be apprehended or understood outside this sort of inquiry taken as a whole.

Neither metaphysics nor any sort of positivism or empiricism need carry such an attitude or cluster of attitudes, orientation, approach to life, or way of being in the world. That should be, it seems to me, a comforting point to all. The discomfort will be, I hope, that of losing the analogue to the naturalistic fallacy in this domain: the idea that such attitudes can ultimately be justified by advancing factual theses, whether empirical or non-empirical.

That said, let us now compare scientism with materialism as I characterized it earlier. Materialism purports in the first instance to be a thesis about what there is: all is matter. But sympathy with science, with its ever-striving, endless enterprise then poses a dilemma. Horn One: the thesis can be given the form of a specific scientific hypothesis. In that case it cannot be proposed as basis for a lasting philosophy. To be scientific, it must be yoked to science in progress, and so be hostage to the fortunes of future experience and future scientific development. This implies for every empirical thesis the prospect of being given up eventually, however well-grounded it may be in present science. Horn Two: the thesis can be given the form of a completeness claim for a specific science, say physics. But completeness claimed for today's science reduces to the previous case. Aiming the completeness claim at science in the long run empties it of content, since no one today can know what science will eventually be like.

We already saw J. J. C. Smart uncomfortably straddling the horns of this dilemma. Quite instructive here is Paul Feyerabend's essay on materialism (1963, with 1980 postscript). At the outset we find:

This suggests that any materialist thesis takes the form of a completeness claim for a certain scientific description of the world. It also suggests that philosophical disputes concerning materialism are entirely independent of the content of the thesis in question. If both suggestions are correct, there is already a serious question whether there is any such thing as materialism, the philosophical position construed as thesis.

Feyerabend gives several of the arguments found in various forms in Place, Smart, Lewis, and Armstrong, though without evident awareness of those discussions. (Feigl appears to be a common source with Smart.) He also allows for the possibility that no form of reduction of 'folk' psychology to physics is possible, in which case he favors eliminative materialism. But in the note added in 1980 he arrives at something like the conception of materialism as stance:

The point is that the completeness claim for physics (at whatever stage) is indeed a factual claim (that physics answers all the questions that it is designed to answer, all that can be formulated in its terms) but its role in philosophy is not just that. The factual claim conveys, in some 'Gricean' way, the satisfaction with the restriction to these questions and the dismissal of any wonder not thereby addressed.

Thus materialism shades into scientism. If materialism continues to be officially identified through the content of its theses, and yet gains significance through this intellectual stance, it is -- as I said -- a matter of false consciousness. If instead it is an acknowledged avowal, then arguments surrounding its theses are mere byplay. No wonder they are independent of the specific contents of those theses -- they serve only to remove some apparent obstacles! Feyerabend's choice of Democritean atomism as sample incarnation of the materialism he wanted to defend shows his characteristic courage, in displaying openly and exactly those weaknesses left out of the limelight. Materialism, the philosophical position, is in the main a stance that rests entirely on endorsement, avowal, and dismissal of certain approaches to life.

[4.3] Conclusion: the relation of metaphysics to science

At this point I would like to express my own, empiricist point of view. Firstly, there is no empirical claim which may not be contradicted by the eventual content (product) of science. Therefore, to accept the scientific enterprise at all, as one of rational inquiry to which we are committed, is to adopt an attitude of detachment with respect to any and all empirical claims. All such claims are to be readily surrendered as hostages to the fortunes of future experience. Therefore, it seems to me that they cannot and should not be part of any philosophical position. This goes for 'materialist' theses, but also of course, for 'anti-materialist' theses, whether about psychological functioning, physiological evolution or conservation of matter/energy.

Secondly, there is no non-empirical claim which matters at all to the process of science. The illustration of materialism generalizes to all metaphysics. I include here also theist, dualist, and idealist metaphysics, if any of these be coherent. Suppose some scientist wishes to write something to the effect that there is no God on the first page of his notebook. This may of course simply express some anti-religious sentiment. He could however intend it to be a metaphysical thesis with which all his scientific theories are to be compatible. Then he may write something like "Matter is all there is; God does not exist" and adds something equating God with the Omnipotent, Omnipresent, Omniscient Being as understood in metaphysical debates. (For contrast recall Pascal's amulet text "The God of Abraham, Isaac, and Jacob, not the God of the philosophers"!). When this scientist comes to the end of that notebook, he'll find that the thesis made no difference after all. At least, its content -- as opposed to its emotive force --played no role. The same thing would have happened if he had written its negation down instead.

Perhaps metaphysics has not always been, and perhaps it is not always even now an attempt to provide foundations or extensions of science, purportedly forged in the same mold as science itself, and thus in uneasy balance between rivalry and apologetics with respect to science. Perhaps, if wonder remains once scientific curiosity is satisfied, philosophy still harbors some response thereto; or perhaps, it too must then pack its bags and leave the field for other kinds of play. I do not know. But the picture of science deeply imbued with metaphysics, permeated by presumptive materialism and naturalism and thereby impervious to exploration of rival empirical hypotheses -- that picture is not accurate. In the main, Duhem's view of science is right.

BIBLIOGRAPHY

Armstrong, D. M. A Materialist Theory of Mind. London: Routledge and Kegan Paul, 1968; revised ed. with new preface New York: Routledge, 1993.

Armstrong, D. M. and Norman Malcolm, Consciousness and Causality. Oxford: Blackwell, 1984.

Buber, M. The Eclipse of God. Studies in the Relation between Religion and Philosophy. New York: Humanities Press, 1988.

Bultmann, R. New Testament and Mythology. Philadelphia: Fortress Press, 1989.

Burtt, E. A Metaphysical Foundations of Modern Science. London: Routledge and Kegan Paul, 1924; republished by the Humanities Press, 1952.

Collingwood, R. G. The Idea of Nature. New York: Oxford University Press, 1960.

Duhem, P. The Aim and Structure of Physical Theory. Tr. P. P. Wiener. Princeton: Princeton University Press, 1954.

Fackenheim, E. L. God's Presence in History: Jewish Affirmations and Philosophical Reflections. New York: New York University Press, 1970.

Feyerabend, P. K. Against Method. London: Verso, 1988.

Feyerabend, Paul K. Realism, Rationalism, and Scientific Method. Philosophical Papers, vol. 1. Cambridge: Cambridge University Press, 1981.

Feyerabend, Paul K. "Problems of Empiricism", pp. 145-260 in R. Colodny (ed.) Beyond the Edge of Certainty. Englewood Cliffs, N. J.: Prentice Hall, 1965.

Husserl, E. The Crisis in European Sciences and Transcendental Phenomology. Tr. D. Carr. Evanston, Ill.: Northwestern University Press, 1970.

Kant, I. Critique of Pure Reason. Tr. N. K. Smith. New York: St. Martin's Press, 1965.

Lewis, D. K. "Reduction of mind", pp. 412-431 in S. Guttenplan (ed.) A Companion to the Philosophy of Mind. Oxford: Blackwell, 1994.

Lewis, D. K. "An argument for the identity theory" Journal of Philosophy 63 (1966), 17-25; reprinted with additions in his Philosophical Papers, vol 1. Oxford: Oxford University Press, 1983.

Place, U. T. "Materialism as a scientific hypothesis", Philos. Review 69 (1960), 101-104.

Place, U. T."Is consciousness a brain process?" British Journal of Psychology 47 (1956), 44-50.

Plantinga, Alvin "Methodological naturalism?" (forthcoming in J. van der Meer (ed.) Facets of Faith and Science. Lanham, MA: University Press of America.

Plantinga, Alvin "Naturalism defeated" ms. 1995.

Plantinga, Alvin "Science: Augustinian or Duhemian?" (ms., 1994).

Plantinga, Alvin Warrant and Proper Function. N.Y.: Oxford University Press, 1993.

Smart, J. J. C. "Sensations and brain processes", Philos. Review 68 (1959), 141-156 (reprinted with revisions, pp.189-202 in his Essays Metaphysical and Moral. Oxford: Blackwell, 1987.

Stump, Eleonore "Non-Cartesian substance dualism and materialism without reductionism", Faith and Philosophy (forthcoming).

van Fraassen, Bas C. (1987). "Against Naturalized Empiricism", forthcoming in P. Leonardi and M. Santambrogio (eds.) On Quine. Cambridge Univ. Press, sched. 1995.

van Fraassen, Bas C. "Against Transcendental Empiricism", in T. J. Stapledon (ed.) The Question of Hermeneutics. Dordrecht: Kluwer, 1994.

van Fraassen, Bas C. "The World of Empiricism", pp. 114-134 in Jan Hilgevoort (ed.) Physics and Our View of the World. Cambridge University Press, 1994.