So also are there many ways of unifying them. One way is to note that they share some feature in common, say a concept of truth that applies to them all (§§1.1-1.6). They therefore share the logical prerequisites for truth, however much they differ in how best to conduct the trial by prerequi-sites. This unity of method is compatible with enormous diversity from domain to domain as regards ways of trying to reach agreement on whether the prerequisites have been satisfied, and additionally on whether various claims are true. The domains, or many of them, may well be incommensurable with respect to ways of resolving disagreements, yet entirely commensurable in the sense of having the same truth-concept.

Not that such commensuration is "ultimate," if indeed any feature is. The domains share still other features in common, which on occasion are more significant. For instance, every domain is such that in many contexts we may freely choose to use its vocabulary rather than some other. We may choose to use the language of the marketplace or the bedroom, rather than the language of science. Our choice can have enormous effects on us and others, especially if we persist in using one sort of vocabulary to the exclusion of others. "Man is a self-defining animal," not least because of the effects on himself of his chosen vocabulary. Accepting responsibility for such choices often is far more urgent than pursuit of truth.

Not even conversation is "the ultimate context" within which the domains may be arranged. For there are other legitimate interests and purposes we sometimes pursue that can also result in a unification. From the time of Thales one such interest has been in finding some one kind of thing of which everything else is somehow a manifestation such that the way former is determines how the latter will be. Thales thought the one thing was water. How quaint. Yet for his day it was a brilliant guess. Water was the only substance known to exist in three key forms, liquid, gaseous and solid. Positing water as the stuff of which everything is composed therefore had a certain explanatory power, and a certain power to unify the sciences of the time, which seem to have been rather compartmentalized. What intrigued Thales' contemporaries was his idea that there is one underlying stuff, rather than his belief that it is water. Later philosophers denied that it is water, but retained the presupposition that despite all differences, everything is somehow a manifestation, not necessarily via composition, of some one kind of existent, or some limited variety of kinds, not necessarily kinds of "underlying stuff" or "substance."

This presupposition, in various forms, has inspired efforts to unify the sciences ever since. It has also inspired efforts to unify all the domains, not just the sciences. Many a metaphysician has sought unity in some limited variety of existents or processes, a unity according to which the properties of these unifiers determine the properties of all the diverse things in which they are manifested. Candidates for the basic unifiers have included Anaximander's "indefinite," Parmenides' unchangeable plenum, Democitus' atoms and the void, the Pythagoreans' "numbers," Plato's Forms, Aristotle's matter and substance plus its movers, Decartes' two substances, Spinoza's one substance, Hegel's Spirit, Whitehead's actual occasions of becoming, and so on and on. No wonder Collingwood concluded that Greek monotheism and much Christian theology were meant to express and perpetuate Thales' presupposition, which had to struggle against the polytheism of the day.

What, according to the physicalist, is supposed to make physicalism superior to all these other varieties of metaphysics? First we must ask, "Superior in what respect?" The physicalist's answer must be that it is superior not only as regards unifying the domains in terms of a limited variety of entities manifested in all the diverse things the different domains are specifically about, but as regards unifying them in such a way that how these diverse things are is determined by how the unifying entities are. Thus physicalism is alleged to be superior with respect to achieving an aim shared in common by most metaphysicians from Thales on. Whatever the priority some such metaphysics may enjoy, physicalism included, it is conditional on the interest or value of that aim.

What such metaphysicians share in common is neither some view of mind as the mirror of nature, contrary to Rorty, nor some related foundationalist view of knowledge, nor even some theory of truth. Systematic philosophers engaged in unifying the domains in terms of a limited variety of entities are interested primarily in what there is. They can differ enormously on how we are supposed to know it, and on how truth and our knowing (if any) are related to what there is. They can even agree with Rorty and others (as I do) that all we can ever know is what-there-is-under-a-description.

At the same time they can differ on the status of the vocabulary used to describe the basic or unifying existents. Doubtless most such metaphysicians have supposed the vocabulary must enjoy some unconditional priority, since they often suppose it describes the nature of things, some underlying essence, which provides the ultimate commensuration for all the domains. Some have gone further and supposed the vocabulary should be used in all inquiry. And some have even supposed that all culture should be remade along the lines of a model or paradigm implied, according to them, by the privileged vocabulary in its intended use. When they suppose any of these things, obviously, they commit monopoly (§4.6).

Whatever these philosophers and their opponents may sometimes have thought, positing some unifying entities of which everything else is a manifestation need not entail that there is any such final vocabulary. Physicalism, of all things, affords an illustration. Minimal physicalism does not require that physics describe the nature of things, the only real world, or an underlying essence, even assuming there are essences. A monism of entities is compatible with a pluralism of properties, of worlds, of faces of existence, of values, and of equally privileged vocabularies. It is compatible also with a pluralism of modes of insight and explanation expressible only in other domains. All this is true of other varieties of metaphysics when they do no more than posit certain unifiers. To echo the deliberately paradoxical saying in §4.5, it is quite possible both that nothing but the unifiers exist, and that not everything is nothing but one of the unifiers. A metaphysician must make further assumptions, beyond his inventory of unifiers, to be forced into committing monopoly. Unfortunately many metaphysicians do make the further assumptions.

Minimal physicalism is superior to the views of many of these metaphysicians, then, at least as regards avoiding monopolistic pretensions. But there are still other metaphysicians in this multi-stranded tradition who may also avoid committing monopoly. What is supposed to make physicalism superior to their views? To begin with, it must at least be true that the posited unifying entities do exist. Yet many of those posited by other metaphysicians in the tradition can be seen not to exist, according to the physicalist. For the claim that they do flunks the trial by prerequisites.

The objectivity-prerequisite is especially telling in this regard. In §1.5(vii), and later in §3.4, we saw how the claim that a certain thing exists can be true period only if its identifying properties are invariants, rather than properties of a projection of something else onto a system of reference. We saw further how the non-invariance of ordinary simultaneity would thereby force us to deny objective existence to such things as "the ongoing present" and the "now" with which the present contents of consciousness are simultaneous. Yet nearly all traditional metaphysicians -- including traditional atoms-and-the-void materialists -- posit some kind of unifying entity or event or being or process whose identifying properties are not invariants (as may well be the case with Whitehead's "actual occasions of becoming," at least as they are interpreted by some of the later Whiteheadians). Such unifying entities do not enjoy objective existence. At best they are projections or appearances on our screens of reference. Nor is it at all clear that the spirit of these traditional views could be retained in any revision that takes adequate account of the implications of contemporary physical theories of spacetime.

Nevertheless, some nonphysicalist inventories of unifiers might prove to satisfy the truth-prerequisites, including objectivity. If so, what would be the physicalist's response? Not only must the unifiers exist, it must also be possible to unify all the domains of discourse in terms of them, and to do so in such a way that truths about the unifiers determine the truths in all the domains. But to unify them all one must at least unify the sciences, whereas in fact the truths in the various sciences are not determined by truths about nonphysical entities (or by alleged truths about alleged nonphysical entities). Instead they are determined by truths about mathematical-physical entities. Or so the physicalist is prepared to argue.

I shall say something about the premises of this his argument in a moment. Notice first that with the help of an additional premise, the argument justifies far more than its immediate conclusion. Its immediate conclusion is merely that even if the alleged nonphysical unifiers existed, the domains still would not all be unified in terms of them. Now assume as an additional premise that the domains are all unified in terms of mathematical-physical entities, in such a way that all truth is determined by mathematical-physical truth. Then there is a definite sense in which everything is accounted for on this mathematical-physical basis, as we noted in §4.3. But in that case no further entities would be necessary, and Occam's razor comes into play. We would have no need of the hypothesis of nonphysical unifiers. Hence we would be justified in denying there are any in the first place. This goes far beyond the conclusion merely that the domains would not be unified in terms of them.

Those who would champion nonphysical unifiers must therefore challenge the physicalist's belief that it is not necessary to multiply entities beyond the mathematical-physical. In connection with the above argument, this entails challenging the physicalist's premises. The key premises are two. The first is that the truths in the sciences are not, as a matter of fact, all determined by truths about nonphysical entities. The second is that all truth, not just truth in the sciences, is determined by mathematical-physical truth.

Because it takes a theory to kill a theory, the first can be challenged successfully only by presenting a whole metaphysics -- a competitor to physicalism -- according to which the truths in the sciences are all determined by truths about something nonphysical. This competing view would have to be fully clarified and defended. Its candidates for the unifiers must exist. Its notion of determination would have to be explicated and defended against charges of reduction and monopoly. Finally, we would want to see at least some of the details about how particular truths in a particular science are supposed to be determined by truths about the alleged unifiers, as we do see in the next section in the case of physicalism. Sweeping generalizations would be no more persuasive here than in the service of physicalism. At every step along the way the anti-physicalist would be haunted by the realization that minimal physicalism might already have succeeded admirably in unifying the sciences. Thus he would have to argue extra hard for his extra entities. One such argument might be that they are required for determining truths in domains beyond the sciences.

This brings us to the physicalist's second premise, according to which all truth, not just truth in the sciences, is determined by truth at the level of physics. This second premise is fatally easy to challenge so long as determination entails reduction. It is easy to challenge also if determination entails that the modes of explanation in all the domains must be those employed in physics, or at least those employed in natural science. But as seen, determination entails neither reducibility nor explainability nor the other kinds of monopoly.

Only one challenge remains. How could we ever tell in practice whether particular truths in one domain are determined by truths in another, if not by way of some sort of reducibility or explainability? This question is familiar from §4.3. We address it, at long last, in the next section. But many truths contain metaphors and other figures often thought to defy any physicalist approach. In §5.3 we consider how the physicalist might argue that such truths also are determined ultimately by physical truths. In Chapter 6 we do the same for truths that contain value terms, especially principles of morality. Truths about subjective experience are considered in §§5.4 and 7.1. Still other truths outside the sciences come in for treatment in Chapters 7 and 8.