CHAPTER I

Empiricism and Ontology

§1. Consequences of Necessity. From beginning to end the concept of necessity is the crucial one in this book. Stress will be laid on two facts about necessity. The first concerns its existence: there are compelling reasons for recognizing that there are objectively based physical necessities. The second concerns its consequences: a number of important features of an ontology of the physical world are determined in view of there being such necessities.

Opposed to the view that the necessity of a proton to attract an electron or of salt to dissolve in pure water is objectively based is the view that it is psychologically, conceptually, or linguistically based. What necessity adds beyond truth is found in us, not in things. This rejection of an objectively based physical necessity results more from the adoption of an ontology that is inadequate to support such a necessity than from a direct attack on it. The ontology called for by objectively based necessity will then be a rich one indeed. In it there will be natures, and not just properties had either by nature or accidentally; there will be actions, and not just strings of conditions; and there will be things, and not just collections of conditions. The kind of necessity that can be accounted for when one or another of these factors has been rejected is a kind that can be accounted for only by appeal to psychological, conceptual, or linguistic factors.

An ontology that rejects things, for example, but retains conditions will quite naturally lead to a rejection of an objective base for necessity. If we carefully remove the dog’s conditions of being a quadruped, of being brown, of being overfed, and so on from the dog itself, we are left with such a thin layer of reality when we collect these conditions together that it is no wonder none of them has any necessary consequences in time. These excised conditions simply are what they are and thus they possess no powers to determine anything else.

The discussion of objectively based physical necessity would be clarified if it were recognized that many arguments against it have served only to conceal the fact that an ontology has already been adopted that is incompatible with it. The classical empiricist argument thene used by Hume–is that we do not know what we are talking about when we affirm or deny the necessity, in the sense of the objectively based necessity, of a physical connection, since there is no way of getting an idea of this necessity. It may be suggested that the impossibility of forming such an idea has to do with some limitation of our receptive faculties, but in fact these are not in question. Rather, hidden behind the talk about what ideas can be formed is an ontological principle that by itself eliminates this necessity from the world of entities accessible to knowledge.

The ontological principle is that in this knowable world “impressions” are the primary entities; but impressions share with conditions excised from things the kind of thinness of reality that prompted Hume’s famous maxim: whatever is distinct is separable. For if the primary entities (impressions) are like conditions excised from things, then any one of them could exist without the other–that is, they are separable. So, in the context of an empiricist view of the source of ideas, the rejection of the idea of necessity is based on an ontology whose entities are quite clearly incompatible with necessary connections.¹ Hume’s form of the rejection of necessity thus requires an ontology which restricts the knowable world to impressions and to entities that are dependent on impressions. Conversely, once we have established the need for an objectively based necessity in a world that is accessible to knowledge, Hume’s argument can be inverted to show that his ontology for this world is impoverished.

Granting the connection between necessity and ontology, what broad factors appear decisive in accepting ontologies that exclude necessity? My conjecture is that they have been epistemological factors. One of them is surely the high priority set on the avoidance of error. This priority favors those entities about which, for a given amount of sense experience, one is less likely to be in error, and is thus thought to favor impressions over things. Another factor is the high priority set on the avoidance of dogmatism about the physical world. One should, in view of this preference, reject any ontology that tends to slacken inquiry and to transform hypotheses into dogmas. Allegedly, natures, so conceived that they differ from empirical properties, are such that commitment to them has this negative (dogmatic) effect.

The cogency of both these reasons for an anti-necessitarian ontology is questionable. Why, after all, suppose that the direction of certainty is the direction of being? Perhaps being less likely to be in error about something is an index of its derivative entitative status. I see no reason to adopt either point of view: that entities must be in the domain of the certain or that they must be so hidden that we remain most uncertain about them. Whether entities of some category exist is independent of how much or how little a given amount of experience supports claims about them.

Also, I fail to see the connection between the dogma-inducing powers of belief in entities of a given category and their nonexistence. If indeed believing in natures induces dogmatic slumbers, then this means only th2t special efforts are needed to reinforce the attitude of inquiry. It can hardly mean that natures do not exist. The bogey of dogmatism will, for some time to come, afflict the upholders of natures. But it will eventually be clear that in an epoch of developed scientific thought the doctrine of natures has an important role, even though it must be denied the pretensions that it had in an epoch of less developed science.

Questioning the cogency of these reasons is not enough. What is needed in addition is some positive reason for accepting a necessitarian ontology. By far the most economical way to support it is to find a defense for necessity itself. Then merely unfolding the requirements of necessity would reveal the special elements of the ontology. These elements would not have to be defended one by one on their own quite apart from the support given them from the defense for necessity. The defense for all of them is contained in the defense of necessity. The concept of necessity is a crucial one throughout, not because necessity is seen as important by itself, but because it provides a window onto ontology. Necessity itself is, of course, not an entity, and it is, thus, not part of the ontology developed here. If a grain of salt necessarily dissolves in pure water, there is no entity, necessity, alongside the grain of salt and the dissolving. Rather, it is precisely because the objective base for the necessity of the salt’s dissolving belongs to the ontology that it is a “necessitarian ontology”.

It is important to emphasize again the kind of necessity that is to be defended here. It is a necessity that has a sufficient basis in the world quite apart from the mind. There is, then, no obvious connection, either way, between necessity and analyticity.* Asusme, for the moment, that there is a viable notion of a rule of meaning, and let being analytic mean having a denial that violates a rule of meaning. Then, for the analytic and the necessary to be co-extensive, there would have to be a miraculous correspondence between rules of meaning and the objective factors grounding necessity. As it is, where there is an objective basis for something’s being necessarily a certain way, there may well not have developed, at the intentional level, a corresponding rule of meaning. Thus there would be no analytic sentence corresponding to the necessity. And, conversely, a rule of meaning may mislead us as to the way things are, with the result that there is an analytic sentence corresponding to no necessity. Of course, in a non-necessitarian ontology, the surrogate for necessity will doubtless be analyticity, but analyticity gains little of the philosophical importance of necessity in this way.

§2. Practice and Ontology. What is there to go on in formulating an ontology of the physical world? There are many things that ought to be taken into account. But not all of them will play an equally basic role. For example, we do not want to ignore current scientific theory. But its suggestions are only as strong as its inductive warrant. A philosophical ontology gets its support elsewhere. For the ontologist, current scientific theory is limited to the role of a reminder. For an explicit conflict with current scientific theory would, at least initially, support the suspicion that there had been a failure to achieve the full generality required of an ontology. Of course, it does not follow that the ontologist realizes the aim of generality by looking for entities that are common to many–past and present–scientific theories. Rather, the aim of generality may lead to entities that are not explicitly part of any scientific theory.

Where, then, does ontology look for its content and its support? Limiting ourselves to the suggestions of the content of scientific theory requires us to limit ourselves to the suggestions of a propositional rendering of the physical world. Such a rendering does not exist in isolation from but in conjunction with human practices. At the very least, the content of scientific theories is that of judgments; not only is judging a human practice but it is also a facet of inferential practice. By a practice I mean something more than a type of activity. For a practice involves criteria for determining whether actions of the given type are warranted. (There is no need for the criteria to be the same for the practice to be the same.) So a human practice is a type of human activity subject to criteria of warrantability. The conjunction of theory–i.e., the content of theory–with practices is not just accidental. We cannot then proceed as though the propositional content of theories were altogether isolated from certain practices.

This suggests the possibility of looking to practice for at least some of the content of and perhaps even the support for an ontology. But do practices have ontological suggestions? There are certain familiar cases where they do. The practice of praising and blaming would seem to require that other people be treated as responsible agents. The practice of counting distinct objects would seem to require belief in something other than undivided stuffs. That these practices have ontological suggestions is clear from the fact that it would be incoherent of anyone who engages in them not to grant the existence of certain entities. There are, of course, changes in practices, and, in view of the requirement of generality, we want more from a philosophical ontology than a recapitulation of the practices of a given culture. Moreover, provincial practices could not provide the support needed for an ontology held to be universally valid.

Are there, perhaps, practices without which it would be difficult to conceive of any other practices? If there are, then since the content of scientific theory is never completely separated from practices, these “basic practices” would transcend differences in the content of theories. Their ontological import would provide not only a framework for practice but also for theory. Should there be a theoretical content common to every scientific theory and without which no such theory would be conceivable, this content would, in general, be additional to and not part of the ontology drawn from such a basic practice. The ontological import of a basic practice need not be part of the content of any theory.

The practice of acting on the basis of, that is, in the light of, prior sense experience has a solid claim to being essential to any battery of practices and will be used as the basis for a later defense of physical necessity. We shall agree that actions based on prior experience concern possible combinations of features in new situations that would be similar to features that were the objects of prior experiences. I act on the basis of prior unpleasant experiences when I insist on having a commitment of management in writing. Similarly, I act on the basis of prior experiences of Jones when I accept as true the proposition that his new venture will succeed. I may or may not remember, when I act, prior experiences of broken commitments or of Jones. So remembrance cannot be appealed to in order to distinguish such action from action on habit.

Though a habit is developed by prior experience, to act on habit is not to act on the basis of (that is, not to act in the light of) prior experience. One does something on the basis of prior experience voluntarily, but if the same thing is done purely out of habit the action is removed from the realm of the voluntary. Also, flexibility of response is important if one is acting on the basis of prior experience, whereas action on habit tends to be invariant under what might otherwise be important differences in the agent’s situation. Nonetheless, in acting on the basis of prior experience, the agent need not be propositionally aware of the subtle factors in the situation that influence the choice of the action. Suppose there were no practice of action on the basis of prior sense experience. Prior experience could be utilized only in action on habit. Could there then be any human practices at all?

A consideration of the broadest categories of human practice reveals ways in which practices in these categories are dependent on the practice of acting on the basis of prior sense experience.

(1)The practice of communication shows such a dependence. One communicates with an interlocutor on the basis of certain assumptions made about the interlocutor. The assumptions are quite different for different interlocutors. So the flexibility required is not supplied entirely by a single habit applied to all interlocutors. Yet there is normally not a special habit for each case. One is then acting on the basis of prior experience with interlocutors in making certain of these assumptions. Of course, as familiarity with a stable set of interlocutors grows, the assumptions about each may become habitual. But if one is to communicate and not merely make noises in a less familiar situation, then one needs to estimate what means will be appropriate, on the basis of one’s own or on the basis of someone else’s prior experience. That an entity like the straw scarecrow in Oz was, at first, unable to communicate because it did not know how to work its mouth, rather than because of its limited experience with interlocutors, is an entertaining reversal of the priorities.²

(2)There is a like dependence in the practice of making judgments. The genesis of this practice is bound up with the need to make our responses (at least those that are based on prior experience) to troublesome situations more discriminating through the resources of propositional thought. The point is not simply that judgment is occasioned by problems. If this were all, action based on prior experience would not necessarily play a role. Rather, judgment refines responses that are already based on prior experience. It enters at a point at which the unlearned and habitual responses have already been added to by the more flexible responses based on prior experience. The flight of thought to abstract matters from its role in guiding concrete responses in no way establishes the possibility of judgment apart from having an origin in this concrete role.

(3)Social regimentation, discipline, and punishment clearly belong to a category of practices that involve action on prior experience. One regiments where prior experience indicates that behavior would otherwise take a different course from the one desired. Even regimentation by so-called a priori standards is to be understood in terms of this dependence. For standards come to be called a priori when their long entrenchment obscures the fact that they were originally thought to be warranted in the light of still earlier experience.

(4) Ceremony and art are sometimes held to transcend calculations serving material interests and thus to be independent of action on prior experience. Yet it is obvious that ceremony and art are possible only in an environment in which, if it has not been won, the struggle for survival has at least achieved a temporary freedom from immediate needs for some few individuals. Labor practices are then presupposed. The techniques needed for any labor practices will have been developed by actions based on prior experience.

In all these cases and in any others, I suggest that the possibility of the practice in question rests on the general possibility of the practice of acting on prior experience. The notion of possibility here is not systematized in the ontology that follows. It is a notion of “practical” possibility in the sense that it concerns the interrelations of practices. These practical possibilities, I contend, do not rest on any special scientific assumptions. Nothing is assumed about the make-up of brains, the nature of sound waves, or the nature of matter in the environment that labor struggles with. Thus these possibilities are not undercut by scientific thought experiments. But of course it is assumed that the agents engaging in these practices are physical persons and have a material environment, for it is human practices we are talking about, not those of angels.

An ontology based on a fundamental practice has certain advantages over other ontologies. For example, an epistemological ontology–one whose entities are chosen because we are less likely to make mistakes about them–can at most reveal what is involved in the practice of looking for the experiential bases of claims that go beyond experience. Such a practice is clearly not one that is needed to make all others possible. It is limited to a concern with the grounding of propositional thinking in experience, and many human practices have no dependence on propositional thinking. Still there are those who, without explicitly considering wider alternatives, erect the epistemological criterion into a norm of philosophical conscience.³

A scientific ontology is also limited. This will be so whether we conceive a scientific ontology as one whose entities are required by the content of current science, or by some conception of what the content of science will ultimately be, or even by the content of any science. For, the ontological requirements coming from practice, rather than deriving from a theoretical content, may be missed by a scientific ontology. This will be true even for the last of the above forms of scientific ontology, which is the most general one. Moreover, those features of the world that make practice such that it can be carried on without illusion will be the context for those features, if any, that are implied by the content of a wide variety of scientific theories. This is not to deny that one is doing philosophy when one does scientific ontology in that most general form. The Kantian study of the requirements of the content of any science–of the conditions of the make-up of empirical knowledge–is undoubtedly philosophical. But the undertaking is more special than the one envisaged here since it is concerned exclusively with the content of propositional thought. Even so, our undertaking is limited also, for it considers the implications of only one of possibly many basic human practices.

§3. Moderate Empiricism. The practice of action on prior experience covers very disparate actions. On the one extreme there are actions satisfying physical needs. On the other there are the relatively detached acts of judging to be true certain propositions that claim for as-yet unobserved entities what has been noted for observed entities. Making such a judgment is making an induction, and the canons of induction are relevant to its evaluation. It seems clear that there are prudential considerations involved in evaluating non-judgmental actions that are not relevant to evaluating judgments. I may, for example, warrantedly believe something even though it would be foolish of me to bet my last dollar on it.

Still, for our purposes, there are common factors that are quite important. If the judgment that a proposition is true is an acceptable induction, then the experience on which it is based supports to a certain degree, confirms to a certain degree, or increases by a certain amount the probability of that proposition. Similarly, a non-judgmental action based on prior experience is warranted in view of that experience only if that experience supports, confirms, or increases the probability of its success. The notion of support by experience is, then, integral to the notion of the practice of action on prior experience. Support, of course, is not itself an action, as accepting a proposition is. Still, support is something that a proposition or a project should have in order to be accepted or to be undertaken. But support is only a necessary condition for warranted action. This is obvious when conflicting propositions or projects are equally well supported.

Even though action based on prior experience is warranted only if supported by experience, it does not follow directly that “extreme rationalism” is wrong. For the extreme rationalist, the act of accepting a proposition does not require, for the acceptance to be warranted, that the proposition have support from sense experience. There may be other ways of accepting any proposition than that of proceeding on the basis of prior sense experience. However, once we consider that the practice of action based on prior experience is a fundamental practice, it turns out that extreme rationalism cannot be accepted. There are two steps to the argument for this conclusion.

First, judgment, like other practices, is dependent on the practice of action based on prior experience. There would not be judgments unless some judgments made responses based on prior experience more discriminating. Judgments in this role are an integral part of the more discriminating responses, and they are not things outside those responses. Thus since the responses that are guided by judgment are based on prior experience, the judgments will be based, in part at least, on the prior experience that the responses are based on. It must then be the case that some judgments are based on prior experience.

Second, it has still to be shown that some specific judgments must be made on the basis of prior experience. Could not each judgment that is actually based on prior experience also be made warrantedly on some a priori grounds without any change in the basis of judgments actually made on a priori grounds? If each one of them could be made a priori, would there be any constraint that would make it impossible for all of them to be made a priori? If there is no constraint of this kind, then the practice of judgment would, contrary to fact, be independent of the practice of action on prior experience. In an economy of a given kind, there may be no individual who must be poor, even though in that economy it must be the case that there are poor. Such an economy provides the constraint needed to guarantee poverty despite the possibility of individual betterment. In our case, there is no comparable constraint. If each proposition could be supported for acceptance in an a priori fashion, then all of them could be accepted in an a priori way. This forces us to conclude that some judgments can be supported only in an empirical way. Otherwise, the result of the first step would be false; it would be possible that no judgment would be made on the basis of prior experience. Thus extreme rationalism is false.

It is not necessary to accept the strong empiricist standard that any proposition pertaining to the as-yet unexperienced can be accepted only if supported by, among other things, previous sense experience. The practice of action on prior experience can be fundamental even though some propositions about the as-yet unexperienced are acceptable on the basis of support from a priori considerations. However, the fundamental nature of this practice does require at least a “moderate empiricist” standard. This standard asserts that some propositions about combinations of features in the as-yet unexperienced are acceptable only if they are supported by, among other things, previous sense experience of similar features.

Accepting propositions in a way that accords with moderate empiricism is a practice of a judgmental՞ sort and is therefore limited. We have seen, however, that even non-judgmental action on prior experience requires that there be support, from such experience, for the proposition that the project will succeed. So any ontological requirements resulting from the adoption of moderate empiricism as a standard for accepting propositions about the as-yet unexperienced will also be requirements for the practice of action based on prior experience. Accordingly, I shall attempt to find ontological consequences of moderate empiricism and treat them as parts of an ontology of the practice of action based on prior experience.

This claim of generality for the standard of moderate empiricism might appear unjustified in view of the debated nature of the inductive status of science. Science, of all things, seems to be inductive in nature. Its propositions seem to require empirical support for acceptance. But if science is in fact not inductive, then to take seriously the idea of acceptance based on empirical support as the starting point for an ontology is to risk building the ontology on sand. Popper has argued vigorously that induction has no place in the logic of theoretical science. As an explanatory hypothesis, a theory can be characterized as having passed the text of experience so far, but as an explanatory hypothesis, it is irrelevant to characterize it as judged true or accepted in the light of experience.⁴ Hence it is irrelevant to characterize it as being strongly enough supported or confirmed to be judged true. As explanatory hypotheses, theories are corroborated, not supported or confirmed.

The position I have taken would not be undercut if Popper were right. There is then no call to argue with him. For Popper, the theorist, as theorist, is a contemplative in regard to his theory, so long as it is unrefuted. The theorist will act, on the one hand, to accept simple experiential claims and, on the other hand, to reject a theory conflicting with such claims. But the theorist will not act to judge a theory true. Even if theorizing proceeds on this contemplative level, there are other occupations within the domain of science itself that do involve the act of accepting claims that go beyond previous experience. In this very regard, Popper himself recognizes the difference between theory and technology.⁵

Before the atomic bomb was exploded, scientists were asked if exploding it would destroy the face of the globe. Their report was a negative judgment on the global destructiveness of the bomb, and not merely an appraisal of the past performance of a theory. So even if induction is not deemed relevant to science as limited to the contemplation of a set of explanatory hypotheses, it is relevant to science in other respects. It is even relevant to science as a world-view, for a scientific world-view is a judgment as to how things are, and the judgment is made on the basis of collective experience. There is, then, no reason to suppose we were wrong in taking action based on prior experience as a fundamental practice. And hence no reason to suppose that an ontology based on moderate empiricism is insecurely founded.

When rationalism is so conceived that pure reason is called on to secure only basic principles, rationalism need not conflict with moderate empiricism. Descartes seems to have thought that basic laws could be both discovered and justified a priori. Still, he thought there were many different worlds compatible with these laws. Experience was to play the role of deciding which among these is the actual world.⁶ Any true general proposition not derived from laws alone would have support from experience. That the amino acid alanine occurs most commonly in higher life forms only in the configuration that rotates light to the right would not be a priori since it rests in part on the boundary conditions of the evolving world. This is not to say that general propositions can be empirically supported when all the factors involved are assumed to be contingent. There would be no empirical support, according to the view I shall develop, for the generalization about the configuration of alanine unless the boundary conditions were assumed to unfold their consequences according to laws.

But there are forms of rationalism that do conflict with moderate empiricism. To that extent they give a misleading picture of human practice. Platonic rationalism extends both to the study of physical nature and to the affairs of the citizen. A theoretical claim about physical nature is not, for Plato, capable of being supported by experience. He thought that experiments of the fundamental sort needed to give support to theoretical hypotheses were humanly impossible.⁷ So a physical claim is acceptable, in the sense of being plausible, on the basis of its agreement with some mathematical construction. The Platonic theory of the elements, for example, derives its plausibility from being based on Theaetetus’ construction of the five regular polyhedra. There is no reason to suppose that human practice in general could not adjust to this Platonic practice, in science, of relying on a priori considerations.

But for Plato reliance on the a priori is extended to the concrete affairs of the citizen. The citizen obtains a modicum of assurance about the future because there are “wise” men in the state.⁸ And the wise man is one whose proposals are supported by a vision of the moral forms. No room is made for the experienced man as a moral leader, that is, for the man whose ideas are warranted by his rich experience. Between wisdom, as intellectual vision, and skill, as inflexible habit, there is no place for the competence of experience. There is either action informed by moral patterns or the rigid response of good or bad habit. But can action informed by moral patterns be deemed possible apart from action based on prior experience? Such action would be conceivable in the context of human practice only if at some point the adoption of some moral pattern seemed warranted in terms of certain prior experiences. The over-all scientific and moral rationalism of Plato’s thinking is at odds with the fundamental nature of the practice of action on prior experience, and hence with moderate empiricism.

We have noted that the practice of making judgments about the world on the basis of prior experience is only one aspect of the practice of acting on the basis of prior experience. One can act on the basis of prior experience without affirming the proposition that something will happen or that things are generally of a certain kind. Suppose there are, or were, cultures with no activity of affirming propositions about the world. The notion of support by prior experience would still be relevant so long as there remained action on prior experience. It has been argued that, in the mythopoeic thought of ancient Egypt and Mesopotamia, there was no affirming of propositions about the world.⁹ Whatever was present to awareness was so as part of the emotional context of myth. On this interpretation, descriptive propositions were not among the intellectual components of mythopoeic thought. But even when awareness of nature has this mythopoeic character, it does not follow that action in response to immediate problems is based on myth alone. It may still be appropriate to say that some action is based on prior experience, and hence to say that the notion of support is relevant to an analysis of the evaluation of actions.

A specific concept of experience is needed to prevent this discussion of action based on prior experience and of moderate empiricism from being purely schematic. Certain views of experience are so limiting that the practice of action based on prior experience could not be a fundamental practice. I shall review three of these briefly. But beyond rejecting these views, I shall add nothing that will make the concept of experience more specific.

First, the view that experiences are indubitable is not accepted here. Experiences are indubitable if they must be experiences of just what they seem to be experiences of. If the objects of experience are sufficiently restricted-for example, to what philosophers call sense data-then one might seriously argue that they are indubitable. But then there would be no justification for claiming that the practice of action on prior experience of such objects is basic to all action. So here an experience may seem to be an experience of a certain condition when it is not an experience of that condition. An experience of the condition of being hot-for exampie, the condition of a sample of water of being hot-may seem to be an experience of the condition of being cold of the sample of water.

This is equally true of the experiences that I may refer to in support of a hypothesis. They may well support the hypothesis if they are the experiences that I take them to be. But they may not be experiences of what I take them to be experiences of. Without assuming that experiences are indubitable, we must assume that they are experiences of what we take them to be experiences of when we say that they support an hypothesis. Hence, to say that some propositions are acceptable only if they are supported by experience is not to imply that experience is indubitable.

Second, the view that experiences are invariably propositional is not accepted here. That is, it need not be the case that when I have an experience I judge or in any way entertain a proposition. Otherwise, the entertaining of propositions would be basic to all practices, which it clearly is not. Still, non-propositional experiences can play a role in supporting propositions. For example, the condition of a vixen behaving slyly can, of course, be experienced, and this experience can be appealed to in support of a general hypothesis which implies, among other things, that a condition of this type obtains in the given circumstances.

Now seeing the vixen behaving slyly and seeing it having a bushy tail are perceptions of distinct conditions and thus have distinct objects, even though the same fox has both conditions. Similarly, seeing that the vixen behaves slyly and seeing that the vixen is bushy tailed are perceptions that involve, in some way, distinct propositions. But from this similarity it does not in any way follow that experiencing conditions is the same as the entertaining of propositions. Conditions–such as the water being hot and the vixen behaving slyly–are, as I shall argue below, extra-mental entities, whereas propositions are intentional. So even if experiencing-that is universally judgmental, it in no way follows that experiencing a condition involves making a judgment.

Still, experience is not to be distinguished from thinking by excluding propositions from experience. Many experiences may well involve the entertaining of propositions. A discussion of the distinction between sense experience and propositional thought would involve an elaboration of the notion of an “affect” as that indispensable component of experience that involves the body in experience and that stimulates sense awareness.

Third, the view that experience, when it is of physical things or their conditions, must be of middle-sized, everyday things or their conditions is not accepted here. Moderate empiricism does not then mean that certain propositions are acceptable only if supported by reference to data about middle-sized, everyday things. The important factor in delimiting experience is not the size or familiarity of its object, but the fact that an affect is involved in becoming aware of that object. The affective element can accompany, at least in principle, awareness of entities that are quite disparate. The philosopher’s sense-data and the scientist’s microsystems might also be objects of experience, depending on how the relation between the sources of affectivity and the content of experience is ultimately defined.

§4. Realism and Materialism. The ontology derived by setting out from necessity is put forth as a realistic one. Its elements are to be mind-independent components of the physical world. They are not categories of a mind that organizes nature. The emphasis on practice and experience might seem to suggest, however, that the conclusions must be about the structure of consciousness. Now, in respect to our basic practice, there are two related but distinct questions that could have been asked. First, what must a conscious being be like for the practice of acting on prior experience to be one of its practices? Second, what general views of the physical world would it be reasonable for someone adopting the practice of acting on prior experience to hold? It is the second question that interests us, and although answering it involves setting out from a consideration of practice, it involves ending up with claims about the physical world.

But is it not entirely possible that the world is other than what it is made to seem to be by the practice of acting on the basis of prior experience? Must we not then abandon any categorical commitment to realism for the merely conditional claim that, if the world conforms to the presuppositions of the practice of acting on prior experience, then such and such entities are in the world? To make this retreat would be to ignore the serious consequences of asserting the possibility of a split between this practice and the world. If this practice were of minor importance, if we could limit the number of practices dependent on it, then we might well admit that the world could conflict with the beliefs that this practice makes undeniable. In fact, however, the practice of acting on prior experience is fundamental in respect to other practices. Without it they would be left in the air. Yet to affirm its basic role is to grant the beliefs about the world that are made undeniable by this practice. Nothing in the vast cultural superstructure of practices dependent on it can make legitimate the possibility of conditions conflicting with those beliefs.

The view that this practice cannot be relied on to determine ontology faces a clear dilemma. Suppose there is the possibility of a conflict between the world and the beliefs that are made undeniable by the practice of action on prior experience. Then this possibility can be derived either from considerations arrived at by practices that are independent of the practice of acting on the basis of prior experience or from those arrived at by practices that are dependent on it. In the former case, it is assumed that there are practices that could go on if there were no practice of acting on the basis of prior experience. And this assumption we have denied above. In the latter case, we are asked to grant that by speculative reasoning, for example, the possibility could be derived and warrantedly asserted as a possibility, even when it is recognized that speculative reasoning is a practice that is not independent of its role in guiding action based on prior experience. Thus, since it is limited by the context of action on prior experience, speculative reasoning cannot coherently challenge the beliefs made undeniable by such a practice. Of course, a theory based on a given practice may predict a change, in the world, to conditions that are incompatible with that practice. But then these conditions are not alleged to have existed when the practice was current. More generally, theory may project conditions in which no human practice of any kind is possible. But all that is denied here is that theory can justifiably say that such conditions might hold while there is human practice, of which it itself is an instance. In sum, the possibility of a conflict between the world and a basic practice cannot be justified. There is then no reason to modify the categorical character of our realism based on practice.

Our ontology is not intended as a complete list of all the categories of entities there are. It is thus not offered as a “full” ontology. It is not even intended as a list of all the categories of entities in the physical world, and so it is not offered as a full physical ontology. It can, however, be called a “required” physical ontology in that it is the physical ontology of a practice required by all other practices.

The standpoint from which a required ontology is written can be called “critical.” It is not based on a cultural status quo, be it commonsense or scientific. It involves the assumption that the ontological suggestions of a fundamental practice are more reliable than those of either common sense or current science. The worldview of common sense needs to be scrutinized in the light of other approaches to the world. And speculative alternatives need to be viewed not just in the context of the content of current science but in the context of practices basic to both science and common sense.

The required ontology–that is, the one that is the physical ontology of the practice of action based on prior experience–is distinct from both materialism and idealism. Let us think of materialism here as an ontology that (1) recognizes only those entities that the propositions of physical theory could be true of, (2) interprets these propositions as true or false of entities that the propositions are standardly taken to be about, rather than of the sense data that the positivist takes them to be about, and (3) treats these propositions as incapable of being true or false of minds, actions, and middle-sized (ordinary) things. Thus the entities that the materialist puts in the domain of physical theory are nonmental, connected by relations rather than involved in actions, and devoid of sense qualities.

But the practice of action based on prior experience, and hence the ontology that is required by this practice, is compatible with, even if it does not contain, many entities excluded by materialism. Thus one might consistently hold the required ontology while agreeing with C. S. Lewis that each of the entities that are parts of the myths we make here on earth has actuality somewhere else in the universe. For the benefit of materialists, Lewis conceived of a rational being, the eldil, that inhabited outer space.¹⁰ It had no body in the gross sense, but was a mere luminous presence. The point is that even if there are no eldila, a fully critical ontology does not busy itself with showing that there are none.

On the other hand, idealism, as the ontological negation of physical things, their components, and their conditions, is incompatible with the existence of human practices. These practices are carried out by physical persons in physical settings. The idealist asks us to perform an act of speculative thought resulting in the judgment that there are no human practices but only practices of entities that not only, like eldila, lack ordinary bodies but are completely disembodied. Yet speculative thought itself is a practice that depends on human practices in physical settings. Any judgment arrived at by speculative thought that rejects the physical setting is unwarranted.

Ironically, although idealism prides itself on being critical, it is critical only in a limited epistemological sense; it gives high priority to making only those commitments that, for a given amount of data, are least likely to be in error. If we accept this priority, we may well be trapped in our own minds. But setting this priority is itself uncritical. It is a priority having to do with the practice of making judgments about what is. And this practice exists only in a context of practices that clearly involve a physical setting. The priority is set without regard to dependence on this context.

Materialism is, to an extent, a critical ontology in that it has gotten where it is by going behind the superficially plausible claims of a commonsense ontology. But it exhausts its critical animus with this first step. Its subsequent lapse into dogmatism leaves it with little or no advantage over commonsense ontology. For at least what common sense lacks in theoretical sophistication it makes up by containing practices widely different from the merely descriptive ones of physical theory. It is open to the commonsense ontologist to emphasize both the implications of these practices and those of the merely descriptive apparatus of common sense.

If materialism, as a critique of common sense, based itself exclusively on the deliverances of current physics, it would be clear what the source and strength of the critique are. But as often as not the source of the critique is an idealized extension of current physics. When confronted with the question of whether current quantum theory, with its systems characterized by superpositions of apparently non-actual states, really supports the materialist’s conception of objectively existing material entities, the materialist is apt to wave a hand at the “inelegant mess” of quantum theory and speak about the possibility of new developments in microphysics.¹¹ Those principles that determine how science will come out are, then, just the ontological ones that science was to justify. The ontology turns out to be a dogma, and the critical attitude toward common sense reduces to pretense.

In the writing of Wilfrid Sellars, for example, the critical attitude toward common sense is, he strongly suggests, supported by science. But a closer look leaves one unclear from what quarter it is supported. After seemingly defending the commonsense view of time as involving the perspectival characters of past, present, and future, Sellars advances the higher wisdom of science. “The non-perspectival structure which, as realists, we conceive to underlie and support perspectival temporal discourse is, as yet, a partially covered promissory note the cash for which is to be provided not by metaphysics (McTaggart’s C-series), but by the advance of science (physical theory of time).”¹² If the realist issues promissory notes that are to be made good by science, where does he get his authorization? Not from science, since science makes no predictions on where it will come out. But we are not told where else. Also, Sellars recognizes an event ontology, in distinction from an ontology of physical things, as a regulative ideal. This, however, is only an ideal and not an established belief partly because “science has not yet achieved the very concepts in terms of which such a picture might be formulated.”¹³ To say that science will justify an ontology that undercuts commonsense amounts to nothing unless some authorization outside contemporary science is forthcoming for this statement.

In a required ontology the claims can be traced back to a fundamental practice. They are not left as unsupported dogmas. A materialist ontology can, when it is taken from current physical theory, also claim support from certain practices, as well as from the content of current theory. But these practices will not all be basic. Thus arises the possibility that ontological claims that do receive support in relation to such limited practices might have to be rejected when the context is broadened to include basic practices. Consider, for example, the following claim. “Speaking as a philosopher, I am quite prepared to say that the commonsense world of physical objects in Space and Time is unreal . . . .”¹⁴ What Sellars takes as support for this view is the greater descriptive and explanatory power of current microtheory in comparison with common sense and the fact that in principle the theoretical expressions of this microtheory could replace commensense expressions in an observation language.¹⁵

It is assumed by Sellars that what is not required by the practices of making descriptions and explanations in terms that can be observational, when these practices are considered by themselves, is not required by anything. But one can make this assumption only by systematically ignoring the fact that description and explanation have a setting amidst other practices and that they are dependent on other practices. Perhaps commonsense objects do not belong to a full physical ontology, but all the relevant considerations should be in before they are excluded.

* Hence, it is not to be objected that we have begun to speak freely of necessity without having bothered to clear up the numerous troubles that are associated with analyticity.