I am quite interested in winning new friends for the views that this book defends – as many new friends as possible. Accordingly, I have written for a group that is somewhat more inclusive than the community of professional philosophers. I hope the book will find readers among undergraduate philosophy students and also among computer scientists, neuroscientists, and psychologists. (Some readers may find that the second half of Chapter 2 and the middle of Chapter 3 are a bit taxing. I hope they will persevere. A partial understanding of this material is entirely sufficient for an adequate grasp of the main themes of the book.)

My main debts are to Ivan Fox and Sydney Shoemaker, each of whom has been extremely generous with his time and wisdom. They have saved me from many errors, and pointed the way toward many ideas that would otherwise have eluded me. Although it is somewhat paradoxical, since they both disagree with many of my main contentions, it is nonetheless true that such value as the book may have is largely due to them.

I have received encouragement and valuable help of other kinds from Anthony L. Brueckner, Anil Gupta, Richard Lee, William G. Lycan, R.J. Nelson, Hilary Putnam, Vola Shulkin, Lynne Spellman, and Chris Swoyer. Each has played an important role at several stages. I am also indebted to Clyde L. Hardin and Brian McLaughlin, who commented on the penultimate draft.

This chapter is concerned with questions about the forms and limits of our cognitive access to visual sensations, and with questions about the semantic properties of the concepts on which such access depends.

It is convenient to distinguish at the outset between two types of visual sensations. First, there are sensations that have the power to induce us to form perceptual beliefs about the visually perceptible properties of extramental objects and events. The members of this group, which will hereafter be called belief-generating sensations, include all of the sensations that occur in the course of everyday veridical visual perception. But the group has other members as well. Specifically, because we can be led to form perceptual beliefs – or, if you prefer, quasiperceptual beliefs – when we are dreaming and also when we are hallucinating, the group includes sensations that are involved in dreams and hallucinations. Second, there are visual sensations that lack the power to induce us to form perceptual beliefs. These sensations include after-images of various kinds and also the highly amorphous visual sensations that come to the fore when we close our eyes. I will refer to members of this second group as perceptually barren sensations.

In defining belief-generating sensations I said that they are sensations that have the power to generate perceptual beliefs.

My goal in Part 2 is to convince the reader that type materialism deserves to be taken seriously. I will work toward this goal by establishing three propositions. First, in this chapter I will show that, under certain assumptions, type materialism is to be preferred to two of its most significant rivals – dualism and a view that is often called the double-aspect theory. Second, in the next chapter, I will argue for the proposition that its other main rival, a view known as functionalism, suffers from serious defects. And finally, two chapters hence, I will establish that the main objections to type materialism are misguided.

When J.J.C. Smart and other materialists of the 1950s and early 1960s set out to defend their favorite doctrines, they were typically more concerned to answer objections than to construct positive supporting arguments. It seems to have been generally felt that materialism has a certain intrinsic plausibility that competing theories lack, and that as a result, once the objections to materialism were answered, the burden of proof would shift to the shoulders of the advocates of other theories. Thus, instead of giving carefully formulated positive arguments, the materialists of Smart's era relied mainly on sketchy appeals to simplicity and terse complaints about the obscurity and messiness of competing views.

This pattern has persisted to the present. Very little has been done to improve the arguments of the materialists of Smart's day.

THE SIGNIFICANCE OF VAGUENESS

Contemporary philosophy is indebted to Peter Unger for reviving the Sorites paradox. Unger argues that few, if any, of the things in our standard ontology really exist. Unfortunately, his arguments have not been taken seriously enough. I suspect that this is because he has been given too extreme an interpretation by some and too trivial an interpretation by others. In proclaiming himself a nihilist, he gives the impression that he believes that nothing exists – all of reality comes to nothing. It would seem that anyone who holds this position is too divorced from perception and reason to be taken seriously. But the Sorites paradox need not commit anyone to this extreme a nihilism. Even Unger is prepared to admit that something exists, but he thinks that in order to find out what that something is, and in order to be able to talk about that which exists in any detail, we must go through a fundamental conceptual change. He asserts that he is not prepared to offer any workable alternative conceptualization, but he does not claim that there could not be one.

On the other hand, Sorites arguments can instead be understood as being relevant only to language and not to reality. It may be admitted that the words ‘heap’ and ‘stone’ and ‘person’ do not apply to anything because of their vagueness. Yet this fact should not lead us to doubt the existence of the things that we had thought we could refer to with the words ‘heap’, ‘stone’, and ‘person’.

TO PARAPHRASE OR NOT?

I have argued that the standard ontology is false and should be replaced by the hunk ontology. Of the four alternative ways to handle vagueness, three of them were unacceptable. However, the alternative that I end up accepting is likely to seem far more extreme than the three that I rejected. My goal in this chapter is to eliminate some of the feeling of extremism associated with the rejection of the standard ontology. One approach, common in contemporary philosophy, would be to present a paraphrase. In this case I would not have to reject the standard ontology so much as redescribe it. The goal would be to interpret our everyday utterances in a way that commits their users only to the objects of the true ontology. If done correctly, those everyday sentences that seem to be true will in fact be true once they are paraphrased, and those that seem to be false will in fact be false once they are paraphrased.

There are some questions as to what follows from our ability or inability to produce an adequate paraphrase. If the paraphrase project can be completed, does that show that we have been talking about hunks all along, or does it only show that our old false nonhunk talk can be replaced by our new true hunk talk? If the latter, why is this valuable? Is it because it presents a means for getting us to start saying true things, or because it provides an explanation of why it is not so bad for us to keep saying false things?

COINCIDENT ENTITIES

Coincident entities, if there were such things, would be distinct objects occupying exactly the same space at exactly the same time. The most difficult examples of coincident entities to explain away are those in which the two objects in question exist for exactly the same time and in exactly the same space for all the time that they exist, but differ in their modal properties. Consider two lumps of clay, one of which is shaped like the bottom half of a statue and the other like the top half of a statue. At the moment that those lumps are stuck together two distinct objects are brought into existence. One is a larger lump of clay composed of the two lumps. The other is a statue. When the original two lumps are again separated both the larger lump and the statue cease to exist. The larger lump and the statue are spatially and temporally coincident, but they are not identical. For instance, it is true of the lump that it could have had a completely different shape, but this is not true of the statue.

In spite of such very plausible examples, coincident entities would be very odd things. If there were any, they would have to have every spatial region in common at the time at which they were coincident. At that time, therefore, they would have exactly the same molecular substructure.

THE GENERAL CAMP

The ontology of physical objects I will defend in this work is that of four-dimensional hunks of matter. Some of these hunks are temporal parts of others. Thus, I place myself in the same general camp as Willard Van Orman Quine, John Perry, and David Lewis. Lewis mentions a common objection to such an ontology, and begins to answer it:

This text is devoted to developing an ontology of four-dimensional hunks of matter. I argue that every filled region of spacetime is exactly filled by one such object and that any one of these objects has its actual spatiotemporal configuration and location at every world at which it exists. This ontology should be contrasted with what I take to be our standard ontology, according to which one and the same three-dimensional object exists in its entirety at several times and at several worlds, having a different spatiotemporal shape and location at many of these other worlds. My arguments can be taken to support either of the following conclusions: (A) the standard ontology should be rejected and the hunk ontology accepted; (B) the standard ontology must really be the hunk ontology rather than the ontology of three-dimensional objects that exist at worlds at which they have different configurations and locations. For the most part I present my arguments as a defense of (A), but I will also show how to read them as a defense of (B).

Since I hope to argue for the hunk ontology over the “standard ontology,” it is natural that this book should contain both a constructive project and a destructive project, though it is crucial that the two projects not be completely separated. Much of the constructive project is closely connected to the work of David Lewis and much of the destructive project is closely connected to the work of Peter Unger.

The action of opening a door consists in the agent's voluntarily exerting parts of her body – her arm and hand, let us suppose – in such a way that that action (the voluntary exertion of the body) causes the door to open. It is possible, in principle, to open a door without using any voluntary exertion to do so. Conceivably, a person's brain could be so wired to a door that merely by mentally saying “Open sesame!”, and without any exertion of her body, she could cause the door to open. There are ways in which one can act on or with one's body without engaging in any voluntary exertion. (For example, sexual arousal can be produced by forming appropriate mental images.) And our lives are filled with mental actions – mentally saying things, forming mental images, and the like – that, though they may affect the body, do not in themselves and in virtue of their very notion include any bodily event, whether voluntary exertion or other. But it is nonetheless true that actions done by voluntarily exerting the body, together with voluntary exertions themselves, comprise most of the actions that we have occasion to consider explicitly. Our voluntary exertions of our bodies are a central and especially important sort of action. We must understand what they are if we are to understand most sorts of actions we talk about.

Is freedom of the will compatible with determinism? I think not, and I would like to explain why.

First, let me explain what I take this venerable question to mean. By freedom of the will is meant freedom of action. I have freedom of action at a given moment if more than one alternative action is then open to me. Two or more actions are alternatives if it is logically impossible for me to do more than one of them at the same time. Two or more alternatives are open to me at a given moment if which of them I do next is entirely up to my choice at that moment: Nothing that exists up to that moment stands in the way of my doing next any one of the alternatives.

OUR IMPRESSION OF FREEDOM DOES NOT MAKE US FREE

We all continually have the impression that our wills are thus free. For example, at most moments while I am sitting and talking or listening to someone, I have the impression that there are several alternative things I could do next with my right hand: gesture with it, scratch my head with it, put it on my lap, put it in my pocket, and so on. For each of these alternatives, it seems to me that nothing at all up to that moment stands in the way of my making it the next thing I do with my right hand; it seems to me that what has happened hitherto, the situation at that moment, leaves each of those alternatives still open to me to perform.

The philosophy of action deals with the notion of action that applies only to beings who have wills. (The words action and act are, of course, applied to other sorts of entities – we speak of the action of the acid on the metal and of how the pistons act to move the drive shaft – but what action and act mean in such applications is not a concern here.) We take ourselves, people, to be the paradigms of enwilled agents. That we are beings who act is a fundamental fact about us. It is as important as the fact that we are cognizers, beings who know and believe. (These two aspects are, of course, thoroughly interconnected: The knowledge and beliefs we have and those we want influence our actions, and our actions influence the beliefs we come to have.) Yet, though philosophers in the Western tradition from Socrates on have had much to say about action, it is only in the last three decades or so that the theory of action has come to be thought of as a distinct branch of philosophy, on a par with the theory of knowledge. This may be because there was a tendency to think of action as a subconcern of philosophizing about rationality and morality. And indeed, very important questions about action arise there, such questions as what it is to choose one's actions rationally and what makes it the case that a person is morally responsible for an action.

What is it for a person to act? It is easy to give examples. I act when I voluntarily move my limbs, when I open a door, when I speak or write, press keys on the keyboard, slice a melon, or throw a ball, when I mentally say a word or mentally rotate a visual image. But not all events or states of which a person is the subject are actions. There are, for instance, perceptions, sensations, desires, beliefs, feelings, unbidden thoughts, faintings, sneezings, tremblings, reflex actions, and states of passivity. What distinguishes actions from these other sorts of things? What is the mark of action? Answering this question is not so easy.

REFORMULATING THE QUESTION

It is useful to have a standard way of referring (in English) to particular personal events and states, including actions. A suitable form of a singular noun phrase can be derived from any indicative active sentence that predicates of a person some event or state simply by changing the subject to its possessive form and changing the verb to the present participle. Thus the sentences “Sue suffered during the race” and “Tom started the engine five minutes ago” become, respectively, the singular noun phrases “Sue's suffering during the race” and “Tom's starting the engine five minutes ago”.

To be the canonical designator I want, such a noun phrase must pick out a particular personal event or state uniquely.

The preceding chapter argued for incompatibilism. In this chapter, I want to rebut two arguments against incompatibilism that have been put forward from time to time. Incompatibilism can be defined this way: Any free action must be an undetermined event. A free action is one such that, until the time of its occurrence, the agent had it open to her to perform some alternative action (or to be inactive) instead. An undetermined event is one that was not nomically necessitated by the antecedent state of the world. (Hence, a determined event is one that was nomically necessitated by its antecedents.) An event was nomically necessitated by the antecedent state of the world if and only if the antecedent state together with the laws of nature determined that that event, rather than some alternative, would occur.

The most widely supported argument against incompatibilism, to which I will give by far the larger response, combines the consideration that a free action can be influenced by the agent's intentions, desires, and beliefs – can have an explanation in terms of reasons for which the agent did it – with the assumption that only a determined event can have such an explanation. My response to this argument will be to counter the assumption by offering an adeterministic or anomic account of such explanations. The other argument does not assume that reasons explanations are deterministic (nor does it assume the contradictory), but simply claims that where we have an undetermined action we do not have an agent in control of (determining) what her action is to be: We do not have an action that the agent chooses, freely or otherwise.