Doing things with logic

We have already discussed various jobs that relevant logic can do. It provides us with a theory of situated inference, a theory of implication, and the basis of a theory of conditionals. In this chapter we will use those theories to do other work that is of interest to philosophers, mathematicians and computer scientists.

Dyadic deontic logic

In chapter 6 above, we briefly discussed deontic logic. There we discussed the operator, ‘it ought to be that’. There are also logics that contain a dyadic deontic operator, ‘it ought to be that … on the condition that’. This is usually written in formal notation as ‘O(_/_)’. The formula, ‘O(B/A)’ is read ‘It ought to be that B on the condition that A’. In this section, I will argue that the dyadic deontic operator should be taken to be some sort of relevant counterfactual.

The link between counterfactual conditionals and dyadic deontic operators is quite striking. Like a counterfactual, the connection in a dyadic deontic statement between antecedent and consequence is defeasible. For instance, suppose that my friend Kevin asks to borrow a length of rope from me. I know that, on the condition that Kevin is a friend and that he asks to borrow a rope, I should lend it to him. But suppose that I then discover good evidence that Kevin is extremely depressed, so much so that I fear that he is suicidal. Then my obligation to lend Kevin the rope is annulled, since I fear that he may hang himself with it.

Introduction

Consider again Adams' examples of indicative and counterfactual conditionals:

Someone asserting the first of these sentences would assume that Kennedy was killed, but would not assume that Oswald did it. Someone asserting the second, on the other hand, assumes that Oswald did kill Kennedy.

We have a theory of how to evaluate the indicative conditional. We look at the circumstances that the people involved in the conversation in which it is stated can legitimately envision. If those people are taking this conditional seriously, then they will limit themselves to possible circumstances in which Kennedy is murdered. In those in which Oswald is not the murderer, someone else is. Thus the conditional is true.

I suggest that there isn't a very big difference between counterfactuals and indicative conditionals. I think that the way we evaluate them is very similar. In the case of the Kennedy counterfactual above, what we do is imagine ourselves in a situation in which we have all the salient information but not the information that Kennedy has been murdered (or even the information that he is dead). We then ask whether the indicative conditional ‘If Oswald does not kill Kennedy, someone else will’ is true in this situation.

The reason that we hesitate about answering this question is that we don't know what all the salient information is. We don't know what the CIA, Cuban Intelligence, the KGB, and the Mafia were planning.

INTRODUCTION

In recent years, doctrines of retentive physicalism that eschew comprehensive claims of nonphysical-to-physical type identity have faced the charge that they epiphenomenalize both mental phenomena in particular and special- and honorary-scientific phenomena in general. Roughly, the objection is that if, for every special- and honorary-scientific phenomenon, there is a physical phenomenon sufficient for it (as physicalism requires), and if all such underlying physical phenomena are completely caused by earlier physical phenomena in strict accordance with physical laws, then special- and honorary-scientific phenomena are just riding piggyback on the physical phenomena; it is the physical phenomena that are doing all the real causal work, and the appearance of causation among special- or honorary-scientific phenomena is just an illusion. So, because special- and honorary-scientific causation is surely not an illusion, any doctrine of retentive physicalism that implies that it is thereby faces the apparently damning objection that it has to deny an obvious truth.

In fact, it is not clear just how damning such an objection would really be. Imagine a doctrine of retentive physicalism that is committed to repudiating special- and honorary-scientific causation as illusory, but which can nevertheless explain why we would still believe in such causation, even if it did not exist; suppose, for example, that this doctrine of physicalism can predict the holding, given the actual physical facts, of just those special- and honorary-scientific regularities that we do in fact observe, and suppose also that we have a strong psychological propensity to infer from such regularities the existence of special- and honorary-scientific causal relations.

Gilbert Ryle once remarked that “there is no such animal as ‘Science’” (1954, 71). His point, of course, was not to deny the obvious existence of science but rather to emphasize the plurality of the sciences. Philosophers have sometimes made it seem as if there were only one science, namely, physics. But even a casual perusal of a university course directory reveals that there are plenty of others. For example, consider meteorology, geology, zoology, biochemistry, neurophysiology, psychology, sociology, ecology, and molecular biology, not to mention honorary sciences such as folk psychology and folk physics. Each of the many sciences has its own characteristic theoretical vocabulary with which, to the extent that it gets things right, it describes a characteristic domain of objects, events, and properties. But the existence of the many sciences presents a problem: how are the many sciences related to one another? And how is the domain of objects, events, and properties proprietary to each science related to the proprietary domains of the others? Do the many sciences somehow speak of different aspects of the same things? Or do they address themselves to distinct segments of reality? If so, do these distinct segments of reality exist quite independently of one another, save perhaps for relations of spatiotemporal contiguity, or do some segments depend in interesting ways upon others?

ORIENTATION

The main aim of this first chapter is simply to provide a clear and tolerably precise formulation of realization physicalism, the version of physicalism whose consequences and plausibility the remainder of the book examines. I postpone until chapter 5 and 6 the question of whether there is actually any evidence for or against it, contenting myself here with getting onto the table a formulation of physicalism definite enough to serve as a rallying point for its friends and a target for its foes. Doing so, however, also yields two desirable by-products: first, it rebuts the charge that physicalism cannot even be formulated adequately (so that any search for evidence in its support is premature); and, second, it reveals realization physicalism's thoroughly a posteriori character.

To get an intuitive grasp of what realizationism claims about the world, consider a humble can opener. Surely there is some good sense in which a can opener is a purely physical object. And yet can openers are in one clear sense not physical: fundamental physics does not speak of can openers as such, because “can opener” is not a predicate of any physical theory. So what is the sense in which can openers are physical? Well, it is plausible upon reflection to say that all it takes for there to be a can opener is for there to exist some object that meets a certain job description: that of having the capacity to help in the opening of cans (and perhaps of being designed or at least deliberately used for this purpose).

I began to think seriously about physicalism in the sense of this book in 1987, when I started work on my doctoral dissertation at Oxford. But by the time the dissertation had been accepted, in 1990, I had entirely lost confidence in the radically eliminativist version of physicalism that it defended. The result of my subsequent efforts to determine what was reasonable and what was not in my earlier views was a series of journal articles on physicalism; and when I began work on the present book, in the fall of 1998, I conceived my task as little more than that of revising those earlier papers, linking them up, and supplementing them here and there.

I conceived wrong. When I reviewed my earlier papers, they somehow seemed to say a lot less than I had remembered; and what they did say was often not quite right. So I decided essentially to start again. I also decided that I had to make an empirical case for physicalism, and in some detail, partly because I was irritated by the charge that adherence to physicalism is entirely unwarranted, but partly also because I had in the meantime grown skeptical about the possibility of a priori philosophical knowledge, and it seemed ludicrous to leave off the project of defending physicalism at exactly the point where, by my own account, it just starts to get interesting.

INTRODUCTION

According to realization physicalism, every causal or contingent token – past, present, or future – that falls within the scope of (R) is either a token of a physical type or else a physically realized token of a functional type; so realization physicalism is false if there is even one causal or contingent token – past, present, or future – that falls within the scope of (R) and that is neither a token of a physical type nor a physically realized token of a functional type. Let direct evidence against realizationism be any evidence that there exists precisely such a token – any evidence that (a) there exists a causal or contingent token of some type, T, such that T is neither (b) a physical type nor (c) a functional type all of whose tokens are in fact physically realized. And let indirect evidence against realization physicalism be any evidence against it that is not direct.

Whether there exists any direct evidence against realizationism is an entirely a posteriori matter. For it is obviously a posteriori whether there actually exists a causal or contingent token of some type, T. Moreover, as argued in chapter 1, it is also a posteriori whether T is (i.e., is identical with) some physical or functional type; in particular, this issue cannot be settled by appeal to (e.g.) thought experiments intended to validate some claim of conceivability and thereby of metaphysical possibility.

INTRODUCTION

The purpose of this chapter is to defend answers to a pair of questions of which the first is this:

1. Is realization physicalism reductive? That is, does it entail the reducibility of the nonphysical to the physical (given plausible additional premises concerning, say, the character of reducibility)?

My answer to question (1) is that we need to distinguish between importantly different senses of “reduction” and its cognates and that, although realizationism turns out to be nonreductive in some of these senses, it turns out to be reductive, and importantly so, in others. If this answer is right, then realization physicalism opposes the drift of nearly all the prophysicalist literature of the past twenty-five years, which has been toward a version of physicalism intended not to be reductive in any important sense (see, e.g., Fodor 1974, Boyd 1980, Post 1987).

There is, however, an almost universal consensus within professional philosophy that reductionism of pretty much any kind is false; there is also passionate opposition to reductionism from a wider intellectual public. Hence, given my answer to (I), I need also to address this second question:

1. (2) How damaging are the reductionist commitments of realization physicalism?

My answer to question (2) is that the forms of reductionism to which realizationism is committed are at least not open to armchair objections – objections that are a priori or based on empirical evidence available to a casual observer.