Before any sort of analysis or criticism of Taylor’s argument can be undertaken, we have to try to get clear on just what Taylor is claiming, and on the best ways perspicuously to represent those claims. For one thing, it is vital to recognize that the kinds of implications and relations treated of here by Taylor must be regarded as physical and causal, not logical, even though Taylor maintains, confusingly, that “Our problem has been formulated without any reference whatever to causation,”8 in an attempt to head off potential objections about causes working in only one temporal direction. The thing to see is that the “→” acts differently in:I-5) (OrderO→Battle B)
from the way it acts in, say:I-6) ((p ∧ q) → p).
The arrow in (I-6) is the arrow of material implication and expresses what Hume would call a mere “relation among ideas”; it’s uninformative. In contrast, though, the arrow in (I-5) tells us something about the world. There is nothing about the “concept” of my giving order O that contains or logically entails the occurrence of battle B tomorrow. The relationship asserted by (I-5) to obtain expresses a “matter of fact”: O implies and is sufficient for B precisely because there is some nomic, causal, physical dependence-relation between O and B that rendersBanecessary consequence of O.
It is also important to see, though, that the internal relations of sufficiency and necessity between antecedents and consequents in (I-5) and (I-6) are identical in form, if not in force. That is, (~Battle B → ~Order O), just as (∼p → ∼(p ∧ q)). To explicate the clear difference in force, we need say only that the conjunction of (p ∧ q) and ∼p constitutes or implies a logical impossibility (a contradiction), while the conjunction of O and ∼B under Taylor’s scenario constitutes or implies some sort of physical, or causal, or non-logical impossibility, some state of affairs that, together with the background facts of the case, would violate some of the rules we regard as characterizing the way the physical world operates and the way causes work in the physical world.
These considerations should make it clear that we also have to be careful to distinguish the kinds of modalities Taylor is concerned with in his argument. I claim that the modalities we are dealing with here are physical. It is fairly uncontroversial to draw a distinction between alethic, logical modalities (logical necessity and logical possibility), on the one hand, and causal, physical modalities (physical necessity and physical possibility), on the other. That there are significant differences between the two kinds of modalities is easy to see. If it is logically impossible for me to be both a human being and a quartz crystal, it is physically impossible for me to be both a human being and a quartz crystal. But, although it is physically impossible for me to fly unaided by flapping my arms wildly, it is not logically impossible for me to do so. We usually say that logical necessity is “stronger” than physical necessity, and that logical possibility is “weaker” than physical possibility.
But there is a further distinction to be drawn, that between two at least potentially different types of physical modality. There seems to be a difference between what is just physically possible in general and what is physically possible for a given agent to do in a given set of circumstances. The former might be termed physical possibility simpliciter. It concerns what is and is not consistent with the laws of nature per se. (It is physically impossible simpliciter to travel faster than the speed of light. It is physically necessary simpliciter that for every action there is an equal and opposite reaction.) The latter type of modality, on the other hand, might be termed situational physical modality. It concerns the modal character of events, actions, and states of affairs, taking into account not only general and unchanging physical laws, but also the situations and circumstances that can affect what is possible and necessary at certain places at certain times. For instance, exactly three weeks ago it was situationally physically possible for me, at3:50pm, to lay both hands on the front wall of Amherst College’s Johnson Chapel. Today it is not now possible for me to lay both hands on the Chapel at 3:50, because it is now 3:49.30, and I am not even in Massachusetts. It is, of course, physically possible simpliciter for a human being to lay hands on our Chapel, and this fact never changes, so long as there are hands and a Johnson Chapel. Any constraints on the physical possibility of touching the chapel would have to be situational constraints.
These sorts of considerations might lead one to argue that there is really no such thing as a physical modality simpliciter, divorced from the context of a time-and-situation. This may or may not be true, but it is not very important for my purposes here; my aim has been to introduce and characterize the notion of situational physical possibility. I would ask the reader merely to recognize the intuitive difference in modal character and modal force between: “It is not possible for me to be both a human being and a quartz crystal”; “It is not possible for me to travel faster than the speed of light”; and “It is not possible for me, now in Champaign, Illinois, to be touching a building in Massachusetts thirty seconds from now.”
The reason why the distinctions are important has to do with considerations of time and circumstance. Logical modalities are determined by the model-theoretic characteristics of formal systems (or, if you prefer, the “meanings of logical particles”). The logical modalities that hold in our world we regard as eternal and unchanging. Physical modalities simpliciter, if there are such things, are determined exclusively by those laws we regard as characterizing the invariable natural processes of the physical world. Since we may reasonably assume that these laws do not change over time, we may regard physical modalities simpliciter, if there are any such beasts, as also eternal and unchanging. Situational physical modalities, however, are determined by the physical laws under which our world operates, and by the sets of prevailing circumstances that bear on what is in fact possible at a given place and time. And since sets of prevailing circumstances quite obviously vary with time and with the situations that obtain at different times, we can see that situational physical modalities enter into relationships with times and situations-at-times that logical modalities and physical modalities simpliciter, by their very natures, do not.
Given these sorts of distinctions, we should be able to see that the kinds of modalities we are concerned with in an analysis of the Taylor problem must be regarded as situational physical modalities. The alleged entailment-relation between O and B is not logical, and there is no contradiction in (O ∧ ∼B). And O does not physically ensure B simpliciter; rather it does so because of, in Taylor’s own words, “the totality of other conditions prevailing,”9 and because of the stipulated causal efficacy of O with respect to B in the situation characterized. The “can” and “cannot” variously applied to the admiral’s possibilities of action concern what is possible, under certain laws, for the admiral to do, given the presence or absence at certain times of certain other states of affairs. I thus claim that the modalities we are concerned to examine in the context of Taylor’s argument are the truly interesting kind of physical modalities, the situational modalities. All references to what is “possible” and “necessary” in the remainder of this discussion of the Taylor problem should, if things are not made explicit, be understood to refer to situational physical modalities. If “□” and “◊” are standardly used to stand for logical necessity and possibility, respectively, here “□” and “◊” will be used to denote the kinds of situational necessity and possibility with which I understand Taylor to be at least implicitly concerned.[1]
As I’ve roughly sketched them, situational physical modalities can be seen to enter into relationships with time that most other types of modalities do not. And since, as can be demonstrated, these relations quickly become extremely complex, confusions in semantic evaluations of physical-modal propositions can arise very easily. A goal of this essay will be to show that the Taylor problem is really nothing more than a classic instance of such a confusion.
II. THE TAYLOR LITERATURE: SOME PROMINENT REPLIES TO TAYLOR, AND WHY THEY HAVEN’T WORKED VERY WELL.
/>
There can be little doubt that Taylor’s argument is a strange one. How licit is an argument from linguistic, semantic, and logical premises to a thoroughly metaphysical conclusion? It is precisely this move from semantics to metaphysics that I am ultimately going to attack in this essay.
Most of the philosophers who have published responses to Taylor, however, have not seen the answer to his puzzle as lying in this direction. Taylor offers an argument for an obviously unattractive and anti-intuitive conclusion. Almost every philosopher who has contributed to the immense Taylor literature has agreed on the need to reject that conclusion. But where most have tried to justify that rejection by disallowing one or more of the presuppositions that serve as the explicit or implicit premises in Taylor’s argument, here I am going to try to bend over backwards to accept Taylor’s premises, to grant him everything he seems to want in the argument, and then to show that the conclusion he desires still does not follow validly from that argument. This is the project in outline. What I propose to do first is to look briefly at a representative sample of the published attempts to solve the Taylor problem—almost all of which seek to disallow premises—and to try to point out why they have not met with complete success, in order to shed light on some of the motivation behind the different kind of analysis I will be attempting.
One approach to defusing the Taylor argument—the one Taylor himself appears to favor in his Metaphysics—is to attack presupposition 1, to deny, as we saw Łukasiewicz do, that LEM/PB is properly applied to future contingents. Taylor writes that this escape from his conclusion might be effected either by assigning both B and B′ some intermediate truth-value between 1 and 0 that still combines in disjunction to yield the alethically-necessary truth of (B ∨ B′), or by asserting, by Destructive Dilemma out of the original order-and-battle argument, that both B and B′ are in fact false. “One need not,” Taylor says, “reject also the law of contradiction by seeking this means of escape, for the law of contradiction assures only that, concerning any statement, not both it and its contradictory can be true.”10 But Steven Cahn, the most vocal public defender of Taylor’s original paper, “Fatalism,” points out that the problem with this approach is that since B and B′ are presented as necessary for O and O′, respectively, then, if neither B nor B′ is actually true, then necessary conditions for both O and O′ fail to obtain, and it looks as though the poor admiral is able neither to give the order nor not give it, which is obviously even worse than the original fatalistic conclusion.11 Cahn argues quite convincingly that rejecting the first presupposition is not the way out of the problem, and Taylor did in fact accept the conclusions of Cahn’s paper in a short commentary appearing in the same journal.12
It is not surprising that the Taylor-presupposition which has come in for the most sustained vigorous attack is his fifth. A popular sort of criticism has focused on the relations between necessary conditions for actions and agents’ power or ability to perform actions. Taylor himself acknowledges that the oldest sort of criticism of his type of fatalistic argument has been the sort that maintains that the occurrence of the battle tomorrow is a necessary condition only for the giving of order O, not for the admiral’s power or ability or freedom to give order O.13 Under this analysis of (O → B), the denial of B would have as a modus tollens consequence only the fact that I the admiral did not give the order, not that I couldn’t have. Taylor rejects this sort of objection on the grounds that he never in fact asserted what it objects to. The necessity-relation between B and O is but the logical mirror-image of the sufficiency-relation between O and B. Taylor maintains that the occurrence of the battle tomorrow is obviously a necessary condition only of my giving the order today, not of my having the ability to give the order today, for, were the latter true, my mere ability to give the order today would by itself be a sufficient condition for the occurrence of the battle tomorrow, which looks to be absurd (although, if we think about it, it is not incompatible with the fatalist’s collapse of all potentiality into actuality, with his collapse of what we “can” do into what we “must” do). The point is that Taylor insists that he is talking only about the actual giving of the order, and maintaining that this giving cannot take place in the absence of some state of affairs necessary for its taking place, with the “cannot” here referring to what we might call circumstantial constraint, something which I will argue does indeed have a very legitimate bearing on situational physical modalities.
Just what Taylor means by “circumstantial constraint” may become clearer in the context of a related objection to presupposition 5 advanced by Bruce Aune.14 How proper is it to say that I “cannot,” have not the ability to, perform an act if I am absent a condition necessary for that act? For instance, it is often a necessary condition of my doing p that I try to do p, but do we say that in those cases when I do not try to do p it follows that I lack the power or ability to do p? In order for me to play tennis, it is necessary that I have a tennis racket, but does the absence of a racket mean that I lack the ability to play tennis? Quite clearly not. But Taylor and others15 have reasonably replied that the “cannot” of the Taylor problem is a “cannot” of circumstances, not really of ability, that it represents an exterior, not an interior, limitation. (This sort of distinction goes at least as far back as Hobbes’s Leviathan (M.IV. pp. 272-278).) A world-class pole-vaulter in a room with a ceiling ten feet high “cannot” pole-vault eleven feet. Rod Laver “cannot” play tennis without a racket. This does not mean that they lack the ability to do these things (where “ability” here means some sort of blend of know-how and physical wherewithal), but Taylor maintains plausibly that one’s power to perform an action is often a function of exterior circumstances, too, and some of these circumstances appear under his reasoning to be the presence or absence at appropriate times of causal consequences of the action in question.
There are other well-known objections to presupposition 5. John Turk Saunders has argued that the premise is covertly dependent on a “radical revision” of our ordinary-language use of such words as “power” and “ability,” and that it is only if we accept these radical revisions that Taylor is able to push his unattractive conclusion through.16 Saunders offers the following sort of counter-argument.II-1) A necessary condition of my swimming at t1 is (according to Taylor) the water being turbulent at t2.
II-2) The water is not turbulent at t2.
II-3) Therefore it is not in my power to swim at t1 (according to Taylor).
II-4) BUT I have at t1 the ability, the know-how and wherewithal, to swim.
II-5) AND Conditions for the free exercise of that ability are normal at t1: I’m not drugged, bound, conked on the head, the waters are swimmable, etc.
II-6) SO I do have the power to swim at t1.
Saunders claims that (II-6), which follows directly from (II-4) and (II-5), is, “if we use language in ordinary ways,” manifestly true, and that thus (II-3) is false, and since (II-3) depends entirely on Taylor’s presupposition 5, that presupposition must also be false, but that Taylor somehow implicitly tries to present presupposition 5 as “analytically true,” and thus tries to force “linguistic reform” on us.
Saunders’ objections are not particularly effective against the fatalist, and they appear to be ineffective for many of the same reasons that have frustrated so many other Taylor-critics. It is not hard to see that the fatalist can respond to Saunders by claiming that the only reason why we accept (II-6) as true rather than (II-3) is that we regard the conditions named in (II-4) and (II-5) as conditions somehow more “necessary” for being able to swim at t1 than the condition named in (II-1), and that the only reason we do this is because (II-4) and (II-5) must obtain before the swim, and water-turbulence must obtain after, which does nothing really but deny that future events can be necessary for present events, which is not terribly helpful without an argument to back up the claim against Taylor’s apparent demonstration that future events are sometimes necessary for present events. The pole-vault
er’s low ceiling might very well appear more constraining than the would-be swimmer’s non-turbulence-at-t 2, but this is only for reasons of temporal priority, which Taylor claims do not really matter. Given the quite plausible and standardly accepted symmetry of presupposition 4, Taylor can argue that a denial that a future state of affairs can be causally necessary for a present state of affairs would be the same as a denial that a present state of affairs can be causally sufficient for a future state of affairs, which would of course be absurd. Saunders’ only option here would thus be to turn his attack on presupposition 4, which he wisely does not do, or simply to invoke intuitions about the importance of temporal priority in causal necessity-relations, intuitions which the fatalist here clearly does not share, and will claim require some sort of argument ... which Saunders does not appear to have. Saunders may, of course, retreat to his original sociological claims about the way people ordinarily use language, but the fatalist can reply that the fact that Taylor’s argument and conclusion do not accord with ordinary language, that the conclusion might entail a revision in our use of certain terms connected with action and power and ability, may simply mean so much the worse for our ordinary use of language, a language which, used as it ordinarily is by people like us who are as a mass obviously not fatalists, might very well on Taylor’s view be in need of some metaphysically-motivated revision.
It is important to see why Saunders’ seemingly plausible objections do not really succeed in refuting the fatalist. They amount to the claim that Taylor’s argument has implications which go against our intuitions about the world and about language. But see that the fatalist does not share our intuitions. He has metaphysical intuitions of his own about the way the world operates and the way language ought to be used to characterize those operations. He also has an argument for his intuitions: here, Taylor’s. Because intuitions are obviously not refutations, mere claims that the premises or conclusion of the Taylor argument has counterintuitive implications or requires counterintuitive reasoning of some sort cannot by themselves refute the argument. Yet most of the best known approaches to the Taylor problem seem to reduce to just these sorts of claims.