The argument is as follows:1. At T1 NC issues order O, or if he does not issue order O, then he issues order O′.
2. His issuing order O at T1 is a sufficient condition for a naval battle occurring at T2 (assuming T2 to be exactly one day after T1).
3. Therefore, a necessary condition for his issuing order O at T1 is the occurrence of a naval battle at T2.
4. His issuing order O′ at T 1 is a sufficient condition for no naval battle occurring at T2.
5. Therefore, a necessary condition for his issuing order O′ at T1 is no naval battle occurring at T2.
6. But at T2 it is true or, if not true, then false, that a naval battle occurs at T2.
7. If it is true at T2that a naval battle occurs at T2, then a necessary condition is lacking for his having issued order O′ at T1.
8. If it is false at T2 that a naval battle occurs at T2, then a necessary condition is lacking for his having issue order O at T1.
9. But in either case, a necessary condition is lacking for his having issued one or the other of the two orders.
10. Therefore, one of the orders was such that he could not issue it, and he was forced to issue the other.
One might wish to deny the conclusion of this argument by asserting the following: the argument does not prove that a necessary condition for one of the two orders was lacking at T1, but only that such a condition was lacking at T2. Therefore, until T2, the naval commander had it within his power to issue either command. In other words, since the law of excluded middle has been denied in regard to certain statements about the future, it is neither true nor false at T1, nor at any time prior to T2, that a necessary condition for one of the two orders is lacking at T2.
This reply, however, appears to raise more problems than it solves. In fact, it leads to the conclusion that no action is ever possible. Consider the following argument:1. In order to issue order O at T1, all conditions necessary for the occurrence of O at T1 must be satisfied at T1.
2. In order to issue order O′ at T1, all conditions necessary for the occurrence of O′ at T1 must be satisfied at T1.
3. If order O is issued at T1, then that is a sufficient condition for a naval battle occurring at T2.
4. Therefore, a necessary condition for the issuance of order O at T1 is the occurrence of a naval battle at T2.
5. If order O′ is issued at T1, then that is a sufficient condition for no naval battle occurring at T2.
6. Therefore, a necessary condition for the issuance of order O′ at T1 is no occurrence of a naval battle at T2.
7. But at T1 it is neither true nor false, according to those who wish to deny the law of excluded middle, that a naval battle occurs at T2.
8. In order to issue order O at T1 all conditions necessary for the issuance of that order must then be satisfied, and one of those conditions is that a naval battle occurs at T2. But this condition is not satisfied at T1. Therefore order O cannot be issued at T1.
9. In order to issue order O′ at T1, all conditions necessary for the issuance of that order must then be satisfied, and one of those conditions is that no naval battle occurs at T2. But this condition is not satisfied at T1. Therefore order O′ cannot be issued at T1.
10. Thus, neither order O nor order O′ can be issued at T1.
It appears then that, if one attempts to avoid the conclusion of Taylor’s argument by denying the law of excluded middle in regard to certain statements about the future, then one is led to deny that any action whatever can occur, a conclusion even stranger than that which Taylor’s argument purported to prove.
It may be possible to deny Taylor’s conclusion by either (i) rejecting the law of excluded middle altogether, as it is supposed to apply to any statement whatever, or (ii) rejecting all his other assumptions, including those which explicate the standard way in which the notions of necessary and sufficient conditions are explained in philosophy. The first alternative is plainly absurd, and no one has so far given any reason for taking the second alternative other than simply saying that, by so doing, one might avoid fatalistic conclusions.
NOTES
1 Richard Taylor, “Fatalism,” The Philosophical Review, 71 (1962): 55-66; “Fatalism and Ability,” Analysis, 23, 2 (1962): 25-37; “A Note on Fatalism,” The Philosophical Review, 72 (1963): 497-499.
2 Bruce Aune, “Fatalism and Professor Taylor,” The Philosophical Review, 71 (1962): 512-519; John Turk Saunders, “Professor Taylor on Fatalism,” Analysis , 23, 1 (1962): 1-2; Peter Makepeace, “Fatalism and Ability, II” Analysis, 23, 2 (1962): 27-29; John Turk Saunders, “Fatalism and Linguistic Reform,” Analysis, 23, 2 (1962): 30-31; Raziel Abelson, “Taylor’s Fatal Fallacy,” The Philosophical Review, 72 (1963): 93-96; Richard Sharvy, “A Logical Error in Taylor’s ‘Fatalism’,” Analysis, 23, 4 (1963): 96; John Turk Saunders, “Fatalism and the Logic of ‘Ability’,” Analysis, 24, 1 (1963): 24.
3 A. J. Ayer, The Concept of a Person, and Other Essays (New York, St. Martin’s Press, 1963), pp. 235-268.
4 “Professor Taylor on Fatalism,” p. 2.
5 Aune, op. cit., p. 514.
6 Makepeace, op. cit., p. 28.
7 “Fatalism and Ability”: 26-70.
8 “Fatalism and Linguistic Reform,” p. 31.
11
COMMENT
RICHARD TAYLOR
THESE EXCELLENT critics make several points, all of them good and perceptive. Sharvy’s remarks seem to me answerable within the framework of my assumptions, but the difficulties raised by Cahn, in the second part of his discussion, are admittedly hard.
Both Sharvy and Cahn note that it is hardly a criticism of my argument for fatalism that it has fatalistic implications—a type of criticism that has now become fairly familiar. Indeed, a fatalist could well be described as someone who believes that those actions which are within his power are coextensive with those he performs, i.e., that he is able to perform those and only those actions which he does in fact perform. Spinoza and a good many other philosophers have felt certain that this is true. To point out in one way or another that it follows from my argument is hardly to refute that argument, as Sharvy and Cahn aptly note. It is only to formulate its conclusion.
Sharvy’s second point is that if, as he believes, I “abolish” modal distinctions and treat “It is within my power to do X” as logically equivalent to “I do X,” then my crucial statement, “Either it is not within my power to do X or it is not within my power to refrain from doing X” becomes logically equivalent to “Either I do not do X or I do.” This, Sharvy rightly notes, is not a significant conclusion and has nothing whatever to do with fatalism.
THE JOURNAL OF PHILOSOPHY, VOL. 61, NO. 10, 1964.
That is true. But I never suggested that “It is within my power to do X” is logically equivalent to “I do X.” I have not “abolished modal distinctions,” nor does my argument imply their abolishment. My argument only entails that statements of the form just given always turn out to have the same truth values; in other words, that those actions which an agent is able to perform are in fact identical with those and only those which he performs. This implies no logical equivalence of meanings.
To illustrate this, all men are fatalists with respect to some of the things they and others do. Thus, no one believes it is within the power of a trembling man on the gallows to control his trembling, or to decline to die shortly thereafter; or that a pregnant woman is able to bear a son rather than a daughter unless, in fact, it is a son that she is going to bear; or that it is within one’s power to decline to jump when a shotgun is without forewarning exploded behind him. Yet such assertions of ability are clearly meaningful; they are just false. There is not the slightest contradiction in saying of a trembling man that he is able to refrain from trembling, that it is within the power of a man waiting on the gallows to stand there serenely. It is, however, very likely to be false.
A fatalist says the same about all actions. It is, admittedly, an implication of my argument that it is in ever
y case false to say of an agent that he is able to do what, as it later transpires, he does not do. But it is no contradiction, as it would be to say that he does do what, as it later transpires, he does not do. “It is within my power to do X and also within my power not to do X” is not, then, equivalent to “I do X and I do not do X.” The latter is a contradiction, but the former is not, and they are accordingly not equivalent. But the first of these statements is, according to my argument, false, on any interpretation of X.
Now I have little inclination to accept this implication of my argument—that is, to be a fatalist. I am quite certain, however, that it follows validly from my argument, and the premises from which it follows seem to me, as to most other philosophers, to be quite plainly true.
Now I had suggested that one might avoid fatalism by rejecting the law of excluded middle with respect to certain statements about the future; namely, those whose eventual truth or falsity is contingent upon the as yet future deliberate and free actions of agents. Cahn has invented two frightfully acute arguments in refutation of this and these arguments are, I must confess, painfully difficult to answer without appearing to violate reason itself. Cahn’s arguments thus press fatalism with a vengeance, in a manner which I have not seen heretofore.
All the premises of his first argument are, I think it must be granted, indubitably true. The sixth step, which Cahn has shrewdly avoided expressing in the future tense, is the crucial one, and is clearly undeniable. The only possible reply to this argument is, I believe, suggested by Cahn himself; namely, that the conclusion is perhaps improperly expressed. It should say, I believe, that one of the orders is such that the commander can not, at T2, any longer issue it, that it is no longer within his power to issue the order that he did not issue. This is of course not controversial; it expresses only a fatalism or irrevocableness about the past which everyone believes anyway and which I in no way brought into question.
I am inclined to think, however, that Cahn’s second argument cannot be answered at all. It purports to show that if, as I had suggested, one denies the law of excluded middle with respect to certain statements about the future, then, far from avoiding fatalism, as I had thought, one lands himself in the absurd conclusion that no agent can ever do anything at all! The acumen displayed in this argument seems to me thoroughly stunning. The only way I can see to avoid the disaster to which it leads would be to reject its first two premises, on the ground that to say that some condition, future to T1, is “satisfied” at T1 amounts to saying either (a) that the condition in question is not after all future to T1, as alleged, or (b) that it is true at T1 that it will be satisfied at T2, which begs the question. If I were to press that reply, however, then I am not sure what sense I could still give to one of my own presuppositions; namely, that no agent can perform an action in the absence of some condition, past, present, or future, necessary for its accomplishment.
I feel obliged to concede, therefore, that it may be quite unhelpful to try modifying the traditional interpretation of the law of excluded middle. Perhaps some of my other presuppositions are doubtful, but I can imagine no reason for rejecting any of them other than the one so frequently brought forth; namely, that they seem to have fatalistic implications.
12
FATALISM AND ORDINARY LANGUAGE
JOHN TURK SAUNDERS
RICHARD TAYLOR’S ideas about fatalism have caused quite a stir.* Numerous objections to his thesis have not led him to abandon it. Indeed, the writing of Steven Cahn, who defends Taylor and who has attempted to state the case for fatalism even more strongly than Taylor does, has convinced Taylor that even the modification of the law of excluded middle may not save us from fatalism.1 Since Taylor comments upon Cahn’s article and in no way repudiates Cahn’s replies to Taylor’s critics, I shall feel free, in this paper, to attribute the contents of these replies to Taylor himself. I think that Taylor (and Cahn) are badly confused on this matter, and I shall try to demonstrate this in what follows.
1. Taylor’s Position. Taylor assumes: (P) No agent has the power to perform an act a necessary condition of which is lacking. The term ‘necessary condition’ is so used that if a is a necessary condition of b then a and b are logically unrelated but are empirically related.2 The term ‘power’ is so used that “Joe has the ability to do a” does not entail “Joe has the power to do a.” Suppose, for example, that a necessary condition of my swimming at t3 is the (nearby) water’s being turbulent at t4. Suppose, too, that the water is not turbulent at t4. It follows, given the usual interpretation of the law of excluded middle, that it is not in my power to swim at t3.
THE JOURNAL OF PHILOSOPHY, VOL. 62, NO.8, 1965.
We are tempted, at this point, to reject the argument on the ground that P is false. We are inclined to hold that, although P would be true if it were rephrased to apply only to acts of which a past necessary condition is lacking, it surely cannot be true as it stands: “Surely I could have made the water churn, had I wished to; I didn’t swim, not because I lacked the power to do so, but because I decided not to.” But Taylor has a ready reply. We will all admit (he thinks) that no agent has the power to perform an act a past necessary condition of which is lacking; for to deny this would be to allow that someone has the power to “alter” or “adjust” the past, “to make happen what in fact did not happen.” For the same reason, then, we ought all to admit that no agent has the power to perform an act a future necessary condition of which is lacking; for to deny this would be to allow that someone has the power to alter or adjust the future, to make happen what in fact will not happen. This we must admit so long, at least, as we refuse to tamper with the law of excluded middle.
2. Is This Fatalism? Most of this paper will be devoted to an elucidation of the mistakes that I believe to be involved in Taylor’s position as thus epitomized. First, however, I wish here to indicate that, unbeknownst to him, Taylor has presented no arguments from which fatalism follows as a conclusion. As he himself sees,3 fatalism is, at least, the doctrine (F) that it is never in anyone’s power to do anything other than what he in fact does. But F follows neither from the conclusion of the first argument above [viz., “It is not in my (e.g., Saunders’) power to swim at t3”] nor from that of the second (viz., “No agent has the power to perform an act a future necessary condition of which is lacking”). The latter conclusion, when it is taken together with something like (S) “Every act of every agent is such that all alternative acts have lacking at least one future necessary condition,” will indeed entail F. And it is something at least approaching the strength of S that Taylor appears to accept. Why, otherwise, should he talk as though his arguments regarding the absence of future necessary conditions of acts alternative to those we in fact perform require the acceptance of fatalism? But Taylor has not even attempted to support anything like S, and he is thus mistaken in claiming: “My argument ... entails ... that those actions which an agent is able [i.e., has the power] to perform are in fact identical with those and only those which he performs.”4
Further, it is not at all clear how one could even go about supporting S. Even setting aside the problem of supporting the claim that all acts ever performed by anyone are acts such that all alternative acts have future necessary conditions which are lacking, how are we to support the claim, made about just one particular action of a particular person, that all acts alternative to it have future necessary conditions which are lacking? We must acquire reason to believe that each such alternative act has some future necessary condition which is lacking. But it is not clear that there is a limited number either of alternative acts or of their future necessary conditions. Nor, therefore, is it clear how we could even go about satisfying ourselves on this score. And the task of disconfirming S would appear to be even more hopeless: to do so we must find some act alternative to an actual act such that either it has no future necessary conditions or all of its future necessary conditions are fulfilled. But what could count as evidence for an alternative act’s having n
o future necessary conditions (as opposed to our failing to find them), and what could count as evidence that a specified group of fulfilled future necessary conditions of a certain alternative act exhausts the totality of future necessary conditions of that act?5 I would conclude from the foregoing that, since it is neither confirmable nor disconfirmable, the sort of fatalism in which Taylor is interested is not so much as a meaningful doctrine.6
If Taylor’s arguments, even given their soundness, do not succeed in establishing fatalism and if Taylor’s fatalism is not even meaningful, then what importance may be attached to his arguments? Well, it is surely interesting to inquire whether Taylor is correct in accepting P and (a fortiori) the following:
(PF) No agent has the power to perform an act a future necessary condition of which is lacking.
(PP) No agent has the power to perform an act a past necessary condition of which is lacking.
It is also interesting to inquire whether the mere calmness of the (nearby) water at t4 renders me powerless to swim at t3. And this is the sort of proposition that Taylor’s arguments would support if they were sound. But I shall now go on to argue that his arguments are not sound, that they do not support this sort of proposition, and that P, PF, and PP are all unacceptable.
3. The Ordinary-language Statement PF Is False. Returning to the aquatic example of Section 1, it follows from PF, together with(1) A necessary condition of my swimming at t3 is the water’s being turbulent at t4