Chapter 6. Fate (F) Fatalism is the belief that whatever happens is unavoidable. (55) The first, and most important thing, to note about Taylor s characterization of fatalism is that it is in modal terms, since unavoidable means cannot be avoided, and can is a modal term. One should then ask: exactly what modality is being invoked and how is the introduction of this modality justified? At first glance, Taylor s characterization of fatalism looks a lot like his characterization of determinism in chapter 5: (D) Things [now and in the future] could not be other than they are. But on some investigation and reflection, we saw that as a statement of determinism (D) had to be replaced or made precise by: (D ) If one has a system in state S* at T*, then there is only one state S at t that is nomically possible.
Later in the current chapter Taylor says that fatalism is not dependent upon any presupposition about causal determinism, the coercion of his actions by causes, or anything of this sort. (60) Fatalists generally agree on this point. Even though the thesis of fatalism may superficially look like the thesis of determinism, when you examine them carefully, you see that they are couched in terms of different modalities. Determinism rests on (what we have come to call) nomic necessity. We shall spend a good bit of time trying to ascertain just what other sort of necessity is invoked by fatalism. But the one thing that we can count on at the outset is this: whatever kind of necessity is being invoked by fatalists, it is not nomic necessity. Fatalism and determinism, then, are distinct views, and quite distinct sorts of arguments are offered for and against them when they are up for consideration. In the first three editions of Metaphysics Taylor gave in Chapter 6 a very careful and somewhat intricate argument for fatalism, an argument derived from his (in)famous paper in The Philosophical Review (1962). Below I give a précis and criticism of that argument. (Those interested in literature will want to note that Taylor s argument was the subject of David Foster Wallace s 2
senior thesis, which has recently been published in a book of related essays on fatalism.) The argument becomes considerably muted in the fourth edition, but I will view the material in this chapter as a more literary (or more crafty) attempt at persuasion rooted in the older argument, and I will try to tease out some of the structure of that argument. One clear idea in the story of Osmo is that there are statements about the future, as well as the past, that have truth-value. That is, the ones that accurately describe the future are true and those that are inaccurate are false. This idea, you may recall, was involved in Taylor s earlier claim that the world is perfectly determinate. The linguisticized version of this claim is: Bivalence: Every proposition is either true or, if not true, false. [Note that Taylor refers to bivalence as the law of excluded middle. Logicians usually use this expression for a theorem of classical propositional logic, (Pv~P), which is closely related to bivalence.] Bivalence, then, includes propositions about the states or behaviour of future things (or things in the future, if you prefer), even when such 3
propositions do not express logical truths. Such propositions are usually called future contingent propositions or just future contingents. A typical such proposition might concern my choice for tomorrow s lunch. If I choose (A) to have an apple for lunch tomorrow, then the proposition that I so choose is true and the proposition (P) that I have a piece of pizza is false. If I choose the pizza, then (P) is true and (A) false. If I skip lunch, both are false. Osmo became convinced of fatalism because he came across a book containing a set of true contingent statements about his future. But, as Taylor takes great pains to point out on page 62, while his reading the book was crucial to his coming to believe fatalism, the doctrine is not an epistemic or epistemological doctrine and is completely independent of what one does or does not know. The validity of fatalism, says Taylor, is assured by (1) alone (62), where (1) is the the claim that there existed a set of true statements about his life, both past and future. (62) This is just bivalence. Working quietly along with bivalence in most fatalist arguments is the correspondence view of truth: A statement or proposition is true of it 4
corresponds to the way the world is. This view is usually taken to be so obviously correct that it does not need to be mentioned, though there are other views of truth that have been considered by philosophers from time to time. We will not look any further into alternative views of truth in this course. What I propose to do is to show, following Sobel, that even if one assumes truth as correspondence and full bivalence, no significant or worrisome form of fatalism follows. There is no need, for instance, to deny truth-values to future contingents, as Aristotle did. Having indicated the assumptions or premises of Taylor s argument for fatalism, it is time to say what the argument is. It seems to come down to the following claim: No power in heaven or earth can render false a statement that is true. It has never been done, and never will be. (61) While this statement seems reasonable, we will note that it is ambiguous. First, it makes a modal claim. That is, it contains the word can, which we know to be polysemous. But there is also a second ambiguity that should be noted and disposed of. 5
Is the power spoken of required to render false a statement that is true while leaving it (as initially assumed) true? That is, if it s true that I have the apple for lunch tomorrow, is the power supposed to render it also the case that I do not have the apple. Then the power is required to render ~A true while leaving A true. That is, it must bring it about that (A & ~A) is true. But no power can do this, since (A & ~A) is a straightforward logical contradiction. The falsity of (A & ~A) follows from the truth tables for ~ and &, not from any lack of power in the world or any heavy hand of fate. If fatalism is the doctrine that no one can bring about a logical contradiction, then that doctrine is true, but innocuous. No one--or very few, at any rate-- will have any quarrel with it. But that is not the view that Taylor called fatalism. Recall that (F) Fatalism is the belief that whatever happens is unavoidable. (55) That view is not expressed as ~ (A & ~A). Rather, one might express it as: (A ~ ~A), where A can be any sentence at all. This sentence can be read as saying that if A is true (or does happen), then ~A (A s not happening) is not possible. That seems a reasonable way to understand (F), what fatalists claim. 6
We know that we can re-write this sentence as (A A). But how are we to understand the necessity or possibility operators in these sentences? It seems to me that fatalists must tell us how to understand them and also why, so understood, (A A) is true, if they wish to convince us of their view (or even to render their view comprehensible). This Taylor does not do, leaving us to see what we can do on his behalf. First, we repeat that nomic necessity (or possibility) is not as issue here. Fatalists specifically say so, and so say correctly, since introducing nomic necessity would change the subject from fatalism to determinism. Second, can represent logical necessity (and logical possibility)? If that were so, then fatalism would just be false. If the box is understood to represent logical necessity, then A says that in all logically possible worlds--all worlds that are consistently describable--i have an apple for lunch tomorrow. Even if I do in fact have the apple for lunch tomorrow, it certainly seems as if I can consistently describe a world (an alternative history of this world) in which I go to the Aquatic Centre and skip lunch altogether. In that case, given the truth table for the material conditional, the sentence (A A ) is false--it has a 7
true antecedent but a false consequent. Fatalism simply fails, if the necessity involved is logical necessity. Of course, this does not prove fatalism false. But what is required to prove it true is a way of understanding necessity in which (A A ) is true and also in which the consequent expresses a kind of necessity that conflicts with what ordinary people (that is, non-fatalists) think is or is not within their power. One final note. (A A ) is a theorem of classical logic. It says (for a future contingent) that if A will occur, then A will occur. This is the famous que sera, sera of the Doris Day song. What this sentence expresses is a logical triviality, on a par with what has occurred, has occurred. It is therefore undoubtedly true, for any sentence you choose to substitute for A, but this truth does not say that what will occur must occur. It only repeats that it will. If A is true in any possible world, then so is (A A ). It follows then that (A A ) is true in any modal logic and also that it will be a theorem of any modal logic. But these facts do not establish fatalism. If I do have the apple for lunch tomorrow, then indeed I do have the apple for lunch 8
tomorrow, that is, (A A ). Nothing in this observation renders my having the apple tomorrow inevitable. So, while it is quite easy to justify the assertion (A A ), that assertion does not amount to fatalism. What is required is (A A ), which is a more difficult to justify. Note that what is at issue here, the difference between (A A ) and (A A ), is the scope of a modal operator. 9
Taylor s Six Presuppositions 1. Bivalence: Every proposition is either true or, if not true, false. The Law of Excluded Middle, (P V ~P), is subtly different. The two usually, but not invariably, go together. 2. If any state of affairs A is a sufficient condition for some other state of affairs B, then A cannot occur without B occurring too. A, B are not logically connected. I.e., not A B or B A. Or, perhaps, not logically connected means: ~ (A B), where indicates truth in all possible worlds (logical necessity). So is Taylor s second condition to be represented in the standard manner : A B? But then how are we to understand A cannot occur without B occurring too? This characterization contains a modal term! So, do we write ~ (A & ~B)? That is, ~(A & ~B). That is, (A B). Taylor has denied this above. We can write this, of course, but what sort of necessity is represented by the if it is not logical necessity? Taylor specifically denies that nomic necessity (the sort of necessity one finds in laws of nature) plays a role in his argument for fatalism. I want to show that certain presuppositions made almost universally 10
in contemporary philosophy yield a proof that fatalism is true, without any recourse to theology or physics. It would seem that we should conclude that the best way to understand A cannot occur without B occurring too is the standard way, as a material conditional. The modal talk is just a loose way of communicating a precise idea, in the absence of some indication of what the modality is supposed to be. 3. If any state of affairs A is a necessary condition for some other state of affairs B, then B cannot occur without A occurring too. We can raise questions and reach tentative conclusions about presupposition (3) that parallel those we raised about presupposition (2). 4. If A is a sufficient condition for B, then B is a necessary condition for A (and conversely). This is clearly so if necessary and sufficient conditions are represented by material conditionals. 5. No agent can perform any given act if there is lacking, at the same or any other time, some condition necessary for the occurrence of that act. Let s be sure to distinguish this from 5*: No agent does perform any given act if there is lacking, at the same or any other time, some condition necessary for the occurrence of that act. 11
5* would indeed follow from the standard reading of necessary condition. It is doubtful that 5, as stated above, does. 6. The passage of time, by itself, is not efficacious. Contra Aristotle, who thought that the passage of time itself was destructive. Taylor s Argument(s) The first argument, about a past contingent particular occurrence: P: A naval battle occurred yesterday. P': No naval battle occurred yesterday. [Note that P' is really just ~P. etc.] S: I am about to read a headline describing the battle. S': I am about to read a headline of a different sort. S P, S' P' are assumed. The argument, then, is 1. P It is not in my power to do S' (by presupposition 5). 2. P' It is not in my power to do S. (similar) 3. P V P' (Bivalence) 4. Either it is not in my power to S or it is not in my power to do S'. 12
So the following assertion is proved false: (A) It is within my power to do S and it is within my power to do S'. No one quarrels with this argument, it is said, since we are all fatalists in regard to the past. However, an exactly parallel argument is exhibited with regard to an arbitrary future contingent particular occurrence. Q: A naval battle will occur tomorrow. Q': A naval battle will not occur tomorrow. O: I (the commander) order the battle to take place. O': I order that no battle take place tomorrow. We assume O Q, O' Q' (i.e., ~O ~Q). The argument, then, is 1. Q It is not in my power to do O' (by presupposition 5). 2. Q' It is not in my power to do O. (similar) 3. Q V Q' (Bivalence) 4. Either it is not in my power to O or it is not in my power to do O'. So the following assertion is proved false: (B) It is within my power to do O and it is within my power to do O'. 13
Why is it that hardly anyone is convinced of the falsity of (B) by this argument, although virtually everyone is convinced of the falsity of (A) by an exactly parallel argument? Possible Responses to Taylor s Argument 1. Time makes a difference. But how? 2. One seems to have causation going the wrong way. But there is no mention of causation in the argument. 3. Deny bivalence. Aristotle s way out, most likely. On this view time will by itself have the power to render true or false certain propositions which were hitherto neither, and this is an efficacy of sorts. (p. 230) The passage of time itself must make certain acts that were within our power, when future, not within our power, when past. Presuppositions (1) and (6) are inseparably linked. [4. Question the fixity of the past. Perhaps I can affect the past in limited ways. 14
For instance, perhaps I can now make certain sentences uttered yesterday (but about the future) true. These are so-called soft facts. (Ockham?) Of course, I can t make any event to have happened yesterday. That s a hard fact. 5. Question the cogency of Taylor s argument. With regard to premises 1 in each argument, the premise as stated are supported by presupposition 5, but that presupposition does not follow from or receive justification from the standard reading of necessary condition, which contains no modal terms. Without other justification, presupposition 5 is dubious and maybe downright false. A necessary condition for my jumping six inches high now is, say, my wanting to so jump. That desire is lacking, and I do not jump. We ve agreed, though, that I have the power to do so. Presupposition 5* is justified by the standard reading of necessary condition but it does not justify premises 1 as stated. Of course, similar remarks can be made for premises 2.] 15