Determinism is the thesis that a given state of the world, in conjunction with the laws of nature, determines all future states of the world. Compatibilism is the thesis that it is metaphysically possible that both (a) determinism is true, and (b) sometimes people act freely. One might be a compatibilist and yet doubt determinism, on the basis of quantum mechanics, or doubt that we ever act freely, on the basis of Freudian psychology. Compatibilism is not a doctrine about what actually is the case, only about what is possible.
One of the main recent opponents of compatibilism has been Peter van Inwagen. Van Inwagen first submits that an action is freely performed only if the agent could have done otherwise. (Although this thesis has not gone unchallenged. Frankfurt has devised thought-experiments which purport to show that someone might act freely even though she could not have done otherwise.) Van Inwagen then asks us to consider some action A, done by an agent J at time T. Van Inwagen will argue that if determinism is true, J could not have refrained from doing A. Since the argument relies on no features peculiar to action A or agent J (other than the fact that J had a beginning in time, and did not exist from eternity), the argument will apply in general to all actions, and, if it is successful, will show that determinism is incompatible with being able to do otherwise.
Let P0 be some proposition that expresses the state of the world at some time before the birth of J, and let L be the conjunction of all the laws of physics. Let P be the proposition that expresses the state of the world at T, the time at which J did A. Van Inwagen's argument also relies on the locution "could have rendered false," a relation between an agent and a proposition. Here is the argument.
(1) If determinism is true, the conjunction of P0 and L entails P.
(2) If J had not done A at T, then P would be false.
(3) If (2), then if J could have refrained from doing A at T, J could have rendered P false.
(4) If J could have rendered P false, and if the conjunction of P0 and L entails P, then J could have rendered the conjunction of P0 and L false.
(5) If J could have rendered the conjunction of P0 and L false, then J could have rendered L false.
(6) J could not have rendered L false.
(7) Therefore, if determinism is true, J could not have refrained from doing A at T.
Premise (1) follows from the definition of determinism, so it is unproblematic. Premise (2) is unproblematic as well, for since P expresses the actual state of the world at T, P entails that J does A at T. Premise (3) is similarly unproblematic, since J's not doing A at T suffices for the falsity of P. Premise (4) is also unproblematic; for since the conjunction of P0 and L entails P, P cannot be false unless the conjunction of P0 and L is false.
Premisses (5) and (6) are the tricky pair. Van Inwagen remarks that he thinks that (5) is an instance of an analytic principle: "If Q is a true proposition concerning only states that obtained before S's birth, and if S can render false the conjunction of Q and R, then S can render false R."
It seems doubtful that this principle is analytic. If S were a time-traveler, she might well be able to render false some proposition concerning only states that obtained prior to her birth. Nevertheless, the principle has some degree of plausibility, and if van Inwagen only succeed in showing that determinism is compatible with free will only if time-travel is actual (or if there are beings existing from eternity, as I remarked above), he will have shown a lot.
Premise (6) is also regarded as a conceptual truth by van Inwagen. "I cannot break the laws of physics," as Scotty told Capt. Kirk.
In "Are We Free to Break the Laws?" David Lewis argues that there is an ambiguity in "can render false", and on one reading, (5) is false, and on the other reading (6) is false. Lewis distinguishes between two senses of "can render false the laws of nature." In one sense, I render false the laws of nature if I do something that is, or causes, a law-breaking event. Moving my hand at faster than the speed of light would be a law-breaking event, and my setting up a particle accelerator in which a proton traveled faster than the speed of light would cause a law-breaking event. In this sense says Lewis, I cannot render false a law of nature. Fortunately, the compatibilist is not committed to premise (5) under this reading. If I can render false the conjunction of P0 and L, it does not follow that I can perform an act that would be, or would cause, an event that falsifies L.
What would follow, Lewis maintains, is that if I can render false the conjunction of P0 and L, I can perform some act such that if I were to perform it, a law would be broken. A law would be broken, but my action would not be, nor would it cause, a law-breaking event. On the contrary, a law-breaking event would be the cause of my action: if J had refrained from A, a "divergence miracle" would have occurred prior to T, causing J to refrain from A. On this reading of "can render false," however, the compatibilist is not committed to the truth of (6). Even on this reading, (6) is not undeniably false, of course, but neither is it undeniably true as it would be on the previous reading of "can render false."
Whereas Lewis's compatibilism is a merely defensive compatibilism, content to find responses to arguments such as van Inwagen's, other compatibilists have gone on the offensive, and attempted to produce analyses of the sense of "can" relevant to free will. These analyses trace their roots back to Moore (at least), and usually take some such form as the following: "S can do X" = "If C, S would have done X", where C is generally an open sentence making reference to S, such as "If S had chosen to do X" or "If S had wanted to do X." Keith Lehrer has issued a general challenge to any such compatibilist analysis, based on work of Chisholm's. Lehrer argues as follows.
Lehrer asks us to consider any attempted analysis of "S can X" of the following form, "S would do X, if C." Lehrer then claims that the following are at least consistent with the analysans:
(a) S cannot X, if not-C.
(b) Not-C.
But (a) and (b) would seem to imply "S cannot X" by modus ponens, and thus the proposed analysans is not correct, for it is consistent with (a) and (b), whereas the analysandum is not.
Bruce Aune has claimed that this argument is faulty, in that Lehrer gives us no reason to think that (a) and (b) are actually consistent with "S would do X, if C." Unless we have some sort of argument to show that (a) and (b) are consistent with the proposed analysans, the compatibilist is free to reject Lehrer's assumption that they are.
I think a stronger case can be made against Lehrer's argument. For there is a natural reading of sentences having the form of (a), but for which modus ponens is not valid (given certain readings of the conclusion). One who proposes an analysis of the "can" of freedom can accept the corresponding conditional of form (a), as long as it has this reading.
As an example, consider this natural way to define what it is for an argument to be valid: "The premisses can't all be true, if the conclusion isn't true." But from this sentence (given the reading appropriate to a definition of validity), and the sentence "The conclusion isn't true", "The premisses can't all be true" (reading this as "The premisses are inconsistent") doesn't follow.
Of course, there are contexts in which it would be true to say "The premisses can't all be true", i.e. contexts in which we are holding constant the falsity of the conclusion. But not the context in which "The premisses cannot all be true" means "The premisses are inconsistent." So the proponent of the analysis of freedom may even accept the conclusion "S cannot X", as long as she is outside the context in which "can" is the "can" of freedom which she has proposed an analysis of.
Lewis, David. "Are We Free to Break the Laws?" Theoria 47 (1981), 113-21. Reprinted in David Lewis, Philosophical Papers, Vol. II. Oxford, 1986.
Van Inwagen, Peter. "The Incompatibility of Free Will and Determinism." Philosophical Studies 27 (1975), 185-99. Reprinted in Watson, op. cit.
Back to Brock's Philosophy Page