I just finished Trenton Merricks' recent Phil. Review (2009, 118:1) article "Truth and Freedom". Like most of his stuff, his arguments were crisp and clear. I'll say at the outset that I'm far from being an expert in the freewill/foreknowledge/fatalism literature. (I make that explicit, because I know that some readers of this blog are!)
It was fun to see Merricks apply some of his work on the nature of truth to the fatalism problem. He starts with the mundane, general point that a proposition is true because of what the world is like; i.e., truth depends on the world, not the other way around. Furthermore, this dependence is not a causal one (more on this below).
Here is one way he applies this mundane point to what he calls "The Main Argument", what he takes to be the most serious fatalistic argument. (He briefly discusses other fatalistic arguments, but considers them to be obviously faulty, despite the big names who have endorsed them (pp. 32-33)):
"1) Jones has no choice about: that Jones sits at t was true a thousand years ago.
2) Necessarily, if that Jones sits at t was true a thousand years ago, then Jones sits at time t
3) Jones has no choice about: Jones's sitting at time t." (p. 33).
After arguing that the Main Argument is question-begging, he then argues that (1) is false.
Here is where things got interesting to me. One of the objections to the falsity of (1) is as follows (p. 40). The past is necessary in a special temporal sense, and so Jones can't have a choice about what was true in the past. Merricks discusses three ways to understand the necessity of the past, and considers whether, on these understandings of necessity, this objection holds any water. Here is the third understanding of necessity:
"Suppose, finally, that the past's being necessary means only that our present and future actions cannot cause events in the past... Then the past's being necessary amounts to there being no "backward" causation."
"But... the absence of backward causation fails to entail that no one now has a choice about what the past was like. This is because having a choice about the past truth of a proposition does not require backward causation.
It does not require backward causation because the truistic way in which the truth of a proposition depends on the world is not causal. This is why, for example, truth's dependence on the world does not involve the laws of nature or the transfer of energy...
Or consider the following: abstract objects cause absolutely nothing. This familiar claim about abstract objects might be false. But it is not shown to be false merely by the truism that that abstract objects exist is true (if it is true) because abstract objects exist. That is, that truth's depending on abstract objects does not imply that abstract objects cause something; so it does not imply that abstract objects cause that truth to be true... Therefore the truistic way in which truth depends on the world is not causal. So even though that Jones sits at t was true a thousand years ago because Jones sits (or will sit or did sit) at t, the 'because' here is not causal. And so it is not "backward" causal." (p. 41)
I found this passage to be illuminating, because it pointed to me how truth's dependence on the world is not a causal dependence, and how much of what motivates my intuition for accepting (1) is my intuition that one cannot cause changes in the past.
So understanding the necessity of the past as the point that backward causation is impossible does not underwrite (1). Let me know if I've been too quickly convinced by Merricks on this point. Also, I found most of Merricks' arguments to be compelling, so if others who have read Merricks' article think that there are good objections to his other arguments, I'm interested in hearing those too.