The following is another anti-Molinist argument I’d like to look at today. It might get a little technical, especially since after the argument is made, I list it in quasi-symbolic form. I also stole the title from William Lane Craig!
If Molinism is true, God knows that some proposition P (“S will take A”) is true in W. If God knows that P is true in W, P is true in W. If P is true in W, then proposition Q (“P is true in W”) is true in all possible worlds. If Q is true in all possible worlds, then Q is necessarily true. If Q is necessarily true, it is not possible that Q is false. If it is not possible that Q is false, it is not possible that S will not take A in C. Therefore, If Molinism is true, it is not possible that S will not take A in C.
1. If M, then K.
2. If K, then P in W.
3. If P in W, then Q in all possible worlds.
4. If Q in all possible worlds, then N.
5. If N, then ~ possibly ~Q.
6. If ~ possibly ~Q, then ~ possibly ~P.
7. If M, then ~possibly ~P.
First, I’d like to point out that Molinism and God’s knowledge of propositions are incidental to this argument. That is, while it applies, you could simply start from premise 3. If you do, what you get is the conclusion that “If S will take A, then it is impossible that S will not take A.” So this argument, if sound, and by the rules of logic, means that if anything is true at any world whatsoever, then it is necessarily true. Perhaps this is the objector’s point; perhaps the objector is committed to the proposition that everything whatsoever is determined or fated in this way. Or perhaps the objector did not notice this point, and instead is committed to the point that nothing whatsoever is true. Or else, finally, the objector just didn’t realize this at all, and the argument is unsound. I don’t think it’s good enough for us to leave it at that; I want to examine why it is unsound. Premise (1) is pretty solid. That states that if Molinism is true, then God knows some particular proposition, symbolized by “K” (which particular proposition we are calling “P,” and whose content is that “S will do A”) is true in some particular world W. (2) stays, since it says that if God knows that P is true in W, then P is true in W (we’ll call this proposition [that “P is true in W”] “Q”).
So what about (3)? Is that true? I think so. It just says that if P is true in W (Q), then Q is necessarily true. Why? Let's say the proposition represented by P is specifically that “Randy writes on his blog on September 23, 2014,” and that P (so defined) is true in the actual world, which we will call W-147 (because, why not?).
We can thereby represent a new proposition, called Q. Q states:
Q. P is true in W-147,
Which is just equivalent to Q-translated, which is:
Q-translated: “Randy writes on his blog on September 23, 2014” is true in W-147.
Now suppose we move to world W-148, which is a very similar world to ours, except kangaroos don't exist (because, why not?). Now let's examine all of the truths there are in W-148 (because we have time!). What about Q? Is Q true here? Remember, Q's content is not about W-148, and so nothing in W-148 can have any effect on Q's truth-value. What is sufficient for the truth of Q can be found only in W-147. But since it's true in W-147, and this is an objective truth, it will be true in W-148 that:
In W-147, “Randy writes on his blog on September 23, 2014” is true; which is only really to say that:
Q-translated: “Randy writes on his blog on September 23, 2014” is true in W-147; which is really just saying:
Q. P is true in W-147,
and the process can be repeated in any world you wish, which gains us a necessary truth. But this is really no more controversial, in my opinion, than the fact that whatever is possible is necessarily possible. That is, if something is logically possible, then its impossibility is impossible; if something is logically possible, then at no world is it impossible, for impossibility stretches across worlds (that's just what it means to be impossible--not possible at any worlds; but if it is possible at a world (in the logical sense), then it's possible at every world). Anyway, just as nearly no one is bothered by every possible truth's entailing some necessary truth (namely, that it's possible), so no one should be bothered by something like Q, in my opinion.
(4) and (5) are analyses of necessity, and so I think they should stay.
But why think (6) is true? Q just is the proposition “P is true in W.” What (6) does, however, is distribute the necessity of the entire claim “P is true in W” to a particular part of the claim (namely, P). But it is the well-known and oft-committed modal fallacy to infer the necessity of a particular part of a claim from the necessity of the entire claim itself. Here’s a quick and easy example of Right Necessity (RN) and Wrong Necessity (WN):
RN. Necessarily, either everyone is six feet tall or not everyone is six feet tall.
WN. (Since not everyone is six feet tall) Necessarily, not everyone is six feet tall.
But perhaps their usage of this fallacy is even worse than what has just transpired. For consider the exact nature of the fallacy:
WN1. If necessarily “P is true in W,” then necessarily P.
But this inference seems crazy! There are plenty of counterexamples to this inference. Let’s take what most assume is a contingent truth: that the Solar System has nine planets. Since in our world W, the Solar System does have nine planets (get off my back about it!), it will be true in every world that “The Solar system has nine planets in W.” We should hardly infer, on this basis, then, that it’s a logically necessary truth that what we call the Solar System should have nine planets! If you need an even less controversial example, take the number of dust particles on the moon. It’s true that the number of dust particles the moon will ever have in its entire history of this world entails “The number of dust particles the moon will ever have in its entire history of W is true in W,” where “W” is the actual world and some number is represented by the long introductory clause. That entire claim is a necessary truth, but who wants to thereby infer that the exact number of dust particles on the moon is logically necessary? Perhaps someone, but not most of us.
So there you have it. (6) commits the modal fallacy, and the argument for determinism from anything’s being true at all (which just so happens to apply to Molinism, since it claims things are true at worlds) fails.
 As a non-trivial point, it should be noted that this “will” language is entirely wrongheaded. We should try to use tenseless language to describe truths at possible worlds, since the vast majority of possible worlds are never actualized (in fact, only one of them is), and thus there is no future “will” to speak of in these worlds.
 I know, I know. Pluto is still a planet in my heart!