According to a version of essentiality of origins for events, if an event E is explanatorily prior to an event F, then F could not have occurred without E. Of course, an event qualitatively just like F might have occurred without E, but F itself could not have.
Suppose Molinism is true. For a reductio, suppose God brought it about that George would be shipwrecked, because God believed that
(1) Were George shipwrecked, he would freely behave heroically.
Let F be the event of George's shipwreck. I shall assume, as is plausible, that it is an essential property of F that F is a shipwreck of George's. Let E be the event of God's believing (1) to be true. Then, E is explanatorily prior to F. By essentiality of origins for events, the occurrence of F entails the occurrence of E. But the occurrence of E entails the truth of (1) (by God's essential infallibility).
Therefore, that George is in F entails (1). Likewise, that George is in F entails the antecedent of (1), since it is an essential property of F that F is a shipwreck of George's. Therefore, that George is in F entails that George freely behaves heroically. (If p entails a subjunctive conditional and its antecedent, it entails the consequent, because modus ponens holds in all worlds.) But this means that if George is in F, he cannot but behave heroically, and for libertarian reasons, it follows he does not freely behave heroically. Thus he both does and does not behave freely in F. Therefore, we must reject the possibility of the assumption that God brought about George's shipwreck because God believed (1).
How is E explanatorily prior to F in the case you describe?
F= George is shipwrecked
E= God knows that (1) (If G is sw, then G is heroic).
If E is explanatorily prior to F, then George could not have been shipwrecked unless God knew (1), or George's being shipwrecked entails that God knows (1). But that's false. There are worlds in which George is shipwrecked and he does not behave heroically, and so God does not know (1).
If x brings it about that p because x believes q, then x's believing q is explanatorily prior to p. However, it does not follow that p can't be true unless x believed or knew p. There will, in general, be possible worlds where x still brings about that p for other reasons, and in those worlds, those reasons (or associated beliefs) will be explanatorily prior to p.
But George did not bring it about that he is shipwrecked (or, it did not occur that George is shipwrecked) because God knew (1). God's knowing (1) is consistent with George not being shipwrecked at all, right? So, it's hard to see the priority you're talking about. But maybe I'm missing something.
And going the other way around, God's knowing that (1) is perfectly consistent with George not being shipwrecked. So I don't see the priority in either direction.
Are you assuming that if p is explanatorily prior to q, then p entails q, or q entails p? Neither entailment holds. That a match was struck is explanatorily prior to the match's catching on fire. But there is no entailment in either direction between the explanans and explanandum.
Maybe I'm getting lost. You said this in the post,
According to a version of essentiality of origins for events, if an event E is explanatorily prior to an event F, then F could not have occurred without E. (my emphasis).
What is the force of the modal talk here? You're saying here that it could not happen. I've been saying, on the contrary, that it could happen. I took myself to be disagreeing with this claim. But apparently you don't quite mean 'could not', is that right?
I think Mike is correct here. Your "could not" can be taken to imply fatalism, which, of course, is not entailed by Molinism. There is a possible world in which George does not freely act heroically when shipwrecked. If he were not to, however, then E would not be explanatorily prior to F and (1) would not be part of God's middle knowledge.
Furthermore, this argument assumes that libertarian free-will requires the ability to do otherwise. But there are many who would disagree with this, e.g., Bill Craig. As he states in his article "Robert Adams's New Anti-Molinist Argument", "my being able to refrain from doing A in C is not a necessary condition of my freely doing A in C. [Thomas] Flint's essay on papal infallibility . . . provides a good illustration: though God would not permit the Pope to promulgate false doctrine, nevertheless he freely promulgates correct doctrine."
This seems to me to be correct.
Mike:
F could not have occurred without E. But it is possible for George to have been shipwrecked without E. F is a particular shipwreck event. This particular shipwreck event could not have occurred without E. Another one, perhaps qualitatively just like it, could have.
Anonymous:
There is a possible world where George does not behave heroically in a shipwreck. But there is no possible world where George does not behave heroically in this shipwreck. (Maybe I should have made it explicit that I take shipwrecks to be events.)
This particular shipwreck event could not have occurred without E. Another one, perhaps qualitatively just like it, could have.
That's a little puzzling, Alex. Are you saying that this particular event does not happen in other worlds? And are you adding that, since the event is world-bound, it is not possible that it should have any properties other than those it actually instantiates? Are you saying, for instance, that this particular event could not have occurred a moment earlier, or could not have been different in any way?
Remember that I start the post by assuming essentiality of origins for events. If event A is caused by agent x on the basis of belief B, then that A is caused by x on the basis of B is an aspect of the origins of A. Therefore, it is an essential property of A that it be caused by x on the basis of B.
Could an event exist in more than one world? Sure! After the event has been caused, all kinds of different things can happen. It may even be that the same event can take different lengths of time in different worlds, but it will always start from the same origin.
F could not have occurred without E. But it is possible for George to have been shipwrecked without E. F is a particular shipwreck event. This particular shipwreck event could not have occurred without E.
Ok, then, why couldn't this particular shipwreck occur without God knowing that, were G to shipwreck, G would behave heroically? The counterfactual in (1) is true, but (we agree, I'm sure) not necessarily true. So suppose G shipwrecks. Could God have failed to know (1)? It seems surely so. What would have been known, had G freely acted in some, say, less-than-heroic way, is that (1') were G to shipwreck (same orgin, same event), then G would not have been so heroic. God would have known (1'). In short, what would have been known depends on what G actually does. It's exactly the sort of situation in which we'd expect backtracking.
Let's suppose what I've described is the actual world w0: there, God believes (1), and causes George's shipwreck. The event of George's shipwreck (actually, I think it would be clearer if I said: the event of George's being in a shipwreck) is F.
Suppose that in w1, God doesn't believe (1), but George is still in a shipwreck. Then the origin of George's being in a shipwreck in w1 is different from the origin of George's being in a shipwreck in w0. For in w0, this origin includes God's believing (1), while in w1, this origin doesn't include it. But if the origins are different, then the events must be numerically distinct, by essentiality of origins. Thus, if we let F1 be the event in w1 which is George's being in a shipwreck, it follows that F1 is numerically distinct from F.
I suspect that you find the esentiality of origins for events implausible.
Suppose that in w1, God doesn't believe (1), but George is still in a shipwreck. Then the origin of George's being in a shipwreck in w1 is different from the origin of George's being in a shipwreck in w0. For in w0, this origin includes God's believing (1), while in w1, this origin doesn't include it.
Alex,
It's just difficult to see how God believing (1) is a part of the origin of G's shipwreck. It's not difficult to see how God causing G to shipwreck is part of the origin. God's believing (1) alone bears no interesting relation to G's shipwreck that I can see. It just happens prior to it, but then so does Smith burping in New Guinea. You wouldn't want to say that Smith's burping is part of the origin.
Well, certainly, if I believe p, and therefore produce A, my believing p is part of the origins of A. I believe p, and this causes me to produce A, and so A comes about. How is my believing p then not a part of the origins of A? In my case (though probably not God's), the belief is even a partial cause of my action.
Now, you might understand "origin" in the sense of "immediate origin." But that is not the sense in which I understand it. Explanatory priority is transitive, and the origins of a thing include all the things that are explanatorily prior to it.
Note, too, that essentiality of origins for immediate origins of events implies a more unrestricted essentiality of origins, anyway. If event A causes event B, and event B causes event C, and the immediate cause of an event is essential to the identity of the event, then A is essential to C. Why? Because C couldn't exist without B, by the essentiality of C's immediate origins. And B couldn't exist without A, by the essentiality of B's immediate origins. And even if "is caused by" is not transitive, "cannot exist without" is transitive (that's an easy theorem in modal logic).
But we are talking about the case of God, where the belief in (1) is not causally related to G's being shipwrecked. (And of course, if I belief that p and therefore produce q, q needn't be a causal conseqeunce of my belief (not even partially). I might have produced q even if I had believed ~p. But this is all beside the point.) We have in mind the case of God's belief and it's relation to G's shipwreck. I guess it might just come down to a difference of intuitions about what constitutes the origin of the shipwreck.
Mike,
Do you think that
(1) If (x did A because x believed p), then x's belief that p is part of the origins of A
is true iff x is not God?
Something like (1) seems right. The belief did have a causal role in bringing about A. But in the case of God, as you note at September 30, 2008 9:49 AM, beliefs probably do not have this causal role.
Divine beliefs may not have this causal role, but they surely have an explanatory role. Or at least the Molinist presumably will say things like: "God did A because God believed p." Surely there "God believed p" plays an explanatory role. And remember that I've defined essentiality of origins in terms of explanatory, and not just causal, priority.
I think I see what you have in mind with essentiality of origins. And the proof is clever. One other question. Suppose S = George is sshipwrecked. Now suppose essentiality of origins together with S entails, (S N-> H), where H = George acts heroically. S of course also entails S. So S entails (1).
1. S & (S N-> H)
and you note that,
2. N((S & (S N-> H)) -> H)
Of course that's right. Any world in which George is shipwrecked he acts herocally. But I don't see how freedom is undermined. Since it is also true that,
3. ((S & (S N-> H))-> M~H
That is, given that antecedent it is possible that George does not act heroically. This is just to say that it is possible that George acts in a way that brings about backtracking. In that case, he acts in such a way that ~(S N-> H) comes out true (though he does not cause it to be true).
I have a problem with your formulation of the Molinist position for reduction. You say:
But what God brings about is not that George "would be shipwrecked", but that "George will be shipwrecked." He does that on his knowledge of 1 of course. If the counterfactuals are a part of God's free or natural knowledge, then it is not Molinism.
If this is taken into account it changes things. You say:
"Therefore, that George is in F entails (1)."
Yes, but not by any necessity (at least according to Molina). It of course would be by necessity, if God also had decreed (brought it about that) 1.
What you are saying is that Georg being in a shipwreck entails that God knows that "Where George shipwrecked, he would freely behave heroically." But it only entails that by God's sovereign free decision, based upon his prevolitional (not free) knowledge, which in turn is determined by what George would do freely in a shipwreck. In order to get your argument to work you have to say that God freely knew all counterfactuals of creaturely "freedom." But this is not Molinism. God determines "will", on the basis of his knowledge of "would" determined by George.
I cannot see then, that you have any other argument than:
2) If George would be in a shipwreck, then he would freely act heroic
3) George will be in a shipwreck
4) Therefore George will freely act heroic
The only way I can see you get a necessity is by saying that 1 og 2 is true necessarily. Which would follow of God decreed the counterfactuals, but then it would not be Molinism.
David,
I think you're right (and I'm sure Alex knows) that God does not make any CCF's true (on Molinist assumptions). So, it is clearly 'will' and not 'would'. But his argument is otherwise sound. As I note above, if S = George is shipwrecked, then essentiality of origins together with S does entail, (S N-> H), where H = George acts heroically. And S also entails S. So S entails (1).
1. S & (S N-> H)
and Alex notes that,
2. N((S & (S N-> H)) -> H)
All of that is right. But what does not follow from this is that George does not freely act heroically. It's true that M~H, since it is true that M~(S N->H). It is worth noting that it's a controversial part of Molinism (at least as Tom Flint develops it) that we have this sort of backtracking all over.