In chapter 5 of Divine Providence: The Molinist Account (1998), Thomas Flint defends a response to the grounding objection which he attributes to Alfred Freddoso. According to the Flint-Freddoso line, there are difficulties about future contingents which are exactly parallel to the difficulties about counterfactuals of creaturely freedom, and solutions to the problems about future contingents can be adapted to provide equally plausible solutions to the problems about counterfactuals of creaturely freedom. This claim is false.
The exact formulation of the grounding objection is a little tricky. Some philosophers take it to be based on the (questionable) assumption of some form of truthmaker theory, i.e., the notion that if a sentence/proposition is true then its truth must somehow be grounded in an actually existing concrete entity. This kind of very abstract claim about truth is quite controversial and can easily be rejected by the Molinist. However, the objection can be stated much more compellingly by keeping the focus on free will, which is of course the Molinist’s main concern. The Molinist endorses a negative thesis about freedom, namely, that my action is unfree if that action is determined by anyone or anything other than me. However, if this negative thesis were the Molinist’s whole conception of freedom, then the Molinist would succumb to the randomness objection to libertarianism: she would be unable to distinguish between an indeterministic spasm and a genuinely free action. Accordingly, the Molinist should conjoin to this negative thesis the positive thesis that an action is free only if it follows from my (undetermined) causal activity. But then, according to the Molinist, all of the counterfactuals regarding my free choices are determined and known by God in a manner that is logically independent of my even existing (let alone choosing), so it seems that it is not my undetermined causal activity that makes the counterfactuals true, and the same ought to be true of the subjunctive conditionals with true antecedents (since those would have remained true even if God had decided not to create me). Accordingly, I am not free in any positive sense, since all of my choices are determined by the prior truth of the counterfactuals and not by my spontaneous causal activity.
One response to this objection the Molinist should not make is that the determination in question is okay because it’s not causal determination. If the Molinist made this response, a Thomist or Leibnizian opponent would reply that it is perfectly consistent with their view that our actions might be free from external determination by natural causes (and, indeed, both the Thomist and the Leibnizian will insist that our actions are indeed often free from such external determination). As Leibniz expresses the matter:
Since, moreover, God’s decree consists solely in the resolution he forms, after having compared all possible worlds, to choose that one which is the best, and bring it into existence together with all that this world contains, by means of the all-powerful word Fiat, it is plain to see that this decree changes nothing in the constitution of things: God leaves them just as they were in the state of mere possibility, that is, changing nothing either in their essence or nature, or even in their accidents, which are represented perfectly already in the idea of this possible world. Thus that which is contingent and free remains no less so under the decrees of God than under his prevision. (Theodicy, tr. Huggard, sect. 52)
If the Molinist is to have grounds for rejecting Leibniz’s view, she has to insist that it is not only (natural/secondary) causal determination that interferes with freedom, but any kind of determination whatsoever. Hence determination by the prior truth of counterfactuals of creaturely freedom must, on the Molinist’s view, be inconsistent with freedom.
Now consider the Flint-Freddoso response. According to this response, the issue here is exactly parallel to the issue about future contingents. (Note that Leibniz makes the same claim about his compatibilist response.) It is true now that I will freely eat breakfast tomorrow. But if it is already true now, then doesn’t that mean I won’t be free, since the truth of this proposition determines that I will eat? Note again that the Molinist can’t say that this doesn’t matter because the determination is not causal, or else the Thomist or Leibnizian comes back with a distinction between primary and secondary causation.
Flint argues that a particular solution to the problem of future contingents can be adapted to the counterfactual case. According to this solution, a future claim counts as grounded iff the grounding will happen in the future. Similarly, a counterfactual claim counts as grounded iff the grounding would happen if the antecedent were true. This solution, however, cannot succeed without surrendering the Molinist’s claim to a more robust notion of freedom than the Thomist or Leibnizian, for here we are saying, effectively, the if the antecedent were true I would exercise undetermined causal efficacy to make the consequent true. But this is exactly what Leibniz says: God sees, in that other possible world, that the manner of causation I will exercise will be free causation. By actualizing that world, he doesn’t make the causation any less free. The Molinist now lacks motivation for saying that God couldn’t actualize that other possible world at which I freely take the opposite action in exactly the same circumstances.
Flint’s formulation of the solution to the problem of future contingents is complicated by a desire to remain neutral in the debate between presentists and eternalists in the philosophy of time (or perhaps by an endorsement of presentism – it’s not really clear). Endorsing eternalism makes the solution to the problem of future contingents easier to state, and more plausible. At the same time, it makes it clearer why the parallel solution to the problem about counterfactuals is not plausible. If eternalism is true, then we can say that the future contingent claim is made true by the fact that at that future time I actually do exercise undetermined causal influence and thereby bring it about that I eat breakfast. The future time really exists. (It is true now that it exists, although it is, of course, located in the future.) My free choice really happens at that time. That’s what makes it true. Nice and simple.
Now consider the parallel move for the counterfactuals. Here we’d have to say that it’s because I exercise undetermined causal influence at some other possible world that the counterfactual is true. But note that if it’s enough for me to exercise undetermined causal influence according to some abstract possible world then we’re back at Leibniz: why can’t God just make that world actual without altering the manner of causation I exercise? What we need, if this is going to be parallel to the case of eternalist future contingents, is for me not merely to be represented as exercising undetermined causal power, but actually doing it. This means that, in order for the Molinist to make the parallel move, we need (a) realism about the feasible worlds (but not the other merely possible worlds); and (b) transworld identity across feasible worlds. In other words, we need it to be the case that I myself actually face every choice which it is metaphysically possible that I face. Needless to say, eternalism is much easier to swallow than this. Accordingly, the grounding problem for Molinist counterfactuals is really not parallel to the problem of future contingents.
(Cross-posted at blog.kennypearce.net)
Molinists urge that we can avoid necessitarian conclusion–the conclusion that there is just one possible world–if it is true in some worlds that God is not able to actualize the best world. This is false. The necessitarian conclusion follows from the plausible principle that God must actualize the best possible world, if there is a best possible world. I don’t think it’s difficult to show that there must be a best possible world, so I leave it as an exercise. Here’s the proof contra the Molinist.
1. Necessarily, God actualizes the best possible world. Basic Principle
2. God is essentially omnipotent, omniscient, omnibenevolent & necessarily existing (all as a matter of absolute necessity). Assumption.
3. w is the best possible world. Assumption
4. God actualized w in w. From 2,3
5. It is true in w that necessarily, God actualized w. From 1, 4.
6. Necessarily, God actualized w. From 5, S5
7. w is the only possible world. From 6
8. Necessitarianism is true. From 7.
Fundamental Molinist conditionals of free will about non-existent agents are brutish: they are not grounded in other propositions, nor made true by a truthmaker, lack of a falsemaker and/or the obtaining of properties/relations between entities.
Now, suppose as seems plausible to me that there are precisely two kinds of explanation: constitutive-style and causal-style explanations. Constitutive-style explanations explain a truth by explaining how the truth is grounded: the knife is hot because its molecules have high kinetic energy. Causal-style explanations explain a truth by giving non-grounding conditions that nonetheless in a mysterious but familiar causal or at least causal-like give rise to the holding of the truth.
Now, brutish truths have no constitutive-style explanations. For the constitutive-style explanation involves the describing of a grounding. But brutish truths also have no causal-style explanations. For causal-style explanations involves the describing of causal-style relations between the aspects of the world (in the concrete sense) that ground the explanandum and explanans. (In fact, for this reason, brutish truths not only lack causal-style explanations but are not causal-style explanations for anything else.) So, brutish truths have no explanations.
But if there are true fundamental Molinist conditionals of free will about non-existent agents, there will also be ones that have explanations. For, some, maybe all, free actions can be explained in terms of the reasons the agent had. Thus, Curley accepts the bribe because he wants to be richer. Granted, this is a non-necessitating explanation–that Curley wants to be richer does not entail that he accepts the bribe. But that’s still an explanation, and one of causal-type. And exactly parallel explanations can be given for Molinist conditionals. Thus, Curley would have accepted the bribe in circumstances C because circumstances C includes his wanting to be richer. And presumably this kind of explanation would have held even had Curley never existed, and presumably if Molinism is true, there are such explanations for true conditionals about actually non-existent agents. Thus some fundamental Molinist conditionals of free will about non-existent agents can be explained. But this contradicts their brutishness.
Moreover, presumably some fundamental true Molinist conditionals of free will about non-existent agents explain God’s creative inactions. Thus, perhaps, God did not create Badolf Bitler, because Bitler would have been so much worse than Hitler. But these conditionals do not provide a constitutive-style explanation for such actions. So they provide a causal-style explanation. But they can’t do that, because they’re brutish.
The same argument goes against Merricks-style presentism on which fundamental truths about the past are brutish. But many, perhaps all, fundamental truths about the past are explained by other fundamental truths about the past.
Suppose Molinism is true. We know the truth values of some Molinist counterfactuals because we know that their antecedent and consequent are true. But we also have reason to believe many other Molinist counterfactuals. Absent further evidence, if P(A|C) is high, and C is an appropriate antecedent for a Molinist counterfactual C→A, that gives me reason to believe C→A. It certainly gives me reason to believe C→A if I know C is actually true; for if I know C is true, then if P(A|C) is high, P(A) will be fairly high as well, and so A is probably true, and hence C→A is probably true. But I also have reason to think C→A is true in cases where C is false. For instance, if Jones is the sort of person likely to accede to my minor requests, then I have reason to believe that were I to make such-and-such a minor request, he’d accede to it, and I have reason to believe the conditional whether or not I make the request (at least assuming Molinism is true so that the conditional has non-trivial truth-value).
This suggests that if the objective probability of A on C is high, then the objective probability of C→A is also high. So the Molinist conditional C→A, assuming it’s true, doesn’t seem to be a mere brute fact. It is a fact subject to meaningful probabilistic assignments. But if it’s not a mere brute fact, it seems reasonable to look for an explanation of it. What is that explanation?
Well, maybe we have a probabilistic explanation. Maybe the fact that C makes A probable explains why C→A. But this is weird. It seems that probabilistic explanation is a species of causal explanation (with probabilistic causation). But there is surely no causal explanation of why C→A, at least in worlds where C is not true. (What would the cause be? The truthmaker of C? But C is not true and has no truthmaker.)
I’ll leave it as a puzzle: How is a Molinist to explain the connection between P(A|C) and the probability of the conditional C→A?
For simplicity, I shall ignore the distinction between God talking and Jesus talking. I shall also write “deny” for “deny libertarian-freely” (note: typical libertarians allow for the possibility of free choices that are determined by character and circumstances, as long as the agent had a sufficient role in forming the character through properly indeterministic choices; it is only the latter that I will call “libertarian-free”). Take the case where God tells Peter that Peter will deny him. What divine knowledge was the prophecy based on? Suppose we say: God tells Peter that Peter will deny because God knows that Peter will deny. This would be a simple-foreknowledge (SF) account of prophecy. Now we have an apparent circularity in the order of explanation. God telling Peter that Peter will deny is explanatorily prior (“e-prior”) to Peter’s denial–it affects Peter’s state of mind when choosing whether to deny. But Peter’s denial is, presumably, e-prior to God’s knowing that Peter will deny. (Thomists and Calvinists will likely deny this. And so such Thomists and Calvinists will have no difficulty.) And God’s knowing that Peter will deny is e-prior to the prophecy. So we come full circle.
There is a way out of this argument: God ensures that Peter’s choice whether to deny is causally isolated from Peter’s memory of the prophecy. This breaks the circle, since then God’s prophesying to Peter that Peter will deny will no longer be e-prior to Peter’s denial. Moreover, Scripture says that only after the denials did Peter remember the prophecy, so there is some exegetical ground for supposing some causal isolation.
The difficulty with this SF account of prophecy is that it only makes prophecy possible in cases where the prophecy is isolated from the prophesied event. I shall argue that the Molinist may face a similar problem.
Here is an argument against Molinism, which while valid, is fallacious in an interesting way. This argument is an improved version of one that I have earlier defended.
- God brings it about that x is in circumstances C because of God’s belief that x would freely do A in C. (Hypothesis for reductio)
- If y brings it about that p because of y’s belief that q, then y’s bringing it about that p because of y’s belief that q is causally prior to p’s holding. (Premise)
- If x freely does A in C, then x’s being in C is causally prior to x’s freely doing A. (Premise)
- If E is causally prior to F, and the occurrence of E entails the occurrence of F, then E deterministically causes F. (Premise)
- That x freely does A is not deterministically caused by anything. (Premise)
- p is entailed by its being the case that God does B because of God’s belief that p. (Premise)
- Causal priority is transitive. (Premise)
- God’s bringing it about that x is in circumstances C because of God’s belief that x would freely do A in C is causally prior to x’s freely doing A in C. (By 1, 2, 3 and 7)
- That God brings it about that x is in circumstances C because of God’s belief that x would freely do A in C entails that x freely does A in C. (By 6)
- God’s bringing it about that x is in circumstances C because of God’s belief that x would freely do A in C deterministically causes x’s freely doing A in C. (By 4, 8 and 9)
- 10 contradicts 5.
What is wrong with the argument, I think, is the seemingly innocent (4). Claim (4) commits a mistake that I have identified elsewhere, the mistake of thinking that one can define concepts conjunctively. Deterministically causing is not just a conjunction of causing and logically determining (i.e., entailing), just as causing intentionally is not just a conjunction of causing and intending. The standard example for the latter is something like: George is pointing a gun at Bob and intends to kill Bob, and George’s intention to kill Bob causes his hands to shake and accidentally squeeze the trigger. Then George intended and caused Bob’s death but did not intentionally cause Bob’s death. For x to intentionally cause B, it has to be the case that x intends B and x causes B, but these two facts also have to be related in the right way. Likewise, for A to deterministically cause B, it has to be the case that A causes B and that the occurrence of A entails the occurrence of B, but these two facts also have to be related in the right way.
I don’t have a counterexample to (4). It could even be that (4) is true for some deeper reason. But as it stands, with (4) being presented simply because of its intuitive plausibility, the argument is fallacious in the following sense: Its plausibility rests in part on a cognitive fault of the interlocutor. The cognitive fault is that we have a tendency to accept conjunctive characterizations like (4) when we should always be suspicious of conjunctive characterizations, because just about always one needs the conjuncts to be satisfied in an appropriately related way. I think this may be because our minds automatically assume an appropriate connection between conjuncts, even if a statement does not give one. Consider “He pressed the trigger and the gun went off.” We automatically assume that the speaker is telling us that the gun went off because of the pressing of the trigger. But no such claim is made.
Suppose we fix up (4) by adding that the entailment must be appropriately related to the causal claim. But now (10) cannot be derived, because we don’t have an argument that in that case the appropriate relation holds.
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).
Suppose that a Molinist God creates a world where there is a sequence of 1000 indeterministic throws of a fair coin, and suppose that middle knowledge extends to stochastic non-agential events. (My argument will also apply in the case of Thomist God who determines indeterministic events.) Suppose 514 of the coin throws, let us suppose, are heads and 486 are tails. Consider the fact p that approximately half of the throws landed heads. A standard scientific explanation of p would involve the following facts:
- The coin was fair: heads and tails each had probability 1/2.
- The individual throws of the coin were independent of one another.
- If (1) and (2) hold, then by an appropriate version of the Law of Large Numbers, it is likely that a sequence of 1000 throws of the coin would have approximately half of them be heads.
Fact (3) is a mathematical fact. Facts (1) and (2) are concrete facts about the situation at hand, and both are essential. If (1) is false, we might well expect a different heads-to-tails ratio. If (2) is false, then the Law of Large Numbers need not apply.
But this scientific explanation is unlikely to be correct if Molinism holds. For if Molinism holds, then God in effect controls what sequences of throws come up, by choosing the antecedents of counterfactuals. God makes the choice of sequence based on global providential considerations. Since the sequence is chosen on the basis of considerations of the sequence as a whole, it seems unlikely that the items in the sequence will be independent.
Suppose we say, as I suggested in the previous thread in response to Mike’s related concern, that God deliberately chooses a sequence of events that is statistically apparently random. Then p will still be true–about half of the throws will land heads. However, (2) will not be true, at least not if we condition on God’s choosing a sequence of events that is statistically apparently random. For, if (1) and (2), hold we have a non-zero probability that all the throws will be heads. But conditionally on of (1) and the claim that God chose a sequence of events that was statistically apparently random, we get a zero probability that all the throws will be heads, since if all the throws were heads, the sequence could not be statistically apparently random.
Perhaps we shouldn’t condition on God’s choosing a sequence of events that is statistically apparently random. But if we don’t condition on that, then to check whether (1) and (2) we need to compute the probabilities of all the possible choices God could have made. And we have little reason to think (1) and (2) will hold then.