July 20, 2020

This Post is Old!

The post you are reading is years old and may not represent my current views. I started blogging around the time I first began to study philosophy, age 17. In my view, the point of philosophy is to expose our beliefs to rational scrutiny so we can revise them and get better beliefs that are more likely to be true. That's what I've been up to all these years, and this blog has been part of that process. For my latest thoughts, please see the front page.

Almeida on Unrestricted Actualization

Molinism is the view that God has comprehensive knowledge of what free creatures would freely choose in any possible circumstance in which they might exist and be free. These kinds of propositions are called counterfactuals of creaturely freedom (CCFs). According to the Molinist, God knows these propositions, but cannot choose them. Although they are contingent, they do not depend on God's will. Instead, God exercises providential control by deciding which creatures to create and in which circumstances to place them while knowing what they will freely choose in those circumstances.

Theological determinism is the view that every contingent state of affairs is chosen by God.

Theological Compatibilism is the view that creatures may still be free even if theological determinism is true.

Molinism is supposed to be opposed to theological compatibilism. According to the Molinist, it is incoherent to suppose that God should choose which CCFs are true because CCFs specify what creatures would freely choose, and if God chose them then the creatures wouldn't be free. However, in various different places Mike Almeida has given what is, in effect, an argument from Molinism to theological compatibilism.*

I used to think that Almeida's argument was subject to an easy refutation. However, I've recently realized that the argument points to a rather serious problem in standard formulations of Molinism in recent analytic philosophy (e.g., Plantinga and Flint). That is, what I initially thought was a problematic assumption on Almeida's part also appears to be a commitment of the Molinist. Thus, even if Almeida's view of providence remains quite implausible, he has raised a quite significant problem for contemporary Molinists.

The Argument

Almeida concedes to the libertarian that if God caused an agent's choice then the agent would be unfree. Almeida assumes that if God issued a decree or fiat that the agent should make a particular choice, this would amount to God's causing that choice.** However, according to Almeida, instead of issuing a fiat, God could issue a prediction. In a recent reply to Bruce Langtry, Almeida confirms that it is his view that God's ability to make such a prediction is simply a direct consequence of divine omnipotence, requiring no further defense. However, it is metaphysically necessary that whatever God predicts occurs. Hence, God's predicting that p metaphysically necessitates the truth of p and so, if God can predict that p, then God can ensure that p will occur simply by predicting it.

The Easy Refutation

Almeida does not adequately distinguish between a decree and a prediction, and it is my contention that proper attention to this distinction undermines his argument. (This objection is similar, but not identical, to Langtry's.) In fact, the distinction between a decree and a prediction is a difference in the direction of fit. Once we recognize this point, the argument collapses.

I borrow the phrase 'direction of fit' from the literature on belief and desire. It is a standard view that belief and desire are distinguished (at least partly) in that the direction of fit for belief is mind-to-world and the direction of fit for desire is world-to-mind. That is, believers try to fit their beliefs to the world. This is often thought to be a constitutive norm on belief. Constitutive norms are like the rules of a game. If you don't recognize the force of those rules, you are not playing the game at all. Of course, you can cheat at a game, but when you do so you recognize that certain rules are in force and try to get away with breaking them. If you don't recognize the force of the rules at all (for instance, if you don't care whether you get caught), you're not cheating, you're just not playing that game at all. If you don't recognize the norm that beliefs are meant to conform to reality, you're not playing the game of believing.

By contrast, a desire aims to conform the world to itself. Some philosophers (e.g., the Stoics) have recommended trying to reform our desires so that we only desire things that are readily attainable. But this is not a constitutive norm on desire. A person who doesn't recognize the force of this rule may still be desiring. In desire, you want the world to be a certain way, and you are not (necessarily) making a mistake if the world is not already that way. It's normal to act to satisfy our desires, that is, to try to make the world conform to what we have in mind.

Now, a very similar contrast obtains between prediction and decree. The direction of fit for decree is, as it were, world-to-speech. That is, a person who issues a decree aims to make the world conform to her decree. But prediction is a kind of assertion, and a person who makes an assertion aims to conform her assertion to the world. The direction of fit is speech-to-world.

Lying, on this view, is a kind of cheating. The liar recognizes the force of the rule that says he should conform his assertions to the world but tries to get away with violating it. On the other hand, as Frankfurt famously argued, a person who doesn't even recognize the norm of truth is not lying, or even really asserting, at all. A person like that is just bullshitting (as Frankfurt calls it).

Now, necessarily, if there are any norms that are in any way binding upon God, then God adheres to them perfectly. But if God issues predictions or decrees, then there are such norms: the norm that predictions (as a kind of assertions) conform to the world, and the norm that the world conforms to decrees. But it follows from this that if God predicts that p, then it is because p that God predicts that p. God's prediction conforms to a logically prior fact. If God aimed to make p true by saying that p, then God's saying that p would be a decree and not a prediction.

What this shows is that in order for God to predict that I will freely eat toast tomorrow, it must be true that I will freely eat toast tomorrow prior to God's prediction. The fact that I will freely eat toast tomorrow must (partly) explain why God issued this prediction.

There are, of course, paradoxes about 'self-fulfilling prophecies' in the neighborhood: the prediction, if it is made before an audience, will influence the audience's behavior in various ways. Hence, if the predicted outcome depends on the audience's behavior then the issuance of the prediction may influence whether or not the predicted outcome occurs. However, these puzzle cases are not relevant for Almeida's argument. In these cases, the prediction causally influences the outcome, and this is precisely what Almeida is trying to avoid.

Almeida's Revenge?

I used to think that the easy refutation above was all that needed to be said about Almeida's argument. But I recently realized that, along the way to his conclusion, Almeida references some features of the expositions of Molinism in recent analytic philosophy that give some additional force to his argument.

Molinists like Alvin Plantinga and Tom Flint distinguish between strong actualization and weak actualization. The states of affairs God strongly actualizes are those God directly brings about (e.g., the occurrence of the Big Bang). God, according to these philosophers, cannot strongly actualize a created agent's free choice. This, according to them, is a contradiction. However, God knows what free agents would choose in any circumstance. Hence, if God brings it about that an agent faces a certain free choice God will thereby bring it about that the agent chooses in a certain way, since God knows 'in advance' that this will be the outcome of placing the agent in that circumstance. Now, Plantinga and Flint assume that for every possible world there is some state of affairs T that is the most inclusive state of affairs that God strongly actualizes in that world, and they assume that, by strongly actualizing T, God weakly actualizes that precise world, in all its detail. They further assume that which states of affairs God can strongly actualize does not vary from one world to another.

But now consider: there is a possible world at which God predicts that I will freely eat toast tomorrow. At that world, certainly God's own prediction is part of what God strongly actualizes. So if the states of affairs God can strongly actualize does not vary from world to world, then God can strongly actualize the state of affairs God's predicting that I will freely eat toast tomorrow at every possible world. However, there is no possible world at which God's predictions are false. Hence, there is a state of affairs God can strongly actualize which is such that, necessarily, if God actualizes it, then I freely eat toast tomorrow. It follows that in any possible world, regardless of the CCFs, God can ensure that I freely eat toast tomorrow.

Morals of the Story

Almeida does not adequately distinguish between predictions and decrees. When such a distinction is drawn, it appears to undermine Almeida's view. However, other features of standard presentations of Molinism in recent analytic philosophy appear to lead directly to Almeida's view. What is to be done?

One way or another, this shows that standard Molinism doesn't work, for two reasons. First, like Almeida, standard Molinism must implausibly reject the idea that God's predictions must necessarily be based on the predicted facts, rather than vice versa. Second, if we're willing to swallow that view about predictions then Almeida's argument goes through and Molinism collapses into theological compatibilism—a view it was specifically designed to avoid.

However, the Molinist argument for the claim that, in any possible world, God can strongly actualize the state of affairs my freely choosing to eat toast tomorrow depended on a quite specific assumption: that what we might call the maximal strongly actualizable states of affairs do not vary from world to world. In fact, the nature of prediction together with God's essential perfect truthfulness guarantee that the truth of a proposition must be (logically/explanatorially) prior to God's asserting that proposition. Since God's own assertions are strongly actualized by God, it follows that which states of affairs God can strongly actualize varies from world to world.

I conclude by noting one more tricky issue here. The argument of the previous paragraph assumes that God's essential perfect truthfulness means that God cannot say what isn't so. This, of course, raises issues about the compatibility of omnipotence with necessary moral perfection. Issues of this sort are among the reasons I have previously given in arguing that use of 'can' or 'able to' in connection with God is a kind of loose talk that is best avoided in metaphysics. It is misleading to talk about what God can or cannot do. Instead, we should simply say that God has infinite (unlimited) power. But this (perhaps rather idiosyncratic) view of mine requires rethinking this entire debate.


* I'm not sure that this is quite how Almeida himself views the significance of his argument.

** I have repeatedly attacked this assumption about the working of creation and providence.

Posted by Kenny at July 20, 2020 5:39 PM
TrackBack URL for this entry: https://blog.kennypearce.net/admin/mt-tb.cgi/874

Post a comment

Return to blog.kennypearce.net