A Short Dialogue on Plantinga's Free Will Defense (Revised and Expanded a Bit)

A. There is at least one metaphysically possible world at which every single person God could've created suffers from transworld depravity (TWD).[1]

B. How could that be?

A. We can support the thesis as follows: (1) Each creaturely essence is transworld depraved at some possible world or other. Therefore, (2) there's bound to be one possible world at which every creaturely essence is transworld depraved.

B. Why think that? First, the notion of transworld depravity relies on the notion of counterfactuals of creaturely (libertarian) freedom (CCFs). But there are powerful reasons to think that the notion of a CCF is incoherent.

Second, (1) would entail that no possible creature is essentially morally perfect. But if you allow that a god could be essentially morally perfect, then there is pressure on the proponent of the free will defense to give a principled basis for why a created being cannot be essentially morally perfect as well.

Finally, the inference from (1) to (2) commits a modal operator shift fallacy:

1. (x)(CEx  --> <>TWDx)
2. <>(x)(CEx  --> TWDx)

Such is the same illicit pattern of inference involved in reasoning that if each book in the library is such that it's possible to read it in a single day, then it's possible to read each book in the library in a single day. In fact, the inference from (1) to (2) relies on the unargued assumption of the truth of Interworld Plenitude. But if Intraworld Plenitude better captures the distribution of counterfactuals of creaturely freedom, then not only is the inference from (1) to (2) demonstrably false, but (2) is false as well. And as it turns out, Plantinga and others have (to date) failed to show that the former is more plausible than the latter. Therefore, not only is the inference from (1) to (2) undercut, but (2) is undercut as well.

Third, there are burden-shifting grounds against (2).   For what grounds the counterfactuals of freedom here? Presumably, it's the character or nature of a given being. This seems to be the most plausible explanation for why it's supposed to be that the redeemed in heaven never do wrong.  So, for example, the Christian tradition speaks of the renovation of the "heart" of the Christian into one that, like God's, finds sin repulsive -- or at least uninteresting --, and which delights in what is right and good. But if so, then all that's required to create a free being that wouldn't sin is to endow them with such a "heart", and with a sanctified and glorified nature or character. (This needn't obviously require that there is no possible world at which such creatures freely do wrong, but only that, relative to the possible world in question, there is no "close" world (in the Lewis/Stalnaker sense) at which they do wrong).  But, prima facie, it's possible for an omnipotent being to create beings and endow them with such a nature or character at every possible world at which the former exists -- indeed, it's a standardly held view among Christians that God will perform such a transformation on the saints when they reach the afterlife.  But if so, then there is no metaphysically possible world at which every creaturely essence God could create suffers from transworld depravity. And if that's right, then Plantinga's free will defense is unsuccessful.

A. I'm not sure you're right about that. For suppose God creates creatures with a "heavenly" character from the get-go. Then since they didn't form their character through their free choices, they're not free at all. But whether or not we say such beings are free, perhaps we should say that it's better to be able to shape one's character through free choices than to come "ready-made" as it were with a morally perfect character.

B. If you're right, then it would seem to follow that since God is supposed to be morally perfect essentially, he's had his morally perfect character or nature from the get-go, in which case he's not free. Furthermore, the same assumptions seem to imply that God's essentially morally perfect character is inferior to those whose character was shaped by their free choices.

A. Nevertheless, I still think there's a crucial difference here between a perfect-from-the-get-go God and a perfect-from-the-get-go created person. For unlike the created person, God is the ultimate source of his actions. That is, he acts on his own reasons, and nothing external to God determines his actions. Because of this, God is free and morally responsible. By contrast, if God created persons with a morally perfect nature, they could not be the ultimate source of their actions. Rather, God would be the ultimate source, as he would be an external cause of their nature, which in turn would ensure that their actions are always good. The only way for God to create free creatures, then, is to create them with the ability to choose between good and evil. Therefore, while the freedom of created beings requires the ability to do evil, God's freedom does not.[2] 

 B.  That's an ingenious and elegant proposal, but I'm not persuaded. Here's why. Consider two finite persons, Alpha and Beta[3]. Alpha and Beta are both morally perfect, and thus unable to do what is morally wrong. They differ, however, in their origins: while Alpha was made to be morally perfect by external causes (e.g., heredity and environment, God, etc.), Beta was not. Rather, Beta just popped into existence. Now if the account of freedom and responsibility you propose is correct, then we should say that while Beta is a free and morally responsible agent, Alpha is not. For while Alpha's nature was caused by an external source, no external source caused Beta's nature. And because of this, we should evaluate the following subjunctive conditionals differently:

1. If Alpha existed, she'd be free and morally responsible. (F)
2. If Beta existed, she'd be free and morally responsible.   (T)

But this doesn't seem right: whether their natures had an external cause doesn't seem to make a difference to the issue of whether they're free or responsible.  What matters here is that the actions of both are due to natures they did not create and for which they are not responsible. Thus, either being is free and morally responsible just in case the other one is. The proposal is therefore an unsuccessful reply to the criticism of the free will defense at issue.

A. That's an equally ingenious and elegant rejoinder, but I'm not persuaded. For the scenario depicted in your thought experiment seems metaphysically impossible. For it seems metaphysically impossible for a being to just pop into existence. And if it's not metaphysically possible, it can't undermine my account.[4]

B. Whether the scenario is metaphysically possible or not is irrelevant. For per impossible arguments clearly have epistemic force, and have proper use in the evaluation of counterpossible conditionals.[5] Therefore, even if the thought experiment should turn out to depict a metaphysically impossible scenario, we can construe the subjunctive conditionals above as counterpossible conditionals, and we can use the thought experiment to evaluate them. The rejoinder is therefore unsuccessful.[6]

The moral seems to be this: According to orthodox Christian theism, there are at least three sets of free beings: (i) the fallen creatures on earth now, (ii) the redeemed and glorified creatures in heaven, and (iii) God. But the problem is that Plantinga's free will defense gives an account of free beings in (i) that's prima facie incompatible with an account of free beings in (ii) and (iii). Plantinga's free will defense is therefore an unsuccessful response to the logical problem of evil.
[1] Plantinga's claim here is one about the counterfactuals of creaturely freedom. He is arguing that there is a metaphysically possible world W at which every single one of the (non-denumerably?) infinitely many libertarianly-free creaturely essences he could create at W would do wrong at least once; the would-counterfactuals of freedom of each such creaturely essence are "messed up" at W, so that although there may be metaphysically possible worlds W*1-W*n where each such creature always freely does right, none of those worlds are "close" (in the Lewis-Stalnaker sense) to W. As such, those are metaphyically possible-yet-infeasible worlds.
[2] Cf.Morriston, Wes. "What is so Good About Moral Freedom?", The Philosophical Quarterly 50:3 (2000), pp. 343-358. 
[3] The following is a paraphrase of Ibid
[4] Ibid.
[5] Indeed, counterpossible reasoning had better be legitimate; otherwise it'd be impossible in principle to evaluate competing philosophical theses, each of which is necessarily true if true at all.
[6] Ibid.


James said...

Well, if there's a possible world where the proposition P(x) holds, and P(x) doesn't entail a contradiction with P(y) (or anything P(y) entails), mustn't there be a possible world where P(x and y) holds?

And doesn't it seem plausible that "x is TWD" doesn't contradict "y is TWD" unless x = y?

exapologist said...

Hi James,

Good questions. My own worry here is that the existence of an Anselmian being would affect the structure of modal space, so that some things that are intrinsically metaphysically possible are yet extrinsically metaphysically impossible. In particular, the essential moral perfection of an Anselmian being might constrain its will in such a way that it can't cause or permit a world at which there are only transworld depraved creatures.

exapologist said...

Update: I'd add that accepting the inference from (1) to (2) relies on an unsupported acceptance of Interworld Plenitude over Intraworld Plenitude.

wissam h said...

Your argument is very similar to Quentin Smith's Logical Argument from Evil. Morriston made the same points before as well. I find the arguments pretty persuasive, but I found the transcript of a lecture by Inwagen which goes against them.

I thought I'd share: http://www.giffordlectures.org/Browse.asp?PubID=TPTPOE&Volume=0&Issue=0&ArticleID=7