A defense (in Plantinga's sense) against the logical problem of evil requires two components: a metaphysical component, which claims that a certain scenario is logically possible, and a value component, which claims that if the scenario in question were actual then it would be consistent with God's goodness to weakly actualize a world containing evil. In Plantinga's Free Will Defense (FWD), the scenario in question is one in which every creaturely essence suffers from transworld depravity (TWD). Now, in both The Nature of Necessity and God, Freedom, and Evil Plantinga's focus is squarely on the metaphysical component, defending the coherence of Molinism and the possibility of every creaturely essence suffering from TWD. The value component is almost completely ignored. Plantinga supposes that, if every creaturely essence suffered from TWD, then God would create a world with evil, and this would not in any way impugn his goodness. But why does Plantinga think this? I suppose he probably endorses:
(1) God's perfect goodness consists in his actualizing the best world he can
and
(2) If every creaturely essence suffered TWD, then the best world God could actualize would contain some evil.
However, both of these claims are open to question. First, a non-consequentialist (such as myself) will deny that (1) is any kind of obvious consequence of moral goodness. It might hold in the case of God, but it needn't hold in general. Second, (2) implies that creating Molinist-free creatures who sometimes go wrong is better than creating compatibilist-free creatures who always go right. (The plausibility of (2) also depends on how bad the choices the creatures are going to make will be, so we might want to be more precise about that.) This could be controversial as well.
I want to suggest an alternative value component for the FWD which I take to be better than the one I think Plantinga implicitly endorses.
Call a moral theory accommodationist if it holds that I ought to accommodate my actions to the immorality of others, even to the extent that, in an immoral world, I am sometimes obligated to do things I would be prohibited from doing in a moral world. Act utilitarianism seems to be accommodationist. For instance, it recommends that, if someone credibly promises to kill ten people unless I kill one, I should kill the one person, which I would be prohibited from doing were it not for the immorality of others. (There are complications related to punishing wrongdoers and so forth, but let's set these aside for now and assume that the distinction between moral theories I am getting at can be made out somehow.)
A moral theory is anti-accommodationist if it says that my obligations and prohibitions are not altered by the immorality of others. (This is consistent with claiming that the relative value of actions which are merely permitted changes.) Some versions of rule utilitarianism are anti-accommodationist: they say I should follow the rules such that it would be best if everyone followed them, and I should follow these rules even though others won't. Obviously deontological theories tend to be anti-accommodationist.
Now suppose we adopt an anti-accommodationist moral theory, and apply it to divine moral perfection. Now suppose that there is only one way for the BPW to come about, namely, for God to strongly actualize a certain set S of states of affairs which includes the existence of Molinist-free creatures, and for those creatures to then make all and only the morally correct decisions. Suppose, further, that, if it were the case that (if God should actualize S, then the creatures would make the right decisions, leading to the BPW), then God would be obligated to actualize S. Finally, suppose that all creaturely essences suffer TWD. The TWD of the creaturely essences is the immorality of others, so (by anti-accommodationism) God's moral obligations don't change to accommodate it. Therefore, God is still obligated to actualize S, even though creatures will make wrong decisions, so that God's actualizing S will not lead to the BPW. Thus, if anti-accommodationism is true, and the BPW contains Molinist-free creatures who make only right actions, and God is obligated to bring about the BPW if he can, then if all creaturely essences (or even only all essences which are instantiated in the BPW!) suffer TWD, then God will be obligated (!) to weakly actualize a world containing moral evil, even if he could have weakly actualized a better world.
Put in a more loose and intuitive way, my idea is this: Plantinga thinks that God can't bring about the BPW all by himself, without the cooperation of free creatures, and that there is nothing God can do to ensure that he will receive the cooperation of free creatures. If this is right, then perhaps God is nevertheless obligated (or at least permitted) to do his part toward the actualization of the BPW even if he knows creatures won't do theirs. If this was the case then what we would expect, in the scenario where all the creaturely essences instantiated in the BPW suffer TWD, is that God would create a world which was such that if all the creatures had always decided rightly, the BPW would have been actual, but which is nevertheless very different from the BPW.
One trouble with my approach is that, if it's really true that God is obligated to do this then, although we started out trying to make a mere defense, we end up committed to offering our story as (part of) a theodicy. More specifically, we end up committed to the claim that, in fact,
(3) if all creatures had always made only right decisions, the BPW would have been actual.
Is (3) plausible? Perhaps not, though I do like it rather better than the claim that, given how the counterfactuals of freedom were, this is the best world God could have created, which is a claim we would have to make if Plantinga's FWD were to become (part of) a theodicy. Note that this needn't commit us to the claim that we exist in the BPW, since the decisions of earlier creatures effect which later creatures exist. Also, I have strong anti-accommodationist intuitions, which makes the moral account I've just outlined more attractive to me than Plantinga's. Actually, if we could make some kind of sense out of a historical (even if not fully literal) Fall of Man, then (b) might start looking pretty good. Of course, that's an extremely hard thing to do.
To end on an anti-climax, I have serious reservations about the metaphysical component of Plantinga's FWD to begin with. However, I think it is worth thinking more seriously than has been done so far about the value component, and I think that the alternative sketched here is superior to the original.
(cross-posted at blog.kennypearce.net)
This is a cool proposal, and I am sympathetic with the main point of it--the idea that we can and should look for alternative ways of construing the value component. But I am not sure i understand your suggested alternative.
Is anti-accommodationism compatible with contrary-to-duty obligations? For example, if I lie I acquire a new duty. And if I hear you lie (and know you lied, etc), I also seem to acquire a duty I would not have had absent your lying. So, my duties seem to be altered because of the immorality of someone else.
Kenny:
1. I am not clear on what anti-accommodationism is. A naive version would prohibit punishing or morally criticizing people. (It's impermissible to morally criticize people unless they've done wrong.) I can't think of how to make a non-naive version that isn't ad hoc. Maybe, though, one can talk of degrees of accommodationism in moral theories. Thus, act utilitarianism is more accommodationist than rule utilitarianism.
2. I do think Plantinga needs more than TWD. Suppose, for instance, that the counditionals come out so badly that in every feasible world every creature chooses the worst option in every significantly free choice. Then the arguer from evil would have a pretty decent case that God should actualize a world without freedom.
So what Plantinga needs is the thesis that possibly in every feasible world containing a significantly free choice there is at least one wrong choice and there is a feasible world containing a significantly free choice which is worth creating. This seems pretty plausible, if one thinks TWD is posisble. Given the possibility of TWD, it sure seems plausible that it is possible that (a) every feasible world containing at least one significantly free choice contains at last one wrong choice and (b) at least one world exhibiting the following features is feasible:
- The first significantly free creature freely does wrong in her first choice, repents by God's grace a minute later, and grows to great righteousness.
- All other significantly free creatures are descendants of her. There are infinitely many of them, and they all live lives of moral perfection and great artistic and intellectual sophistication.
- There is no suffering except the deserved suffering of the first free creature when she regrets her sin. Apart from this minute of regret, all conscious beings live lives of great joy. No conscious being dies.
It seems very plausible that no world lacking significantly free creatures would be preferable to a world like this.
Why is it plausible that if TWD is possible, then this scenario is possible? Because we only have reason to think TWD is possible provided we believe in some sort of a broad independence principle for the truth values of conditionals of free will, one that allows a very wide variety of combinations of truth values to be assigned to them. And if we have that, then it should be possible that every world containing significant freedom contains at least one sin and a world like the above is feasible.
A better definition of the accommodationist/anti-accommodationist contrast is certainly needed. As far as punishment, moral criticism, etc., there is (for me, at least) an intuitive contrast between accommodating evil and responding to evil. So perhaps the distinction I want to make is about whether you should alter your actions in ways that would otherwise violate your obligations based on the expectation that you or someone else will do evil in the future. (The really objectionable cases - and these show up regularly in objections to act utilitarianism - are cases where I do some small evil now because I believe that if I don't then I myself will perform some greater evil later.) Of course, if God is atemporal, then that would be a reason for thinking that the distinction doesn't apply to him.
Alex, I think you are right that, insofar as we are trying to respond only to the claim (made, e.g., by Mackie) that the existence of any evil at all proves the non-existence of God then a case like yours is sufficient, and surely if Molinism is coherent then that scenario is possible. So Plantinga could have made his case with a much weaker claim than possibly all creaturely essences suffer TWD.
"So perhaps the distinction I want to make is about whether you should alter your actions in ways that would otherwise violate your obligations based on the expectation that you or someone else will do evil in the future."
1. Suppose I know that Smith will physically attack me some time over the next year, so I hire a bodyguard. I thereby perform an action that would otherwise violate my obligations to spend my money prudently (or to give to the needy, or the like). But given the expectation of an attack, the action is permissible. This generalizes to many cases where one's resource allocation would be immoral but for the expectation of wrongdoing by self or others.
2. In two minutes, Jones will introduce poison A into the city's drinking water. I have poison B. It turns out that B is an antidote for A and A is an antidote for B. So I introduce B into the water.
3. I assert that somebody will sin today. In so doing, I have performed an action that would be impermissible except on the expectation that someone will sin today.
4. You can see that Jones is about to try to impermissibly reveal out something that will cause grave harm to many. So you clamp your hand over his mouth. If he weren't about to try to reveal this, clamping your hand over his mouth would be a violation of our etiquette-based moral duties of interpersonal respect.
5. In ten seconds a device will be triggered which will kill everyone who is conscious. I grab a brick and hit you on the head just hard enough to make you temporarily unconscious.
The distinction is, as I said, not easy to formulate. Here's a paradigmatic statement of accommodationism:
Perhaps you (Alex) are right that accommodationism should be seen as coming in degrees. Hobbes is at the extreme degree: he holds that virtually all of our moral obligations are suspended in situations where we cannot reasonably expect others to fulfill their moral obligations. Perhaps no one has ever endorsed the most extreme conceivable positions on either side, namely, (a) that we have no obligations at all unless others can be expected to fulfill their obligations, or (b) that the expectation that others will not fulfill their obligations has no effect whatsoever on what course of action we ought to take. Hobbes comes pretty close to (a), but he holds that we still have an obligation to self-preservation, even in the state of nature, so even he doesn't go all the way.
I share the intuition that all of 1-5 are examples of cases in which the action would be impermissible, but is made permissible, or in some of the cases even obligatory, by the reasonable expectation (in some cases a higher degree of rational confidence than that may be required) that others will act wrongly. So I don't want to endorse anything nearly as strong as (b). Of course we need to be careful with wide vs. narrow scope conditional obligations (see Schroeder on hypothetical imperatives). If we are talking only about wide scope obligations [O(if p then do A)], then it may be that, even in 1-5, there is no change in our obligations at all when we expect others to behave badly; if the obligations are narrow scope [if p then O(do A)] then what obligations we have will change.
On the other hand, even act utilitarianism could be formulated in terms of exceptionless wide scope conditional obligations (or just one universally quantified wide scope obligation: O(for any action A, if A uniquely maximizes happiness, then do A)), so this won't capture the distinction I want either.
In the end, it may be that the best we can do is to say that there is a spectrum of views and that there is nothing to be gained by trying to draw a precise line on it. I take it that this was your (Alex's) original suggestion.
Here's an only somewhat vague formulation that might be helpful:
Anti-Accommodationism =df. The view that it is frequently (or always) the case that one ought to act as if others will fulfill their obligations, even if one knows or reasonably believes that they will not.
Accommodationism =df. The view that it is rarely (or never) the case that one ought to act as if others will fulfill their obligations, when one knows or reasonably believes that they will not.
The vagueness of course enters in 'frequently' and 'rarely,' but some vagueness seems unavoidable. So we still get a spectrum of views as discussed.
One concern with these definitions is that deontologism will render expectations about what others will do irrelevant in a wide variety of cases, and so, with respect to these cases, will satisfy both definitions vacuously. This does, however, capture the relevant difference between act utilitarianism and rule utilitarianism, and as long as a given deontological theory has some non-vacuous cases (as I think a plausible deontological view will) it can be classified one way or the other (subject to vagueness) as well. However, contrary to what I was originally thinking, there will be both accommodationist and anti-accommodationist deontological views, and it's not clear which side Kant, or most contemporary Kantians, will end up on.
In the context of the value component it is I think important not to consider creation as separate from its creator. The value of creation resides not only in how it is, but also in that it is created by God. The world does not lie outside of God, and God does not lie outside of the world. Thus how God takes part in creation is one major ground of its value. Plantinga somewhere suggests that the greatest created worlds are those in which the vastly beautiful event of atonement obtains, and thus only worlds in which evil, indeed serious evil, at first exists.
My comment on the accommodationist versus anti-accommodationist moral theories is that in my view morality is not about obligations and prohibitions, especially in the case of God. I do not consider morality to be about conformity to rules, and I find it very natural that ethicists should have failed to produce a mechanistic or rule-based theory of ethics. Rather I consider morality to be fundamentally creative, i.e. the creative (and therefore to some degree unexpected) expression of the love of a good person. Now the theist necessary holds that our world is the one God wished to create, and it is I think productive to try to understand why the greatest conceivable being would wish to create our kind of world, and not so much why our kind of world is by itself better than any other kind.
Finally, and in this same context, I hold that to think of what world God “could” create is misleading, since the meaning of “could” applies only to imperfect persons. The only open question is what world God would wish to create.
Another question in the vicinity is whether accommodation to reasonable expectations of counterfactual free actions falls under accommodationism.
Presumably the situation for the Molinist optimalist looks something like this. Let w0 be the best world (or at least a very good world, much better than ours). w0 is not feasible. Let w1 be our world. Let c(w) be the core of w--say, the conjunction of all the propositions at w that God strongly actualizes (it's a bit more complicated than that, but let's just suppose so). Then God knows that if he were to actualize c(w0), he'd get a world w0' (such that c(w0')=c(w0)) worse than w1, which is what he'd get were he to actualize c(w1). This is because the bad choices that creatures would make were he to actualize c(w0) would produce something bad, while the better choices that creatures would make were he to actualize c(w1) would lead to a better overall world.
Thus, God accommodates himself not to the actual bad choices in the actual world w1, but to the counterfactual bad choices in w0', where w0' is the unique feasible world with the same core as w0.
Now such accommodationism seems quite unproblematic. It is surely quite unproblematic to avoid choices that one knows would lead to others behaving particularly badly.