I've wanted to discussthis paper on hell (PDF) by Ted Sider (via Scottish Nous) for some time. I cannot count on two hands all the things that are wrong or confused or purely arbitrary about the argument, but I don't want to nitpick. After all, the doctrine of hell is a complicated thing; tackling it in a single paper requires a few simplifying assumptions and shortcuts. I do want to talk a bit on how it relates to a recent argument of interest in metaethics.
The basic target of Sider's paper is a doctrine of hell that holds:
Dichotomy: There are two and only two states of the afterlife (heaven and hell).
Badness: Hell is much worse than Heaven (i.e., there is a sharp and serious difference between heaven and hell)
Non-universality: Some go to heaven, some go to hell.
Control: God determines the criterion for who goes to hell.
The argument is that this violates divine justice: "any just criterion must judge created beings according to a standard that comes in degrees, or admits of borderline cases; but no such criterion can remain simultaneously just — or at least non-arbitrary — and consistent with the nature of the afterlife just described" (p. 2). The idea is that justice must be proportionate to the factors upon which the criterion depends; people very similar in the relevant way must be treated very similarly. Because of this, someone who holds the doctrine of hell described above (and the doctrine that God metes out consequences with strict proportional justice) must conclude that the criterion cannot depend on a matter of degree. Sider recognizes that these conditions are rather restrictive; as he specifically notes, they mean that his argument doesn't work against pessimistic universalism (the doctrine that we will all be damned), or against optimistic universalism (the doctrine that we will all be saved), 1 or Calvinism. Nor does it work against Muslim doctrines of hell. It also doesn't work against Catholic doctrines of hell, since I know of no Catholic version of the doctrine that holds that God determines these consequences with strict proportional justice. Indeed, I don't know of any doctrine of hell that involves such a view; they usually involve more than one principle, i.e., more than proportional justice. Indeed, I take it that the traditional view on both Catholic and Protestant sides is that God's acting according to proportional justice is conditional on more fundamental attributes, which proportional justice necessarily presupposes even to exist. (On such a view, God is not violating proportional justice when he seems superficially to deviate from the principle of proportional justice, because in such cases the conditions that are necessary for the principle even to be relevant haven't really been met.) So I'm not sure who is supposed to be holding the inconsistent pentad (the four conditions and the strict proportional justice requirement). But presumably somebody could be found, somewhere, so we can just consider the argument itself.
This argument is actually very similar to an argument that has recently been discussed in metaethics, which comes to the conclusion that it is impossible to be moral. The proof begins with two metaethical principles: 2
PE (Principle of Equality) If A and B are the same in every morally relevant respect, then A and B must receive the same moral treatment.
PDT (Principle of Differential Treatment) If A and B differ in some morally relevant respect, then A and B must receive different moral treatment.
Let's take a particular case to show how this sort of proof runs.
(1) A is a person.
(2) C is a non-person.
(3) B is indistinguishable from A with respect to the properties relevant to being a person.
(4) B is indistinguishable from C with respect to the properties relevant to being a non-person.
(5) B and A must receive the same moral treatment.
(6) B and C must receive the same moral treatment.
(7) A and C must receive the same moral treatment.
(8) A and C must receive different moral treatment.
(5) follows from (1)+(3)+PE; (6) follows from (2)+(4)+PE; (7) follows from (5)+(6)+transitivity; (8) follows from (1)+(2)+PDT. Since (7)+(8) is a contradiction, we seem to have a problem. This problem can be generalized quite widely; the problem that is here noted is a problem that will arise in the case of any vague moral properties, including justice.
It can easily be seen that Sider's argument is just a version of this. His principle of proportionate justice is a version of PE; PDT is not explicitly stated, but is required by Sider's talk about the criterion; and Sider's argument also builds on vagueness. In fact, his method is exactly described by the above argument, too. The basic version here is:
(1') A is saved.
(2') C is damned.
(3') B is indistinguishable from A with regard to the properties relevant to being saved.
(4') B is indistinguishable from C with regard to the properties relevant to being damned.
(5') B and A must receive the same moral treatment.
(6') B and C must receive the same moral treatment.
(7') A and C must receive the same moral treatment.
(8') A and C must receive different moral treatment.
The derivations are basically the same. (5') follows from (1'), (3'), and the principle of proportionate justice; (6') follows from (2'), (4'), and the principle of proportionate justice; (7') follows from (5'), (6'), and transitivity; and (8') follows from (1'),(2'), and the ex hypothesi assumption of the criterion for the damned and the saved.
We find, then, that Sider's argument doesn't identify an issue with the doctrine of hell itself; it identifies a problem that arises in any case of vagueness about moral properties. This in itself should be enough to make us doubt the argument. It's also important to note, however, that Sider has to have a much stronger argument than the person putting forward the Moral Impossibility Proof. The Moral Impossibility Proof requires only an interpretation of (3) and (4) and their cognates that appeals to indistinguishability for us; the Proof is as serious if it only applies to us as it is if it applies to everyone. Sider, however, must argue (and does not really argue) for an interpretation of (3') and (4') in which the relevant cases are indistinguishable to omniscience. So if we accept Sider's argument at all, we seem committed to a form of deep moral nihilism; in fact, the impossibility of proportional justice, or of any morality, is more easily proven this way than is the impossibility of hell.
Of course, as Almeida notes with respect to the Impossibility Proof, the argument relies on an equivocation in (3) and (4). I refer you to his discussion, but here's a hint: Does (3) imply that it is true that B is in the range of determinate persons, or does it imply that it is not false that B is in the range of determinate persons? The equivocation is made all the more serious in Sider's argument, since Sider quickly forgets that he needs to show not just similarity but morally relevant similarity, where 'morally relevant' is determined relative to a criterion selected by a divine intellect. But I thought it was interesting that Sider puts forward, as if it were an argument against hell, an argument that actually would have to be taken as a general problem for all sorts of moral attributions.
1 I find, by the way, that most people assume that universalists don't have a doctrine of hell. In the Christian tradition this is manifestly false; it's actually difficult to find universalists without a rather substantive doctrine of hell. Indeed, Christian universalists tend to have a doctrine of hell very similar to other Christians'. All they do is qualify it, e.g., by making it temporary (and thus blurring it into a sort of purgatory) or hypothetical (as what would be possible if God did not save us out of his superabundant mercy but gave us what we deserve).
2 In what follows I will be following Michael Almeida's exposition of the Impossibility Proof ("Is It Impossible to be Moral?" Dialogue: Canadian Philosophical Review 44 (2005) 3-13). The application to Sider, of course, is my own.