blog.kennypearce.net

Skeptical Theism and the 'Beforehand-Switch'

I return now from my hiatus to blog through the last three chapters of Sobel's Logic and Theism. There are two chapters on arguments against the existence of God, mostly focused on arguments from evil, and one on Pascalian wagers.

In chapter 11, section 4, Sobel presents what he takes to be Hume's evidential argument from evil, and discusses skeptical theist responses to it. Now, in general, the dialectic between the evidential arguer from evil and the skeptical theist goes something like this: the evidential arguer from evil says, a perfect being would probably create a world with very little (or no) evil, but this world has lots of evil, so it was probably not created by a perfect being. The skeptical theist responds, how should I (who am imperfect) know what a perfect being would do?

Now, according to Sobel, Hume's version of the evidential argument from evil includes an important maneuver which he calls the 'beforehand-switch.' The idea is that a hypothesis is best evaluated by considering what we would predict, using the hypothesis, if we didn't already know the outcome. So, for instance, we evaluate Newton's Law of Universal Gravitation by pretending we don't already know the orbits of the planets and using the law to predict them. Suppose, says Sobel's Hume, that we do the same for the hypothesis that the world was created by a perfect being. We pretend we don't know anything about what the world is like, and try to predict what world the perfect being will create. Everyone, it seems, should agree that we would predict that the world would have a lot less pain and suffering in it, and it is surely unreasonable of the skeptical theist to deny that this would be our prediction and that we would be justified in making this prediction.

In general I'm a little hesitant about skeptical theism. I worry that the skepticism will spread. However, it seems to me that the skeptical theist is actually in a very strong position here. The skeptical theist should say that, while we would indeed predict that the world would have much less pain and suffering in it, we would have very strong reasons for being extremely hesitant about our predictions. If we want to keep up the comparison to physics I introduced earlier, we can say that it is as though we have a physical theory which involves intractable mathematical problems, and have made liberal use of approximations and simplifying assumptions of questionable validity in order to come up with what can only be called a guess. The fact that the guess we make is the only reasonable one to make given our capacities and the information that we have does not make it any more than a guess. As a result, it should have little effect on our overall evaluation of theism.

To see why this is so, let's think about what we would actually do if we were trying to put ourselves in the 'beforehand' position and say what kind of world we think would be created. It seems to me that we would simply come up with the best sort of world we could think of. However, it ought to be obvious that the best sort of world that could be created by a perfect being would be radically different from the best sort of world I can think of! Here are just a few reasons, which are commonly cited in the skeptical theism literature:

1. I'm not logically omniscient. There are a priori necessities that I can't see, so I don't know what is necessarily involved in the things I'm putting into the world I'm imagining. The world I'm imagining might not actually be possible, or my explicit description might imply some really bad things that I haven't thought of.


2. A posteriori necessities. In addition to a priori necessities, it is now widely held that there are a posteriori necessities, such as the proposition that water is H₂O. If there are any of these at all, then there are lots of them that I don't know, so, again, my world may not really be possible, or might implicitly include evils I don't know about.


3. I don't know all the actual kinds of goods and evils. There are lots of things in the universe that I don't know about, which probably includes various goods and evils. Since I don't know about them, I don't know how good or evil they are, in comparison to the goods and evils I do know about, so I can't rank them. As a result, there may be evils of great magnitude implicitly included in my world, or goods of great magnitude left out of it.


4. I don't know all the merely possible kinds of goods and evils. If I were really going to make an informed judgment about what kind of world would be created, I would need to know not just all the actual kinds of goods and evils, but all the possible kinds. I don't.

The reason Sobel's reconstruction of Hume's argument falls to these skeptical considerations is that it asks us to consider the world as a whole, something that we are not well equipped to do. Here is a strategy for an evidential argument from evil that wouldn't have this problem: Evil is a widespread and undeniable phenomenon. It is very difficult to come up with an explanation for this phenomenon in terms of the theistic hypothesis. As a result, any competing hypothesis which easily and satisfactorily explains evil is, ceteris paribus, to be preferred to theism. This kind of argument comes to a more modest conclusion, but is, I think, more difficult to diffuse by raising skeptical worries. It is because of this sort of consideration that I think evil should be regarded as a genuine problem for, but not a decisive refutation of, theism.

(cross-posted at The Prosblogion)