Thursday, April 20, 2006

Contrastive Questions

Consider the following questions:

Q1. Why does anything at all exist?
Q2. Why does anything at all exist (rather than nothing existing at all)?

In the previous post I gave an argument for the claim that these questions are not equivalent, despite their verbal similarity -- cases of overdetermination. Discussing the matter with Richard in the comments, it occurred to me that there's a related argument for this claim. The argument is this. Q1 is about a member of a class; Q2 is about the class.

It's clear that Q2 is about the class 'anything at all exists', because it is opposed to a contrast class, 'nothing existing at all'. We see this with other questions:

P1. Why did the window break?
P2'. Why did the window break (rather than not break)?

To answer P1, all we have to do is explain this particular positive case: the window actually broke. Our answer to P2', however, has to mark off all relevant positive cases (all cases in which the window would have broken) from the contradictory cases (all cases in which the window would not have broken). (I add the word 'relevant' because usually we don't hold ourselves to a fully rigorous standard in answering questions like P2'. If our answer to P2' fails to give us the information required to mark of the positive class from its contradictory, our answer fails to give us the information needed in order to say why the window broke rather than not (it would just give us information needed in order to say why the window broke).

If questions of form 2 (Why P rather than not P?) typically call for an answer not merely about distinguishing one class from another; and questions of form 1 (Why P? understood noncontrastively) do not typically require such an answer, the questions are not equivalent. We do know, in fact, that this is usually the case with contrastive questions. If I ask "Why P rather than Q?" I am asking for something that distinguishes P-cases, one of which obtains, from Q-cases, none of which do. If I ask "Why P?" noncontrastively, I am simply asking for the explanation of a particular P-case. So why would one deny that questions of form 2 are simply a specific form falling under the more general "Why P rather than Q?"

This argument is related to the overdetermination argument in that overdetermination is the easist place to see that one question requires a class-relevant answer and the other requires a case-relevant answer. But the argument is slightly different in that it shows more clearly the mistake of those who want to conflate Q1 and Q2. They assume that Q2 means:

X: What is the explanation of this: P and not not-P?

When in reality Q2 means:

Y: What is the explanation of this: (any) case of P obtaining rather than (any) case of not-P?

X and Y are not at all the same question; and this is true even if we are dealing with classes involving only one case each. In X, P is not a class; in Y it is.