According to the naive counterfactual analysis of dispositions (NCA), o is disposed to M when S if and only if, if it were the case that S, o would M. Unfortunately, NCA is too nice and simple to be true and counterexamples to both sides of the biconditional abound. These include (on the "if" side) finks (the device that would turn a dead wire into a live one if it were to be touched by a conductor) and masks (the carefully wrapped but nonetheless fragile Ming vase) and (on the "only if" side) mimicks (the golden chalice hated by a wizard who would destroy it, if something where to touch it).
As a result of these counterexamples, some have abandoned NCA in favour of some different analysis, others have tried to fix it. Both projects, however, have proved to be quite tricky. Nevertheless, I still hope NCA can be fixed (it's too nice to give it up). The idea I'm exploring right now is that there is a common theme to all counterexamples to NCA. In all of them something is interfering with o's disposition to M when S. So, to avoid the counterexamples NCA should be fixed by adding 'unless something interferes with o's disposition to M when S'. Now, of course, this cannot be the whole story unless we are also able to give an analysis of 'something interferes with o's disposition to M when S' without mentioning 'o's disposition to M when S' otherwise our analysis would simply be circular (and this is far from being an easy task but I'll leave my suggestion for doing so for future post).
Now, the problem is that, as far as I can see, this general strategy seems to be quite obvious and yet, to my knowledge, no one has tried to pursue it so far. So, am I missing something? Have there been any attempts to pursue this general strategy I don't know of? And, if not, is this due to the fact that there is something clearly wrong with it (or is just due to the difficulty of analyzing the concept of interference in non-dispositional terms)? (One thing that could seem to be wrong is that in the case of mimicks there would seem to be no disposition to interfere with (and that is exactly the problem). However, I think this problem can be dealt with by claiming that there is, in fact, a disposition that is being interefered with--i.e. the chalyce's sturdiness. And that if nothing was interfering with that disposition the chalice would not appear to be fragile.)
I think the problem with trying to come up with a conditional analysis that avoids finks, masks and antidotes consists in what you've pointed out - that analyzing the concept of 'interference' in non-dispositional terms ends up making things rather murky.
ReplyDeleteSome prominent examples of this are in so called 'hedged conditional' accounts, where a CP clause is embedded into the dispositional conditional itself - such as Mumford's 'conditional conditionals'; Other authors who have attempted to deal with CP conditions within dispositional conditionals are Bird and Ellis.
If these kinds of accounts are flawed, and it seems that they are (Schrenk complains that no matter what happens, "CP clauses shoot up like mushrooms"), I do not see how a hedged conditional of the form 'unless the disposition is not interfered with' could do us any good. Most likely, this kind of account would face the familiar CP worries as well.
But I am very interested to hear what you propose.
"So, to avoid the counterexamples NCA should be fixed by adding 'unless something interferes with o's disposition to M when S'."
ReplyDeleteThere seem to be two quick problems. (1) It might be that there is an e that interferes with o's disposition to M when S, but also an e' that instantaneously restores o's disposition to M when S. (2) it might be true that there is no e that interferes with o's disposition to M when S, but that there would be such an e were S to occur (I think that's the typical assumption anyway). But I think this might result in too many dispositions. Suppose I claim that my car is disposed not to start because it is true that, were I to turn the key & some e occurs that prevents the starter from turning over, my car would not start. Because I add special conditions in the antecedent, I'm no longer describing a disposition of the car. Or, at least, I'm not describing a disposition not to start, but some other disposition. You might be doing something similar (i.e., describing the disposition to M when there are no interfering e’s, for instance, as opposed to the disposition to M).
Christopher,
ReplyDeleteI think the problem with trying to come up with a conditional analysis that avoids finks, masks and antidotes consists in what you've pointed out - that analyzing the concept of 'interference' in non-dispositional terms ends up making things rather murky.
Well, we'll see what you think of my proposal, which certainly is far from simple but I hope it doesn't make things murky.
Most likely, this kind of account would face the familiar CP worries as well
I don't think it does. The problem with CP accounts is that the content of CP clauses cannot be clearly specified. I intend to give a clear (if quite cumbersome) definition of interference. The question is whether it will fall prey to the kinds of worries Mike mentions in his comment.
Mike,
I don't see why (1) would be a problem. The definition says that if o were to M if it were the case that S o is disposed to M when S. Since the unless is meant to work as an inclusive 'or', in the case you have discussed the object would still come out as having the disposition. Don't you think so?
(2) it might be true that there is no e that interferes with o's disposition to M when S, but that there would be such an e were S to occur
It's true that the interference would usually swing into action only if S were to occur, but the interference is usually there before S occurs. There is typically someone (e.g. the wizard) or something (e.g. the fink) waiting in the wings and it seems that if we were able to counterfactually remove them (and all other interferences), the disposition would manifest itself if the stimulus condition were to obtain. (The question it's going to be whether my definition of interference can remove all and only such interfering factors)