## Friday, July 31, 2009

### Dispositions and Interferences (Part II)

In Part I of this post, I suggested that the simple counterfactual analysis of disposition (SCA) may be saved from the usual counterexamples by introducing clauses to the effect that nothing interferes with o's disposition to M (or not M) when S.

More specifically, the "intereference free" counterfactual analysis (IFCA) would maintain that:

(IFCA) o is disposed to M when S iff:
1. (If it were the case that S, o would M AND it is not the case that something interferes with o's not being disposed to M when S) OR
2. Something interferes with o's being disposed to M when S.

As I noted, this analysis would be circular unless one were able to provide an analysis of 'x interferes with o's disposition (not) to M when S' without employing the notion of disposition.
This is my a first stab at doing so. (be warned that it's more than a bit convoluted)

(Interference)
--> For all ks, Ik interferes with o's being disposed to M when S iff:
1. It is the case that I1 and … and Ik and … and In,
2. For each j, if it were the case that not (I1 or … or I(j–1) or I(j+1) or … or  In), then it would not be the case that, if it were that S, then o would M.
3. There is some property G such that o has G and if it were the case that not-(I1 or … or In), then it would be the case that: (3.1.) if it were the case that S and o retained G, o would M, and (3.2.) it is not the case that, if it were the case that not-S, then it would be the case that M and (3.3.) it is not the case that, if it were the case that S and O did not retain G, then o would M.
4. There is no property H such that it is not the case that o has H, and, if it were the case that not-(I1 or … or In), then o would have H and, if o didn’t have H, then it would not be the case that, if it were that S, o would M.

As far as I can see, this can deal with all the usual counterexamples to (SCA). For example, there being an (inverse) fink attached to this live wire comes out as interfering with the wire's disposition to conduct electricity when touched by a conductor (had the fink not been there, the wire would have conducted electricity when touched by a conductor) and there being a chalice-hating wizard interferes with the chalice's disposition not to break when touched (because had there been no wizard, the chalice would not have broken when touched).

(Question A) Am I wrong in thinking that IFCA avoids the standard counterexamples to SCA?
(Question B) Can anyone think of any new counterexamples lurking in the background? (My spidey senses tell me that there is a whole battery of them just waiting to be thought of... :-))

One last thing: I am assuming that properties are sparse. So, in (IFCA 4.), H cannot be something along the lines of being such that no chalice-hating wizard is around or the likes, for I take there is no such property to be had. However, H can be something along the lines of being made of glass (So that the fact that, for example, the live wire is not made of glass does not come out as interfering with its disposition to conduct electricity when touched by a conductor).

1. Perhaps I don't understand the conditions, but it seems like you've only ruled out interference from some fixed finite set I_1,... I_n. If so, then some other thing could interfere.

2. You are right, Tim. It would be better to say I_1, I_2, ...

3. The difficulty now is going to be coming up with a complete, exhaustive list of interfering conditions.

Given how interfering conditions interact (defeaters can themselves be defeated) it's not clear that you can even specify *one* interfering condition in a single sentence, without resorting to ceteris paribus clause. And there may be sizeable infinities of such interfering conditions. Ultimately, a complete, exhaustive list of interfering conditions is a rather ugly class of possible worlds.

You also need to avoid "grue"-like dispositions. If any ugly class of possible worlds can be thought of as a list of interfering conditions, then everything will have lots of weird dispositions. I guess you hope that sparseness will let you get round this problem, but I'm not sure how it will help.

4. Hi Tim,

As far as I can see, the list doesn't need to be an exhaustive list of all possible interfering conditions--it only needs to be a list of all actual interfering conditions. In other words, you describe me a situation that is a counterexample to SCA and I'll show you how it is no counterexample to ICFA because something in your description would count as interfering with the disposition in question.

(Btw, I'm now dissatisfied with the way I was trying to deal with mimicks in the post. I now believe that the second conjunct of (IFCA 1) should be 'it is not the case that something interferes with o's not being disposed to M when S')

5. In addition to my last remark, I should mention that 'something interferes with o's not being disposed to M when S' needs a completely different analysis from 'something interferes with o's being disposed to M when S', but I'll leave that to the side for the moment.

6. I've two comments about counterfactual interference; for convenience, I'll do two separate posts. First, consider two cases:

(1) Glass 1 is *relatively* fragile. It is dropped to the floor and it smashes. But if some minimal cushioning had been put on the ground, Glass 1 wouldn't have smashed.

(2) Glass 2 *very* fragile glass. It is dropped to the floor and it smashes. If some minimal cushioning had been put on the ground, Glass 2 still would have smashed.

If you want to say that these two glasses have (somewhat) different dispositions, you need to talk about counterfactual interference conditions. Perhaps you're not thinking about these cases, since they aren't counterexamples to any account of dispositions. But an account of dispositions does need some way to explain why the dispositions are different.

7. Here's my second comment on counterfactual interference. Your third condition says something like:

"There is some property G such that o has G and if it were the case that not-[big conjunction of interference conditions], then it would be the case that..."

If you only list actual interference conditions in there, your analysis will give the wrong answer. For in possible world W, perhaps none of the conditions which frustrate the disposition in the actual world @ occur. But nonetheless, in W, some *other* conditions interfere with the disposition.

8. Tim @6:03AM,

I don't think this has much to do with interferences. It has to do with the fact that some dispositions are gradable (if the two glasses were equally fragile why would the cushion interfere with the fragility of one but not the other). Of course, this means that one would need an account of what is for Glass 2 to be more fragile than Glass 1, but this may not be particularly hard to achieve--one could claim that Glass 2 is (strictly) more fragile than Glass 1 iff there are circumstances in which Glass 2 would break but Glass 1 would not but there are no circumstances in which Glass 1 would break but GLass 2 would not.

9. Tim @6:05am,

I don't think this is right. On the standard Lewis-Stalnaker semantics for counterfactuals, the closest possible worlds at which the antecedent is true are ones that are as similar as possible to the actual world, so, in general, I don't think there is space for some other interfering conditions to arise in the absence of the actual ones.

Moreover, (Interference 2) requires that in the absence of the actual interfering conditions, the counterfactual 'if it were that S, o would M' would hold.

10. Gabriele @1:00pm. Right, but here you talked about non-actual breaking conditions. And to talk fully about non-actual breaking conditions, you need to talk about non-actual interfering conditions not obtaining. (BTW, and it's probably not relevant: I doubt fragility is a linear-ordering; two things can be fragile in different ways.)

I suspect you'll answer this worry just as you answered me @6:05pm, so:

Gabriele @1:10pm. I guess this puts us firmly into the territory of what you mean by "closest possible world". Could you give me a general reason to think that the closest possible world in which the actual interfering conditions do not obtain is one in which no other interfering conditions obtain?

Invoking (Interference 2) just highlights my concern, I think. My worry is precisely that something can have a disposition to break even though, in the absence of actual interfering conditions, "if it were that S, o would M" is false (because in all the "closest" worlds, something non-actual interferes).

11. Hi Tim,

I suspect you'll answer this worry just as you answered me @6:05pm

Right!

Could you give me a general reason to think that the closest possible world in which the actual interfering conditions do not obtain is one in which no other interfering conditions obtain?

The general reason is that according to (Interference 3) "[...]if it were the case that not-(I1 and … and In), then it would be the case that, if it were the case that S [...], o would M" cannot hold if the closest possible world at which it is not the case that (I1 and … and In) is one at which something else interferes with o's disposition to M when S because, if something else did interfere, it would not be the case that if it were the case that S, o would M. In other words, the definition catches all "active" interfering conditions as well as all interfering conditions that are "waiting in the wings".

Invoking (Interference 2) just highlights my concern

Ooops, sorry! That was a typo--I meant (Interference 3).

12. I agree completely with the claim: "If something else interfered with o's disposition to M, it would not be the case that, if it were the case that S, o would M."

Nevertheless, I worry that the following situation might obtain:

(A) o has a disposition to M; and
(B) were it the case that S and not-(I_1...), something else, J, would prevent o from M-ing, where (I_1...) contains all actual states of affairs interfering with o's disposition, and J is not an actual state of affairs.

This worry is ruled out if we treat your conditions (1)-(4) as *defining* what it is to have a disposition. But, precisely because I do have this worry, I doubt that your conditions *capture* the prior notion (if there is one) of what it is to have a disposition.

So, what I'm after is a reason (independent from the stipulative conditions (1)-(4)) that explains why I should not think that (A) and (B) are conjointly possible.

13. Hi Tim,

Okay, let's see. Let's assume that
(C) it is not the case that J and
(D) if it were the case that J, then it would not be the case that, if it were the case that S, o would M.

Now, you are claiming that it could be the case that if it is the case that 'If it were the case that that S and not-(I_1...), it would be the case that J'. But in virtue of what would this counterfactual be true?

As far as I can see, there are three classes of cases:

(i) There is some actual X, such that, if it were the case that that S and/or not-(I_1...), X would bring about J.
(ii) Nothing would bring about J. J would obtain simply as a result of S's obtaining.
(iii) Nothing would bring about J. J would obtain simply as a result of the actual interferences' not obtaining.

In (i), X itself would count as an interference.
In (ii), it is not clear to me that o is disposed to S when M. As the obtaining of S itself would prevent o from Ming.
I find the case (iii) hard to imagine concretely (but this may be due to lack of imagination). We have assumed that there is no X that would bring about J in case the actual interferences did not obtain. So, J obtains simply in virtue of the non-obtaining of the actual interfering conditions, but I don't see how this can be the case unless 'not-(I_1...)' simply entails 'J'. This could be simply because 'J' is the negation of one the Is or of a conjunction of them, in which case, the disposition would always be interfered with. But, if this is the case, I'm not sure if I would be inclined to believe that o is disposed to S when M.

It would be really helpful if you could come up with some example that does not fall under (i) but that is such that the obtaining of the J is not just entailed by the non-obtaining of the Is.

14. Mmmm those three kinds of case are very helpfus.

Think I agree with your analysis of case (ii).

Not sure what I think about case (iii).

I suppose I had in mind something close to instances of case (i). My thought is that the plausible way to remove each actual I_n might be to invoke X as a destroyer of the I_ns, which brings about J. But why does X need to be actual (as in your case (i))?

15. Think I agree with your analysis of case (ii).

Good ;-)

Not sure what I think about case (iii).

As I said, me neither, but that's why they are the only ones I am somewhat worried about.

My thought is that the plausible way to remove each actual I_n might be to invoke X as a destroyer of the I_ns, which brings about J. But why does X need to be actual (as in your case (i))?

Well, most people seem to think that a miracle should bring about not-(I_1, ...). (I am on record arguing against that but, in my view, if X were not actual, you would still need something to bring about X (well, in fact I'm also on record arguing something quite close to that), but discussing this point would lead us far astray.)

16. What is the motivation for going through all of the admitted trouble to get around interference or masking?

At the moment I am persuaded by "Dispositions and Habituals" on www.princeton.edy/~fara/research.html which is recapped in the beginning of the most recent paper posted.

17. One of the main motivations is that there seem to be more than a grain of truth in the conditional analysis of dispositions.

This is not the place to offer a proper criticism of Fara's account, but as far as I can see it has a lot of problems (some of which Fara seems to candidly admit). One of them is that it does not avoid mimicks. Another is that it is not clear if it deals satisfactorily with unmanifested dispositions.

18. Not sure this will ever be read, sorry Gabriele but I didn't follow this thread. I just wanted to mention Sungho Choi's recent work, which appears to go exactly in the direction you're exploring. Probably you are already aware of Choi's papers, but I would be curious to know what your opinion is, and in what way you think your suggestions differ from his.

19. Yes, it was read ;-) As far as I know, Choi seems to think that some of the things I call interferences can be avoided by specifying the stimulus condition, the problem is that one should so so for every possilbe interference and I don't think one can do so. Moreover, I think that stimulus condition are just okay the way they are. As far as I can see, the stimulus condition of this pill's poisonousness is its being ingested not its being ingested without having ingested an antidote. But maybe I'm not familiar with the papers by Choi you are thinking about? I'd be curious to know which papers you are thinking of and to hear more about what you think the similarities between Choi's proposal and mine are.

20. In his “Dispositional Properties and Counterfactual Conditionals” (Mind 117: 795-841, 2008), Choi seems to avoid your worry about stimuli by specifying that dispositions are manifested in response to ‘being situated in a stimulating circumstance’ rather than to a stimulus. This allows him to say (e.g., fn. 3), that the stimulus can be taken in its canonical form, and the absence of interfering factors considered part of an appropriate stimulating circumstance (of which the stimulus is one component).
Now, of course there are differences in the details and the formulation, but if this neutralizes your worry, then the clauses in your IFCA seems to be doing the same work as the specified stimulating circumstances plus direct process in Choi’s version of SCA, but in a perhaps more convoluted way (and also invoking ‘negative’ dispositions such as sturdiness?). In other words, if the problem with Choi’s proposal is just that he needs to specify the stimulus but this goes against our intuitions about what a stimulus is, he seems to have a response. Otherwise, what is the problem? You claim that Choi has to specify all the possible interferences, but I am not sure I see why he has to do this while you don’t. If anything, you seem to feel the need to define interferences explicitly while he doesn’t.
Incidentally, Choi (same paper, and “The Conditional Analysis of Dispositions and the Intrinsic Dispositions Thesis” (Philosophy and Phenomenological Research 78: 563-590, 2009)) spends some time explaining why he believes that there is a genuine difference between, say, a disposition to break when struck and a disposition to break when struck in the absence of a glass-loving sorcerer (this has to do with the exact definition of nomically intrinsic dispositions).
I am sure there is something I am missing, but maybe it would be helpful if you could spell out what you take to be new in your account a bit more?

21. Hi Matteo,

Thanks for the pointer--that's definitely a paper I should not have missed. (For some reason it doesn't come up on philpapers (???)). From what you say, however, Choi does not seem to be doing something that different from saying 'absent all interfering conditions' and I take it that the standard worry about 'ceteris paribus'-style approaches--i.e. that they are vacuous or circular--seems to apply to his proposal if he does not have an account of what it takes for something to interfere with a dispositions.

Btw, I don't think I need sturdiness any longer, although I do not have any problem with "negative" (I would say continuously manifested) dispositions. (Btw, why does sturdiness worry you? I think it's a perfectly respectable disposition and I don't see any principled reason to consider it a "negative" disposition. The fact that we tend to describe it by using a negative manifestation (a disposition not to break) does not mean that it is in any real sense negative. Every object that is not in the process of breaking seem to manifest a disposition not to break or to keep its parts together if you prefer to avoid a negative characterization of the condition of manifestation).

22. Hi Gabriele, happy to have been helpful!
I couldn't say whether or not he's entirely successful, but in the Mind paper Choi surely spends a lot of time addressing all the usual problems with SCA, including CP clauses, circularity etc. You'll judge for yourself...
As for 'negative' dispositions, I agree with you that they are respectable properties, I just meant that I am not sure I see why you need the 'o's disposition not to M when S' addition to your 1.

23. I still need that to avoid cases of mimicking such as that of the chalice-hating sorcerer. In those cases we want to say that the chalice is not disposed to break when struck but something interferes with its not being so disposed.

24. In any case, I hope to have a presentable draft of the paper ready to post soon and hopefully things will be clearer. (I have changed my mind on a some details since writing these posts).