Get out the vote

Following up on Marcus' comment about appeals to intuition and commonsense, I'd like to ask you, fellow pupae, what you think about the following complaint against experimental philosophy.

Critics of experimental philosophy often say that philosophical disputes cannot be resolved by voting. (I just had a referee make this complaint about a paper of mine.) What do you make of this charge?

I've always found it rather strange for two reasons. First, to the best of my knowledge, no experimental philosopher has argued as follows: In response to thought experiment E, most people say that p; therefore, p.

Second, isn't it the case that appeals to intuition (i.e., "it seems to me that p; therefore, p," as in "it seems to me that it is morally wrong to push the fat man off the bridge; therefore, it is morally wrong to push the fat man off the bridge") are a form of voting? It's just that only one philosopher gets to vote and only his vote counts.

Comments

[Clayton]

Hi Moti,

On the second point, it isn't clear to me that appeals to intuition should be understood on the voting model. It seems that votes can establish something independent from them, but that's where the analogy breaks down. It isn't obvious that when something is intuitive, there's a seeming or intuition that is offered as some sort of independent reason to accept p. I don't _infer_ that it is wrong to push the man of the bridge from the premise that it seems to me that it's wrong. Instead, the wrongness of pushing is intuitively evident and my belief that it is wrong is justified by the (apparent) fact that is intuitively evident. We might model intuitive justification on one model we have for observational knowledge--both perception and intuition can make facts evident and justify beliefs without supplying them with evidence beyond the fact (or apparent) fact one believes.

I'm sure there are other ways of construing appeals to intuition beyond this, but this seems like one way to understand appeals to intuition.

[Brad Cokelet]

Here is a go at making sense of the voting objection:

It takes some work to go from a first take on a case to a considered judgment about what it true of it, and the later is presumably the kind of thing that could count as an intuition in the construction of a philosophic theory (a point Rawlsians often make in discussing reflective equilibrium).

Experimental philosophy is notorious for posing questions in quick, dirty, and isolated judgment contexts so what they yield are not really peoples considered judgments about carefully and clearly considered cases, but off the cuff judgments of (often) indeterminately or confusedly understood cases.

Moreover, the quick and dirty judgments are often solicited from people who do not have the habit of (or skills to) self-critically form reflective judgments. Of course philosophers are not all absolute masters in doing this, but I take it that we each aspire to do the best we can, and we have more practice and skill at it than some.

So trying to settle a philosophic debate by appeal to the results of an experimental study of "intuitions" that fit the above characterization, is like trying to figure out who is the most fit to represent the US on the world stage by taking a vote of the uninformed public walking out of the grocery store on questions like "Would Obama do a good job representing the US in the middle east". I don't think there is a lot of wisdom in those kinds of crowds myself.

[Brad Cokelet]

I should say, of course, that not all the work that sails under the flag of experimental philosophy can be convicted in the way I suggest...even if my complaint has some merit

[Moti Mizrahi]

Hi Clayton,

I was trying to be sarcastic in making the point about appeals to intuition and voting. But I do think that appeals to intuition involve some sort of inference. If we agree that intuitive judgments in response to thought experiments are seemings (i.e., “in case C, it seems to S that p”), then I think we should also agree that there must be something that connects the seeming that p with p (the two are not the same). I don’t see what else could connect them other than inference.

To put it another way, upon considering case C, S makes the intuitive judgment p. The fact that S responds to C by judging that p is a fact about S, not C. But we are not interested in facts about S. We want to know what is true about C. To get from facts about S (e.g., “it seems to S that p”) to facts about C (e.g., “pushing the fat man is wrong"), we need inference.

[Moti Mizrahi]

Thanks for your comment, Brad.

Here are a few worries:

1. Intuitions—whatever they are—are supposed to be immediate, non-inferential reactions to cases. That is, in response to case C, I immediately judge that p. (Of course, one then needs inference to get from one’s reaction to C to the conclusion that the reaction is true of C. See my comment in response to Clayton.) Now, it seems to me that if we are talking about considered judgments or reflective judgments, we are no longer talking about intuitive judgments, and hence we are no longer talking about appeals to intuition. You may be right in pointing out that philosophers consider more carefully the assumptions that are involved in appealing to intuitions. But that doesn’t mean that their appeals to intuition are any better than those of “the folk.” (This raises the question of philosophical expertise, which is a difficult one.)

2. I actually think that even philosophers are not sufficiently critical about their appeals to intuitions. Take any example of a thought experiment that is supposed to elicit a certain intuitive judgment. Usually, when philosophers use this method (AKA, the method of cases), they take the intuitive judgment elicited by the thought experiment as a given and do not consider the implicit assumptions involved; for example, that intuitions are reliable truth-trackers, that the case is not misleading, etc.

3. To the best of my knowledge, no experimental philosopher has argued that the philosophical debate between proponents of p and proponents of not-p is a closed case because most people say that p. The arguments made by experimental philosophers are usually much more sophisticated than that.

[Brad Cokelet]

Thanks for the response. I'm not saying I am 100% behind this objection, but here are some follow ups..

re (1): That makes sense, but can't the considered in 'considered judgment' involve simply being very careful to clarify the kind of specific case you have in mind and all its relevant aspects. This might involve thinking about other cases and how they differ. That sort of thinking does not, I take it, have to undermine the idea that the judgment (or preceptual-type intuition) we make about the clarified case is non-inferential. The idea, then, would be that philosophers are a lot better at thinking about the various specific different cases, language we can use to distinguish them, what differentiates them, etc. I agree this is contentious, but I suspect that some assumption about philosophical expertise of his sort is behind the objections, and many may feel that the onus is on experimental philosophers to show otherwise (I think is something like what T. Williamson argues in a recent paper in metaphilosophy).

re(2): Yeah, I don't think they are worried about these general methodological issues, but I do think good philosophers are worried about forming a theory on a one-sided diet of examples. And there are lots of cases of people being nervous about leaving a debate hinging on a difference of intuitions and a recognition that independent argument is needed (e.g. Smith vs. Bink on judgment internalism)

re(3): fair enough. but don't some x-philosophers argue as follows: you believe that your intuition that p about case c supports the conclusion that p (etc), but lots of people don't share your intuition, and that casts doubt on the support yours gives the conclusion. And this sort of argument only works if the "intuitions" being reported come from as good sources as the one being contested...and that is what it what I am saying proponents of the voter argument probably doubt.

Maybe the idea is that in politics we are ok with the votes of the less than best informed having an impact because we care about legitimacy, not just getting the best result; but in philosophy we just care about getting the best result.

In any case, I can see why this sort of objection is annoying to get.

[Clayton]

Hi Moti,

There are lots and lots of issues here, so there's a better than fair chance that I'll get in a tangle.

You wrote:
"If we agree that intuitive judgments in response to thought experiments are seemings (i.e., “in case C, it seems to S that p”), then I think we should also agree that there must be something that connects the seeming that p with p (the two are not the same). I don’t see what else could connect them other than inference."

Are the two not the same? There's a sense in which seeing that p and p aren't the same, but the latter is included in the former. Something connects p to the seeing that p, but it isn't inference. Why can't we say the same with intuitive knowledge? Like observational knowledge, it's supposed (by some) to be immediate or non-inferential. There's some story about connection that explains how observational belief connects to observed fact that doesn't involve inference. Let's just steal that for intuition.

You also wrote:
"To put it another way, upon considering case C, S makes the intuitive judgment p. The fact that S responds to C by judging that p is a fact about S, not C. But we are not interested in facts about S. We want to know what is true about C. To get from facts about S (e.g., “it seems to S that p”) to facts about C (e.g., “pushing the fat man is wrong"), we need inference."

So, that's a way to look at it and I'm not convinced that it's the wrong way, but I'm also not convinced that the alternatives are wrong. Consider the case with vision. Upon opening the window, the array of things below become visible to you. The fact that you notice the birds on the fence is a fact about you, but it is a relational fact. It is also a fact about the birds. Now, with logic. Upon considering the proof, the relation between premise and conclusion is one you can see. The fact that you notice the relation between them is a fact about you (in some sense), but it is a relational fact. It is also a fact about the relations between premise and conclusion. Now, with intuition. Upon considering a case, the fact that this is an instance of F-ness is evident to you. The fact that you notice that F-ness is present is a fact about you (in some sense), but it is a relational fact.

Alright, so we can run a sort of argument by analogy. I suspect that the X-Philes reject this picture. Why? Careful empirical research? No, of course not, it's a combination of ideology, prior commitment, the failure to take seriously the pioneering work of Cook Wilson. (That's sort of tongue in cheek.) To the extent that I have a dog in this fight, it's this: whether intuition, perception, etc. serves as a source of non-inferential knowledge shouldn't be separated from the question as to whether these serve as a source of reasons/evidence.

[Moti Mizrahi]

Thanks for the follow-up comments, Clayton.

First, I’d like to say that I am not speaking on behalf of experimental philosophers.

Second, I actually think that you are right about the analogy between intuition and perception. So I agree that perceptual seemings and intellectual seemings are both immediate and non-inferential. I think we disagree about how reliable truth-trackers seemings, whether sensory or intellectual, are. So, you say that when S has a perceptual seeming that F, that is a relational fact about S and F. But it is only a relational *fact* if S is really seeing F. Similarly, when S has an intellectual seeming that C is a case of F, that is a relational *fact* about C and F provided that C is really a case of F. In other words, from “It seems to S that p” it necessarily follows that “S has an intellectual seeming that p,” but it doesn’t necessarily follow that p. S could be wrong about what s/he intuitively judges. We need independent reasons to think that intellectual seemings are reliable truth-trackers.

Finally, to return to the voting objection, I’d like to point out that, if there are alternative ways of construing appeals to intuition (as we have discussed), and it’s not clear which way is better, then critics of experimental philosophy don’t get to use the voting objection as a decisive objection against experimental work. They get to use it only if experimental work *presupposes* a view of appeals to intuition that is inferior to others. Since that is hardly clear, the voting objection loses its force.

[Clayton]

Hi Moti,

"First, I’d like to say that I am not speaking on behalf of experimental philosophers."

That's alright, I don't think I'm speaking on behalf of myself. Wish I had a view to defend here, but it's such a tricky area.

On the intuition/perception analogy, I think I'd want to say this. There are lots of views of perceptual knowledge and I cannot think of any good reason not to try to model an account of intuitive knowledge on any of the standard accounts of perceptual knowledge. In perception, I'm pretty sympathetic to disjunctivist views. I'll register some possible disagreements with your second point. (They reflect some of my views about perception, which I suspect differ a bit from standard views...)

* "So I agree that perceptual seemings and intellectual seemings are both immediate and non-inferential."

While I'd want to allow that the beliefs we form in the relevant cases are non-inferentially justified, I'm not all that happy to talk about perceptual and intellectual seemings. In the perceptual case, I'd want to say this: when it looks to someone as if p, in the good case, p holds and it is evident to the perceiver that it does. I don't know if I'd want to say that there's any mental intermediary common to success and failure that people have in mind when they talk of 'seemings'. Does this matter? It might, because you wrote:

* you say that when S has a perceptual seeming that F, that is a relational fact about S and F. But it is only a relational *fact* if S is really seeing F. Similarly, when S has an intellectual seeming that C is a case of F, that is a relational *fact* about C and F provided that C is really a case of F. In other words, from “It seems to S that p” it necessarily follows that “S has an intellectual seeming that p,” but it doesn’t necessarily follow that p. S could be wrong about what s/he intuitively judges.

While I agree that it can seem to you that p even if ~p, I don't want to sign up for the view that the reason you have to believe p can only be the fact that it seems to you that p (or, the seeming, if you like). In the good case, the fact that p itself is the basis you have for your belief. Chisholm once defended a version of foundationalism on which properly basic beliefs are beliefs based on the very fact that the belief concerns. He thought that properly basic beliefs had to concern your own mental life, but he never offered any remotely plausible argument for that restriction. The plausibility of that view seems to derive from a kind of fetishism people have about introspection. Both introspection and perception are fallible capacities, but when exercised properly, they put us in touch with the facts directly. When that happens, I don't think we do need independent reason to take the apparent fact to be a fact.

I guess what I'd want to see is an argument that when certain facts (that don't have to do with our reactions or take on the facts) obtain, _they_ can't be what appear to us to hold. Without that, I don't think I'd want to be pushed into accepting the view that we don't have the capacity to form beliefs directly in response to certain kinds of facts. In the case of introspection, we can form beliefs about our own mental lives without having those beliefs based on an appearing or seeming that represents some further aspect of our mental life. Introspection is nevertheless fallible. I'd say the same for perception and possibly for intuition. [For independent reasons, I'd worry that the resistance to this view is fueled by a sort of antecedent skepticism.]

[Moti Mizrahi]

Hi Clayton,

You write: “I guess what I'd want to see is an argument that when certain facts (that don't have to do with our reactions or take on the facts) obtain, _they_ can't be what appear to us to hold. Without that, I don't think I'd want to be pushed into accepting the view that we don't have the capacity to form beliefs directly in response to certain kinds of facts.”

So here’s a first stab at such an argument:

1. The fact that p can be the basis for S’s belief that p only if S knows that s/he is in a good case.
2. But S cannot know whether s/he is in a “good case” or not.
3. (Therefore) The fact that p cannot be the basis for S’s belief that p.

In other words, when we say that the fact that p is the basis for S’s belief that p, we assume that S cannot be wrong about p. But, as you seem to admit, it can seem to S that p even if not-p. So, the only way I see to get from “it seems to S that p” to p is through the implicit assumption that S is in a good case, and so S cannot be wrong about p. But that assumption seems unwarranted to me, not for global skeptical reasons, but for more ordinary reasons, such as framing effects, and other kinds of biases.

This argument bears on the method of cases (i.e., appeals to intuition) directly because intuition pumps (i.e., the sort of thought experiments that elicit the sort of intuitions that then get used as premises in philosophical arguments) are not “good cases.” Rather, they are rather bizarre and very misleading cases.

[Clayton]

Hi Moti,

This has been both fun and helpful (for me, at any rate). If I had some time, I'd try to write something up on this. In response to the argument on offer, I'd question (1) and (2).

If 'good case' is understood as the case in which you have knowledge (in this case, non-inferential intuitive knowledge), then I'd say that if S knows non-inferentially that p, S's belief about p is based on the fact that p. If we reject KK, then I'd want to reject (1) on the grounds that you don't have to know that you know to have the relevant sort of knowledge and that having the relevant sort of knowledge involves having a belief formed directly in response to the fact believed to hold true.

As for (2), I think there are probably plenty of cases where we do know that we're in the good case--in the case where we judge that it's permissible to turn the trolley away from the 5 knowing that it would travel down a track and kill 1, we don't just know that this is permissible, we can easily know that we know this.

What I'd reject is the idea that your belief about p can only be based on the fact that p if you cannot be wrong about p. Just as introspection might be fallible and you still can (sometimes) base your introspective judgments directly on facts about your own mental life, I wouldn't argue from the fact that I have a fallible recognitional capacity that when I exercise that capacity my judgments about Fness are based on something other than the fact that the relevant x is an F. It seems to me that there are lots of cases like this, cases where a fallible capacity is exercised in such a way as to produce non-inferential knowledge. Maybe knowledge of the position of your own limbs is like this. The judgment that legs are crossed as I type this is one that is produced by a fallible capacity for determining the position of my limbs but isn't based on something other than facts about limb position.