If you ask a group of people a philosophical question or present them with a challenging thought experiment, it’s often the case that they will express very different views about it. We quite naturally interpret this as different people having different intuitions. But is that what’s going on?
In “Conflicting Intuitions,” published earlier this year, Joshua Knobe (Yale) takes up this question. Having imagined posing a question about free will to a group of students who express disagreement over its answer, he writes:
[I]t might be that some people have compatibilist intuitions, while others have incompatibilist intuitions, and similarly for each of the other traditional questions of philosophy. Then the puzzlement we feel regarding these questions might be explained in terms of the conflict between the intuitions of different people.
But that’s not the only possible explanation:
I will argue that experimental research has been providing evidence for a very different hypothesis. On this hypothesis, the difficulty is in large part a matter of individual people having conflicting intuitions. Thus, when a person encounters a thought experiment designed to get at one of the traditional questions of philosophy, she will often have an intuition pulling her in one direction but also an intuition pulling her in the exact opposite direction. This might seem like an obvious hypothesis, and perhaps it is. Still, I think that its implications haven’t been fully appreciated. If true, it would have important implications both empirically and philosophically.
On the empirical side:
The conflicting intuitions hypothesis has important implications when it comes to what exactly our theories about people’s intuition are supposed to explain…
If we assume that the difference in responses is due to different people having different intuitions, we would naturally assume that the theory should explain why different people have different intuitions regarding this case. So we might think that our theory of the underlying cognitive processes should explain why [some] have one intuition and the other[s] have the opposite intuition.
However, if we conclude that the difference in responses is due to each individual person having conflicting intuitions, there is an important sense in which there is more that needs explaining. An adequate theory would not merely have to explain why half of the participants have the intuition that it is a case of causation; it would have to explain why those people are conflicted…
There are various issues on the philosophical side. Here’s one, mentioned after the discussion of some empirical work on cases about which respondents say that apparently conflicting intuitions are both correct:
Independent of anything that could be settled through further empirical work on ordinary intuitions, we need to know whether there is some way of working out a philosophical theory that makes sense of what people seem to be saying in these cases. For example, within the philosophy of law, we can ask whether there is any way of developing a theory on which it can be right to say something like: there is a sense in which this action clearly violates the law, but there is a deeper sense in which it does not violate the law. Similarly, within the work on personal identity, we can ask whether there is any way of developing a theory on which it can be right to say: there is a sense in which this person is clearly Phineas, but there is also a deeper sense in which he is not truly Phineas.
Here again, I suspect we will arrive at different answers for different philosophical issues. For some, we may find that it is possible to develop such a theory, and philosophical work along these lines might lead to important new insights. For others, we will presumably arrive at the opposite conclusion. We will find that there is simply no possible way for both intuitions to be true.
Additionally, “we would need a clear answer to the… question as to how conflicting intuitions can have a larger philosophical significance even when we recognize that there is no possible way that both intuitions can be true.”
The whole article is here. Discussion welcome.
Glad to see attention to this issue, though I’ll note that back in 2013 (“Experimental Attacks on Intuitions and Answers”, PPR) I argued for no-conflicting-intuitions explanations of the data collected in x-phi studies.
As far as I can tell, that earlier paper’s argument against the straight intuitions hypothesis also tells against Josh’s conflicting-intuitions hypothesis. I welcome hearing if others think this assessment is mistaken.
Edit! “argued for *alternatives to* the no-conflicting-intuitions explanation”
One of my main concerns with your answer to x-phi studies is that it makes having an intuition a difficult thing to pick out in the stream of consciouness. Since it can be confused with inference, guesses, emotional reaction, and an array of other mental states in x-phi studies, why it cannot also be confused in ordinary philosophical investigation? I think that I’m intuitively justified to believe that in the gettier scenario the progatonist doesn’t have knowledge, but, actually, no, maybe I’m inferentially justified to believe that, or even not justified at all because it was only a guess. Putting in other terms, not only there is an underdetermination of the mental states by survey, but also underdetermination of the mental states by the difficulty of picking it out even in ordinary philosophicla investigation.
Edit: my question is the following: what makes ordinary philosophical investigation less prone to the same problem that you are pointing in the x-phi surveys?
Fair question, Vinicius. I consider the worry in §3 and address it via what I call the First and Second Observations, which I take to be on firm footing. The cases that support the basic psychological distinctions they observe are relatively humdrum and seem quite clear. I agree that it’s far less clear what the best theory is. I also agree that the distinctions at issue can go unheeded in practice for various reasons (e.g., some haven’t paused to notice; others haven’t appreciated their significance; still others may be hindered by familiar impediments) — that’s why §8 encourages greater sensitivity within philosophical practice, speaking to the question in your edit.
But this thread is about Josh’s interesting paper. Let me reiterate that I very much welcome its exploration of alternatives to the no-conflicting-intuitions orthodoxy. My worry isn’t simply that it rests on the problematic assumption that the studies tell us something about intuition. By fixating on contents, it misses the chance to consider plausible alternatives, including extant ones, that highlight other attitudes.
“Independent of anything that could be settled through further empirical work on ordinary intuitions, we need to know whether there is some way of working out a philosophical theory that makes sense of what people seem to be saying in these cases.[…] For some, we may find that it is possible to develop such a theory, and philosophical work along these lines might lead to important new insights. For others, we will presumably arrive at the opposite conclusion. We will find that there is simply no possible way for both intuitions to be true.”
A straightforward reading of this is a call for investigation into the application of reasoning models built on non-classical logics to traditional subjects in philosophy (right?). Knobe’s framing makes it sound like he is calling for suggesting something new to explore. But isn’t it just the field sometimes called “philosophical logic”?
I don’t think the call is for non-classical logics. I interpreted more as a call for non-classical theories of meaning, such that the words being used to express these judgments have multiple meanings for the people using them, such that one claim can be true with one meaning, and the other claim can be true with the other meaning, even though these claims would be contradictory if they were expressed with the same meaning. But I think it’s not so much about words, as about whatever people are using as constituents of their thoughts.
I guess I can see that reading of the call, but then I just know even less what to do with the call. Consider that, at least as far as I can tell, a theory of meaning interprets either sentences (utterances) or else propositions*.
Option 1, the sentences: in a large language corpus, there are techniques to identify where words are polysemous. But (to Knobe’s point, I guess) these statistical method treat each instance as independent (as opposed to treating each author as independent), so are silent about polysemy within anyone’s head. In principle these methods could be applied to a single individual to generate evidence of polysemy in a single head, but it seems experimentally infeasible to get that much data. Still, maybe the suggestion is just a speculation about the existence of some new statistical analysis for smaller datasets, which is nearly as capable. In any case, I don’t think this is a likely interpretation of Knobe’s call, as there is no indication that the philosophy he calls for is philosophy relevant to the foundations of statistical analysis.
Option 2, the propositions: if there is going to be any connection between a theory of polysemous propositions and sentence-level utterances, it will be dependent on specification of a reasoning model assumed to apply in the subject’s head, because we need a story of how the subject chooses the utterance. Unless classical logic is assumed when defining the background set of plausible reasoning models, there is going to be at least a serious modeling degeneracy between scenarios in which subjects’ utterances are explained in terms of non-classical reasoning and scenarios in which subjects’ utterances are explained in terms of polysemous propositions featuring in that reasoning. This degeneracy is at the level of empirical analysis; but I suspect there is a deeper issue here. It wouldn’t surprise me if there are very general conditions under which we are guaranteed to be able to simulate a non-classical reasoning model in a classical logic with polysemous propositions, and vice versa.
That last point may be an interesting line of new work. But it still strikes me as just more of an existing tradition of work in non-classical logic. Unless I’m missing something. (Or unless I’m overestimating what gets worked on in the world of research on non-classical logic.)
~~~~
*I know it is open as a matter of cognitive science whether the constituents of subjects’ thoughts are (necessarily) propositional. But I’m not sure how to make progress on a polysemous theory of meaning, except by assuming they are.
This is a startling hypothesis, because if one and the same individual can have conflicting intuitions about cases, then a philosophically adept questioner should be able to elicit those conflicting intuitions in a way that would leave the individual with no clear sense of how to resolve the conflict in their own thinking. One would expect that such an individual, faced with a heightened awareness of the conflict within their own thinking, would feel puzzled, confused, and frustrated.
But the history of philosophy shows us no cases in which adept questioners left their interlocutors in aporia. So, I think we should reject the hypothesis.
Mad Max – That’s Bait