Our books

Become a Fan

« Chalmers on Philosophical (Non-)Progress | Main | "Peer Review" »



Feed You can follow this conversation by subscribing to the comment feed for this post.


I am utterly baffled by Dennett's historical examples, in particular the way he seems to think that they are *either* "logically sound arguments" *or* "stories" intended to "structure our thinking". Why is this a forced choice? Isn't every argument intended to structure our thinking and provoke imaginative reflection?

While his references to Plato and Kant are totally obscure, the passage on Hobbes shows that this 'distinction' is illusory. Hobbes' argument, like Rawls', proceeds from a hypothetical situation to the formation of a kind of contract. The conclusion is that it is rational to choose the kind of contract, insofar as one's situation properly resembles the hypothetical one. And there are quite a few philosophers who think that Hobbes' argument is powerful and that it might even succeed (i.e. Gauthier), just as there are many who think that Rawls was on to something.

Of course these are "stories" that aim to structure our thought. They are also arguments straight out of rational choice theory. Moti, I think your ire is properly directed at Chinese-room-style thought experiments, where the whole thing really does hinge on our having some conceptual intuition about "understanding" (or "knowledge" in the case of Mary). But I would urge you to take a softer view towards the "stories" found in ethics and political theory, where perfectly respectable arguments can and do have a certain narrative structure.


Hi Moti,

While I think arguments are great, I don't think we can equate things that provide rational support with arguments. Something has to get them going and it can't be arguments all the way down. An intuition pump can provide support for a claim so long as it provides _some_ support for a claim and it can do that even if the support it provides is nowhere near adequate for belief. If it literally provided no support at all to anyone ever, I'd be surprised.

Marcus Arvan

Moti: I think I agree with your overall characterization, which I take to be Dennett's as well. Intuition pumps are useful as a kind of starting point: of making what *appears* to be problem or some such especially vivid. The task of argument is then to examine carefully and rigorously *whether* the intuitions involved withstand scrutiny. For instance, the Chinese Room *appears* to raise a very important worry about merely syntactic engines. But (or so I think) rigorous reflection on the example sheds light on how syntactic engines *can* be semantic engines as well. And so I think a proper understanding of the pump can generate real knowledge, but not the pump all on its own.

I think I also agree that philosophers too often lose sight of this. For instance, Kripke's Godel/Schmidt intuition was long taken to *establish* a non-Russellian/non-Fregean theory of proper names. But I don't think it shows any such thing. It's just *an* intuition, and not one that everyone shares (for my part, I think names can do whatever the hell we want them to. We can use them to refer directly, indirectly, through definite descriptions, etc.). In short, I'm a Wittgensteinian. There many language games we can play with names. Direct reference is only one such game, and although it may the dominant game or "meaning of names" in the minds of some analytic philosophers, the intuition is little more than an obsession with *one* particular use of names generated by a false and pernicious background assumption: that words like names must have *a* literal meaning rather than many (which is what I take Wittgenstein's point about language games to explode).

So, yeah, anyway I completely agree with you. Intuitions are fine to invoke to discuss and frame a problem, but they are not, on their own, any evidence whatsoever before they are subject to rigorous examination. And this is too often ignored.


I'm finding these remarks interesting because a colleague and I recently went a few rounds with a reviewer over our use of intuitions pumps. The reviewer didn't think "intuition pump" was an appropriate term, but we cited Dennett's use of the term, and explained that we were offering a set of thought experiments that did not so much serve as claims but as structures for the problems we were examining that brought interesting issues to the foreground.

Really, it seems to me that intuition pumps are the most honest way to use thought experiments. I have seen various criticisms that the thought experiments employed in ethical arguments do not really demonstrate or support a claim because they are designed by the author to generate that claim. There is a notion that thought experiments are "cooked" to produce certain results, and I can understand why that would be. Nevertheless, thought experiments do invite reflection on intuitions or gut reactions, and they allow us to check our abstract thinking for consistency and coherent.

All of that is to say that thought experiments are better framed as intuition pumps, ways to get us thinking about a problem to highlight issues that we may otherwise let fade into the background.

Justin Caouette

What Clayton said.

Moti Mizrahi

Thanks for the comments, everyone.

Nick: I agree that some intuition pumps are worse than others. I also share your puzzlement about Dennett’s choice of examples. I am not sure I would count Hobbes’ state of nature as an intuition pump. And the same goes for Rawls. As I see it, the main difference between these hypothetical cases and something like the Chinese Room is that the former appeal to rational choices that can then be examined by reference to rational choice theory (as you point out), whereas the latter simply appeals to idiosyncratic intuitions.

Clayton: I am not sure I understand your distinction between support and argument. As I understand it, support for a claim can be conclusive, as in deductive arguments, or defeasible, as in non-deductive arguments. In both cases, however, to support a claim is to argue for it. Do you have something else in mind?

Marcus: We agree too much lately. :)

Justin: What Marcus said. :)


"Clayton: I am not sure I understand your distinction between support and argument. As I understand it, support for a claim can be conclusive, as in deductive arguments, or defeasible, as in non-deductive arguments. In both cases, however, to support a claim is to argue for it. Do you have something else in mind?"

Arguments can provide rational support, but arguments aren't necessary for rational support (non-inferential justification seems to be a possibility) and it seems that they can only transmit the support that supports the premises (reductios are an exception to this general rule). I know some authors would argue that arguments cannot provide any rational support unless intuitions could also provide support for accepting a premise or accepting that the premises support a conclusion. I don't buy these sorts of arguments, but there's probably a middle ground somewhere.

This idea that intuitions never provide any support at all seems to my mind to be a pretty radical position. What's the argument for it? There might be one, but I don't know what it is. Wouldn't any such argument have to show that the evidential probability of the relevant claims will never be changed by an intuition? Nothing about the unreliability or diversity of intuition would establish anything that strong.

Moti Mizrahi

Thanks for your reply, Clayton.

I know you prefer not to talk about intuitions in terms of inference. (Although I would say that I am simply making explicit certain patterns of reasoning that are taken for granted, which is part of what philosophers do, I take it.) So, for the sake of argument, suppose you are right that intuitions can provide prima facie justification for beliefs. In that case, intuitions are defeasible evidence for beliefs, i.e., the support intuitions provide for beliefs can be defeated by further evidence. Now, consider a case in which one philosopher intuits that p and another intuits that not-p, as is often the case. To use Pollock’s terminology, the intuition that p is a rebutting defeater of the intuition that not-p and the intuition that not-p is a rebutting defeater of the intuition that p. Now we need a method other than appealing to intuitions to figure out whether p or not-p. But if we need another method, why bother with intuitions? As a method of fixing belief, then, appealing to intuitions is useless and/or redundant.

By the way, if you ask for an argument to the effect that appealing to intuitions is not a good method of fixing belief, then I think that the burden of proof is on those who introduce a new method to show that the method works. For example, think of Galileo and the telescope. Imagine that Galileo said to his critics “Show me that the telescope is *not* a reliable instrument of celestial observation.”


Hi Moti,

Thanks for your response!

It's an interesting question whether disagreement defeat the justification provided by intuition. It might do so sometimes, but I doubt that it does so invariably. There are tricky issues about peers and equality of epistemic position that need to be taken account of. The most skeptical view of disagreement seems to be an equal weight view and even that doesn't seem to apply to lots of the interesting cases. It certainly cannot sustain the argument for the claim that intuitions never provide any evidence. (Of course, even if the justification intuition provides to belief isn't defeated, there are dialectical concerns that are perfectly appropriate.)

The analogy with the telescope is interesting, but do you really want the intuition mongers to be on the same side historically as Galileo?

I sometimes worry about blanket intuition skepticism. Substitute thought or line of reasoning for intuition. They're fallible, defeasible, contentious, etc., but they're often good enough in spite of this. When you say, "By the way, if you ask for an argument to the effect that appealing to intuitions is not a good method of fixing belief, then I think that the burden of proof is on those who introduce a new method to show that the method works." My immediate response is that the method isn't new. I'd also say that intuition is an important contribution to the method of reflective equilibrium and I have yet to see any argument for the conclusion that intuitions cannot provide a minimal degree of support. Once you have that, it can be amplified in a variety of ways just like uncertain observations can.

Notice, I haven't claimed that an appeal to intuition should settle a contentious issue in philosophy. There are lots of ways of avoiding that commitment without chucking intuition out the window.

Dan Dennis

I like your articles on Intuition in The Reasoner, Moti, and recommend them to others who are interested in intuitions.

Could it be that a valid way to use stories (aka ‘intuition pumps’) is to help us construct new ideas, concepts etc. and to convey them to others. We may then employ those ideas, concepts, etc in arguments.

Perhaps the problem is that those using these stories commonly think that once they have told their story they have done sufficient work.

Alex Hughes

The title of this post surprised me precisely because I have found it enormously useful to think of thought experiments as compressed arguments.

A thought experiment works roughly like this: One entertains a scenario and judges that if it were actual, then such-and-such would be true. [The clock is stopped; you don't know -- even though you believe truly and with justification!] We are not normally particularly explicitly aware (at first) of what it is about the scenario that inspires the judgment -- which is why, I suppose, we tend to call the judgment "intuitive".

It is perhaps best to think of a description of a thought experiment that does not perspicuously validate the judgment as enthymemetic. The job of the theoretician is to carefully describe the salient features of the scenario -- and their interaction with background assumptions -- in order to construct an argument that would validate the link between the scenario and the judgment. [Validation is the best case scenario! Often enough we find that we cannot construct a validating argument -- perhaps because the "scenario" is somehow incoherent.]

If I am right, then thought experiments provide warrant for judgments in the same way as arguments (because they are arguments). [Though it is something of a vexed question how arguments provide warrant.] If we think of them as something other than compressed arguments, then it is indeed something of a puzzle to say what their proper role might be.

But even if thought experiments are not arguments, it is the job of the theoretician to turn them into arguments!

Moti Mizrahi

Clayton, thanks for the follow-up. As I see it, the problem with appealing to intuition as a method of fixing belief is not merely that the method does not eliminate disagreements but rather that the method provides evidence for contradictory beliefs. That is, assuming for the sake of argument that intuitions provide prima facie justification, which I don’t think they do but I am willing to grant for the sake of argument, then when equally competent philosophers use this method, it yields evidence for both p and not-p. But any method of fixing belief that does that is useless as a method of fixing belief. Please note that this argument is different from the one I make in the Reasoner paper. There I argue that the method is unreliable.

Dan, thanks very much for your kind words. I am inclined to agree with you that intuition pumps can play a useful role in the context of discovery. But when we switch to the context of justification, we must realize that appeals to intuition are not good arguments.

Alex, thanks for your comment. I would say that the hypothetical scenarios themselves are not arguments. The intuitive judgments philosophers make in response to hypothetical scenarios are then used as premises in arguments. Any reconstruction of this kind of reasoning, however, would have to include a move from ‘it intuitively appears (or intellectually seems) that p’ to ‘p’. It is this move that is problematic.

Alex Hughes


I should be clearer: the scenario is not the argument. Rather: the transition from the scenario (there is a stopped clock, you believe on its basis that it is noon, it is noon) to the intuitive judgment (you don't know it is noon and you are justified in thinking it is noon) is the argument. Intellectual seeming plays no more of a role here than in any argument. A reconstruction would articulate (among other things) the features of the clock that determine why believing on the basis of its position is justified, why doing so in this context doesn't yield knowledge, etc.

Notice that thought experiments are objected to in ways in which arguments are. "Actually, your belief would *not* be justified in the relevant way. Here's why." or "Actually, you *would* know! Here's why".

Dan Dennis

Hi Moti

Yes, I agree

Moti Mizrahi

Alex: I am not sure what you mean by "Intellectual seeming plays no more of a role here than in any argument." I hope you can elaborate on that. As I see it, a philosopher considers a hypothetical case C and intuits that p. From 'it seems that p', s/he gets 'p', which s/he then uses as a premise in an argument (e.g., an argument against JTB that goes something like this: "If JTB = K, then S knows that p in C; S doesn't know that p in C [*this is the intuitive judgment*]; therefore, JTB ≠ K"). The question, then, is this: if a philosopher intuits that p is p true or more likely to be true? I think not.


Hi Moti,

I think you've raised an interesting problem for the intuition monger (mongerer?), but I don't think the problems articulated yet quite identify any fatal problem for the view.

For a start, I struggle to think of a method that won't provide evidence for both p and ~p. I suppose that this will depend upon some tricky issues about individuating methods (or processes, sources, whatever your preferred terminology). Expert testimony will provide evidence for p and for ~ p (that's why you got to ask the _right_ experts, as Malcolm Tucker reminds us...). Observation will often provide us with evidence for p and for ~p. Maybe introspection won't? Even that seems a bit doubtful.

For another, I wonder whether it's a good idea to think of intuition as a homogenous category. I can't think of any reason to think that everything that falls under the heading of 'intuition' is generated by the same psychological processes. Moreover, I can't think of any reason to think that everything generated by these processes ought to be assessed on par. After all, there are appropriate and inappropriate circumstances for relying on observation and testimony. Why shouldn't similar issues arise for intuition. Just as one shouldn't argue from the fact that there are conflicting perceptual appearances to the conclusion that observation provides no evidence and one shouldn't argue from the fact that one can try to rely on experience in circumstances where experience isn't a reliable guide to the conclusion that experience cannot provide justification, I fear that the intuition fear mongering often appeals to 'principles' that cannot be consistently applied to things like perception or testimony without leading to some rather implausible sceptical conclusions. (I celebrate, however, your scepticism about seemings. I don't think torturing our beloved language is a great way of doing epistemology or mind.)

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been saved. Comments are moderated and will not appear until approved by the author. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Comments are moderated, and will not appear until the author has approved them.

Your Information

(Name and email address are required. Email address will not be displayed with the comment.)

Job-market reporting thread

Current Job-Market Discussion Thread

Job ads crowdsourcing thread

Philosophers in Industry Directory