Thought Experiments and Philosophical Method
In an interview at 3:am Magazine, Richard Marshall presses Philip Kitcher (Columbia) on his criticism of a priori, thought-experiment-driven approaches to philosophy. Marshall says that a criticism of Kitcher’s view is that it “would end much typical philosophical investigation.” Kitcher replies:
Thought experiments work when, and only when, they call into action cognitive capacities that might reliably deliver the conclusions drawn. When the question posed is imprecise, your thought experiment is typically useless. But even more crucial is the fact that the stripped-down scenarios many philosophers love simply don’t mesh with our intellectual skills. The story rules out by fiat the kinds of reactions we naturally have in the situation described. Think of the trolley problem in which you are asked to decide whether to push the fat man off the bridge. If you imagine yourself – seriously imagine yourself – in the situation, you’d look around for alternatives, you’d consider talking to the fat man, volunteering to jump with him, etc. etc. None of that is allowed. So you’re offered a forced choice about which most people I know are profoundly uneasy. The “data” delivered are just the poor quality evidence any reputable investigator would worry about using. (I like Joshua Greene’s fundamental idea of investigating people’s reactions; but I do wish he’d present them with better questions.)
Philosophers love to appeal to their “intuitions” about these puzzle cases. They seem to think they have access to little nuggets of wisdom. We’d all be much better off if the phrase “My intuition is …” were replaced by “Given my evolved psychological adaptations and my distinctive enculturation, when faced by this perplexing scenario, I find myself, more or less tentatively, inclined to say …” Maybe there are occasions in which the cases bring out some previously unnoticed facet of the meaning of a word. But, for a pragmatist like me, the important issues concern the words we might deploy to achieve our purposes, rather than the language we actually use.
If the intuition-mongering were abandoned, would that be the end of philosophy? It would be the end of a certain style of philosophy – a style that has cut philosophy off, not only from the humanities but from every other branch of inquiry and culture. (In my view, most of current Anglophone philosophy is quite reasonably seen as an ingrown conversation pursued by very intelligent people with very strange interests.) But it would hardly stop the kinds of investigation that the giants of the past engaged in. In my view, we ought to replace the notion of analytic philosophy by that of synthetic philosophy. Philosophers ought to aspire to know lots of different things and to forge useful synthetic perspectives.
Lots to chew on in here, not least of which is the parenthetical—“In my view, most of current Anglophone philosophy is quite reasonably seen as an ingrown conversation pursued by very intelligent people with very strange interests.“ Intuitions about any of this?
(image: detail of “Girl Before a Mirror” by Pablo Picasso)
Given my evolved psychological adaptations and my distinctive enculturation, I find myself, more or less tentatively, inclined to agree with the middle paragraph. The final bit seems to me mostly speculation mixed with some less than convincing assessments of how the discipline currently stands (for instance, I’m not sure philosophers’ interests are any more strange than those of very specialized biologists, or historians of pre-[insert historical event] [insert geographical area]).
Most controversial for me is this: “But even more crucial is the fact that the stripped-down scenarios many philosophers love simply don’t mesh with our intellectual skills. The story rules out by fiat the kinds of reactions we naturally have in the situation described.” Now, on some potentially defensible view of thought experiments, it is exactly the fact that they rule out our natural reactions to situations that prove their utility. The thought being that our intellectual skills, as they stand, aren’t especially useful in getting to the bottom of whatever topic is being investigated. So let’s pump some unnatural, forced intuitions! Those might go a long way in getting us out of intellectual ruts and pernicious philosophical habits.
An argument against implicit ceteris peribus clauses based on the claim that “they create forced, awkward, non-natural choices” cannot be a good argument. That is the point. And if they are too restrictive, just play with the clauses to allow or deny the responses desired for the particular purpose. It seems easier to tweak the them then to get rid of them entirely.
This sort of skepticism about thought experiments (or at least those which are “unrealistic”) is not really original with Kitcher. Other than the brandishing of the “pragamatism”
talisman no real alternate method is proposed. And for the good reason that there is no other method with which to address the central topics of philosophy such as the nature of knowledge, evidence, personal identity, justice, consciousness, the good, truth, etc.
Skepticism about thought experiments threatens to transmute into a far more global skepticism and tends to be self-undermining.
“there is no other method with which to address the central topics of philosophy such as the nature of knowledge, evidence, personal identity, justice, consciousness, the good, truth, etc.”
!!!
recent grad [4]
Could you elaborate?
There are large numbers of philosophers who try to address those topics in empirical or pragmatist ways. (There’s even a book with the title “Real People: Personal Identity Without Thought Experiments.”) Perhaps you don’t think these methods are, in the end, any good. But that is not the same as saying that there are no such alternatives.
Thanks, recent grad [4, 6]. You are certainly correct that there are philosophers who “try to address these topics in empirical or pragmatist ways” (and Wilkes may be a good example). I don’t, myself, think Wilkes ends up really addressing (rather than avoiding) the question of the *essence* of personal identity. If we want to know the metaphysically necessary and sufficient conditions for personal identity we have to depart from actuality (and even nomic necessity) in ways that Wilkes and her fellow travelers are loathe to do and that looks to require appeal to intuitions about merely possible (hypothetical) cases. S
Isn’t the reason you’d look around for alternatives to pushing the fat man, consider talking to him, etc. just that pushing him would be obviously wrong? It seems to me that Kitcher has, and relies on, the very intuition he’s trying to debunk.
I think there is an important difference between saying (i) that pushing the fat man as your first plan of action is wrong and (ii) that pushing the fat man is wrong even when the only other choice is to let multiple people die. I don’t think the fact that Kitcher has the first intuition is evidence that he’s relying also on the second.
“Philosophers ought to aspire to know lots of different things and to forge useful synthetic perspectives.”
The “useful” should be in bold caps along with the “ought”. Pragmatism all the way down it seems, and I’m not unsympathetic to that.
However, though I’m also skeptical that a priori arguments/thought experiments produce some special province of knowledge, I do think that such arguments have a significant “paring” function in taking concepts that claim to be full-blooded truths and cutting them down to size, or at least making them more manageable. In my own area of interest in free will and moral responsibility, Harry Frankfurt’s attempted take-down of alternative possibilities was a completely novel recasting of compatibilism, as was Ginet’s/van Inwagen’s consequence argument for showing the logical results of a concept of determinism. Even more recent empirically-loaded work like Neil Levy’s Hard Luck uses data to show that logically neither incompatibilism nor compatibilism can surmount skepticism about anything like basic responsibility, but in the tradition of hard logic to produce significant results.
I think Kitcher’s critique of thought experiments would be quite good if we used our intuitions about them as decisive refutations of philosophical theories. If a theory said one thing about a weird case, and intuition said something else, and we threw out the theory, then that’d be bad methodology because our intuitions about odd cases might not be accurate. But at least a lot of the time that’s not really what happens, it seems to me.
Take Thomson’s thought experiments in ‘A defense of abortion’. If we rested our views about abortion just on intuitions about cases involving violinists, people seeds and expanding children in confined spaces, that’d be bad. But it’s enough that Thomson’s examples make us realize that the debate has been focused only on whether embryos are people, rather than on under what circumstances autonomous agents can be expected to make large sacrifices for others. Consideration of the weird cases prompts us to think in a more nuanced way about the normal cases. Once you’ve thrown away the ladder, Kitcher’s critique of the ladder isn’t really important anymore.
You could say similar things about other thought experiments. Newcomb’s problem makes us realize that evidential and causal decision theory can give different verdicts, even in real-life cases. Kripke and Putnam’s thought experiments made us consider alternatives to descriptive theories of reference. The fat man on the bridge, Jim and the Indians and Parfit’s mineshaft can also prompt re-examination of theories in ways that can lead to more satisfying treatments of everyday cases. Kitcher is probably right about brain-swap thought experiments though; those are just silly.
I wrote about this some years ago. Permit me to quote a pertinent section:
Unlike sporadic bouts of purely discursive activity such as casual conversations, philosophy seminars are not simply ad hoc departures from instrumental language-use. Philosophy is systemically and consistently isolated from instrumental discourse, from the concrete circumstances in time and space that give words determinate meanings. Even if philosophers have perfect memories, and all words in philosophy are taken from natural languages, it would not matter. The memory of how to use an ordinary expression is of little use in an extra-ordinary setting.
In typical non-philosophical settings, real-world empirical criteria would filter out inappropriate linguistic usages (for example, someone referring to a fork as a ‘spoon’), but in philosophy there are no such tests. There are only logico-grammatical tests in the a priori realm of thought experiments; tests which are under-determinative of linguistic sense because in them, philosophy does not use language to do things in the physical world (other than manipulate purported symbols), or demonstrate sufficient continuity with natural languages (beyond using the same words in the same order) to ride on the latter’s claim to meaning.
The arbitrariness of thought experiments is greatly compounded by the philosophical practice of pitting them against each other in an increasingly complex imagined landscape, which barely makes contact with the real world via instrumental use. It is this arbitrariness that elicited Peter Hacker’s pithy characterization of thought experiments as “no more experiments than monopoly-money is money” (Hacker 2005, 27).
Extract from ‘Why Philosophy Fails’ by Ben Gibran
Reference: Hacker, Peter. ‘Analytic Philosophy: Beyond the Linguistic Turn and Back Again.’ Unpublished Manuscript, 1 April 2005. http://info.sjc.ox.ac.uk/scr/hacker/docs/ (filename: ‘Beyond the linguistic turn’, accessed November 26, 2011).
It’s quite strange that Kitcher seems to be restricting the cases where intuitions might show anything to cases in which we are investigating our language as such. It seems to me that even if he wants to claim that philosophers are doing it all wrong, he still should at least accept that intuitions can show something about ourselves — as the creatures adapted to the enviroment the way we are. He has been engaging in the debate about scientific explanation a lot himself, and in the texts I know he never objected to the flagpole example that it was just showing something about our use of the linguistic expression “explanation”. The flagpole example seems to show something about explanations as such as well, because we are the ultimate judges about whether a proposed explanation does explain things to us or not. This might — with certain metaethical assumptions — be different for intuitions about ethics, but at least the claim that intuitions do not help us to go beyond language never made any sense to me.
Kitcher’s view here (applied to ethics as it is in his review of Parfit) seems like it could only be correct if, as a matter of fact, we know almost nothing about what would be right or wrong should some unusual, streamlined case turn out to be actual. But in fact we know many things about what would be right or wrong under very bizarre conditions.
But in fact we do know a huge amount about what is right or wrong in bizarre cases. This is why we are able to watch movies or read books with unusual or surprising plots without constantly thinking “Oh my god, I have no idea whether these characters are acting rightly or wrongly!”
SPOILER ALERT if you haven’t seen the wonderful Henry Fonda movie Fail Safe:
In Fail Safe, the Fonda as the president is faced with the decision of what to do if an accidental US nuclear strike against the Soviet Union succeeds in blowing up Moscow. He decides he will respond by dropping a nuclear bomb on New York, so that the Russians don’t counterattack and destroy the US. It’s made clear in the film that this is the only deal the Russians will accept. If he doesn’t destroy New York, the Russians will kill off two entire countries.
I have never encountered this situation or anything like it in real life. But I know that when Fonda decides to blow up NYC, he makes the correct decision. Call this an intuition if you like. It’s something I know to be true. And if I’m arguing ethics with someone, I’m going to be perfectly happy to use this piece of knowledge as a premise in my arguments. (That’s all that “intuition” usually means in moral philosophy: a moral intuition is a moral fact most of us are in a position to know which is being used as a premise.)
If Kitcher wants to say that I don’t know whether Henry Fonda makes the right choice in that movie–which I think he would have to say in order to maintain his anti-thought-experiment view… well, if he wants to say that, I strongly disagree with him. I would have to conclude that his conscience is a very different sort of thing from my own conscience, if it is silent on that sort of hypothetical question.
Dave Baker wrote: “Kitcher’s view here (applied to ethics as it is in his review of Parfit) seems like it could only be correct if, as a matter of fact, we know almost nothing about what would be right or wrong should some unusual, streamlined case turn out to be actual. But in fact we know many things about what would be right or wrong under very bizarre conditions.”
I don’t think the problem is simply that our response to a hypothetical case may differ from our response to a similar real-life situation. As Dave rightly points out, our predictions about how we (or even others) would act is fairly accurate a lot of the time. However, thought experiments in philosophy are not primarily about making predictions.
In philosophical ethics, the general role of thought experiments is to highlight apparent inconsistencies in how we react to different hypothetical situations. The rationale behind this approach is that such inconsistencies should prompt us to examine and revise our beliefs with a view to making them more consistent. The problem is, there is no evidence that thought experiments actually do that, to any significant extent, or that consistency is an over-arching principle in everyday decision-making.
I don’t personally know of anyone who has changed their views on abortion or meat-eating because of philosophical thought-experiments and arguments. There are people who claim to have been influenced this way, but actually verifying that claim is a tricky business (and I suspect that there are few people who would even make the claim, and far fewer for whom the claim is reasonably justified).
From a philosopher’s point of view, we often appear to act inconsistently. Many philosophers find this problematic because they assume that practical moral reasoning is simply ratiocinating from moral beliefs. But there is no evidence that we consciously apply moral rules when we decide what to do in real life. A lot of the time, we just act, and it takes a sociologist or psychologist to show us that we’re unconsciously following a pattern that has to do, for example, with our social status or gender. I think that’s why Kitcher mentioned the importance of empirical data for ethics.
Of course, if we are asked for some of our moral beliefs, we can rattle off a list, but these are usually non-contentious broad principles that don’t really explain the fine grain of how I decide what advice to give a friend with marital problems, or how much I should give to which charity. Empirical data can often have real impact on such decisions, but philosophical thought-experiments less so, it seems.
This is probably because philosophical theories tend to ignore the social and psychological factors that subconsciously influence our behavior, so end up not engaging the actual decision-making mechanisms that have a real prospect of changing our minds. Hence Kitcher’s advice that philosophers “ought to aspire to know lots of different things [especially in the sciences] and to forge useful synthetic perspectives.”
Nice example David (if I may). And may I spoilier add the President knows his wife is in New York. And I still agree with you.
Personally, I’m sick of all these epistemologists and their perception-mongering — as if they had unfiltered access to the causes of phenomenal experience!
We’d all be much better off if the phrase “I see a …” were replaced by “Given my evolved psychological adaptations and my distinctive enculturation, when saddled with a potentially delusive experience, I find myself, more or less tentatively, apparently in the presence of …”
Ben, I think the hard-to-change nature of people’s views on ethics is a pretty interesting issue. I would just say that I see my job as coming up with arguments (for interesting views) that *should* convince people if they are thinking in the right way. If I saw my job as coming up with arguments that convince people whether they are thinking rationally or not, I’d probably study rhetoric or propaganda or something. Anyway, my point above was just that arguments appealing to well-constucted thought experiments should be convincing. If they don’t in fact convince people, that’s an important topic for further research, but not an indictment of the thought-experiment arguments as philosophy.
In a sense the thought-experiments are tools to test statements and arguments that support them. They may take the “evolved psychological adaptations and my distinctive enculturation” to a test drive. Is that devoid of philosophic purpose? Of a purpose of educate in critical evaluating the beliefs people have?