Thursday, December 09, 2004

Experimental Philosophy Naturalized

When I read this post over at Majikthise on Quine's naturalized epistemology, it got me all nostalgic for a moment. I remember, when I first left philosophy for cognitive psychology, I used to tell people (especially my old philosophy friends) that what I was doing was experimental philosophy. This was before I had heard of the emerging area of analytic philosophy that refers to itself by that name, of course, but even philosophers like Stephen Stich weren't the first to use the label. In fact, I had stolen it from Voltaire (as quoted by Horkheimer and Adorno), who had used it when he referred to Bacon as the "first experimental philosopher." But that's really what I thought I was doing - experimental philosophy. In my mind, the sort of research I was doing and would do was both the data-driven end of philosophy of mind, and the metaphilosophy or metaepistemology (I stole that term, too, from Ernst Nagel and Richard Brandt) behind all epistemological inquiry. Even if we conceptualize epistemology as a normative discipline, as I think most philosophers have done over the last 100 years, the experimental metaepistemology that cognitive psychologists and others in related fields do is invaluable. To have a proper "theory of knowledge," and to understand epistemological terms like "warrant," "reference," and "knowledge" itself, we have to first understand the perceptual and conceptual capacities of knowing subjects, and the ways in which those capacities limit the types and amounts of information that subjects can take as input, and the types of information that they can produce as output. For instance, taking seriously what we've learned from cognitive psychology forces us to treat knowing subjects, and therefore knowledge, as embodied, situated, and ecologically limited.

If you've read this far, you probably recognize that my view of my own discipline isn't too different from Quine's. Here is a passage from Quine that Lindsay quotes in her post:
The old epistemology aspired to contain, in a sense, natural science; it would construct its somehow from sense data. Epistemology in its new setting, conversely, is contained in natural science, as a chapter of psychology. But the old containment remains valid, too, in its way. We are studying how the human subject of our study posits bodies and projects his physicis from his data, and we appreciate that our position in the world is just like his. Our very epistemological enterprise, therefore, and the psychology wherein it is a component chapter, and the whole of natural science wherein psychology is a component book—all of this is our own construction or projection from simulations like those we were meting out to our epistemological subject. Thus there is reciprocal containment, though containment in different senses: epistemology in natural science and natural science in epistemology.—W.V.O. Quine, Epistemology Naturalized, pg 25.
Lindsay uses this passage, and the sentiments therein, as a launching point for a discussion of what is today (rather than in the 18th century) called experimental philosophy. I, however, wonder if experimental philosophy, as its practiced by philosophers like Stich, is really what Quine had in mind. The general method of experimental philosophers seems to be to present people (college students, most often) with stories, and to answer a few questions related to important philosophical issues, designed to get people to tell them what they think is going on. This practice is meant to solve the problem of philosophers doing their own individual thought experiments (experiments with an N of 1, as Lindsay puts it) made up of problematic (and often counterfactual) situations, and using their intuitions to develop philosophical pictures of the concepts involved in those situations. What experimental philosophy, in this sense, can do is a.) show that there are several possible intuitions (each of which is held by members of different populations, perhaps), and then b.) investigate the extent to which those different intuitions are justified.

What I wonder is, can experimental philosophy do this? And is this really even something that needs to be done? To answer the second question first, I'm pretty sure the fact that different philosophers, usually from within the same population, have different intuitions about the same situations renders experiments designed to show that this is true for non-philosophers as well fairly superfluous. One can think of the last 50 years (or 2500 years, depending on your preference) of epistemological inquiry as an extended experiment with an N in the thousands, and the data should force us to admit that there is a diversity of intuitions about most, if not all of the sorts of epistemological problems and concepts that philosophers consider.

My answer to the first question, which asks whether experimental philosophy can tell us anything about the intuitions epistemologists use, is just this side of no. For one, I don't think this is the sort of thing that Quine had in mind. It's certainly nothing like the way psychology is done by psychologists. Outside of clinical and social psychology (which I seriously doubt experimental philosophers want to take as their methodological models), psychologists almost never conduct experiments designed to get participants to tell them what they think about something. That's what people in marketing departments do. Sure, in cognitive psychology, self-report measures are used, and with frequency in some sub-areas, but in most cases, what the researcher is interested in is not so much what a person says he or she thinks about the stimuli, but the cognitive and percpetual mechanisms that underlie what he or she thinks about them. For example, some psychologists (I'm not one of them) believe that they can understand the subtelties of the phenomenological character of different types of memory, or memory for different types of information, by asking people to indicate whether theythey "remember" seeing the information during the learning phase (i.e., they remember the presentation itself), or only "know" that they saw it (i.e., they have a feeling that they saw it, but don't remember the presentation of that item) - the so-called "remember-know" procedure (see this paper for an example of its use). In most cases, though, psychologists tend to be wary of self-report measures, and prefer indirect measures of the phenomena being studied (even the remember-know procedure is pretty indirect). This is in part because experiment participants tend to say what they think you want to hear, but it's also because much of what is interesting, scientifically, about cognition is not intrpospectively available. The consensus among cognitive psychologists is that the science of cognition has to rely primarily on third-person methodologies, using first-person reports for primarily abductive purposes.

One criticism of experimental philosophy and its methods, therefore, is that it doesn't look a heck of a lot like the rest of experimental psychology. Then there's the question of what the hell the results of experimental philosophy experiments demonstrate. For one, they tend to present fairly strange circumstances, the sort that philosophers use all the time, but which nonphilosophers will probably never consider (except in drunken conversations with philosophy majors late at night in dorm rooms). Of course, this is a criticism of many cognitive psychology experiments as well, but the awkwardness of those is usually caused by the indirect nature of the experimental measure. In the case of experimental philosophy, the awkwardness is intentional, and the measurement direct. Do peoples' intuitions about reference in unusual counterfactual scenarios really tell us about their intuitions in ordinary situations? Philosophers are trained to use certain pathological cases to help them develop generalizable intuitions; undergraduate social work majors are not. Then there's the types of questions experimental philosophers ask about their awkward situations. A common criticism from people like J. L. Austin and Norman Malcolm is that philosophers use words that ordinary people never do to refer to ordinary things, and that they use words that ordinary people use to refer to things that ordinary people never refer to. Are we really getting at peoples' intuitions about "intentionality" and "reference" by asking them questions with those terms, which they probably never use (how often, really, do ordinary people say, "I intended to do that?") in those contexts?

I think the sort of thing that I do is the real experimental philosophy (actually, I think all science is experimental philosophy in a way), and that if philosophers want to develop better intuitions, then it's the sort of work they should be doing, or at least paying attention to. If you want better intuitions about reference, you don't go out and ask a bunch of people what their intuitions about reference are. Instead, you go out and look at how people refer to things, and then use that information to develop more educated intuitions about reference. Seriously, isn't that what philosophers who use other branches of philosophy do? When philosophers want better intuitions about time, they don't go out and ask a bunch of people what they think about time. Instead, they look at what physics experiments and theories have to say about time, and use that knowledge to develop better intuitions. Granted, time is supposed to be an objective property of the world, while reference is at least partially subjective (it takes a subject and an object to do any real referring), but I think the principle is still the same.

Have I ranted long enough? Probably too long. I'll just shut up now.

13 comments:

Anonymous said...

Outside of clinical and social psychology (which I seriously doubt experimental philosophers want to take as their methodological models), psychologists almost never conduct experiments designed to get participants to tell them what they think about something. So, just to rekindle the rant, what do you think of sociologists who seem to do this all the time and then draw implications from it? I've long been critical of sociology for that sort of thing precisely because I don't think we are good judges of our own beliefs. That has a long history of course, going back to Plato. (The dialog we read in college on it was the Alcibiades although it is one of the disputed dialogs - but it's still quite good) The idea is that in a dialog with others we use that encounter with them as a mirror to see our own soul.

My view is that philosophy has, in many ways, lost that dialog character of philosophy it once had. But then I say that and people tell me that's what publishing is supposed to accomplish, along with the questions at the end of paper presentations. Still something seems wrong. 

Posted by Clark

Anonymous said...

Survey research in sociology and other disciplines has its place. Most of the time, the variables with which sociologists are concerned simply aren't amenable to real experimental methods. Still, you have to take any self-report measures with a big grain of salt, for reasons that people who study response biases have been talking about for decades.

As for philosophy, I think the other place where is retains its dialog character is in the talks and presentations that philosophers give to other philosophers. Of course, those are often glorified pissing contests, in philosophy and every other academic discipline that I've heard talks in, but at least it offers the oppurtunity of interactive study. Plus, I think very few academics publish in scholarly journals without spending a significant amount of time discussing the ideas with colleagues. In science, this is what lab meetings are for, and it's what makes science work much more smoothly. Maybe philosophers should have lab meetings (maybe some do). 

Posted by Chris

Anonymous said...

As far as I'm concerned, you can't go on too long ranting about this subject; I went through your post with an "Absolutely" and an "Amen to that" on almost every point.

I think a lot of the problems with experimental philosophy derive from the counterexample method which is used so much (to the point of abuse, IMHO) in contemporary philosophy. It's only under such a method that 'intuitions' begin to play such a central role in actual philosophical reasoning: the whole pressure of the reasoning ends up forcing philosophers to try to persuade people that what they really think about example E is such-and-such rather than so-and-so. That people recognize this is a problem, and are trying to deal with it, is a step forward; but the approach taken in 'experimental philosophy' is often just a more elaborate and pretentious method for doing the same thing, rather than a real reformation of the practice.

Plus, it has never been very clear to me what one gets out of it, beyond comparison of one's own hunches with other people's. Suppose you put forward a sophism and survey people's intuitions about it. Conceivably this might help give you some hint about why it is so deceptive, once you've resolved the sophism; but it really won't go very far toward resolving the sophism itself. I think there's some value in the 'experimental philosophy' movement in that it potentially provides a restraint and checking method for certain things; but this is really quite a modest gain.

My thought about philosophical loss of dialogue is that it's due to the relatively common tendency to separate teaching philosophy from doing it, as if they were different things. After all, that's what a lot of the older dialogue was in the first place: doing philosophy by teaching it. 

Posted by Brandon

Anonymous said...

I agree Brandon. One of my favorite philosophy classes used a Socratic method. One seat had to state their question from that week's reading. The next seat had to answer this question. Then the next seat raised a question. This counted for I think 15% of the grade. But it really fostered a way of doing philosophy that some of my other classes never engendered.

Further I know a lot of people hate the Socratic method - especially lawyers. Even if it apparently is still very common in law school. (Not being a lawyer I don't know - but I have heard law student rants)
 

Posted by Clark

Anonymous said...

Chris, I just found your post. Thank you so much for your feedback. I'll write a proper reply soon.  

Posted by Lindsay Beyerstein

Anonymous said...

Hey Lindsay, no problem. I'm never convinced anyone wants to read my comments. Otherwise, I'd let people know about them when I'm responding to something they've said. 

Posted by Chris

Anonymous said...

As a civilian to philosophy and science, I have what will surely strike you as a naive question, but nevertheless:

Isn't a good, succinct definition of any kind of science "experimental philosophy?"

It strikes me that perhaps the field of cognitive science might be deemed "experimental epistemology" but that seems close to redundant, yes? because nearly all true assertions about "how we know things" would surely require empirical confirmation before they could be called true.

But then, what do I know? 

Posted by tristero

Anonymous said...

Well, the sciences originated out of philosophy, and the term was applied to Bacon, who was hardly a psychologist. I tried to imply somewhere in there that things like physics could be experimental philosophy as well, but did so poorly 

Posted by Chrsi

Anonymous said...

I got here via a link to Lindsay's paper about NE from the experimental philosophy webpage - which may well taint the following remarks :)

Whilst I find Chris' remarks congenial I think the situation regarding experimental philosophy (EP) is more complex than his discussion suggests. Surely it is right that part of a real epistemology must come from an understanding of the capacities of inquirers (I prefer animal cognition myself but drinking with too many cognitive psychologists will do that to a person!) But, it is also the case that key concepts of epistemology are deployed explicitly by we humans in various attempts to influence one another - Now as Hacking has shown us (see his Social Construction of what?) humans reify social constructions in perverse ways at approximately the speed of light. So in addition to our 'raw' capacities as inquirers we also need an account of how our community deploys the terms used to evaluate inquiry *and* some story about how these two areas interact. (I owe this picture of epistemology as a whole to Fred Schmitt at IU Bloomington - who is the best epistemologist I know).

Back to EP - i think that it's best seen as an attempt to do the basically sociological project of sorting our how terms of inquiry evaluation get used in various places. This view I think fits with Chris' views about sociology, EP and survey questionaires (Note - on the subject of the value of questionaires - go see Kinsey and weep - his questionaires make him a hero but his fieldwork makes him a pervert apparently - aaaaaaaaaaaaaaaaaaaaarrrrrrrrrrrrrrrrrrgh). Now I don't think that Weinberg, Nichols and Stich (WNS) would disagree with Chris regarding the primative nature of their reseach to this point but that surely is part of its appeal - if they are getting results that challenge traditional philosophy with the limited resources they've deployed we must surely be optimistic about what a richer research project might accomplish.

In the end, it seems to me, we're looking for a way to integrate Chris' cog sci with the WNS work rather than to rule either as *the* experimental philosophy. 

Posted by Yelworcs

Anonymous said...

Yelworcs, interesting. I see a lot of very good research on how people use concepts, how their concepts change over time, how social, affective, and even epistemological factors influence representations, etc., and it all uses more indirect cognitive or social cognition methods. I don't think trying to understand the complex ways in which concepts are represented, used, etc., is a bad thing, I just think it's already done, and in some cases done well.

One of the criticisms of traditional concept research is that it relies on stimuli that are about as far from ecological validity as Sydney is from Texas, but over the last few years, there has been some excellent concept research using hybrid methods that have a great deal of ecological validity. I'll probably post on this at more length sometime, but I do talk a little about Medin's cross-cultural research, research on the differences between the concepts of experts and novices, etc., in various posts here and there.

Anyway, my point being, if all they want is a survey of the various beliefs people have within and across cultures about the definitions of certain philosophical terms (or if not the definitions, then how they think these terms apply to abnormal circumstances), then Experimental Philosophy is probably the right way to go about it, but I've gotten the impression that's not what I take the motives to be. 

Posted by Chris

Anonymous said...

Chris,

I know bugger all about psychological work on concepts - so I'd be more than pleased to be educated. May I, in fact, ask for that? :)

As for the motives of EP - I can only speak for my own corner of that enterprise but the work of Weinberg Nichols and Stich is aimed in large part at just the job you mention (getting clear about how various folk think about key philosophical examples/notions). There are clearly other motives within the EP community and they need to be evaluated on their own merits but I suspect we should stay away from blanket evaluations of an enterprise as catholic as EP.

That said, let me offer a completely unsubstantiated speculation. Chris, mentions Bacon as an early EP (now I'm a little worried that he's really a philosopher of experiment rather than an EP but let that rest), the guts of Bacon's project was to *improve* his communities methods of inquiry - it's this goal that I want to think about. It seems to me that the improvement goal (IG) has both empirical and normative dimensions - it's meant to be an improvement and that's normative but we find and evaluate candidate methods empirically (marvellous term that - let's me fudge a bunch of *really* hard questions). Now as things currently stand, it seems to me that IG gets a raw deal - philosophers worry about the normative but are no longer seriously trying to improve our methods of inquiry - cognitive scientists on the other hand appear to be mapping our actual methods of inquiry without too much concern with which ones are "best". My speculation therefore is that much EP arises out of the tug of IG and the lack of any obvious place to pursue it (IG that is) .....

Clearly more needs to be said here but it's xmas and my roof is leaking so I'll leave it here! 

Posted by Yelworcs

usadrug said...

Hi, i think the other place where is retains its dialog character is in the talks and presentations that philosophers give to other philosophers.
...........................
usa drugstore - my favorite pharmacy.

Unknown said...

thanks