x-phi + a-phi = phi

This is the mathematical equation for reconciling philosophical methods that Dave Chalmers put up at the end of his excellent talk that closed the ‘Experimental Philosophy Meets Conceptual Analysis’ conference here at ANU.

Dave’s talk raised some interesting questions about what exactly experimental philosophy is and what exactly conceptual analysis is. ‘A-phi’ refers to ‘a priori philosophy’, which Dave suggested better captures the traditional projects of analytic philosophy better than ‘conceptual analysis’ does. After all, it does seem as though many ‘experimental philosophers’ are engaged in conceptual analysis of a sort, so using these as dividing lines might be a bit misleading.

A distinction was also drawn between ‘experimental philosophy’ and ’empirical philosophy’; my work on moral disagreement was cited as an example of the latter, since I didn’t actually do any of the experiments myself. The kind of work done by Joshua Knobe and Shaun Nichols, on the other hand, falls squarely into the former category.

Finally, ‘negative’ experimental philosophy was distinguished from ‘positive’ experimental philosophy; the former, Dave claimed, seeks to raise doubt about the philosophical status of intuition by finding disagreement in those intuitions, while the latter attempts to detect interesting and informative patterns in intuitions. I’m a bit skeptical of the stability of this distinction. My work on disagreement initially seems to fall into the former category, but I think it can also be argued that cross-cultural variation in moral beliefs is, in its own way, an ‘interesting and informative pattern’. So it seems that it’s possible to read one paper in both of those ways.

The overall message of the conference was extremely positive; the point made by lots of speakers (including myself) was that experimental, empirical, and a priori philosophy are all part of the same project, and that these methods can usefully complement one another. Thus while aspects of a question may require empirical or experimental philosophy- as in the case of moral disagreement, for example- other aspects of these questions may require more straightforwardly a priori inquiry — such as the question of whether expressivism is a conceptually coherent view.

Slides of Dave Chalmers’ talk should be available on his website soon, although they were not at the time of this writing. Those of you with access to GoBleen will be able to see the slides from my talk, ‘Moral Disagreement: Empirical and Conceptual Issues.’ In the meantime, feel free to use this space to discuss: what is at issue between experimental philosophers and their a priorist counterparts? Can’t we all just get along?

Advertisements

23 Responses to x-phi + a-phi = phi

  1. Dustin says:

    “A distinction was also drawn between experimental philosophy and empirical philosophy. My work on moral disagreement was cited as an example of the latter, since I didn’t actually do any of the experiments myself.”

    So the distinction is over whether the philosopher actually performed the experiment she’s citing herself? How is this an important distinction?

    “What is the issue between experimental philosophers and their a prioristic counterparts?”

    Who are their a prioristic counterparts? It seems to me that the VAST majority of analytic philosophy falls squarely within the ’empirical philosophy’ camp: most analytic philosophy is done from the armchair indeed… with stacks of ’empirical’ articles sitting on the coffee table.

  2. nate charlow says:

    Dustin, the distinction does seem to be a sociological fact, which is some evidence of its significance. The fact that Knobe et al. actually gather their own data seems to be an interesting and fairly unique feature of their work. (I think we can agree about that without committing ourselves to the special value of that work, which may be what you’re worried about.)

    I don’t know what counts for you as an “’empirical’ article” but it seems dubious to say that “the VAST majority of [current? historical?] analytic philosophy” pays sufficient attention to empirical work. It’s impossible to evaluate your claim directly, but here are some data points. Alex’s work identifies an entire line of inquiry in analytic philosophy that rests on what appears to be a false empirical assumption. Speaking of what I know best, a lot of what goes on in philosophy of language would benefit from a more careful consideration of related work in linguistics (here it’s not necessarily that the philosophical work is relying on dubious empirical assumptions, but that the philosophical work isn’t explicit enough about what its empirical commitments are).

  3. alexandra says:

    yes, it’s an interesting question what exactly counts as an ’empirical article’. now, one thing to think is that any article that makes use of intuition is empirical, since the author is reporting an empirical fact about him or herself. but this can’t be right, because then all of philosophy would be empirical work! also notice that ’empirical’ and ‘not philosophical’ are not co-extensive; work in logic and math is non-philosophical, perhaps, without being empirical.

    however, i think dustin is right about this: traditionally, philosophers have not drawn a line between work in, say, ethics or perception or conceptual content and facts about human psychology. when one reads hume or hobbes or nietzsche or plato or aristotle, it seems that these authors are interested to a large degree in human nature, which is of course an empirical field of study. it’s an interesting historical question when, exactly, the assumption began that empirical work in human psychology was irrelevant to philosophy. (if indeed this assumption exists; speaking personally, i know that i am sometimes asked, when reading a psychology article, what it could possibly have to do with philosophy, but on the other hand, i suspect few philosophers would want to deny outright that empirical research has any potential relevance to philosophy.)

    i do share dustin’s skepticism about the value of a distinction between ‘experimental’ and ’empirical’ philosophy. knobe asked people for their intuitions about intentional actions in a way that no one in psychology had thought to do before; in these kinds of experiments, it may be useful for the philosopher to carry out her own experimental work if no psychologists are asking the questions she needs answered. on the other hand, someone like myself may find ample experimental evidence in experimental economics, social and cultural psychology, etc., and so may not feel it necessary to carry out my own experiments. having said that, it’s entirely possible that i may one day come up with a question that i need to carry out my own experiments to answer!

    finally, the slides of dave’s talk are available on his website, if you’d like to check them out.

  4. simon cullen says:

    If your readers wanna get a taste of the conference talks, I’ve put a few of the recordings which worked out on my website: http://discursively.org/epca/ . It was nice meeting you! Simon.

  5. dtlocke says:

    “The fact that Knobe et al. actually gather their own data seems to be an interesting and fairly unique feature of their work.”

    Philosophically interesting?

    I think that, when conducted properly, the empirical work can definitely be of use. I think that, yes, philosophers have given arguments that have rested (in part) on false empirical claims (so have scientists, but no one is criticizing them for not being empirical enough). My only trouble is with the suggestion that empirical philosophy is something importantly different, rather than an extension of what philosophers have already been doing.

  6. Alex,
    First, I am jealous I couldn’t make it to the party down under. I was also particularly surprised to see you with a glass of wine in your hand! Who woulda thunk it? In any event, I just posted a link to the paper that Eddy and I wrote that explores the various projects within experimental philosophy. Some of the readers of this blog might find it useful. If you’re interested, wander over to the X-Phi blog and take a peek!

  7. alexandra says:

    simon: thanks for the link! it was great to meet you as well, and i look forward to crossing paths again soon.

    thomas: yes, experimenting is hard work, and the key thing is to remain hydrated!

    dustin: yes, scientists have made false empirical claims. but notice that since such claims are routinely tested, they are often discovered to be false in a much shorter time period than philosophical claims. so, it wasn’t until Doris’ work on social psychology and virtue ethics came along that we had good empirical evidence that some of the claims made by certain virtue ethicists were empirically false. in addition, many philosophers take their own intuition to be all the evidence required to substantiate a claim. thus Richard Boyd feels perfectly confident asserting that “agreement on non-moral facts… would eliminate almost all moral disagreement” and does not feel compelled to provide any evidence at all in favor of this claim! bad philosophy? unlikely, since this reliance on one’s own intuition characterizes philosophers such as Kripke, Anscombe, Stroud, and many others. so all in all, pretty par for the course.

    let me also say that, speaking for myself, i don’t have much stake in the claim that what i am doing is part of some radical break with traditional philosophers. i suspect that some of the ‘x-phi’ crowd share my feeling that i would be completely content to simply do ‘philosophy’ full stop. and yet i am often asked how my work constitutes philosophy given its reliance on empirical and interdisciplinary data. so it’s not that i am trying to argue that my work, and that of the experimental philosophers in general, constitutes some new and radically different way of doing philosophy. on the contrary, joshua knobe made the point several times at the conference that what we are doing represents a return to traditional philosophy, which was concerned primarily with questions about human nature and its place in the world. rather, the question of whether and in what way this is philosophy at all is one typically raised by detractors of experimental philosophy.

    sounds like we agree more than we disagree here, which i couldn’t be happier about!

    also please do check out dave’s blog and website for comments on the conference and slides, respectively.

  8. dtlocke says:

    Alex,

    Yes, I do think we agree more than we disagree, and I’m happy about that too. But perhaps here is something important that we disagree about:

    “many philosophers take their own intuition to be all the evidence required to substantiate a claim. thus Richard Boyd feels perfectly confident asserting that “agreement on non-moral facts… would eliminate almost all moral disagreement” and does not feel compelled to provide any evidence at all in favor of this claim! bad philosophy? unlikely, since this reliance on one’s own intuition characterizes philosophers such as Kripke, Anscombe, Stroud, and many others. so all in all, pretty par for the course”

    Does Richard Boyd really think that his intuition is all that’s needed to establish the claim that “agreement on non-moral facts… would eliminate almost all moral disagreement”? If so, then you’ve got your example, but you’ve also got the exception that proves the rule, for that would be very BAD philosophy indeed. Notice what the object of the Boyd’s alleged intuition is. It is the proposition that

    1) Agreement on non-moral facts… would eliminate almost all moral disagreement.

    But is that something people really rely on intuitions to determine? I don’t think so. I suspect that Boyd is really making something of an inference to the best explanation of apparent moral disagreement—he finds it more likely that apparent moral disagreement is due to disagreement on non-moral issues than genuine moral disagreement. What evidence does he have? His life experiences of course. Is this very compelling evidence? Of course not. Should we do some empirical research? Of course we should. Would Boyd consider the data yielded by this research (if collected properly) as relevant to his claim? Of course he would, or, in any case, he should.

    But most importantly: would it be “par for the course” for him to think his intuitions settled the matter? Of course not! You argue that it would be par for the course by citing “Kripke, Anscombe, Stroud, and many others” as examples of philosophers who rely on intuitions, rather than conducting empirical research. Yes, there are many philosophers that rely on intuitions to settle certain matters. But WHICH claims are being supported by those intuitions in the case of these other philosophers? Are they claims like “Agreement on non-moral facts… would eliminate almost all moral disagreement”, or are they claims like “Aristotle (counterfactually) might not have taught Alexander the Great”?

    It seems to me that there’s a pretty big difference. The first claim is a straightforward empirical claim about human psychology. Thus, to find out whether it’s true, empirically study human psychology. But how, exactly, are we to empirically determine the truth of the second claim? Is there some research that Kripke can be faulted for not doing? (Keep in mind here the relevant sense of “might”.)

  9. Shen-yi Liao says:

    If Kripke’s aim was to explain how people use proper names to refer, then he might look at how people actually use proper names to refer. If I remember correctly, I think in N&N he does mention that he has asked a few people about the Godel/Schmidt case, so it is not as if he does not recognize the empiricity (is that even a word) of the claim, but a systematic study might have been useful. I would think that is the research that he could be faulted for not doing.

  10. dtlocke says:

    Hi Sam.

    Would you mind being a little more specific about the “systematic study” you have in mind? How would it be carried out? How does one “look at how people actually use proper names to refer”?

    Notice also that the content of Kripke’s intuition was not “people use proper names to refer in such and such a way” (i.e., not in the way they would if a certain sort of decriptivism were true). Rather, the content of his intuition was

    (*) Aristotle might not have been the teacher of Alexander the Great”.

    From (*) and other such claims, Kripke REASONED TO the conclusion that the description theory (or a particular version of the description theory) of reference was incorrect. How is Kripke supposed to empirically test (*)?

  11. Shen-yi Liao says:

    Dustin: Hi. I think the paper I linked to shows at least one experiment one could run. I am not sure if this is the job of philosophers or psychologists. Seems to me to be a job of semanticists who record many patterns of usage of proper names and infer to the most plausible theory and then test them, et cetera. But I am quite ignorant of what linguists do, so perhaps someone like Nate can correct me on this.

    I think the Aristotle case is a difficult one, and I am not sure if I have anything to say about it right now, but here is what Kripke says about the Godel/Schmidt case:

    “So, since the man who discovered the incompleteness of arithmetic is in fact Schmidt, we, when we talk about ‘Godel’, are in fact always referring to Schmidt. But it seems to me that we are not. We simply are not.” (84)

    I am not sure there is any reasoning there, besides the intellectual seeming (“intuition”) highlighted. It seems that the intuition he presents here is a pretty straightforward claim

    ($) “Godel” does not refer to Schmidt.

    If everyone agrees with ($), then that is pretty good evidence against descriptive theory. If many people do not agree, then it is not good evidence. So before we can build a theory out of the evidence, wouldn’t it be prudent to see if the evidence is any good?

    So I guess here is what I want to say about the Aristotle case. I assume when you say reason, you have in mind inference to the best explanation rather than logical deduction. In which case, we should try to make sure we have sufficient data to make that inference.

    I mean, here is one way to understand what Kripke is doing that seems pretty in line with what is done in other areas. He collected some preliminary data, his own intuitions, and devised a hypothesis from that. Then other people are free to collect their own data an show whether the preliminary data were representative or not. And we decide whether to accept the hypothesis, as a result.

  12. alexandra says:

    in ‘from metaphysics to ethics’ (pp 36-37) frank jackson writes:

    “i am sometimes asked… why, if conceptual analysis is concerned to elucidate what governs our classificatory practice, don’t i advocate doing serious opinion polls on people’s responses to various cases? my answer is that i do– when it is necessary. everyone who presents the gettier cases to a class of students is doing their own bit of fieldwork, and we all know the answer they get in a vast majority of cases. but it is also true that often we know that our own case is typical and so can generalize from it to others.”

    this last sentence, i think, is exactly the assumption that experimental philosophers are concerned to dispel. because, as weinberg, nichols and stich discovered (see their paper ‘metaskepticism: meditations in ethnoepistemology’), our case might be typical for university students in the USA and Europe, but it is in fact NOT typical for students in East Asia. so either philosophical concepts are local somehow- a conclusion jackson may or may not be happy with- or we can’t generalize as often as we thought.

  13. nate charlow says:

    Dustin, the argument you cite is indeed one of the arguments you find in N&N. There are some others as well. Including arguments in which Kripke appears to directly intuit that we are referring to so-and-so, and then uses this to reject any theory of names that implies that we are not referring to so-and-so. Sam mentions one in his reply to you.

    Concerning your question, Kripke’s causal theory of reference clearly makes empirical assumptions (which seem to be false) about the character of our referential practices.

  14. Dustin says:

    Hi Sam. You write:

    “It seems that the intuition he presents here is a pretty straightforward claim

    ($) “Godel” does not refer to Schmidt.

    If everyone agrees with ($), then that is pretty good evidence against descriptive theory. If many people do not agree, then it is not good evidence. So before we can build a theory out of the evidence, wouldn’t it be prudent to see if the evidence is any good?”

    Huh? If the question Kripke is trying to answer is “what do people BELIEVE the name ‘Godel’ refers to?”, then, yes, surveying people to find out what they believe is a good idea. But that is not Kripke’s question. His question is “what DOES the name ‘Godel’ refer to?” How is asking people whether they BELIEVE the term refers to Schmidt (in the case described) relevant to answering that question?

    Now, as far as RHETORIC goes, Kripke would do well to find out if people agree that ‘Godel’ does not refer to Schmidt (in the case described). This, I take it, was his aim in “asking a few people about the Godel/Schmidt case.” He simply wanted to know if people agreed with a premise of his argument. If they did not, there would be no use in assuming it. He would either need to not rely on ($) or argue for it first. But if they did agree, he could simply assume it… as he did, because they did.

    It’s important to keep in mind that ($) here is Kripke’s premise, and not

    ($`) People agree that “Godel” does not refer to Schmidt.

  15. Dustin says:

    Hi Nate. You write:

    “Concerning your question, Kripke’s causal theory of reference clearly makes empirical assumptions (which seem to be false) about the character of our referential practices.”

    That sounds right. But could you give me a specific example of an assumption that Kripke’s theory makes that he can be faulted for not having empirically tested?

  16. Dustin says:

    Oh, and I guess it should be one that he relied on intuition to establish, since that’s what at issue here.

  17. Shen-yi Liao says:

    A point of clarification. I did not have in mind a survey asking people what they BELIEVE the name ‘godel’ refer to. That would be akin to asking them what theory of reference they subscribe to, and they are probably not at all aware of the answer. Instead, you look at how people in fact use names. You present them with the same cases that Kripke uses to draw his intuition from, and you survey their responses. That is what I gather Mallon, et al have done. Insofar as reference is part of our linguistic practice, rather than some out-there fact, surveying people’s linguistic practice seems like a sensible first step to figure out their implicit theory of reference. That is, a first step in figuring out what DOES the name ‘Godel’ refer to.

    Re: “[Kripke] simply wanted to know if people agreed with a premise of his argument. If they did not, there would be no use in assuming it. … But if they did agree, he could simply assume it… as he did, because they did.” I think Alex’s most recent comment, the one containing a similar-sounding Frank Jackson quote, serves as a response. Kripke’s argument crucially relies on ($), and he does not argue for ($). What empirical studies might show is that he does not have the right to generalize from his own intuition and assume the truth of ($).

  18. Dustin says:

    Re: “Kripke’s argument crucially relies on ($), and he does not argue for ($).”

    Yes. All non-circular arguments rely on some assumptions that are not argued for.

    Re: “What empirical studies might show is that he does not have the right to generalize from his own intuition and assume the truth of ($).”

    I’m sorry, but I just don’t understand what you mean here. If Kripke “generalized from his own intuition” the conclusion of his generalization would be “everyone (most everyone) has the intuition that ‘Godel’ does not refer to Schmidt”. But that is NOT a premise of Kripke’s argument. The premise of his argument is ($), not ($`) or even

    ($“) Everyone (most everyone) has the intuition that ‘Godel’ does not refer to Schmidt.

    Re: “Insofar as reference is part of our linguistic practice, rather than some out-there fact, surveying people’s linguistic practice seems like a sensible first step to figure out their implicit theory of reference. That is, a first step in figuring out what DOES the name ‘Godel’ refer to.”

    Two worries:

    Minor worry: how would such a survey work?
    Major worry: what’s with the “that is”?

    Regarding the major worry, since when does ‘their implicit theory of reference’ determine what ‘Godel’ refers to? You seem to be making a pretty big assumption about reference, namely, that ‘their implicit theory of reference’ determines reference. To assume THAT begs the question against Kripke’s causal theory of reference. Also, how are we to empirically determine whether that assumption is true? Empirical research? How would such research go? Would it not rest on that, or some other assumption about reference?

  19. […] Experimental Philosophy Meets Conceptual Analysis Jump to Comments Are experimental, empirical, and a priori philosophy all part of the same project? Check out the discussion at Go Grue! […]

  20. alexandra says:

    those of you who are interested in this discussion might also want to check out the discussion going on at

    experimentalphilosophy.typepad.com (this should be a link but who knows?)

    where edouard machery (incidentally, one of the authors of the godel/schmidt study) has a really interesting post about what the project of experimental philosophy is and is not, and about the role of empirical claims in philosophical debate. highly recommended reading.

    will attempt a more substantive comment when i am more fully caffienated (sp? more coffee!)

  21. Dustin says:

    Here is Knobe’s description of the Godel/Schmidt study and its relevance to philosophy:

    “Readers of this story almost universally agreed that the word ‘Gödel’ did not, in fact, refer to Schmidt. Any theory which declared Schmidt to be the referent of ‘Gödel’ was therefore assumed to be incorrect.
    A question arises, however, as to whether everyone shares this intuition or whether it is only shared by the kinds of people who normally read Anglo-American philosophy. The philosophers Edouard Machery, Ron Mallon, Shaun Nichols and Stephen Stich recently ran an empirical study to address this question. All subjects were given the story of Gödel and Schmidt. But the design of the study included a surprising twist. Some subjects were Americans; others were residents of Hong Kong. As expected, the American subjects shared the intuitions of most analytic philosophers. But the Hong Kong subjects showed a quite different pattern of responses. Among subjects in Hong Kong, the majority said that the word ‘Gödel’ did indeed refer to Schmidt.
    This recent result — along with similar results from studies in ethics and epistemology — suggest that Asian people may not share many of the intuitions on which widely accepted philosophical theories have been based.”

    Even IF (and this is a pretty big if) this study shows that not everyone shares Kripke’s intuition, how is this relevant to Kripke’s argument. Again, his premise is not

    $“) Everyone has the intuition that ‘Godel’ does not refer to Schmidt.

    Nor, importantly, is the premise of Kripke’s argument:

    $“`) Kripke has the intuition that ‘Godel’ does not refer to Schmidt.

    The premise of his argument is

    ($) ‘Godel’ does not refer to Schmidt.

    What I want to know is this: how is the (alleged) falsity of ($“) relevant to the truth/falisty of ($)?

    Here’s one possibility: the (alleged) falsity of ($“) is evidence that Kripke’s intuition that ($) is true is unreliable. Thus, since Kripke’s belief that ($) is true is based on his intuition that ($) is true, his belief lacks justification (warrant, evidence, what have you). If this is the answer one wants to defend, then here is one’s challenge:

    (1) Why is the (alleged) falsity of ($“) relevant to the reliability of Kripke’s intuition and
    (2) Why is the reliabilty of Kripke’s intuition relevant to whether he is justified in holding a belief based on it?

    Does it just SEEM to you that the (alleged) falsity of ($“) is relevant to the reliability of Kripke’s intuition? Does it just SEEM to you that the reliability of Kripke’s intuition is relevant to whether he is justified in holding a belief based on it?

    It SEEMS to me that this process must stop somewhere. So the question is where? Where Kripke stopped or where the x-phiers stop? I suggest that it is where Kripke stopped, because I find the claim that ‘Godel’ does not refer to Schmidt much more intuitively plausible than the assumption that the (alleged) falsity of ($“) is relevant to the reliability of Kripke’s intuition and the assumption that the reliability of Kripke’s intuition is relevant to whether he is justified in holding a belief based on it.

    I should say that I do in fact find those other assumptions pretty darn intuitively plausible. And I don’t think I need to reject one of them quite yet, because I don’t think that the study actually shows that ($“) is false–there’s just too many other (plausible) explanations for the results.

  22. Steve C. says:

    Hmm… I have sympathies on both sides here. Dustin, it seems like Sam is making a reasonable point. Take

    ($*) Everyone has the intuition that ‘Godel’ refers to Schmidt.

    Clearly ($*) is false. But do you agree that if ($*) were true, it would speak against the truth of ($)?

    If so, then consider

    ($**) Everyone (with the exception of Kripke, three people Kripke surveyed, and Dustin Locke) has the intuition that ‘Godel’ refers to Schmidt.

    This is also false, of course. But if this were true, and established by empirical research, then it seems like that would also speak against the truth of ($), and that the five of you should distrust your intuitions. If all that is correct, then it seems like empirical data about people’s intuitions is indeed relevant to our consideration of ($). Mind you, the data might vindicate Kripke’s claim. Does that sound reasonable?

    Of course, based on what I’ve heard about the study (which I haven’t examined directly, I confess), like you Dustin, I’m not yet convinced that Kripke needs to recant his assertion of ($). (Is anyone claiming that he should??) If divergence of intuitions about ($) moves along (rather than cutting across) cultural lines, then might different notions of reference be at work? If so, the study may actually end up supplying confirmation of ($) since the relevant cultural group did agree with ($). In that case, perhaps Kripke was justified in making the assertion with little more than his own (strong) intuition and a tiny bit of informal polling. I’m not so much arguing for this, as I’m entertaining it. I’d be curious to hear people’s thoughts.

  23. Steve C. says:

    Ok, this comes from the intro of the Machery/Mallon/Nichols/Stich study:
    “These results constitute prima facie evidence that semantic intuitions vary from culture to culture, and the paper argues that this fact raises questions about the nature of the philosophical enterprise of developing a theory of reference.”

    Sounds reasonable.

%d bloggers like this: