Wednesday, July 30, 2008

The Linguistic U-Turn

In the early to mid-20th century, anglophone philosophy famously took a "linguistic turn". Debates about the mind were recast as debates about the terms or concepts we use to describe the mind; debates about free will were recast as debates about the term or concept "freedom"; debates about knowledge were recast as debates about how we do or should use the word "knowledge"; etc. Wittgenstein famously suggested that most philosophical debates were not substantive but rather confusions that arise "when language goes on holiday" (Philosophical Investigations 38); diagnose the linguistic confusion, dissipate the feeling of a problem. J.L. Austin endorsed a vision of philosophy as often largely a matter of extracting the wisdom inherent in the subtleties of ordinary language usage. Bertrand Russell and Rudolf Carnap set philosophers the task of providing logical analyses of ordinary and scientific language and concepts. By no means, of course, did all philosophers go over to an entirely linguistic view of philosophy (not even Austin, Russell, or Carnap), but the shift was substantial and profound. In broad-sweeping histories, the linguistic turn is often characterized as the single most important characteristic or contribution of 20th century philosophy.

I view the linguistic turn partly sociologically -- as driven, in part, by philosophy's need to distinguish itself from rising empirical disciplines, especially psychology, as departmental and disciplinary boundaries grew sharper in anglophone universities. The linguistic turn worked to insulate philosophical discussion from empirical science: Psychologists study the mind, we philosophers study in contrast the concept of the mind. Physicists study matter, we study in contrast the concept of the material. "Analytic" philosophers could thus justify their ignorance of and disconnection from empirical work.

This move, however, could only work when psychology was in its youth and dominated by a behaviorist focus on simple behaviors and reinforcement mechanisms (and, even earlier, introspective psychophysics). As psychology has matured, it has become quite evident -- as indeed it should have been evident all along -- that questions about our words and concepts are themselves also empirical questions and so subject to psychological (and linguistic) study. This has become especially clear recently, I think, with the rise of "experimental philosophers" who empirically test, using the methods of psychology, philosophers' and ordinary folks' intuitions about the application of philosophical terms. (In fact, Austin himself was fairly empirical in his study of ordinary language, reading through dictionaries, informally surveying students and colleagues.)

A priori, armchair philosophy is thus in an awkward position. The 20th-century justification for analytic philosophy as a distinct a priori discipline appears to be collapsing. I don't think a prioristic philosophers will want to hold on much longer to the view that philosophy is really about linguistic and conceptual analysis. It's clear that psychology and linguistics will soon be able to (perhaps already can) analyze our philosophical concepts and language better than armchair philosophers. Philosophers who want to reserve a space for substantive a priori knowledge through "philosophical intuition", then, have a tough metaphilosophical task cut out for them. George Bealer, Laurence BonJour, and others have been working at it, but I can't say that I myself find their results so far very satisfying.


  1. I just started re-reading Russell's An Outline of Philosophy and after a few chapters I came to exactly this conclusion and thought of your work in particular as rising out of the need that Russell identifies at the beginning of the work.

    I don't need to remind you that Russell said specifically that the place for philosophy that he was defining was a limited one based on the lack of advancement in psychology--specifically it seems developmental psychology--to that time.

    I don't think that Russell would be too horribly convinced by your argument here that analytic philosophy was doomed though. Analytic philosophy is as much about the quality and form of the inquiry as it is its subject matter. I think Russell would have simply moved over to some of the newer questions opening up under the weight of biological, chemical, and physical scrutiny.

    He may have also feltcompelled to put more emphasis on morality and aesthetics as worthy of philosophical much out of professional interest as intellectual necessity.

  2. Nicely written. Since you've tagged it 'metaphilosophy', i.e. philosophical reflection about doing philosophy, and since you're emphasizing empirical, scientific method: shouldn't this view rather emerge as conclusion of empirical research, and count as true or false only after empirical verification/falsification? (Empirical sociology, in this case, e.g. studies that show how the reputation of philosophers had declined, and how that trend had been reversed after turning 'linguistic', or something like that.) Otherwise it would itself look a bit like armchair (meta-)philosophy.

  3. Jimpanzee: I'd forgotten Russell had said that, thanks for the tip!

    Leif: Yes, it is empirically grounded -- in the sociology of academia and in facts (both psychological and sociological) about what psychologists seem to do well.

  4. Some good points here about phil. of psych (although I have my doubts about "experimental philosophy" so-called, but that's for another day).

    This "division between fields" issue came up a lot when I started looking into recent philosophical research into pain. Some of it is highly empirically informed (Valerie Hardcastle's The Myth of Pain, for one), while other stuff is maddeningly "conceptual" in the worst of senses. And with regret, I point you to some of the papers in Murat Aydede's anothology Pain.... parts of that book are excellent, but other parts go down very nitpicky linguistic dead ends, largely ignoring decades of (empirical!) research on the very subject coming out of psych and medical departments.

    Anyone doing "philosophy of perception"? I'm sure the same problems crop up there too.

  5. Great post! Two quick points of dissent, though. (i) I think Russell was one of the leading advocates for a naturalized study of language and the mind. This is one of the reasons why he was so opposed to the later Wittgenstein and people like Austin. See, for example, the end of Inquiry: "Some modern philosophers hold that we know much about language, but nothing about anything else. This view forgets that language is an empirical phenomena like another [sic], and that a man who is metaphysically agnostic must deny that he knows when he uses a word. For my part, I believe that, partly by means of the study of syntax, we can arrive at considerable knowledge concerning the structure of the world."

    (ii) I think we can still hold out some hope for a priori knowledge, but just not a priori philosophical knowledge. Mathematics and logic are hard to make sense of without some a priori basis.

  6. You might be right that conceptual analysis is a task better pursued by psychologists, linguists, and experimental philosophers than by philosophers. But isn't there some room for the a priori in two other Carnapian projects? I have in mind: (1) explicating (as opposed to analyzing) terms from ordinary language and other disciplines, and (2) axiomatizing and reducing existing theories in order to render their ontologies more transparent.

    It's not clear to what extent (1) and (2) can be pursued without any specialized empirical knowledge. But certainly there are some cases - interpretations of probability, the formalizations of modal and counterfactual sentences, the semantic paradoxes - in which philosophical explication can do some good without coming into any significant with non-common empirical knowledge. And maybe even Aufbau-type projects, which would fall under (2), can be pursued at least to a certain point from the armchair. There is also a clear sense in which axiomatizing and reducing specific scientific theories, a la Hartry Field, is an a priori activity. I imagine that's not where your main beef lies, though, since that does require specialized empirical knowledge.

  7. Check my current thoughts on about Experimental Philosophy as Anthropology.

    I think that philosophy has been amazingly ingrained in the history of ideas, and hasn't much bothered to note that we are bodies: body in interaction with others' bodies - children of m/others whom we "join" and study their presentation of the world in the context of what i call the Question-Response System.

    We have vastly underestimated the human, and it may be time to study how we "measure" the world, in order to begin to appreciate how complicated is the human body: faces, hands, out-of-balance postures requiring so much self-study.

    We "emerge" from relationships with our m/others to become the individual "self/I" - and...

    Beyond dualism, beyond the limiting ideas of Plato et al, we become more interesting to ourselves.

    Harvey Sarles

  8. You and I (and others here) have had the debate about the main substance of this post a number of times now, so I'll set it aside today and just ask about Williamson.

    I'm surprised to see him on the list of people who are looking for a substantive, independent, a priori investigation into philosophy via a faculty of philosophical intuition. As I read Williamson, he thinks that philosophy is continuous with the other sciences, that there is no special faculty of philosophical intuition or special philosophical subject matter, and that there is no apriority.

  9. Until we do not see mathematics, logic or formal disciplnes as "natural" or empirical investigations, and then make sense of how its "knowledge" is produce via empirical questions, that corner of philosophical theorizing will still spring and blossom.

    On the other hand, i believe there is room for everyone. Even for more empirically oriented individuals they have to launched hipothesis or design experiments under working assumptions to prove the null hypothesis. It is in this phase of their work where a little bit of armchair suppossitions is at work. Perhaps we have to concede this small and reserve space for that kind of philosopher.

    Neverthless, the core message and warning of Eric, the old methods are collapsising, it is essentially true.

  10. Thanks for the great comments, folks!

    Holyoke: It does seem to me that there's a similar division in the philosophy of perception, and I'd evaluate the work more or less analogously -- though I always think there's room for really cool a priori conceptual philosophy (e.g. David Lewis), which if nothing else displays an interesting conceptual network.

    Chris: You may be right; I can't claim to be a Russell expert! I'm not sure the quote *quite* shows that Russell would welcome empirical investigation of language as a key to metaphysics; and I would part ways with him to the extent he things we learn any kind of robust metaphysical facts about the world by studying language. (As I've argued in previous posts, I think we learn about our minds and our concepts.)

    Iolasov: You are absolutely right, and this post, because the ideas are so compressed, makes it seem as though I'm farther from Carnap than I really am. I do think that "explication" is one of the major contributions philosophers can make (and I myself offer an explication of the concept of belief in my 2002 paper on belief); but I also think that explication generally has to be thoroughly empirically informed, and that it's often better done by the scientists using the terms than by philosophers. One exception you note is mathematical concepts; I'd add moral concepts. In neither case are substantive claims about the world, at stake, I think -- at least not in quite the same way as with most concepts; I do both areas are more empirical than is generally supposed. I'm not as sure of the value of axiomatization....

  11. Harvey: Thanks for the interesting comments and the link!

    Jonathan: You're right, of course. Williamson is a more complicated case, and I've now axed his name from the company of Bealer and BonJour. I'll probably do a separate post on Williamson before long.

    Thanks for the kind remarks, Anibal!

  12. Interesting post. Your quasi-sociological description of what motivated the linguistic turn sounds like Rorty's: "The idea [behind the Ling Turn] was to mark off a space for a priori knowledge into which neither sociology nor history nor art nor natural science could intrude... The replacement of ‘mind’ or ‘experience’ by ‘meaning’ was supposed to insure the purity and autonomy of philosophy by providing it with a nonempirical subject matter."

    Rorty further explains that two crucial premises lay behind the attempt to replace experience or consciousness or mind with language: "First, the two terms had an equally large scope—both delimited the entire domain of human inquiry, of topics available to human study. Second, the notions of ‘language’ and ‘meaning’ seemed, at the beginning of the century, immune to the naturalizing process." The second crucial premise failed to hold out and by the time Quine was finished naturalizing epistemology philosophy had lost its sense of purpose. So says Rorty.

    So it sounds like your view of the matter has the considerable merit of being in keeping with one of philosophy's best provocateurs.

  13. I'd be interested in seeing how you do separate out this view from Rorty's Eric.

    I was also surprised to see Williamson in the list and look forward to your post on his views. He's one of those folks I love reading and tend to agree with the views he's arguing for but often have trouble with his arguments themselves. But I find myself always reading blog posts that discuss him.

  14. Thanks for the comments and encouragement, Colin and Clark! I've read some of Rorty's work on the linguistic turn (such as the intro to his volume of that title), but that was several years ago now, and I didn't realize how much I was echoing him. I'll have to go back and revisit that. I do part ways with Rorty on truth and on his positive vision for philosophy. I'm baking some ideas now for a post on what the role of a philosopher should or could be, if it's not to discover a priori truths. Here I'll be less relativist and more scientific realist than Rorty.

    Williamson I see as a skilled and sophisticated defender of 20th-century armchair philosophy, though one who wisely adds caveats and compromises in important places. As I mentioned, more on him in a future post!

  15. Great stuff, all.

    I wonder to what extent analysis and explication just are empirical. To voice some Quinean concerns (certainly also in the spirit of Rorty), why think that any analysis or explication (or axiomatization, for that matter) is immune to empirical revision? The very process of figuring out just what the correct or intuitive or most elegant analysis amounts to is inexorably intertwined with empirical data. To put things in more baldly Quinean terms, I don't see how one can avoid this empirical intertwining without a well-justified analytic/synthetic distinction. No doubt I'm sympathetic with the claim that there isn't one (NOTE: not that there cannot be one), but I wonder what might be said in support. Recall that 'well entrenched' does not mean 'analytic'; the trouble is spelling out a real difference here.

    I think analysis (yes, even in the armchair!), explication, axiomatization, etc. have a place in philosophy and science, but we need to recognize that they are not immune to revision, that counter-arguments to "analytic" or "a priori" claims are not "incoherent" or "mere subject changing" (at least not without further evidence), and that when we can, we should bring empirical evidence and the theoretical "values" of simplicity, successful prediction, elegance, etc. to bear on everything we do in philosophy. That leaves a reasonable (and important) space for the clarifying methods of the linguistic turn--looking at language is better than "peering inward" or "accessing the forms"--without liscensing a philosophical safe-haven for old-fashioned armchair intuition mongering.

    (Don't know where all this heady rhetoric is coming from! Must be the tropical storm we're having today in Houston. ;>) )


  16. I have a reply to Josh's comment, but it got too long and irrelevant for this thread.

    it is.

  17. I agree with you completely, Jonathan. Rhetorically, though, I downpedal the Quine: I think it's still worth saying that there is an analytic-synethetic distinction -- it's just that what we choose to treat as "analytic" is up for revision in light of empirical evidence. That is, I can if I want for scientific or philosophical purposes define "hot" as following analytically from "star" -- but then it's a pragmatic (i.e. empirical-cum-normative) question whether that's a good thing to build into one's definition of a star.

  18. iolasov--

    I'm not sure why the stronger forms of explication are needed to make my point. Aren't the weaker forms enough to vindicate Quine? The question, as I see it, is what's the force, in philosophical debate, of claiming that something is analytic? If it's just to say, "I hereby stipulate that by 't' I will mean 't*'", then why should anyone else care that you've done so? The interest we have in a particular explication is whether or not it captures what we take ourselves to mean in particular contexts, or if it captures something of theoretical importance. And justifying those claims, I contend, involves empirical evidence, broadly construed to include the theoretical "virtues." So stipulating is fine, but to engage anyone else in discussion, you need to step back into the empirical world. Stipulating is just not enough to philosophers to shut up! (Is anything? ;>) )

    But in any event, if every explication is theory-relative, and theories are empirically revisable, isn't that enough to undermine strong claims of analyticity? Or did I miss something (likely, due to my tin-ear for this stuff!)?

    I do, as I said, think that arm-chair reflection has a place in philosophy. I think that our everyday mental terms, for example, get their meanings from their place in folk-psychological theory. And given that I am one of the folk, I can consult my own intuitions about what that theory says. But I do not have infallible access to this theory--it is largely implicit and "below the surface." And folk theory itself is open to revision. Our folk meanings change over time in response to new knowledge, fashions, social dynamics, etc. So I take it that I am a source of evidence concerning the meaning of my mental state terms, but a defeasible one. I might stipulate that by 'consciousness' I mean 'object of higher-order thought.' And I might even claim that this is obvious to me given my reflection on possible scenarios. But that settles nothing, in my opinion, in consciousness studies, alas. Empirical work, at both the "conceptual" (read: 'theoretical') and experimental level, mixed with the usual constraints of elegance, predictive success, simplicity, etc., is required to say something of interest to other philosophers and scientists.

    Eric: Now there's an a/s distinction that even a Quinean can love!


  19. Thanks for that comment, Josh! I think we're more or less in agreement -- and maybe not so far from iolasov, either, taking a broad view.

    I do hope Quineans love my modest take on the a/s distinction! I wouldn't mind if Carnapians loved it too.

  20. I agree that we're in broad agreement here. I think we agree that what makes a good explication is that it "captures something of theoretical importance" (was it Prior who first called good explicata "good things to mean"?), that for that reason the best explications in most cases must be (among other things) empirically well-informed, and that, at least at a pragmatic "language-building" level, claims to analyticity are always up for revision. I suspect the disagreement is in what we mean by "theory", what it is for theories to be empirically revised, and maybe in our appraisals of how nice it would be if theories could be expressed with fixed definitions for most of their special terms.

  21. Eric,
    I'm sorry I'm coming to your post so late, and now with the comments above, I'm less clear about what has actually been said. Let me, however, make a few comments:
    1. I think it's really very important for a number of the linguistic philosophers that they felt they had found a way to solve philosophical problems, and this certainly was something that was sustaining ordinary language philosophy.
    2. A lot of the work was about the logic or the grammar, and less about what a lot of psychologists mean by concepts. Grammatical investigations are, of course, open to revision, but I'm not sure that, e.g., Pinker's use of examples has any more authority behind it than Ryle's or Austin's. It is after all part of the theory motivating Pinker that competent speakers are going to agree (more or less).
    3. I suspect that a lot of philosophical study of the mind is replete with concepts that aren't going to be investigated empirically because no one but philosophers tries to use them in anything like a rigorous fashion. For example, sadly, philosophy of perception is heavily concerned with hallucinations; that concern motivates ontologies that, I am totally certain, are not going to be found in vision science and, though they are said to connect with what ordinary people think, probably don't.

  22. Just some points.

    The linguistic turn was taken with Frege's context principle as Dummett points out.

    The linguistic turn is over and the philosophy of mind currently dominates as Williamson and Searle have pointed out.

    The experimental philosophers should accept the fact that attacking the armchair is nothing new -ask the naturalists.

  23. Thanks for the comment, Joel! I don't think the linguistic turn has been completely reversed. I think this is a very live issue, right now, at the heart of metaphilosophy -- as evidenced by the debate between those who think philosophy is mostly about concepts (and via them, the world) vs. those who think that philosophy more directly about the world. The armchair method seems much more promising if philosophy is about concepts.

    I agree that in the 20th century many "naturalists" attacked the armchair. But even some of the most eminent naturalists - Quine, Carnap - were still pretty armchair, from the perspective of recent experimental philosophy and cognitive science!

  24. Eric, it seems to me like experimental philosophy is something like putting naturalism to practice.

    I have an issue with what appears to be by some in the experimental community of taking the armchair to be a priori ridden, hence, very unreliable. In fact, I believe, armchair philosophy is synthetic, hence, reliable. Williamson, perhaps Papineau, would concur. (Though, I think, Knobe, Nichols and Kornblith think that intuitions are not a priori but nonetheless unreliable.)

  25. My own take is not that the a priori is unreliable but rather that it doesn't get you much. But much of what we want to say as philosophers isn't a priori. A philosopher claims X. Is X true by definition or as a matter of logical or mathematical fact? Maybe, but then the question is whether anything in the world is well modeled by X-like terminology; which is not an a priori question. Is X true not by virtue of logic alone? Well, then unless it's a very bland truth, with which no one would reasonably disagree (such as that people of both sexes exist), odds are that evidence not available in the armchair will be pertinent to evaluating it.