Is Natural Language Philosophically Relevant? (J. Stanley)

I regularly encounter philosophers who are puzzled about how language could be relevant in shedding light on philosophical problems that are not primarily about language. But I have genuine problems understanding their befuddlement. I don’t think of appealing to facts about language as a special kind of methodology. Rather, I think of it as a source of evidence that is not subject to many of the familiar worries that arise with (say) the methodology of intuitions. Language allows us insight into distinctions to which our explicit theories may blind us. As Austin writes in Sense and Sensibilia, “…the distinctions embodied in our vast and, for the most part, relatively ancient stock of ordinary words are neither few or always very obvious, and almost never just arbitrary.” Hannah Arendt begins her argument that there is a distinction between labor and work in The Human Condition by appeal to the fact that “every European language, ancient and modern, contains two etymologically unrelated words for what we have to come to think of as the same activity, and retains them in the face of their persistent synonymous usage.” And later in the same chapter, she writes “It is language, and the fundamental human experiences underlying it, rather than theory, that teaches us that the things of the world, among which the vita activa spends itself, are of a very different nature and produced by quite different kinds of activities.” I don’t see anything objectionable about these sorts of appeals to language; they can provide legitimate sources of evidence. Furthermore, though philosophers nowadays are more likely to appeal to grammar than to etymology, I also don’t see anything in contemporary philosophy that appeals in language in anything other than this kind of way.

Comments

[Marcus Rossberg]

Might it be that some of these puzzled philosophers are surprised about the authority that some of our colleagues are willing to grant to natural language? One line for demanding caution when taking natural language phenomena as evidence might run as follows: natural languages are riddled with inconsistencies, prejudice, unwarranted reifications, vagueness and other features we would prefer to keep our philosophical theories free of. If one takes natural language as a source of inspiration for philosophical work, one had better be careful not to import these. A whole bunch of (largely American based) twentieth-century philosophers warned against the hazards of (largely British) ordinary language philosophy along these lines while granting that natural language is a valid (and maybe the only) starting point for philosophical inquiry. Another valid concern might be that a natural language phenomenon should not be taken as last word on a matter: philosophy should be critical of the prejudices found in natural language. Revisionary and "counter-intuitive" theories must be admissible.

[Steve Mc Kay]

There are two domains where natural language is clearly a significant source of evidence. First, linguistic analyses and linguistic judgments are relevant data points when trying to build a theory of language. The problem, however, is that many philosophers see linguistic analyses and judgments as the sole sources of evidence in this domain. This is a mistaken assumption, especially when the focus is limited to folk linguistic intuitions in ordinary contexts (in constrast elicited intuitions in experimental contexts). Second, natural language provides important information to achieve some understanding of the various folk theories (e.g., folk-psychology, folk-biology, folk physics) that we intuitively rely upon to make judgments. It is much less obvious that natural language is a store of useful and reliable evidence when it comes to theoretical issues in metaphysics and epistemology. Complex theoretical problems almost always require that we build new concepts and theoretical frameworks that are either only vaguely related or entirely at odds with the folk concepts and intuitions built in natural language expressions. So while natural language might be a source of evidence in these domains, it has to be argued for on a case by case basis.

[Jeffrey C. King]

I think there are somewhat different kinds of appeals made to linguistic phenomena in giving philosophical arguments that are important and legitimate. Sometimes in adopting a position in a different area of philosophy, one incurs commitments in philosophy of language. If these commitments are implausible, that's a strike against one's position. Two examples (there are MANY): 1. (epistemology) claiming that what is expressed by a knowledge ascription varies with context; 2)(ethics) claiming that some sentences don't express propositions (or have truth conditions) but rather "express attitudes".

In arguing for or against claims like these, considerations in philosophy of language need to be in the foreground. (This is not of course to deny that in taking a position in one area of philosophy, including philosophy of language, one may incur commitments in other areas of philosophy that need to be examined.)

[cc o madagain]

As proposed in the last comment, we often take the language that we have inherited to be an expression of views that we have inherited. And so we might be less interested in considering what views we have inherited (the opinions of the many) and more what views we ought to hold. Finding out what views we ought to hold would be alot easier if we didn't have to construct them out of the views we already hold (hence we often work from the hypothesis that thought can proceed independently of language), and hence studying what we already say about things might seem contrary to the goal of philosophy for many.

The rebuttal, of course, is to argue that thought is hemmed in by language. But on that reply, the upshot seems to be that we should study how language works in order to know how thought works - not that we should consider what views we already hold in order to know what views we should hold.

A response to this latter point could be that we can't criticize our views unless we know what they are, and so we should study our ways of saying things in order to do this. And that's a valid point. But at this stage, taking on board both of these points, (that thought is hemmed in by language and that our only resource on what we think is our linguistic practices) it becomes mysterious how we amend our views, as that would suggest an ability to go beyond language which we haven't now allowed ourselves. Hence, perhaps, McDowell's wrestling in Mind and World.

[Barry Lam]

Jason, you say "Rather, I think of it as a source of evidence that is not subject to many of the familiar worries that arise with (say) the methodology of intuitions."

I didn't think there was a distinction between methodologies here. Aren't we using the methodology of intuitions about language whenever we use the methodology of intuitions? God help me if I take seriously my undergraduate student's intuitions about knowledge rather than their intuitions about "knowledge".

[Martin Lin]

C.C. O Madagain says "Finding out what views we ought to hold would be alot easier if we didn't have to construct them out of the views we already hold (hence we often work from the hypothesis that thought can proceed independently of language), and hence studying what we already say about things might seem contrary to the goal of philosophy for many."

But I don't see how it is possible to rationally come to new beliefs without constructing them (in some sense)out of what we already believe. Evidence only tells us how to revise our beliefs given what we currently believe, not absolutely. Indeed, nothing is even evidence at all absolutely.

Is it so wrong for philosophers to have respect for common sense? Philosophy does, after all, have a dismal track record when it comes to loggerheads with common sense. Philosophers have concluded that nobody knows anything, that there are no people, that time is an illusion, etc. Has anyone remained convinced for long?

Physics, to take just one example, is in a different position. Physical theories normally win acceptance after ample empirical confirmation. And when physics has contradicted common sense, it has often resulted in fruitful lines of inquiry and opened the door to future successes. Physics has the bona fides to contradict common sense. Philosophy typically proceeds on a more slender basis. What is there to choose between two rival philosophical theories? Unless one of them is incoherent and the other is not, it is often considerations like simplicity, universality, and elegance that persuade us to accept one or another. Are you willing to abandon common sense on the basis of such considerations? I'm not. Indeed, by my lights, that one philosophical theory requires that we abandon common sense while another preserves it is an excellent reason to reject the former and accept the latter. It is not, for all that, always obvious just what is part of common sense and what isn't. Consideration of what we say and under what circumstances can shed light on that. So consideration of natural language can provide a valuable constraint on philosophy.

[Iain Martel]

Martin Lin makes a good point about the relative track records of science and philosophy when they come up against common sense. But it seems to me that a lot of the problems with relying on natural language as a source of philosophical evidence come up precisely in those cases where common sense and science are opposed. By siding with common sense, as encoded in natural language, philosophy often seems to end up on the wrong side.

In metaphysics this happens all the time. The philosophy of time is a prime example, where a great deal of confusion arises, in my view, from the simple fact that our language requires sentences to contain tensed verbs. Much of the philosophy of causation runs into conflict with science, in large part because of reliance on the ways in which we use causal language, which is riddled with inconsistencies. I suspect that the philosophy of mind is likewise held back greatly by overreliance on ordinary-language based intuitions that hark back to pre-scientific theories of the mind.

In such cases, my sympathies lie far more with those who sought to rationally reconstruct the language of science than with those who looked primarily to our ordinary-language usage. However, it seems that the kind of case Jason Stanley cites is different here: the kinds of phenomena illuminated in these cases are cultural phenomena, where the phenomena under investigation and the language we use to talk about them are much more deeply intertwined. In such cases, presumably both scientists and philosophers should be paying close attention to the nuances of natural language.

[kris]

Jason Stanley writes: 'I regularly encounter philosophers who are puzzled about how language could be relevant in shedding light on philosophical problems that are not primarily about language.'

I smell a straw man here. For one thing, we aren't told who says such things- a good indicator of the presence of a straw man. Now, perhaps I'm being a bit unfair here; after all, Professor Stanley is writing a blog entry and not a journal paper.

Still, I would guess that the philosophers that Professor Stanley is referring to have a more subtle position concerning the value of natural language philosophy than Professor Stanley suggests. Surely every philosopher recognizes that thinking about the everyday meaning and use of philosophical terms is "relevant in shedding light" on any and every philosophical problem. But that's not saying much about whether natural langauge philosophy can take you very far in answering -for example- metaphysical questions.

Perhaps I'm missing something here.

[Ralph Wedgwood]

One thing that is presumably not philosophically relevant is anything that is a purely idiosyncratic feature of English. (I take it that no one will defend the proposition that English provides some unique insight into the philosophical truth that is not provided by e.g. Spanish or Chinese or Quechua!) So philosophers intent on canvassing linguistic evidence for and against philosophical theories would be well advised to consider a range of languages, not just English.

[Martin Lin]

Ian, I would have thought that deference is transitive. If common sense should defer to science and philosophy should to defer to common sense, then surely philosophy should defer to science.

[Eric Schwitzgebel]

I am among those skeptical of ordinary language approaches to philosophy. For some of the reasons discussed above and in the original post, it seems to me that distinctions in ordinary language *may* be interesting and useful, but they also may reveal confusions and historical prejudices. For this reason, ordinary language distinctions should be taken no more seriously than many other sources of information.

What I object to is philosophers taking ordinary language distinctions and intuitions as indicative of some special set of "metaphysical" facts with an a priori status that empirical evidence is powerless to overthrow. What ordinary language distinctions give us is an *optional* set of *concepts* to be evaluated pragmatically for their usefulness in negotiating the empirically given world.

[John Chipman]

Ralph Wedgwood-
To qualify what you wrote a bit- the fact that there are idiosyncratic yet intelligible features in every language might be philosophically relevant.

[Lisa Shapiro]

It strikes me that there is a kind of naivete about the actual practices of scientists in some of the above discussion. While it is often the case that new and groundbreaking scientific theories (say, physical theories) run counter to common sense, in defending those theories extraordinary efforts are made to render those theories intuitive (in a colloquial sense), typically by offering metaphors and images that capture the common-sensical imagination. Think of the seventeenth century mechanist appeal to clocks, or the elegance of Feynman diagrams. While part of what scientists do is offer up a way for lay persons to understand what they are up to (and so to change what is common sensical), you might also think that the theory is about the world, rather than simply a mathematical abstraction, just insofar as it can connect with common sense. You might think the same thing about philosophical theories. If a philosophical theory cannot manage to connect with the common sensical imagination, well, then it's simply a theory, and not necessarily helpful in furthering the goal of philosophy, which I take to be, as polyannish as it might sound, enriching our understanding of ourselves. I don't mean to be privileging common sense, but rather to assert its relevance (and with it the relevance of natural language).

[Jason Stanley]

It seems to me that there are two different issues here. One is the status of common sense, presumably as reflected in the method of intuitions (which are, after all, just reports of firmly held beliefs about the subject matter). Another is the status of appeals to language of the sort one finds all over philosophy, from the British ordinary language philosophers to continental philosophy to contemporary analytic philosophy. It seems to me that many of the latter sort of appeals are often in the service of *correcting* currently pervasive beliefs (i.e. common sense) by pointing out that our language reflects a different point of view that our current explicit and immediate beliefs about the subject-matter. To take an example dear to my heart, common sense may tell us that knowing-how isn't a species of knowing-that, but a close investigation of language might reveal we don't actually speak that way. That may lead us to overturn common sense, in the same way that a scientific theory would. Or common sense might lead us to think that there is only one form of labor, but etymology might reveal that some of our ancestors made distinctions between kinds of labor to which we are now blind.

[Kevin Schutte]

My response to this question is along the same lines as the comment provided by kris. Professor Stanley frames his question by asking about people who question "how language could be relevant in shedding light on philosophical problems that are not primarily about language." I can certainly think of philosophers who question whether "language could be relevant..." because they think that the problems that ARE primarily about language need to be solved FIRST. But Stanley seems to be thinking of people who question whether "language could be relevant..." and yet proceed to work on those very philosophical problems without considering the concerns raised by language. In order to adequately respond to Stanley's concerns, I'd need to see a clearer example of such a philosopher, one who dismisses natural language concerns tout court rather than merely dismissing them often. Surely we can all think of philosophers who think that many concerns raised by language are irrelevant to specific problems, but I'm having a hard time thinking of any philosophers who think that ALL concerns raised by language are irrelevant to some specific problem.

[Matt Weiner]

Jason, I'm not sure that appeals to common sense are as pervasive as you say. For one thing, I don't think appeals to intuition are always and everywhere appeals to common sense. They're appeals to some sort of rock-bottom judgment that isn't backed up by reasons, perhaps, but not necessarily judgments made by the Volk; sometimes only philosophically trained people are expected to share or even formulate these judgments. My favorite example is Davidson's claim* that it is a deliverance of intuition that 'All emerires are grue' is lawlike; whatever faculty delivers judgments in a sentence that contains three technical terms in six words, it's certainly not common sense! But Davidson himself may have intuitions about that claim, and other philosophers may share them.

So I'm not sure that those who appeal to intuition (or other non-empirical philosophical methods) will necessarily be moved by the linguistic data, because I'm not sure that they'll see themselves as holding pervasive beliefs that can be corrected by closer observation of the way we talk. Anscombe's argument that practical knowledge is distinct from theoretical knowledge doesn't rely on an appeal to common sense, and I don't think Ryle's arguments about knowing how and knowing that do either (though I could be wrong, since I'm not that familiar with Ryle). So presented with an investigation of language that reveals that we don't speak as though knowing-how is different from knowing-that, they might respond, "What of it? We aren't concerned with how we speak." If their arguments aren't beholden to common sense, they might not be beholden to the way that we usually speak either. Just as a Cartesian skeptic who argues that we don't know certain ordinary claims won't be impressed that we speak as if we know them.

*In "Mental Events," p. 218 in the 1980 edition of Essays on Actions and Events.

[marc moffett]

It does a disservice to those who have worked hard to isolate an epistemically significant notion of intuition (whether you accept the view or not), to simply characterize them as nothing more than "firmly held beliefs." Belief (as Bealer, Huemer and others have argued) is neither necessary nor sufficient for intuition.

I am curious, however, why you think linguistic analysis of the sort invoked in the know-how paper is more epistemically secure than concrete-case intuitions such as the Gettier intuitions? After all, while I know plenty of smart linguists who reject your particular theory of the syntactic structure of know-how attributions, it is very rare to find someone who understands the Gettier cases but who doesn't have the Gettier intuitions (i.e., that they are not cases of knowledge).

[Jason Stanley]

Marc,

Sorry, I shouldn't have imported my own suspicions about the literature on intuitions into my brief remark. As to the main question you raise -- of course if someone appeals to a theoretically controversial claim about the etymology, grammar, or semantics of natural language to support a point, then their point only has enough significance as their premises have theoretical support. But you are misrepresenting the situation vis-a-vis knowing-how. Sure, some linguists deny the existence of PRO, and have other views about infinitives. But I know of no smart linguist who has argued that the syntactic structure of sentences like "John knows how to vote" is significantly different than the syntactic structure of "John knows why to vote Democratic." Yet Ryle would classify the latter as a species of knowing-that, and the former not. So while it's true that we give a specific syntax and semantics for knowing-how constructions, one which obviously incurs various theoretical commitments, we did so just to show it could be done. Our general conclusions are consistent with views that are widely shared (or, as we also argue, should be widely shared) across theoretical frameworks in syntax and semantics. And that's why the conclusions have the plausibility they do.

Though I'm certainly not theoretically committed to suspicion of the method of intuitions, my concern with the method of intuitions is that some of the concerns raised by Stich-type folks will turn out to be vindicated. In contrast, the kind of work (say) Williamson and I did with knowing-how is not subject to these concerns. Framing effects are irrelevant in doing syntax.

[jonathan weinberg]

Jason: "Though I'm certainly not theoretically committed to suspicion of the method of intuitions, my concern with the method of intuitions is that some of the concerns raised by Stich-type folks will turn out to be vindicated. In contrast, the kind of work (say) Williamson and I did with knowing-how is not subject to these concerns."

As a "Stich-type folk", let me just chime in with my agreement here. I think that nothing in our arguments should be taken as undermining the appeal to real linguistics in one's philosophizing (of which I find the Stanley & Williamson work in question to be an excellent exemplar).

Marc: "...it is very rare to find someone who understands the Gettier cases but who doesn't have the Gettier intuitions (i.e., that they are not cases of knowledge)."

Only if you don't ask people who come from other parts of the globe! ;-)

[marc moffett]

Jason, thanks. The clarification as to the level of generality (theoretical nuetrality) required to make language-based considerations epistemically safe is an important one. Sorry if I misrepresented you. But by parity wouldn't the same sort of care in isolating the relevant class of "seemings" also helps to alleviate some of the Stitch-style concerns? After all, it doesn't follow from the fact that some intuitions (typically about marginal cases, as I recall) are susceptible to framing effects, that all intuitions are susceptible to them. As with perceptual evidence, one must take note of the cognitive conditions under which a given judgment is made.

[jonathan weinberg]

"After all, it doesn't follow from the fact that some intuitions (typically about marginal cases, as I recall) are susceptible to framing effects, that all intuitions are susceptible to them. As with perceptual evidence, one must take note of the cognitive conditions under which a given judgment is made." But what does it mean, in practical terms, for intuitionists to try to follow this advice? Unlike our perceptual practices, and linguist's practices with the intuitions they make use of, we just don't have anything like a working theory of the competence of philosophical intuitions. We know not to intuit while drunk, say, or while heavily distracted by another cognitive task. But beyond that, we just don't know very much about the conditions that are or are not conducive to good intuiting. So the basic idea here (of taking cognitive conditions into account) is a good one, but not currently implementable.

Relatedly, what counts as a marginal case is something that we don't seem to have any sort of working consensus on in philosophy, either. The main case that my students and I found a framing effect on, for example, was a version of Lehrer's Mr. Truetemp. I don't think that Lehrer considered it a marginal case, nor have people seemed to consider similar anti-externalist cases (like BonJour's various clairvoyants) to be problematically marginal. And, as I flagged above, we found significant cross-ethnic variation in a Gettier case, which I figure epistemologists would not in general count as marginal. Part of what we need to do -- and have not yet done, as a profession -- is to get a much better sense of what the application boundaries of of our intuitive capacities really are.

[marc moffett]

Thanks, Jonathan. I don't want to hi-jack the thread with a debate over the relative merits of x-phil. (As you may know, I have dabbled in some experimental stuff myself, though I think it is only feasible to draw very weak conclusions from it.) My primary point was just that I don't see any very good reason for thinking that intuition--properly circumscribed--is more prone to error than linguistic analysis--properly circumscribed. In fact, I think that most interesting cases require some sort of balancing act between the two.

[Peter Ludlow]

Jason, whether appeal to language is a method or a source of evidence is kind of beside the point (assuming I can wrap my mind around the idea that routinely appealing to a particular kind of evidence is not a method -- a difficulty Barry Lam raises above).

I think that skepticism about *any* philosophical methodology/source of evidence is to be expected. If you have a method/source of evidence, expect that you'll have to defend it -- especially if you are working in their problem space. These peeps aren't puzzled by your appeal to linguistic data -- they are just challenging you. This is what philosophers do.

A partial answer to the challenge is to point out that they deploy your method all the time, whether they know it or not. I believe this is what Jeff is getting at above. Anyone who says that strictly speaking x is the case, but that they can speak as if x is not the case because of assertability conditions or pragmatics or whatever is neck deep in the semantics/pragmatics distinction and hence they are engaged in a form of linguistic methodology. Example: if someone says that we don't know x, but we are entitled to speak as though we do, they are playing that game. If someone says that everything in by backwards light cone is a cause, but pragmatics entitles me to overlook some of those events when ascribing causes they are playing that game. If someone says that a sentence is not doing the usual thing of expressing truth conditions but is rather expressing attitudes, again, they are playing a form of that game.

The reason we appeal to linguistic judgments (can we *please* retire the word 'intuitions' here?) is that we have crafted a methodology in which we believe that linguistic judgments can play a helpful role in probing the boundaries between the semantics/pragmatics divide among other things. The judgments can do this because they can help shed light on what is genuinely semantic and what is not. These are linguistic questions, so it is no surprise that linguistic judgments are illuminating.

The best way to convince someone to play your game is to show them they are already playing it. Then you can start talking about what is required to play it well.

[jonathan weinberg]

I don't think we're threadjacking here -- the question on the table was the comparative methodological merits of the appeal to linguistics on the one hand, and analysis-by-intuitions on the other. And some experimental philosophy work is directly relevant to our evaluation of the latter. In general, if we find ourselves needing to appeal to premises like the one you appealed to earlier, of the form "it is very rare to find someone that...", then we're already up to our armpits in empirical commitments, and they're commitments that we should have no expectation to be able to answer well from our own informal sense of things. If we really want to know whether premises like that are true, then, in Peter's terms, we're already playing the experimental philosopher's game. And I think what evidence there is suggests a pretty significant difference between the kind of resources available to the syntactician on the one hand, and the philosophical analyst on the other.