What the Contextualist REALLY Says about Disputes

There’s a certain understanding as to what contextualists think or say about disputes between skeptics and their opponents that seems to be very widely assumed. And here it pops up again, in Mark Richard’s (very well worth reading) recent article, “Contextualism and Relativism” (Philosophical Studies 119 (2004): 215-242):

Suppose a confrontation between a skeptic with high standards, and Moore, who has low standards. The skeptic says

(1) You don’t know that you have hands.

Contextualism tells us that the content (and thus the extension) of ‘knows’ in the skeptic’s context is determined by the standards that his context provides. Since he, unlike Moore, has high standards, Moore and the claim that he has hands just don’t make the cut. The skeptic’s utterance of (1) is true: that is, Moore doesn’t know that he has hands.

Of course, when Moore utters

(2) I know that I have hands

the standards in his context are the relevant standards, and so, given his low standards, he speaks truly. So Moore knows that he has hands after all. But how can that be? Didn’t the skeptic just establish that Moore doesn’t know that he has hands? Well, says the contextualist, what the skeptic said was true. But since ‘know’ is contextually sensitive, (2) doesn’t say the same thing, when Moore uses it, as does

(3) You know that you have hands

when the skeptic uses it. So there’s nothing contradictory about the skeptic’s being able to use (1) truly while Moore can so use (2).

One feels that something is awry. One wants to say that when the skeptic and Moore argue with each other…

(pp. 215-216)

Indeed, something is awry here — something beyond what Richard has in mind: This is not at all what the contextualist says about such confrontations…

At least this contextualist — but also other contextualists I know. Richard here assumes that the contextualist opts for what in my “Single Scoreboard Semantics” (Philosophical Studies 119 (2004): 1-21) I call the “multiple scoreboards” view, on which, in such debates, the truth-conditions of each speaker’s use of ‘know(s)’ is governed by that speaker’s own “personally indicated content.” But I have always thought that, and have written as if, some sort of “single scoreboard” view is instead correct — and I thought the same was true of David Lewis, Peter Unger, and Stewart Cohen. On a single scoreboard view, appropriately enough, there is in such a debate a single scoreboard that governs the truth-conditions for both speakers’ uses of ‘know(s)’. The “score” can be affected by conversational moves made by either speaker, and there are tricky questions to answer about what that single score is in cases where the two speakers are pushing the score in opposite directions, but such a view has the result that the skeptic is denying just what her opponent is affirming — the two really are contradicting one another.

OK, but what then what is that single score in the debates we are considering? Who is speaking truthfully — the skeptic or her opponent?

Well, there are many possible single scoreboard views for the contextualist to choose among, several of which I describe and briefly discuss in the aforementioned “Single Scoreboard Semantics.” This post, in fact, is a shameless advertisement for that paper, about which I’d be grateful to receive comments. Yes, it’s already published, but I’ve continued to think about the issues addressed there in the 2+ years since I actually wrote that paper, and I plan to revisit them very soon. Those with subscriber access to Philosophical Studies can get “SSS” — and Richard’s paper, and several other papers on contextualism from the UMass Contextualism conference — by clicking here, then clicking on “PREVIOUS ISSUES”, then clicking on Vol. 119 “Issue 1-2, May 2004”. For those who do not have such access, the conference web site (click here) has preliminary versions of many of the papers posted, and I have the penultimate draft of “SSS” posted on my web site here (pdf) and here (word).

Note that — as Richard sees well — the issue here is not confined to ‘knows’, but extends to other context-sensitive terms, including many that are obviously context-sensitive. How tall must something or someone be to count as “tall” in a given context, how far out does a given use of “here” reach? There can be debates & disagreements turning on these matters, too, where the parties to the debate sure seem to contradict one another, they intend to contradict one another, and they take themselves to be contradicting one another, though the “personally indicated content” of each seems different. On “single scoreboard” views, they really are contradicting one another, and the various single scoreboard views about ‘know(s)’ I discuss in “SSS” can be applied to these other terms as well.

So, at last, which single scoreboards view is correct, and which party to the dispute is speaking truthfully? On the particular single scoreboard view I opt for, so far from it being the case that both the skeptic and her opponent are speaking the truth, I hold that neither of them speaks truthfully. (And, as I report in the paper, when I asked Stew, he was inclined toward the same answer.) On the “Gap view” I opt for, in such cases of debate, a claim that “S knows that P” is true (and “S doesn’t know that P” is false) iff S meets the personally indicated standards of both parties to the debate; “S knows that P” is false (and “S doesn’t know that P” is true) iff S fails to meet either of the personally indicated standards; and “S knows that P” goes truth-value-less (and so does “S doesn’t know that P”) where S meets the personally indicated standards of one, but not of the other, speaker.

Why do I believe that? See “Single Scoreboard Semantics.” But whether or not you look at that paper, I’ll be happy if, having read this post, you at least now realize that it’s simply wrong to generally say that contextualists deal with cases of dispute by making what both parties say come out true.

If you happen not to like the gap view (I know that many out there are gapophobic), don’t take that out on contextualism. Just tell me how you handle analogous disputes turning on, say, how far out “here” reaches (surely you’re not an invariantist about that!), and I’ll tell you what to say about the dispute between the skeptic and her opponent.


What the Contextualist REALLY Says about Disputes — 6 Comments

  1. Keith, I hadn’t seen the Richard paper yet, but when I read the quote I was taken aback. This is precisely the misunderstanding of contextualism that students first leap to, and need cautionary remarks concerning to make sure they understand the view correctly.

    There is an issue in the neighborhood here, if there is a need to explain how Moore and the skeptic are always and everywhere disagreeing. That, of course, is simply the denial of contextualism. Richard’s last comment suggests that he may have that claim in mind, but then the discussion ought to go a different direction, and in any case, nothing that comes before the last line of the quote would be relevant to this point, anyway.

    I think one of the problems with getting things straight about contextualism is that there are so many possible contextualist positions that it is easy to confuse a particular version of it with the view itself. The position Richard complains about is a version of the view, but you’re right that it’s one that none of the main exponents of the view endorse.

  2. Jon: Agreed: the multiple scoreboards view is a (possible) contextualist view. But I’m not sure I know of any contextualists who hold it. But I may be forgetting or misconstruing someone. Perhaps some contextualist will come forward and ‘fess up, or some invariantist will say that MS is the form of contextualism they think is best, or someone can remind me of some contextualist who opted for the view.

    Where the confusion comes in is probably here: standard contextualists typically appeal to the fact that we don’t “know” according to what we can here call the “absolute” standards that… well, as I typically put it, that the skeptic’s maneuvers have at least some tendency to put into place, even though we do “know” according to ordinary standards. It’s perhaps easy to read that as saying that the skeptic’s claims really are true, and that the Moorean’s claims are also true, even in the context of a debate with the skeptic? Or maybe MS is just the version of the view that comes first to mind?

  3. Keith, maybe it’s the argument from ignorance and glosses on it that incline toward this mistake. If one uses that argument to explain the view, one has to say that there is no context in which both premises are true. Instead, I say, in the skeptical context, one premise is true, and in the moorean context, the other one is true. And immediately, students think that the skeptic and Moore are talking past each other… It has something to do, I think, with labelling the contexts as skeptical and nonskeptical, leading one to expect that the skeptic is always in a skeptical context and Moore in a moorean context–sounds so clear and obvious it seems nearly tautological!

  4. I think you’re both being a little harsh on Mark, though I think he could have stated this claim more cleanly. The SSS approach will mean that when Moore and the sceptic are in the same room, or on the same stage, debating with each other, then they are disagreeing. I took it the intuition Mark is appealing to is the intuition that when Moore (or, better still, an everyday Moorean) makes a knowledge claim in his own study, and the sceptic writes in her paper “What a stupid Moorean, claiming to know things”, they are having a disagreement. Now contextualism doesn’t agree with that, which is obviously a cost of contextualism, related to the famous costs to do with the need for (regular) contextualists to adopt semantic blindness as an explanation.

  5. Brian, you’re right that there is a cost to contextualism in having to posit semantic blindness, but I think you’re way of taking the passage doesn’t fit a couple of the claims made. Mark describes “a confrontation between” the skeptic and Moore, and characterizes the skeptic’s standards as “high” and Moore’s as “low”. This way of talking about contexts makes them sound much too subjective, and if there is a confrontation between the two, that suggests that they share a context.

    But the point about semantic blindness is the one to focus on, regardless of how we should take the above passage. It’s the serious issue behind the temptation to give MSS interpretations of contextualism. I’m not sure what Stew’s particular reason was for giving up on describing his view as an “indexical” theory, but the semantic blindness needed here is a good reason to distinguish contextualism about knowledge attributions from the contextual character of indexical expressions. Competence in recognizing the varying semantic import of indexical expressions from one mouth to another is a requirement on having mastered the language, and in this way epistemological contextualism is not like the contextualism involved in indexicals.

  6. Brian:

    1. Too harsh on Mark?

    MR writes of various conversational situations, including some like you describe. But in the passage I quote, he certainly writes as if he’s dealing with a direct, in-person confrontation between a skeptic and a Moore — he not only describes it as a “confrontation”, but also writes, “…when the skeptic and Moore argue with each other…”. It’s wrt such direct disputes that the intuition that they are contradicting one another is strongest — and MR takes advantage of that strong intuition. And one of the most misleading myths about contextualism is that it takes the “eveybody’s speaking the truth” approach even to such disputes. So it seems worth pointing out that this is indeed a myth.

    2. One-Way Disputes

    What you bring up is an intermediate case — where S2, in a different conversation from another speaker S1, says that S1 was wrong (or stupid or whatever) in claiming that S knows, since S didn’t know. Here, the intuition of contradiction may not be as strong as in the case of a direct disagreement, but can be pretty strong, nonetheless. But the contextualist can — and *this* contextualist is inclined to –treat these as “linked contexts”, and apply something like the gap view here, too.

    I’ll explain by use of a case involving how far out “here” reaches — to erase any impression that this is a conundrum special to contextualists about “know(s)”. So, S1 is involved in a discussion of whether Frank is “here”, like the one I describe in “SSS,” except that all parties to the discussion are using “here” in the same way — say, to cover the hotel in Atlanta where the APA meetings are being held. S2 is passing by the table S1 is sitting at, and so overhears as S1 expresses the opinion that “Frank is not here.” The next night (at the second APA “smoker”), S2 is sitting at the same table that S1 was sitting at the night before, and is discussing whether Frank was “here” the previous night. S1 is not present. But in this new conversation, at least at its start, everybody is using “here” to cover a larger area — maybe the whole city of Atlanta, rather than just the hotel where the APAs are being held. S2 expresses the opinion that “Frank was here.” So far, so good. But now S2 starts talking about S1 and how wrong/stupid S1 was: “S1 said that Frank wasn’t here. But she was SO wrong. OF COURSE, he was here.”

    Now, once S2 starts talking about S1 that way, I get very tempted to “link” S2’s context to S1’s. For S2 has now clearly expressed *two* intentions regarding his use of “here” that turn out to be in conflict: He intends to use “here” to refer to certain area — say the city of Atlanta — and he also intends to be using it to contradict S1. In this situation, I’m inclined to take both of these clearly expressed intentions very seriously, and go gappy on the content of S2’s use of “here,” at least when he’s talking about S1, intending to contradict her:
    –“Frank was here” is true (and “Frank was not here” is false) iff Frank was in the hotel (and thereby satisfied both of the relevant extensions of “here”);
    –“Frank was here” is false (and “Frank was not here” is true) iff Frank was not even in Atlanta (and thereby failed to satisfy either of the relevant extensions)
    –“Frank was here” is truth-value-less (and so is “Frank was not here”) iff Frank was in Atlanta, but not in the hotel (and thereby satisfied one, but not the other, of the relevant extensions)

    Note that in such cases of “one-way” disagreement, if you take my approach, you don’t get as clean a contradiction as you get in cases of two-way disputes. For the above gappy truth conditions govern only S2’s confused use, but not S1’s use, of “here”. So you don’t get S2 affirming the same gappy thing that S1 denies, for S1’s use of “here” doesn’t go gappy in the relevant way. But you do get the result that no matter where Frank was, S1’s denial and S2’s affirmation can’t both be true.

    I’m inclined to take the same approach to cases of one-way disagreement — where a speaker in one context clearly expresses the intention to be contradicting another speaker in another context, not realizing that there’s a difference in standards — in cases involving “knows.” Not all contextualists will agree with this approach — just like not everyone will agree with the analogous treatment of the one-way “here” dispute that I favor. But it’s certainly far from clear that THE contextualist position on cases of one-way disagreements is that both parties are speaking the truth. (Though such cases have not yet been discussed in print, to the best of my knowledge.)

    3. The Obvious Cost to Contextualism, “Related to the Famous Costs to do with the Need for (Regular) Contextualists to Adopt Semantic Blindness as an Explanation”?

    Where it really is clear (at least I take it to be) that THE contextualist position is that both speakers are speaking truthfully are cases where S1 is in a low-standards context, S2 is in a high-standards context, and neither’s use of “know(s)” is governed by any intention to be contradicting some speaker in a different context. (For a more complete description of the types of case pairs in question, about which this really is the contextualist position, see my “The Ordinary Language Basis for Contextualism and the New Invariantism,” which, like “Single Scoreboard Semantics,” and many others of my papers, is available on-line at http://pantheon.yale.edu/~kd47/OLP.htm ) Where such pairs of cases are properly constructed (see again “OL Basis” for some directions), the intuition that the two speakers are contradicting one another is … well, let me just say for now (more below), not nearly as strong. Now, I certainly admit that the thought that they are *not* contradicting one another seems not as intuitive here as in analogous cases turning on such matters as how tall one must be to count as “tall”, or how far out “here” reaches. And I suppose at least a big part of the explanation for this is that the relevant aspect of context-sensitivity just isn’t as clear to speakers in the case of “know(s)” as in the case of “tall” or “here”. So, if understand your use of “semantic blindness” correctly, I do indeed here appeal to such semantic blindness.

    But this appeal to semantic blindness seems not to be a relative cost to contextualism (relative to invariantism). For here speakers go both ways. Try it on your students some time. Give the real test case pairs for contextualism vs. invariantism (as described in “OL Basis”) and ask them whether the two speakers (who, recall, will be participating in separate conversations) are contradicting one another. In fact, as far as I’m concerned, you can even bias your question against contextualism: “Obviously, these two speakers are contradicting one another. Which one of them is speaking truthfully?” If my experience is any guide, if you have a large enough group of students, one of them will overcome the suggestion, give the answer that they both are speaking truthfully, and will even give a proto-contextualist analysis of the situation, saying something along the lines of how they mean different things by “knows”, and then, at least if you have the right kind of pair of cases, about half, or often significantly more than half, of the students will agree with this (proto-) contextualist suggestion, some of them treating it as obvious. Now, often at least, many others will reject this contextualist suggestion. About those speakers who go contextualist here, the invariantist will have to say they are blind to the invariantism of their very own word “know” (to mimic some of the rhetoric one tends to hear invariantists use in connection with the issue of “semantic blindness”: “Are we REALLY to believe that speakers are THAT blind to the meaning of their VERY OWN word?!!!), just as the contextualist will say that those who go invariantist here are blind to the context-sensitivity. Apparently this semantic blindness (that revealed by intutions about whether various affirmations involving “know(s)” contradict various denials of “knowledge”) is something we’re just plain stuck with in either case: If contextualism is correct, it is so in a way that isn’t obvious to all ordinary speakers; and if invariantism is correct, it is so in a way many ordinary speakers are quite blind to.

Leave a Reply

Your email address will not be published. Required fields are marked *