Sunday, January 06, 2008

Davidsonian philosophy of language

In an introductory article on forcing, Timothy Chow mentions something he calls "exposition problems," which are the problems of presenting some material in such a way that it is perspicuous, clear, explained, and learnable. He thinks that forcing presents an open exposition problem. I just read through Ramberg's Donald Davidson's Philosophy of Language and it goes a way towards an answer to the exposition problem for Davidsonian philosophy of language. With the exception of the incommensurability chapter towards the end, it is remarkably clear and quite helpful. I'm not sure if it would be perspicuous to someone coming to it without having read at least some of the Davidson articles. If you have read them it does a good job of displaying the unity of Davidson's thought on language which is not always apparent when, say, one juxtaposes "Truth and Meaning" and "A Nice Derangement of Epitaphs". Ramberg isn't doing straight Davidson exposition though and the volume of quotation is rather meager. He does succeed in presenting Davidson's ideas in a coherent, unified, perspicuous manner that, at least for me, made things gel. One of the things that he emphasized is that interpretation is a process that is supposed to result continuously in the revision of theories of truth rather than a single theory. This is maybe easier to see in "Nice Derangement" than the early stuff. I don't know if Davidson made this explicit anywhere though. Anecdotally, I heard someone say that Davidson endorsed this book as a better explanation of his theory than he ever gave.

I came across something while reading this that reminded of a claim Davidson makes which I've never quite gotten. He claims that in order to interpret someone you have to treat their beliefs as mostly true. Since beliefs are mostly true there isn't the possibility of systematic error of the kind skepticism points to. Ramberg didn't say much about this that clarified why this is so. He may have said some things in relation to the principle of charity that are relevant and I suspect there is a connection to his rejection of the principle of humanity (aim to maximize intelligibility rather than agreement). However, it seems like if I ran into a modern Don Quixote, who took cars to be metal horses and who took my apartment to be a castle and me to be a coffee bean, I could interpret his (bizarre) behavior even though it seems like most everything he says is false. It may take a little while for enough of his knights errant tale to come out, but it seems like his speech would be interpretable. Despite the fact that most of what he says is false, one would be able to work out the ways in which it is false, thereby making sense of him. Maybe the idea is supposed to be that there is a lot more that he believes that is true, or at least that you take to be true, that is semantically connected to what he says, though not made explicit in his speech behavior (possibly implicit in his nonverbal behavior). This other stuff must, for the most part be true, in order for us to make sense of him. But if my Don is under the impression that he is floating above the surface of Mars, many of these background beliefs go false too. It seems like I'd be able to interpret him, with some difficulty, yet his beliefs are systematically mistaken. I don't think I could interpret him if I didn't take him ah treating most of his beliefs as true. This, however, isn't what Davidson claims. He thinks that it would be impossible to interpret someone unless you treated them as having mostly true beliefs. So, I am stuck.


Daniel Lindquist said...

There is more than just Don Quixote's utterances which can count for evidence of his beliefs. If we are to understand Don Quixote as holding that he believes himself to be floating above the surface of Mars, then we must understand his actions to be generally reasonable for someone who holds themself to be floating above the surface of Mars: Perhaps he tethers himself to the ground to avoid blowing away in the wind, or he wears an oxygen tank (since Mars has no breathable atmosphere). But this would mean that Don Quixote would believe (truly) that strong winds could blow away someone who was floating above the ground, and that supplementary oxygen is needed for humans in Martian environs. So the false belief about whether or not he's on Mars (and floating) presupposes many true beliefs, if we are to be able to attribute it to him. If Don Quixote merely says that he is floating above the surface of Mars, while behaving like a normal person (who is neither floating nor on Mars), then we oughtn't attribute the belief that Don Quixote is floating above the surface of Mars to Don Quixote; he espouses the belief, but does not actually hold it -- Don Quixote is irrational. (This means that belief that p on someone's part does not entail belief that they believe that p. But I think that's just right, though I've seen it argued against. Self-knowledge is often hard-won.)

(If Don Quixote says he's floating above the surface of Mars, while behaving in a way which is not reasonable either for a normal human on Earth's surface nor for one floating above the surface of Mars, then it seems reasonable that Quixote means something other than "floating" and "Mars" when he claims to be "floating above the surface of Mars". For instance, his actions might become more comprehensible on the hypothesis that he thinks he's walking along the ocean floor on one of the seas of one of Jupiter's moons. In which case we must attribute to him various true beliefs about ocean floors, Io, etc. Or if we can find no way to make sense of Don Quixote's behavior, then we can't very well attribute false beliefs to him, either. For we haven't the foggiest what's going on in his head, in that instance. And if we are to suppose that he really is acting in a way which can be understood, though we have not been able to do it, then we must suppose that whatever the final account turns out to be, it will attribute mostly true beliefs to Don Quixote. Or, if we suspect that there is no sense at all to what he is doing (perhaps he is a full-blown schizophrenic), then he's no longer the sort of being we can plausibly attribute beliefs to -- and so again, he can't have mostly false beliefs.)

I don't see that Davidson would have any reason to reject the idea that most of a person's utterances might be false. We can easily conjure up science-fiction scenarios in which this is the case: Have someone learn to speak a language, but then spend several years saying nothing but "I am a fruit bat" while carrying on the rest of his life normally (albeit taciturnly). Trivially, most of what this person says is false (assuming he's not actually a fruit bat). But he still may have mostly true beliefs -- he just doesn't talk about them anymore, since he got into his fruit-bat phase.

Davidson can allow for someone to be "systematically mistaken" in particular areas but not in all areas. If someone was smellblind, then I should expect them to be "systematically mistaken" if they tried their hand at cooking; it would all taste lousy to anyone who could smell in addition to taste their food. The clusters of false beliefs depend on clusters of true beliefs for us to be able to understand them as having the content we take them to have. Thus the smellblind cook would know about how to use ovens, what is and is not poisonous, whether something is similar to a dish he's encountered before or not, etc. though he doesn't know squat about whether or not a particular dish is tasty.

So, yes, I think it's right to say that "the idea is supposed to be that there is a lot more that he believes that is true, or at least that you take to be true, that is semantically connected to what he says, though not made explicit in his speech behavior". Beliefs are interpreted alongside actions; I can't understand any of what you do unless I understand some of what you think, and vice-versa. So Don Quixote (or the guy in that Onion article) are plausibly self-deluded to an extreme extent, but still they have a great bulk of true beliefs.

(One question I'm not so clear on is whether most of someone's beliefs have to be regarded as true, or just a great many of them. I suppose the idea is that any false belief will rely on some true beliefs (plural) in the background to give it its content, and so false beliefs can't be a majority. But I'm not sure why the same body of true beliefs couldn't serve as the background to a great many false beliefs, thus allowing for false beliefs to end up as the majority of a person's beliefs (though there are still a great many true beliefs, which should be enough to leave the skeptic out of business). Though I suspect that trying to figure out "how many beliefs" someone has is nonsense, and so also would be the question of whether most or only some of them were true -- there's no counting, so there's no question of which is the greater quantity. But it seems like it might be desirable to allow that some people are wrong about "darned well everything", which is at least in a superficial tension with the principle of charity. If there's no problem with speaking of "a great many" beliefs rather than "most" beliefs when formulating the principle of charity, then the tension would dissolve.)

Duck said...

What Daniel said; but let me repeat and/or add a couple of things.

When your interlocutor is deluded, it may indeed seem odd to think of him as having mostly true beliefs. But when you ascribe particular false beliefs to him (again, that he is floating on Mars or that those things are metal horses), you are ascribing meanings as well ("Mars" means Mars, etc.). (If you don't know what he means, then you don't know what he believes either – so as far as you're concerned, maybe he's too Froot Loops to believe anything.) In ascribing meanings to someone, you give the contents of their utterances (and beliefs) in your own language: he believes that cars are metal horses. But in order to believe that, he has to know enough about metal and horses, by your lights, for the ascription to make sense.

So it's not so much that you're ascribing true beliefs to him (which you are), but that you're explaining the particular disagreements (beliefs about cars or Mars) in terms of your shared beliefs about the one and only objective world which is causing the sensations of both of you (beliefs about metal, or horses, or even that thing there). If you don't take him to have enough true beliefs about the latter, then you have no right to attribute him those particular false beliefs about the former. To interpret someone is to see the two of you as occupying a shared world. (The later Davidson is key here – "Three Varieties of Knowledge," "Meaning, Truth, and Evidence," etc.)

Now the skeptic will rush to point out (as Barry Stroud does, as if on cue) that just because our boy shares a great number of your beliefs, that doesn't mean that they're true. After all (on his view of the matter), maybe you're both massively deceived. (Maybe we all are.) But on my view that's not a coherent attitude to take w/r/t one's own beliefs. To believe something at all is to believe that it is true. So to be a belief at all, for an interpreter, is to be part of a set of beliefs (about the one and only objective world) which is mostly true. If it didn't come out that way, you wouldn't have any reason to ascribe beliefs (and meanings) at all. (BTW, I always say "mostly" true rather than that "a great many" are true, but I don't think it matters. I do think that if one gets the point being made here, there's no reason to doubt that they're "mostly" true, so that's one way of checking to see if the point indeed got through.)

Davidson himself waffles about whether he is 1) refuting skepticism as false; 2) rejecting it as incoherent; or 3) with Rorty, telling the skeptic to go away. I say "waffles" because I take him to believe that one should really pick one and stick with it. But I'm perfectly happy to look at it in whichever way is appropriate in the context.

Oh, and what the heck is "forcing"?