Wednesday, January 28, 2009

Moral intuition and linguistic intuition

Here's a disanalogy between moral intuition and linguistic intuition (for example, intuitions about what a word means or whether a particular construction is grammatical). I'm sure that something like this is true, though I may not be talking about it right. And who knows, maybe it's more controversial than I think...

We can imagine a community where everybody across all times has the same moral intuitions, and they're all wrong. But we can't imagine a community where everybody across all times has the same linguistic intuitions, and they're all wrong. If the community of Spanish speakers regards it as intuitive that 'arroz' means 'rice' in Spanish, that's what 'arroz' means in Spanish. When we imagine them all using it to mean 'beef', we're just imagining a situation in which 'arroz' means beef in Spanish. However, if all Spanish speakers (or all Puritans, if we want to make this be a community of moral co-believers rather than a linguistic community) thought it was wrong to use birth control, there still might be nothing wrong with using birth control. This is because the linguistic intuitions of the community play a role in constituting the language, while the moral intuitions of the community do not constitute morality.

I'm just using intuition in the sense of 'pretheoretical judgment' here. Obviously if you say it's a presentation of necessary truth to your nous or something that'll mess up the example.

14 comments:

Anonymous said...

I don't agree that everyone in a linguistic community couldn't get it wrong. I think the practice of the members of a community is what helps constitute their language, not their intuitions. People are often unable to give accurate definitions of words they can use perfectly competently. People will claim certain constructions are ungrammatical even though they use them.

Also, about the specific example, the intuition that "arroz" means rice, I think you need to explain what you mean by this intuition more clearly. (Obviously you don't mean the intuition that the Spanish word "arroz" means the same as the English word "rice".)

Richard said...

I wonder. Suppose everyone thinks (prima facie, without reflection) that 'whale' means a kind of giant fish. Further suppose the following dispositional fact: if they were to think about it more, they would repudiate this initial judgment, and acknowledge that it's at least conceivable that whales aren't fish. But as it happens, nobody in history ever does think about it more. They all think that 'whale', in their language, means a kind of giant fish. Are they right? It seems not.

Prima facie intuitions are not enough; linguistic dispositions (or idealized intuitions) play at least as large a role in constituting the language. Due to failures of introspection or imagination, members of a linguistic community may fail to realize that they are disposed to use a word in certain ways, so their initial "intuition" to the contrary - even if universal - may be mistaken.

Richard said...

(Oops, cross-posted with Anon.)

Bryan Pickel said...

Hi Neil,

I take it that by 'linguistic intuition' you meant more than just pre-theoretic judgments about how words should be translated, but also judgments about when to use words in context. Imagine that native Spanish speakers insist the 'arroz' means the same as, say, 'engineering', but still use the word exactly as they do now. It seems to me that there is then no pressure to suppose that they are right about this. I take it then, that you meant to include speaker judgments about when sentences are true. So the position is something like this: speakers can't be wrong about all sentences.

Supposing that's what you mean, I don't understand the point. Sure speakers can't be wrong about all of their linguistic judgments about the truth conditions of their utterances. But that seems consistent with the view that speakers can be wrong with isolated segments about their judgments. Imagine for instance a group of speakers with the same empirical information that I have who still insisted on calling a group of women 'witches' and treating them as such. These speakers further insist that the women are engaged in all sorts of nefarious behavior no matter how much empirical evidence we present them to the contrary. Imagine finally, that these speakers also make similar judgments about who would be a witch in other possible worlds. That is, the speakers can't even imagine a world in which the witchohood and nefarious behavior attributions aren't true of these women.

Now, you might say that since interpreting these speakers as meaning what we do by 'witch' would require them to be systematically wrong about a large number of their sentence including their modal sentences, we should instead interpret them as meaning something else. But this doesn't seem clear to me, if enough of their practices involving their word 'witch' fit with our understanding of the word, we should just interpret them as meaning what we do. So, I take this example to illustrate that people can go wrong about a large class of their pre-theoretic judgments about a large class of sentences.

But now let's go back to your example of a morally deviant culture. People in this culture use the word, say, 'wrong' in different circumstances from those which we would. But, it is embedded in their practices the right way. For example, they systematically prevent people from engaging in actions which they call 'wrong', etc. Further, on this interpretive hypothesis, the rest of their judgments are largely true. In such a scenario, it seems right to call interpret them as meaning the same thing we do by 'wrong' and just making false judgments about which actions are wrong. So I don't quite see the contrast you're looking for.

Neil Sinhababu said...

Anonymous:

People are often unable to give accurate definitions of words they can use perfectly competently.

True, but this just seems to me a matter of not being good at building theories that fit all the intuitive data. Maybe I shouldn't do this, but in the armchair psychology of language that I'm working with, our linguistic intuitions not only turn out judgments like '"arroz" means that white grain', but help us structure what we say when we talk, and thus play a major role in generating use.

Neil Sinhababu said...

Well, Richard, if everyone's semantic intuitions matched up 'whale' and 'giant fish', I'd say whale meant giant fish.

But that's presumably not the whole story about actual people's semantic intuitions. Sure, they might've thought that whales were giant fish, but it's not clear that this was in the semantics. These are natural kind terms, and they've got a substantial demonstrative element. So the real semantics is something like 'whales are those big things that live in the ocean, for example that creature and his cousins, and that creature and her cousins...' or something like that.

You could test for this by asking: "What if it turns out that whales are actually the descendants of land animals who went to the water, so they're actually in the land mammal family tree and not the fish family tree?" If the answer there is "Doesn't matter, whales have to be fish" then "whale" really does mean "giant fish" in your interlocutor's language. But if they say, "huh, well, if that were true I guess they wouldn't be fish, but that sure would be weird!" then they actually don't think 'whale' means giant fish.

Maybe this is what you mean when you say "linguistic dispositions (or idealized intuitions) play at least as large a role in constituting the language". If so, yeah, I mean to be including those idealized intuitions too.

Neil Sinhababu said...

I take it that by 'linguistic intuition' you meant more than just pre-theoretic judgments about how words should be translated, but also judgments about when to use words in context.

Yeah, as I was saying to anonymous, I think there's basically one mental state that plays both of these roles.

the speakers can't even imagine a world in which the witchohood and nefarious behavior attributions aren't true of these women.

I'm a little confused about this case. They think these women necessarily behave in a nefarious manner? Other than the god of the medievals, I don't know anyone who necessarily performs a certain behavior or necessarily has a certain moral status. At a certain point it's hard to figure out what to make of people's linguistic practices if their modal intuitions are really screwy.

Richard said...

Neil - yes, those dispositions are exactly what I was talking about.

You wrote: "we can't imagine a community where everybody across all times has the same linguistic intuitions [pretheoretical judgments], and they're all wrong."

But we can. I offered a counterexample. Everyone across all times makes the same pretheoretical judgment that 'whale' means a kind of giant fish, and they're all wrong. The reason they're mistaken is that if they were fully informed and thinking clearly -- so as to consider the neglected scenarios you would want to ask them about -- then they would acknowledge their error.

But now the supposed disanalogy with the moral case is less clear. Presumably people who are fully informed and thinking clearly will not make rational or moral errors either. According to ideal agent theories, at least, the judgments people would make in this idealized situation are pretty much constitutive of moral truth, just as in the semantic case.

So, what's the difference? (I suspect we can get away with much less idealization in the semantic case than is required in the moral case. But this requires some argument. The difference between the two is at least not so clear cut as your initial presentation makes out.)

Neil Sinhababu said...

Presumably people who are fully informed and thinking clearly will not make rational or moral errors either.

Here we disagree. I don't see how full factual information and clear thinking is going to get, say, a firm Kantian to accept that we should tell a lie to the murderer at the door.

I haven't seen an ideal agent theory that I regard as particularly impressive. Either they don't idealize enough and have all the problems of subjectivism, or they idealize so heavily that you might as well just forget the agent who's being idealized. (Brandt's cognitive psychotherapy stuff falls in the latter box.)

I suspect we can get away with much less idealization in the semantic case than is required in the moral case

This could be my point, though we'd have to precisify the 'much less' to figure out whether it is or not. If we're talking about really intense moral idealizations where we end up transforming the agent's moral intuitions, I'm willing to say that my point was just that we need much less idealization in the semantic case. But these are the ways of building an ideal agent moral theory where I think you might as well forget the agent you're idealizing and go over to some kind of straightforward realist story.

Neil Sinhababu said...

I'm thinking that when I express this view in the future, I should talk about moral/linguistic dispositions rather than moral/linguistic intuitions.

Charles said...

To the text above:

1.
What do you mean with the sentence: "We can imagine a community where everybody across all times has the same moral intuitions, and they're all wrong."
How can moral intuitions be wrong?

2.
Do intuitions really play a role in constituting a language?
("This is because the linguistic intuitions of the community play a role in constituting the language...")
For we have rules that constitute languages - and rules maybe are enough.
Intuitions are normally being asked for in the analysis or interpretation of certain questions (or for philosophial-problem- constitution).
Maybe the same with moral intuitions: intuitions play a role in analysis, not in constitution.

Neil Sinhababu said...

1) Well, Charles, some people have the intuition that homosexuality is wrong. Others have the intuition that homosexuality is not wrong. At least one party's intuitions are wrong.

2) Yeah, this is why I should probably talk about dispositions. See the comment before yours.

Charles said...

I do not think that one can really speak of truth and falsehood here.
(But that is a problem of moral philosophy...)

A comment to my previous comment:
Probably (here) analysis can be seen as (i) a form of constitution or as (ii) a mean for constition - a refined constitution.

Charles said...

Probably (here) analysis could ...