a pragmatic understanding of language, which doesn't address the question of realism — Tom Storm
That was the idea, yes, but I'm not sure it excludes what we want out of realism. This is precisely a question about the cognitive capacities and behavior of language-users. One reason to focus on learning when faced with such an issue is to "catch it in the act." Children are the ones who have to manage this mapping somehow; if it's a real thing (heh) then they're the ones who have to connect "ball" in their mouth to ball in their hand.
Put another way, if you're going to see it anywhere, you'll see it there, so look at the research on language learning and if that's not what it looks like, then this mapping is a myth.
I can give a small example of what I have in mind -- I think I'm remembering this from Rosch's prototype theory of concept acquisition. If you imagine a bunch of concepts arranged along a scale of abstractness, something like
cocker spaniel-
dog-
mammal, then children tend to come into that scale in the middle, learning
dog before the more specific or the more general.
Now we can ask how this partial language maps onto a partial world.
Dog applies to every breed, and adults are fine with that. But what about in the other direction? Indeed children will over-generalize their use of a concept while they lack the more general term, so, if
dog is the first mammal concept they acquire, or the first four-legged mammal, they'll apply it as if it were what we use
mammal for: cows are "doggies", cats are "doggies", and so on. (In
Monsters, Inc Boo calls Sully "kitty" -- those guys at Pixar are smart.)
Realism finds its clearest expression in the model-theoretic description of language, where you have a complete, closed set of symbols and a complete, closed set of objects, and they are matched up to each other according to some scheme. (It might be more precise to talk about systems of differences among symbols and among objects.) But to talk about natural languages, you have to allow the collection of symbols to grow, and allow the collections of objects that satisfy those symbols to shift, because the satisfaction scheme shifts, most dramatically when the collection of symbols is still small, but growing rapidly, as it is with children.
This is just one approach I remember a bit of, and only a tiny start on confronting the issue of realism using this research. What do we say about the child seeing a field full of cows and excitedly announcing "doggie!" or "doggies!"? One thing is clear, that the child would
not have been "trained" to say this, because that's not what adults say, so an account that passes by issues of categorization is missing something. Is it plausible to focus only on categorizing the communicative situation, and describe the child as thinking,
this is an appropriate occasion for uttering "doggie"? There's still over-generalization, but it's different. --- And what happens when the child does acquire
cow but still doesn't have
mammal? Does that mean cows are, to them, a kind a of dog?
One thing is clear from trying to write about this: hard as it is, it's easier to talk about a partial language than a partial world, but I think we have to find a way to get at the latter as well. If you don't know anything about chess and watch a game, you see everything the players do, under one scheme of description, but I really want to say that "black's king is in check" is not a possible fact for that observer -- neither true nor false if you must -- and you could describe this as not being able to categorize positions by whether black's king is in check. We might say the observer's world is not partial in the sense that it has less stuff in it, but that it makes slightly less sense. But it's also true that the observer cannot
see check, and so there is something in the world of the players that is not part of the observer's world. (I think James somewhere gives an example of a dog, seeing perfectly the interactions among humans but attaching necessarily different meaning to it.)
So there's some stuff about realism.