Before I forget: Why universal syntax means means mentalese is dumb (maybe)
Before I forget, I wanted to tell you about one of the things Whitney Wood and I were talking about.
As I mentioned, Whit does research on (natural language) syntax. I'm skeptical about the value of a lot of syntax research and told him so, which put us into a conversation which he later described as "defending his existence" from me. I didn't really intend to do this, but those of you who know me too well will recognize that I sometimes have this affect on people.
"Now I'm no fan of the Sapir-Whorf hypothesis," he said politically, "but I think syntax tells us a lot about the way we think. There are all these universal features of human syntax [like the existence of verbs and nouns in all languages]. From a pure syntax perspective, there's no reason for it to be that way; it's a total mystery."
Or he said something like that. Then I pulled a tentative hypothesis out of my ass that I ended up liking so much that I became convinced of it by the end of the evening.
First, a digression. Those of you who I've talked to about this sort of thing before know that I think that the idea that concepts/sentences are structured in some way that's similar to words/propositions is dumb. This attitude makes me scowl at everything from Kantian discursive concepts to Fodorian Mentalese. And I just want to point of that it is a symptom of this kind of thinking that any sort of explanation for the structure of language is impossible. Mentalese could be, in theory, any arbitrary kind of coding. [Yes, there are some major counter arguments to be made here, but I'm going to run with this because my wrists hurt]
But look what happens when you deny Mentalese. You are stuck with thoughts which inhabit different sensory and cognitive modalities. You concept of 'dog' ceases to be some description of dogs, but associations between the way dogs look, the way they sound, the way they smell, what you can do with a dog, etc. Everything about your DOG concept is wrapped up with ones intentionality toward dogs.
Meanwhile, you are faced (either as an individual language learner, or as a linguistic community, or as a species--I'll be agnostic about all that here) with the task of communicating with others. Again, I'm going to borrow the vaguely Merleau-Pontian line that communication involves the inhabiting of one another's intentions. Again, this doesn't make any sense in a Mentalese model of thought, but it makes perfect sense if you understand concepts and beliefs as modally distributed. For me to communicate that there is a dog over there to you, I need to make you share with me an imagining of a dog, an spatial and muscular orientation of yourself by which you understand what "over there" means in terms of your potential for action, and so on.
What is language, then? It's solving a terrific problem. This messy network of mentally graven images and sounds and actions needs to be transmitted somehow into a brief sequence of grunts and hisses. It's the most important problem of information compression to ever be faced by humankind.
The solution, then, is the same as in all problems of encoding. To be efficient, the language needs to take advantage of the invariants in the message being sent. The most common messages need to be most easily sent; the most essential structures of thought must be mirrored in the structure of the code.
And so we have an explanation for syntactic universals. The fundamental duality in our experience between the world as declaratively and sensually known and the world as infused with our own agency (between sensory process and motor cortex, between explicit and implicit memory) emerges in our linguistic codes as the noun/verb divide.
"What about verb nominalization?" he asked--that's when a verb is used as a noun, as in "the kick." Of course: we can witness the actions of other, or our own actions, and experience them as objects. One would predict then that we should not be able to verbify nouns.
"But that happens all the time."
"Like when?"
"Like 'to pen'."
I told him I thought that was the exception that proved the rule.
That's it--just wanted to get that off my chest.
As I mentioned, Whit does research on (natural language) syntax. I'm skeptical about the value of a lot of syntax research and told him so, which put us into a conversation which he later described as "defending his existence" from me. I didn't really intend to do this, but those of you who know me too well will recognize that I sometimes have this affect on people.
"Now I'm no fan of the Sapir-Whorf hypothesis," he said politically, "but I think syntax tells us a lot about the way we think. There are all these universal features of human syntax [like the existence of verbs and nouns in all languages]. From a pure syntax perspective, there's no reason for it to be that way; it's a total mystery."
Or he said something like that. Then I pulled a tentative hypothesis out of my ass that I ended up liking so much that I became convinced of it by the end of the evening.
First, a digression. Those of you who I've talked to about this sort of thing before know that I think that the idea that concepts/sentences are structured in some way that's similar to words/propositions is dumb. This attitude makes me scowl at everything from Kantian discursive concepts to Fodorian Mentalese. And I just want to point of that it is a symptom of this kind of thinking that any sort of explanation for the structure of language is impossible. Mentalese could be, in theory, any arbitrary kind of coding. [Yes, there are some major counter arguments to be made here, but I'm going to run with this because my wrists hurt]
But look what happens when you deny Mentalese. You are stuck with thoughts which inhabit different sensory and cognitive modalities. You concept of 'dog' ceases to be some description of dogs, but associations between the way dogs look, the way they sound, the way they smell, what you can do with a dog, etc. Everything about your DOG concept is wrapped up with ones intentionality toward dogs.
Meanwhile, you are faced (either as an individual language learner, or as a linguistic community, or as a species--I'll be agnostic about all that here) with the task of communicating with others. Again, I'm going to borrow the vaguely Merleau-Pontian line that communication involves the inhabiting of one another's intentions. Again, this doesn't make any sense in a Mentalese model of thought, but it makes perfect sense if you understand concepts and beliefs as modally distributed. For me to communicate that there is a dog over there to you, I need to make you share with me an imagining of a dog, an spatial and muscular orientation of yourself by which you understand what "over there" means in terms of your potential for action, and so on.
What is language, then? It's solving a terrific problem. This messy network of mentally graven images and sounds and actions needs to be transmitted somehow into a brief sequence of grunts and hisses. It's the most important problem of information compression to ever be faced by humankind.
The solution, then, is the same as in all problems of encoding. To be efficient, the language needs to take advantage of the invariants in the message being sent. The most common messages need to be most easily sent; the most essential structures of thought must be mirrored in the structure of the code.
And so we have an explanation for syntactic universals. The fundamental duality in our experience between the world as declaratively and sensually known and the world as infused with our own agency (between sensory process and motor cortex, between explicit and implicit memory) emerges in our linguistic codes as the noun/verb divide.
"What about verb nominalization?" he asked--that's when a verb is used as a noun, as in "the kick." Of course: we can witness the actions of other, or our own actions, and experience them as objects. One would predict then that we should not be able to verbify nouns.
"But that happens all the time."
"Like when?"
"Like 'to pen'."
I told him I thought that was the exception that proved the rule.
That's it--just wanted to get that off my chest.