Thursday, October 11, 2007

Paca tells an academic story

The field of linguistics has been given birth to several times. At least one of those times is often said to have occurred in 1786, when Sir William Jones gave a lecture to the British Royal Society. The meeting went something like this:

Sir William: Hallo, Sir Chaps!

Society Chaps: Hallo, gov'ner.

Will: So, my dear royal society lads, I've been tossing about India on a royal old holiday-

Chaps: Right'o, Billy Boy. Lovely place, that India. I think we should colonize it.

Will: Yes, yes, jolly good. They'll be happy that we did. But you know, I looked at some Sanskrit after me cricket match and, blimey, if it doesn't look a might bit like Greek.

Chaps: Shove off! You're out of your blooming mind, you are.

Will: No, no. I think it does. In fact, some of the words seem awfully like Greek, Latin, and Persian.

Chaps: He's off his chum.

Will: No, no. I think, you know maybe, I was wonderin' if they might not have once been the same language.

Chaps: The same WHAT?!

Will: Sorry, sorry. I meant the same bloody language.

Chaps: Ohhhh. Yeah, and we're all descended from monkeys, too. Uproarious laughter

And so the idea of Indo-European was born and Europe was off to the races from then on, studying how languages were connected and how they might change. There was a shift in the early 20th century to more and more sophisticated ideas of how languages change, which meant better understanding what languages are, but historical linguistics was arguably the driving force.

Over in the New World, linguistics of the early 20th century was a branch of anthropology, to a large degree, with a focus on learning and documenting Native American languages. Due to this, the emphasis was clearly on language as a cultural thing. There's enormous variability in language and a child learns one by participating in that culture. This focus on learning language in the late 40s had gotten merged to some degree with the popular form of psychology at the time, behaviorism, most often associated with B.F. Skinner. Behaviorists' focus was also on learning with all that work on conditioning and making dogs salivate and what not, as you may remember from your Intro to Psych class (and which made me think I never wanted to be a psychologist). Behaviorism also forcefully argued that the mind was not something you could study scientifically. No one's ever seen a mind after all. But you can study behaviors, and people sort of become these input-output machines; you give them a stimulus, they associate this stimulus with some other stimulus, and respond.

Then along came Noam Chomsky who killed behaviorism and changed the focus of linguistics to this day. He did it with a book review. Let me pull it out of my file cabinet. Here we go: A review of B.F. Skinner's Verbal Behavior by Noam Chomsky in the journal Language, volume 35, issue 1, 1959. Here's a critical sentence from page 2.

"One would naturally expect that prediction of the behavior of a complex organism (or machine) would require, in addition to information about external stimulation, knowledge of the internal structure of the organism, the ways in which it processes information and organizes its own behavior. These characteristics of the organism are in general a complicated product of inborn structure, the genetically determined course of maturation, and past experience."

Chomsky proceeds to argue that you can never understand language in a behaviorist model because it specifically exludes the internal structure of the organism. You need to know the biology of humans and, particularly, you need to know their minds. Moreover, and this is THE critical shift, from studying language, we linguists can actually figure out the mind. We are going to look at the forms that languages can take from all over the world and see what they have in common. We are then going to create a theory about our language data, which we will call a grammar; in fact, it is the Universal Grammar, as it is a grammar to explain all forms that language takes. This is the real term, not something from Star Trek. This universal grammar is the set of rules that underly all languages from Albanian to Zulu, and this Universal Grammar is the biology that Chomsky thought was missing from Skinner's account.

Indeed, the fundamentals of language are never learned at all -- take that, you early 20th century American linguists! UG is a mental organ, but an organ just like the heart or lungs or brain. Universal Grammar is part of our very genes and language grows inside a child just like a heart does. Now, Chomsky's not a fool; he's wickedly smart. He's not saying that Hindi or Navajo or Portuguese is in our genes, but all languages follow this set of universal rules and that set of rules very much is. It's one of the things that makes us specifically human. I've previously written up a lengthy post about the arguments for this. Here's the link.

And here's where I finally come in.

Much of the work in the Chomskyan tradition continues to assert that it is revealing our biology and the results of our generic endowment, as they refine the principles of Universal Grammar. There are books and journals and conferences that have recently taken on the name "biolinguistics". But let's go back to how we find out what our biology is in this tradition that's lasted almost 50 years now, because I don't think it truly works.

The method is, we take a bunch of linguistic data, the possible grammatical structures of a language, and we look for the pattern. That's what virtually every single class I've taken here has taught me to do, and that's what we all did on this blog just a week ago with the child data. Take a bunch of things that B said, and then see what the pattern is. Ohhh!! It has to do with stress! When I did that, I just introduced the notion of Stress into our theory. And if I kept looking at language after language and they all had stress, I might start to wonder if Stress is part of the Universal Grammar somehow with which children are born. If the data holds up, I might know that everyone has stress, but maybe everyone learned it. If it is learned, then it's not inborn with our genes. And so I think of a way that child might learn stress from the language she hears. If I don't find evidence that it IS possible to learn stress from the sentences around her, then stress is part of Universal Grammar. This is called a Poverty of the Stimulus argument, because the quality of the stimulus (the language we hear as children) is poor, and the argument goes: Propose a principle that fits a pattern in language, show that it cannot be learned, and conclude that it is part of biology.

But notice something a bit odd about all this. I keep talking about language forms and structures over and over. Here's the form of a sentence in Urdu; here's one in Tongan. And now we propose a theory of abstract rules that generates these possible human sentences and doesn't generate any others. Forms, forms, forms. And yet biology isn't forms really, is it? It's actions. There's no biological rule that says, "humans cannot run faster than 27.3 mph." Instead, it just gives you a bunch of muscles and a brain to move them. And you move them step by step as fast as you can and you go however fast you go.

Philosophical interlude. If you ever had to read Plato in an Intro to Philosophy class, you probably learned about two things. One is the Cave and how we've got shadows dancing around and all that stuff. The other is likely to be that Plato believed in Forms with a Capital F (also called Ideas often). Forms were something like the universal structures, the ideal form, to which the real world was only an approximation. However, Plato had a problem with his Forms. It wasn't clear how anyone could ever learn one. There's a famous scene (well, famous depending on the crowd you hang with) in a dialogue called the Meno, in which Socrates is hanging out with Meno and they are trying to figure out what Virtue might be, and if it might be learned. But they cannot figure out any way Virtue can be learned, and so Socrates proposes the idea that maybe it's not learned at all. Maybe, it's somehow inborn in us. He then has an extended example of a boy correctly answering geometry questions even though he's never been taught geometry. It's the first Poverty of the Stimulus argument, and Chomsky knows it because he often calls the problem he is trying to solve Plato's Problem.

I believe that Aristotle, however, solved Plato's problem a long time ago, or at least he points the way out. He was trying to understand cause and effect, and he ended up proposing four types of causes. One is the final cause and it is why we do something. If I go to the store to buy a gallon of milk, my purpose is to buy milk and it is the final cause. There's also the material cause and it's just what I am made of. Flesh, bones, hunks of burning love, etc. Next up, we have.... drum roll please.... the formal cause. This is the forms that the thing will take. The formal cause of an oak tree is the structure of an oak. The formal cause of me going to the store would be all the descriptions of store-going behavior, things like, how fast people walk, acceleration, the sort of directions I take. Structures and forms. The last one is called the efficient cause, efficient because it produces the effect. The efficient cause is the billiard ball hitting another one, causing it to roll away. The efficient cause of me going to the store is putting one foot forward and then another. It's what is making me change from just standing to walking. An efficient cause initiates the change. It's a procedure and a real action that makes something move.

Now let's put it all back together.

Right now we know very, very little about how our biology allows us to speak and understand language. At best we can name areas of the brain that light up when we ask questions or read a relative clause or try to remember a word. That's great and all, but it tells us very little about what the heck our brain is doing when it's all lit up like Church Lady back in college. But whatever the biological mechanism is it's going to be something that takes one bit of neural information and moves it over here causing a change. It's going to be something like our running example earlier, where the brain caused our muscles to move. It is not going to be a rule like "No faster than 27.6 mph." It will be efficient in Arisototle's terms, not formal in Aristotle's terms.

Another example is that of drawing a circle. To draw a circle, you have to draw a line of a certain length. How long? Time for Intro to Geometry now. It's the diameter times pi, right? That's the best formal description of the length of a circle's outside line, and it is a really really useful thing to know, as half the world's engineering is based on it. However, it's not clear that that perfect formal description of a circle has much of anything to do with people actually drawing circles. The way I draw a circle is to try to eye the center and then I just move my pencil in a curve, doing my best to stay the same distance from the center the whole time. Then I stop when I hit the beginning of my line. Someone else might imagine a circle in their head and create the image in their mind on the paper. There are probably several possibilities. But do any of them need to involve the notion of the number pi? The ideal formal model of the circumference of a circle, while a great contribution to knowledge and science, might have little to do with drawing actual circles. And if we look for evidence of how the length of a circumference is learned, using our best formal model of 'pi times the diameter', lo and behold, we may find no evidence that it is learned at all. But do we conclude that the ratio pi is in our genes? Not at all. It's not learned, because it's not used to draw circles in the first place, whatever the model's other merits.

Classic linguistic theory of the last 50 years has been building models that are just formal. But biology is efficient. It's about the steps from one place to the next. The fact that a formal model cannot be learned does not reveal our biology, because our biology doesn't use that model. It uses an efficient one. Whatever our biology does when we use language, it's going to be a series of steps to take a thought and turn it into speech or a series of steps to take a sound wave and figure out what it means. If the linguistic theory is not about those steps, it isn't moving us towards knowledge of the missing internal parts of the organism that Chomsky so long ago said was missing from behaviorism.

There you go. One and a half million caveats need to be inserted here, but this is really long already. Also, the exciting breakthrough I wahoo'd about is essentially about 2-3 paragraphs of this. Everything else is to transport you folk to the place where you might understand those paragraphs.

8 comments:

Robin S. said...

Hi paca,

I've just read through this, and it's really fascinating. Are these points:

"whatever the biological mechanism is it's going to be something that takes one bit of neural information and moves it over here causing a change. It's going to be something like our running example earlier, where the brain caused our muscles to move. It is not going to be a rule like "No faster than 27.6 mph." It will be efficient in Arisototle's terms, not formal in Aristotle's terms,"

"Classic linguistic theory of the last 50 years has been building models that are just formal. But biology is efficient. It's about the steps from one place to the next. The fact that a formal model cannot be learned does not reveal our biology, because our biology doesn't use that model. It uses an efficient one."

..the crux of your dissertation/idea? If so, I can see more clearly now what you talked about earlier, concerning the marriage of philosophy and linguistics.

pacatrue said...

Hi, robin s. Wow, someone actually read this!?

Yes, the paragraphs you quoted are very much the main idea. That was the wahoo idea. This won't be my dissertation, so people don't have to hear about it for the next year and a half. However, I do hope to get an article out of it, so it might come up a bit more.

Ello said...

Paca, you made my small brain hurt with information overload. You are some sort of linguistical genius boy aren't you!

Robin S. said...

Hey there, paca,

I told you I love this stuff - keep it coming, please. Edification - it's a very good thing!

So what's the dissertation on, then? And what's the relationship between it, and the wahoo idea?

Mamaebeth said...

i read through it... although, it was difficult.
i took some linguistics as part of my Anthropolgy degree and i took some philosophy as part of my religious studies degree,
BUT this was so way over my head.

i'm glad you have it figured out though. i tend to grasp concepts in the abstract and then it takes me a while to get them to the concrete... so i understand the "wahoo"

Church Lady said...

This was interesting to read. What I gathered is you're trying to map the cognitive functioning of the brain. I think. Hmmm...I think I get some of it.
I took a linguistics course in grad school. The professor believed that language ability is what distinguishes man from ape. I disagreed, and still do, though it's not my field. (I posted something about Koko a while back). I think first you have to define language. I guess I have a broader definition that's closer to communication.

Fascinating post!

Sammy Jankis said...

I'm often surprised at how leading pioneers and thinkers appear in the study of such a broad array of fields. Chomsky, the great philosphers, Skinner, etc. appear over and over in linguistics, psychology, philosophy, sociology and other fields and their work is used to generate a variety of different theories and utilizations. It is really inspiring that one person's work can be used for so many varied purposes, and probably a far greater number than the person was originally addressing when formulating the ideas.

pacatrue said...

Yeah, I agree, Sammy. The philosopher Alfred North Whitehead famously stated that all of Western thought is a series of footnotes to Plato. It's largely true.

Chomsky himself is one of the 10 most cited people in all of academia and perhaps the only one alive. I actually think this is somewhat damaging to the field of linguistics, but that's a post for another time.