Linguistics and Cognitive Science

April 17, 2006 10:40 AM

What sorts of things has linguistics added to the field of cognitive science?

Im looking for concrete examples of ideas and concepts that the field of linguistics has contributed to the field of cognitive science. Just a basic idea of the sorts of things that linguists theorize about how the mind works, the sorts of research they do etc.

posted by tranceformer to writing & language (15 comments total)

Well, I know (from somewhere) that babies process vocabulary in the same parts of the brain as adults, but use their entire brain to process grammar. Grammar becomes more localized when they mature, which is why it's much more difficult for older people to learn a new language than babies.

You might also want to look into the Sapir-Worf hypothesis, but that's more of a liguistic-cultural thing.
posted by borkingchikapa at 11:00 AM on April 17, 2006

What stands out in my mind, from one of my classes from about 3 years ago, is the idea of phoneme solidifying over time.

For example: Someone raised in a country in asia learnes to speak their language, and as he/she gets older different parts of the brain are maped out for different phonemes, or sounds. Then that person comes to the US and tries to learn english. In english there is a /ll/ sound, that most asian languages do not have. So when that person tires to say 'hello' they say 'herro'....but they aren't only saying the /ll/ sound as /rr/, they are also hearing it that way, so it seems totaly normal to them. Over time their auditory cortex has been mapped out for different sounds in different areas, sounds that don't exist in their cortex fall onto sounds that are already there.

hence, once you have these areas defined, it becomes much harder to learn a new language.

I don't know if that made sense to you, but it makes sense in my head ...
posted by TheDude at 11:17 AM on April 17, 2006

i have a strong recollection - but perhaps incorrect/warped - of languagehat telling me that linguistics was not a science. so i suspect there might not be much. no doubt he'll be a long in a mo...
posted by andrew cooke at 11:45 AM on April 17, 2006

The intersection of cognitive science and linguistics is called cognitive neurolinguistics (or cognitive neuroscience of language -- CNSL). Whole journals have been devoted to the subject (e.g., Brain & Language).

It's tough to seperate out language from all the other cognitive processes. The brain has no module in it that says "language only." For instance, Broca's area has long been thought to be the seat of syntax, but recent research has shown that it also plays an important role in working memory. Are these two separate functions or just two sides of the same coin? There comes a point when it really stops looking like linguistics and starts looking like brain science.

The research I am involved with uses event-related potentials (ERP), a tool in EEG, to examine how the brain does language. It turns out that a characteristic "blip" is produced by the brain whenever it encounters a semantic anomaly, or another when it encounters a syntactic one. We can show people sentences and record what kinds of blips are produced and gain some insight into how they are processing these sentences.

Another example, from another domain of linguistics is the McGurk effect. The McGurk effect provides evidence that speech perception doesn't only take input from the auditory modality, but the visual one as well. This video should thoroughly confound you.

CNSL has also given us clues as to how native and non-native speakers of a language use the language neurologically and what strategies both employ in speech.

I could go on about this all day. If you have any specific questions, feel free to ask.
posted by pealco at 11:46 AM on April 17, 2006

i have a strong recollection - but perhaps incorrect/warped - of languagehat telling me that linguistics was not a science.

I don't remember, I may have said something like that, but it would depend on the context and what kind of science we're talking about. It's not a hard science like physics or chemistry, but if (say) geography is a science, then so is linguistics. But I personally wouldn't say it's made a huge contribution to "the field of cognitive science"; on the other hand, my area of linguistics was historical/etymological, so the cognitive stuff is not really my thing. In any case, TheDude's idea sounds in the right ballpark; just be sure to ignore any Chomskyan bullshit about genetics and universals.
posted by languagehat at 11:50 AM on April 17, 2006

just be sure to ignore any Chomskyan bullshit about genetics and universals

Languagehat: Why do you say that? I also vaguely recall you making some similar comments about The Language Instinct. I ask because I'm genuinely interested and want to educate myself about this.
posted by pealco at 11:55 AM on April 17, 2006

I remember learning about two types of aphasia, and how it was discovered that each type was indicative of damage to a different area of the brain.

Aphasia (in case you didn't know) is a language disorder. In Wernecke's aphasia, you get long gobbledy-gook sentences; in Broca's aphasia, you get short sentences with words left out.

On reflection, it's a question in my mind whether aphasia was originally studied by linguists, but I thought of it because I originally studied aphasia in linguistics class. Not that that makes much difference.

This example may be the reverse of what you asked for; It may be that cognitive scientists discovered & studied Broca's and Wernicke's areas in the brain, which led to insights about the way we process language and language itself.
posted by amtho at 12:25 PM on April 17, 2006

Exactly right, amtho. It seems that cognitive science has informed (what is classically thought of as) linguistics more than the other way around.

I know my syntax professors have sometimes cited psycholinguistic evidence to show why we do some operation one way over another way.
posted by pealco at 12:35 PM on April 17, 2006

"...just be sure to ignore any Chomskyan bullshit about genetics and universals."

Yeah, a lot of linguistic theory seems to be pretty rattletrap when situated in real language settings. You'd be doing yourself a disservice, however, if you ignored the fact that Chomsky was instrumental in ushering out behaviorism as the dominant model for human cognitive processes -- particularly language. This contribution alone is immensely significant.

Also, what TheDude mentioned was covered previously on AskMe: it's called categorical perception.
posted by cog_nate at 12:36 PM on April 17, 2006

Hmm. Well, cognitive science is by definition a broad field, but in general I see two primary ways of studying cognition: either directly evaluating what the brain does, by ERP or fMRI or some other physiological measure, like pealco's research, or computationally modeling what the brain does, more in the abstract. The latter is (sort of) what I do. So here's another f'rinstance:

I do research in lexical selection and retrieval. That is, when you see an object (a cat), how does your brain retrieve the linguistic label CAT? The average person knows many thousands of names for objects - how do we pick the right one, and weed out the rest? What is the cognitive "algorithm" by which this weeding-out process takes place?

We know from behavioral experiments that when people hear "ca-", lots of possibilities are active in the mind, like CAT, CANDY, CARROT, and so on. At the same time, to a lesser degree, things that are semantically related to these are also activated, e.g., DOG, LOLLIPOP, TURNIP. We are constantly entertaining an ever-changing network of lexical possibilities.

In a related vein, there is lots of linguistic research done on speech errors - when we make mistakes, they're very nonrandom. If we meant DOG we're more likely to mess up and say CAT or LOG than something like SHOE.

So this all leads to interesting questions that branches of cognitive science try to answer: how do we model these similarity networks? Are words that are similar in language (in semantics, phonology, whatever) simiarly represented in the brain?

As some other posters have said, though, this also isn't really a case of one field influencing another. It's just the confluence of two fields: how the mind works, and how languages work. So, I apologize if I missed your point.
posted by miagaille at 1:12 PM on April 17, 2006

It's not a hard science like physics or chemistry, but if (say) geography is a science, then so is linguistics

I think theoretical linguists have a somewhat different perspective. Linguistics is nothing like geography, at least what I know of geography. Linguistics is an empirical science -- the data consists of native speaker judgments about what sentences are grammatical, how words are pronounced, and under what conditions sentences are true. This is the kind of data, by the way, that can be reproduced and tested under experimental situations. The goal is to form a theory that correctly predicts the data as we find it. This may not be a hard science, but it is much harder than geography, or even something like sociology.

A consequence of this approach is that many linguists do not theorize directly about the mind. A correct theory that makes the right predictions will of course correspond to a correct theory of how a human brain produces and processes language, but there is no need to theorize about the brain directly -- a mathematical model is in many cases enough. Linguistics is to psycholinguistics/neurolinguistics as formal language theory/algorithms are to the actual implementations of computer programs, except that in this case we've been handed a computer to figure out from scratch. In this kind of situation the mathematical model based on the data can be just as useful (or more useful) in informing the brain-studies than vice versa.

There are many linguists and psychologists who work at the interface -- some examples are Chuck Clifton and Lyn Frazier at UMass, Ted Gibson at MIT, and Sheila Blumstein at Brown (to pick a random sample). The best way to understand the contribution of linguistics to cognitive science is probably actually to read some work by these people, and others like them.
posted by advil at 2:04 PM on April 17, 2006

It depends on what you mean by "cognitive science". There are all sorts of subfields in linguistics (psycholinguistics, semantics, pragmatics, computational linguistics) that are typically thought of as being a part of cogsci. Any advances in those areas would perforce be advances in cognitive science. Cognitive science includes high-level psychological (and linguistic) studies of the mind as well as the low-level neurological studies. (Or: what miagalle said)

What you really want to ask, I think, is what our models of natural language processing have added to our understanding of the brain. My impression is... not too much yet. The point of the cogsci project, I think, is that our linguistic and psychological theories put top-down constraints on the type of structures we look for in the brain, and reciprocally, our study of structures in the brain influences our models of linguistic and psychological processing. But we certainly haven't synched things up.
posted by painquale at 2:06 PM on April 17, 2006

Languagehat: Why do you say that?

Because it's what I believe. I freely admit I'm not up on recent developments, and it's possible great strides have been made, but when I was au courant all such claims turned out to be pure pulled-from-the-ass hoo-ha upon investigation. People love to make resounding claims about The Nature of Mind/Brain/Humanity; they don't so much love doing the painstaking research and analysis necessary to eliminate the (inherently much more likely) scenario wherein their brilliant idea is wrong, wrong, wrong. That's why I prefer the more philological end of the field: many people may think it's boring to root around among documents and/or field recordings looking for rules and exceptions, but at least you can be confident you know what you're doing and aren't spouting nonsense. When, like Chomsky, you have to totally revamp your theory every few years because your last version turned out to be hoo-ha, you have to have a high tolerance for embarrassment to keep at it.
posted by languagehat at 2:48 PM on April 17, 2006

People who would like a single-perspective attempt at answering this question would probably enjoy reading Steven Pinker's The Language Instinct. It really does try to illuminate the contributions of cog sci to linguistics, from one very particular point of view.

It so happens that that point of view is what our languagehat calls "Chomskyan bullshit about genetics and universals;" this is another way of saying that the conclusions, though stimulating to read about, are widely disputed.
posted by ikkyu2 at 3:13 PM on April 17, 2006

thanks for responses. btw what i mean by cognitive science is this
posted by tranceformer at 7:35 AM on April 19, 2006