Machines Like Us interviews: Daniel L. Everett

Wow! Thats pretty weird! I gotta say I feel a great deal of cognitive dissonance with Dan Everett's thesis that there is no language instinct! Although Chomsky is a pretty weird character with some way-out-there ideas, I always thought he was right on with the language instinct idea.

As a Gestaltist, I am more of a nature than a nurture kind of a guy. And then when Everett claims that even interpreting photographs as 3-D scenes is a learned skill rather than an intuitive process, he really lost me! Nobody has to be taught that a little shadow under a cartoon object indicates that it is hovering over the ground. Nobody had to teach us "this symbol means hovering." We learned it as a kid the first time we saw it, and learned "hey, this little shadow has the effect of making it look like it is hovering!" It IS all innate, although of course there are aspects that are also learned.

Here's what makes more sense to me. This is a case similar to Margaret Mead. She was a researcher sent to some remote Pacific islands back in the 30's or 40's to live with some primitive people (she was one of the very first) and she came home with this extraordinary hypothesis that her native people did not have family relations the way everyone else does, but they had sex willy-nilly with whomever whenever they fancied. It was an earth-shattering hypothesis with profound implications that had all the journals a-twitter for many decades, until many many years later (the LIE has gone three times around the world before the TRUTH even had its boots on!) it was discovered that these were pre-conceived notions that Margaret Mead projected onto her people. They picked up on subtle cues of hers that that was the kind of story she liked to hear. Her eyes would light up and she got all excited whenever they told her things she wanted to hear, and so they learned to humor her, probably elbowing each other furtively and rolling their eyes, trying to out-do each other in coming up with anecdotes that made Margaret Mead happy.

Think about how much more interesting and significant Dan Everett's "work" is if it shatters a widely held established theory of language. There is a strong incentive to make earthshaking discoveries even if there are none there. It does not even have to be deliberate -- Margaret Mead was totally unaware of how she projected her preconceived notions on her people. She really was sincere in her belief to the end.

The case that clinched me on the language instinct was the fact that slaves from different places who spoke different languages, when all brought together in a foreign land, would all learn a "pidgin" (e.g. pidgin English -- me no talk good but me say what need say). But in the very next generation, the children of those slaves, spontaneously and instinctively developed an elaborate grammar for their parents' pidgin, and this has happened more than once in different places and different times. It would take a LOT of counter-evidence to dis-convince me of that evidence!

And what clinched me on the innateness of interpreting photographs was Gestalt theory and Gestalt illusions -- like the Poggendorf illusion where one line looks longer than the other due to the influence of a pair of converging lines -- you don't even need to see it as a 3-D scene to be fooled by the illusion.

I applaud Everett for finally seeing the folly of religion -- that was an interesting part of the interview. But the fact that he was fooled by it for so long might suggest that he is susceptible to getting stuck in small-minded paradigms ignorant of the larger surrounding BIG PICTURE perspective.

It is often suggested that I have been hoaxed like Mead apparently was or that one way or the other the motivation to say something beyond the facts in order to become famous or some such has led me astray. My claims are being tested and debated by many people, including quite a few who have gone to the Pirahas. And the conclusions on perception, numbers, grammar, and so on are quite compatible with a large number of theories and proposals.

To take one example, the claim that recognition of 2D representations is not innate is not original to me. There is a group of psychologists at MIT and Stanford, among other places who believe exactly that this could be the case for every human being. And our initial tests in the field with Piraha subjects, myself along with MIT psychologists Ted Gibson and Evelina Fedorenko and Stanford psychologist Mike Frank, are consistent with this view. The results of these experiments should be published in 2010. More experiments are planned. The conclusion seems to be, clearly, that photographic interpretation is not innate. Whatever 'clinched it' for anyone, it is an empirical hypothesis that must be tested empirically. Tests show that it is learned.

On the innateness of Language, as Mike Tomasello of the Max Planck Institute for Evolutionary Anthropology puts it in the current issue of GEO Magazine (German), 'Universal Grammar is dead.' I would say that only a minority or at best a very small minority of linguists believe in UG or a Language Instinct. Data like Piraha support this, but people's opinions on this are based on many empirical studies, especially like Tomasello's comparative primate research. To better appreciate my arguments and the evidence for them, I suggest reading my "Don't sleep there are snakes: life and language in the Amazonian jungle" (Vintage). I address these issues in more detail in the forthcoming "Cognitive Fire: Language as a Cultural Tool".

More than 25 linguists, anthropologists, psychologists, journalists, and even one philosopher have been with me to the Pirahas to test these ideas.

A movie on this research with lots of footage of the Pirahas and experiments with them will be released in 2010, 'The Grammar of Happiness', from Essential Media and Entertainment of Australia. Many linguists and others will be interviewed about the issues on camera.

Also, there has been a debate on this for some time. The June 2009 issue of the journal Language, the main journal of linguistics, dedicates about 100 pages to a debate on my ideas. The hypotheses are serious, based not only on my own experience, but so far standing up to testing by a wide range of scientists.

Dan Everett

Hi Dan,

I'm sorry if my tone was brusque, and I apologize for the final insensitive ad hominem attack. Norm sent me the link to your interview, and I responded with an informal email back to Norm, originally intended as a private exchange, although I consented to Norm's request to post it. Nevertheless, I would have worded it more sensitively if I had planned to release it publicly, and I would never have made the final remark, which was intended at least half in jest.

I should also explain that linguistics is not my specialty, I am more into visual perception and Gestalt theory, so my views on Chomsky are those of a dabbler, I defer to your greater literacy in that field, and I acknowledge your responses to my critiques.

I will say however that I stand by the general message of my commentary, I do believe that language and visual interpretation are more hard-wired than learned. In the case of visual perception the Gestaltists conducted numerous detailed experiments that were pretty conclusive in my mind. But I also acknowledge that many continue to differ on this question, and the nature v.s. nurture debate is one that seems to swing back and forth like a pendulum, taking turns alternately coming into and out of fashion. So the fact that many prestigeous researchers side with you on the nurture side of things is not persuasive. I have noticed however that just as some people tend to lean politically liberal and others conservative, and some tend to be believers while others tend atheist, so too I find that some people lean toward the nature, and others the nurture side of the debate, and as in political and religious debates, people are rarely persuaded by reason and data that support the other side. Its a paradigm thing, it has more to do with the initial assumptions we bring to the debate in the first place. Your intimate experience with the religious world will surely have aquainted you with that phenomenon!

My own experience with the intractability of paradigm debates is illustrated graphically in this cartoon, you may find it amusing.

http://cns-alumni.bu.edu/~slehar/cartoonepist/cartoonepist.html

Even if I disagree with your conclusions, I applaud your earnestness and dedication, and I enjoyed reading the interview. Good luck with your work!

Steve

Wow, I never knew linguistics was so important, and had such heady consequences concerning human cognition. Mr. Everett seems to be quite a character who has done some remarkable things. Amazing what some people do with their lives. I always figured that our capacity for language was in-born, but after reading the interview now I'm not so sure. I wonder if it is even possible to think without language. I mean, it feels impossible for me to have a thought without thinking of words that accompany that thought.

It seems to me that when one first learns a new language or teaches a first language to an infant, one's natural inclination is to begin by naming objects: nose, ear, crayon, car. It makes sense to suppose that the first languages came about in order to fulfill the same basic need: to describe objects to others during times of necessity, such as while coordinating the hunt of wild game. From there, maybe the task became one of refining and clarifying meaning with the invention of adjectives, verbs, and other grammatical structures. Mr. Everett, if you happen to read this: does that sound like a plausible explanation for why language evolved?

Dear Cindy - your proposal for the origin of language is a very plausible one. I look into the nature and origin of language more in my book, to appear in late 2010 or early 2011, Cognitive Fire: Language as a Cultural Tool.

Steve - if scientists are unable to turn these issues into an empirical debate, there is not much point in having or offering opinions on either side. Universal Grammar these days makes no predictions at all. It is simply a way of talking. I discuss this in a long article in the journal Language, published in June of this year. If you want to place your faith in an idea that is nonempirical, that is up to you. Part of the problem is that up until now, the press and the popular literature on language has been dominated by people sympathetic to Chomsky, largely a matter of personalities, rather than research. I cited Tomasello's work, because it is richly empirical, not because Tomasello is a big name (he certainly is, though). In any case, the interest in all of this discussion is how to best understand the data we encounter. All the best. Dan

I find myself being quite receptive to what Professor Everett said in this fascinating interview. My own (in no way authoritative, but also not unique) views on the brain are that it is, centrally, a learning system that learns patterns between different inputs and outputs occurring over time, which then uses these patterns to determine optimum outputs. I have always found the idea of "universal grammar" difficult to reconcile with such a view, where all the structures that would give any kind of grammar context have to be built from the ground up by experience. A newborn child does not understand the concept of "being late for work", and few of us would claim that any such language structure in a newborn child's brain could represent such a thing: That would be fairly uncontroversial. I think, however, that the idea of some kind of universal grammar system, with some kind of "option switches" which set up a specific grammar has the same problem. I think you would need to start with an awfully large amount of sophistication about how the world was represented to give such a system any context - and without such sophistication, without some structures that map onto the complexity of the world in some way there at the start - I don't see what you would anchor this universal grammar to. I am not making the mistake of here of thinking that children are supposed to start with a preformed language: I think the problem arises even with a pre-built language making system. Whatever neurological system you have to implement this universal grammar, it would have to build something which then interacts with all kinds of structures that have not been developed yet. It would seem to me to be like trying to write some computer code to make some very high level change to a system which has not yet been programmed – like someone in 1965 trying to make a webpage that would work in the year 2009, on the basis of what was expected to happen in computing. I hope I have made some kind of sense with that. I find it easier to think that to describe.

I will try to give an analogy:

Suppose understanding modern art somehow became critical for survival in our society, to such an extent that we tried to genetically engineer humans (using some futuristic science) to have an inbuilt modern art understanding ability. I don't mean, by this, the ability to learn to understand modern art. I mean some system in which human brains, at the start, have some genetically hardwired system which will turn into a full modern art understanding system later, and this system has switches which will be set for how people will view various artistic movements. I think many people would find it implausible that such a thing could be built. It could only make sense if there is a certain amount of rigidity in how brains develop - if they all go down some common path to a great extent.

I do find the subject important. It clearly relates to that very important subject of consciousness. Is consciousness what most people imagine it to be? What if "you" are not really in the driving seat? What if your conscious experience is simply your brain's own model, or narrative, of some much deeper decision making processes. Maybe one big advantage of consciousness is that it turns billions of neural switching events and chaotic, churning computation into some kind of "narrative" so that mental states and intentions can be communicated to others? I am not saying that this is the case, or that it would be the only purpose for consciousness - I can easily imagine that consciousness is mainly that part of my model of the world which deals with my own behavior - but the obvious relationship between narrative and consciousness, and the fact that both language and consciousness seem to involve attaching symbols to a mess of things in the world, suggests to me that anyone who is interested in consciousness should be interested in language. On that basis, I think this was a very useful interview.

RE: Everett vs. Chomsky debates!?

Finally, I had an opportunity of reviewing the above Everett-Nason interview, that I promised myself here: http://www.physforum.com/index.php?showtopic=27192&view=findpost&p=44094... and post my comment here: http://www.physforum.com/index.php?showtopic=14583&view=findpost&p=44212... "Evolution/creation Of Our Human Intelligence, Consciousness, etc. -- RE: Everett vs. Chomsky: Armchair vs. anthropological linguistics -- A review of Everett-Nason interview!" (PhysForumEU; January 3).

Best wishes, Mong 1/4/10usct11:12a; practical science-philosophy critic; author "Decoding Scientism" and "Consciousness & the Subconscious" (works in progress since July 2007), "Gods, Genes, Conscience" (2006: http://www.iuniverse.com/bookstore/book_detail.asp?isbn=0595379907 ) and "Gods, Genes, Conscience: Global Dialogues Now" (blogging avidly since 2006: http://www2.blogger.com/profile/18303146609950569778 ).

If I recall correctly human stereoscopic 3d vision is limited with distance in natural scenes, if this is so after a certain distance and also outside the visual region where both eyes share the visual field, the image must be equivalent to a 2 dimensional image that requires interpretation from the brain to be understood as representing 3 dimensions.

Now, again, assuming that I recalled correctly, that would seem to imply that an inability to understand photographs[which are in general 2 dimensional images representing a 3d world similar to the images outside stereoscopic range] would mean a.) that during development there is little or no exposure to objects|scenes outside this distance, b.) maybe the scenes|objects to which there is exposure are limited enough that no general ability to interpret 2d images emerges, c.) some neurological defect causing an inability to do this that should also affect the internal representation|understanding of the real world occurs, d.) that I recalled incorrectly the information about stereopsis.

As for the ability to spontaneously generate language, it is my understanding that there have been cases where deaf children who've been allowed to congregate have been able to within reasonably short period generate what seems like true novel sign languages of their own.

Thanks for the great interview, Norm and Dr. Everett both!

Responding to Steve Lehar above: regarding pidgins and creoles, I don't think they are evidence for the innateness of language. The children who spontaneously craft a creole from the pidgin of their parents have already been exposed to sophisticated grammars, because their parents would certainly speak to their children in their mother tongue. So from the interactions of the children as a group, a new grammar indeed emerges - but surely the rules of that grammar bear strong resemblances to those of the "mother" languages. Which grammatical conventions are selected for the new language is most certainly an emergent process, one that is directed probably by the playfulness and inventiveness of children. I suspect that groups of children can pull off this remarkable feat for the same reason that they can learn established languages with much less difficulty than adults (namely, that adult brains are much less plastic than childrens').

In other words, all that is going on (not to trivialize such an amazing and sophisticated phenomenon) is that children are re-working or playing with the "linguistic technology" they have already been exposed to. It would indeed be extremely compelling evidence for UG if you could demonstrate that the only language those children spoke was the pidgin - but that is extremely unlikely.

The same argument goes for the emergence of new sign languages - the population from which the new language emerges has already been exposed to sophisticated grammars from "mother" languages.

The arguments for Universal Grammar based on creoles are based largely on the work of Derek Bickerton. Although Bickerton's work is interesting, it certainly hasn't garnered a lot of support among specialists in Creole languages. In their book on language contact (http://www.amazon.com/Language-Contact-Creolization-Genetic-Linguistics/...) Sally Thomason (Michigan) and Terry Kaufman (Pitt) refute Bickerton's arguments convincingly, showing that Creole similarities are based on shared features with the languages in contact with the creole languages.

Dan Everett

I wish I had a dollar for every time my arguments have been "refuted...convincingly". The reasons those arguments "have not garnered a lot of support among specialists in Creole languages" has much more to do with ideology than with science. Rather than doing science by majority vote, maybe you'd like to explain just one way in which my arguments have been "convincingly refuted". Note for openers that the statement "Creole similarities are based on shared features with the languages in contact" is a non-starter. It's flat-out untrue. (Come to that, Thomason and Kaufman are not "specialists in Creole languages" but historical linguists with a limited knowledge of Creoles.)

Bickerton's reply is fair enough. I do not deny, since I myself have experienced the same kind of ideological criticisms, that it is entirely possible that Bickerton's work has been criticized for ideological reasons at times. Communities of scientists can often be like the drivers of Conestoga wagons across the US in the 19th century - they circle at the first sign of threat.

Sally Thomason and Terry Kaufman have both published widely on a variety of topics in linguistics, Sally in particular has published some major work on creole languages, in addition to their historical work. Kaufman is one of the greatest living field linguists and one of the greatest experts in the world on languages of the Americas, especially Mayan and Epi-Olmec. Their criticisms, since I know them both well, are not based on anything ideological. They just found Bickerton's work on creoles for the most part to be unconvincing, to put it diplomatically.

But the point of my mentioning Bickerton was not to discredit him. He has done a lot of interesting synthesis work and made a few interesting proposals of his own (none widely adopted) about language evolution. I simply made the point to make it clear that although creole languages are often cited as support for a universal grammar, the evidence is far from clear and in fact there are solid counteranalyses of the data that do not agree with the creole-universal grammar link.

Bickerton's work should be read by all people seriously interested in the issues. He is an engaging writer.

- Dan Everett

Thanks, Dan, for those suitably hedged encomia. I guess this is neither the time nor the place for the extended discussion that any serious exploration of these issues would involve. On the other hand, I am more than happy to debate the issues, whether in Creoles or in language evolutyion, with anyone, any time, in any forum, any medium, etc, etc. I don't expect any takers, though. It's one thing to write articles at a safe distance, another to go mano a mano with the monster himself.

Hi Derek -

As a layperson, I'm interested in how you would counter the supposition I made above that the emergent grammars of a creole are sourced ultimately in the parent languages of the children who invent it. I'm assuming you wouldn't suggest that the children who invented the creole had no exposure to other languages, so I'm not sure how you could argue for UG in this scenario, given that they'd already developed proficiency in the sophisticated grammars of their parent tongues.

Thanks,
Terren

Derek,

I am the monster in any debate with you or anyone else. Let's get it on. But in some other forum.

Dan