Visit ELTBOOKS - all Western ELT Books with 20% discount (Japan only)

May 25, 2011

Green alien shows language hardwired into our brains

Teaching-English-in-Japan-Education-Eigo-Noto" Research by Johns Hopkins University has produced evidence confirming Noam Chomsky's fifty-year old hypothesis that language has been hardwired by evolution into the human brain at birth.

In the study, a small, green, cartoonish “alien informant” named Glermi taught participants an artificial nanolanguage named Verblog designed to violate these natural word order rules of language. Other groups were taught a variation of Verblog that used word order combinations commonly found in human languages.

The adult learners quite easily learned the variants but failed to learn pure Verblog. Researchers claim that this was because learners’ brains “knew” in some sense that the Verblog word order was extremely unlikely, just as predicted by Chomsky a half-century ago.

See also:
John Hopkins press release
Noam Chomsky's universal grammar

Share this:  

« Previous | Main | Next »


Sounded a bit fishy. Read into this (the linked article) and uncovered that they only tested adults with this illogical grammar and the results said that the adults had difficulty in learning this difficult grammar. Is this really a revelation?  If we want to study the possible innateness of grammar we need to test children who are pretty much still unaware of the concept of language learning. If the testers created a village where the society used this "illogical" language to communicate, I'm pretty sure of two things: (1) the language will evolve to match the daily needs of the village and (2) children born in that village will learn that evolving language natively.

Language is kneaded and chipped away by the daily needs of the society that it being used in. Adults having trouble with an illogically designed artificial language? How about a logically designed language then, such as Esperanto? Without a living culture to knead and chip away at the language on a daily basis, the language will not take flight; it will never really be a usable language.  It will be doomed to failure, as in the case with the logically designed Esperanto. Similarly, since there is no major culture in Japan that practices the spontaneous usage, kneading, and chipping away at English, Japan is still not an English speaking country.

The teaching of that "illogical"  artificial language to those adults without providing an authentic culture to foster further development of the language (ironically for us) parallels English teaching in Japan, doesn't it? I believe this article teaches us the importance of meshing culture in with language learning. It reinforces the need for teachers to mesh culture with their language teaching - something that we probably all need to do more of while we are in our EFL contexts. As for the results discussed in the linked article, I'd say that because (1) they only tested adults and (2) did not allow the language to evolve into something culturally meaningful for the learners, the results are not very significant in the real world.

Couple of questions:

1) Doesn't the fact that adults found the variants easy but Verblog difficult (because of its contravention of universal grammar) show anything? It would seem to be in line with the theory, at least. Had the results not shown a difference between the ease of learning the variants and Verblog, surely that would have been evidence of the invalidity of the UG theory? Having a test that could disprove the UG hypothesis seems like good practice, and therefore these results have some value.

2) Your village example: if it were possible to create such a situation, presumably the UG theory would dictate that language would evolve away from Verblog to a variant consistent with UG. Would you agree?

Hi Martin,

Thanks for the great questions! Lets look at each of them.

1) Doesn't the fact that adults found the variants easy but Verblog difficult (because of its contravention of universal grammar) show anything?

It does. As you point out, it at least does not disprove UG -so that counts for something. But not very much. Why use adults for this? From what we know about the difference between children's vs adult learning, adults often try to consciously analyze the new language and end up using their prefrontal cortex much more than children do. Children just suck up language at an alarming rate and are not really aware of the learning. This is obviously ideal and therefore this is the zone that these researchers should be working with. Anyway, it makes total sense, therefore, that adults would find this illogical language (designed to be different from what they are used to) difficult to learn. The real question that needs to be addressed is --will babies be able to learn Verblog usage and grow into it "like natives" would?

2)Your village example: if it were possible to create such a situation, presumably the UG theory would dictate that language would evolve away from Verblog to a variant consistent with UG. Would you agree?

Yes, I agree. But we must realize that it is unfalsifiable (circular argument). Let me explain what I mean by this. This circular nature is an inherent problem with UG theory logic. Human language evolves on a daily basis (inclusive of grammar usage). This is a fact that is not contested. However, UG theorists may look at all those changes, across all our languages, across all those centuries, and say that all of those combined and compounding changes conform to UG principles. The opposing view (gaining speed these days) is that the massive differences between all of the languages across all of those centuries of language usage is much too complex to explain away with UG theory.

If I understand the experiment correctly, they realized that most languages put modifiers either in front of noun/verbs or after noun/verbs, but not both (and other similar "commonality" observations). So what they did was to develop a language that sometimes put modifiers before, and sometimes after nouns/verbs (and other similar designs). I assume Verblog was not entirely chaotic; they must have had a system worked out as to when the modifier went in front, etc. Most languages may indeed not be that confusing, but if a culture had a socio-cultural evolutionary reason to put modifiers in a varying set of locations, I'm sure the language would grow to accommodate the complex needs of that culture and more importantly, to people of that culture, it would not appear confusing because for them, there was a reasonable reason to using language in that way. We are not birds, but we have learned to fly -likewise, if we had a socio-cultural reason to create a unique grammar, we could and would do so. So why do I call this circular? Because inevitably, any modifications to Verblog would be called "a variant consistent with UG" or at least be seen as a language that was evolving conformity to UG (whatever that means). As long as UG mandates that ALL human language fits into UG, it remains unfalsifiable within that context, so within that contextI must say 'yes' to your question.

However, this discourse parallels the nature vs nurture question. This is also a bad question. In fact, it is an invalid question. The truth is - it is nature AND nurture. So, what we are born with AND our immediate context determines what we are at any given moment. Naturally, language is a byproduct of "nature AND nurture". So, while I must agree that we use human languages because we are in fact born as humans, we must all also agree that it is the context that affects our language usage.

Are you not contradicting yourself here? On the one hand you accept that the experiment would be able to disprove UG (if Verblog could be learned as easily as its variants), but on the other hand you suggest that UG is not a valid theory because it unfalsifiable.

Whilst human language might "evolve" on a daily basis the hardwired language systems in the brain evolved over a much longer period. So effectively all languages will be developed from deep UG rules. I'm not familar with those rules in detail, but presumably if they have been explicated then (1) any language found that contravened them would falsify the theory, and (2) being able to learn artificial languages that contravene the theory as easily as those that don't would also falsify it.

Whereas studies that showed the opposite, as in this one, would support the theory.

I think Robert said that the experiment while not really proving UG does NOT disprove UG either. Your last paragraph here is guilty of begging the question. You say "Whilst human language might "evolve" on a daily basis the hardwired language systems in the brain evolved over a much longer period. So effectively all languages will be developed from deep UG rules." You are assuming that UG is valid and therefore all language develop upon those rules. However, logically we cannot do that. Rather than assuming UG holds true, one should see if there are any other explanations as to why the adults could not easily learn the language. Once all other possibilities are negated (such as motivation, language distance, innate language learning ability and so on) can we assume that UG explains the difficulty. However, we don't know if this has occurred.

Additionally there are two fundamental problems with the claims of this study as it is written in the article. First it tested adults and then make claims as to how children learn, which Robert has pointed out are two different things. Secondly, the adults found it more difficult to learn the non-logical language, but they did learn some. The fact that any learning took place at all may possibly discredit UG rather than support it.

In my mind the big problem is that they tested one group (adults) and then made claims for another (children). This article, and possibly the authors of the study, are guilty of making unsupported assumptions based upon faulty comparisons. In order to to be correct they would first have to prove that adults and children learn language the same way. Then they would have to show that both adults and children have the same level of difficulty learning the language with the illogical grammar. But then they are still left with the thorny problem that "some" learning did take place. The question then is what caused it to be more difficult? In looking into that question UG could be one possible explanation. Unfortunately the study is far from doing this.


You're right, Robert didn't say that it disproved UG -- and I didn't say that he said that. I said that he accepted that the trial had the ability to disprove UG but that in the same message he called UG unfalsifiable. He can't have it both ways.

On your second point, I didn't take UG as valid, I just stated the element of the hypothesis that would allow it to be falsified and then suggested ways for it to be falsified.

This study had clear potential to falsify UG theory but didn't. As I'm sure you're aware, in science the idea is to try and falsify a hypothesis, not prove it.

As for the issue of the trial design, I'm not decided but don't have time to look at it in detail at the moment (hopefully they've published all the trial design info somwhere.)

Hi Martin (and Terry),

Thanks for the reply. Let's not confuse my opinions about UG with my opinions about the validity of the article's study -they are separate. I apologize if my wording seemed to mesh the two together. Let's focus on UG for the moment. As for me, it is a big stretch to believe in UG -today. There was a time when I seriously considered it though. I think that if you are serious about its implications as a researcher, you can't really 'partially' believe in UG, you are either in the camp or outside of it (especially from the UG perspective), so in that sense I am 'out' today. As I stated above, I don't think that it is possible to disprove UG because, whether by design or not, it is fundamentally unfalsifiable ('universal' is 'all-inclusive', so any human language is automatically incorporated into the system by definition).

I also agree with what Terry mentioned about the article.

Martin, you discuss "hardwired language systems in the brain" - I think we may disagree here. Due to studies in neuroplasticity, current researchers in the field don't use these terms much any more. Our brain is truly amazing at finding new ways to think about things (create new neuro-networks). All those old stories about the right brain being hardwired for such and such, and left-brained people being so and so, are for the most part, outdated. Read Immordino-Yang. Fascinating stuff! -she's done work with kids that have had hemispherectomies performed on them (half of their brain removed). At first guess, I think most people would think that these kids would simply not survive the operation (removing 50% of the brain!), and even if they did survive, they would be severely retarded. Who wouldn't have assumed this? As it turns out though -to everyone's amazement- these kids (now much older) are fluent, do well in academics and are artistic; functionality from both sides of their brains are being demonstrated with only their remaining halves intact (be it their left or their right side remaining). The secret? They had good tutoring! Utterly amazing (especially from our older frames of mind). So, now we know better about hardwiring. Our minds are fantastically plastic instead.

On the other hand, we also know that children born with fully functional and healthy brains who are neglected ( such as in very poor and neglectful foster homes) very often show terrible signs of mental retardation -they often don't learn social skills and their language capabilities suffer tremendously. In fact, even their body sizes can be much smaller -as though their bodies have simply given up on growing. Photos of such children are jaw-dropping-horrific. Humans are massively social creatures. WIthout society and culture, we literally do not grow to our potential, physically and mentally.

Let's put some major concepts together to come to a conclusion. (1) You can do remarkably well with literally only half a brain if you have good tutoring (a high support context). (2) On the other hand, if you are born with a full, healthy brain, your growth will likely be severely hampered if you are neglected during the rearing years. You will probably not learn to use language as well as others at your age and it is unlikely that you will excel in academics. And (3), everyone is born with differing physical/mental potential -the nurturing that follows significantly affects the outcomes of the individual, while at the same time affecting the very nature of the individual, too. We now know that immediate context actually turns DNA gene expression on and off like a faucet. So, it seems fair to say that the equation is literally "nature AND nurture" and not "nature vs nurture".

With all this in mind, I have found it very difficult to agree with a hardwired UG these days, so this is where I stand on UG. As for the problems with the article, I think what I posted earlier, along with the addition of Terry's great additions, and my current clarification, pretty much closes the deal. Thank you for participating, Martin and Terry!

Yes you're right. Yes it may have the potential to falsify the trial design but like you implied, until we can examine it we really don't know what variables were accounted for.

An interesting study though.


(1) By "hardwired" researchers here mean an inate abilty for grammar that exists as a product of evolution. That innate ability would express itself, obviously, in every human language, i.e. universally.

This is the view taken by Stephen Pinker of Harvard (author of "The Language Instinct" and "The Blank Slate")and the term "hardwired" is used in the press release from John Hopkins, supposedly a world-class research university that is publicizing results by modern researchers. So I think we can say it is used by current researchers.

Are you saying we don't have an inate language ability that is a product of evolution? A "hardwired" language ability?

(2) In order to falsify the UG theory, you would have to find just one language that didn't follow the rules of universal grammar. And demonstrating that non-UG artificial languages are just as easy to learn as ones that do would seriously undermine the theory as well. So UG seems falsifiable to me.

Unless of course UG proponents keep moving the goal posts about what the what the valid expressions of the deep structure of UG are...

Hello again Martin,

(1) I'm not sure what the purpose of this pursuit is Martin. Your questions have departed from the original topic. Moreover, I find it funny that you choose Pinker in defense. Pinker (who is a fabulous writer and a great public speaker), has views that are hotly debated. He's been famously boxing relentlessly with Gladwell for quite a while in public. But perhaps more importantly, even Chomsky and Pinker don't always see eye to eye. So I find Pinker to be an odd choice for you.

You ask these questions, "Are you saying we don't have an innate language ability that is a product of evolution? A 'hardwired' language ability?" as if it were a unique view to have. I took this from Pinker's own wiki to save typing:

(quote begins) Pinker's books, The Language Instinct, How the Mind Works, Words and Rules, The Blank Slate, and The Stuff of Thought combine cognitive science with behavioral genetics and evolutionary psychology. The Language Instinct has been criticized by Geoffrey Sampson in his book, The 'Language Instinct' Debate.[19] The assumptions underlying the nativist view have also been subject to sustained criticism in Jeffrey Elman's Rethinking Innateness: A Connectionist Perspective on Development (Neural Networks and Connectionist Modeling), which defends the connectionist approach that Pinker has criticized. (quote ends)

It should be sufficient to say that questions on these lines are within long standing debates. There are tons of words that have been professionally/academically exchanged on this topic. No reason to mimic that here.

(2) The fact that UG proponents have not given up on UG, sort of answers your second question doesn't it?

UG proponents have not found a single language that did not conform to UG, or else they would have given up on UG by now, right? Sarcastically I ask, "How can that be!? hmmm…"

You could just replace the 'Universal' in Universal Grammar with 'Human' and call it a Human Grammar, and still make all the same conformity arguments. From the outside it seems ridiculous doesn't it? Why wouldn't all human languages end up using a Human Grammar? (Of course all human languages conform to a Human Grammar. How could you falsify this?)

Martin, you know where I stand on UG, and you know where I stand on the article's shortcomings. I think this conversation has run its course -especially because we are not actually talking about the article anymore.

Well, Robert, first you contradicted yourself about whether UG was falsifiable or not (which is pretty relevant issue given people are doing trials that they believe can falsify it) and then you went on to confidently but erroneously assert that the term "hardwired" isn't used by current researchers. I thought I'd point out that this wasn't the case. To repeat: the press release about the research in question (done by very current researchers) uses the term "hardwired". Pinker, who is fairly prominent in this area, as you agree, uses it. Your attempt there to suggest that the theory of an innate language abillity hardwired into the brain by evolution has somehow been discredited to the point that current researchers won't even use term just won't wash.

In the face of me pointing out that UG can be falsified by trials with artificial languages just like the one in the article, you respond "We are not actually talking about the article anymore." and attempt to close the discussion down through suggesting we are off topic.

Let's see... the article is about a trial's results supporting UG and I was talking about... issues raised by the article relating to UG and the potential falsifiabilty of UG by this trial through the use of artifical languages... I think that's pretty much on topic, if you'll allow me to be so bold -- so you seem to have made another erroneous assertion. Although now as a result of having to correct these errors of yours I am indeed straying off topic... :-(

One last minor factual correction: the quote isn't from "Pinker's own wiki" it is from the Wikipedia article on Pinker. That's rather different I'm sure you'll agree -- and I'm sure you'll also agree on the importance of being precise in these matters.

Your arguments about "Human Grammar" etc. show a complete misunderstanding of UG, in my opinion, but I understand that you want to close the conversation down now, so I'll be taking my leave of this one.

Thanks for the discussion.

Recent Headlines


Hedbanz / Charades
Helene J Uchida
Think Read Write
David Lisgo


Useful Links


  • ALTEThe Association of Language Testers in Europe
  • British CouncilInformation, resources, and links to other sites
  • Cambridge English Language AssessmentInformation on examination and qualifications for teachers and students
  • IALICInternational Association for Languages and Intercultural Communication
  • IALLTInternational Association for Language Learning Technology
  • IATEFL"Linking, developing and supporting" ELT professionals worldwide
  • TESOL IncPublishing, connecting, events and career development for teachers




World Today