P is for Primate language

24 07 2011

I’ve just seen this somewhat dispiriting documentary about Nim, one of a number of primates who have been sequestered, domesticated, scrutinised, feted, and ultimately abandoned in the name of linguistic research.  Even the shots of the Columbia University forecourt that I walk through every day failed to enliven a story of wanton cruelty, institutional pettiness, dodgy science and bad hair.

The film charts a succession of sudden, traumatic abductions, starting when baby Nim was snatched screaming from his mother’s arms. Over a period of several years, with only humans to interact with, the young chimp was taught to sign, using an adapted form of  American Sign Language, and acquiring a working vocabulary of several hundred words. When he outgrew his cute and cuddly stage, and/or when the funding ran out, he was packed off to a sort of primate Guantánamo Bay. The story is only slightly redeemed by the efforts of one of his former minders to track him down. Even in his hoary old age, Nim still retains a trace of his former competence, pathetically signing ‘play’ from within the bars of his prison.

Columbia University

Frustratingly, the film hardly touches on the linguistic controversies that fuelled this research. In the 1970s, when this unhappy story took place, the debate as to whether only humans are innately equipped with a modular language acquisition device (LAD) was still a fairly hot issue. Not for nothing was Nim named Nim Chimpsky, in (cute) recognition of Noam Chomsky’s role as the leading protagonist of the debate.

What was at stake was this: if highly intelligent apes, exposed to a similar linguistic environment as human children, could acquire an extensive lexicon, but fail to develop even the rudiments of a ‘grammar’, this would go some way towards supporting the view that humans are uniquely hard-wired for language acquisition. On the other hand, if evidence of syntax, however primitive, could be demonstrated, Chomsky’s notion of a ‘Universal Grammar’ (UG) would either need to be extended to nonhuman primates, or it would need to be re-evaluated entirely.

And the findings? Nim’s vocabulary was impressive, but more impressive still was his ability to form two-sign, three-sign, and even longer strings: MORE EAT, HUG NIM, BANANA EAT ME NIM, etc. Moreover, a superficial analysis of the data would suggest that Nim was operating according to some kind of embryonic grammar, producing word order patterns not dissimilar to those of human children’s first utterances. For example, he consistently placed the sign for MORE in front of the sign it modified:  MORE TICKLE, MORE DRINK, etc. But, as Jean Aitchison (1983) notes, “a closer analysis showed that the appearance of order was an illusion. Nim simply had a statistical preference for placing certain words in certain places, while other words showed no such preference” (p.55).

However, as Roger Brown(1973) argued, with regard to similar results for Washoe, an earlier case study of primate signing, “While appropriate order can be used as evidence for the intention to express semantic relations, the lack of such order does not establish the absence of such intentions” (p. 41). This is because the use of appropriate word order, of the verb-object type, for example, as in GIVE BALL, is not strictly necessary, since the context in which the utterances are generated usually resolves any ambiguity. That is to say, the pragmatics of the situation renders syntax redundant. But if that is the case, why do (human) children show evidence of a proto-syntax right from the start?

In the end, we don’t seem to be much the wiser as to whether the higher primates have a rudimentary LAD, despite all the anguish that was inflicted in trying to find one. Nor, for that matter, do we really know whether humans have an LAD either, or whether their faculty for language acquisition isn’t just a spin-off of their vastly more developed cognitive capacities.

What we do know is that the chimpanzees who have been studied do not use their linguistic capacities in the same way as humans, even very young ones, do. Nim, for example, rarely initiated a conversation, and was unable to grasp the basics of turn-taking. As Aitchison (1983, p. 57) concludes, “Nim did not use his signs in the structured, creative, social way that is characteristic of human children”.

In fact, Nim’s ‘language’ was simply a more elaborated version of the way chimpanzees use gestures and vocalizations in the wild: to regulate two-way social interactions such as grooming, feeding, and play. As Tomasello (2003, p. 11) puts it, nonhuman primate communication functions “almost exclusively for imperative motives, to request a behavior of others, not to share attention or information with others in a disinterested manner”.

As someone once said, “Your dog can tell you he is hungry, but not that his father was poor but happy”.


Aitchison, J. (1983). The Articulate Mammal: An Introduction to Psycholinguistics (2nd edn). New York: Universe Books.

Brown, R. (1973). A First Language: The Early Stages. Cambridge, MA.: Harvard University Press.

Tomasello, M. (2003) Constructing a Language: A Usage-based Theory of Language Acquisition. Cambridge, MA.: Harvard University Press.



27 responses

24 07 2011

Hello, Scott.

Thanks for the always interesting blog post.

Chomsky’s positing of a universal grammar and an innate LAD still seems to provoke huge contention even today. Daniel Everett, an “anthropological linguist”, conducted research into the Pirahã language, which is spoken by a hunter-gatherer tribe in the Amazon, and came to to conclusion that Chomsky had it wrong all along.

Everett claims that the Pirahã language has no grammatical recursion, an absence of relative clause, and no lexis with which to describe specific numbers or particular colours. These observations, among others, led him to directly dispute the notion that human beings are necessarily hard-wired to follow certain linguistic rules and that language is an inborn mechanism.

The implications of his research apparently also contributed to Everett’s eventual rejection of Christianity.

Not surprisingly, Everett’s research is itself controversial and contested. Chomsky, with whom Everett used to work and get along swimmingly, allegedly later branded him a charlatan. (Obviously, even major linguists aren’t above a bit of back-stabbing now and then.)

You can hear Everett himself give a succinct overview of his findings here on the excellent Philosophy Bites website:


I’m interested to know on which side of the debate you yourself lie, Scott?

24 07 2011
Scott Thornbury

Thanks for being the first to comment, Wes.

Yes, Chomsky has always been fairly dismissive of any evidence that might challenge his theory of the innate and exclusively human nature of universal grammar. After all, it would take only one proven example of grammatical recursion or syntactic embedding in animal language, or one example of a human language that doesn’t have these features, to throw the whole elaborate architecture of transformational generative grammar into doubt.

Thus, he was equally dismissive of projects like the Nim project. In a frequently cited article in the New York Times, on the subject of animal language, he is quoted as saying that “attempting to teach linguistic skills to animals is irrational — like trying to teach people to flap their arms and fly”.

He goes on:

“Humans can fly about 30 feet — that’s what they do in the Olympics,” he said in an interview. “Is that flying? The question is totally meaningless. In fact the analogy to flying is misleading because when humans fly 30 feet, the organs they’re using are kind of homologous to the ones that chickens and eagles use.” Arms and wings, in other words, arise from the same branch of the evolutionary tree. “Whatever the chimps are doing is not even homologous as far as we know,” he said. There is no evidence that the chimpanzee utterances emerge from anything like the “language organ””

(Johnson, G. 1995. Chimp Talk Debate: Is It Really Language? The New York Times. June 6, 1995, Section C; Page 1; Column 1; Science Desk).

As for my own ‘position’ on UG, etc, the clue is perhaps in my having referenced Michael Tomasello’s brilliant book, Constructing a Language. But I won’t elaborate until I am nearer my voice recognition software, so I can quote a hefty chunk of the Introduction for you!

24 07 2011

Dear Scott,
This post takes me back to the first module of my MA studies, Language Theory, and the heated debates the “innateness hypothesis” always generated.

The research conducted on the Bonobo chimpanzee, Kanzi, appeared to be the most convincing on film but then on further analysis it became clear that a lot of passion and enthusiasm can greatly shape the way results are presented and then interpreted by the researchers and audience alike.

Your opening sentence communicates the way I also felt after watching the disturbing footage of Genie and then learning about what happened in that story once the funding ran out.

Looking forward to further comments,


24 07 2011

Some references for Kanzi:
http://www.youtube.com/watch?v=LxmbjLoUnhk + Savage-Rumbaugh, S. and Levin, R. 1994. Kanzi: The Ape at the Brink of the Human Mind. UK: Transworld Publishers Ltd.

For Genie:

24 07 2011
Wolfgang Butzkamm

Excellent summary of a long standing and controversial issue. Thank you.

24 07 2011
J.J. Sunset

An enlightening summer read: Kenneally, C (2007). The First Word: The Search for the Origins of Language. New York: Viking.

“Studies in subjects as diverse as ape gestures, parrot cognition, numerical and grammatical competence in monkeys, and self-awareness in dolphins are changing our understanding of how language came about -not as a sudden, dramatic shift in our genes but as a gradual process involving somatic, neurological, and cultural innovations.”
Mere survival at the basis of language evolution rather than a sudden Big Grammar-Gene Bang, so to speak…

24 07 2011
Scott Thornbury

I’ve just seen this Guardian summary of Nim’s story, as – presumably – the film is about to launch in the UK:


25 07 2011
Alastair Grant

Hi Scott et al,

As always, a great topic. I remember reading about Nim a few years ago and it also struck me that the wretched creature died once funding dried up like some has-been rock star. Nothing new there I guess, although this case and Genie’s highlight the possibility we could just re-lable the LAD as a panzootic capacity (necessity) for communication, for food, help etc without thinking that either unfortunate being “needed” to learn English.

Genie’s is a fascinating story. It seems that her capacity or otherwise for acquiring language was less to do with any kind of “critical age” and more to do with the abuse she suffered as a child, resulting in a condition called psychogenic dwarfism, whereby any kind of growth, mental and physical, is stunted. This, the research team seemed to totally overlook:


What both cases have in common is the somwhat imperialist stance that, instead of drawing out methods of effective communication, a method was being forced in! How ridiculous and unhelpful. Indeed, Genie’s mother ended up suing the original research team – http://en.wikipedia.org/wiki/Genie_%28feral_child%29

How much do we, as teachers, impress the same templates on our learners?

25 07 2011
Scott Thornbury

Thanks Alistair – you raise an interesting question about the nature of the teaching itself, and its effect on the (language) learning of both Nim and Genie. In the Nim Project film, we see a number of ‘teaching’ sequences which are, predictably, ‘Direct Method’ in style – using the ‘here-and-now’ to contextualise the sign being taught (obviously grammar-translation wasn’t an option!), and requiring a fair bit of effort on the part of the teacher to focus Nim’s attention. Presumably a fair amount of repetition and recycling was necessary too, although how systematic this was we don’t know. We do know that the ‘classroom’ was often the real world – the garden of the Bronx estate where the project was eventually located – but that also ‘lessons’ took place in a tiny, featureless room at Columbia University, a requirement (apparently) that the teachers themselves hated, not least because there were so few ‘affordances’ for learning, and it was consequently very dificult to focus Nim’s attention.

We don’t know whether Nim’s teachers attempted to initiate ‘conversations’ about the world beyond the immediate context, e.g. about people or things that were physiclaly absent, or even about the past. So we don’t know whether the nature of Nim’s language is (partly at least) a function of the nature of his instruction, rather than the limitations of his cognitive capacity (a question we might also ask of our own students’ failure to achieve a target-like competence!)

25 07 2011
Thomas Ewens


I’ve just finished reading Pinker’s ‘The Language Instinct’ which is highly critical, almost dismissive, of projects which try to teach sign language to animals. His argument is certainly that Nim’s failure to learn the language (indeed the failure of any animal to learn ASL) was due to lack of cognitive capacity.

There has been a lot of debate on this blog (and elsewhere) about humanistic language teaching and, seperately, about the use of technology.

Maybe we should view the people who think they can teach language to animals with the same suspicion that we have for those who think that we can teach language using technology (at the expense of human values).

This (humanistic language teaching and technology) is actually the theme of a conference presentation I’m developing at the moment (very exciting for a newbie like me too). I had never thought about animalistic language teaching before.

26 07 2011
James Quartley

if you’ve read Pinker’s book you may want to balance his conclusions by reading Tomasello’s (see Scott’s post above) or Geoffrey Sampson’s (http://www.grsampson.net/AMiu.html) ‘Language Instinct Debate’.

26 07 2011
Scott Thornbury

Thanks, James, for mentioning the Sampson book, which I confess to having read but not remembering a lot about, at least in terms of his arguments. I seem to remember being disappointed that he wasn´t up to speed on the ‘usage-based theories’ of acquisition that Tomasello, for example, is a key spokesman for. To remind you (and here’s a chunk of An A-Z of ELT shamelessly cut-and-pasted):

Usage-based is a way of describing those theories of second language acquisition (SLA) that argue that acquisition occurs primarily through engaging in communication, i.e. through usage. According to this view, the learner’s internal grammar is abstracted from the cumulative data to which they are exposed, and which forms their history of actual language use. The bulk of learning is implicit, and is a direct effect of the frequency of encounters with an item. Only when the learner comes up against a problem, e.g. when communication breaks down, does the learner need to take conscious control of learning. Usage-based theories reject the mentalist view that language acquisition can be explained only by reference to some innate language learning faculty (or universal grammar). Instead, they appeal to general learning processes, which include:

• pattern extraction: automatically and implicitly, learners extract patterns from prior input;
• “tallying”: they also infer statistical frequencies and sequential probabilities of occurrence of phonological, lexical and grammatical features of the L2;
• association learning: associations between co-occurring elements of the language are gradually strengthened as neural networks are weighted to reflect the frequency and distribution of patterns in the input;
• “chunking”: sets of already formed associations that have been stored in memory are welded into larger units;
• memory: patterns and sequences that are extracted from input are rehearsed in short-term memory, and eventually establish themselves in long-term memory, where they serve as a “data-base” for the acquisition of grammar.

All of these processes are optimised where there is maximum exposure and use, suggesting that language teaching that is consistent with communicative principles will serve learners well.

27 07 2011
James Quartley

Communicative principles! Shhhhhh! Don’t let Michael Swan hear you!

Sampson sits at the opposite end of the nature/nurture continuum, which, perhaps, as an advocate of nurture, clouds his view and makes his view seem as ‘extreme’ as Pinker’s nativist one. However, he does tear down many of the nativist assumptions and in so doing clears away a lot of the acceptance of the nativist view as some how incontrovertibly true.

It certainly challenged me to view Chomsky ‘off his pedestal’ (something I wouldn’t have dared to do in my youth) and i then found many of his (Chomsky’s) arguments as relatively groundless opinions or weakly constructed. Sampson’s book was a difficult read, a kin to losing your religious faith, but ultimately liberating, even if it doesn’t give all the answers. It does a good job of illuminating the house of cards that nativism is built upon.

27 07 2011
Thomas Ewens

As I understand it, psycholinguistics is moving on from the idea that there is a definable ‘universal grammar’ true of all languages. I’ve read second hand about a study (reported in Nature) which suggested that grammar evolves over time within language families and that therefore language is part of general human cognition rather than a specialised module within the brain. Can anyone explain a bit more about this?

You know, I have an MEd in TESOL and Applied Linguistics, but I’m often astonished (and a bit depressed) at how little I still know about the subject.

Thanks for the references, I’ll try to hunt them down. It seems I have a lot of reading to do!

27 07 2011
James Quartley

Firstly, you could go for a ‘skeleton version’ of Sampson’s case, written by him and titled ‘There is No Language Instinct’ available at http://www.grsampson.net/Atin.html

Then there is an article by Cowley (2001) ‘The baby, the bathwater and the “language instinct” debate’ in Language Sciences (No. 23 (2001) 69-91). If you have Athens access or equivalent, it is easy to find and download as a pdf.
It eschews both Pinker and Sampson’s arguments and to quote him; “Instead of analyzing language in to form-based units, we can treat it as an aspect of social life deriving from a capacity to contextualize experience”.

27 07 2011
Scott Thornbury

As I understand it, the notion of ‘linguistic universals’ – e.g. that all languages have verbs, objects, vowels, etc – is slightly different to Chomsky’s concept of ‘universal grammar’, which argues that humans are born with an innate language learning capacity configured as a set of principles and parameters. UG accounts for deep structure, while linguistic universals are the surface features that languages have in common. As Geoffrey Sampson, in another book (Schools of Linguistics, 1980) argues, “An argument from linguistic universals to innate mental mechanisms could only possibly work provided that the linguistic universals were not ones for which more obvious explanations are available”.(p. 239).

Thus, the argument that, because all languages have objects, objects are part of our grammatical hardwiring, can be countered with the argument that humans experience the world in terms of agents having effects on things or people, and hence we talk about the things or people affected as the objects of verbs. That is to say, the category of object is only hard-wired in as much as our experience of the world determines our cognitive structures. (Whether these cognitive structures are inherited or acquired is, presumably, moot).

Whether nonhuman primates perceive the world in this way, and whether therefore they see a correlation between verb-like signs and object-like signs (as in HUG NIM), and whether this relationship is encoded in a fixed word order, is another moot point.

27 07 2011

Dear Scott and all,
I’d like to add another reading to the expanding list being put forward here:

Evans, N. and Levinson, S. 2009. “The myth of language universals: Language diversity and its importance for cognitive science”. In Behavioral and Brain Sciences (2009) 32, pp. 429-492.

It is a fascinating read and the open peer commentary that follows it sees Michael Tomasello declare “universal grammar is dead” and Steven Pinker and Ray Jackendorff clarify what UG is for them.


27 07 2011

Apologies for the late reply. I think the comments here cover as much as can be said about the LAD/UG debate, but I’d like to offer this tidbit.

Scott wrote: As someone once said, “Your dog can tell you he is hungry, but not that his father was poor but happy”.

I was reading this poem last night, The Span of Life, by Robert Frost

The old dog barks backwards without getting up.
I can remember when he was a pup.

Frost’s poem is an example of countermotion: the first line, a statement but not a poem by itself, runs in countermotion to the second line, which comments on the first to create the poem. (For fun, you might try reciting the poem to notice how clever it really is in terms of meter, etc.)

The experiments with Nim might not have resolved the debate you’ve highlighted (LAD), so the documentary stands as a sort of statement with your observations about the treatment of Nim as commentary. Thanks for the countermotion, Scott.

28 07 2011
Scott Thornbury

Thanks, Rob, for that doggy comment. On the subject of dogs and language, I feel I have to pass on this joke:

A man went into a pub and told the barman that his dog was a talking dog. ‘I bet you five pounds I can get him to answer any question you like to ask him,’ he said.
‘OK,’ said the barman. ‘How are you?’
‘Rough!’ barked the dog.
‘And what do you call the top of a house?’
‘Roof!’ barked the dog again.
‘Right,’ said the barman. ‘Now then, who was the manager of the English football team in the Mexico World Cup?’
‘Rough!’ barked the dog.
‘What a load of rubbish!’ said the barman. ‘Give me my five pounds.’
The man gave the barman the money and started to leave the pub. On the way through the door the dog turned to the man and said,
‘I remember now. It was Alf Ramsey!!’

(Girling, B. 1990. The Great Puffin Joke Directory. Harmondsworth: Puffin Books. p. 163).

28 07 2011
Gareth Knight

We need to forget the mechanistic views of the human being where we’re as good as the words we produce and as clever as the language we create. We have much more at our core. It’s time to understand consciousness. I look forward to receiving the thoughts of Nim.

Guy walks into a bar with a duck on his head. The bartender looks up and says “Where did you get that ape?” Guy says, “This isn’t an ape, it’s a duck”. Bartender says “I was TALKING to the duck”.

28 07 2011
Stephanie Ashford

Why is it always a GUY who walks into a pub? Why always a barMAN? Even Robert Frost’s dog is a HE.

Sorry, I know this is quite off-topic. But it does make me wonder…

28 07 2011

Stephanie, I assume your question is rhetorical, since the answer is obvious, ie, We live in a patriarchy. Patriarchy affects our language genres, eg, jokes and stories. Of course it doesn’t have to be that way, and awareness-raising is the first step; however, in my experience, just as many women as men use ‘guy’ in these a person-walks-into-a-bar jokes.

At least it’s bar-tender here in PC America. 🙂


28 07 2011

“So, a woman went into a bar and told the barwoman that her cat was a talking cat”

Or more PC..

“A person went into a drinking establishment and told a person behind the bar that his or her unisex feline domesticated animal was able to speak”

Jokes lose something when adapted I think. Those classic introductions just put you in the mood and you know they’ll lead to a punchline and yes, perhaps they will be sexist or possibly not PC but if we get rid of all those subjects will there be anything to laugh about?

28 07 2011

Hello everyone!
Wes, James and Scott have explored the relationship between Chomsky’s UG and primate language. I would like to put my two cents in with some ideas from Michael Gazzaniga’s Human: The Science Behind what makes us unique (2009) about Chomsky and his view of language. It seems Chomsky had proposed that human language is a discontinuous feature (discontinuity theory proponents argue that “some elements of behavior and mental traits are qualitatively unique to a given species and share no evolutionary heritage with other living species or archaic species”). Nevertheless, as Savage-Rumbaugh states “the significance of Kanzi’s [Cf. Nathan’s post] possession of certain elements of language is, however, enormous. As the ape brain is just one-third the size of the human brain, we should accept the detection of no more than a few elements of language as evidence of continuity” (p. 164).
I think this is interesting for language scholars and teachers because it allows us to draw comparisons between animal language and human language -Otherwise, we would be talking about two essentially different phenomena and comparisons would be pointless. Instead, analyzing differences (as Scott does at the end of the post) is very enriching. We can also learn about the origin of language by analyzing animal language (as Gazzaniga does in his book)

30 07 2011
Scott Thornbury

I must admit that, if there is anything to be surmised from the research into animal language, it is that the discontinuity thesis might be seriously flawed.

For me, the key is in the Aitchison comment, quoted in my original post, to the effect that “Nim simply had a statistical preference for placing certain words in certain places”. Usage-based acquisiiton theory would argue that all language acquisition is simply a matter having a preference “for placing certain words in certain places”, i.e. of extrapolating frequently occuring patterns from the data (rather than accessing pre-determined grammatical parameters), and that what Nim was doing is what human children (and second language learners do), i.e. registering certain probablistic tendencies in the input feed, and constructing a mental grammar that reflects these tendencies. The fact that Nim hadn’t generalised these patterns across all two-sign combinations doesn’t necessarily mean that the basic strategy wasn’t operating – only that it was operating weakly, or that it needed yet more data. If so, this would argue for a (non-species specific) continuity – rather than a discontinuity – in terms of a ‘language instinct’.

11 08 2011

Hi, thanks to all for a fascinating discussion. I’m quite ignorant of these things myself, so forgive the potential banality of my questions of which there are three:

1) In the Tomasello quote above it’s stated that the usage theory applies to second-language aquisition, but the implication seems to be that the same process occurs in first language acquisition. Is this the case?

2) If it is, how does it get around the seemingly fundamental difference between the two processet due to the “fossilization” effect?

3) The same quote claims that usage-theory makes an “appeal to general learning processes,” How does this relate to the “emergentist” theories such as the “competition model” which (according to wikipedia) posits that “language is learned through the competition of basic cognitive mechanisms in the presence of a rich linguistic environment.” I’d also appreciate it if those in the know could suggest reading matter regarding this…

Those are the questions. These are the idle speculations:

It’s claimed that this theory serves as “an alternative to strict nativist and empiricist theories”, (presumably Tomasello et al) but it seems to me that usage theory is quite compatible with the idea of emergence (as defined by complexity theory), i.e. that the complexity of human language is a result of the combined cognitive operations of the human mind, operations which may exist in other animals to a lesser degree, but due to the immensly greater capacity of the human brain to perform these operations, we take a leap into another level of complexity. As Hegel might have put it – quantity changes into quality.

As far as I can tell the nativist proposition is really the same thing, but fossilised into some sort of “language organ” – ie the qualiative leap took the form of the evolutionary development of some sort of neurobiological “device”, as opposed to an ongoing cognitive process.

For example, a striking parellel occured to me between what Tomasello describes as the process where “associations between co-occurring elements of the language are gradually strengthened as neural networks are weighted to reflect the frequency and distribution of patterns in the input” and the concept of “attractors” in chaos theory: “a set towards which a dynamical system evolves over time”.

If, like me, you tried to disguise your ignorance of linguistics by using philosophical terms, you could say that the “controvesy” is the same old battle between “being” and “becoming”.

11 08 2011

There also seems to be some sort of fundamental contradiction between the innativist “poverty of the stimulus” and the emergentist “rich linguistic environment”, tho I may be completely wrong. Anyone for tennis on this one?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: