Join 3,555 readers in helping fund MetaFilter (Hide)


Talk with the animals?
August 21, 2014 10:22 AM   Subscribe

A surprisingly dramatic world of lawsuits, mass resignations, and dysfunctional relationships between humans and apes. From Koko and Kanzi to Chantek and Nim Chimpsky, research into human-ape communications used to be all the rage. Nowadays, not so much. What happened?
posted by gottabefunky (33 comments total) 24 users marked this as a favorite

 
If you ever wanted to understand all that is wrong with primate-based research into language and cognition, Project Nim is what you want to watch. Also a great primer of the sexual politics of the liberal Boomer generation.
posted by Nevin at 10:48 AM on August 21 [3 favorites]


What happened?

Too much monkey business.
posted by yoink at 10:58 AM on August 21 [2 favorites]


Project Nim is an excellent movie, though it ended up enraging me. It will tell you more about humans than chimps.

It was made by the same director who did Man on Wire. If you liked that movie, you'll probably like this one.
posted by hurdy gurdy girl at 11:05 AM on August 21


What Project Nim didn't cover though was the huge problems regarding researchers becoming too connected to their ape subjects, and as a result losing objectivity when it comes to interpreting the actions and abilities of the apes in question. I feel like this is clear in the anecdote the piece you link to opens with regarding Koko's reaction to the death of Robin Williams where Koko's carers are projecting human response on to a Gorilla who met the actor once 13 years ago. I'm a researcher in language evolution and as a result have read most of the literature surrounding the studies where an ape has been brought up by the researches writing the papers. There is a stark difference between the objectivity displayed and how the results are presented compared with studies which study ape behaviour in the wild, they mostly rely on anecdotes rather than statistically robust findings. I think the drop off in this kind of study is precisely for that reason, it makes for bad objectivity and therefore bad science, which is something the media won't pick up on because they love a story where the ape can think, feel and talk like a human.
posted by hanachronism at 11:05 AM on August 21 [11 favorites]


Terrible ethics, shitty (politics-driven) science. That is all.
posted by spaceman_spiff at 11:15 AM on August 21


We figured out what they were saying and we didn't like it.
posted by Renoroc at 11:20 AM on August 21 [1 favorite]


Project Nim was a particularly egregiously bad study, though, it should be mentioned, and not really representative of the field.
posted by Navelgazer at 11:27 AM on August 21


Project Nim is one of the single most depressing documentaries I've ever seen in my entire life.
posted by desjardins at 11:37 AM on August 21


I feel like this is clear in the anecdote the piece you link to opens with regarding Koko's reaction to the death of Robin Williams where Koko's carers are projecting human response on to a Gorilla who met the actor once 13 years ago.

Wow--that really is some astonishing bullshit.
posted by yoink at 11:37 AM on August 21 [1 favorite]


What Project Nim didn't cover though was the huge problems regarding researchers becoming too connected to their ape subjects, and as a result losing objectivity when it comes to interpreting the actions and abilities of the apes in question

Is there another documentary about Nim Chimpsky? Because the one I watched definitely identified major problems in research assumptions, research methodology, and research ethics. There is no way you could come away from watching that movie and still think that Herbert S. Terrace's "results" provided any meaningful addition to our understanding of primate cognition and language acquisition. Terrace does come across as an ape, though.
posted by Nevin at 11:39 AM on August 21 [4 favorites]


Not to mention all the "hey Koko *really* wants to see your nipples I bet she's bored with mine so c'mon pop 'em out on this porch we'll all totally turn around" (paraphrasing) stuff.

O_o (for the workplace culture, not at Koko)
posted by Earthtopus at 11:41 AM on August 21 [3 favorites]


You're getting mixed up with Conan O'Brien, surely?
posted by Nevin at 11:42 AM on August 21 [1 favorite]


Is there another documentary about Nim Chimpsky? Because the one I watched definitely identified major problems in research assumptions, research methodology, and research ethics. There is no way you could come away from watching that movie and still think that Herbert S. Terrace's "results" provided any meaningful addition to our understanding of primate cognition and language acquisition. Terrace does come across as an ape, though.

No, I guess I watched the same one, I just remember being disappointed by the lack of objective linguistic content in the film, commentary by other linguists or content covering other ape studies. There is certainly a awful lot of interesting stuff that was absent, but I guess I wanted the whole thing to be about the linguistic aspect, because that's where my interest lies.
posted by hanachronism at 11:52 AM on August 21


Gorilla Won't Stop Saying 'Gorilla' In Sign Language
posted by usonian at 12:25 PM on August 21 [3 favorites]


There's a lot of relevant stuff linked in this post from 2011.
posted by hanachronism at 12:27 PM on August 21 [1 favorite]


With the way things are going in most places on the planet, the only place where certain animals will survive will be in captivity. And that's a problem because we don't know how to do that without abusing them physically and/or mentally.

It is perhaps telling that we haven't figured captivity out for our fellow humans, either, so why should we be surprised we haven't for our primate cousins.
posted by tommasz at 1:34 PM on August 21 [1 favorite]


The article features this quote:

These apes are able to communicate with humans, and this alone is a testament to primate cognition.

That's a misleading characterization of these studies. The sign language studies are basically just teaching apes to perform tasks in exchange for treats, the same way that you can teach a dog to shake. That doesn't mean that the dog is communicating some desire to shake your hand when you say the command word, it's that the dog associates doing that with getting a treat and therefore does it.

If you look at the actual raw data from these studies (which the researchers are usually hesitant to provide), it's usually the animal pushing buttons or using signs that are associated with stimulus that they like (such as food or toys) over and over again because they have been trained to know that good things happen when they those buttons or make those movements. Similarly, when researches show the ape something that the ape has seen before like doll and the ape signs/pushes the word doll, it's because originally the ape just pushed random buttons/made random signs until it hit the one for doll and was given a reward. Sometimes during this random button pushing/signing where the ape is fishing for a reward from the researches there ends up being some seemingly intelligent new series of signs that actually fit the situation just by chance, which researchers cherry pick and use as evidence that the ape is using language.

It's all operant conditioning and all of the results they claim come from the researchers either directly training the apes to do things that seem intelligent or projecting intelligence onto randomness. The apes in these studies are "communicating" in exactly the same way that a mouse trained to navigate a maze is "communicating" that it wants food by navigating the maze. There are plenty of ways to study the intelligence of various types of animals, but these researchers aren't really studying anything.
posted by burnmp3s at 1:40 PM on August 21 [7 favorites]


burnmp3s, what's interesting about that is whether language acquisition in humans is itself just consistently applied operant conditioning also. I mean, a kid finally says "banana" and they get their favorite food--is there a Cartesian stage where a banana appears first? Are they interacting with the concept of "banana"? Sure, maybe later, but only after persistent reenforcement that sometimes accidentally leads to complex thoughts which can seem profound to adults, but ultimately are just strings of known sounds to the child.

If gorillas can connect certain sounds to things consistently, they are right there with us, playing language games…well, proto language games. That's really interesting but also merely foundational to performing complex speech acts, which happens spontaneously at first but is constantly encouraged by cultural standards of interaction.
posted by zinful at 3:40 PM on August 21 [1 favorite]


I knew Duane Rumbaugh, Sue Savage-Rumbaugh, and Herb Terrace personally, the Rumbaughs when Duane was chairman of my department at Georgia State and he and Sue were teaching Lana Yerkish at the Yerkes Primate Research Center in Atlanta, Terrace as graduate advisor at Harvard (dept. of experimental psych, which was B. F. Skinner's department for those who dig operant conditioning.)

The short but correct answer to the "What happened?" question is that, emotional attachment aside, the research arrived at a clear conclusion, and it was not an interesting one. So good new researchers quit wanting to work on the topic.

The conclusion: no matter what you do or how you teach, chimps (or any other great ape) don't learn to talk like people. They pick up individual "words" very well (however "word" may be defined in a given training regime) and they make a (very) small amount of progress at stringing words together. But not into anything like a sentence, because the order of the "words" varies pretty much at random. If you had visited Lana at Yerkes, and she had seen through the glass partition that you were holding a coke, she might ask for some by going to her panel of lexigram buttons* and punching out, perhaps, Give Lana Coke. But, equally likely, Coke Lana Give, Lana Coke Give, or Lana Give Coke. That last one did not mean Lana was going to give you some coke. The notion that word order in a sentence can change the meaning of the component words is beyond where chimps can go. They just don't do syntax, which means they just don't do language as humans experience it.

*which would now, inevitably, be called icons

While Nim did learn 125 signs, Terrace concluded that he had not acquired anything the researchers were prepared to designate worthy of the name "language" (as defined by Noam Chomsky) although he had learned to repeat his trainers' signs in appropriate contexts. Language is defined as a "doubly articulated" system, in which signs are formed for objects and states and then combined syntactically, in ways that determine how their meanings will be understood. For example, "man bites dog" and "dog bites man" use the same set of words but because of their ordering will be understood by speakers of English as denoting very different meanings. - Wikipedia, Nim Chimpsky


> The apes in these studies are "communicating" in exactly the same way that a mouse
> trained to navigate a maze is "communicating" that it wants food by navigating the
> maze.

There are qualitative differences among various species in what they are able to learn. Not even Fred Skinner could teach deictic anchoring ("self-recognition") to a flatworm.

But there is also great continuity in cognitive ability across species. When Gallup first reported his mirror test results for self-recognition I thought no species other than people and great apes would ever pass. But that hasn't turned out to be so. Capuchin monkeys (small New World primates) have passed, elephants have passed, some cetaceans have, and at least one bird.

Given a time machine (and the great-grandmother of all research grants) I would love to be able to test all the primates on the line of descent to humans, starting with Aegyptopithecus and going through H. neanderthalensis just to see where in the sequence this particular cognitive ability starts to appear.
posted by jfuller at 3:44 PM on August 21 [19 favorites]


That's a clear and helpful summary jfuller. If you (or other linguist familiar with the controversy) is still here though, can I ask a question? I've never quite understood the Terrencean jump from "apes cannot be taught human syntax" to "therefore it's all operant conditioning and Clever Hans and screw it". If Nim or Washoe or who have you can learn that a symbol represents a concept, use that symbol in new contexts and combinations, and (allegedly) pass it on to other chimps - why is that not scientifically interesting?

I live about five miles from the old Institute for Primate Research (they raise sheep there now). I've talked to people who remember the weird old days when chimps were boarded with half the OU psych department and would randomly turn up at grocery stores and Sunday morning church services. I've chatted with a sweet old lady who had uncomplimentary things to say about Prof. Lemmon (the monstrous shadowy guy with the cattle prod, for Project Nim watchers) and his iron-fisted, chimp-related cult of personality. I've heard Bob Ingersoll (Nim's genial hippie grad-student buddy) talk a couple of times, and bought an incredibly ugly lime-green tshirt from him that I can never wear in public. And the one thing that's weird about the whole signing-chimp saga* is how it ended abruptly with this stark divide between "We communicated with the chimps" and "No, you didn't, experimental error, pfft". I guess I just don't get why the concept of syntactical word ordering seems to be the utter and complete deal-breaker that the entire field of study lives or dies by.


*Besides literally everything else
posted by ormondsacker at 6:10 PM on August 21 [2 favorites]


On the other hand, science is ramping up on communicating with dophins.

Come on in, the water's fine, I Like Seaweed.
posted by IndigoJones at 6:47 PM on August 21


I maintain that since we can never assume a shared experiential context with another species that we will never know if we are in fact communicating with any individual of that species at any level of abstraction.

Likewise, we will be unlikely to be able to communicate complex abstractions with any alien species we may ever encounter in the future, science fiction notwithstanding.

I've been debunking the Talking Chimpanzee Fantasy to students for 20 years. It never gets easier. They so want to believe it is possible, and are so conditioned by anthropomorphizing ideology that they really do resist, often, being told that honeybee dances are closer to human language than anything we have ever witnessed in a non-human primate or cetacean.

Mind you my view is agnostic on whether animals other than humans are capable of abstract thought or complex intra-species communication with novelty, prevarication, etc. I just maintain human beings can only ever infer non-behaviorist explanations for animal behavior, and we are very prone to inference by wishful projection.
posted by spitbull at 8:43 PM on August 21 [3 favorites]


It's only...it's only...
posted by carping demon at 9:34 PM on August 21


> can I ask a question? I've never quite understood the Terrencean jump from "apes cannot be
> taught human syntax" to "therefore it's all operant conditioning and Clever Hans and screw it"

ormondsacker, will you check back here tomorrow on this one? It's after midnight and I'm falling over.


> I maintain that since we can never assume a shared experiential context with another species
> that we will never know if we are in fact communicating with any individual of that species at any
> level of abstraction.

My only problem with this point of view is that I can't clearly see a reason for drawing the line here (between individuals of different species) rather than over there (between individuals of the same species.) In one direction lies the "Of course Fluffy talks to me. All the time. We understand each other perfectly!" style of touchy-feely anthropomorpism, which can't be right; but in the other direction lies solipsism, which can't be right either.
posted by jfuller at 9:44 PM on August 21 [2 favorites]


This is anecdotal and second-hand, but back in my college days (god, that was a while ago) my then-girlfriend was studying anthropology and had done a lot of work with great apes. This would have been around 2002 or a bit earlier, and by then the Nim Project was already understood to be a travesty.

She would talk about how the apes she was around would be given paints and canvass, and make "art" at about the same level as a young toddler, but would also name these paintings things like "pretty bird," and she swore that sometimes you could see what the gorillas were going for in their paintings.

Uh-huh, I think, and I'm still skeptical, because titling something is such an abstract concept and there are just so many layers of possible and probably miscommunication there. Oh, the ape knows that paintings are supposed to be representative, or even can be? And knows what it means to title the smearings it just made on a canvass? And isn't just remarking on something else, and the carers aren't just assuming that's a title? ANd you aren't then just recontextualizing the smearings in light of the title? Okay.

But there was another one of her stories that was much more convincing to me. She had held a conversation in sign language with a gorilla. That by itself is pretty cool, but there was another level to it. The gorilla in question had not been taught sign language by humans. It was an orphan who had been "adopted" by a childless "mother" gorilla who had been taught by humans, and who taught sign language to her adopted son herself. Mother and son reportedly communicated with one another with sign language, as well as with the researchers/carers.

This is largely why I don't have a lot of respect for Noam Chomsky, whose ideas seem to largely come to him immaculately rather than through any real empiricism and show profound post-hoc confirmation bias, but then, I suffer the same sort of confirmation bias about my opinion of him when reading him and/or about him, so take that with as many grains of salt as you need. In any case, that story made me believe that apes are capable of interspecies communication (to the degree that the novel I'm currently working on repeatedly touches on this topic, though not without hearing out the skeptics and not coming to a definite conclusion on the matter.)

Another more recent ex also worked with apes, in Kentucky. Her most notable story was with one unpopular female ape who kept a collection of photos of a human movie star and would furiously masturbate often while staring at them. That probably doesn't mean much but now you've got that image in your head so I feel like I've accomplished something, at least.
posted by Navelgazer at 10:03 PM on August 21 [3 favorites]


Jfuller, as an anthropologist and not a psychologist, my simple answer is because members of our own species can confirm that their experience conforms to an abstract idea by using language.

It's a catch 22 that is elegantly solved by considering language an innate, species specific adaptation of Homo sapiens, and proved by the fact that we dominate all other species on this planet, live and thrive in ecological niches to which we are not adapted, and have been altering other species' genomes since the dawn of agriculture in deliberate ways.

A good chance to say how much I enjoy your contributions to these sorts of threads.
posted by spitbull at 3:21 AM on August 22 [1 favorite]


If you (or other linguist familiar with the controversy) is still here though, can I ask a question? I've never quite understood the Terrencean jump from "apes cannot be taught human syntax" to "therefore it's all operant conditioning and Clever Hans and screw it". If Nim or Washoe or who have you can learn that a symbol represents a concept, use that symbol in new contexts and combinations, and (allegedly) pass it on to other chimps - why is that not scientifically interesting?

Actually this is an issue of great contention within the field of evolutionary linguistics as to what it is that actually makes human language unique and special. The whole syntax is special hypothesis is probably the one that gets bandied around most often because it's Chompsky's preferred hypothesis. See Hauser, Chomsky & Fitch (2002) for an article that really pushes this idea. (Interestingly, Hauser has since been disgraced for academic misconduct, and is now pushing the view point (with Chomsky) that empirical research is actually useless in the field of language evolution, which is interesting since he's banned from doing it unsupervised, and Chomsky has always been uninterested in empiricism as Navelgazer said.) But there are other researchers who think that the thing that makes us special is our ability to make inferences and recognise communicative intent (people like Thom Scott Philips and Michael Tomasello), or that the special thing about humans is our abilities to connect arbitrary symbols to meanings (most recently Denis Bouchard in this book (I've just written a review of this book which will appear soon if you're interested)). But in answer to your question, it is scientifically interesting but depending on who you're talking to and what their favourite hypothesis is.
posted by hanachronism at 3:33 AM on August 22 [3 favorites]


By the bye, the "species specificity" of Language from my anthropological point of view lies in its adaptation to and for human social organization, achieved primarily through non-referential capacities, as well as in Language's general capacity for communicating meta-referential abstraction, for which syntax is the engine but not the sole explanation. In my Boasian and Peircian traditions we'd say it is the "dual patterning" that bridges concept and phonation more broadly, making phonology just as important as syntax, and non-referential (poetic, phatic, co native, expressive, and meta-linguistic, etc. in Jakobson's scheme) communication inseparable from reference as such, and mind and body conjoined in the faculty of voicing, not merely cognizing, intentional and emergent meaningful communication that can conjure possible (past, future, untrue, fantastic, etc.) as well as actual, co-present worlds. For this reason "languages" are not assimilable to "Language" as such, and the diversity of human linguistic behavior is as interesting and relevant to its evolutionary natural history as its universality or innateness or location in the mind/brain schema. None of that is less important either, but it is a natural empirical fact that other species do not communicate with anything like the capacity for subjecting evolutionary destiny to intentional manipulation (for good or ill) that humans have done, albeit for a remarkably short period of time no matter which estimate you use or whether you're a discrete-ist or a gradualist about when and how the faculties in question emerged.


Therefore it is a strange approach to science to study something artificial as if it were a priori natural for humans and our fellow apes to communicate via sign language, as stupid in the end as expecting gorillas to learn to talk before the mechanism of phonation was understood to preclude that.

We are still witnessing the ebbing of the over-confident Chomskian revolution, from which an entire bullshit enterprise of pop science ev psych has been spun off as a cocksure extension that human behavior generally would yield to an account whose triumph was rooted in the same simplistic mind/body and individual/group binaries as the crude Bloomfieldian and Skinnerian claptrap it replaced.

Progress lies in the study of how people (and other species) actually do communicate, including with individuals of other species by the way, in the natural and actual social world. Sapir was right all along.
posted by spitbull at 7:18 AM on August 22 [2 favorites]


Another way of putting it: almost all of the ape language stuff tells us way more about people than gorillas or bonobos or chimps. And what it says about naked apes is that we still like to paint (other) animals on the cave wall at night.
posted by spitbull at 7:35 AM on August 22 [2 favorites]


I've never quite understood the Terrencean jump from "apes cannot be taught human syntax" to "therefore it's all operant conditioning and Clever Hans and screw it". If Nim or Washoe or who have you can learn that a symbol represents a concept, use that symbol in new contexts and combinations, and (allegedly) pass it on to other chimps - why is that not scientifically interesting?

The problem is that unless the apes show some real evidence of higher level connection between the words and ideas, there is no way to differentiate "learn that a symbol represents a concept" from "learn to take an action when given a cue to receive a reward" and "use that symbol in new contexts and combinations" from "randomly perform actions until rewarded".

The main problem is that these symbols and signs have nothing to do with language when the apes themselves use it. For example, let's say I had a bunch of an ape's favorite treats and I trained the ape that every time they use the sign for "murder" they get a treat. Or maybe I just reward the ape for signing "murder" when a particular other ape is around. If the ape starts signing "murder" all the time whenever the other ape is there, does that mean that the ape has learned that what the word means and is now thinking violent thoughts about the other ape? No, the ape is just repeating actions that were trained with no knowledge other than what cues that action can be used after to get a reward.

Putting the layer of human-like semantic meaning on top of these signs only clouds the issue, and it doesn't help that the researchers in these types of studies believe that the apes are capable of much more complex understanding of language than there is any evidence for. If you want to find out if apes can teach trained behavior to other apes that can be a good study, but don't pretend they are teaching each other anything more complex than "this is how you get treats".
posted by burnmp3s at 9:01 AM on August 22 [2 favorites]


ormondsacker:

> I've never quite understood the Terrencean jump from "apes cannot be taught human syntax" to "therefore it's all
> operant conditioning and Clever Hans and screw it".

For Prof. Terrace himself there is some relevant history. He got his PhD from Harvard's Department of Psychology at a time when there was a deep philosophical and emotional split at the H school between experimental (read "rats and pigeons in training cages") psych and the rest of it. The Department of Psychology and the much larger and more popular Department of Social Relations were both housed in campus eyesore William James Hall. The tippy-top floor right under the projecting roof is an observation balcony surrounding a conference room. Under that, the top two real floors were psych, including one floor for the department and one floor for the rats and pigeons. The entire rest of the building was social relations. Soc rel had some brilliant faculty but it was also notoriously the touchy-feeliest department on campus with an undergraduate program of study that was almost impossible to fail out of. Psych, by contrast, was the department of B. F. Skinner, Mr. Operant Conditioning himself. In those circumstances psych and its graduates became more and more hardcore. "If you can't show your effect in your data record you haven't shown it, period." Terrace's scientific attitudes (with which, honestly, I am very sympathetic) came directly from that environment. Note, this is not to say that the department made people that way. It was more a case of attracting applicants who were already like that and then making them more so.

The effect the chimp-language team hoped to show--clearly, in the data, anectotes and impressions don't count--was humanlike (as Chomsky defined it) or at least compellingly proto-humanlike linguistic performance, and they couldn't show it, so they reported a null result.

Here is a two-part exchange of letters between Terrace and philosopher (and animal rights advocate) Peter Singer in The New York Review of Books (one, two) in which the contrasting attitudes are on full display. (I rate the exchange a draw.)
posted by jfuller at 12:26 PM on August 22 [4 favorites]


Interresting discussion, and thanks to jfuller for the great commentary. It's amazing to think of the work of skinner not just in the context of science or the american climate as a whole, but also of the skinner box he literally worked in :)

A few additional thoughts:

1. We have a personification problem in that we personify humans. Animals are not people; rather, people are animals. I have not researched language much but I truly believe there's a lot less special, and a lot more coincidental, about it than we like to think.

2., and I'm talking mostly about pets here and less about apes, but I think a lot of animals are mentally incapacitated by being locked in a room for their entire lives. I'd love to see a study that compared the intelligence of a street rat to a lab rat, or of an indoor cat to an outdoor cat. No wonder they don't talk if they've got nothing to talk about.
posted by rebent at 11:25 AM on August 25


> We have a personification problem in that we personify humans.

Not to mention personifying not merely humans and not merely animals but even things like ships. She canna' take any more, captain. She's gonna BLOWWW." One of the (many) unprovable things I believe is that the tendency to see persons as persons is evolutionarily wired in, and therefore inevitably generalizes to things that are not persons (live pets, fuzzy stuffed toys, hugable trees, clouds, nation-states, hypothetical supernatural beings) and is thus not to be "cured" by any sort of argument. In fact it seems as if we don't have enough of this tendency, considering how obviously easy it is to view others who are human as less than human or not really people at all.

A philosopher named Strawson even argued in what is IMHO a very profound book that human beings should be considered as logical primitives and not analyzed further. tl;dr version: go with the flow about this, we're wired up to think that way anyhow, nothing good will come of trying (and failing) to resist it.
posted by jfuller at 1:36 PM on August 25


« Older http://www.codefights.com/...  |  Middle East Peace Potential th... Newer »


You are not currently logged in. Log in or create a new account to post comments.