Tag Archives: Columbia University

Project Nim – The Untold Story

Cinema’s around the world have been showing “Project Nim“, a documentary about a chimpanzee raised like a human child in the 1970s. However, the documentary misses some key points about the research behind Project Nim, and what was learnt along the way. Herbert S. Terrace was instrumental in this research, directing the project at Columbia University. He has kindly allowed us to reproduce a piece that sets the record straight on the rationale and conclusions of his research.


Project Nim – The Untold Story

During the mid-70’s, I directed a project at Columbia University that tried to teach Nim, an infant chimpanzee, to use American Sign Language (ASL). Project Nim, a documentary currently running nationally, loosely describes the project.

It shows scenes of Nim interacting with his teachers after he was flown to New York from his birthplace at the Primate Institute in Norman, Oklahoma, when he was two weeks old.  Viewers would have to close their eyes not to appreciate Nim’s loveable personality and endearing antics, but they would be hard pressed to see the science on which the project was based.

That’s because the documentary pointedly avoids that topic, which is a shame because the research provided groundbreaking scientific insights into what chimpanzees can tell us about the evolution of language.

The project’s goal was to determine whether a non-human primate could learn the essence of human language: the use of grammatical rules to create particular meanings.  Positive evidence would undermine the then popular claim of continuum between chimp communication and human language. Negative evidence would undermine this claim and require us to ask why chimpanzees are unable to learn a grammatical language.

By the time Nim was almost five, I had collected enough data on his multi-sign combinations to determine the extent of his grammatical knowledge. In September 1977, I ended the project’s research and charted a plane to return Nim to his birthplace.  At five, he would also benefit from being with other chimpanzees because of his size.  If he had continued to live with his teachers, it was likely that one of them would be seriously injured because of his innate aggressiveness, which could manifest itself at the slightest provocation despite his generally benign disposition.

When I returned to New York, I began to analyze the signs that Nim had ostensibly learned during a period of 27 months in which his teachers recorded more than 20,000 multi-sign sequences. While analyzing those sequences I obtained quantitative evidence that Nim could indeed create grammatical sentences and that I had the most powerful evidence to date that a chimpanzee could construct particular meanings by using a grammatical rule.  This would have been exciting news by any standard.

But while preparing my findings for publication, the research took a decidedly different turn. After reviewing a video I’d seen at least a dozen times previously, I noticed that Nim’s positive results were caused by a simple artifact. Nim’s teachers signed what he signed but a quarter of a second earlier. This meant that Nim’s signs were mainly imitative and not spontaneous.  How had other scientists and I missed these prompts while watching the video previously? I realized that when I observed Nim sign producing signs, either in person or on videotape, my attention was always riveted on his hands because I thought they were making history and I didn’t want to miss a second of it. I saw the same symbiotic relationship while viewing other tapes of Nim and movies of other apes (e.g., Washoe and Koko) that had purportedly learned ASL.

This analysis made me wonder why Nim signed at all.  The answer was immediately obvious — to obtain rewards that he couldn’t obtain otherwise. Equally important, I noticed that Nim never signed to start a conversation. He only signed with the expectation of obtaining a reward — e.g., food, candy, drinks and nothing more. Like a child, Nim learned, what I refer to as Language 1. What Nim never learned was Language 2, the ability to converse with someone to convey information that was not about her basic needs, e.g., I just saw Mary or I’m going to the library tomorrow. Language 1 consists of uni-directional imperative statements; Language 2, of bi-directional declarative statements between a speaker and listener. By revealing the true nature of Nim’s signing, the research confirmed the efficacy of unbiased scientific inquiry.

I published my “negative” results, at that time contrarian, in the journal Science and in a book entitled Nim. My conclusions about Nim’s signing have withstood the test of time. But in the film Project Nim, my findings are very briefly described as a failure without explaining why and what failure meant. Apparently the director didn’t understand the difference between negative results and a failure. More important, my results were disparagingly cited as the reason I had returned Nim to Oklahoma, even though I hadn’t discovered the true nature of Nim’s signing until a year after he had been returned to his birthplace.

For me, the omission of the scientific validity of the actual project diminishes the film’s credibility. Also, the director clearly missed an opportunity to educate the public about the science involved in the actual research with Nim. The project proved once again the importance of negative results that they can inspire questions that lead to significant positive results, as was the case with the discovery that space wasn’t filled with ether and that inanimate matter did not give rise to animate matter.

A few years after I had returned Nim to the Primate Institute, it went bankrupt and he was sold for medical research. But thanks to Nim’s signing ability, I was able, with the pro bono assistance of a lawyer, Henry Herrmann, and Bob Ingersoll, a graduate student, to rescue him and place him in an animal sanctuary run by philanthropist Cleveland Amory.  Nim lived there with a mate named Sally and three playmates, Midge, Kitty and LuLu, until he died from a heart attack at age 27.

In the end, Nim’s inability to learn a language deepened our understanding of the basic difference between the minds of humans and apes. Most important, apes lack a “theory of mind”, the ability to perceive what another ape is thinking. Without that ability, it is impossible for their signing to rise above the level of begging to conversation, the essence of human language.

As charming as Nim was, he was not human and to anthropomorphize him as such is not only bad science but also dangerous sociology, e.g., chimpanzees are fully capable of maiming and killing humans.  Nim himself hurt several people as he matured.  Nonetheless, he was clearly special. He was a remarkable creature from the living tree of evolution, as are his threatened relatives, and he should be greatly respected for sharing himself and his abilities in the pursuit of what it means to be human.

Herbert S. Terrace
Professor of Psychology & Psychiatry
Columbia University


Hopping rabbits herald breakthrough in tissue engineering

A team of NIH-funded scientists and veterinarians at Columbia University, the University of Missouri, Clemson University, and the Medical University of South Carolina, have this week announced a significant advance in tissue engineering, for the first time they have used cutting–edge tissue engineering technology to produced a moving joint, in this case the hip, in rabbits.  A press release on the NIH website discusses the work in some detail, and those with a subscription can read the original research article in the Lancet.  This is not the first paper to describe the production of bone or cartilage using tissue engineering, but it is the first time that the two tissues have been regenerated together to produce a moveable joint, and represents a significant step forward in terms of the complexity of tissue that can now be engineered.

Rabbits are a popular experimental model for the study of bone repair and regeneration; the structure of their bones is very similar to that seen in larger animals including humans, for example unlike some smaller rodents they have structures known as  Haversian canals that affect bone growth and repair, while their size allows more complex surgery than is possible with smaller rodents.

Tissue engineering techniques we have discussed previously, such as the artificial lung, involved seeding a scaffold, were created by stripping cells from donor tissue, seeding with stem cells, and then allowing the cells to grow in vitro to produce a functioning organ. The technique reported this week differs in that the scaffold was made from an artificial bio-polymer, and rather than implant stem cells into the scaffold and growing the tissue in vitro, they coated the scaffolds with the growth factor known as TGFβ3 and then implanted it into the rabbits. TGFβ3 attracts bone and cartilage precursor cells to the scaffold, where they multiply and after a few weeks have formed a functioning joint.  When they compared scaffolds coated with TGFβ3 to bare scaffolds, they observed that more precursor cells were recruited to the scaffold when TGFβ3 was present, and that the rabbits transplanted with TGFβ3-coated scaffolds moved more easily when assessed one to two months after surgery, indeed the joints were able to support the weight of the rabbits without any limping.

Rabbits play an important role in medical research. Image courtesy of Understanding Animal Research.

This technique is significantly simpler than those approaches that require stem cell seeding and in vitro growth prior to transplant, and might be especially useful for younger hip transplant patients, individuals aged 65 or younger. Younger patients would be expected to recover more quickly, have fewer co-morbidities that would be aggravated by staying in bed for a prolonged time to allow the tissue to regenerate, and would benefit more from not having to have hip operations every 10-15 years as is currently the case with metal hip joints.  For more elderly patients metal hip joints are likely to remain the best option.

So does this technique replace that used in the tissue engineering studies we have previously discussed? Well, the answer is no, for some applications either approach might work, but for others, for example the artery and lung transplants, the tissue needs to be capable of functioning immediately following transplant. One aspect that is being evaluated elsewhere is the use of biopolymer scaffolds, which are being used with stem cells to produce replacement blood vessels, and may provide a more flexible and reliable alternative to the use of decellularized tissue.

It’s an interesting development, and one that again highlights how quickly things are happening in the field of tissue engineering. Of course it will be some time before clinical trials in humans start, before then this technique must be evaluated in a larger animal, probably a pig, to determine whether tissue regeneration on the scaffold is rapid and effective enough in a model of comparable sizes to humans. Only if these tests are successful will this technique warrant evaluation in a human clinical trial.

Paul Browne