Kanzi by Savage-Rumbaugh and Nim Chimpsky by Terrace

For comparison with the Gardner and Gardner study I have researched two other studies on language acquisition.

These were of 'Kanzi' by Savage-Rumbaugh, started in 1980 (and still going) and of 'Nim Chimpsky' by Terrace in 1983. Nim Chimpsky was experimented with because Terrace had seen videos of Washoe's 'communication' and was sceptical toward its reliability. Terrace decided to re-do the study himself using the same methods but observing and judging the results more strictly.Terrace attempted to teach Nim American Sign Language in a deaf environment (so as to avoid alienating the ape with verbal language which might inhibit its learning). Terrace used the 'moulding and imitation' method as had been used with Washoe. This entailed teaching Nim to imitate his experimenters by moulding his hands with theirs to match the correct sign.

He was trained from the age of 9 months and Terrace was determined to make his experiment stricter than Gardner and Gardners'. During the 44 month duration of Nim's training he seemed to learn over 125 signs.Most of these signs were proper nouns but he also learned many verbs and adjectives as well as a small number of pronouns and propositions. Savage Rumbaugh criticised Terrace's methods by saying they were too strict and that Nim's progress may have been inhibited by the lack of a personal relationship as she felt she had with Kanzi.

The method and initiation of Kanzi's training by Savage Rumbaugh was quite different. Kanzi's acting mother 'Matata' was a research subject in language acquisition studies. When Kanzi was six months old Matata was taken away to breed with another Bonobo ape.When left alone he seemed miraculously to correctly employ all 10 lexigrams that were on his mother's learning keyboard. Although the experimenters had not intended to teach Kanzi he had been present at all his mother's lessons (though he had not seemed to be paying any attention at the time). Not only did he seem to understand the lexigrams but also the spoken English words they represented.

As he seemed to have skipped such a large stage in initial language acquisition they never employed reward based training.Their assumption was that Kanzi did it for its own sake. Later experiments helped Kanzi enlarge his vocabulary by accompanying words and lexigrams with gestures, pictures, video tapes and behavioural activities that made manifest the intentions which underlay the communications. Savage Rumbaugh believed that she'd accomplished a great deal. No ape had ever before been able to determine that a specific word corresponded to either a picture or printed symbol.

This is called 'symbol-to-symbol' transfer and has never before been reported in a non-human species.Kanzi was apparently able to converse freely about travel, finding food and what he might wish to play. Today he knows over 200 words productively and 500 receptively. He tested 100% accuracy on words in his vocabulary. He could also understand these words in differing accents and in synthesized speech. He also demonstrated unequivocal understanding of thousands of spoken sentences with complex structures including embedded phrases, pronouns, case markers and absent referents.

Although Savage Rumbaugh believed that Kanzi was communicating for his own choice critics argue that he had little or no idea of what the experimenters were saying and was not using the signs symbolically. According to Savage-Rumbaugh, Kanzi's linguistic ability was tested as being the same as a two year old child's. Kanzi also used the same techniques of communication as Barret found in human children in 1989. Both children and Kanzi used real words as well as 'non-real' words. Gardner and Gardner recorded that Washoe babbled (non-words) as children do, using untaught signs to communicate certain desires.

Words were used expressively and directedly. Nelson, in 1973, found that human children within the two year age bracket had a larger receptive vocabulary than expressive vocabulary. This is to say they understood more words than they could express. Kanzi was tested to understand 500 words, but could only use 200.

Gardner and Gardner also believed that Washoe understood more signs than she could produce. This shows a correlation of learning between humans and apes. Savage-Rumbaugh's case study on Kanzi seems to me to be the most convincing and reliable exploration into apes using human language.Gardner and Gardner's study was greatly discredited by Terrace's version of the same experiment.

In the only videotapes of Washoe 'communicating', Terrace thought he recognised Washoe merely replicating signs that had been shown moments before. From Terrace's own study, he concluded that his ape was not really communicating, but going through the motions he's learned through operant conditioning to acquire food. However, Kanzi did not learn through operant conditioning. In fact, Kanzi seemed to learn without Savage-Rumbaugh's knowledge or understanding.Kanzi would have been fed as usual, whether he took an interest in communicating or not.

Also, Kanzi's teaching was more real in the terms of the child learning language. Whereas Washoe and Nim Chimpsky started their training when they were already many months old, Kanzi had been present for lessons practically from birth. As a new-born child has adults talking at them to much the same effect, so had Kanzi. I've little doubt that Terrace was incorrect in his findings that both Nim and Washoe lacked real language skills.When testing creativity (this entailed changing the order of words in sentences to have different meanings) both chimpanzees produced poor results.

In similar tests carried out by Savage-Rumbaugh, Kanzi produced results on a par with a human child (the child was Alea Murphy). By and large, both child and ape correctly understood three-quarters of the sentences presented to them. The difference may have been the style of teaching (ASL vs. lexigrams and speech) or it may have been the age at which training began, or it may have been that Kanzi was not a chimpanzee by a Bonobo ape.It could be simply that Bonobo apes have a more innate capability toward language.

Savage-Rumbaugh went on to study more Bonobos, which she believed all produced better results than Nim and Washoe. It seems to me that if a two year old child does something, it is called child language whereas if an ape is tested to definitely do the same thing, it is not called language. I believe we call what human infants communicate as initial language because we know that ultimately they shall grow to be fully grown human beings, using full blown human language.We also know that apes shall never grow up to be adult users of language and so there is an unfair bias in the interpretation of the data concerning language acquisition in relation to apes. Although there has been extensive research carried out on how children acquire language, 'language acquisition' is a part of psychology.

Psychology, by definition, is not a science. No one can be certain as to how children acquire language and so no one can be certain as to how or if apes can acquire it. It seems to me that tests on the ability to recognise and apply language are the only reliable indicators for whether it is present.