Search Holah

Core Studies


 

 

Kanzi has his own up-to-date home page here. You can watch videos and find out more about his vocabulary which now includes more than 500 words.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Medication interaction list comprar kamagra sin receta tellings.

Home > Cognitive > Savage-Rumbaugh et al. (animal language) Study

Savage-Rumbaugh et al. (animal language)

Savage-Rumbaugh, S., MacDonald, K., Sevcik, R. A., Hopkins, W. D. and Rubert, E. (1986) Spontaneous symbol acquisition and communication use by pygmy chimpanzees (Pan paniscus)

 

Background

This article describes the initial results of an attempt to investigate the language acquisition abilities of two pygmy chimpanzees (Kanzi and Mulika) and contrasts these results with the abilities of two common chimpanzees (Sherman and Austin).

The researchers were investigating whether pygmy chimpanzees (Pan paniscus) could spontaneously use symbols to communicate with people. The researchers found that unlike common chimpanzees (Pan troglodytes) the pygmy chimpanzees could acquire the use of symbols by observing others use these symbols in everyday communications without any explicit training.

Furthermore the researchers demonstrated that pygmy chimpanzees were beginning to comprehend spoken English words and could identify lexigrams upon hearing the spoken words. In contrast common chimpanzees were unable to do this.

The authors suggest that pygmy chimpanzees exhibit symbolic and auditory perceptual skills that are distinctly different from those of common chimpanzees.

Previous research such as a project carried out by Gardner and Gardner have demonstrated the ability of chimpanzees to acquire language. For example the Gardeners demonstrated that Washoe, a common chimpanzee, could produce symbols, such as sign language, to bring about a change, such as asking for a tickle. This type of language learning is called associated symbol learning.

Associative symbol learning occurs when an individual learns to associate specific symbols with specific objects, for example naming pictures in a book.

It has also been demonstrated that when chimpanzees are systematically taught how to request, label and comprehend objects that a more sophisticated type of learning can appear.

For example, Savage-Rumbaugh was able to demonstrate that two common chimpanzees called Sherman and Austin were able to use representational (or referential) symbol learning.

Representational (or referential) symbol learning occurs when an individual can refer to the objects when there are no contextual cues to provide association, for example being able to respond to the request ?get your ball? even if the ball is in another room.

Savage-Rumbaugh argues that it is this shift from associative to representational learning symbol usage that is the critical issue for the proper understanding of language use in apes.

Lock (1980) has demonstrated that human children pass through various phases in their use of symbol usage which demonstrates the transition from associational to representational usage.

Past research with apes has demonstrated evidence of associative symbol learning but representational usage has only been demonstrated after extensive training with Sherman and Austin, two common chimpanzees (Pan troglodytes). For example, it took a year and a half of practice before Sherman and Austin could travel unaccompanied to another room and to return with a requested item.

Although apes lack the vocal apparatus to produce human-like speech there is still a question about whether they can understand human speech. Previous attempts at teaching language comprehension have involved training chimpanzees a few phrases that have been drilled in contextually specific situations. However, true language comprehension involves more than just responding to words. For example a dog can respond to the command ?sit?.

One of the obvious barriers in training apes to use human language is their difficulty in producing speech. However they may still be able to comprehend it. True comprehension is more than just responding to words, for example a dog can respond to the command ?sit?. Children acquire true comprehension without any training, unlike dogs. Another characteristic of human language is the ability, when directed, to select a specific object from a group, that is, provide a differential response on cue.

 

Aim

The aim of this article was to describe the initial results of the first longitudinal attempt to investigate the language acquisition capacity of a pygmy chimpanzee and to contrast it with common chimpanzees.

 

 

Method/Procedure

This article describes the data gathered from a longitudinal study of the spontaneous language acquisition of two pygmy chimpanzees (nowadays often referred to as Bonobo chimps). The study is on-going but this article reports just a 17 month period of language acquisition across a ten year time span.

Although the researchers state that this project was not an experiment, comparisons were made between the pygmy chimpanzees and two common chimpanzees that were taught language with a similar visual graphic symbol system.

Therefore the researchers ended up having a quasi experimental design, whereby the naturally occurring independent variable was the species of chimpanzee (pygmy and common chimpanzee) and the dependent variable was their language acquisition.

Subjects

The principal subject was a male pygmy chimpanzee called Kanzi who was aged 30-47 months during the time of this report. This particular species of great ape is rare both in captivity and in the wild (Zaire). Observational studies of Pan paniscus suggest they are a more social species than other apes and display more highly developed social skills such as food sharing. Other research has also found indications that they might be a brighter species. This would suggest they might be better able to acquire language.

Kanzi was born in captivity and was assigned to the Language Research Center at 6 months of age along with his mother (Matata), a wild caught chimp. Matata had served as a subject in previous studies of non-verbal communication

A second pygmy chimpanzee was also studied called Mulika and she was Kanzi?s younger sister. Mulika was aged 11-21 months during the period of this report.

Both Kanzi and Mulika spent several hours a day with their mother and were attached to her, but they appeared to prefer human company.

Two other chimpanzees (common chimpanzees ? Pan troglodytes) are included in this report as comparisons ? Austin and Sherman. They were part of earlier training programmes. They were assigned to the Language Project at 1.5 years and 2.5 years of age respectively, and were removed from their mothers before coming to the Language Research Project.

The small number of subjects is inevitable because of the time required in conducting an in-depth longitudinal study.

Communication system

Kanzi and Mulika used a visual symbol system consisting of geometric symbols (lexigrams) which brighten when touched. These symbols were on an electronic keyboard. A speech synthesiser was added when it became apparent that Kanzi could comprehend words, so the appropriate words were spoken when a symbol is touched.

Attempts were made to use battery powered lexigrams for outdoor use although they did not prove rugged enough to stand up under constant field use. Therefore, for most of the time in the field Kanzi and Mulika used a pointing board, which is a thin laminated panel containing photographs of all of the lexigrams.

Kanzi, Mulika and the people with them could point to the lexigram that they wished to use. The researchers would also use spoken English when pointing to lexigrams. These gestures were spontaneous.

Researchers also used ASL (American Sign Language) gestures to accompany the lexigrams. Approximately 100 ASL gestures were used by the experimenters and none of the experimenters were fluent in ASL.

Early Rearing and Exposure to Lexigrams

Kanzi was first exposed to the use of symbols, gestures and human speech at the age of 6 months as he watched the interactions between his mother (Matata) and her keepers. Kanzi remained with Matata constantly until he was two and a half years old, and during that time attempts were made to teach Matata symbols.

Importantly, no attempts were made to teach Kanzi symbols directly. Although Kanzi was allowed to observe Matata?s training sessions, he was generally kept otherwise occupied as he tended to distract Matata?s attention

Beginning at one and a half years of age Kanzi did show an interest in symbols. For example he would light keys sporadically and then run towards the vending devise suggesting that he had learned that touching symbols caused this devise to dispense food. Although Kanzi did not engage in any behaviours that suggested that he knew that specific symbols were associated with specific foods

Kanzi also spontaneously began to use the chase lexigram to initiate chase games. This behaviour appeared shortly after he spontaneously began to use a gesture (hand clapping) for the same purpose.

Kanzi was separated from his mother Matata at two and a half years so that she could take part in a breeding programme. When she returned four month later Kanzi had developed a preference for human company even though he was allowed to be with his mother as much as he liked.

Kanzi?s sister Mulika was born nine months later and Kanzi enjoyed spending time with her. At four months of age Mulika developed an eye infection and was taken away for treatment. When she returned she also chose to stay with human companions most of the time. Mulika did not have the opportunity to observe Matata during symbol training but she did observe Kanzi using symbols.

Rearing and Exposure to Lexigrams Following Separation from Matata

The rearing environment of Kanzi and Mulika was similar to that of Sherman and Austin following their separation from Matata. For example, they were with people who use the visual symbol system around them throughout the day for a wide variety of communication and a number of the same teachers have worked with all 5 animals (Sherman, Austin, Matata, Kanzi and Mulika).

However there were some major differences between the environments and exposure to lexigrams.

1. Unlike Kanzi and Mulika, Sherman and Austin were trained to use lexigrams. However later on Sherman and Austin did acquire some symbols through observation.

2. Sherman and Austin?s keyboard was not equipped with a speech synthesiser because tests revealed that unlike Kanzi and Mulika they did not understand spoken English words.

3. Sherman and Austin did not use a keyboard outside of the laboratory because their specificity of symbol use tended to decrease when they used a board that did not require a discrete response

No formal training was ever attempted with Kanzi and Mulika and food was never made contingent on symbol acquisition because they could (a) use symbols without specific training, (b) identify symbols regardless of location, (c) did not tend to confuse symbols for similar items (such as apple, orange and banana), and (d) comprehended spoken words.

Instead of formal training, people modelled symbol use during their communications with each other and around Kanzi and Mulika. During all daily activities (playing, eating, resting, travelling in woods, etc.) people commented on their activities both verbally and visually by pointing to an appropriate lexigram on a keyboard. For example, if they were engaged in a tickling bout, the teacher would comment "(teacher's name) tickle Kanzi" both via the keyboard and vocally. Sometimes Kanzi and Mulika observed people using the keyboard, and at other times they ignored the keyboard usage.

Naturalistic outdoor environment

During the warmer months of the year food was placed daily at 17 named locations within the 55 acre forest that surrounds the Language Research Centre. Food was not available to them in the laboratory during this time so they had to travel from place to place during the day with their human companions to obtain it.

During these warmer months most of the day was spent travelling outdoors. One or more foods were taken to each location, for example, bananas and juice were taken to the tree-house and peaches were taken to the lookout.

Early on when Kanzi knew only a few food symbols he was provided with photographs of foods which were placed on the ground at each location. When Kanzi wanted to travel to a different location he could indicate this by selecting a photograph and he was then taken to the location where the food could be found.

Within 4 months Kanzi learned where all the foods were located and could select a photograph when he wanted to go to that location and also guide others to the appropriate location.

Later Kanzi could use the keyboard to announce where he wanted to go and photographs were eventually not used. Mulika started travelling with Kanzi in the woods at 6 months of age with Kanzi often carrying her for short distances.

By her second summer Mulika had also begun to use symbols to initiate travel herself and was old enough to share some of the foods.

Kanzi would normally initiate travel to specific locations himself but if he didn?t indicate a desire to go anywhere travel was initiated by people.

If Kanzi did not choose to eat the food on arrival he was asked if he would like it placed in his backpack. Kanzi was able to make positive gestures such as shoving the food towards the pack (positive gestures were often accompanied by positive vocalization).

Typically after a half day?s travel Kanzi would have a number of different foods in his backpack. A word such as Juice could now refer to either the juice in his pack or to a location where juice can be found. If Kanzi now indicated juice he could be queried as to whether he wanted to go the tree-house or drink the juice in his pack. If he wanted to go to the treehouse he could then gesture in the direction of the treehouse or if he wanted the juice in the backpack he would approach the pack and touch it.

A blind test was run with an experimenter, who had never been in the woods, four months after the foods were first placed in the woods. The experimenter (over two afternoons) travelled from place to place with Kanzi, at Kanzi?s initiative. Kanzi was able to successfully select each location by pointing to a photograph or lexigram from a selection and guide the experimenter to the location he had selected.

Indoor Environment

During the day Kanzi and Mulika helped in a variety of domestic activities such as doing the laundry. They would attempt to engage others in games such as hide, and often spontaneously helped in simple projects such as wiping up spills. At night Kanzi often asked to watch TV. These indoor activities are similar to those that characterised Sherman and Austin?s free time.

Data recording

When the chimpanzees used a lexigram indoors on the computer this was automatically recorded. Outdoors the record was made by hand and entered into the computer at the end of each day. This meant there was a complete record of Kanzi?s utterances from 30-47 months of age and for Mulika from 11-21 months.

Each utterance was classified when it occurred.

Firstly it was classified as correct or incorrect.

Secondly it was recorded as spontaneous, imitated or structured.

Spontaneous utterances are those initiated by Kanzi or Mulika with no prior prompting.

Imitated utterances are those that include any part of a companion?s prior keyboard utterances.

Structured utterances are those that are initiated by a question, request or object. Structured questions were used to determine whether the chimpanzees could give a specific answer.

Vocabulary acquisition criterion

The criterion used for determining whether a word should be listed as a member of Kanzi?s and Mulika?s vocabulary was that the utterance was made in a spontaneous situation. For example, Kanzi might indicate he wanted to go to the tree-house and this would be verified if he then took the researcher to this location, producing a positive concordance score.

When there was a behavioural demonstration of the correspondence between symbol and object this was considered to be sufficient for producing a concordance score. For a symbol to be classified as being part of Kanzi?s and Mulika?s vocabulary it had to occur spontaneously on 9 out of 10 occasions followed by a concordance of 9 out of 10 additional occurrences.

To asses the accuracy of the observations which were made in real time, an analysis was carried out of a 4.5 hours of videotape. The scoring was done by two different observers, with one observer scoring the behaviour real time and the other scoring the tape. The real time observer did not know that their scoring was being tested for reliability.

There was 100% agreement with regard to the 37 lexigrams used and whether they were used correctly in context.

There was only one disagreement about whether an utterance was spontaneous or structured.

In addition the videotape observer recorded an extra nine utterances.

No utterances were scored by the real-time observer that were not seen on the videotape.

Tests of productive and receptive capacities

Kanzi and Mulika were tested informally, in everyday situations. For example, If Kanzi was playing with some keys he may have been asked what they were. Importantly though, such tests were not used as drills, nor were they associated with rewards. Furthermore, these informal tests were only used with lexigrams that Kanzi and Mulika knew

At the end of the period covered by this report, Kanzi and Mulika were also formally tested on all the items in their vocabulary. These tests were done formally to ensure that their performance was not due to contextual cues or inadvertent glances.

These tests were highly controlled. For example the order of presentation and the location of stimuli were carefully controlled in order to preclude knowledge on the part of the experimenter that might inadvertently bias the response. No test item was ever repeated during a given session, and no two trials were ever the same. The test items and alternatives were randomly determined on each trial with the requirements that neither the alternatives nor the test items were repeated on consecutive trials and that each item serve as both an alternative as well as a test item. Either three or four alternatives were present on every trial.

Photograph to lexigram. This test consisted of showing the subject a photograph who was asked to select from a set of three alternatives the proper lexigram for that photograph. During this test the alternatives were not visible to the experimenter.

Spoken English to photograph. This test consisted of the subject listening to the spoken English word and then selecting the appropriate photograph from a set of three alternatives. The English word was usually presented in a sentence.

Spoken English to lexigram. This test consisted of the subject listening to the spoken English word and then selecting the appropriate lexigram from a set of three alternatives. The English word was usually presented in a sentence.

Synthesised speech to lexigram. This test followed the format of the above test but the word was produced by a synthesiser. The word was produced twice by the synthesiser and the purpose of this test was to demonstrate that the subject was not responding to the intonation of natural human speech. Only Kanzi was used as a subject for this test.

Similar formal tests were also carried out on Sherman and Austin. However because Sherman and Austin were unable to select photographs in response to spoken English they were not tested on the photographs and they were not tested on spoken English because this task was thought to be difficult for them. Unlike Kanzi and Mulika, Sherman and Austin had to be rewarded with food to complete the tasks.

 

 

Results

Untutored Gestural Usage: Some Observations

Between 6 and 16 months of age, both Kanzi and Mulika spontaneously began to use gestures to communicate preferred directions of travel and actions they wished to have performed. For example, they used an outstretched arm and hand to point toward areas to which they wished to be carried. Gestures were often accompanied by vocalizations that served to orient the listener?s attention and convey affect. These gestures were more advanced than those of their mother Matata.

Although similar gestures were observed in Sherman and Austin they appeared at a later age and were not accompanied by effective vocalisations. Kanzi and Mulika?s gestures were more explicit than those used by Sherman and Austin. For example, when Mulika wanted a balloon blown up she placed it in a person?s hand and then pointed to the person?s mouth and even pushed the balloon towards their mouth.

First Appearance of Lexigram Usage for Communicative Purposes

Note that there was no direct attempt to teach Kanzi to use the lexigrams. Kanzi (aged 2.5 years) started using the lexigrams immediately after his mother went away for breeding purposes.

Although it was assumed that Kanzi had learned from observing his mother that lighting symbols was a way of obtaining food, it was not assumed that he had learned any referential relation between particular symbols and particular foods. However, Kanzi did demonstrate that he had already learned some symbols such as apple, banana, chase, juice and ball. Kanzi must have learned the meaning of these symbols by watching his mother

Mulika began using symbols at 12 months, much earlier than Kanzi. Although her progress was slower, suggesting again that Kanzi had learned many things that did not become evident until his mother departed.

Mulika?s symbol usage was also different to Kanzi?s and that she used a single symbol for many different things. Fir example, she used particular symbols (such as milk) for a variety of all-purpose things such as asking to be picked up or requests for food or even for milk.

At about 14 months Mulika began using a number of lexigrams appropriately ? her new words over the next few months were milk, surprise, Matata, peanut, hotdog, cake, mushroom, melon, cherry, banana, jelly, go and blueberry. Mulika occasionally reverted to using milk as an all-purpose communication.

Neither Kanzi or Mulika had difficulty identifying a lexigram when it was in a new position or on another keyboard.

It was observed that both chimpanzees started using a new term in an associative contest first before they started to use it in a contest free situation. For example, the symbol ?strawberries? was introduced to Kanzi when one day he was eating mushrooms on mushroom trail. He and a researcher then went to a new place where strawberries could be found. Initially Kanzi?s spontaneous usage of strawberries was restricted to where he first imitated the sign. Eventually his use of the sign strawberries extended beyond the original context and Kanzi could use the sign to indicate what he wanted to eat and where he wanted to travel. It seems likely that children go through the same initial process of associative usage before being able to use symbols independent of context.

Progress During the 17 Months After the Initial Separation from Matata

Kanzi did make rapid progress over the 17 months after the initial separation from Mulika. The table below shows Kanzi?s single word acquisition from age 30 to 46 months based on the concordance measure. The table also shows which symbols dropped out of the vocabulary after having initially met the criteria.

In total, during the period covered by this report, Kanzi acquired 46 words and Mulika 37.

Mulika?s initial rate of acquisition was slower than Kanzi?s. This was probably because Kanzi had actually acquired some words in the period before his mother?s departure but not produced them. After his mother left Kanzi suddenly started producing words which he probably had acquired before so it appeared that he learned more quickly.

As is the case with human children Kanzi?s symbol comprehension typically preceded his symbol production. Overall, the receptive criterion (comprehension) was reached before the productive criterion for 63% of the words in Kanzi's vocabulary.

Spontaneous utterances consistently account for more than 80% of Kanzi's single-word and combinatorial utterances

Prompted, imitated, or partially imitated utterances accounted for only 11% of Kanzi's total corpus

Mulika's rate of spontaneous utterance production is slightly lower and her rate of imitated utterances somewhat higher.

Kanzi and Mulika's symbol usage had been accurate the majority of the time.

Combinations

Kanzi?s multi-symbol utterances (combinations) appeared within the first month of spontaneous keyboard usage. Combinations were far less frequent than single symbol utterances.

Across the first 17 months, Kanzi produced a total of 2,540 nominated combinations and 265 prompted or partially imitated combinations and all but ten were judged to be appropriate and understandable; 764 were only ever produced once.

Kanzi?s combinations accounted for only 6% of his total utterances during this period.

Compared to a common chimp called Nim (another chimpanzee trained to use human language) this is not as impressive. Nim produced 19,000 combinations during a similar period of time. However Kanzi?s combinations were more impressive in that they provided more information. For example, Kanzi could produce combinations like ?ice water go? (with ?go? conveyed by gesture) to ask some one to get ice water for him.

An interesting aspect of Kanzi?s three-item utterances was that he used many of them to specify individuals other than himself as the agent or beneficiary of actions. Of Kanzi's three word combinations, 36% were used to specify a beneficiary other than himself, whereas none of Nim's combinations functioned in this manner. Nim?s most frequent combinations related to food whereas Kanzi?s were more likely to relate to games (e.g. by indicating ?chase grab? at the keyboard and then taking one person?s hand and pushing it to the second person)

The Issue of Imitation

The researchers were interested to compare Kanzi?s and Mulika?s rates of imitated utterances with those of human children, Nim and Sherman and Austin.

Like human children Kanzi and Mulika imitated most often when they were learning new words. Their proportion of imitated to spontaneous utterances was similar to children?s with a much higher percentage of utterances being spontaneous than imitated. This means that most of the time Kanzi and Mulika are actually adding new information to a new situation.

Even though the majority of Sherman?s and Austin?s utterances were also spontaneous in that the teacher did not make a query for a response the teacher did generally do something to elicit a response such as place food in a tool site. Therefore Sherman and Austin?s utterances were more closely linked to situations promoted by actions on the part of the teacher than was the case for Kanzi and Mulika.

Nim demonstrated a much lower proportion of spontaneous to imitated utterances.

Formal tests

Even though the formal tests were novel to Kanzi and Mulika they did well on these tests from the start. They seemed to understand that the experimenters were not communicating about something that was going to happen as they touched a symbol but rather were posing a specific question.

Kanzi's and Mulika's could select lexigrams in response to the spoken English word, select photographs in response to the spoken English word, and select photographs when shown lexigrams.

At the time these tests were given to Mulika, it was thought that she knew only a few lexigrams well enough to accurately select them out of context in a formal test. However, her test results quickly revealed that her capacities had been underestimated and illustrated that she knew 42 symbols, a number of which she had not yet used herself.

Kanzi did have some difficulties when the word was produced by a synthesiser although even the researchers found it difficult understanding some of these words.

When Sherman and Austin were tested they were initially confused because they anticipated that, when they identified an object, they would then get the object (e.g. a banana). Sherman and Austin were able to select the correct photograph when shown the lexigram; however, their performance dropped to chance when they were asked to select the correct photograph in response to a spoken English word. They also seemed not to like these trials and attempted to avoid them by requesting to go elsewhere; and at times, they refused to respond. Both Sherman and Austin repeatedly scratched themselves over their entire bodies (a behaviour that signals frustration in chimpanzees) during English trials, but not at all during lexigram trials.

Announcement and Verification of Travel Plans

When Kanzi was about three years old a ?blind? test of the foraging sites was arranged with someone who had never been in the wooded area and therefore could offer no cues. Kanzi was able to choose a symbol or photograph from a selection and then lead the experimenter to the correct location. En route, Kanzi often pointed to the photograph or to the symbol, as though to remind himself and the experimenter where they were headed

When he was showing the ?blind? visitor around he directed him to the back of the 55 acres, an area where he was not usually allowed to go, presumably just to explore this area of dense bush. He then directed them back to the trial. There were two locations not selected by Kanzi; the visitor used spoken English to ask Kanzi to lead him to these places as well, which he was able to do. It seems as though Kanzi took advantage of the blind experimenter's na?vet? to go to places that he did not normally get to travel. After exploring this area, Kanzi led the experimenter back to the trail and on the correct location.

This test was not given to Mulika because, at the time of this report, she did not like to travel without Kanzi.

General Observations Regarding Kanzi and Mulika's Symbol Usage

Like other chimpanzees that have been taught language, both Kanzi and Mulika were able to made generalisations beyond the particular meanings of a symbol. For example, Kanzi used tomato to refer to a variety of small round red fruit (e.g. strawberries and cherries). Similarly, Mulika used apple to refer to plums.

Kanzi was also able to use symbols in different ways depending upon the context. For example, juice was used as a food name and also to refer to the location (treehouse) where juice is typically found.

Kanzi frequently took the lexigram keyboard and went off by himself to use it, as if he wanted to practice. For example, he often pointed to hide and then covered himself with the keyboard or blankets. When the experimenters attempted to interact with Kanzi during such activities he would end the interaction and depart. The researchers believed that he did so because such activities were not meant by Kanzi to be social activities.

Kanzi also used the keyboard to indicate that he wanted to behave differently. For example, if he was taken inside when he did not want to he could respond ?no play yard? to indicate that he did not want to be where he was.

 

 

Explanations

The researchers concluded that pygmy chimpanzees can spontaneously acquire human language.

They give four comparisons between common chimpanzees and pygmy chimpanzees.

Firstly, common chimpanzees like Sherman and Austin needed to be trained to use symbols whereas the pygmy chimpanzees acquired their symbol use spontaneously through observation. Kanzi and Mulika did not need to be taught to differentiate between naming and requesting, nor did they need to be taught receptive skills. Unlike Austin and Sherman, Kanzi's and Mulika's receptive skills typically preceded their productive skills. Because they learned words before they began to use them, when usage did appear, it was usually appropriate from the start, just as is the case with normal children. The researchers argued that Matata did not acquire symbols spontaneously because there may be a critical age in the pygmy chimpanzee, beyond which acquisition of a symbol is very difficult.

Secondly, pygmy chimpanzees comprehend spoken English, whereas common chimpanzees do not. Mulika and Kanzi were not taught to respond to English commands and no one said single words to them over and over again. They only ever heard normal language and extracted words from this. Their ability to comprehend spoken words was completely unforeseen when the study began and gives the pygmy chimps a distinct advantage when using lexigrams.

Thirdly, the pygmy chimpanzees and the common chimpanzees differed in the way they used general and specific symbols. Common chimpanzees easily generalise symbols but could not easily learn to specify. For example, they may use ?juice? and ?coke? interchangeably. However, Pygmy chimpanzees spontaneously acquired the ability to specify. For example, Kanzi might call coke ?juice? only if coke is a novel item and no symbol for coke has been used before. Furthermore, once this difference between symbols had been acquired he would not make such mistakes. For example, once the symbol for a coke was assigned, the mistake was not made again.

Fourthly, Sherman and Austin never formed requests in which someone other than themselves was the beneficiary of the request. The same was true of Nim. Kanzi however was able to request that A act on B, when he was neither A nor B. The researchers argue that this particular ability to interact suggests that Kanzi was following complex rules and therefore could lead acquiring the rules of syntax.

 

 

Evaluation

The research method could be described as a longitudinal case study and therefore allows in depth data to be collected and allows development to be studied over time. For example, every utterance made by Kanzi was documented by the researchers over a 17 month period.

It is also possible to argue that the study was high in ecological validity as Kanzi and the researchers could roam from place to place around the 55 acre site. However the ecological validity can also be questioned as the subjects were not reared in their natural environment.

The data was gathered under rigorous controls such as the formal tests and are therefore are less likely to be open to bias and subjectivity. This improves both reliability and validity. Similarly the data gathered were quantitative allowing for analysis and comparisons between chimpanzees to be made. Qualitative data were also collected which improves the richness of the study.

It is possible to criticise the ethical nature of this study. Is it necessary to study chimpanzees in a human environment and to test their language skills in such a formal way? Note though that the normal ethical guidelines do not apply to non human animals.

The researchers did note that they were generalising from a very small sample of chimpanzees and a study of more subjects would have to be carried out for the study to be more representative.

Many researchers still doubt whether Kanzi and Mulika were using language in the complex way that humans do.