The use of alternative artificial languages to communicate with apes

Several alternative approaches to teaching apes a language have been elaborated basing on special devices and thus allowed to compare language abilities of animals more objectively.

Premack (1971, 1976) elaborated an artificial language in which the words were plastic figures, which varied in shape, size, texture, and colour. Plastic words were arbitrary. For example, red square meant “banana”, a black T-shaped figure denoted the colour “yellow”, and a pale blue star meant “insert”. Sentences could be formed by placing the tokens, which had a metal backing, in a vertical line on a magnetic board. Sarah, Premack’s brightest chimpanzee, was taught nouns, verbs, adjectives, pronouns, and quantifiers; she was also taught same-difference, negation, and compound sentences. Consequently, Sarah taught to perform complex tasks by constructing sentences and following instructions composed of plastic words. If she created a correct sentence in response to an answer, then the trainer would place a word representing “correct” on the board, praised the chimpanzee verbally and perhaps give her a small reward. On trials when the sentence was wrong, a word representing “incorrect” was placed on the board, and the trainer said something like “No, you dummy”. To test Sarah's view of words, Premack presented her with an apple and a set of features (for example, round vs. square and red vs. green). Then she was presented with her “word”for apple, and the same set of features. She chose the correct features for both the real apple and her word for apple, a light-blue plastic triangle (Premack, 1976).

Sarah became proficient in the use of about 130 words, including complicated ideas such as “colour of”, “same”, “different”, and “if…then”. The questions she was presented with were formed by putting the token for “query” in front of a sentence:

QUERY RED COLOUR OF APPLE (are apples red?)

“Reading” the sentences Sarah could answer questions or obey instructions such as “Sarah insert apple red dish banana green dish”.

In one task two coloured cards might be placed one on top of the other, on the table, and a sentence on the board would ask “? red on green” (is red on green?). Sarah correctly answered by placing “yes” or “no” token on the board in dependence on what card was placed on top. Premack (1983) concluded that, like human words, plastic words used by chimpanzees possessed both representational and communicative functions.

Premack’s technique has been developed further a project supervised by Rumbaugh (1977), in which the symbols serving as words - or “lexigrams” as they were called- were displayed on a keyboard connected to a computer. Pressing a key resulted on its symbol being projected onto a screen above the consol. The language of lexigrams, each of which represented one word, was elaborated in Yerks Regional Primate Research in Atlanta, Georgia, and called Yerkish. The symbols were arbitrary and formed by combinations of geometric figures on different colour backgrounds, For example, a small solid circle inside a larger diamond on a purple background was a symbol for Lana, the first chimpanzee who was trained to use this variant of language. Lana used the symbols primarily to ask for things (Please machine give apple Lana). The purpose of the computer was to keep a record of Lana’s statements and to dispense films, slides, music, food, and drinks, when requested. However, Lana’s requests not always could be met; for instance, she once asked “Please machine tickle Lana”.

Lana started using NO as a protest (for example, when someone else was drinking a Coke and she did not have one) after having learned it as a negation (" it is not true that..."). Lana acquired many linguistic-type skills for which she had received no specific training, which showed her ability to abstract and generalize. For example, she spontaneously used THIS to refer to things for which she had no name, and she invented names for things by combining lexigrams in new ways. Lana’s ability to generalize showes that her system has semanticity, that is, she understands that a symbol refers to a certain type of object, not just one particular thing. For example, she was taught the word MORE in connection with an axtra ration of fruit-juice. Within a few days, she was readily attaching the symbol for MORE to other types of food and drink (“more milk”, “more bred”). Lana also showed some evidence of creativity. For example, she was taught the words “put” and “in” in connection with putting a ball into a bowl or box. Soon after, Tim, one of her trainers, was late with her morning drink of milk. Lana spontaneously made the request “Tim put milk in machine”. This shows not only creativity, but also displacement. In addition Lana used descriptive phrases such as “banana which is green” for cucumber or “apple which is orange” for orange (Rumbaugh and Gill, 1977).

Later two other chimpanzees, Sherman and Austin, educated in Yerkish, demonstrated the social use of language (Savage-Rambaugh, 1986; Rumbaugh and Savage-Rumbaugh, 1994). They have been trained to ask for tools to open food containers using lexigrams and to request and share food with one another. Similar results were demonstrated by chimpanzees in experiments of Schastnyi and Firsov (1961). Two chimpanzees were housed in adjacent cages. One of them was hungry but was lucky to have a toy in his cage. Another chimpanzee was fed and after that had been given bananas. When a hungry individual held a token that meant “food” to his mate, the satisfied guy accepted a token in exchange for a banana. A short time later he “asked” a toy in exchange for a token and had also been satisfied. These chimpanzees had never been taught to exchange by food and toys, so they acted spontaneously.

One of the most interesting studies of language learning in primates is that of a bonobo (or pygmy chimpanzee) called Kanzi (Rumbaugh and Savage-Rumbaugh, 1994; Savage-Rumbaugh and Brakke, 1996). Initially, researchers began by studying Kanzi’s foster mother Matata with lexigrams but she failed to learn Yerkish well. However, even though no attempt was made to teach little Kanzi, he started to use lexigrams to request and name things. As soon as Kanzi started to use the lexigram keyboard, he received special training. Instead of taking part in formal training sessions for a restricted amount of time each day, he received constant attention and spent his time as a full member of the human group that worked in laboratory. Humans both spoke with him and communicated on a portable lexigram keyboard. Using this keyboard Kanzi has communicated with humans more and more successfully. For instance, he would press the lexigram for " tickle" and then point to the person he'd want to tickle him. He followed strict order of action first and object second, just the opposite of the rule in English of the noun first then the verb second. This difference in communication indicates that not only does he seem to be imposing rules for how he " speaks" but that he made up this rule which is contrary to how he had been spoken to or taught, indicating productivity of language use (see Fig. IX- 2).

Kanzi went further by demonstrating his ability to discriminate on the basis of placement of an object. He was asked to bring the ball into the room from the outside. There was a ball inside already. The researchers wanted to observe whether or not he understood the difference between the balls and the words (inside and outside). He went directly to the ball outside and brought it in. This action (together with his reactions to many similar events) indicated that not only did he understand the difference between the two balls and the two words, but he also understood where the object was versus where he was.

Kanzi continued to receive training in the use of lexigrams, but these were now associated with the appropriate speech sounds. He was able to make associations between the sound of an English word and the picture of the object the word represented. For example, Kanzi listened to the voice coming from headphones he had on that asked him to give the picture of a specific object (i.e. a mushroom) to the person in the room with him. This person was not informed about what Kanzi was told to do, in order to avoid giving him any cues to the correct response. Kanzi was very successful at the task of associating word to picture.

After several years of training, a series of different carefully controlled tests showed that Kanzi could both understand and produce sentences. For example, he was able to follow instructions such as “Make the doggy bite the snake” operating with his dolls. In general, Kanzi displayed ability to request activities that involve other individuals like " Person 1 tickle Person 2". All the other language-trained apes only use themselves as an agent or a recipient of actions or objects. Kanzi was accurate in his understanding of location of items. For instance, if he was asked to get a pine cone from the fridge, he did so even if there was another pine cone in full view.

An important difference between Kanzi and the primates on previous studies is that much of his communication is spontaneous and not made in response to his trainers. He used lexigrams not to ask something but just to comment his actions like children do. For instance, he indicated “blanket” lexigram and began to play hide-and-seek, covering his head with the blanket. When eating an apple or a piece of melon, he would point to the corresponding lexigrams thus making his comments of the situation, not asking for food (Savage-Rumbaugh et al., 1986, 1993, 1998).

Researchers claim that Kanzi’s comprehension skills are equivalent to those of a child that is 2, 5 years old and that Kanzi’s communications, made with lexigrams and gestures, were structured according to rules, so Kanzi used a simple version of grammar (a “proto-grammar”). For sentences containing an agent and an action, the action was consistently placed second, whereas for sentences containing an object and an action, the action came first (Rumbaugh and Savage-Rumbaugh, 1994; Savage-Rumbaugh and Lewin, 1994).

Savage-Rumbaugh and her colleagues have raised other chimpanzees and bonobos in a similar fashion to Kanzi with close success and, given a failure with older individuals, have suggested that there is a sensitive period early in primate development when apes need to be exposed to language. Researchers believe apes assimilating intermediary language are communicating very complex things. It is worth to note that many of the questions raised by critics about the linguistic abilities of apes are not relevant to Kanzi and other bonobos. It is still unclear whether their achievements can be attributed to species difference between Pan troglodytes and Pan paniscus or to peculiarities of the experimental paradigm.