Before DNA could be decoded, it was suspected, correctly as things turned out, that humans were more closely related to chimpanzees than to any other primate. So chimpanzees represent the most obvious species for comparative research. In 1967 it seems to have occurred to Robert Allen and Beatrix Gardner to try and debunk the entire apparatus of Chomskyan linguistics by teaching language to a chimpanzee using the behaviorist technique which Chomsky had argued in 1959 was incapable in principle of accounting for human language. Other had tried to teach chimpanzees language before, and failed. The Gardners reasoned that if chimpanzees, obviously very smart and dextrous animals, could not be taught speech, it might be possible to teach them American sign language. They found a chimpanzee, known as Washoe. The Gardner’s student, Roger Fouts, showed that it was more effective to let Washoe learn by simple observation – the way that human babies naturally learn to sign or talk. And the behaviorist teaching was abandoned. By the time of Washoe’s death from disease in 2007  at the age of 42, she had learnt some 350 signs,  had signed “YOU ME OUT GO”, “YOU ME TIME EAT?” “BABY IN MY DRINK” (of a  doll in her cup), and signed correct and appropriate answers to questions including “WHAT THAT?'” “WHO THAT SHOE'” ‘WHAT COLOUR?” “WHO COME?”. She might seem to have taken something like the first steps towards language, implicitly querying the claim that language is human-specific. Soon afterwards a series of other similar experiments with apes were carried out, with varying degrees of succcess. The experimentation with a gorilla has not been as tightly scrutinised as that with chimpanzees.

The successful teaching of chimpanzees is expert, intense, prolonged, and highly structured. One of the teachers, Laura Ann Petitto once told me that on her count it took 4,000 presentations to teach a chimpanzee one item.

Washoe’s achievement

Any animal capable of distinguishing between more than one call has to be able to put two things into the same box, a meaning and a way of externalising it.

There is a single branching here. the externalisation involved one or more sensori-motor gestures, which could be seen or heard, implemented with the hands, face or head, vocal hoops or howls or by the shaking of trees. The meaning could be pragmatic, like a greeting or a curse, or semantic, like an expression. of feeling. The expression does not look or sound like what it symbolised. Singly branched calls could be of any length and complexity, for any purpose. The relation between the two atomic elements of expression and meaning was essentially arbitrary, with neither the externalisation nor the semantics more important than the other.

Washoe’s expressions involve more than one branching. But there is no reason for postulating anything other than a simple linear order, as by the way they are heard. Similarly there is no way of classifying them.

All non-human systems seem to be limited in this way, greatly restricting the repertoire of signs.

But Washoe’s achievements can be misunderstood. In some respects the structural complexity of the Washoe’s language was that of a normally developing child of around two. In respect of her use of pronoun forms equivalent to you and me, her language was more like that of a child of two and a half. Although there is some significant complexity in what Washoe signed and understood of others signing, it is of a different order to that of the human three year old. In particular there is no sign of embedding clauses within clauses, as by “I want to go out”.

A hundred or a thousand times better than what came before

It seems reasonable to assume that at the point when modern human ancestors diverged from modern chimpanzee ancestors, seemingly forsaking the relative safety of life in the trees for the greater opportunities and greater risks of life on the ground, they continued to use the same system as their erstwhile conspecifics. A recent paper by Cat Hobaiter suggests that modern chimpanzees command about 80 meaningful signs and gestures.  It would seem unlikely that at the point when modern human ancestors diverged the ancestors of modern chiimpanzees commanded any more than 80.

Cat Hobaiter (2023) estimates that chimpanzees, seemingly the most prolific of non-humans, have a maximum repertoire of some 80 meaningful signs, gestures, hoots. Current work is on how these signs can be modified for any purpose (personal communication).

Assuming that modern chimpanzees continue to communicate as they did when human ancestors diverged from modern chimpanzee ancestors, such a branching must have characterised their communication at that point. But by the evidence of the different levels of success of chimpanzee babies trying to learn a human word and human babies doing the same, there is a clearly significant difference here.

By the age of three the normal human infant is learning 10 or so new words every waking hour, learning in one week more words than Washoe learnt in her whole life. In contrast to Washoe, the normally developing human child can seize upon a new word if it seems to be interesting from hearing it just once. And the lexicon can be expanded throughout life with no maximum limit. A vocabulary of 10,000 root forms, 30 times greater than Washoe’s, would be small. A vocabulary ten times greater is quite possible.

As far as words or signs are concerned, the human mind is a vacuum cleaner. The rate at which the human infant is learning new words or signs suggests that these have to be represented in a special way, differently from any process in any non-human mind. By the proposal here, the human difference is by the principle which Diana Archangeli (1984) called ‘Radical Underspecification‘. This hugely economises on the storage requited for the lexicon. This is quite different, I propose, from anything that happened in Washoe’s mind as she learnt her 350 signs from the Gardners and their students.

Other animals

Exceptional dogs have been trained to recognise more than a thousand words and simple word combinations. Most dogs recognise less than a dozen commands.

There are numerous research programs investigating communication between monkeys, dolphins, and other cetaceans.

Less intensely scrutinised research has been done with gorillas.

Only us

Human babies are not just better learners. They are in the order of a thousand times better. The simplest explanation of a difference on this scale between humans and all other animals is that the human capacity for language owes something to the human genome. Humans are differently wired.

Any adaptation capable of enhancing the fitness of the organism in this way has to be simple, accidental, and such that it can be genetically encoded. It must have evolved to be the way it is.