When a grunt means 'Sorry'

On the eve of a major conference on the evolution of behaviour in primates and man, Robin Dunbar explains why he believes animals are capable of culture

I think, therefore I am," declared the 17th-century philosopher-mathematician Rene Descartes, adding by way of afterthought that, since animals obviously did not speak, they could not think and therefore certainly didn't have souls. We have lived in the them-and-us of Descartes' shadow ever since. Nowhere has his influence been more intrusive than in the social sciences, where conventional wisdom has always insisted that the great divide between humans and other animals makes the latter totally inappropriate as models for the study of human behaviour. The great markers that set us apart from the brute beasts are culture and language.

The argument, of course, hinges on the uniqueness of these two key phenomena. The result has sometimes been near-farcical efforts to defend the honour of our species against upstart claims that mere beasts might aspire to such a noble condition. Every attempt to show that some animal or other possesses language or culture has been met with a counter claim that has tried to shift the goal posts by redefining the terms. Man-the-tool-user rapidly became man the-tool-maker when it became apparent that many species of animals do in fact use tools.

So what is this culture of which we are so defensive? Forty years ago, the American anthropologists A. L. Kroeber and C. Kluckhohn surveyed the literature and emerged with some 40 different definitions in current usage by anthropologists and social scientists. By and large, these seem to break down into three major classes of definition. Culture consists of ideas in people's minds (social rules, patterns of ritual, beliefs, etc); culture consists of artefacts that are the products of those minds (so-called material culture like tools, pottery and its decorations, clothing, etc); culture is language and its products (high culture in the everyday sense, everything from Shakespeare to Bob Marley). The last, of course, brings us back to that other unique pillar of the human condition, language.

Apart from some inherent circularities (only humans have language, therefore only humans can have culture because culture is language), most of these definitions beg questions about the uniqueness of human behaviour. Are animals' minds really empty? Do they have no beliefs about the world? Are the hammers and anvils that chimpanzees use for cracking nuts bona fide instances of material culture?

Bill McGrew (now at Miami University at Oxford, Ohio) has been a vigorous critic of the culture-as-artefacts school of human uniqueness. In his book Chimpanzee Material Culture, he challenges the advocates of this view to show why the chimpanzee's tool kit fails to meet the definitions they readily accept for humans. Three decades of intensive field work in Africa has uncovered a long list of natural and manufactured tools that chimpanzees use, ranging from hammers to probes, fishing tools to sponges. Were we to lose the labels from such exhibits in a museum, he insists, we would be hard pressed to tell whether they had been manufactured by humans or apes. In only one respect does the chimpanzee tool kit differ from that of pre-technological human societies, and this is in the absence of vessels of storage and the construction of traps.

Two other widely touted examples of animal culture have long since entered into popular mythology. One is the way bluetits learned to remove the cardboard discs that once capped British milk bottles: during the 1940s, the habit spread throughout much of southern England over a period of a decade or so. The other is the habit of washing sand off sweet potatoes that spread through a troop of Japanese macaques once the habit had been invented by a young female named Imo.

Both examples have, however, received hard knocks at the hands of psychologists during the last few years. Several careful reconsiderations of the data have pointed out that, for a culturally learned behaviour, the rate of transmission through the population was remarkably slow in both cases. It took literally decades for Imo's potato-washing to spread to the rest of the troop; even then, only animals that were younger than Imo learned to copy the habit. The old dogs never learned new tricks.

In fact, it seems that in most cases these new habits spread by a much simpler process. An observer animal's attention is drawn to a problem by the behaviour of the tutor, and it then learns the solution to the problem for itself by a process of trial and error. In humans, the tutor would teach the observer both the nature of the problem and the solution.

Observations of this kind have led psychologists like Mike Tomasello of Emory University at Atlanta, Georgia, to doubt whether any animal has true culture. But before we leap to premature conclusions, we might bear in mind the questions that are being asked. Tomasello is interested in the mechanisms of transmission; primatologists like McGrew are interested in what the animals actually do. By any conventional operational definition of culture, chimpanzees have culture, but we may legitimately doubt whether they learn it in quite the same way as we do.

It is obvious that what we often view as culture in humans is deeply embedded in language. We use language to describe, to teach, to intone our rituals. Animals, as Descartes observed, do not. Yet they are not dumb. Dogs bark, monkeys chatter. Conventional wisdom has always insisted that these are merely the direct products of the underlying emotions. Dogs bark because that is the kind of noise their vocal tract produces when they reach a certain level of excitement. While humans too produce similar kinds of vocalisations (screams and grunts), they also produce sound chains that are arbitrary yet meaningful. We can easily dismiss the much vaunted waggle dance by which honey bees notify each other of the direction and distance of nectar sources because they are specific to a very particular situation. Honey bees do not enquire after each other's health or sympathise over a misfortune.

Yet, recent research suggests that, when it comes to monkeys and apes, it may be necessary to turn this conventional wisdom on its head. Dorothy Cheney and Robert Seyfarth, of the University of Pennsylvania, carried out a series of ingenious experiments on wild vervet monkeys in Kenya's Amboseli National Park. By playing vocalisations of known individuals from hidden speakers, they have been able to demonstrate quite uncontroversially that vervet vocalisations convey considerable information that is independent of the behaviour of the vocalisers. Vervets reliably use calls to refer to specific kinds of predators (leopards versus birds of prey versus snakes). They know from minor differences in sound whether a grunt is a comment on what someone else is about to do or on something it has seen, whether they are being approached by a dominant animal or a subordinate. In their recent work in Botswana, they have demonstrated that baboons use grunts in a way that amounts to an apology in order to mollify an ally they have previously offended. And all this in what was once thought to be a simple all-purpose grunt.

There is, it seems, much more to animals's vocalisations than we had supposed. Like the proverbial visitor to China, the naive observer hears only a jumble of sounds where in fact something much more complex is going on. We have been, and still are, mere beginners when it comes to deciphering the languages of other species.

More impressive still are the achievements of the language-trained chimpanzees. Around a dozen chimpanzees, a gorilla and an orang utan have now been trained to use a variety of languages. These have included ASL (American Sign Language) and at least two languages in which arbitrary physical symbols stand for concepts (in the one case, magnetised plastic shapes assembled on a board, in the other a specially adapted computer keyboard). The chimps, in particular, have demonstrated remarkable abilities, responding to instructions and answering questions at the cognitive level of young children.

Worse still, most of these have been matched by an African grey parrot named Alex who uses spoken English to communicate.

There remains, however, one crucial stumbling block for animals. The ability to engage in the higher forms of culture that we associate with religious ritual, literature and even science depends on the ability to step outside oneself to see the world from an independent perspective. It requires being able to ask not just "What happened?'' but also "Why did it have to be that way?'' Animals, it seems, take the world as it comes. Only humans seem able to detach themselves from their own parochial concerns to imagine that things could be other than they are. Only then is it possible to ask the all important "Why?'' questions that adults find so infuriating in children.

In the social context, this ability to stand back from the way things are is referred to as possessing a "theory of mind''. It underpins our ability to understand another person's beliefs and the way we use this knowledge to exploit and manipulate each other. Children do not possess it at birth: they acquire the ability at around four years of age. In fact, some humans (such as autistic people) never acquire it. Neither sophisticated lying nor fictive play are possible until a child has acquired theory of mind. Without it, fictional literature is impossible and both science and religion, with their need to imagine impossible worlds, are out of the question.

It is equally clear that no animals reach this exalted state of mind. Monkeys can, of course, engage in deception, but it is deception of the kind that three-year-old children are good at. They can read another's behaviour well enough to exploit them, but they cannot understand that another individual can hold beliefs that are different from their own. The one exception, yet again, seems to be the chimpanzee.

Sanjida O'Connell, one of my postgraduate students, has been testing chimpanzees with the same kinds of tests that psychologists have used to identify theory of mind in children. These involve determining whether the child realises that someone else can have a belief about a situation that the child knows is false. O'Connell developed a mechanical analogue of these verbal tests. The chimp is taught that food will be placed in one of a row of boxes indicated by a marker (visible only to the chimp) that the experimenter placed above the box. On a later trial, however, the marker slips inadvertently to another box while the experimenter is not looking. Which box will the chimp opt for: the one where the marker is now or the one where the experimenter still believes the marker to be? The latter choice is only possible if the chimp can attribute false beliefs to others.

Chimpanzees pass the test by opening the original box, but monkeys fail it. More importantly, autistic people and very young children (both of whom lack theory of mind) also fail it. To be sure chimps are not as good as normal five-year-olds who rapidly go from strength to strength once they have acquired the basic skill. But it is strong evidence that, if a divide can meaningfully be said to exist, then chimpanzees lie on our side of it.

All this, of course, raises serious moral and legal issues. Should we defend the rights of chimpanzees (and perhaps other great apes) in the way we would not hesitate to do for humans who lack the capacity to be morally responsible?

Should they be kept in captivity against their will? Ought the law on habeas corpus to apply to them? Should they be used for invasive biomedical research, especially when the beneficiaries are humans and not other chimpanzees? The implications are a legal nightmare, but it may well be one we are going to have to grapple with. It is also a welfare nightmare, because most animals that have been in captivity could not survive in the wild and we would inevitably have to provide them with sheltered accommodation for the rest of their lives.

The substantive point is surely that the continued insistence that culture is a phenomenon that sets humans apart from the rest of creation seems to smack more of generic chauvinism than anything else. There are, of course, aspects of human culture that are not found in other species, just as there are aspects of language that appear to be unique to humans. These are but the high points on what in reality is a continuum. And therein perhaps lies part of the problem: humans seem to find it extraordinarily difficult to think in terms of continua, preferring instead to deal in simple dichotomies, of them-and-us. We should recognise that neither language nor culture are unitary phenomena and that we share many of the processes underpinning them with at least some of our fellow creatures.

Robin Dunbar is professor of psychology at the University of Liverpool. His book The Trouble With Science is reviewed on page 20, and he has organised a conference, "Evolution of Social Behaviour Patterns in Primates and Man", in conjunction with the British Academy and the Royal Society, which will take place at the latter on April 4-5.