Interview: Out of Africa

These days there are a number of different terms for scientists who study animal behaviour from comparative psychologist, to ethologist to behavioural ecologist. What exactly would you call yourself?
I would say I study animal cognition. I started life as a cognitive psychologist. When I was a student at university I found comparative psychology interminably boring because it seemed to be based on a misunderstanding of evolution – using a few kinds of animals in the laboratory as if they stood for past stages of our evolution. Although I realise nowadays people do comparative psychology in a much better way, I still can't bring myself to call myself a 'comparative psychologist'. In fact I don't think there is an 'ist' for what I do.

So if you started out in cognitive psychology, what was it that got you into studying animals like apes and elephants? Did you have pets as a child or was it purely academic interest?
It was purely academic. I'd always been keen on animals and loved watching animals for my pleasure – but I always kept that separate from my academic work. For my PhD I studied memory in the complex tasks of everyday life. At that time cognitive psychology was very much paradigm-driven and lab-based – people just studied what somebody else had studied before in the lab, adding a bit more. And I felt quite strongly, from an academic perspective, that we were losing a lot of tricks by not looking at everyday skilled behaviour. So in a way I was predisposed to discover ethology which looks at these sorts of things – it's the study of animal behaviour in naturalistic settings. And because ethologists study animal behaviour, that got me into the field.

One of the things that you are well known for is your work with Andy Whiten on the notion of 'Machiavellian intelligence'. Does this suggest that the driving force for the evolution of social intelligence is in order for one animal to manipulate and control others? And, if so, do you see it as a dark explanation of why we became so clever? 
No, I don't think so. By the way I should say it certainly wasn't our idea. The idea of human intelligence having an origin in social complexity and manoeuvres in society, there were a number of routes to that and Nick Humphrey was probably the most prominent figure in the development of that idea. The idea that natural selection should tend to make animals manipulate each other wasn't new either: that was really down to Richard Dawkins and John Krebs. The manipulation of others that evolution automatically produces is very much at the genetic level, and that may or may not result in dark nastiness at the interpersonal level. It may also result in the reverse – cooperative, empathic, kindly behaviour. The term Machiavellian intelligence has often led to misunderstandings because people, especially if they haven't read the original Machiavelli, think it just means being nasty. The genetic 'aim' is always selfish, but the machinery of how society works can equally go either way. We were just as impressed by animals being highly cooperative and building up support networks, in what you might call a nice way. They weren't just being cruel and unpleasant. But you could of course say that that is even worse – they are being nice to get their own way. Certainly that's what Machiavelli talks about. He talked about the Prince needing to be the most trustworthy, kind, reliable and well-liked member of society – until he needs to stop being so. Then he should change. And I wonder if we should expect anybody ever to be totally prosocial.

One area that intrigues me is your findings that there are convergences in cognition abilities between primates and elephants. Primates and elephants are very distantly related – so how can they have evolved similar cognitive abilities?
Actually it's the fact that they are so distantly related that makes it so interesting. I've felt very strongly for some time now that those of us who study primates – our close relatives – are missing something. It's not that we are on the wrong track, because studying the abilities of our relatives will help us to reconstruct the historical sequence of how we became like we are. But, in a sense, that's the bad news because the evolution of an ability that you study by this comparative approach only happened once in the shared lineage. Say we find a trait in chimps, gorillas and us: that means it's not uniquely human, and helps us to pin down when it evolved. But if we want to understand what good it was, what function it served and what circumstances allowed it to happen then we need to study other, ideally multiple cases. So we need to look at animals as unrelated to each other as possible. Animals such as elephants are marvellous because they are so distantly related to us. They are closely related to aardvarks and hyraxes and manatees. But unlike those animals, elephants appear to be smart. Elephants have converged with primates in some of their abilities, including social intelligence – and that can tell us about the evolutionary causes.

You mention how clever elephants are compared to many other mammals. Could you rank order animals in terms of intelligence or would that be far too simplistic?
I think simplistic is the wrong word. It would be very hard work and it might not be very useful. But I think there's nothing wrong with it in principle. It's a bit like rank ordering animals in terms of how fast they are. Speed has a lot of different uses – the use that cheetahs put it to when they are running down antelope and the use that peregrine falcons put it to when they are diving to kill pigeons are quite similar. The use that other animals make of speed may be very different, like avoiding predators. But they are all animals that use speed for evolutionary benefits. Well, I think intelligence is just like that. Cognitive ability has pinnacles of its evolution – hopefully not just the one we are on. Elephants, maybe dolphins and maybe all sorts of animals we haven't thought of have also reached their own pinnacles. So there's nothing wrong with rank ordering animals in terms of intelligence, but doing that would not mean that evolution is progressive with me on top!

Is it true that elephants are able to recognise which tribe a human belongs to by the colour and smell of their clothes?
Yes – it's quite interesting how we got into that. The researchers who had been studying elephants for many many years, in the Amboseli project set up by Cynthia Moss, had published research covering all sorts of aspects of their behaviour, and if you just read the papers you'd get the impression elephants were just another African animal. But if you talk to the researchers, none of them actually thought that. They thought elephants are very different, and uncannily more like humans than some big fat animal with a trunk and tusks ought to be. So it was talking to them that led me to getting a grant with my colleague Lucy Bates to look at all the anecdotal data – that is, the narrative records that researchers had simply put in their field notebooks, because they seemed important but weren't quite sure what to do with. So we categorised those data and it led us to make predictions and eventually even do various experiments.

For instance, one interesting thing that came out was that we found elephants really don't like cows. Now cows don't do any harm to elephants – but sometimes elephants kill cows. This didn't appear to make sense. But the thing about cows is that they are kept by the Maasai people and sometimes the Maasai people spear elephants – it's the young men showing off. Elephants live a long time and they maybe remember these nasty events. So we wondered whether the elephants were distinguishing among groups of people in the way that you or I would. And whether they were doing this on the basis of indirect cues like cows. So we did experiments with indirect cues to the Maasai. We used garments that had been worn by the Maasai, who are pastoralists and other garments that looked the same but had been worn by other people, the village living Kamba people that are agriculturalist. We put them in places where the elephants might happen to sniff the scent drifting down-wind. The reaction was dramatic, and the difference showed the elephants certainly were using scent to categorise people: they ran off if they detected Maasai, but not Kamba. We also found they were sensitive to the colour red, which is favoured by the Maasai. When we put red garments out the elephants would attack them – but ignore white garments. So they really can distinguish between different groups of people based on garments colour or a distinctive smell.

Another African animal you have found something interesting about is the feeding behaviour of gorillas.
When I first got the chance to see gorillas it struck me how very much like chimpanzees they were, in lots of ways. But at the time gorillas were not seen as like chimpanzees: they were big, dumb, cute animals that did boring things. Not like their clever little cousins, chimps, that did smart things like use tools and make use of Machiavellian intelligence. But in fact gorillas have to eat very difficult foods and although they don't use tools in the process, they do process their food. We found they devised very elaborate processing routines – rather like the sorts of things we do to turn unpleasant dirty, spiky vegetables into something we can eat. Although they don't cook, of course, they do deal with the physical defences of the plants in really clever ways. Our analysis showed they can construct hierarchically organised programmes of manual activity – almost syntactically – and develop new and highly organised manual behaviour in ways that correspond to the things that chimps do with tools. For instance, to eat nettle leaves with least pain, they strip the leaves off a stem with a hand cupped in a cone shape, strip several more by using their outer fingers to hold on to what they've got already, then twist off the leaf stems from the whole bunch with the other hand, and finally pull out the leaves and fold them over the thumb into a little parcel – a nettle sandwich – to pop into the mouth. It's neat, it works, but it's not part of the gorilla's genetic repertoire – most gorillas don't even live in places where nettles grow.

What about apes using gestures to communicate – don't orangutans use a lot of gestures?
It's really the same story for chimps, gorillas and orangs. We find that they all use gestures in very intentional, goal-directed ways. That is, they have a clear idea in mind of what they want from the individuals they are gesturing to, and they will persist or elaborate until they get what they want. This very intentional use of gestures has been found in all of the great apes now. A student of mine – Erica Cartmill – did some nice experimental work with orangutans. She set it up so that the person to whom the orang was gesturing, who was supposed to hand over the really delicious pile of bananas, would sometimes get it slightly wrong and hand over half the pile of bananas, or get it totally wrong and hand over, say, celery or something else that didn't taste as good. And what the orangs did is exactly what a person would do. When the person was on the right track the orang would speed up the rate of gesturing – but using the same gestures. Whereas if the person seemed to be completely dumb and handed over the wrong thing entirely the orang would use completely different gestures.

You've worked on imitation. We tend to think that aping someone is quite an easy thing to do, but imitation is far more complicated than that. 
Yes it is more complicated, and it's also the case that the word 'imitation' is used in about six different ways. Imitative learning – the sort of learning that you'd do when someone comes to fix your car and you watch them carefully and you think, 'OK I see you undo this, then you undo that, and you blow it out, and put it all back in the same order' – that sort of imitative learning seems to be very rare in the animal kingdom. Although I would argue that the great apes can do it. They can't do it from one presentation like that,  and we sometimes can, but that might just be a memory limitation on their part as we have much larger brains. There are many other meanings of the word imitation, such as matching each other's body posture when they are sitting together. We are probably doing that right now. Animals do show just as much of that one as we do. So some types of imitation are more special than others, and imitative learning – in the sense of constructing new behaviours from observation – is really quite challenging even for the great apes.

All psychologists are influenced by those that came before them – whose shoulders do you stand on?
Well, I really got into this through cognitive psychology and I was very lucky that as an undergrad I was supervised by Donald Broadbent. He was endlessly fascinated by how things worked and was happy to waste his time with a mere undergraduate. And I went on to be supervised by John Morton, a fantastic inspiration and a wonderful cognitive theorist. He was very good at encouraging my half-formed ideas to be dragged out. One of his rules was you should always be interesting – you don't have to always be right! I would say they were the people who gave me a fascination with how behaviour works, and got me constructing mechanistic theories of things that seemed airy-fairy psychological attributes.

But most of my career moves have been luck really. Turning up at St Andrews meant I was able to change from studying Cambridge housewives to looking at African apes and elephants. That sort of change invigorates me. I have been lucky that I have been able to move on quite a lot from where I started and avoid getting stale.

Animal cognition has come a long way in the last 20 or 30 years. What do you think the next step is on the horizon?
I think in recent years there have been a lot of discoveries of things animals can do. On the whole animal cognition is going in the right direction. It used to be regarded as all about piling up 'wow that's amazing!' things that animals can do, but in reality it's about discovering whether similarities are convergences or a result of common descent, to piece together how and why evolution has channelled cognition in different species.

I would like the next step to be to use the animal data just as developmental psychologists have used normal development and developmental disorders to take apart the abilities of children. So rather than say 'yes, chimpanzees can learn imitatively', 'yes, gorillas can communicate with intentional gestures', and 'yes, elephants can categorise people into subtypes' we might ask 'how do these cognitive abilities actually work?' Do they work in the same way in different species? Do they work in the same way in different periods of life? And really bring the mechanistic approach to bear in the way that developmental psychologists do.

Finally, what is the next step for Dick Byrne?
I must say, having spent many years working on non-human primates, I found it a real treat to work on the Amboseli elephant population. It's a marvellous project to be part of because the research assistants who work there all of the time know all the individuals in the population and their history. Imagine going out in the field with someone who can recognise 1400 different individual elephants! So there is so much to be done in terms of field observations and experiments with these elephant – and I'm personally looking forward to helping with some of that. Of course, someone has to go out and do all of this hard work on the African savannah with Kilimanjaro above the shimmering heat haze in the distance, why shouldn't it be me…?