SCIENCE AT THE SHINE DOME canberra 2 - 4 may 2007

Symposium: Development and evolution of higher cognition in animals

Friday, 4 May 2007

Panel discussion and summing up

Chair: Professor Lesley Rogers

Lesley Rogers: We can take some more general questions, but I might just answer the question that was asked after Nathan Emery’s talk about how we might cope with the idea that this tool-using occurs very rarely.

I was reminded of the work done on orang-utans; it used to be thought that they rarely use tools in the natural environment, although rehabilitating ones use them a lot. It wasn’t until somebody ventured into the deeper jungles of Sumatra that they came across orang-utans living in an area where they have resources of insects in notch-holes, and they found that the orang-utans here actually manufactured a probing tool to insert into the holes – much as do the crows in New Caledonia.

That made me think: we know so little about animals in their natural environment – we are really only beginning to scrape the surface – and it is only recently that people have gone into the natural environment with their eyes more open, perhaps, to the possibility of seeing these behaviours. In my opinion, we might well turn up many more examples of tool-using in future, as animals are researched in their natural conditions more strenuously. It was Gavin Hunt who stumbled across the example of the crow, and that wasn’t very many years ago. So there is probably a lot more out there waiting for us.

Question 1: May I just reinforce what you have just said? I think the one thing that we are most ignorant about is the information processing habitats of animals – what it is that they need to comprehend about the world and what makes them good survivors.

On a more general level across the different talks today, what seems to be clear is that there are different aspects of cognition that now come under experimental scrutiny. For instance, in these comparisons between birds and bees and primates, what is often forgotten is that a lot of our cognition has to do with social structure. There are these habitats in which tool use, for instance, is extremely useful, but there are others where that isn’t a cognitive problem. And then there are social environments where foraging may not be the big problem but sharing food and its meaning for the whole society is the important thing.

So that is, in a way, what was asked earlier: what is our definition of cognition if we know so little about the different cognitive habitats animals live in?

Patrick Bateson: Listening to the talks I have been very struck by how, a lot of the time, the ideas about cognition are top-down; they come from human ideas about how animals process things. Russell Gray made the point very strongly, in his talk, that one should think about it in other ways, and I was thinking how important it is to start thinking about simple rules which can generate this very complex kind of behaviour. It was obviously true for the honeybees and it may well be true for the scrub-jays and all of the species we have been hearing about, that it is really important to go the other way round and start to think how you can generate complicated behaviour from simple rules.

Question 2: Today we saw some very good examples of very strong empirical evidence showing how these animals can generate quite abstract pattern-recognition skills. One thing that didn’t come through to me was how much, in their learning, they get might stuck upon those particular patterns that they pick up. So if the patterns have an element of non-stationarity to them, do they still keep fixing to the early lessons they have picked up on?

Russell Gray: I can’t comment on the bees, but I can comment on the New Caledonian crows, in that metatool task. In the first 10 trials they had been using a small tool to get the big tool out of the box. When we put the big tool out and had the small tool in the box, what all of them did was to take the big tool over to the box, look in, and then say, ‘Nah,’ and go straight off with the big tool. So they were able to respond flexibly but there was an element of past associations there as well.

Patrick Bateson: I would just say that we all know it is possible to learn a motor skill but if we get a bad habit it is very difficult to get rid of it. The problem with a lot of the study that is done in the laboratory is that the animal is still around and virtually has to be reused again and again, and there is a real problem of overtraining, so that you get animals that are used to doing a particular kind of thing and then you try to train them on something else and they do badly. So I am quite sceptical when people fail to show that animals do different things consecutively. So there is a real problem of interpretation, I think.

Question 3: This is probably more of a comment than a question, but in response to what Patrick was saying: if we strip away our personal phenomenology when we are looking at animal cognition, perhaps we could also do the same when we look at our own cognition, so that we can learn things from the way the animals perform the tasks, and get a greater understanding of exactly the way that we make decisions or use tools, rather than thinking of it in a higher phenomenological sense.

Question 4: The bird species that were considered today, with the exception of the magpie, are all corvids. That is probably not a mistake; they are, I think, universally known or suspected to be damn smart. Biogeographers would also say that they are, since the Pleistocene, extraordinarily successful, in that they have colonised the world. So it is entirely possible that this higher cognition that has drawn people to look at corvids also explains their distribution and success across the Earth.

A question for Nicola now: could you remind me, as you probably know this. On the margins of where these corvids live there is a high-altitude one, Clark’s nutcracker. That thing is known to cache some phenomenal number of pine nuts, and then snow lands on them and they dig them out. Can you remind me of the numbers we are talking about here, so we can get some feeling for just what kind of memory these things really have?

Nicola Clayton: An individual bird may hide as many as 33,000 items in about 6500 different locations. That’s one bird, in one autumn. The claim is that in the wild they recover typically about 80 per cent of those seeds, and an experiment in the laboratory – I feel very sorry for the poor PhD student who did this –was looking at how long their memory was. Earlier in the discussion we brought up how long memories might be, and I was thinking, ‘Gosh, this pales into insignificance compared with the nutcracker.’ This poor person made the retention intervals longer and longer and longer, and found that with a gap of 285 days there was no evidence of forgetting. Can you imagine the poor PhD student’s thesis!

Question 5: I want to make a somewhat different comment reflecting Patrick Bateson’s comments of a moment ago. Koehler was studying the apes on Tenerife during the First World War, out of which came his claim that here was a new form of learning called ‘insight’ – insight was essentially seeing the relationships between things rather than the individual elements, hence the beginnings of the Gestalt movement.

But what is not generally known, and I have seen it mentioned in no textbook in the last 25 years or so, is a paper involving two chimps, which appeared in the American Journal of Psychology some time in the late 1940s or early ’50s. One chimp was reared in the usual conditions of an enclosure, but at no time was it allowed to handle sticks. (So, as it were, no beating about the bush!) And the other chimp, out of sight of the first one – when the first one was put to bed – was given lots of sticks to play with in the enclosure. The two chimps were then given the raking-in-the-banana test, in which a banana was put outside the cage and they needed to put sticks together to get it. Now, Rana – the great favourite of Koehler’s apes – would sit down with his head in his hands, as if having a think, and then get up, walk over, put the sticks together and rake in the banana. And the animal who had had experience of sticks did the same thing like a shot. The animal who hadn’t, never did it – even after numerous trials and attempts to do it. And in due course in his last paper Wolfgang Koehler withdrew the notion of insight as a special form of learning, as opposed to the old Thorndykian trial and error learning.

So there was a case where we didn’t know enough about those animals in order to draw proper conclusions about them. They came in to the Tenerife enclosure without any knowledge at all of where they had been before or what they had been doing.

Question 6: I am interested to know at what point evidence from a non-human species becomes germane to a human phenomenon. For example, we heard quite a bit this morning about imprinting and the previous exposure breeding a preference. In human beings the mere exposure effect, if you have been exposed even briefly in a test to a word or a picture, if you see it again apparently there is a more positive valence given to that. So again familiarity breeds liking, or a preference. At what stage does evidence within the animal world become germane to a mechanism in something that looks rather like that in the human case?

Giorgio Vallortigara: I think that if you look at the basic toolkit of cognitive mechanisms that could be observed in the very young infant, you can say that they have from the start some sort of representation of the physical properties of objects, as it is revealed by, for instance, occlusional phenomena. They have from the start some very prelinguistic representation of number, for small numerosity and also for large numerosity, and they have also a representation of space.

I think that this toolkit of basic core domains – knowledge, abilities – is shared among vertebrates. It is quite common. It could be observed in a variety of other species, and probably through very little experience. At least in very young chicks these core domain abilities are apparently there from the start. That I think would be quite common, and probably is the root on which the higher forms of cognition can be built.

Gisela Kaplan: There is another reply I would like to give from the idea of vocal learning. That is one area where we are now learning, so to speak, from birds in order to understand human mechanisms. So it has gone the other way round, because you need auditory feedback mechanisms in order to do that. Secondly, there is a second propensity or ability of birds to repair damage in the brain in case of impairment of a vocal output. Neurons can migrate and change shape and reissue functions that were destroyed, and basically repair that impairment.

In the 1930s, much of bird song was discussed in music departments, whereas it is now discussed in medical faculties. There has been a tremendous shift from musical appreciation of bird song to the newer scientific basis of the mechanisms of vocalisations. If, of course, we solve and understand those questions, this has enormous implications for treatment of damage from strokes and other trauma or disease. So we are learning from those mechanisms in birds.

My criticism today, which I will put very cautiously and very gently but I am actually very passionate about this, is that so far the study of songbirds has relied on songbirds that produce song and sound for a particular reason: not for very complex communication but rather to attract a female. That means some bird may sing beautifully but say in essence the same as a bird species that doesn’t sing all that much. It says, ‘I’m healthy, I’m well, aren’t I beautiful, take me’ – a sort of mating arrangement in a hurry, in populations where you have snowfall and short breeding seasons. But it may not be a very useful model for us to understand vocal mechanisms more generally, unless somebody wants to argue that human speech development is solely motivated by reproduction! Some people may cynically say that is so, but that is probably not a useful premise.

So what we are looking for, and what I think I have found in the magpie – and we are now working in teams on this, together with one of Russell Gray’s colleagues, Martin Wild, and with Roderick Suthers, from the United States – is firstly to understand what happens, and what are the mechanisms and the outputs, in a bird species that sings all year round. In fact, the magpie continues to vocalise in ways similar to the ways we vocalise, because we are a very chatty species when you think about it. And, secondly, we are trying to understand what happens in a species about which vocal plasticity and brain plasticity are already shown. That means it is something that is maintained as a lifelong possibility.

So these are germane, very important for us. Another important point is, as Russell said, that not all useful models, such as the song control system, can be found in the primate line and it knocks quite a few theories on the head as well.

Question 7: I have a response to Nicky Clayton’s 264 days or whatever it was. I was unfortunate enough to be one of those PhD students! I did an experiment with groups of fishes, looking at a novel avoidance response, and I gave up after 11 months – 100 per cent recollection at 11 months. I had other things to do! I think the longevity of memory in animals can be absolutely astonishing.

Question 8: I would like to hear a little bit more from Dr Kaplan about magpies. You spoke about individual magpies and the search for a mate and so forth. What about interactions between magpies? I am very struck some mornings by the way in which several magpies will get up in a tree, one will start singing and they will all join in, in a great chorus, as if they are really enjoying the morning.

Gisela Kaplan: There is one very nice paper that has been done in a French laboratory to show the cognitive aspects of audition in starlings. One of the many studies by Hulse, Belgie and Isabelle George and others has shown that, in auditory perception, one hemisphere attends to distant vocal cues by conspecifics of the same group, and distinguishes those from foreign other birds, while close communication happens with the other hemisphere. And although there is a good deal of overlap, much more so than in the visual system, there is a clear distinction of identifying individuals by their auditory signals only. So there is no way a magpie from a neighbouring territory could sing its way into the territory and masquerade as somebody who belonged there.

We have complex vocal behaviour not only from songbirds but also from parrots. Research on budgerigars that was carried out in the 1930s again showed that budgerigars had a memory for quarter-tone differences in a message, even after seven months. Budgerigars aren’t even songbirds and they don’t produce particularly beautiful tonal quality, but their ability to recognise the difference is enormous.

So that is one part of your question. The other part of your question is the social part. When they sing together, very often that is a confirmation of very strong group cohesion, because this is part of the strategy that makes them survive and keep their territory. It is certainly a complex social order as well.

Patrick Bateson: I was just reflecting on how the work on animals might actually impact on the education of children. A point has come up in recent questions now which I think is highly relevant. It relates to the teacher from Queensland who asked this morning about what he should do for his students.

One point which I think is very important is how, in many schools in Western countries, the opportunity for play has been steadily reduced. So playing fields have been got rid of, time for play has been reduced, time for breaks or recesses has been taken away. And parents have got increasingly ambitious for their children – they want them to learn calculus at the age of five, or 10 languages at the age of six. This does take away the kind of thing that we now know is actually very important, in terms of exploring all sorts of different possibilities, not only socially but in terms of object play and imaginative play. So I think that if one had any message for the educational establishment it would be to give children more time for play.

Question 9: When I saw the program for today I was surprised that there was no mention of high-minded cetaceans, given all of the focus on the inquiry into whales and whaling and so forth on their intelligence versus ours, because of their large brain size. Has there been much development in the field of cognition in cetaceans and so forth? Will they save us?

Lesley Rogers: I am perhaps not the best person to answer this, but there is certainly ongoing research looking at brain size, learning capabilities and so on in cetaceans, and also in invertebrates – cephalopods and so on – looking at their abilities to solve problems, and people have been quite stunned by what they can do.

I might say, though, just personally, that I think sometimes the story about the intelligence of cetaceans might be rather overrated, or be pushed beyond what we actually do know. But they certainly have extraordinary abilities and they have rather extraordinary brains. I remember when we visited your laboratory, Giorgio, we saw that Paolo Zucca had a preserved dolphin brain. It is highly convoluted, with a huge surface area packed into that space, and it was quite stunning to actually see that.

One thing we haven’t touched on in any detail is what all this means to animal welfare, but I guess each one of us might take away some messages and start to think about it.

(Click on image for a larger version)

I put in this slide because it reminded me about animal welfare. This is a photograph that appeared in the Grafton local paper, showing a branding iron of 150 years ago, used to brand cattle in the Northern Tablelands. It is now on display at the New South Wales Parliament House. The reason it caught my eye is the word ‘SAD’, in that it is sad that this sort of treatment ever occurred for animals. But in fact the SAD stands for Sir Stuart Alexander Donaldson, who was the first Premier of New South Wales and had large cattle holdings in the Northern Tablelands area. He was also noted because he was the last person in Australia to engage in a pistol duel – so, other claims to fame!

While we have moved on from treating animals in these sorts of ways, there is surely a wealth of considerations ahead of us now. As we know more about the cognition of animals, many of the practices that we do in laboratories need to be modified. How much do animals know and suffer when they see other members of their species treated in certain ways in the same laboratory? When animals are lined up for the guillotine to provide material for neuroscience and so on, how much stress is caused to other animals in the same room that either see or hear this happening? The same applies, of course, to agricultural practices.

I think there is a wide-open forum there of issues that we need to consider, and will need to consider in the future as more of this information from our research comes in and filters through to consideration of the way we treat animals in various ways, in our homes, on our farms and in our laboratories.

Question 10: What you have just said reminds me of Jeremy Bentham, who said it is not a question of whether animals can think, it is not a question of whether they can communicate. The question is whether they can suffer.

Lesley Rogers: Yes, that is a very good point to make.

It remains simply to thank all of the people that have spoken today – we have had some very fascinating and stunning talks, and I have been very pleased to hear everything that has been said.