SCIENCE AT THE SHINE DOME canberra 2 - 4 may 2007

Symposium: Development and evolution of higher cognition in animals

Friday, 4 May 2007

General questions of all speakers

Question 1: I would like to ask a question of Nathan Emery. If you get a novel widespread opportunity – for example, the simultaneous appearance of garbage bins through the British Isles – do you expect that your birds would have clever innovators and that their ideas would be disseminated, or do you think, at the other end of the spectrum, that they would all get the idea of lifting the liners and getting the food?

Nathan Emery: Do you mean something like the blue tits, opening the bottle tops?

Question 1 (cont.): Yes.

Nathan Emery: It does seem strange that it has just been seen in this one service station. I don’t think anybody has looked elsewhere. As you can see from the video, I think, they do seem to pay close attention to what others are doing. We don’t know the relationships of those individuals, because we have not studied them – it is just an anecdote – but I would imagine that there would be quite similar rapid social dissemination, because obviously this is such a high-calorific food that it is actually quite important for them.

It is actually quite a big deal, I think, for rooks, which tend generally to be quite neophobic, to encroach into these environments. So there is something good there for them and I imagine that that factor would allow it to spread, as would their being sociable.

Question 1 (cont.): With the blue tits, was there just one ‘genius’ that all the rest followed?

Nathan Emery: I’m not sure. I don’t think anybody would know that, even Robert Hinde. I think that they sample the environment, find that this is a very nice resource – it’s not too difficult to gain access through a tin foil top.

Question 2: This question is to Professor Srinivasan. In one of the pictures you showed us, there was a discrimination test involving a disc and a ring on random backgrounds. I was wondering whether audience members were supposed to be able to see the ring against the random background, because I couldn’t.

Mandyam Srinivasan: You might have noticed that I had shown the background at a slightly lower contrast than the actual object. The object had black and white pixels, whereas the background had white and grey pixels. That was just to make it a little more visible to us humans, but in reality the background had exactly the same contrast as the foreground. In other words, the texture in the background was exactly the same as the texture in the foreground, so it was totally camouflaged. The only way you could really pull out or break the camouflage was to actually move relative to the whole constellation, and then make the object ‘pop out’ against the background, because the object moved at a slightly different rate as compared with the background.

Question 2 (cont.): So we were supposed to be able to see it, were we?

Mandyam Srinivasan: You were supposed to be able to see it, just to visualise the experiment. But in reality, if I had shown it to you the way it was, you wouldn’t have seen it.

Question 3: I apologise for introducing a possibly practical note into an academic discussion, but I would like to ask a question of Professor Kaplan. I live in an inner-city suburb and our backyard is the preferred habitat for a large number of magpies, with obvious results for our garden furniture and sometimes for us. You mentioned an alarm or danger signal made by magpies. Would it be possible to buy a CD of this signal?

Gisela Kaplan: Any strange ‘magpie’ would be attacked, indeed. But it wouldn’t take the magpies very long to know who put it out there! And then the origin would be attacked. So that might not be the best strategy.

In order to be cognitively complex we would expect that they are able to auditorily discriminate the origin of the caller. In alarm calls in particular there seems to be also a social order. You may recall that I mentioned in my talk that alarm calls, and very timid ones, appear only at seven months of age. So there is clearly a developmental aspect to the appearance of certain types of vocalisations. In the timid alarm call a parent or the adult birds will reinforce that by either then alarm-calling themselves or by totally ignoring the youngster and walking away. That particular youngster will then eventually know whether it had vocalised correctly or incorrectly. It may be very similar to the way vervet monkey vocal development of particular referential signalling is shaped socially by adults.

Question 4: I would like to ask a question of Professor Srinivasan. If you were to do one of your clever experiments, first using an Italian bee and then an Australian native bee, and then a New Zealand native bee, which one would come out on top?

Mandyam Srinivasan: It is very difficult to decide. I suppose the native bees have a different social structure compared with Italian bees, so that has to be taken into account. Many bees are less social, they don’t seem to have as highly evolved a dance language as the Italian bee. But it would be very hard to decide. It would be really interesting to try that, for sure.

Question 5: I was rather struck by the wide range of the memory spans that we heard about today. I read recently, incidentally, that people who are working with goldfish find that they have a short-term memory of only nine seconds, and it takes about 18 seconds to go round the bowl. You sometimes feel sorry for goldfish, that they must get very bored, but apparently we don’t need to feel sorry because it is new to them all the time.

In Dr Srinivasan’s talk I think he mentioned in one case a memory span of about seven seconds, but then I think he went on to speak about days and months. Am I correct? There seems to be a very large range.

Mandyam Srinivasan: If I could clarify one aspect of that: the short-term memory, remembering that stimulus, was a different kind of memory. That was working memory. That is like a temporary scratch pad. It is the kind of memory you would use when you looked up a number in a phone book and then went to dial. But long-term memory is something where you learn the task itself, and learning the task in that case involves learning the notion that you have got to do the matching. That long-term memory can reside for a long period of time. That is a very long-term memory. But this working memory is the little scratch pad you use to scribble down exactly what you have noted down now, in order to meet the task.

Giorgio Vallortigara: I think Professor Srinivasan has already explained that in our object memory task, the chick was facing a working memory task. The memory should last as long as the chick has to retain a memory, and then it can erase that and go for another trial.

Obviously, one issue in this sort of task is whether the chick is maintaining some sort of orientation towards the object in order to facilitate its memory – for instance, toward the screen. So, for instance, we check for any difference using a transparent covering or an opaque covering, in such a way as to avoid any visual contact with the target during the delay. This makes a lot of difference, because when there is an opaque screen the chicks are able to maintain memory up to 60 seconds, but if there is a transparent one, even though it is not apparent that they are maintaining any visual contact by looking at the video recording, they are still able to maintain memory for up to three minutes.

Question 6: This is a question to Professor Srinivasan. You have told us a lot about very clever things that bees, with those tiny brains, are able to do, and you have referred to ‘top-down’ processing being evident from that sort of behaviour. One of the things which we do, and which I know and have published on in the past, is to have the capacity as human infants, as early as eight weeks, when presented with an object of a certain size which is then placed further away, to respond to that size in the same way as if it were close – even though the projection of the retina of the eye, as you would know, is much greater when it is near than when it is far. Is this so-called perceptual size constancy likely to be manifested also in the bee? Have you checked?

Mandyam Srinivasan: It is a very interesting question. Other people have looked at that. The ‘innate’ – the most hard-wired – response seems to be one where they learn the size of the object, the angle of subtense of the object. If you train a bee, for example, to find a food reward at a particular location and you have a piece of blue cardboard placed a certain distance away from it, and if you then take away the food and see where the bee comes and looks, it will look at the correct spot where the food used to be. But if you now double the size of the piece of cardboard, it will search at twice the distance away. So it is really learning the angle of subtense of the stimulus very well.

Question 6 (cont.): So it is essentially a size–distance type phenomenon?

Mandyam Srinivasan: Right, but you can train that away. If you can train the bee, ‘Look, you have really got to come at the right distance, regardless of the size of the card,’ then they will learn to use other cues such as motion parallax to judge the distance. That’s something you have to train them to do.

Question 7: I want to ask Dr Srinivasan a question too. In your experiment with the rose-scented feeder and the lemon-scented feeder, when you put the scent into the hive did only the trained bees come out, or did all the bees come out?

Mandyam Srinivasan: Just the trained bees.

Question 7 (cont.): If I could just follow that up with a general question: I wanted to get an impression of how much, then, these experiments are looking at individuals and individual cognition. How much has been done to see whether other animals copy? If you put the bee back in, do other bees learn? Is there communication between groups?

Mandyam Srinivasan: That is fascinating. That has not been looked at very much. Of course there is communication going on within the hive through the dance – a bee that finds a good food source will come back home and dance, and convey information about the distance and the direction in which to go to get to the food source, and the other bees will pick up that information. Again the interesting question there is just how much of this dance is really preprogrammed, and how much of it is actually learned. No-one really knows, but the impression now is that the basic properties are built in. It could be refined with age, but a bee doesn’t have very long to live – it forages for only three weeks – so if you have got to learn something, you have to learn it very fast.

The other question of whether one bee can learn from another bee through imitation has not been looked at yet. The only thing people have observed so far, as far as I know, is that if a bee sees a lot of other bees clumping at a particular food source, it will go towards that as well, thinking there is a good chance that there is some food there. But learning by imitation has not been looked at in bees, and that is a very open question.

Question 8: I am a teacher from Queensland. I wonder if any of the panel – who I think have all touched on ways of training animals to perform higher cognitive functions – would care to give me any advice on how to develop higher cognition in my students.

Patrick Bateson: Get ’em young!

Gisela Kaplan: I just want to give you some hope by referring to magpies. The production of vocalisation and the acquisition of complexity is a spontaneous and improvised act. I think humans and birds, particularly these very accomplished songbirds, are capable of that. Babbling in humans is spontaneous, it doesn’t have to be taught. And as to the formulation of language, apparently the latest study, which came out just this year, said that human babies in the first two weeks of life can detect differences in sound quality within a quarter of a tone, and distinguish between languages. They are already formulating an auditory soundscape of language, and so do magpies.

Among the latest theories of learning, one of the more notorious ones practised right now in Victoria is called ‘incidental learning’. You send your students to school to ‘incidentally learn’ something. I find this a very interesting concept and a misunderstanding of spontaneity.

But there is certainly a lot of spontaneity that can be developed. I don’t know whether that helps.

Question 9: This is a comment. It is a shame we don’t have a fish cognition person speaking, but we have a book that was published last year, Fish Cognition and Behaviour, if people are interested in learning and memory in fishes. Their long-term memory is fabulous, so we need to dispose of the three-second memory theory.

My question is really for Gisela Kaplan, about the whole idea of why magpies might be mimicking. We did some work with lyrebirds – it was based in Edinburgh – and we had the idea that it might actually be niche-filling, so that although they do mimic predators, they don’t just mimic predators. They mimic a whole bunch of different animals that are occupying the area around them. Although another predator may not be fooled by the sound of the same species, perhaps birds have trouble identifying inter-specific calls clearly enough to say that there are those birds present or not. So maybe mimicking lots of different species gives the impression that the niche is filled, that perhaps competition would be high. I wonder if you might want to comment on that idea.

Gisela Kaplan: That is a very interesting suggestion. There have been some studies in South Africa that have shown that mimicking predator vocalisations may aid learning about predators. Unfortunately, there are very different forms of mimicry between the different species and these have not been widely explored. Interestingly, lyrebirds will also mimic car horns and the click of cameras – inanimate objects, in other words. We haven’t found this commonly in other species, but it appears that male lyrebirds add any sound that could add interest and thus help attract a female – anything will do. And they will memorise those songs for the coming season, and will add on to those in the Albert Cycle, like pearls on a string. So the male lyrebird’s repertoire is lengthening over time and, while getting more complex, I don’t think that it has any other function outside that of attracting a female.

But if you consider some other bird species – and Alec Chisholm, in his 1948 publication, suggested that there are many Australian songbird species that may mimic in one way or another – then there may well be a case in some species of niche filling. Niche filling may occur in the sense that the mimicking of sounds may suggest that more birds are occupying a particular patch than is actually the case, and therefore perhaps dissuade others from moving in. That would be a very clever strategy. I don’t think there is a body of knowledge that has been able to show this as a widespread strategy to date. The magpie’s mimicry that I was referring to in my talk is unusual in a sense because it is so very selective and is practised and improved. We have very few studies, let’s be honest, and it is partly so because the evidence is difficult to collect (10 years were needed to collect the evidence in magpies so far). Hence, what we may now think is a very novel and amazing feat may in future be found to be very common.

But we think of mimicry in animals as mindless; if we think of mimicry in humans, we don’t think of it as mindless, we think of mimicry as imitation. I would suggest also to the teacher in Queensland that half of our learning strategies involve imitation learning. We imitate in language, so that we get the intonation right; we imitate even in motor activities. And it has been shown very well in, let’s say, great apes, that juveniles have motor imitation of the adults’ movements – which, in an orang-utan, by the way, is a survival mechanism, because unless they know exactly how and where to put feet and arms they will fall from great heights and die. They have to learn to negotiate their arboreal environment, so that kind of ‘imitation’ or ‘true imitation’, as it is sometimes called, is a vital learning tool. And yet, when we come to vocal mimicry, we sometimes presume that it is mindless. I think we have to be very careful not to throw the baby out with the bathwater in this sense. Obviously, sound imitation is an important part of learning both in humans and in many songbirds that have been investigated. Where the watershed is between mimicry as mindless and mimicry as imitation, as learning and as cognitively complex is the point at which an organism can use that information of the mimicked sequence independently and can apply this in contexts that are novel. I think that these lines of inquiry are just beginning, so yours was an interesting question.