SCIENCE AT THE SHINE DOME canberra 2 - 4 may 2007
Symposium: Development and evolution of higher cognition in animals
Friday, 4 May 2007
Professor Nicola Clayton
Professor of Comparative Cognition, Department of Experimental Psychology and Director of Studies in Natural Sciences (Biological), Clare College Department of Experimental Psychology, University of Cambridge, UK
Nicola Clayton is Professor of Comparative Cognition in the Department of Experimental Psychology at the University of Cambridge and Clare College, Cambridge. She received her undergraduate degree in Zoology at the University of Oxford and her doctorate in animal behaviour at St. Andrews University. Her research interests are in the development and evolution of cognition. Nicky’s work is mainly with members of the crow family (e.g. ravens and jays), and comparisons between the crows and apes including humans. She is on the editorial board of Proceedings of the Royal Society B, Biological Reviews and Public Library of Science One.
Memories of tomorrow: Do animals remember the past and plan for the future?
| According to the mental time travel hypothesis only humans can mentally dissociate themselves from the present, travelling backwards in time to recollect specific past events about what happened where and when (episodic memory) and travelling forwards in time to anticipate future needs (future planning). Studies on the behaviour of food-caching western scrub-jays question this assumption. In terms of retrospective cognition, these birds remember the ‘what, where and when’ of specific past caching episodes, they keep track of how long ago they cached different types of perishable foods that decay at different rates, and also remember whether another individual was present at the time of caching, and if so, which bird was watching when. Recent work demonstrates that the jays also make provision for a future need, caching more food in places in which they will not be given breakfast the next morning than in places where they will be receive breakfast the next morning even though there is plenty of food available to them when they cache the food. Taken together these results challenge the mental time travel hypothesis by showing that some elements of both retrospective and prospective mental time travel appear not to be uniquely human. |
What I want to talk about today is the paradoxical notion of ‘memories of tomorrow’, or specifically the claim that many people have made – and when I say ‘people’ I am referring not just to scientists, philosophers and psychologists but also to many other people (I will give you one example in a moment) – that animals differ from humans in one very fundamental way, other than in language. Animals, it is claimed, are stuck in time, they live in the more-or-less extended present; what they can’t do is go back in time to reminisce about the past. Nor can they think about the future and imagine various kinds of scenarios.
Obviously, if this were true then it would have many blessings, because if you are stuck in time then you have no worries or regrets about the past, nor do you have any anxiety about the future. However, what I am going to argue is that it is not the case that animals are necessarily stuck in the present. I am going to give you evidence from one particular member of the corvid family, the western scrub-jay. This is one member of the crow family which includes the ravens, crows, rooks and jackdaws.
Why do people make this claim? Well, the first quote I want to give you is actually not from a scientist at all but from a poet, Robbie Burns. It is in a poem called ‘Ode to a Mouse’. In this poem, Burns is lamenting the fact that it is dusk and he has just ploughed through a mouse’s nest. He sees this poor little field mouse wander off into the distance, and he feels absolutely wracked with guilt because he realises that, as a consequence of his ploughing over the nest, this mouse is surely going to die. It is not going to make it through the night. Then, riddled with guilt, he turns to the mouse and suddenly an idea occurs to him and he says to the mouse:
The present only touches thee:
But och! I backward cast my e’e
On prospects drear!
An’ forward, tho’ I canna see,
I guess an’ fear!
Robbie Burns, then, believed that mice and other animals were indeed stuck in time, and he also, obviously, appreciated the welfare blessings of being stuck in time, or the welfare implications if this were not true.
(Click on image for a larger version)
This has received scientific endorsement within psychology by Endel Tulving, who is the father of the work on episodic memory in humans, our ability to remember specific events that have happened to us in the past. He argues that this kind of memory, which he calls ‘remembering’ – remembering specific personal past experiences, as opposed to acquiring and storing other kinds of information which you might think of as knowledge of facts about the world – this ability to remember is something that is universally familiar in human beings, or at least in healthy human beings. But he says that it is unique, that although other members of the animal kingdom can learn and benefit from experience, and solve problems, make decisions and so on, they cannot travel back into the past in their own minds.
More recently, this has received endorsement from Thomas Suddendorf and Michael Corballis, who have come up with what they call the ‘mental time travel hypothesis’. Their hypothesis is that this ability to travel in the mind’s eye is indeed something that is unique to humans, and therefore constitutes a discontinuity between humans and other animals.
Before I tell you about some of the work that we have been doing, in which I argue that animals are not stuck in time, I just want to step back for a minute and put it in perspective. I want to say why I think people have said that this is so uniquely human, because those of you who are biologists or naturalists may be thinking, ‘Well, isn’t it obvious? Of course animals remember, and why wouldn’t they think about the future? Why on Earth would this be such a uniquely human thing?’
(Click on image for a larger version)
I think it is because many people who have talked about this ability in humans have defined it based on phenomenological criteria which are necessarily language based, or the way in which we assess these things ourselves is language based. There are three features, really.
The first is that, when we think about the past, it is accompanied by a feeling of re-experiencing a specific past event. It is a bit like rewinding the videotape in your mind, when you go back to imagine what happened, where and when.
The second is that it is accompanied by an awareness that the event happened in the past. If I were to reminisce with friends about what happened last week or three weeks ago or whatever, I would be very aware that what I am talking about is in the past, even though, at the time of talking, I am in the present.
The third feature is a realisation that these are one’s own memories. In other words, you can think of it as you being the author of the memory. We may all have a slightly different take on what has happened today – we would all be the individual authors – even though the actual event that occurred was the same.
It is very difficult to know how to assess whether or not animals have a feeling of re-experiencing the past. I don’t know how to ask my scrub-jays, for example, whether or not they have feelings of re-experiencing, nor whether they are aware of the past or realise that these are one’s own memories. That is, at least, in the absence of any agreed behavioural markers of awareness in non-linguistic beings; it is by talking that we assess this in one another.
(Click on image for a larger version)
When in doubt, in cognition, I like to resort to Gary Larsen cartoons. Here is one which to me depicts what episodic memory is. It reads:
day, remember exactly where they were and what
they were doing when they heard that Bambi’s
mother had been shot.
In trying to find a way to test these abilities in animals, I am trying to come up with a definition based not on phenomenology but on behaviour, in order to make it empirically tractable.
Episodic memory, then, in terms of behavioural criteria, would be recollecting what happened, where and when, on the basis of a single past experience. And in acknowledgment that this may or may not be the same as episodic memories that we talk about in humans, accompanied by these phenomenological criteria, I am going to call this ‘episodic-like’ memory, to mean these behavioural criteria of remembering what happened, where and when.
What about the other side of the coin, the thinking about the future? Well, that might be anticipating what might happen, to inform current decisions – so, the here and now – in order to make a plan for the future. In the Larsen cartoon it might be what they might plan to do to avoid suffering the same fate as poor old Bambi’s mum. That would be the future planning criterion.
Most studies on animal memory have certainly not made this distinction and really looked at the ability of animals to remember what happened, where and when, as opposed to more general information about what happened or where it happened. Very few of them have actually looked at the ‘when’ component of whether the animals could remember things that had happened at different times in the past.
It seemed to me, 11 years ago when I first took up a professorship at Davis, California, and embarked on this project, that it might be quite useful to take an ethological approach and look at the natural behaviour of the animal.
(Click on image for a larger version)
One species that is particularly prevalent there is the scrub-jays, and you can see in green on this map where they live, in the west of the US and going down into Mexico.
One of the things about these birds is that they hide food. This is a behaviour we usually call food caching, based on the French word cacher, to hide. The thing about the jays is that they hide all kinds of different types of food, including various invertebrates – worms and things – that perish, that degrade, over time. It was well established before I started working on them that these birds, and many other food-caching animals, have very accurate, long-lasting memories for where they have hidden their caches, even if it is just based on one trial. Remember that part of the definition of this episodic-like memory is what happened, where and when, on the basis of a single past experience. So it seemed to me that the scrub-jays were a potentially good candidate, because clearly they must be able to remember where their nuts were, in order to have these accurate memories. The question was whether they could also remember which types of food – that could be my ‘what’ – and, using this perishability, how long ago. (The birds love things like worms when they are fresh, and they hate them when they have gone bad.)
We thought we could also look at the prospective component, because the whole point of caching food now is for future consumption. It is for being able to eat it later, ‘when I come back hungry’.
The lucky thing about these birds is that they were all over campus where I was working at the time, but you can easily bring them into captivity and they will readily cache in captivity. This picture shows the sort of set-up that we used, with Sweetie-Pie, one of my prima donna, princess scrub-jays, in the act of caching. We give them food items that they love to eat and will readily hide, and we give them ‘caching trays’, actually an ice cube tray filled with sand or corn kibble. We make each one distinct with Lego Duplo bricks.
[A video was shown, with the following commentary.] What I want to do now is to show you a little video of the caching behaviour. What you are going to see is a bird with a bowl of worms – these are the Belgian truffles of the scrub-jay world, they absolutely adore them – and her action in caching. She has buried a beakful of worms in the sand. She has a quick look around, eats a couple, picks up another beakful and makes a second cache.
So it is very easy to quantify this behaviour and see how many items have been hidden, and where they have been hidden.
(Click on image for a larger version)
In the first experiment, we tried to get at this what-where-when memory by having two kinds of foods: peanuts and worms. One kind, the worms, degrades over time, and the other stays fresh.
To try and show you the basics of the experiment, I am doing this here with plasticine models courtesy of Nathan Emery – a dab hand with the plasticine! The idea behind the experiment is that the birds get to cache peanuts and worms, and on some days they get their trays back four hours later, in which case they can recover the peanuts and the worms while both are still fresh. They love worms, so they tend to have a slight preference for recovering the worms, but they eat peanuts as well.
But after a longer gap between caching and recovery, 124 hours – this is so it is at the same time of day but now five days later plus four hours, as opposed to four hours later on the same day – the worms have now degraded. They are shown here as being black, and the birds don’t like them; they spit them out. So now the birds should recover their peanuts.
What they are being trained in, if you like, is how long it takes for the worms to decay.
We contrast this group of birds with a control group who get to cache their peanuts and worms, just as these ones do, but for the second group – which we call Replenish – the worms never decay.
So the question is what happens if we then suddenly give them a probe trial, a test trial, of memory, in which we let them cache as normal but then remove all the food prior to recovery so they can’t use olfactory cues to detect where the food is. We can look at where they search, and the idea is that if they can remember how long ago they cached their peanuts and worms, and where they hid them, then the birds in the group at the bottom right of this slide, who have learned that worms decay over time, should preferentially search for the worms at the short interval of four hours – but they should actually switch and search instead for peanuts at the long interval, if they can remember that at the long interval the worms would have degraded. Our control group, however, the Replenish group whose worms are always fresh, should continue to search preferentially for worms at both intervals.
(Click on image for a larger version)
And that indeed is what you see. We have here the percentage of birds that search for worms (W) and peanuts (P), and you can see that whereas the Replenish group show a consistent preference to search for worms, the Degrade group do not. They search for worms if there was a short gap between caching and recovery, when the worms would still be fresh, and instead switch to peanuts if the worms would have degraded.
In other words, we are arguing that this demonstrates that the jays remember that specific caching event of caching these worms and peanuts, in terms of what happened, where and when.
Jays, like us, are social creatures. They hang out with lots of other birds. A problem for these birds in being social is that they don’t only hide their own food, but they also steal other birds’ food, and consequently other jays also steal their food. This is sometimes called pilfering, and this thievery is rife among the scrub-jay world.
In a series of experiments we have shown that the birds go to great lengths to protect their caches from being stolen by other birds. We capitalised on that to ask a second question about their episodic-like memory: could they keep track of who was watching?
(Click on image for a larger version)
So in this second experiment – again you see more plasticine models by Nathan Emery – we have two caching events. The caching bird hides in one tray in the presence of a second bird, Observer A, who is in another cage and can have a good look but can’t actually get to the caches. And there is another tray there, labelled arbitrarily Tray B, which is covered over. So Observer A can see the caching bird hiding food in Tray A, and Tray B is there but with no caching going on.
Just before or just after that there is another caching event. Our same caching bird now gets to hide food again, in the other tray, Tray B, in front of a different bird, Observer B. So Observer A and Observer B each watch a different caching event.
(Click on image for a larger version)
Then the question is: what will our caching bird do at recovery? Will it bother to protect its caches?
We have four conditions. Some days the cacher has the opportunity to recover from the trays in front of Observer A, sometimes in front of Observer B, sometimes in front of a naïve individual – I will call that Bird C – who didn’t witness either caching event, and sometimes in private.
(Click on image for a larger version)
What you see is that when they come back to recover their caches in private – with no other birds present – they eat about half of the food they have hidden and they re-cache or re-hide the rest of it. So, yes, they protect their caches.
If one of the observers is present, then they protect the caches that that observer saw them hide. For Observer A, that was Tray A, and for Observer B, it was Tray B. If a naïve bird had been present, who hadn’t witnessed either of the caching events, they would essentially withhold information, re-hiding very little of the food from either tray.
At this point we thought it would make sense that you would re-hide your food if you had been watched and come back in private, because, by definition, if you move the food the observer won’t know the new location.
(Click on image for a larger version)
That seemed to make sense, and indeed they put the items in new sites that by definition the observer doesn’t know about, because they are now in private. And each item is moved just once to that new place.
However, when an observer is present at the time of recovery, when they are re-hiding the food, they do something a bit different. In fact, when an observer is present they are just as likely to re-hide the food in the old cache sites, which the observer has already seen, as opposed to any new ones. And in fact each item is moved, often several times, up to six times.
(Click on image for a larger version)
Just to put this in perspective: in the video you saw a picture a bit like the one on this slide. This is what they do when they are re-caching in private. The worms are fully visible.
(Click on image for a larger version)
(Click on image for a larger version)
But we think that this re-caching when an observer is present is a bit like the Shell Game. It is a way of protecting caches, because it is very difficult to know where the caches end up.
Here you can see a re-enactment of this. Notice that the worms are not visible in the beak, so it is not even clear whether the bird has actually moved a worm from that site or not.
(Click on image for a larger version)
All you can know – if you look carefully you can perhaps see the shadow there – is that there is a bit of a bulge. This is the privacy of the pouch. It is a bit like my handbag: you might think that there is something interesting in there, but you can’t see what it is.
So we actually think that this is a way of protecting their caches, making it much less certain that the observer will know the final destination of the cache. At any rate, I think it shows that the birds do indeed keep track of who was watching, when.
But the whole point of this re-caching is really that it is geared towards the future. After all, there is no benefit to a bird in re-hiding food that it is never ever going to get back in the future. So to what extent might this actually engage planning behaviour? Obviously, there are many future-oriented behaviours which may be nothing to do with planning at all.
(Click on image for a larger version)
In terms of trying to establish behavioural criteria, it seemed to me that there are probably three things that you need to show.
The first is that the behaviour must be flexible, by which I simply mean sensitive to its consequences, so that you can be certain that it is not simply a pre-programmed response.
Secondly, according to the mental time travel hypothesis, the point about future versus current needs is that animals, it is claimed, ‘cannot anticipate future needs…and are therefore bound to a present that is defined by their current motivational state’. So it seemed to me that the second thing was to check that they are actually doing it for a future, not a current time.
The final thing is that it shouldn’t just be because the anticipatory act has been reinforced – in other words, it is not just through learning.
(Click on image for a larger version)
In order to address these issues, we began by capitalising on the fact that if the birds have cached peanuts and worms, and their worms have degraded, then they only bother to recover the peanuts. We gave them a series of trials to see whether, if they only ever got trials in which their worms had degraded, they would learn to stop caching the worms – because even though at the time they are cached they are still fresh, and nothing at the time of caching says, ‘Don’t cache them now,’ if they are sensitive to past recoveries then they should learn, ‘There’s no point caching my worms, even though they’re fresh now, because they won’t be in the future.’ And you can see that yes indeed, they learned quite rapidly to stop caching the worms.
(Click on image for a larger version)
In terms of looking at current versus future motivational needs, we capitalised on something in psychology that is called specific satiety. You can think of it as the dessert principle. It is the idea that having eaten one type of food until you are sated, you don’t want any more of that food but you will quite happily eat a new one, in just the same way as, having had a main meal, most of us find room for a dessert.
The idea is that if you give animals food A to eat, say, and then give them a choice of A and B, they will always preferentially eat the non pre-fed food, food B. Caching birds will also preferentially cache the non pre-fed food. So we had one group of animals who were pre-fed one of the two foods – in the example here it is food A – and then the idea was that they should continue eating and caching the non pre-fed food, in this case food B. I am going to call that the Same Group, because they are pre-fed the same food at caching as they are at recovery.
We contrast this with a Different Group who are pre-fed a different food at recovery from that which they are fed at the time of caching. If they are stuck in time, they should do just the same as the Same Group and continue caching the non pre-fed food. But if they can think about the future, or are sensitive to the future conditions, then they should realise that actually, at the time of recovery, having been pre-fed B, it’s A that they’re going to want. And so the prediction is that, if they can plan for the future, although they should continue eating the non pre-fed food B, they should switch to caching the pre-fed food type A.
(Click on image for a larger version)
Here you see the amount of eating, and you see that, as you would predict for both groups, they preferentially eat the non pre-fed food. And you see that, whilst the Same Group continue to cache the non pre-fed food, importantly the Different Group actually switch to caching the pre-fed food – suggesting that yes indeed, they are doing this for the motivational need for the future, at recovery, rather than at caching.
(Click on image for a larger version)
Finally, putting it all together in terms of forethought, we asked whether they could plan for tomorrow’s breakfast. In this experiment, during training they get no opportunity to cache, so they get no opportunity to learn anything about where are good places to cache and bad places. And that is the reason they are given powdered food: so that they can’t cache it. There is nothing for them to grab hold of and fly off to.
They get training, and during these six days they get to explore three little chambers and eat the powdered food.
After they have gone to bed on some nights, they wake up the next morning and they find themselves in one of the outside chambers. On some days they end up in one compartment, they’re hungry and breakfast is served. On other days, they wake up in the other compartment and, unlucky, they’re hungry but there ain’t any breakfast. They don’t know beforehand which place they are going to find themselves in, in the morning. All they learn is that they get three days on which they get breakfast in one place, and three days when they end up in the other place and there’s no breakfast.
Then, all of a sudden, on the evening of the sixth day, we give them a novel test. Now we give them items that they can cache, and we ask where they do it. The point is that if you can plan for breakfast, then having learned that one room you get breakfast in and one room you don’t, and given that you don’t know which room you are going to end up in, what you should do is put the majority of the items in the no-breakfast room, in case that’s where you find yourself in the morning, when hungry.
(Click on image for a larger version)
And that is what they do.
(Click on image for a larger version)
Finally, we asked not only whether they would be sensitive to which room serves breakfast but whether they could also plan what and where to cache for breakfast, basing the ideas on the specific satiety, or dessert, principle effects.
Now, rather than having breakfast in one room and no breakfast in the other, they get different breakfasts. On some days they end up in one room and they get powdered peanuts for breakfast; on other days they end up in the other room and they get powdered kibble. They don’t know which room they’re going to end up in. So, on test – I’m sure you’ve guessed it – there are cacheable peanuts and there are kibbles, and the question is where they hide the food.
If they can plan, then not knowing whether they will end up in the peanut room or the kibble room they should put food in both, but they should put predominantly kibbles in the peanut room and peanuts in the kibble room. After all, in the kibble room they will get as much in kibbles as they want in the morning, so what they want is the other food.
(Click on image for a larger version)
And that is what they do. In both rooms you see that they place more of the different than the same food.
(Click on image for a larger version)
Taking these things together, then, I am arguing that they are sensitive to the past, the present and also the future. They remember the what, where and when of the specific caching episodes; they will cache for a future motivational need, independent of their current needs; and they can anticipate future conditions on the basis of past recoveries, from protecting their caches from potential pilfering thieves to planning for breakfast.
(Click on image for a larger version)
There is a famous song by Robyn Hitchcock that says, ‘If you want to know what the future holds, the black crow knows.’ Well, I’m not sure about the black crows – the jury is out on that one at the moment – but I would certainly argue that the scrub-jays do.
(Click on image for a larger version)
It just remains for me to thank the people in the lab who have done these experiments with me, particularly Tony Dickinson and Nathan Emery, and our students Sérgio Correia, Joanna Dally, Caroline Raby and Dean Alexis, and to thank the birds, of course, and the funding bodies who have so kindly funded us.
Discussion
Question 1: The caching provides an obvious opportunity to test these kinds of ideas about mental time travel, but to what extent do you think that the fact that these birds in particular can do this might be a consequence of the fact that they are birds that do cache? Is being able to do this the kind of thing that that kind of ecology encourages?
Nicola Clayton: I guess there are two ways of assessing that issue. One is: is it domain specific? In other words, can caching birds only do it with respect to caching? I am afraid I don’t know the answer to that question. Part of the problem, particularly with the planning stuff, is that it is very difficult to think of a well-controlled, simple but easy way to do it without caching. We thought maybe you could ask how they plan for where to sleep; the problem with that is that you would have a very simple learning explanation for it, because they would simply have to choose the room associated with food.
The other way of looking at it would be to ask whether species that cache food are special, and compare cachers and non-cachers. And again I would love to do that but at the moment what we are grappling with is trying to think of a really nice behaviour we could use that isn’t caching, and then, hopefully, we could tackle both aspects of the same coin.
Question 2: Nicky, I am sure you know this experiment, but there are some people at Oxford who looked at great tits, giving them either continuous food or unpredictable food. And the great tits that got the unpredictable food put on fat, whereas the others didn’t. The question is this: could your jays be, in fact, using a conditional rule of this kind? Would there be a difference between what the great tits do in Oxford and what your jays are doing in Cambridge?
Nicola Clayton: There are a couple of ways to answer that. People have followed up the great tit experiments, and asked what happens if you do that with cachers by providing them with predictable and unpredictable foods. You find that the more unpredictable the environment, the more they cache. So one thing cachers do, rather than just putting it on as body fat, is to put some of their fat reserves elsewhere – which, obviously, has other advantages for them.
In terms of whether it is a conditional rule or not, I suppose what we have tried to do in all these experiments is to have just one opportunity for learning, so that you would have to say either that it isn’t a conditioned response or that the learning occurs very rapidly. But I don’t think you could rule out the latter, except in the case of the planning for breakfast. There it is a novel test. You could have a conditioned association about where the food is in the place but the actual decision of the caching would have to be separate, because it is untrained.
In terms of the social work, there it is very easy to have conditioned rules about who has been a thief in the past or not, and there it just depends on how complex your conditioned rules have to get.
Question 3: I am thinking of a study somewhere in Canada on eastern coastal crows who had learned to collect mussels and shells at low tide, cache them and then consume them at high tide. The same species further inland didn’t ever show that kind of behaviour. The person who did the research argued that they were obviously capable of looking into the future and undertaking some future planning because all their food sources would be hidden.
Would there not be a simpler explanation for that as a spontaneous adaptation? You see those, almost like a stimulus-response thing, and you collect as many as you can, and then obviously you are suddenly trapped, and through the site of learning you know when it is exposed you need to cache them. But how do you prove or show that in such a thing there is an understanding of a future event, or planning, rather than a spontaneous response to say that when there is much food you have got to collect it?
Nicola Clayton: I think that with field observations it is very, very difficult, because you just don’t have the level of control that you would need, at least to satisfy the criteria for most psychologists. Different types of audience have different kinds of criteria for what will and will not make them satisfied scientifically, but I am just going to deal with the psychological one.
I think there are some things you have to be able to rule out. As I said with my three points, you really need to be able to show that such a behaviour is flexible – that is the first one, that it isn’t just triggered by a stimulus, no matter what.
Secondly, it has to be for a future need, independent of the current one. So the analogy that Thomas Suddendorf always makes is to say that a sated lion is no threat to nearby zebras, whereas a sated human is. The idea is that the lion, being sated, doesn’t want any food and doesn’t appreciate that it will be hungry in the future, unlike the human. (Of course, that is assuming that it is a carnivorous human who wants to eat zebras. I wouldn’t be any threat to zebras either, because I am vegetarian, but we will leave that aside!)
Thirdly, you have to show that it is not simply through extensive training of the anticipatory act.
Any time you see it in the wild, you would never know how you could rule those things out. I think that is the difficulty with it.
Question 4: In experiments with honeybees, we showed that the honeybee can learn what to do and when. This has been published. In the second project we found that the honeybee knows what to do, where to go, and when. The problem is that the time zoning, the temporal aspect, for biological reasons is 24 hours – the morning, the afternoon and the next day or the day after. So is this a sort of episodic memory?
Nicola Clayton: It is an interesting debate, isn’t it, because it is really about what you mean by ‘when’, and whether you are willing to have a circadian rhythm supply that ‘when’ information. Most psychologists would say, I think, that that is not good enough – you need to be able to show independence of circadian rhythm. For example, there has been a debate in the literature over the past couple of years about whether or not rats could remember what, where and when. There were some very nice experiments done by Stephanie Babb, in Jon Crystal’s lab, showing that rats could remember which type of food was available where, and how long ago. But theirs was also with a circadian rhythm, and therefore many people criticised it, saying, ‘This isn’t a “when”. They simply have this biological rhythm that is providing the “when”.’
Then they redid the experiment, controlling for time of day – which is, incidentally, why I used those four and 124 hours. For me it was to get rid of time of day as a possible explanation for the results. In the case of the rats, they found that they could continue to do it, and therefore they called it episodic-like memory.
So I think that in the honeybees if you could show remembering on the basis of a single past experience, by providing a ‘when’ that was a relative length of time and couldn’t be cued by time of day, then you could say that that was episodic-like memory.
Question 5: This comes at perhaps 90° but it might be useful. Could I ask you first: did you offer the jays peanuts in or out of the shell? There is a reason for my asking this.
Nicola Clayton: In those experiments they were out of the shell.
Question 5 (cont.): Just quickly: as a small boy I used to do experiments with these little creatures with peanuts in the shell, and learned something that I thought at the time was interesting. I have only now, thanks to your talk, realised why it was. They were able, once habituated in my household, to select peanuts in the shell that had one, two or sometimes three peanuts in the shell. I watched as they would uniformly select three of the three-peanut shell individuals, if they were available, and cart them off, and then come back and go to the number twos and ones. They were very predictable.
I wasn’t entirely satisfied with that, so I selected two- and three-peanut shells and weighed them, and put out more than they could carry, and watched as they picked them up and dropped them again and then finally selected the ones they wanted. This suggested that not only can they think back and forward, they can do sums in their heads pretty well! They carted off the heavier ones, maximising their food – for what that’s worth.
Nicola Clayton: That’s wonderful. Thank you for sharing that with me. They do have experience with the shelled peanuts; we have just never done it in a test. But the one with number is particularly interesting for me because, although we have not done it yet, one of the next kinds of experiments we wanted to do was actually to see if they are sensitive to the number of items that they have cached. So if they come back and they have put five worms in a place, and suddenly they discover there are only four there, are they bothered? And what happens if they suddenly find 10 there? Are they surprised? Are there ways you could look at that? I think it is a very interesting question.
