04 November 2008
NewScientist.com news service
ROVER stares up at you, apparently yearning for a pat on the head. But behind those endearing doggy eyes, his mind is whirring, recalling the excitement of yesterday's walk: long grass, a sloping field falling away into woodland and a rabbit running for safety. With a gentle wag of his tail, he smiles at the memory, then wonders if he's got time for a nap before today's outing.
An animal couldn't do that, could it? For most of us, the instinctive answer is no: the ability to reconstruct past events in our mind and imagine ourselves in future scenarios is something we can do, but other creatures cannot. And yet, while there is no evidence that dogs can think in this way, recent studies have suggested that some birds, great apes and a few species in between are not "stuck in the present" after all, but may be capable of thinking backwards and forwards in time.
The issue is getting researchers of human and animal minds rather hot under the collar. Critics argue that what looks like memory or forward thinking is nothing more than instinct or learned behaviour, and insist that there is no convincing evidence that non-human animals can remember their past or contemplate the future. As a result, every paper claiming to demonstrate the ability in animals is fiercely debated.
The idea builds on the work of renowned neuroscientist Endel Tulving, who pioneered the study of human episodic memory - the recall of our autobiographical past. Tulving wrote in 1983 of how the human mind is capable of travelling through time, but it wasn't until 1997 that the idea started to receive serious attention. That was when Thomas Suddendorf and Michael Corballis at the University of Auckland in New Zealand coined the phrase "mental time travel", and argued that it could be one of the key cognitive features that distinguishes humans from the rest of the animal kingdom.
"People typically speak about language as the thing that sets us apart," says Suddendorf, now at the University of Queensland in Brisbane, Australia. "But the content for communication is, to a large extent, provided by our ability to project ourselves backwards and forwards in time," he says.
Many researchers working on animal cognition, however, believe that some species can indeed remember their past and plan for the future. Proving that this is the case is notoriously difficult. In studies of humans, memories and thoughts about the future are measured by asking the volunteer to verbalise what they are thinking or what motivated a decision. Animals, of course, cannot do the same, which makes it difficult to separate the repetition of a learned behaviour from evidence of true memory or planning.
A key question is whether animals can recall experiencing an event at a specific time and place in a manner similar to human episodic memory. This is a major bone of contention in the debate. As Tulving saw it, human episodic memory requires self-awareness - the ability to imagine oneself in the past, as opposed to merely remembering what happened when and where. By this definition, the existence of animal episodic memory is virtually impossible to prove. Comparative psychologists who research animal memory, however, work on the basis of "episodic-like memory". This requires only that the animal can remember what it did, where, and when.
Several studies have shown this to varying degrees. Animal trainers working with a pair of bottlenose dolphins, for example, have found that they are able to remember what they did in the immediate past. After being trained to perform dozens of different tricks in response to specific hand signals, the dolphins were asked, using another hand signal, to repeat the trick they last performed. Both dolphins could do this easily - one, a female called Elele, had a 100 per cent success rate. Similarly, experiments with pigeons and rats have shown that they are able to peck or nose at a particular symbol to indicate what behaviour they have just carried out.
In 1999, primatologist Emil Menzel from Stony Brook University in New York demonstrated even more impressive memory skills with Panzee, a chimp who had been taught to communicate with researchers by pointing to symbols. In experiments Menzel hid food around Panzee's enclosure as the chimp watched. Up to 16 hours later she was able to guide her keepers to the hidden treats using symbols and gestures, even though they had no idea what had been hidden or where.
Western scrub jays, a member of the crow family with a particular talent for cacheing and retrieving food, go one better. In 1998 Nicola Clayton and Anthony Dickinson at the University of Cambridge demonstrated that these birds can remember not only where they have cached a tasty morsel, but also what they have cached and when. If, for example, a wax worm - the caviar of the scrub jay world - remains buried for too long, it begins to rot. The birds seem to know this and do not bother recovering worms likely to have passed their use-by date (Nature, vol 395, p 272).
Yet because there is no way of asking these species how they come to such decisions, it is not clear that they are reliving their past. While the jays clearly had some knowledge of what they had experienced and where, it is possible that they acquired this not by going on a mental journey through time, but by keeping track of how long ago they buried the food. The same can be said for Panzee's skills, says Suddendorf. "I may know where my car keys probably are without necessarily remembering having put them there," he says.
There are similar problems with proving that animals can think about their future. In 2006, Nicholas Mulcahy and Josep Call of the Max Planck Institute for Evolutionary Anthropology (MPIEA) in Leipzig, Germany, tested the foresight of two of our closest living relatives - bonobos and orang-utans (Science, vol 312, p 1038). First, they taught the apes to use a tool, and then carry it into a "reward room" where they could use it to obtain a reward of grape juice.
Once the apes were trained, the team shut off access to the reward room and gave them the choice of eight tools - two suitable for obtaining the reward and six unsuitable ones - before ushering them empty-handed into a waiting room, from where the animals could see all the tools being taken away. An hour of so later, they were let back in and the door opened to the reward room once more, though of course they had no tool to enable them to obtain the coveted grape juice.
In subsequent trials, many of them solved the problem. They apparently realised that, while they had the opportunity, they should grab a suitable tool and take it into the waiting room in order that they could use it when the reward room was opened.
"This demonstrates that apes can select, transport and save a tool not for immediate use, but to use it in the near future," says Call, who is director of the Wolfgang Köhler Primate Research Centre at MPIEA. However, though this suggests that non-human primates might be capable of foresight, the extent of their cognitive talents is still up for debate.
The feats of Call's primates are impressive, says Suddendorf, but they still fall short of real mental time travel. He believes that what looks like planning could actually be glorified classical conditioning, driven by a thirst for fruit juice at the time of the experiment and by having recently learned to use a tool for a specific purpose.
This argument stems from the so-called "Bischof-Köhler hypothesis", based on the work of Norbert Bischof and Doris Bischof-Köhler of the Ludwig Maximilian University in Munich, Germany. In the late 1970s they suggested that while humans are capable of using their experience to think about the future - planning a meal when not hungry, for example - animals are only able to act based on their current motivational state.
One possible counter-example comes from work done in Ivory Coast in the 1980s by primatologists Christophe Boesch and Hedwige Boesch. They observed chimpanzees who carried around a large stone as they went in search of nuts, then used it to crack them open. Though some believe this to be an example of animal foresight, it runs into the same problems as Call's orang-utans. "This is quite a remarkable form of future anticipation, but it is guided by their present appetite for nuts," says Bischof. Once full, the chimps cast the stones aside, seemingly unaware that they might need them in the future. If they were able to envisage being hungry again, we might expect them to take more care of these scarce and therefore valuable stones, he says.
In response to these criticisms, several researchers have designed experiments to explicitly test the Bischof-Köhler hypothesis. Some say they have produced evidence that, in certain situations, some animals are indeed able to act on future needs and not just immediate ones.
In one such experiment, Clayton and her colleagues built on their work with scrub jay memory by investigating the bird's talent for thinking ahead (Nature, vol 445, p 919). They gave scrub jays access to two different chambers and trained them to expect that on some - but not all - mornings they would get a breakfast of pine nuts in a specific chamber. The other chamber never contained anything. Then, one evening, the researchers opened up a third area containing an abundance of pine nuts. The jays rapidly set about moving them into the chamber that they had learned was always empty.
Whilst this seemed like good evidence of forward planning, it was still possible that the animals were hungry at the time of the experiment, so a current motivational state of hunger could not be ruled out. To address this uncertainty, the same team designed another experiment.
If tomorrow comes
This time they gave the birds a hearty meal of either pine nuts or kibble before giving them the opportunity to cache from a mixture of the two foods. They found that birds that had gorged on pine nuts showed a clear preference for eating and cacheing kibble, and vice versa, showing that the birds will reliably choose something other than what they have just eaten. Next, the team trained the birds that had earlier gorged on one food, for example kibble, to expect a meal of the opposite, in this case pine nuts, just before being allowed to recover their caches (see diagram). After a number of repetitions of this experimental set-up, birds which were fed kibble, then allowed to cache, then fed pine nuts and then allowed to retrieve their caches began to cache kibble, the food that would be of most value after their meal of pine nuts.
For Clayton and her colleagues, here was a serious challenge to the Bischof-Köhler hypothesis (Current Biology, vol 17, p 856). Their birds seemed to be able to anticipate, act on and satisfy a future motivational need.
There are other examples that challenge the Bischof-Köhler hypothesis. William Roberts and Miriam Naqshbandi, both at the University of Western Ontario in London, Canada, gave squirrel monkeys a choice of one or four dates - one of their favourite fruits. Predictably, the animals almost always chose four. Next, the researchers took the water bottles away from the monkeys just before they made their choice, returning them after just half an hour if the animals selected one date and denying them water for 3 hours if they had plumped for four. The animals soon began to show a preference for one rather than four dates (Journal of Comparative Psychology, vol 120, p 345). Roberts reckons that this suggests the animals can anticipate that they will be thirsty in the future despite not being thirsty at the time they made the choice.
For Suddendorf, however, no experiment has yet managed to refute the Bischof-Köhler hypothesis. The most likely explanation of the squirrel monkey behaviour, he suggests, is that they learned to associate taking four dates with the discomfort of thirst. This, he says, is not the same as foresight.
He is also underwhelmed by the scrub jay findings. "Humans are virtually unlimited in the kind of things we can remember and plan for. There's no evidence yet that the scrub jays do anything but the cacheing and retrieval of food." And they are not even very good at that, he says. "Consider that, in the laboratory setting, there is in fact no point in cacheing food, given that humans feed the birds," he and Corballis wrote in a critique of the scrub jay research.
Clayton calls this criticism "ridiculous" and defends her experiments in a recent article in Animal Behaviour. She points out that even in captivity, the jays need to cache food not only because no one may come to feed them tomorrow, but also to prevent it being stolen by other birds. "Given the future is never certain, a bit of insurance is always a good thing," says Clayton. Plus, it's always worth burying treats like wax worms. "I often cache my after-dinner mints, even if I don't want them at the time," she says.
Clayton and Suddendorf can, at least, agree on one thing - that no one has yet been able to demonstrate that these and other animals are creating mental pictures of their past or future. But, as Clayton notes, "you can't use absence of evidence as evidence of absence".
It remains to be seen whether anyone will find a way to produce this kind of evidence, or whether the two sides can agree on what would constitute mental time travel in animals. But as an ever-increasing number of studies delve into the minds of animals, we can expect to find out a whole lot more about how they understand their worlds. This could have far-reaching consequences for the way we think about - and treat - them. After all, they may have plans of their own.
Henry Nicholls is a freelance science journalist based in London
From issue 2680 of New Scientist magazine, 04 November 2008, page 32-35