Ants Are First Non-Humans to Teach, Study Says
Monday, January 16, 2006
The ants are tiny and usually nest between rocks in the south coast of England. Transformed into research subjects at the University of Bristol, they raced along a tabletop foraging for food -- and then, remarkably, returned to guide others.
Time and again, followers trailed behind leaders, darting this way and that along the route, presumably to memorize landmarks. Once a follower got its bearings, it tapped the leader with its antennae, prompting the lesson to literally proceed to the next step.
The ants were only looking for food, but the researchers said the careful way the leaders led followers -- thereby turning them into leaders in their own right -- marked the Temnothorax albipennis ant as the very first example of a non-human animal exhibiting teaching behavior.
"Within the field of animal behavior, we would say an animal is a teacher if it modifies behavior in the presence of another, at cost to itself, so another individual can learn more quickly," said Nigel R. Franks, professor of animal behavior and ecology, whose paper on the ant educators was published last week in the journal Nature.
But defining even common behaviors such as teaching is complex, and it is even harder to understand what is happening in the brains of other animals. So it is no surprise that the paper has sparked debate over what constitutes learning and teaching in the non-human world.
Franks said careful analysis, and a great many hours of videotape, proved that the ants were teaching one another. For example, if the leaders were to race to the food on their own, he said, they would reach it four times as fast than if they had a follower tagging along. Even grabbing a follower by the mouth and physically lugging it to the food, as the ants are sometimes prone to do, was three times as fast as the teaching exercise. But the ants persisted with the tutorial, Franks said, presumably because followers that were carried were trucked with their heads turned upside down facing backward -- hardly the best vantage point from which to master a new route.
The ants appeared to follow pedagogical techniques that good human teachers have used for centuries. The lesson was highly interactive and proceeded at a pace set by the followers. If the gap between leader and follower increased too much, the leader slowed down. If it was too close, the leader accelerated.
"Tandem running is an example of teaching, to our knowledge the first in a non-human animal, that involves bidirectional feedback between teacher and pupil," wrote Franks and graduate student Tom Richardson, who spent countless hours poring over videotape.
No sooner was the paper published, of course, than another educator (this one at Harvard) pooh-poohed it. Marc D. Hauser, a psychologist and biologist and one of the scientists who came up with the definition of teaching, said it was unclear whether the ants had learned a new skill or merely acquired new information. Mere communication of information is commonplace in the animal world, Hauser noted. Consider a species, for example, that uses alarm calls to warn fellow members about the presence of a predator. Sounding the alarm can be costly, because the animal may draw the attention of the predator to itself. But it allows others to flee to safety.
"Would you call this teaching?" Hauser asked in an e-mail. "The caller incurs a cost. The naive animals gain a benefit and new knowledge that better enables them to learn about the predator's location than if the caller had not called. This happens throughout the animal kingdom, but we don't call it teaching, even though it is clearly transfer of information."
Hauser cited the work of another scientist, Tim Caro, who found that cheetah mothers that take their cubs along on hunts gradually allow their cubs to do more of the hunting -- going, for example, from killing a gazelle and allowing young cubs to eat to merely tripping the gazelle and letting the cubs finish it off. At one level, Hauser said, such behavior might be called teaching -- except the mother was not really teaching the cubs to hunt but merely facilitating various stages of learning.
Psychologists study animal behavior in part to understand the evolutionary roots of human behavior, Hauser said. The challenge in understanding whether other animals truly teach one another, he added, is that human teaching involves a "theory of mind" -- teachers are aware that students don't know something. He questioned whether Franks's leader ants really knew that the follower ants were ignorant.
Could they simply have been following an instinctive rule to proceed when the followers tapped them on the legs or abdomen? And did leaders that led the way to food -- only to find that it had been removed by the experimenter -- incur the wrath of followers?
That, Hauser said, would suggest that the follower ant actually knew the leader was more knowledgeable and not merely following an instinctive routine itself.
Franks responded by saying that the two-way communication between the ants was quite different than merely sounding an alarm about a predator. And, he added, the follower ant often did not use the same direct route on its return trip. Once led to food, ants found new paths back to the nest, Franks and Richardson found, and those paths were sometimes more direct than the route that leaders had shown them.
In other words, Franks said, the teaching appeared to give follower ants more than just information; it generally increased their knowledge of the foraging environment.
Bennett G. Galef Jr., a psychologist who studies animal behavior and social learning at McMaster University in Canada, sided with Franks. He said ants were unlikely to have a "theory of mind" -- meaning that leaders and followers may well have been following instinctive routines that were not based on an understanding of what was happening in another ant's brain.
He warned that scientists may be barking up the wrong tree when they look not only for examples of humanlike behavior among other animals but humanlike thinking that underlies such behavior. Animals may behave in ways similar to humans without a similar cognitive system, he said, so the behavior is not necessarily a good guide into how humans came to think the way they do.
"You can get the same functional product without the same underlying mental states," Galef said. "That serves as a caveat to us all."