A Different Kind of Smart

Zola the crow is about to face a test that has baffled animals from canaries to dogs.

She’s a wild New Caledonian crow, and for the first time, she’s seeing a tidbit of meat dangling on a long string tied to a stick. She perches on the stick, bends down, grabs the string with her beak and pulls. But the string is too long. The meat still hangs out of reach.

In similar tests, dogs, pigeons and many other species routinely falter. Some nibble at the string or keep tugging and dropping the same segment. Some pull at a string that’s not connected to food just as readily as a string that is. Eventually many get the hang of reeling in the tidbit, but they seem to learn by trial and error.

Zola, however, does not fumble. On her first attempt, she anchors the first length of string by stepping on it and immediately bends down again for the next segment. With several more pulls and steps, Zola reels in the treat.

Watching the crow, says Russell Gray, one of the researchers behind the string-pulling experiment, “people say, ‘Wow, it had a flash of insight.’ ” At first glance it seems Zola mentally worked through the problem as a human might, devising a solution in an aha moment.

But Gray, of the University of Auckland in New Zealand, has had enough of such supposed animal geniuses. Asking whether the crow solves problems in the same way a human would isn’t a useful question, he says. He warns of a roller coaster that scientists and animal lovers alike can get stuck on: first getting excited and romanticizing a clever animal’s accomplishments, then crashing into disappointment when some killjoy comes up with a mundane explanation that’s not humanlike at all.

Gray is looking for a way to get off the roller coaster. In Zola’s case, he and his colleagues devised several different variations on the string-pulling test that would never trip up a human, and the crow’s smooth performances fell apart. Whatever Zola was doing to solve the puzzle, Gray says, it’s not full, humanlike insight.

That may disappoint some people, but not Gray. “Often we learn the most when we see what we can change that makes the apparently impressive performance collapse,” he says.

He and a handful of other researchers are studying not only what animals can do, but also what they can’t. Forget the animal Einsteins — give Gray the not-so-miraculous beasts that ace one version of a test but flunk another.

After all, seeing an animal succeed at a mental challenge reveals little about how it evolved that capacity. Evolution doesn’t proceed by astonishing leaps, but by baby steps. “I’m interested in halfway scenarios, intermediate scenarios,” Gray says. These modest capabilities, he argues, offer the richest inspiration for understanding the small steps that build up into the rich diversity of animals’ mental powers.

Clever creatures

That’s not to say that scientists haven’t been looking for signs that animals have humanlike thought processes. Recent decades have seen a flood of reports that animals share some degree of capabilities once assumed to be uniquely human. Recently hatched chicks manage simple addition and subtraction, correctly keeping track of which of two hidden groups of familiar objects is larger. Foraging rock ants look as if they’re among the very few animals to show true teaching behavior. Sheep have sophisticated powers of facial recognition and can remember 50 of their fellows for two years. Black bears can learn to sort images into categories, such as bears versus humans. Dolphins can use tools, carrying sponges that protect their sensitive snouts while foraging.

None other than Charles Darwin noted many examples of humanlike cleverness in animals, which he celebrated as support for evolution’s tenet of shared deep ancestry, says Sara Shettleworth of the University of Toronto.

The unintended result of Darwin’s remarks was such uncritical enthusiasm for anecdotes about clever animals, however, that a backlash struck as early as 1894. That year, British psychologist C. Lloyd Morgan published what’s called Morgan’s canon, the principle that suggestions of humanlike mental processes behind an animal’s behavior should be rejected if a simpler explanation will do.

Still, people seem to maintain certain expectations, especially when it comes to birds and mammals. “We somehow want to prove they are as ‘smart’ as people,” Shettleworth says. We want a bird that masters a vexing string to be employing human-style insight.

Aha moments

New Caledonian crows face the high end of these expectations, as possibly the second-best toolmakers on the planet.

Their tools are hooked sticks or strips made from spike-edged leaves, and they use them in the wild to winkle grubs out of crevices. Gray first saw the process on a cold morning in a mountain forest in New Caledonia, an island chain east of Australia. Over the course of days, he and crow researcher Gavin Hunt had gotten wild crows used to finding meat tidbits in holes in a log. Once the birds were checking the log reliably, the researchers placed a spiky tropical pandanus plant beside the log and hid behind a blind.

A crow arrived. It hopped onto the pandanus plant, grabbed the spiked edge of one of the long straplike leaves and began a series of ripping motions. Instead of just tearing away one long strip, the bird ripped and nipped in a sequence to create a slanting stair-step edge on a leaf segment with a narrow point and a wide base. The process took only seconds. Then the bird dipped the narrow end of its leaf strip into a hole in the log, fished up the meat with the leaf-edge spikes, swallowed its prize and flew off.

“That was my ‘oh wow’ moment,” Gray says. After the crow had vanished, he picked up the tool the bird had left behind. “I had a go, and I couldn’t do it,” he recalls. Fishing the meat out was tricky. It turned out that Gray was moving the leaf shard too forcefully instead of gently stroking the spines against the treat.

The crows’ deft physical manipulation was what inspired Gray and Auckland colleague Alex Taylor to test Zola and other wild crows to see if they employed the seemingly insightful string-pulling solutions that some ravens, kea parrots and other brainiac birds are known to employ. Three of four crows passed that test on the first try, so next the researchers set out to test the crows’ limits.

Gray and Taylor set up a platform instead of a perch, which limited what the crows could see while pulling the string. The birds could investigate the string and the dangling meat from the sides but had to hop onto the platform and pull the string up through a slot. The supposedly insightful toolmakers had a terrible time. Out of four birds that had never confronted a dangling tidbit, only one hauled in the treat, and that was on the fifth try. Another bird failed at 10 opportunities, pulling at the string 188 times but never stepping on it.

In another test, researchers laid the string on a table in S-curve loops. The birds could see the meat, but they wouldn’t see it moving closer until they’d pulled enough times to reach the last segment of string.

An animal with true insight, in theory, would recognize that continuing to pull the string would eventually pull in the meat. But in this setup, the birds “completely fail,” Gray says. Some gave the string a tug at first, but only one kept hauling. And that bird was just as happy to pull on a string not connected to meat as on one that was, Gray, Taylor and their colleagues reported in the Proceedings of the Royal Society B in a 2012 paper titled “An end to insight?”

After seeing all this, the researchers proposed that Zola and the other crows had solved the first test — the perch with the hanging string — not by insight in the human sense, but through an enhanced ability to pay attention. New Caledonian crows, which do have relatively large brains for their body size, may be able to notice and absorb in detail the consequences of what they’re doing. Reaching down to grab the dangling string isn’t a big change from normal poking and exploring. And when the meat rises a bit, the birds absorb the positive feedback and take another step-pull.

As mental prowess goes, Gray says, “that’s not a miracle, just a small tweak in cognition.”

Physics test

Researchers have done a similar kind of tweaking using experiments based on Aesop’s fables. In one of the old tales, a thirsty crow finds a jug partly full of water but can’t reach down far enough for a drink. So the bird plops stones into the jug until the water level rises.

Nathan Emery and Chris Bird of Queen Mary, University of London taught real-life rooks a version of this trick. The researchers gave birds a tube partly filled with water and a waxworm bobbing on the water’s surface. The rooks readily dropped stones into the water until they could grab the treat. (Orangutans in lab tests have solved the problem in their own way, taking mouthfuls of water from their drinking supply and spitting into the tube to raise the water level.)

To see if the behavior extended to a related group of birds, Nicola Clayton and her colleagues tested Eurasian jays. “The birds often walk around the tubes having a good look first,” says Clayton, who studies the evolution of animal cognition at the University of Cambridge in England. Soon two of five jays began to drop stones into the water to score a waxworm. Those two also learned a preference for dropping in pieces of rubber that sink instead of foam chunks that float uselessly on the surface. “This is especially striking because when we tested children, the children don’t pass this version of the task until quite late in development,” Clayton says. One 5-year-old grasped the value of sinking objects, but overall the successful children averaged more than 8 years of age.

Then researchers devised a counterintuitive set-up, offering three tubes partly filled with water. The treat floated in the middle one, but that tube was too narrow for a stone. The only way a jay or child could score the treat (kids got tokens to exchange for stickers instead of waxworms) was to drop stones into one of the outer tubes that had a hidden connection to the narrow middle tube. Jays just didn’t get it, but a substantial number of 8- to 10-year-olds did, although “most of them didn’t understand why the setup worked,” Clayton says. “They attributed it to magic.”

Neither the jays nor the kids managed the trickiest tasks the way an adult human would. But like Gray, Clayton is intrigued by the partial successes. In the last test, she speculates, children may be better able than jays to accept the counterintuitive quirk of the secretly connected tube. “Without a belief in magic,” she says, “jays fail to figure it out.”

Reading minds

Besides studying how birds solve physical problems, Clayton has tested the notion that a bird can imagine, in some sense, what’s going on in another animal’s head. People have this ability, called theory of mind, but proposing, as Clayton does, that the Western scrub-jay can infer what another bird is thinking is a striking conclusion.

Scrub-jays cache food, and possess prodigious powers for remembering where. They also steal from each others’ caches, with higher-ranking birds tending to steal from lower-ranked birds. Clayton has found that if a bird with a larcenous past knows it’s being watched as it stashes a tidbit, it’s likely to later shift the cache to an unobserved location. This suggests that the birds have something like a theory of mind, Clayton says, because they understand that the bird watching them may come steal their stores. Yet nonthieves aren’t as likely to recache the food. So jays that steal may project their own behavior onto other birds that are watching them.

Elske van der Vaart of the University of Amsterdam has been looking for a simpler explanation. Maybe the birds are not relying on something even close to a human’s theory of mind, she and her colleagues suggested in PLOS ONE in 2012. Maybe all the hiding and rehiding is just a side-effect of something as simple as stress. Being watched is stressful, the researchers say, as is failing to find a cache. In experiments with virtual birds in a computer simulation, flustered individuals that were being watched and following  a simple rule (they cached as far away as possible from observers) hid and rehid their hoard much as real scrub-jays do.

But real birds don’t behave like simulated birds, Clayton and Cambridge colleague James Thom reported in January in PLOS ONE. Given a chance to hide peanuts in ice cube trays, birds cached about the same number of treats in both more and less stressful conditions. “Sometimes the simplest explanation is not the best,” Clayton says.

So the debate about theory of mind continues. Van der Vaart says the supposedly serene ice cube–tray situations might have held hidden stresses that confounded the results. And other predictions from computer simulations still need testing. “I certainly do think it’s possible that [crows and related species] could have something like a theory of mind, and it would be very exciting if they did,” she says. “But right now, I don’t think we know enough to be able to say one way or the other.”

Flub factor

Research on elephant insight and chimps’ understanding of the physical world has approached the question of limits from the other direction, with scientists tweaking tests that animals normally flub to discover what specific factors let them improve.

Chimps may not understand how the physical world works well enough to attribute phenomena to underlying causes such as gravity, solidity and other such qualities — or at least that’s been a long-standing proposal. Amanda Seed of the University of St. Andrews in Scotland isn’t so sure. “The difference between human and ape folk physics may not be as clear-cut as that,” she says.

In a classic lab test, chimps seem not to grasp basic cause and effect. To coax a treat out of a device called a trap tube, an animal has to use a stick to poke the treat to one end. If nudged in the wrong direction, the food tumbles irretrievably into a hole. “Chimps appeared not very good at telling that their food would fall into a trap,” Seed says.

In one of Seed’s first chimp experiments, she redesigned the trap so the chimps could poke with their fingers instead of a stick. In a video of the new experiment, a chimp stands in front of a clear plastic box and without much ado, pokes a finger through a series of little holes, working a tidbit along in the correct direction and safely out of the box.

The problem may have been the same one faced by floundering human pool players who spend too much time watching the cue instead of the ball. Chimps may not have a different conception of surfaces and holes than people do, but rather a different capacity to focus attention or remember. Without the tool to distract them, they may absorb more of what poking around in the tube is actually doing.

Revealing such hidden animal talents requires devising the right kind of test, which often takes a bit of ingenuity. Preston Foerder of the University of Tennessee at Chattanooga was testing for insight in elephants, for instance, by putting food out of reach and providing a stick as a tool for getting it. Working with three elephants at the Smithsonian National Zoological Park in Washington, D.C., Foerder found that they readily picked up a stick. But instead of pointing it at hard-to-reach food, they banged the walls, scratched themselves and threw it around. “This was about three months of research off and on, and I was commuting from New York City to do it,” he says. “Then I had my own insight.”

Foerder moved the food and sticks outdoors and provided a cube or tub that could be moved over to the food if an elephant wanted to stand on it. In its seventh session of straining toward the food, 7-year-old elephant Kandula moved the cube into position as a stepstool and snagged some fruit (SN Online: 8/24/11).

“Elephants are more olfactory than visual,” Foerder says, and sniff with their trunks. When holding a stick, their trunks face the wrong direction for detecting what the stick is poking, and the trunk openings may even be closed. The experience points out that people may have to step outside their primate biases to get an idea of what another animal can do, Foerder says.

In the end, experiments that test animals’ cognition by determining when they succeed and when they fail may reveal more about human minds than other species’. Whether humankind truly wants to find all it looks for isn’t so clear. Homo sapiens is hardly modest about its brainpower, perhaps wanting to discover a bit of mental kinship while remaining mental kings.

Blurring the boundaries
New studies suggest that a variety of animals may have some version of the fancy mental powers once ascribed only to humans. It’s not easy to discern what’s going through another species’ mind, so these bold claims will need big proof. At the same time, a growing body of research suggests that humans share a number of largely unconscious, so-called irrational processes with other animals. — Susan Milius

tobkatrina/Istockphoto

Honeybees | Abstract concepts
The miniature brains of honeybees may be able to learn abstract concepts such as “same versus different” or “above versus below.” In experiments, researchers teach bees these notions by training them to fly toward a symbol that represents a concept, for example illustrating the idea of “above” with an icon sitting above a line. Researchers show the bees symbols that look different but have the same meaning of “above,” and eventually bees tend to make the right choice when they see a new symbol. The insects seem to be learning something beyond mere association of a specific symbol with a reward, and in one recent experiment were even able to learn two abstract concepts, left/right and above/below, simultaneously, Martin Giurfa of Paul Sabatier University in France and colleagues reported in 2012.

zzoom/Istockphoto

Dolphins | Syntax
Language may be one of humankind’s most glorious achievements, but decades of experiments suggest that bottle-nosed dolphins can learn sophisticated communication. Louis Herman of the University of Hawaii and his colleagues taught dolphins an invented sign language that requires an understanding of syntax — a way to string together words and phrases to create meaning. The dolphins could distinguish between commands such as “person surfboard fetch” to bring a surfboard to a person and “surfboard person fetch” to tow a person riding a surfboard. Now, Vincent Janik of the University of St. Andrews in Scotland has used microphones to eavesdrop on dolphin talk in the wild. Individual dolphins develop distinctive whistles that they use when meeting at sea, he and a colleague reported in 2012.

GlobalP/Istockphoto

Primates | Fuzzy math
Chimps are natural accountants, says Michael Beran of Georgia State University in Atlanta. They can roughly track how a group of objects changes as objects are added and subtracted, like in a bank account. And in a new test, Elizabeth Brannon of Duke University and her colleagues found that rhesus monkeys and college students do about equally well on a test of adding without being allowed to explicitly count objects. When she and a colleague gave both kinds of primates mere glimpses of two sets of dots — not visible long enough for the humans to actually count them — both did pretty well at selecting a picture that represented the sum of glimpsed sets. This shared ability to do approximate arithmetic in a flash probably reflects the evolution of cognitive abilities in the animal family tree, Brannon says.