April 15: Crows and their relatives, like jays and rooks, are definitely in the gifted class when it comes to the kinds of cognitive puzzles that scientists cook up.
They recognise human faces. They make tools to suit a given problem.
Sometimes they seem, as humans like to say, almost human. But the last common ancestor of humans and crows lived perhaps 300 million years ago, and was almost certainly no intellectual giant.
So the higher levels of crow and primate intelligence evolved on separate tracks, but somehow reached some of the same destinations. And scientists are now asking what crows can’t do, as one way to understand how they learn and how their intelligence works.
A useful tool for this research comes from an ancient Greek (or perhaps Ethiopian), the fabulist known as Aesop. One of his stories is about a thirsty crow that drops pebbles into a pitcher to raise the level of water high enough so that it can get a drink.
Researchers have modified this task by adding a floating morsel of food to a tube with water and seeing which creatures solve the problem of using stones to raise the water enough to get the food. It can be used for a variety of species because it’s new to all of them.
“No animal has a natural predisposition to drop stones to change water levels,” said Sarah Jelbert, a doctoral student at Auckland University in New Zealand who works with crows.
New Caledonian crows, rooks, Eurasian jays and humans (past age 5) can do it, said Jelbert, who noted that great apes could do a slightly different version.
But in the latest experiment to test the crows, Jelbert, working with Alex Taylor and Russell Gray of Auckland and Lucy Cheke and Nicola Clayton of the University of Cambridge in England, found some clear limitations to what the crows can learn. And those limitations provide some hints to how they think.
The birds, Jelbert and her colleagues reported in PLOS One last month, were wild New Caledonian crows trapped for the experiment and then released.
The crows were first trained to pick up stones; this is not something they do in the wild. They dropped the stones into a dry tube to gain a reward. Then they took the Aesop test, in several different situations. The birds learned not to drop the stones in a tube of sand with a treat.
And they correctly chose sinking objects rather than floating ones, and solid rather than hollow objects, to drop in the water.
But if part of the tube apparatus was hidden, the birds could not learn. They also seemed unable to learn that the water would rise more quickly with fewer stones in a narrow tube.
This suggested two things, said Jelbert. They weren’t just learning abstract rules, because otherwise they would have been able to learn where to drop the stones to make the water rise even if they couldn’t see what was going on.
And second, the need to see the results of the behaviour suggested that they did seem to have “a level of causal understanding.” These were just hints, though, in terms of understanding how crows learn and think, Jelbert said. “We’re still very much at the beginning.”
Amanda Seed, who studies animal cognition at the University of St Andrews in Scotland, said in an email that the results were intriguing but still left open the question of whether the crows grasped cause and effect.
“The experiment raises a really interesting question: Why is intermediate visual feedback so important for learning?” she wrote. “But whether or not a representation of causality comes into this remains to be seen.”