The Crow: Black, beaky, beady ... brilliant

Betty the crow managed to bend a piece of wire to make a hook, which she then used to get food. Does this bit of ingenuity put her above man's close relatives? Charles Arthur reports

Crows are smart! Or so it would seem, to judge by our reaction to the doings of Betty, the New Caledonian crow that demonstrated on film that it could bend a piece of wire to create a hook, which it then used to get some food. (If you like, you can watch the film for yourself on the internet at the address below.)

What Betty did was quantitatively different from the sort of exploits one sees in most other animals. It is not a pigeon pushing a lever to produce food, or a rat negotiating a maze that it has been shown before, or even a chimpanzee pointing to a picture of something it wants. It is an animal and a tool – the sort of thing we thought we humans excelled at. (Actually, we still do.)

Rudimentary tool use has been reported before: some African chimpanzees have been observed selecting and using stones to open nuts, and monkeys are known to use sticks to fish edible ants and termites out of their nests. But this is different. It's an animal making a tool. Most of all, an animal that had not seen wires being bent to make things before.

"Toolmaking and tool use has always been considered one of the diagnostics of a superior intelligence," said Professor Alex Kacelnik of Oxford University's behavioural ecology research group, which carried out the experiment. "Now a bird is shown to have greater sophistication than many closer relatives of us humans."

Most remarkable was that Betty "had no model to imitate, no opportunity for hook-making to emerge ... she had seen and used supplied wire hooks before but had not seen the process of bending".

So should we now be chummying up to the New Caledonian crows, and forget the dolphins, whales, apes, dogs, cats and other animals in which we have tried to perceive intelligence?

It is very tempting to say yes. But while it is easy to give in to anthropomorphism – assigning human characteristics to animals – it is a temptation that has to be resisted.

Indeed, it is a temptation that Professor Kacelnik has resisted before. Last November a team at the University of Cambridge did an experiment with American scrub jays and found that animals which had themselves stolen items from others jays were more worried about it happening to them: if they were being watched by another bird while they buried their food, they would return and rebury it; those who had no history of thievery left their own buried items alone, watched or not.

What can it mean? One suggestion was that the birds might be "putting themselves in others' shoes" – also known to psychologists as "theory of mind", which only humans definitely exhibit. Professor Kacelnik said: "This doesn't allow us to say that the jays were aware of a causal relationship but they behave as though they are."

However, mind the gap – the one between behaviour and intention. That is where many dreams of finding smart animals disappear. Go back to the 19th century and you will find Clever Hans, the show horse that was claimed to be able to do maths by tapping its foot with the answer when asked a question – say, add five and six. It was exceptionally good at it. Except that one odd thing emerged: if the people watching did not know the answer to the question, Hans would get the answer wrong.

What was going on? Hans was not doing maths; he was reading the subtle body language of the people around him, who would unconsciously relax when he reached the answer they knew was right, and give him a sugar lump. Human "psychics" on premium phone lines do much the same, detecting vocal signs and giving feedback about as useful as Hans's. They get rewarded in hard cash, though.

The question of whether animals "think" is a subtle one. Of course they do. The problem is that they do not think anything like we do – and that makes it hard to express what is going on in their heads.

We know we think because we have language. What we do not think is that any other animals have their own language. The science writer Stephen Budiansky comments that language is "a rocket that has allowed us to escape the gravitational pull of biological adaptation". Other animals pass messages, but it is not the same as language that contains concepts. It is reaction to the moment. That is the difference between the thinking that we humans do, and the cognition – it is the best word – that animals like Betty perform.

That is not to say such cognition is not valuable, or even surprising at times. In his BBC TV series The Life of Birds, David Attenborough observed: "No living bird is truly stupid. Each generation of birds that leaves the protection of its parents to become independent has the inborn genetic information that will help it to survive in the outside world and the skills that it has learned from its parents. They would never have met the challenge of evolution without some degree of native cunning. It's just that some have much more than others."

The same goes for any group of animals; do we call a leopard dim because all it is good for is chasing game, and it is really never performed well in wire-bending experiments?

The problem actually comes down to that thing we use so effortlessly, as we seem to be genetically programmed to do – language. We say that Betty bent the wire. This implies that she saw a series of steps to perform, and mapped them out. But how does Betty see it? It is an unimaginable concept, though one could imagine that in the bird's neurons there was some memory of hook-shaped things and what they did, and another memory of how those hook-shaped things felt like the same material as this; then something sparked, and the wire was being bent.

A more interesting experiment now would be to see what Betty, or her brethren, made of wires that were either too thick to bend, or too thin to hold a shape; and also with plastics that feel unlike metal but hold a shape. It's an experiment that's waiting to be done. And it would certainly sort the thinking sheep from the cognitive goats.

But Betty can rest easy. After all, the experiment was repeated 10 times, and she succeeded nine times. She's not going hungry.

You can see Betty make the hook at www.sciencemag.org/feature/data/crow/index.html