Crows have human-like powers of inference, study says

A smart species of bird called the New Caledonian crow can make inferences about the behavior of hidden animals in its environment, an ability previously observed only in humans, according to a new study published Monday by an international team of researchers in the Proceedings of the National Academy of Sciences.

The capability is obviously adaptive for wild animals. As an example, the researchers describe what a bird might see from its perch: “Imagine a bird looking down on a monkey moving through a forest canopy. Generally, the bird will be able to observe both the monkey moving and the canopy shaking at the same time. Sometimes, however, when the canopy is thick overhead, the bird may only observe that, against a background of stationery leaves, there are waves of moving leaves that can start and stop abruptly.”

Humans can infer that an animal is causing the waves of moving leaves, and, depending on the pattern, may even be able to tell a monkey is causing the movement. What’s more, studies have shown that humans possess the ability to make these inferences from infancy. But can other animals do the same?

It turns out, at least in the case of the New Caledonian crow, that they can. In the new study, the researchers first allowed the crows to use a tool in their enclosure to fish some food out of a box (tool making and tool use are abilities the crows are already famous for).


The researchers then moved the box so it was near a corner of the enclosure that also had what’s called a “hide” -- basically a sheet with a slit in it that the researchers could hide behind while poking a stick out of the slit, a sort of ramshackle Wizard of Oz setup.

The box was positioned so that the researchers’ stick would come near the box when it was poked out, scaring the bird and discouraging it from feeding.

The researchers then exposed the crows to two conditions. In the first one, two humans walked into the enclosure. One stood in the open while the other hid behind the sheet and poked the stick through, near the box. Then both researchers left the enclosure.

In the second condition, only one human entered the enclosure, standing in the open away from the hide. Nonetheless, the stick poked through the slit and began to move (it was rigged to be movable from outside the enclosure). Then the single human left the enclosure.


After both conditions, the researchers recorded standardized measures of how cautious the crows were being in their feeding.

The key difference between the two conditions is that in the second condition, no human was observed leaving the hide after the stick was poked through.

If the birds had no ability to infer that there was a person hiding behind the sheet and moving the stick, their behavior should show that they would simply get progressively more comfortable with the hide and the stick, becoming less cautious over time regardless of whether humans exited the hide before feeding.

But if the birds could infer something about the humans’ behavior while they were hidden, they should become highly cautious of feeding after the second condition, because they did not observe a human’s departure after the stick moved -- so the stick might move again at any time, striking the birds during feeding.


That increased caution is exactly what the researchers found: The birds were about 1 1/2 times more likely to break off feeding to inspect the hide after the second condition than the first.

The researchers say that this provides strong evidence that the New Caledonian crows possess the ability to infer the presence of “hidden causal agents” -- living beings responsible for the movement of an inanimate object while hidden, like a stick or shaking tree leaves.

It’s another ability knocked off the “humans only” list. And now that they’ve found it in crows, the researchers say, it may well be found in other animals too, allowing scientists to begin to trace the evolution of this piece of complex cognition.

Return to the Science Now blog.