Read our COVID-19 research and news.

Dig deep. New Caledonian crows are known to use tools; they can also reason out the presence of hidden mechanisms.

Mick Sibley

Whodunit? Crows Ask That Question, Too

Imagine hearing a distant roll of thunder and wondering what caused it. Even asking that question is a sign that you, like all humans, can perform a type of sophisticated thinking known as "causal reasoning"—inferring that mechanisms you can't see may be responsible for something. But humans aren't alone in this ability: New Caledonian crows can also reason about hidden mechanisms, or "causal agents," a team of scientists report today in the Proceedings of the National Academy of Sciences. It's the first time that this cognitive ability has been experimentally demonstrated in a species other than humans, and the method may help scientists understand how this type of reasoning evolved, the researchers say.

Causal reasoning is "one of the most powerful human abilities," says Alison Gopnik, a psychologist at the University of California, Berkeley, who was not involved in the study. "It's at the root of our understanding of the world and one another." Indeed, it is the key mental ability for many things humans do, including inventing, making, and using tools. We develop this ability early in life: A 2007 study in Developmental Psychology reported that human infants as young as 7 months old understand that when a beanbag is tossed from behind a screen, something or someone must have thrown it. The infants infer that a "causal agent" must be involved in the motion of the flying beanbag.

But why should this ability be limited to humans? "It seems like it would make good sense for crows and many other animals to be able to distinguish between the wind rustling tree limbs and an unseen animal crashing through the canopy," says Alex Taylor, an evolutionary psychologist at the University of Auckland in New Zealand and the lead author of the new study. Because New Caledonian crows are also inventive and skillful tool-users, Taylor and his colleagues thought the birds might have causal reasoning skills similar to those of humans.

Working on Mare Island in New Caledonia, the scientists captured eight wild crows (five adults and three juveniles) and housed them inside a large outdoor aviary. Over the next few days, the birds used a slender stick to extract food from a small box placed on a table in the aviary. Then the plot thickened: The scientists placed the food box close to a blue tarp large enough for a person to hide behind. The researchers also set up a large stick that could be poked through the tarp and waved around by a human outside the aviary pulling on a string; the moving stick posed a danger to the birds if they tried to extract the food.

Behind the curtain. New Caledonian crows respond to the "hidden causal agent" and "unknown causal agent" conditions. The number of inspections of the hideaway by the crow is noted in the top left-hand corner for each condition.
Credit: Alex Taylor

The crows in the aviary then observed two different situations. In one, the "hidden causal agent" scenario, the crows saw a human enter the blind. Then, a few moments later, the stick poked through the tarp and moved back and forth 15 times. The human then exited the blind. In the "unknown causal agent" scenario, the crows saw only the stick as it emerged from the tarp and moved back and forth 15 times.

In both situations, a human also stood next to the table in the aviary, so the crows never tried to get the food. And in both cases, when the visible human left, the crows began to remove their food from the box. Yet the crows' behavior differed depending on whether they had seen a human come and go from the blind. If the birds had seen a human stepping out of the blind, they seldom gave the stick so much as a glance as they dug out their food. But the crows that saw the stick move but no one emerge from the blind were nervous: They often stopped probing for food and studied the blue tarp and stick—apparently suspecting that someone or something unknown had caused the stick to move and that it might move again. Some even flew away from the setup.

Together, the tests show that the crows are "capable of causal reasoning," Taylor says. "We expected the crows to initially be scared of the moving stick. Instead, they only became scared when they could not attribute the movement to a hidden human—which suggests the crows were reasoning that the stick's movement was caused by that human." The crows, he says, apparently don't expect an inanimate object to move on its own, just as infants don't expect beanbags to be tossed through the air by a toy block.

"It's an extremely clever study," says John Marzluff, a wildlife biologist at the University of Washington, Seattle. "Using a controlled experiment, they've validated what many crow hunters know—that crows keep track of hunters in blinds. Even if the crows [in this study] never see a person push the stick, they connect the dots between the location of a person and the actions they associate with people."

The study "makes important new advances in our understanding of the extent to which nonhuman animals may be capable of causal reasoning and offers the potential to open this whole area up to scientific inquiry in animals," adds Nicola Clayton, an experimental psychologist at the University of Cambridge in the United Kingdom. Because it suggests that causal reasoning evolved in parallel in humans and crows, the work may even "help solve the fascinating question of just how and why our human intelligence evolved," Gopnik says.