Crows Understand a Fundamental Part of Logical Reasoning

Crows are far more rational than we had realized. New research shows that wild New Caledonian crows can compete with 7-year-old children when it comes to understanding causality, or how one action causes another.

Round objects always roll down hills. Dense objects sink. What goes up must come down. Researchers have used various experimental methods to assess whether non-human animals are capable of that sort of causal reasoning, but there are so many different tasks, and so many species whose success is so tightly dependent on the particular task chosen, that it's hard to make comparisons between species.

For example, there's a widely used experiment called the trap-tube task. An animal uses a stick to push or pull a food reward out of a transparent tube, avoiding a small well in the center of the tube where the food would become stuck. Oddly, great apes were more successful when they could pull, rather than push, the food from the tube. Either orangutans, gorillas, chimpanzees, and bonobos have a very limited understanding of causality, or the task isn't really measuring what it's supposed to measure. The second explanation is more likely, especially since there are cases where human adults understand that sort of causality and still somehow manage to fail some versions of the task.

The Aesop's fable paradigm, as it is known among animal behavior researchers, in which an animal has to drop stones into a tube of water to bring a floating food reward close enough to grasp, has so far proven to be a more useful to task to assess causal reasoning and intuitive or "folk" physics.

It's very similar to another incredibly productive paradigm, the "floating peanut task," in which apes or children need to spit water into an empty tube with a food reward at the bottom to reach their snack.

It was in 2009 when corvid researcher Nathan Emery first showed that rooks could solve the task. Follow-up experiments designed to further probe their ability showed not only that they implicitly understood something about displacement, but also that liquids operate differently from solids. They were more likely to drop stones into a tube containing water than one containing sawdust. They also understood something about efficiency, opting to use fewer larger stones than to waste their time with lots of smaller ones. Further studies demonstrated that human children and other corvids – Eurasian jays and New Caledonian crows – could reason about displacement as well.

Now, a group of researchers from New Zealand's University of Auckland and the UK's University of Cambridge, led by psychologist Sarah A. Jelbert, have given a group of wild New Caledonian crows (Corvus moneduloides) a few new variations on the Aesop's fable paradigm to see just how sophisticated their causal reasoning is.

The different experiments are shown in the video above. In the first experiment, the researchers replicate an earlier finding that the crows preferentially drop stones into water rather than sand to retrieve their snack. The second and third experiments pitted heavier and lighter objects, and solid blocks against hollow ones. The fourth experiment pitted narrow tubes against wide ones. Would the birds realize that fewer objects were necessary to displace water in the narrow tube?

The fifth experiment was a variation on the fourth, with narrow and wide tubes as before. However, the water level in the narrow tube was so low to begin with that the crows would never have been able raise it high enough to grab the reward, meaning that the only useful option was to use the wider tube.

The birds passed the test in all cases except for the fourth experiment. They were able to retrieve the food from the wide tube in that task; it just took a bit more work. It's a failure, but a qualified one. That they succeeded the fifth experiment shows that they do understand something about the diameter of the tubes, even if it isn't always shown by their behavior.

The final experiment, called the "U-tube task" was a bit different. The crows were presented with a set of three tubes. The food reward was floating in the center one, but it was too narrow to drop items. Instead, one of the outer tubes was connected to the central tube, but the connection was concealed. To pass the test, the crows would have to try dropping objects into one of the side tubes, notice that the water level in the central tube would be raised, and then learn that the effect would only occur for one of the side tubes, not both. The crows failed miserably at this test, just randomly dropping objects into either of the side tubes. Not a single crow got it right.

New Caledonian crows are very clever, but not quite as clever as eight-year-old children, who fairly effortlessly passed the U-tube task in other experiments.

While a coup for corvids, this study adds to a growing body of research that demonstrates just how useful the Aesop's fable paradigm is. For one thing, it doesn't rely on a behavior that only some species perform in the wild, and not others, nor does it require direct involvement of demonstration from human researchers. "Any animal capable of picking up stones could potentially participate," Jelbert says. And it doesn't take hundreds of trials to measure performance. Since it is extraordinarily unlikely that wild animals routinely drop stones to raise water levels, the task isn't somehow more ecologically relevant or familiar for some species than others.

And their failure in the U-tube task is also useful. Is that failure limited to New Caledonian crows? To all corvids? All birds? If other species fail, it might reveal that learning the counter-intuitive causal nature of the task is a uniquely human ability. Failure is often just as enlightening as success when it comes to research in animal behavior.

Jelbert S.A., Taylor A.H., Cheke L.G., Clayton N.S., Gray R.D. & Addessi E. (2014). Using the Aesop's Fable Paradigm to Investigate Causal Understanding of Water Displacement by New Caledonian Crows, PLoS ONE, 9 (3) e92895. DOI: 10.1371/journal.pone.0092895.s003