trickster logic —

Scientists investigate why crows are so playful

New experiments reveal a complex link between crow play and tool use.

A New Caledonian crow uses a serrated leaf edge to pull grubs out of a hole in a log.
Enlarge / A New Caledonian crow uses a serrated leaf edge to pull grubs out of a hole in a log.
Mark Sibley

Crows share an interesting set of behaviors with humans: they like to play, and they often use tools. We know that humans play to learn. When toddlers knock over a pile of blocks, they're developing the ability to build and measure objects in the real world. The question is, do crows play for the same reason? An international team of cognitive scientists played with some crows to find out. What they discovered gives us a new understanding of crow consciousness, but it still leaves a lot of questions unanswered.

Lund University cognitive science researcher Megan Lambert and her colleagues designed three experiments to figure out whether there's a relationship between crow play and their ability to use tools to solve puzzles. It's well-documented that wild New Caledonian crows make a variety of tools, from hooked sticks to specially-prepared leaf edges, to pull insects out of hard-to-reach spots in trees. But crows have also been observed doing all kinds of weird things with tools, often for what seems like the pursuit of fun.

A crow sleds down a roof using a plastic lid.

In the YouTube video above, you can see a crow in Russia using a plastic lid to sled down a snowy roof. Researchers call these shenanigans "unrewarded object exploration." The crow doesn't get a "reward" because nothing about this activity aids its survival. Its only reward is the fun of sliding down a roof. But maybe, Lambert and her colleagues speculated, this type of seemingly goofy activity might actually lead to better tool use later on. The bird is learning about slipperiness, after all, and we even see it figuring out that it can't slide on the roof unless there's enough snow underneath the lid.

Toys into tools

The real question is whether there's a direct relationship between playing with something and then using it as a tool. The first two experiments that the researchers designed were attempts to answer this question. Researchers tested six wild New Caledonian crows, who had been caught five weeks before the study. For comparison, they also tested eight kea, born in captivity. Kea are known for their playful exploration of objects, but they don't use tools in the wild.

Both groups of birds were given several objects to play with: cubes of varying colors, half of which were heavy and half of which were light; and ropes of varying colors, half of which were rigid (they had a wire in the center) and half of which were flexible.

From the paper: (a) Block set: individuals must select a heavy block (weight corresponds with pattern) to insert into the opening of the box to collapse the platform inside. (b) Rope set: subjects must choose a rigid rope (rigidity corresponds with color) to push out the weighted box in the middle of the tube. Training objects for each set are shown in the insets of each image.
Enlarge / From the paper: (a) Block set: individuals must select a heavy block (weight corresponds with pattern) to insert into the opening of the box to collapse the platform inside. (b) Rope set: subjects must choose a rigid rope (rigidity corresponds with color) to push out the weighted box in the middle of the tube. Training objects for each set are shown in the insets of each image.
Royal Society

Once they'd played for a while, the birds were trained on a task in which they had to use a heavy object (not a cube) to collapse a platform and get a food reward. Then they had to learn to use a rigid stick (not a rope) to poke food out of a tube. So now the birds were familiar with their "toys" and with tasks that required tools to get food.

The researchers wondered whether the birds would change how they played with their toys once they knew how the food task worked. Maybe they would look more closely at the heavy cubes or rigid ropes? The researchers gave the birds some toys within view of the food platform, to keep the task fresh in the animals' minds. But none of the birds changed how they played with their blocks and ropes. There was no indication that the birds were trying to find tools that would help them get food later.

In the last phase of testing, birds were given 10 chances to pick a tool that would let them get at food on the platform or in the tube. In each of the 10 rounds, researchers presented the birds with two familiar toys that they'd played with before. To get the food, they had to choose either a heavy block or a rigid rope. The results were mixed.

During these phases of the experiment, crows and kea played with toys, then turned them into tools to get food from a food platform.
Enlarge / During these phases of the experiment, crows and kea played with toys, then turned them into tools to get food from a food platform.
Royal Society

Four of the 12 birds (two kea and two crows) performed far above chance when it came to the ropes. None of the birds performed above chance when it came to the blocks (in a separate experiment, one kea and one crow learned to choose the right blocks when they were color-coded, so it's likely they were going by color rather than weight). In a third round of experiments, in which the birds were not allowed to play with the tools, the researchers found that none of the birds performed above chance.

These experiments yielded several interesting insights. One is that the birds clearly did learn about the physical properties of objects by playing with them. When the birds were allowed to play with the tools before tackling the food test, they often made better choices about which tool to use. And yet not all of the birds were equally adept at turning toys into tools. The birds' behaviors were highly individualistic, despite the fact that they were all presented with exactly the same test conditions.

In an unrelated tool use test, a crow shows understanding of water displacement. It uses rocks to raise the water level in a beaker so it can get at food floating on top.

Indirect learning

What this suggests, say the researchers in a recent paper for Royal Society Open Science, is that the link between play and tool use is indirect. The two are clearly related, because the birds who played with tools were much better at using those tools in a food-finding task. But there was also huge variability between the birds, suggesting that they were not all getting the same thing out of play.

Importantly, the birds were not using play as a way of honing their skills on tool-using tasks. Write Lambert and her colleagues:

These results support the hypothesis that unrewarded object exploration provides information about object properties or affordances which can then be used to solve problems, but [they] do not speak to other potentially overlapping functions of exploration such as honing manual skills or generating novel behavioral sequences. Given its apparent costliness, it is likely that exploration confers myriad benefits...

New Caledonian crow subjects may apply information generated from their exploration of novel objects to select functional tools in a later problem-solving task; however, we have no evidence that that they engage in strategic exploration to gain information about the functional properties of objects with respect to a problem-solving task.

Playing is a "costly" activity because birds spend a lot of time doing it when they could be getting food, finding shelter, or doing other things to make survival more likely. From an evolutionary perspective, there has to be some benefit to play if it costs so much. But that benefit, as these researchers discovered, is complex and oblique. Crows played with the ropes because it was fun. Getting better at poking food out of a tube was only a secondary effect.

Crows may play simply because it helps them gain generalized problem-solving skills. Of course, that doesn't entirely explain one of the often-documented habits of crows, which involves goading cats into fights. In the video below, you can see how a crow pokes and pecks at two cats until they fight, then eggs them on.

A crow gets two cats to fight, then makes things even worse.

Similar videos show crows working together to get cats to fight, and tweaking dogs' tails to make them freak out. In a sense, the crows are treating these unwitting mammals as tools. They've learned the exact things that will drive cats and dogs mad (namely, pecking their backs and tails), and seem to enjoy the results.

What, exactly, is the "reward" for doing this? What are the birds learning in a generalized sense that might help them survive in other situations? The answer, as Lambert and her team discovered, is fairly ambiguous. There may not be any specific thing that the birds are learning from these activities. Possibly all they get is momentary amusement at the idea that they can make other animals do things. This might give crows a better understanding of how to manipulate objects and mammals to get food.

But perhaps further research will reveal that "unrewarded object exploration" is its own reward. Especially if it means that a pair of annoying cats gets tricked into smacking each other around.

Royal Society Open Science, 2017. DOI: 10.1098/rsos.170652  (About DOIs).

Channel Ars Technica