Ducks might be capable of abstract thought, say scientists

Humans might not be such standout thinkers after all. Ducklings just might be capable of abstract thought too.

A new study suggests that ducklings can learn, and imprint on, the relationship between objects, considering more than just the characteristics of the objects themselves. And that could disrupt what we think about thought.

When a duckling hatches, it quickly identifies its mother, or what it thinks is its mother, and begins to follow her around. Researchers harnessed that behavior for a new experiment described in a paper published Thursday in the journal Science. 

Antone Martinho, a doctoral student at the University of Oxford, was curious about whether these little fuzzy babies learned and remembered their mother using a simple visual memory or whether imprinting was representative of more complex cognitive abilities.

So the researchers set up an experiment intended to test if the ducklings could learn a concept. The concept the ducklings were supposed to learn was relational, that is, whether two objects had the quality of "same" or "different." 

Just hours after a duckling hatched, it was placed in a training enclosure with two objects attached to a string, moving around the enclosure in a circle. Those objects were either the same, perhaps two spheres, or different shapes, like a cone and a cylinder. All the objects were red. 

As the pair of objects moved around the enclosure, the ducklings began to imprint on them and follow them around. 

Then, after a half-hour break, the duckling was put into a test enclosure where it encountered two moving pairs of objects. One pair was the same shape, perhaps two pyramids, and the other pair was different, for example a cube and a prism. The duck had seen none of those shapes in the training enclosure.

The researchers watched to see whether the duckling followed around the 'same' pair or the 'different' pair. 

And 32 of 47 tested ducklings followed the pair that matched the relationship of the pair they had imprinted on in the training enclosure, suggesting that 68 percent of the ducklings had accurately learned the relational concept of 'same' versus 'different.'

The researchers repeated the experiment with 66 fresh ducklings, using spheres that differed in color rather than variously shaped objects. And the accuracy was about the same: 45 ducklings preferred to follow the objects with the same relational concept as the ones they imprinted on.

"We really did go into this expecting them not to be able to do this," Mr. Martinho tells The Christian Science Monitor. "We were shocked not only that they did it, but also that they did it with such accuracy."

Jennifer Vonk, director of the laboratory of cognitive origins at Oakland University in Michigan who was not part of the study, isn't so sure that this experiment actually tested if the ducklings were capable of processing the relationship between the paired objects. Instead, she says, perhaps the ducks could be imprinting on more or less variability in the pair as an overall configuration, or one unit.

If that's the case, she says, "then that's really attending to a perceptual feature and showing a perceptual preference. And attending to a single feature isn't really considered abstract reasoning."

Dr. Vonk says her hesitation hinges on whether or not it would make sense for ducks to be able to reason about the relationship between objects. In other words, how being able to understand sameness and difference helps ducklings know which mother duck to follow, for example.

"There's definitely something interesting going on here," she tells the Monitor, "but I do think that we have to think further about why this would be useful or adaptive to engage in this kind of reasoning."

Ducks aren't the only animals studied by researchers looking for evidence of such concept learning.

"Previously we've seen this ability in things like primates and in parrots and in crows – animals that we generally think of as quite intelligent," Martinho says. "But by seeing it in ducklings, which are only distantly related to parrots and crows, that makes the suggestion that probably a wide variety of vertebrates are going to have this ability."

With this ability apparently cropping up among a broad variety of animals, perhaps such abstract thinking is actually a convergent trait, independently evolved in distinct species under different pressures. 

"Our one project can't say whether it is ancestral or convergent," Martinho says, "but it certainly does reinforce the usefulness of this ability."

Furthermore, finding that other animals may be able to grapple with abstract ideas to varying degrees highlights that humans are animals, too. 

"We often think of things like ability to form abstract thoughts as one of the things that makes humans special," Martinho says. "That's not necessarily untrue, but the presence or absence of abstract thought is probably not what makes humans special."

Instead, he explains, "it's probably a question of degrees," of just how abstract those thoughts are and what we can do with those thoughts.