They might not understand what they have in common with brothers of the Night’s Watch or why “as the crow flies” means traveling in a straight line, but hooded crows are capable of analogical reasoning. According to a study published in Current Biology this week, they can solve higher-order relational matching tasks like humans and other primates.
Analogical reasoning was once thought to be unique to us. It’s the basis of categorization, creative problem solving, and also scientific discovery. But how do you test this? Previous work showed that “relational matching-to-sample” (RMTS) basically captures the essence of analogy, visually. If a sample pair was, say, AA, then picking BB would be correct. If, on the other hand, the sample pair was CD, then EF would be the right test pair to pick. None of the items in the test pairs physically match the items in the sample pair -- that way only relational sameness or differentness is used to help with the choices. Apes and monkeys have successfully learned RMTS.
Now, an international team led by Edward Wasserman from the University of Iowa wanted to see if non-primates can do it too. After all, crows and ravens have long been heralded for their intelligence. They trained two hooded crows (Corvus cornix) to identify items by color, shape, and number in what’s called “identity matching-to-sample” (IMTS). The birds were placed in a wire mesh cage with a plastic tray containing three cards and two cups. The card in the middle served as the sample card. The cups on either side were covered with the other two cards: One matched the sample (in the color, shape, or number of items pictured), while the other didn’t. The cup with the card that matched the sample card contained two mealworms as a reward.
In the second part of the experiment, the birds were tested with relational matching pairs. A sample card with two same-sized circles, for example, means they should pick the test card with two same-sized squares -- and not two different-sized circles. To the right, you can see a crow making the correct selection during a relationship matching trial, while study author Anna Smirnova of Lomonosov Moscow State University takes note. The birds picked the correct card more than three-quarters of the time, Science News reports.
“What the crows have done is a phenomenal feat,” Wasserman says in a news release. Furthermore, the crows displayed this response spontaneously, without ever having been trained on RMTS. They had only been trained on IMTS, which didn’t involve understanding relationships between items. “That is the crux of the discovery,” Wasserman adds. “Honestly, if it was only by brute force that the crows showed this learning, then it would have been an impressive result. But this feat was spontaneous.”
And that’s something to crow about!
Images: Lomonosov Moscow State University
3
