A study has shown that birds learn how to use tools through interacting with objects, but that’s not their intention – it’s all incidental. It’s a bit like how human infants explore the world, which is big news for our understanding of animal intelligence.
Researchers at the universities of York and St Andrews found that birds are better at choosing tools when they are given the chance to interact with different objects beforehand. This suggests the birds were learning something about the properties of the objects as they interacted with them. The tentative conclusion of the study is significant: by exploring things around them, certain animals learn in a similar way to human infants.
This type of latent learning, which occurs without reinforcement, is thought to be particularly important for more flexible forms of tool use or creative problem-solving.
"A lot of the birds that performed above chance level were younger birds, so this suggests that there’s a critical time during development where they are learning more,” says Megan Lambert, author of the study which has been published in the journal Royal Society Open Science. "This can be likened to human infants, who are much more curious and inquisitive about objects, as they use these to learn about the world."
For the study, New Caledonian crows and kea parrot’s, who are proficient tool-using birds, were presented with different objects: heavy and light blocks in one instance, and then rigid and flexible ropes in another. The objects were given visual markers to differentiate them; for instance, the light blocks had circles, and the heavy blocks were striped.
The researchers then gave the birds a task to complete using different, but very similar, objects. The aim was to teach them that heavy blocks push a platform down to reveal food, and that rigid ropes can be used to push food out of a tube. Next, they were given the chance to explore the objects again, with the task apparatus in view.
"All we wanted them to learn was that the task needed something heavy. The idea is that if they did learn that from the training when they explored the objects for a second time, they would pick up and throw the objects around more to gauge their weight," Lambert says.
Finally, they were given ten test trials, during which they had a choice between two objects (heavy and light blocks and rigid and flexible ropes) and had to choose one to complete the task with.
The results showed that of the 14 birds tested, six performed significantly above chance when they could explore the objects beforehand. Conversely, none of them performed above chance when they had no opportunity to explore the objects before the task. The team also found that the birds did not change their exploration tactics - such as exploring them for longer, or spending more time picking them up – once they learned that objects could be used as tools.
"Our results are tenuous because it not all birds learnt how to use the objects," Lambert says. "However, at the moment it seems that being particularly exploratory and interacting with objects gives the birds more opportunities to learn things. But, they aren’t necessarily adjusting how they explore to learn more."
In other words, the birds did not seem to actively seek information about the objects, but rather learned about their properties incidentally when they played with them. “For me the most interesting thing is that it seems as if interacting with objects is providing important information that they can use later. It would be really interesting to do it with more exploratory species, and I would predict it would be similar results," Lambert says.
Kea spent significantly more time exploring the objects than the crows. This is interesting because Kea use tools in captivity but does not use tools in the wild, whereas crows use them in both. For Lambert, it’s another step in a complex mission to understand more about animal intelligence. "There were a lot of birds that weren’t able to solve it, so trying to tease that apart is our future goal," she explains.
This article was originally published by WIRED UK