A report of the research is published this week in the journal PLoS ONE.[A video related to the research is available here].
Dr Alice Auersperg, who led the study, said: ‘We confronted members of both species with a ‘Multi Access Box’ in which a food reward could be acquired by applying one of four different methods: pulling a string tied around the reward, opening a window by pulling a hook-shaped lever, inserting marbles into a curved pathway so that they rolled down and knocked the food reward out of place, or inserting a stick-like tool into the side of the box to poke the food out.’
She adds: ‘Our aim was to examine the birds’ performance in a battery of different tasks. We allowed them to choose which task to tackle first, and once they were proficient with it, we blocked it to see how long and in which way they proceeded to the next solution.’
Six kea and five New Caledonian crows participated in the study, but only one of each, called Kermit (a parrot) and Uek (a crow) respectively, mastered all four tasks. For the kea the most difficult task was using a stick as a tool. This is not surprising, since kea do not do this in the wild, but, as Dr Auersperg explains, technical understanding was not their only obstacle: ‘The kea’s curved beak makes it impossible for them to hold a stick aligned with the head. Kermit overcame this constraint by developing a complex technique, involving several steps: he first held one end of the stick against the box opening, holding it laterally inside its beak. Thereafter he exchanged his beak with his foot, still pressing the tool’s end against the opening. Finally, he grabbed the other tool end with his beak and pushed the tool through the opening, manoeuvring it until the stick’s tip hit the reward. Kermit’s behaviour gave a strong impression of acting towards a purpose’.
The most difficult task for New Caledonian crows was instead opening the window by pulling a hook. Dr Auguste von Bayern, a co-author from Oxford University’s Department of Zoology, explains that this difficulty too should not be attributed exclusively to cognitive difficulties in understanding the task: ‘New Caledonian crows frequently use tools for food extraction, and new research shows that they also use them to explore potentially dangerous objects at a safe distance. Thus, for them, directly touching and pulling from novel objects is very threatening. Their difficulties seem related more to fear of contact than failure to understand: once the female crow Uek succeeded in opening the window, instead of putting its head inside the box to collect the food as the kea did in this task, she used a stick to poke at the food from a safe distance.’
Professor Alex Kacelnik from Oxford University’s Department of Zoology, another member of the team, comments: ‘In the early days of comparative cognition research scientists were content by demonstrating which wonderful tasks their subjects could solve, as if competing to show that their study species was particularly clever. Not anymore. Many skills previously thought to be confined to humans have been demonstrated in other species, so that the critical task now is to understand how they do it, whether different species solve similar physical problems by different routes, and which factors in their ecology led to different cognitive specialisms.
‘The list of abilities where humans have lost exclusivity is large, and includes extraordinary feats of memory, tool making, multi-tool use, cultural transmission of skills and vocal communication, orientation, and so on. Our task is finding out how they do it, how and why natural selection leads to differing skills, and what are the limits of each species’ performance.’
He points out how Kermit’s quick acquisition of a complex sequence of behaviors aimed at achieving a well-defined goal is reminiscent of similarly complex behavior in New Caledonian crows; such as the use of up to three tools in a logical sequence displayed by Betty, the mother of present top crow Uek.
Professor Kacelnik said: ‘Such sequences cannot be learned by acting at random and simply repeating whole sequences that frequently yield a positive outcome: there is simply not enough opportunity for this. The issue reminds us of the problem of data sparcity in human language acquisition. As Chomsky and others noted, the linguistic experience of each human child is insufficient to simply learn human languages by trying utterances, let alone sentences, at random. A constructive process including innate propensities, generalization from experience and cognitive operations on this information must be involved.
‘As with human language acquisition, we do not know exactly what psychological operations the animals use to solve such problems, but they are likely to include heritable propensities and higher-order generalizations supported by experience with individually learned actions.
’For more information contact Alex Kacelnik on +44 (0)1865 271164 or email firstname.lastname@example.org
Video related to the project is available at:http://d3qk4vw19t7z2n.cloudfront.net/Kea_Crow_Suite_Movie_web.mp4
Images related to the project are available at:http://www.ox.ac.uk/media/news_releases_for_journalists/110603.html
Alternatively contact the University of Oxford Press Office on +44 (0)1865 283877 or email email@example.com