Social learning and physical cognition in kea

This project aims at gaining a deeper and more inclusive understanding of animal cognition by investigating the interaction of social learning and technical (physical) cognition in a species that seems to us exceptionally appropriate in this respect. The kea (Nestor notabilis) has proved to have the potential for advanced forms of social learning, it is highly motivated to attend to the foraging actions of conspecifics and is highly neophilic in the lab as well as in the field. Although not known as tool user, its manipulative skills and attendance to effects generated while manipulating objects indicate a high potential of physical cognition. Nevertheless, our previous research has produced ambiguous findings: free-living keas have failed to show convincing cases of social learning and their technical skills are distributed quite heterogeneously between individuals; captive keas seem to perform better in some social learning tasks and also master some challenging technical tasks but fail in others. Thus, three sets of fundamental issues remain unresolved: How does physical cognition and social learning interact? What are their developmental paths of physical and social intelligence? This questions may then provide clues to why do wild and captive keas differ in their competence of social learning and physical cognition? We will investigate these questions by providing captive and wild keas information from conspecific models for the solution of mechanical problems. This information will be provided during different stages of development in physical cognition and for cognitively different demanding mechanical problems. Investigating learning mechanisms of a non-mammalian species in the laboratory and field provides important integrative impact for biology and psychology. In this respect, this unique project provides, in mutual benefit, important knowledge for other national and international projects on comparative cognition. There is also great public interest in intelligence of this worldwide well-known species.

Tool use in nonhuman animals has often been thought to reflect a specific, evolved competence for increased physical intelligence and manipulative skills. However, recent demonstrations of "insightful" tool use in birds that do not naturally use tools have questioned this adaptive view, suggesting instead that some form of domain-general intelligence may be adequate to support sophisticated tool use. In order to investigate if wild and captive animals differ in such tasks, how physical cognition and social learning interact, and how these competences develop, we started a research project in summer 2006 that aimed at gaining such a deeper and more inclusive understanding of animal cognition. To address this, we studied keas (Nestor notabilis), New Zealand mountain parrots renowned for their playful cleverness and problem-solving abilities. Although not known as tool user in the wild, its manipulative skills and attendance to effects generated while manipulating objects indicate a high potential of physical cognition. We hand reared keas, tested them on tasks derived from developmental psychology, and performed social learning experiments. We found an amazing potential in our captive kea to combine objects and to insert objects into tubes also as means to acquire food (tool use). We show that success on the task required prior playful, exploratory experience, and thus demonstrate a role for unrewarded play in the mastery of a key novel behavioural competence. Older birds that failed in a tool-use task, succeeded after being given demonstration by a group member. There was no indication that demonstration accelerated the acquisition of tool-use in younger subjects. This parallels findings in naturally toll-using birds and indicates that the birds need to develop a concept for spatial relations between objects before they can learn socially how to use tools. We radio tracked keas in the wild, and when we tested them on a basic task of producing spatial relations between objects. They explored and played with the experimental objects but, in contrast to their captive conspecifics, failed to combine object. Our results indicate that the keas` sensorimotor intelligence and their playful and active information acquisition make them more likely to become aware about spatial relations between objects and in exploiting them in captivity than in the wild. Beside this general capacity to develop new competences depending on environmental conditions, we also found limitations in keas to handle stick-like tools in an efficient way. Our findings fit well with an extractive forager such as the kea where object exploration provides many advantages but little disadvantages in their natural environment. Thus by combining the developmental and the ecological approach in the study of animal cognition, we gained better insight into the function of manipulative and explorative foraging behaviour, e.g. such as tool use, and its distribution across the animal kingdom.