Reasoning in birds

Corvids and parrots possess advanced cognitive abilities. However, so far the functional differentiation within these groups remains unclear. The "adaptive specialisation hypothesis" suggests a species-specific development adaptive for the particular socio-ecological environment, which may lead to significant qualitative and quantitative differences between species. In contrast, according to a "general process view", evolutionary pressure on developing advanced, but general-purpose cognitive skills would lead to a rather homogenous development across species. Surprisingly little is known about animals` abilities to process incomplete and / or indirect information, although this has been suggested to be one of the most important functions of cognition. The use of partial information may be valuable in a wide range of contexts, i.e. to infer the location of hidden food by knowing where food cannot be found and what food source a competitor most likely had exploited. In fact, most of the cognitive abilities of corvids had been attributed to an arms-race between food cachers and pilferers. To investigate the potential adaptive differentiation of cognition in birds, we will test five species of parrots and corvids for their ability to (i) infer the location of hidden food by knowing where food cannot be found (ii) infer where another bird had found food before (iii) infer where another bird will search for food and (iv) infer unknown individual`s rank position. We selected three corvid species caching food to different degrees: intensively-caching European jays, moderately-caching New Caledonian crows and non-caching jackdaws. Additionally we will test keas and grey parrots, that both do not cache. This comparison may be of interest in its` own right: even though both species possess well-documented cognitive abilities, keas may have evolved different cognitive skills or may apply them differently than grey parrots due to their evolution on an island without mammalian predators. Following the two major frameworks, only minor quantitative differences between the species would favour a "general process view", whereas qualitative differences, particularly in the tasks (i) - (iii), may support the "adaptive specialisation hypothesis". However, since all five species have a complex social organisation, only minor differences may be expected in task (iv).

It was the aim of this project to investigate the ability of birds to draw logical conclusions, using a comparative approach,. Our main study subjects were several species of corvids and parrots, but we also included greylag geese for comparative reasons. Logics and reasoning are particularly interesting as they are regarded as hallmarks of human intelligence; it is still debated whether any non-human animals, let alone birds, exhibit such abilities. Furthermore, we were interested whether the cognitive abilities of our subjects could be explained as adaptive specialization for certain environmental contexts or as generally applicable cognitive tools. Finally, we were interested in the reliability with which the birds use their higher cognitive abilities and whether and how much these could be constrained by phylogenetically older mechanisms. We focused mainly on the ability to reason by exclusion. In other words, can an animal exclude one option based on the information that this option is evidently unrewarding, and infer logically that an alternative option is correct? We applied a variety of different experiments to answer this question. Initially, the animals were confronted with the task to find food through exclusion (if evidently not in A, it must be in B). The food-caching species among the corvids, ravens and carrion crows, were particularly successful, as were the Grey parrots. In contrast, the non-caching jackdaws and the keas failed. In more complicated reasoning tasks, only the Grey parrots remained successful. However, if the task involved the deduction of social relationships rather than the location of hidden food, also the non-caching jackdaws were successful. From these results we conclude that both corvids and parrots are capable of logical inferences, but different selection pressures may have led to the slight divergence of these skills in the two groups. Within the corvids it seems possible that this ability evolved particularly as an adaptive specialization to food caching, whereas this explanation may not hold for parrots. One possible explanation would be that for what reason ever, a greater degree of "general intelligence" skills (i.e. those which can be generalized across contexts) evolved within the parrots. Results and the methods were published in detail in 18 peer reviewed publications and 32 congress contributions. And we are pleased to report that this project provided a great opportunity to integrate, train and mentor a fair number of students and that it provided a base for individual careers and also, the recent positive developments at the University of Vienna (e.g. successfully competing for an FWF PhD college "Cognition and Communication").