What got you interested in biology in the first instance?
Ever since early childhood, I have been fascinated by animals. I wrote in the friendship books of my classmates that I wanted to become an ‘animal researcher’. Later I learned that the proper name is biologist.
Around the time I started at university, I got interested in animal behaviour after reading the books of Dian Fossey, Jane Goodall and Konrad Lorenz. Fortunately, in the early 90s Vienna became a great place to study animal behaviour because of the diversity of fields and approaches covered by different working groups, from behavioural ecology and physiology to human ethology. I took every possible course and soon wondered about the lack of cognitive questions aside in the primate literature. I nevertheless got hooked on such kinds of questions and didn’t have to think for long when I got the chance to join a newly formed group on animal learning and concept formation. They also had a colony of common marmosets, which I chose as subjects for my first scientific studies.
How did you come to study ravens?
As a student, I visited a friend at the Konrad Lorenz Research Station in the Austrian Alps who had just hand-raised a group of ravens. When I saw the playfulness of these birds, I was struck by the richness of their behavioural repertoire — they reminded me more of puppy dogs than of normal songbirds. I was also fascinated by the curious way they can look at you, which reminded me of the looks you often get from primates. So I happily took the offer to study ravens for my PhD. My task seemed simple at the time: just find out how intelligent they are. But I now know that this is a question for a lifelong research project.
Is it true that corvids are the smartest birds?
Corvids have large brains, show a high behavioural flexibility and live in a diverse range of habitats. These features are likely interlinked and indicate considerable flexibility in decision-making. So, yes, they certainly can be considered ‘smart’ birds. But I wouldn’t claim that corvids are the smartest birds. First, advanced cognitive skills are also found in other avian groups, such as parrots, and there are several avian taxa with relatively large brains (for example, woodpeckers) and/or ecological and social complexity (for example, gulls) that have barely been tested for their cognitive skills. And second, it is difficult to say how smart ‘corvids’ are because it depends very much which corvid species we’re talking about. Out of the around 120 described species, only a handful have been studied so far, and even those only in a limited number of contexts and paradigms. For instance, tool use has been investigated predominantly in New Caledonian crows and mental time travel mainly in Western scrub jays. So it is just too early to say.
Is this because research on avian cognition is a relatively recent development, or could it just be that researchers are skilled in picking the best species for their questions?
I think it’s a combination of the two. Cognitive studies on birds in general, and corvids in particular, are relatively new, and within the short time period of roughly 25 years we could not achieve a level of knowledge comparable to the vast literature on primates. Of course, the few labs working on corvids have chosen their model species carefully according to their research questions, and there’s nothing wrong with this. It is just when one tries to take a broad view that this might lead to the impression of cherry-picking the best species for any given skill in corvids (perhaps so that they can rival the competence of another given species or taxon, say chimps or other apes).
So what is left of the idea of corvids being ‘feathered apes’ or ‘primates of the sky’?
The first comparisons were relatively rough, but they were tremendously important for the field. They certainly rocked our intuitive understanding of cognitive evolution, and the assumption that primates are always on top. But it does not really matter which species is ‘smartest’ or which taxonomic group is ‘most advanced’. The critical point about the ‘feathered ape’ idea is to consider the possibility of convergent evolution: that is, that certain socio-ecological selection pressures may have favoured the evolution of similar cognitive skills in phylogenetically distant species. Because the last common ancestor of birds and mammals lived about 300 million years ago, birds such as corvids are perfectly suited to test hypotheses about brain evolution and intelligence — mostly developed based exclusively on primates/mammals — independently of phylogeny. In this respect, it is also worth mentioning that avian brains are very differently organized than mammalian brains, indicating that complex cognitive solutions can be achieved using very different neural structures and circuits (for example, without a neocortex).
What are the next steps in this research field?
We have just scratched the surface of understanding what conditions can drive the evolution of what types of cognition. There is ample room for improvement, from theoretical and conceptual considerations to methodological refinements. One possibility would be to make a better use of the comparative approach within the corvid clade. Corvids vary in feeding ecology (for example, short-term versus long-term hoarders), breeding biology (territorial, communal and cooperative breeders) and social structure (types of social relationships, degree of fission–fusion dynamics), which makes them perfect candidates for testing which factors are important for the evolution of particular skills.
Take food caching, probably the best investigated trait in corvids, where several findings fit a coherent picture: first, long-term hoarding species tend to outperform short-term hoarders in spatial abilities; second, memory for the what-where-when of a caching episode appears to be highly relevant for short-term hoarders that cache a broad spectrum of food; and third, socially competitive species seem also capable of remembering who saw them caching, and implement cache defense strategies accordingly.
Unfortunately, this picture is anything but complete, as we hardly know about the what-where-when memory of long-term hoarders or the observational capacities of birds with different breeding systems and social structures, respectively. In other contexts, things are even less clear. New Caledonian crows, for example, are the only corvids that habitually use and manufacture tools; still, other corvids like rooks seem to be on par with them with respect to tool selectivity and problem solving by analogy, at least when tested in captivity. This shows that cognitive traits may not always be an adaption to very particular socio-ecological conditions. Probably for feeding generalists, the key strategy is to evolve general problem solving capacities that are shaped by experience. Fortunately, with their relatively short maturational period of a few months, corvids are also great model systems for ontogenetical studies.
Why focus on corvids — aren’t you interested in the bigger picture?
Of course. One of my ultimate goals is to substantially contribute to such a broad, cross-taxon picture. My colleagues and I have several ideas for research on a broader taxonomic scale. I also welcome the increasing number of groups that are investigating the socio-cognitive skills of birds other than corvids. The problem is that we have to be extra careful when comparing different taxonomic groups. Standardized test batteries are great, but often preferentially tap into the predispositions of the species they were originally designed for. For instance, primates can use their hands or feet to reach into a tube; birds need to use their beak but may be reluctant to do so if it requires inserting their whole head. Sometimes subtle differences in motivation turn out to be critical. For instance, when ravens find playing with paper strips rewarding independently of whether they contain small pieces of food, this likely contributes to their chance performance in their choices between strips in a support problem task, even in the simplest configurations.
I should add that many of the problems encountered when comparing distantly related species may also apply to comparing members of the same taxonomic group or even the same species in different labs. Prosocial choices in chimpanzees, for example, seem to be highly sensitive to context and task. Even observational studies on different groups may reveal different results, as the post-conflict behaviour in chimpanzees has been shown to function as consolation for the victims of aggression in some groups, but as protection from redirected aggression in other groups. As flexibility is one of the key characteristics of higher cognitive systems, such variability should not be surprising. But understanding the contextual variation of given skills is everything else but simple, and seems likely to become a very hot topic in future research.
Could it be that laboratory data are biased in one or the other direction, and thus do not provide the ‘true’ picture of corvid intelligence?
Sure, but it depends a lot on what you mean by ‘true’ picture. Laboratory results give us insight into the mind of particular subjects tested under particular circumstances and with particular experimental histories. That’s a ‘true’ enough picture, as long as we are careful in not over-generalizing to the entire species or even broader taxonomic groups on the basis of a few results. Unfortunately, this is what is often done. For instance, if one out of ten captive ravens solves a particular problem in a complex task, we may conclude that it is in the range of cognitive capacities of ravens and thus of corvids; however, it says little about how relevant this skill is for ravens, or corvids in general, under daily life conditions.
How do you cope with this problem?
For me personally, the best way to keep laboratory results in context is to also investigate the patterns of interest in the wild. Studying animals under field conditions is quite challenging but it gives us a richer idea of when and how abilities are actually used, and very often inspires set-ups and further questions for the lab. That is why I am particularly proud of the advances made at our field sites, where we have access to a population of about 220 individually marked ravens and almost 300 marked crows now. But both field sites are in human-influenced environments, so some of my colleagues are rightly questioning the generality of the findings obtained under those conditions. So we are back to the question of what is a ‘true’ picture, in this case what is the ‘natural’ environment for highly generalist feeders and scavengers like crows and ravens.
You are a co-founder of the Department of Cognitive Biology: is studying animals in the wild what you mean with a biological approach to cognition?
No, studying animals under field conditions is only one aspect of our approach: we also do plenty of lab work. Our general aim is to foster comparative, evolutionary thinking in cognitive research. While much has been achieved in this respect in the last decades, we see ample room for improvement. Most notably, many theories about the evolution of ‘higher forms’ of cognition remain relatively vague, and many core concepts are biased towards primates and/or constrained by definitions based on human standards. As you have probably guessed, we also strongly support truly fair comparisons, not only between non-human animals but also when comparing the abilities of animals with those of humans. This means, for example, testing humans without verbal instructions. And, of course, we aim to contribute to the integration of different approaches and to the bridging of fields, in my case combining the powerful testing paradigms of psychology with standardized ethological observations under daily life, and also non-invasive physiological measures such as hormone metabolites in saliva or feces. So we intend to promote a very integrative and inter-disciplinary approach to studying cognition and its evolution.
Biography
Thomas Bugnyar is Professor of Cognitive Ethology at the University of Vienna, Austria. Trained in biology, he became interested early on in large-brained animals and the evolution of mind. He studies social behaviour and cognition in both captive and field conditions, focusing on corvids such as ravens and crows. His current work involves studies of social knowledge, networks and ‘politics’, social problem solving and social learning/tradition formation.