The effect of early experience on physical cognition in dogs

Intense research of the last years has shown that dogs have evolved specialized skills for reading human social and communicative behavior during the process of domestication (e.g. Viranyi et al., 2006; Szetei et al., 2003; Gacsi et al., 2004; Viranyi et al., 2004; Miklosi & Soproni, 2006), but usually perform rather poorly in the physical domain (e.g. Watson et al., 2001; Collier-Baker et al., 2004; Osthaus et al., 2005). Altogether, these results suggest the lack of abstract representations and inferential reasoning and at the same time a dependency on human-social cues in dogs (Topal et al., 1997). This dog-specific cognitive evolution is especially interesting if dogs are contrasted with great apes, which have evolved powerful inferential skills to solve problems in the physical world associated with finding food (Call, 2004; Call, 2006). However, apes tested for these kinds of studies usually have extensive experience with physical tasks, whereas dogs usually have only experience in the social domain. Thus, it remains unclear which of these differences between primates and dogs are due to experiences during development rather than real species differences. More importantly the questions arises how much experience can influence cognitive abilities within a certain domain. Flexible and creative behaviour in the physical domain requires the ability to distinguish between functional and non-functional properties of objects, to anticipate the effects that objects have on other objects and the acquisition of physical concepts serving as large functional categories (like gravity, force and connectivity). However, it is unrealistic to expect any individual to be capable of solving entirely novel problems with objects or materials they have never had any experience with. In order to investigate the influence of experience on physical cognition in dogs, we intend to systematically enhance the exposure to problems in the physical domain in a group of dog puppies, whereas a control group is only exposed to simple manipulation problems that do not provide information about physical properties. The adult dogs will then be tested with a number of problems where dogs have failed before (e.g. means-end connection, causal understanding) or have never been tested in (e.g. trap tube). In order to control for the effect of human cues, we conduct these latter tests in the absence of human cues. This design will allow us to investigate our main question of the influence of experience in the physical domain on problem solving abilities in the adult dogs. Testing additional adult dog groups with either manipulation experience or not, we will further ponder the questions of the influence of i) manipulation experience, ii) the presence or absence of a human during the experiment, iii) the development of physical cognition in dogs and iv) dogs` physical cognition in several tasks never tested before with this species (e.g. connectivity, gravity). This project is intended as continuation of a previous European-funded project on the evolution, development and intentional control of imitation (EDICI). It would also benefit from collaborations with the Department of Ethology at the Eötvös Lorand University in Budapest (Dr. Miklosi) and the Institute of Developmental Psychology of the Hungarian Academy of Sciences (Prof. Gergely).

The cognitive abilities of animals have become a focus of research in the past decade since they can shed light on the evolution of our own abilities. Over this period, large differences between species have come to light, in particular in the area of physical cognition (the understanding of regularities in the physical world such as gravity, support or connectivity). To date, it remains unclear whether these differences are mostly a consequence of different selection pressures over the millennia, or whether they can be explained by different experiences that animals typically gather throughout their lives. The aim of this project was therefore to investigate which factors influence the abilities of animals in the domain of physical cognition.When compared to primates, dogs typically perform poorly in physical cognition tasks. With this project, we therefore set out to investigate whether this poor performance of dogs is explained by their limited opportunities to learn about physical rules in their daily lives. For this purpose, we supplied a cohort of dogs between the age of 3 and 18 months with a series of intelligence toys, which gave them opportunities to learn about causal relationships in their environment. Subsequently, these dogs were tested with a battery of 6 physical cognition tasks and compared with a control group of dogs that had not been provided with these extra learning opportunities. Even though the dogs became very proficient with most of the intelligence toys, this did not influence their performance in the physical cognition tasks. We can therefore conclude that, at least for dogs, performance in physical cognition tasks is little influenced by experiences and that dogs instead appear to learn to solve each task from scratch. In an extension of this project, we could show that the individual problem-solving abilities of dogs is little influenced by personality traits or their attachment bond to their owner, but that it is influenced by the dogs level of inhibitory control. Since this project involved following a cohort of dogs over a period of two years, we took the opportunity to investigate also the development of individual differences in behaviour, or personalities, from the first weeks of life until early adulthood. With these data, we could show that early temperament or personality tests performed in the first week of life or at the age of six weeks have little predictive value for individual differences in personality at the age of two years. In addition, we could show that certain personality traits remain stable after the age of six months, whereas other traits can still change markedly in the second year of life.