Title (eng)
Precursors to human theory of mind and action understanding abilities of dogs (Canis familiaris)
Author
Lucrezia Lonardo
Degree supervisor
Ludwig Huber
Description (eng)
PhD thesis - University of Veterinary Medicine Vienna - 2024 The full text is only available to university members. Please log in!
Abstract (eng)
The work presented in this thesis investigated pet dogs’ possible precursors to the human theory of mind and action understanding abilities. The comparative study of these topics can help shed light on the nature and origin of such abilities in humans. It has been suggested that, through a process of convergent evolution, dogs might have evolved socio-cognitive abilities functionally comparable to those of human children. My behavioural experiments, using a non-verbal change of location task, highlighted that dogs are sensitive to very subtle human behavioural cues connected to “knowing”. Indeed, in the context of a search game, they reacted differently to the misleading suggestion coming from an experimenter who knew the correct location of food compared to the misleading suggestion from an experimenter who was mistaken about the location of food. Unexpectedly, more dogs followed the misleading suggestion coming from the mistaken experimenter. However subtle the manipulation that dogs reacted to, this non-verbal task still involved a behavioural difference (whether the experimenter was present or absent in the room during the crucial displacement of food) between conditions, hence it cannot provide strong evidence of false belief understanding. Additionally, several controls are still missing in the canine literature, to rule out the multiple alternative explanations that have been proposed to explain the scarce positive findings of non-verbal, implicit false-belief tasks with infants and apes. For these reasons, while not excluding it, my study also does not strongly support the hypothesis that dogs might represent and attribute mental states to others in a way comparable to what human children do from about 4 years of age on. Instead, dogs’ behaviour in my study was consistent with that of younger infants and chimpanzees tested on “altercentric bias” tasks. Hence, it seems that dogs can at least be influenced by human mental states implicitly. With my eye-tracking experiments, I elucidated how dogs visually attend to object-directed actions performed by others, a foundational component of many ToM abilities. In a first study, I presented a conspecific and a human agent approaching and acting on a toy. I also varied whether the human agent moved normally or performed movements closer to dogs’ own motor possibilities. Examples of the first case are bipedal walking and kicking with a foot or grasping with fingers; examples of the second case are crawling and pushing away using an upper limb or grasping with mouth. Unlike primates (including humans) tested in previous studies, dogs did not predict others’ action targets by directly matching the observed movements on their own motor repertoire, but rather relied on visual cues that they probably learned to associate to salient action effects during their ontogeny. Their gaze arrival times to the target object seemed mainly driven by the saliency of the agent performing the action and were only rarely predictive (i.e., arriving to the object before the agent made contact with it). In a second study, I adapted for dogs Woodward’s paradigm. The results suggested that, unlike apes and humans infants, dogs might preferentially encode the spatial location where agents act rather the identity of the object they act upon. Finally, across my studies, dogs’ looking patterns showed sensitivity to the species and animacy of the agents shown in videos on a computer screen, with longer looking times to a conspecific than to a human agent and to a human agent than to an inanimate, though selfpropelled, object. Overall, these findings highlighted only a few similarities between dog and human behaviour, for example the tendency to “trust” misleading communication. Mostly, however, I found the processing of social stimuli diverged between the two species. Because these findings might be due to methodological challenges and to (not yet clarified) differences between dogs’ and primates’ processing of social stimuli, it might be premature to conclude that, at least in the perspective-taking and action understanding domains, evolution did not shape dogs’ cognition similarly to that of humans. From mine and previous findings, however, it is clear that dogs’ social cognition is extremely well attuned to reading human behaviour and more research is needed to test the convergent evolution hypothesis in the social domain.
Description (deu)
PhD Arbeit - Veterinärmedizinische Universität Wien - 2024 Aus rechtlichen Gründen sind nicht alle Teile dieser Arbeit frei zugänglich. Der Zugriff auf den elektronischen Volltext ist auf Angehörige der Veterinärmedizinischen Universität Wien beschränkt. Bitte einloggen!
Type (eng)
Dissertation
Language
[eng]
Persistent identifier
https://phaidra.vetmeduni.ac.at/o:4676
AC number
AC17679311
DOI
10.34876/tm5r-6b32
Number of pages
196
Date issued
2024
License
All rights reserved