Biology session

The ability of word comprehension is mostly believed to be closely related to the ability of speaking, and thus attributed to humans only. However, in this last decade, a few examples of outstanding minds were revealed among dogs, who became famous for knowing and distinguishing hundreds of human words referring to objects. These dogs proved – in rigorous experimental conditions – to understand indeed the connection between a word and its referent, by attributing object reference to the spoken words. Nevertheless, apart from these few, highly intelligent dogs, we have no information about the ability of dogs in general regarding word comprehension – are they some special dogs with exceptional abilities, or is this an ability being present in dogs in general?

Proceeding from these previous studies, the aim of this study was to receive a more comprehensive conclusion about dogs’ ability to comprehend human words as verbal referents, by – as a first step – testing a larger sample of usual pet dogs for word-object recognition. The criterium for dogs participating in the study was to confidently know the names – according to their owners – of at least 2 objects/toys and be able to fetch them by their names. The experiment consisted of two parts: a fetching and a looking task, where dogs were supposed to fetch or look at the named object. We controlled for human social cues available for the dogs to rule out the possibility of the ‘Clever Hans’ effect.

Results show, that the knowledge of these word-object connections is clearly not stable in these dogs. On the individual level most of them were not able to choose the correct object significantly above chance level in the experimental conditions. This shows, that for owners what often seems like understanding reference, is rather the result of reading human social cues and association with a certain context, both of which were controlled in the experimental conditions. On the group level however, dogs (N=14) chose the named object significantly more often in both the fetching (p=0.006) and the looking (p=0.013) task. This could indicate that dogs might have indeed the ability to understand reference. To become susceptible to learn object names is not their obvious capability, but with the proper training, in context-independent situations, they could have the chance to become quite aware that words can refer to objects, and their ability to understand reference could be manifested.