Dog-logic: inferential reasoning in a two-way choice task and its restricted use

Three experiments were designed to test whether adult pet dogs are able to show inferential reasoning when searching for their toy in a series of two-way choice tasks. The experimenter placed a toy under one of two identical containers and then provided some information by manipulating the covers: either both containers were lifted or just the empty or baited one. There were other trials when the experimenter not only revealed the corresponding container but manipulated also the other one without showing its content. In the second experiment the same conditions were used except that the content of the containers was revealed by strings without any human manipulation. Results of the two studies show that dogs are able to use inferential reasoning by exclusion (i.e. they can find the hidden toy if they have seen where the toy was missing). However, dogs were able to solve the reasoning task only when they could not rely on social-communicative cues (directional gesture and gaze cues) or could not use any other simple discriminative stimuli (movement of a container) for making decisions. This suggests that dogs are often prevented from showing reasoning abilities by pre-existing biases for social or movement cues. Results of the third experiment also support the primary importance of social cueing because in another object-choice task, individuals preferred to choose the ‘socially marked’ container (touching, gaze alternation) to the remotely moved one when they had no visual information about the location of the toy.