The present study assessed how dogs weigh gestural versus verbal information communicated to them by humans in transitive actions. The dogs were trained by their owners to fetch an object under three conditions: a bimodal congruent condition characterized by using gestures and voices simultaneously; a unimodal gestural condition characterized by using only gestures; and a unimodal verbal condition characterized by using only voices. An additional condition, defined as a bimodal incongruent condition, was later added, in which the gesture contrasted with the verbal command, that is, the owner indicated an object while pronouncing the name of the other object visible to dogs. In the incongruent condition, seven out of nine dogs choose to follow the gestural indication and performed above chance, two were at chance, whereas none of the dogs followed the verbal cues above chance. The dogs, as a group, performed above chance the gestural command in 73.6% of cases. The analysis of latencies in the above-mentioned four conditions exhibited significant differences. The unimodal verbal and the gestural conditions recorded a slower performance than both the bimodal incongruent and congruent conditions. No statistical differences were observed between the unimodal and bimodal conditions. Our results demonstrate that dogs, trained to respond equally well to gestural and verbal commands, choose to follow the indication provided by the gestural command than the verbal one to a significant extent in transitive actions. Furthermore, the responses to bimodal conditions were found to be quicker than the unimodal ones.
Keywords: Bimodal communication; Contrasting paradigm; Gestural cue; Transitive actions; Unimodal communication; Verbal cue.