Would dogs copy irrelevant actions from their human caregiver?

Abstract

Companion dogs learn easily from humans, including human behavior, human communication, and some aspects of the human-made environment. They benefit from having the opportunity to learn from humans and are able to spontaneously synchronize their behavior with that of their caregiver. Here, we tested whether pet dogs would show a special form of observational learning, one that has been considered uniquely human. Indeed, humans show overimitation, the faithful copying of causally irrelevant actions, but great apes do not. Because in humans this peculiar form of imitation is strongly motivated by social factors, such as affiliation or conformity, we hypothesized that domesticated and enculturated dogs are more likely than apes are to copy such actions, especially if shown by their affiliated caregiver. Indeed, half of the dogs replicated a causally irrelevant action that was demonstrated by their caregiver, and about the same number did this whether they saw only this action being demonstrated or being demonstrated before or after a causally relevant, functional action. The demonstration of a causally relevant action, one that is immediately followed by access to food, thus does not inhibit the copying of an action that is spatially separated and functionally opaque. Given that the copying frequency in this study was low overall, these results suggest evidence for overimitation in dogs, which might challenge the human uniqueness of this type of social learning.

Keywords: Affiliation; Attention; Companion dogs; Overimitation; Social learning.

Fig. 1.

Experimental setup during attention test and imitation test. AT = attention test; REL = causally relevant action; IRR = causally irrelevant action; D = dog; E = experimenter; O = owner; C1, C2, C3 = video cameras. Solid arrows show the movement trajectory of the owner when demonstrating the IRR+REL action sequence. Dotted arrows show the distances between the dog waiting at the starting position and the objects in the imitation test (see Method section)

Fig. 2.

Two targets of the imitation test. Left: Sliding door affording the causally relevant action (opening). Right: Paper wall with color dots affording causally irrelevant actions (nose touching)