Functions of pointing by humans, and dogs’ responses, during dog-human play between familiar and unfamiliar players.

Although much research focuses on human index finger pointing to hidden items for dogs in experimental settings, there is little research about human pointing in naturalistic interactions. We examined human pointing to dogs during 62 dog-human play interactions, spanning 4.8 hours of videotape, to determine the functions of human pointing and dogs’ responses to that pointing. Participants were 26 humans and 27 dogs. Humans played with their own dog(s) and, almost always, an unfamiliar dog. Seventeen people (16 players and one passerby) pointed for 20 dogs a total of 101 times (once with a foot) during 26 interactions. Most (49.5%) points were toward an object (almost always a ball), to direct attention or action toward the object; 36.6% were to the ground in front of the (almost always familiar) pointer, directing the dog to come, and/or drop a ball the dog held, here; 10.9% directed the dog toward the designated player and/or play area; and 3.0% directed the dog to move away from a ball the dog had dropped. Humans almost always pointed such that the dog could see the point (92.1%), and pointed more with their own than with an unfamiliar dog. Dogs responded appropriately (i.e., did what the pointer requested) for only 24.7% of the visible points, more often for points to the ground than for points to objects. The proportion of dogs’ appropriate responses to visible points was similar for both familiar (30%) and unfamiliar (18%) humans. Six dogs who responded appropriately to some points resisted responding appropriately to others. Future research should examine non-object directed uses of pointing with dogs and their responses in naturalistic and experimental settings, and experimentally assess diverse explanations, including resistance, when dogs and other animals fail standard pointing tasks