Title: The Domestication of Social Cognition in Dogs
Authors: Brian Hare, Michelle Brown, Christina Williamson, Michael Tomasello
Published: November 2002 in Science
These authors conducted four experiments to explore three hypotheses about how domestic dogs may have acquired their adept skill at using human social cues: 1) canids (wolves, foxes, coyotes, dogs and etc.) are adept at exploiting the behavior of both conspecifics (members of the same species) and other species due to hunting requirements of their diet. This hypothesis predicts that many canids would perform as well as dogs on tasks involving social cues from humans. 2) Domestic dogs have much more experience with humans and their social cues due to dogs/humans’ relative closeness from a very young age. This hypothesis would indicate that dog puppies of advancing age would performing increasingly well at tasks involving human social cues due to their cumulative knowledge of living with humans longer. 3) Dogs have undergone a selection pressure during domestication for specific skills involved with using human social cues. This hypothesis predicts that dogs and puppies (over a certain age) would perform equally well on tasks requiring human social cues and would outperform their nearest relative, wolves.
The authors conducted four experiments to explore canid skills in exploiting human social cues for food rewards (with control experiments to address olfactory food detection). In the first experiment, the researchers tested 11 dogs and 11 apes on the ability to correctly choose a container that was gazed at, pointed at and had a small marker placed atop. Dogs were significantly more skilled at this task, demonstrating that dogs are better at exploiting human social cues than humans’ nearest relative.
In the second experiment, 7 dogs and 7 human-reared wolves were tested on the ability to correctly choose a container with a single or combination of the following human social cues: gaze, point and/or tap. Dogs generally outperformed wolves, although the wolves did appear to understand a combination of the cues.
In the third experiment, 5 dogs and 5 human-reared wolves were individually shown food being hidden in a container and then later had to pick the container from memory. Dogs and wolves performed equally well in this experiment.
In the fourth and final experiment, the authors tested 32 dog puppies between the ages of 9 and 26 weeks. The puppies were tested on their ability to pick out a container based on a gaze and point or just a gaze by a human. Some puppies had been reared from a young age by humans while others had been raised in a kennel situation with limited human exposure. There was no difference between age or rearing type on either cue.
From the last three experiments, the authors conclude that their results suggest that dogs’ ability to exploit human social cues was due to selective pressure during the domestication process (hypothesis 3). Because puppies of all ages and human exposure type performed well on tasks using human social cues and while dogs and wolves performed well in the memory task (experiment 3), dogs outperformed wolves at social tasks, the authors found a significant social-cognitive difference between wolves and dogs that suggests an evolutionary explanation rather than the general social skill of canids (hypothesis 1) or a human exposure source (hypothesis 2).
As you may have been able to tell, this article was chock full of interesting experiments and results. Science is a very prestigious journal with one of the highest impact factors of all academic journals. In other words, you don’t get published in Science unless you’ve done some pretty fascinating work!
Although their sample sizes are small (10 to 32 test subjects), these authors make a compelling argument to support the domestication theory of dogs’ ability to read human social skills. Of course, in order for this theory to be fully accepted, these experiments will need to be reproduced (and likely already have, since this research is 12 years old). Unfortunately, given the number of experiments these authors were explaining and the limited word count of Science, not a lot of space was given to elucidating the experimental conditions or even the subjects (except for the puppy experiment where ages were given, the only thing I know about the animals tested was their species). Of course, future researchers can always contact the corresponding author for this information if they’d like to replicate the science.
In my view, this research has very interesting implications for canine welfare. In many areas of the U.S., dogs are legally protected as property, with the exception that owners must provide adequate shelter, food, water and veterinary care. Basically, in our legal system, a dog is a box*.
However, if dogs are cognizant of human social skills, is it humane to treat a dog as a box*? After all, if you scream at a box, no harm is done. If you scream at a box*(i.e. a dog), and the dog feels stress from that social cue, has harm been done? Where and how does a creature’s cognitive abilities play into its welfare requirements?
(*that you can’t outright neglect or abuse)
These findings are super applicable for training purposes. Many common dog training methods involve pairing a human-given hand or verbal cue with a behavior (i.e. a sit) from the dog. This research suggests that this type of training is only possible through the selective pressure that gave dogs skilled at exploiting human social cues an evolutionary edge. Consider that in contrast to training a cat: the same hand/verbal cues may not be as meaningful to felines (who by all accounts never underwent the same domestication process as dogs) than our canines!