Researchers in the United States evaluated the ability of dogs and toddlers to read human social cues and make social inferences about novel individuals and situations.
While dogs outperform human infants on some tasks, human infants outperform dogs on others.
However, the similarities in how dogs and human infants make social inferences are striking, suggesting that the capacity to communicate and understand social information may be a deep evolutionary adaptation that is shared across species.
Previous studies have shown that both dogs and human infants can understand gestures such as pointing, but the new research suggests that they also share the ability to use a variety of social cues to make inferences about the intentions of unfamiliar individuals.
In the first experiment, researchers showed 21 dogs and 36 toddlers a video of a familiar person opening and looking at a box, before placing it at one of two locations.
The children's average age was 22.5 months, while the dogs' average age was 3 years and 2 months.
Following this training, the toddlers and dogs were each tested on their ability to find the object behind a wall, after an adult had shown interest in the object, either through direct gaze, pointing or a combination of gaze and pointing.
Dogs performed significantly better in the gaze-plus-point condition, as did the toddlers, while both performed much worse when gaze and point gestures conflicted.
In a second experiment, the researchers investigated the ability of dogs and toddlers to understand the meaning of human facial expressions.
Both the dogs and human infants were shown pairs of photographs, each of a person either showing an angry or a happy expression.
Participants were trained via an iterative selection task to choose between pairs that were matched or mismatched for expression.
In order to test whether both species were interpreting the expressions correctly, the participants were shown a third set of images and asked to locate another image that matched the emotional expression (angry or happy).
Toddlers performed better than dogs did on the training phase of this task, but there was no difference in the species' ability to categorise the third set of test images.
This was true for both unfamiliar individuals, but not for familiar faces.
Finally, in the third part of the study, researchers tested whether dogs and toddlers could make use of multiple social cues (gaze and head orientation) to make inferences about the location of a hidden object.
Dogs and toddlers were significantly more likely to search for the object in the direction that the person's head was oriented, relative to the direction of the person's gaze.
There was no significant difference between the two species on this task.
Taken together, these results provide the first direct comparison of social cognition abilities between dogs and human infants, suggesting a deep evolutionary adaptation in social communication that is shared across species.
The findings are published in the journal Proceedings of the National Academy of Sciences.
