After glancing at the people in the photo above, what did you think? Perhaps you just thought 'four people', or 'older people', but it's likely if I asked you to describe it you would say, 'two men, two women'. As humans, we love to categorize things. Upon meeting a new person you're likely to put them into a category: perhaps old versus young; tall versus short and most likely, male versus female. Although categorizing people too soon can often lead to unfair or incorrect judgements, the human tendency to categorize is actually a very useful cognitive mechanism. It makes it easier for our brains to process large amounts of information and apply things that are learned in one context to a different context.
Although it seems likely that other animals share this same ability with us, few studies have addressed how other animals' categories might compare to our own. As we've bred dogs to be able to cooperate and communicate well with us, it seems plausible that they might share some of the same categories that we use.
It's already well established that dogs do use categories: they can categorize dog versus non- dog sounds, images of dogs versus landscapes and images of dogs versus other animals. However, scientists wanted to see if they used categories specific to humans. In particular, they wanted to see if dogs categorized people by male and female, by using the two main cues we attend to: visual appearance and voice.
To test this, dogs were sat opposite a speaker, on either side of which stood a man and a woman. The sound of a man or woman's voice was then played through the speaker, and the scientists looked to see whether the dog looked towards the man or the woman. The idea behind this was that if dogs did combine the visual and vocal cues from humans, then they would look at the man when hearing a man's voice and at the woman when hearing a woman's.
Using a total of 51 dogs of 17 breeds, the scientists found that some of the dogs looked at the person of the gender associated with the voice, whereas other dogs did not. Dogs that had come from homes where there were three or more adults were much better at looking at the 'correct' person than dogs that had come from a home with a single owner. This makes sense, as these dogs would have had more opportunity to learn to associate male-type voices with male-type visual cues, and similarly for female-type cues.
Of course, we cannot be sure exactly which visual and auditory cues the dogs attended to. For us, the depth of a person's voice is a good indication of their gender, but men and women's also differ in their vocal intonation when talking to dogs, so it is possible that the dogs used this. Similarly, dogs could use any number of cues to categorize human gender when the person is standing in front of them, including body size or weight, or odour. What this study shows is that dogs can associate at least one aspect of human voice with one cue that can be detected in the presence of a human, that both relate to human gender in some way.
Four people: Eleni Papaioannou
Dog diagram: taken from Ratcliffe et al. (2014)
Ratcliffe, V. F., McComb, K., & Reby, D. (2014). Cross-modal discrimination of human gender by domestic dogs. Animal Behaviour, 91, 126-134.