Following on my post about whether or not dogs have an image of self, Marji of For the Pit Bulls makes a great post about the problems of the “mirror test” and makes an excellent point about what senses a species uses. The mirror test, she points out, is biased in favor of species like humans that use vision as their primary sense. Dogs use smell.
Her post called to mind the behavior my dogs exhibit when we take their collars off after they’ve been wearing them a while. Once a dog here is trained enough that we can get them in the house from outside, on and off the furniture on demand, and interrupt any behaviors (like chasing cats) that are less than desirable by voice command rather than gently taking a collar to redirect the dog to a more appropriate activity, and once they’ve been microchipped, they live collar-free most of the time. Collars go on to leave the house, and come back off when we return.
Anyway, any time I have removed collars, the dogs will preferentially show a deep fascination for sniffing their own collars over the collars worn by another dog. It’s the same kind of fascinated Big Sniff that I get after coming back from visiting other dogs: nose pressed tight against the material, audible snuffs and sniffs, a whole-body sniffing experience. They don’t do it with the collars that the other household dogs wear, only with their own. This definitely suggests to me that Marji is right on the money, and if there’s a “smell mirror” then we might in fact find that dogs recognize themselves in it. But the concept of a “smell mirror” is so alien to people that we’d have trouble even designing the experiments we have used on, say, chimps, where we change the animal’s appearance with a small dot of cosmetics to see if they notice and touch themselves at the altered point.
Something fascinating to think about, anyway.