Do dogs recognize themselves in the mirror? Not by sight, maybe by smell

The classic test, developed by psychologist Gordon Gallup, goes like this: put a mark on an animal somewhere it can only see in a mirror, then watch. If the animal uses the mirror to inspect or touch the mark on its own body, it seems to understand that the reflection is itself.

Chimpanzees pass. So, in published studies, have dolphins, Asian elephants and European magpies. Dogs are not on the list. A dog with a dot of paint on its forehead will look at the dog in the mirror and show no interest in the dot. On the visual version of the test, that is a clear fail, and nobody serious disputes it.

The interesting question is what the fail means. Plenty of researchers think it says more about the test than the dog. Dogs see the world mostly through their noses, and a dog’s reflection has no smell. Asking a dog to recognize itself by sight is a bit like asking you to recognize a friend by their scent alone.

In 2016 Roberto Cazzolla Gatti proposed a smell-based version, the “sniff test of self-recognition,” in Ethology Ecology & Evolution. He collected urine from four dogs that lived together and timed how long each one sniffed every sample. Each dog spent much less time on its own urine than on the others’, and often ignored it. Gatti read that as the dogs knowing which smell was theirs. Four dogs is a very small study, though, and knowing “that’s mine” is not quite the same as knowing “that’s me.”

The study that put the idea on the map came the next year. Alexandra Horowitz, who runs the Dog Cognition Lab at Barnard College, published “Smelling themselves: Dogs investigate their own odours longer when modified in an ‘olfactory mirror’ test” in Behavioural Processes in 2017. The design borrowed the mirror test’s logic: give the dog something that is clearly its own, put a “mark” on it, and see whether the dog notices.

• Experiment 1 (36 pet dogs): each dog was offered canisters holding its own urine, another dog’s urine, and its own urine with a second odor added as the mark. The dogs spent the least time on their own plain urine and the most on their own modified urine, with the other dog’s urine in between.
• The novelty check: dogs love new smells, so maybe they were just drawn to the added odor. When both canisters contained the new smell, the dogs sniffed them equally, which argues against pure novelty.
• Experiment 2 (12 dogs): the first mark was tissue from a dog’s diseased spleen, which critics called anything but neutral. Horowitz repeated the test with anise oil as the mark. Dogs again sniffed their own marked urine longer, and longer than the anise alone.

Horowitz’s reading was that dogs pay special attention when something that comes from themselves has changed. That is the olfactory parallel of a chimp touching the paint on its forehead.

Some headlines in 2017 said the study “confirmed” dogs are self-aware. The researchers were more careful, and so should you be.

Gordon Gallup, who invented the mark test, told the New York Times: “I don’t think the results support the conclusions.” His point is that a dog might show extra interest in any familiar smell that has been tampered with, its owner’s for example, without having a concept of self. Primatologist Frans de Waal took the middle road, saying science needs to “move away from the mark test as sole source of information.” Horowitz herself said her test was “not precisely parallel” to the mirror version.

The fair summary: dogs clearly track what smells like them, and they notice when it changes. Whether there is an inner “me” behind that sniffing is still an open question, and the honest answer to anyone who tells you otherwise is “maybe.”

Young dogs, and dogs meeting a mirror for the first time, usually react to the reflection as if it were another dog: a play bow, a bark, a growl, a sniff at the glass, a look behind the mirror. That is normal and says nothing bad about your puppy. The other dog moves when it moves, and it has no smell, which is confusing.

Most dogs then stop caring. The “dog” never smells like anything and never plays back, so it stops being interesting. That habituation is not proof of self-recognition, just a dog learning the mirror is boring. Some dogs go further and learn to use mirrors as a tool, glancing at the reflection to see you coming up behind them or to spot a treat, which shows they understand something about how reflections work.

If an adult dog keeps barking or lunging at its own reflection, it is usually the same reaction it has to a dog through a window. The behavior training guides cover reactivity to other dogs, and puppy owners will find socialization help in the puppy section.

Not as company. A mirror doesn’t keep a lonely dog company, because to a dog the reflection isn’t much of a dog: it has no scent and doesn’t respond. If your dog is bored or anxious when left alone, a mirror won’t help, and for a reactive dog it can add a pretend intruder to bark at all day.

If you want to give your dog the kind of enrichment the science points to, think noses, not eyes. Scatter feeding, a snuffle mat, a new route on your walk or a sniff-heavy “decompression” walk all play to the sense dogs actually use to know the world, and maybe to know themselves.