[The Lone Ranger]

I'd have to give it some further thought, but my first thought would be to employ a one-way mirror. An incompletely-silvered mirror functions as either a window or a mirror, depending on the lighting conditions.

Ideally, you'd need such a mirror/window set into a wall, with a dog on the other side. That is, you'd want to make sure that there were no olfactory or auditory cues to alert your dog that another dog is nearby, if possible.

Then, by altering the light level on either side of the mirror/window, you could make it so that your dog sees either the other dog or itself, and observe the resulting behavior.

That's what comes up off the top of my head, anyway.



The null hypothesis would be that your dog does not distinguish between an image of herself and an image of another dog, I should think. If she clearly does respond differently to images of other dogs compared to images of herself, that could indicate self-recognition.

That's actually a very widely-studied phenomenon -- whether or not an animal is capable of self-recognition. One way to test self-recognition is to put a dab of paint on a part of an animal's body that it can't see, and then let it look at itself in a mirror. Most animals won't make a connection between the mirror image and themselves, but a chimpanzee will recognize that the image in the mirror is itself, and will immediately begin picking at the discolored fur.

Such self-recognition has been convincingly demonstrated in many primate species, in cetaceans, and in elephants. To my knowledge, it has never been demonstrated in dogs, however. This may be because most dogs are much less visually-oriented than are most primates, however.