Pigs, for example, are very smart and can use mirrors as a kind of tool, but since they don’t care about having a mark on their faces (nor could they really do anything about it (no hands)) the mark test is basically inconclusive.
An old facebook thread had Eliezer argue that being able to identify others with mirrors is cognitively easier than self identification—that it only requires a visual transform on the sensory input instead of correlating your volition with the sensory input to deduce that you are looking at yourself. You also don’t usually see yourself. I think this argument is fairly likely correct—and I strongly disagree with his belief that some sort of cognitive reflectivity is a part of what consciousness/experience/qualia really is.
An old facebook thread had Eliezer argue that being able to identify others with mirrors is cognitively easier than self identification—that it only requires a visual transform on the sensory input instead of correlating your volition with the sensory input to deduce that you are looking at yourself. You also don’t usually see yourself. I think this argument is fairly likely correct—and I strongly disagree with his belief that some sort of cognitive reflectivity is a part of what consciousness/experience/qualia really is.
This is also relevant for Anthony’s comment about dogs and cats.