Below there is my message to the everything-list (see Unconsciousness components):
http://groups.google.com/group/everything-list/t/1dadda0d2b84f605
It is tricky to prove consciousness from behavior. Yet, it seems sometimes to be possible under some mild assumptions. To this end I will briefly describe below an experiment that has been done to prove that a monkey has conscious visual perception (p. 69 in Jeffery A. Gray, Consciousness: Creeping up on the Hard Problem). By the way, I have found the original paper that he cites (pdf seems to be freely available):
N K Logothetis, Single units and conscious vision.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1692419/
See Fig. 4. “The monkeys were taught to pull and hold the left lever whenever a sunburst-like pattern (left-object) was displayed, and to pull and hold the right lever upon presentation of other figures (right-objects), including images of animate objects.”
The experiment is based on binocular rivalry. When each eye sees a different image, then brain cannot merge them into a consistent view. Rather a person experience in such a case either the first image or the second and the images changes periodically. Let me put it this way, the images on retina in both eyes are constant and different, but we experience not two images at once but rather they change periodically.
An assumption. If someone experiences binocular rivalry, then he/she has conscious visual perception.
The question is then how to prove that a monkey experience binocular rivalry. This has been done for example by training a monkey to press correctly different levels when it sees different images. As a whole it is tricky but looks reasonable. Well, this was just an idea and there is some more stuff in the paper.
Clearly one can develop a robot that will claim that it experiences binocular rivalry. Yet, this is in my view an another problem.