So, if you have been around here lately you will have noticed that I have been talking a lot about priming, change blindness and the function of conscious mental states in the higher-order theory. I have been arguing that some recent results on priming effects in change blindness suggest that there is some function for conscious mental states (even/especially for those who like higher-order accounts (of whatever type). David’s response to this has been to admit that this shows that there is some functionality for conscious metal states but then to insist that it is not enough to justify calling it ‘the function of consciousness’ or anything like that. He then points out stuff like this article and argues that change detection is pretty big stuff, maybe even the stuff that you thought might turn out to be The Function of Consciousness but even that can be done unconsciously.

But after thinking about this, I am not sure that the Fernandez et al stuff really shows as much as David thinks that it does.  So, consider the experiemtn that Fernandez et al did as summed up in the figure below (from their paper).

The only difference between the two pictures is whether one sees George or Not-George. Subjects then see figure b and are forced to guess which of the two highlighted bars was the one that changed. The study reports that people pick the correct one even though they say that they did not see the change.

But notice that in figure b subjects are presented with Not-George. They did not check to see what would happen if they presented subjects with George and asked the same question. Mow, though they didn’t do this, the Silverman-Mack experiments predict that George should have been just as good at allowing subjects to perform above chance. This would suggest (it seems to me) that, though the subjects are conscious that there is a difference, they are not conscious of what the difference consists in. When they are conscious of the difference as the difference (that is, when the consciously see the difference) the Silverman-Mack results predict that only Not-Geroge would show any effect. The representation of George would be supressed. So the kind of change detection that happens consciously serves a distict function from the kind that happens unconsciously. Conscious change detection serves to bias the system; inhibiting some representations and thereby enhancing others, unconscious ones don’t. This biasing is important for survival since it helps to determine which representations can be assesed for action (like button pushing) and so this is a function for perceptual consciousness that is pretty important.

Stay tuned…there’s bound to be more of this after the big talk tomorrow!