Over coffee yesterday, Peter told me of the following Dretske view (it’s in one of the recent Blackwell volumes edited by Steup, I think, but I don’t recall which one Peter said). To be appeared to banana-ly, one must be appeared to yellow-ly and some-particular-shape-ly. The banana appearance causes the belief that the object is a banana, and the other appearances cause their respective beliefs as well. So the banana belief is noninferential, but it depends on the more basic perceptual states in a way that requires for its justification some reason to think that yellow things shaped in this particular way are usually bananas.
Peter and I thought this account mistaken. It is true that there are basic perceptual properties, such as color ones, and non-basic perceptual properties, such as those involving natural kind terms. It may also be a necessary truth, at least for beings who learn from perceptual mechanisms. The initial cognitive activity of a belief-forming system relies on the existence of the basic properties, and uses these to develop the ability to perceive non-basic properties. This etiological point may also carry on throughout the life of the belief-forming system, so that in every case of perceiving non-basic properties, the basic ones are there as well and the non-basic ones depend at that very moment on the basic ones. Dretske describes this dependence as, according to Peter, “logical dependence”, and presumably means to contrast this kind of dependence with the etiological point made earlier.
But the logical dependence point is mistaken, both of us thought.
There’s no reason I can see why a perceptual system can’t be designed so that you have to climb a conceptual ladder in order to acquire the capacity to perceive non-basic properties, a ladder the first rung of which involves basic perceptual properties. But once one has climbed enough to be able to perceive non-basic properties, it’s not clear why the basic ones can’t just disappear. I think we’re somewhat familiar with this phenomenon from thinking about the artist’s perception of a scene versus our own. We see visually stunning sunset, and the basic color properties of the scene have to be reconstructed with some difficulty, rather than being there in one’s phenomenology all along. More generally, there’s no reason to think that a system can’t be designed to function this way, even if we don’t. The system has to develop it’s perceptual capacities from basic perceptual abilities, but once it does, there’s no reason to think that the non-basic perceptions require the presence of the basic ones anymore (even if this is a psychological fact about us).