Reference: Mental Representation
Ok. So. Concepts are taken to be images or representations, and the argument is whether they influence our sensory perceptions or not, and to what extent.
Perceptions are taken to be representational, but then the question is "representational of what?" (all representational states have their content in virtue of their phenomenal features -2012 Pitt, David SEP "Mental Representation")
I've been thinking of it all along in the reverse: that all perceptions are conceptual. I.e., that all images, ideas, impressions and sensorily states are token instances of the type "mental state." In this view, concepts are always inert. This is probably a view I've favored because of its easy one-to-one mapping onto a computational explanation of how concepts can manipulate and drive behavior*. But this presupposes that if there is phenomenality, then there is a perception had - which perhaps conflates perceptual experience and perceptual judgment. Moreover, if there is any phenomenality, I've come to assume that it expresses the "activation" of the relevant concept (again, just a token of a possible mental state). Thus the view requires that there be a concept to be had the first place, which baldly contradicts the accepted tradition.
That is, I see concepts as being implicitly representational, if in use - but only some concepts in use are privileged to be phenomenal. It's in this respect that I believe we can find a necessary condition for phenomenal states to penetrate our concepts. My condition for a concept to be phenomenally penetrable is that in a process of concepts representing one another, there must be an exclusive disjunction of possibly represented states which succeed the representing concept in that process.
The requirement of an exclusive disjunction is what leads me to favor sorts of explanations about our perceptions and experiences which incorporate Bayesian procedures in their operation.
Yet, seeing how I can only really claim this by turning tradition on its head, I have to ask myself if this could really be a good way to see things, or if instead I've made some horrible mistake. On the plus side, I now know I was making a definite mistake in how I believed other philosophers understood the problem.
I kind of wish I could rewind and replay the entire conference now.
* I'll need to revisit Sheldon's proposed changes to the File Model, given all this.