Here's a draft of something I've been kicking around for a while, now: Color-Consciousness Conceptualism [link].
The central idea is to explain away allegedly damning data concerning color discrimination. The data in question concern color pairs the members of which are distinguishable in synchronous but not diachronic presentations. This might be thought to give the conceptualist about conscious contents a problem. The thought seems to be this: Since the colors are visually distinguishable, they are clear candidates for conscious contents. However, concepts, especially concepts for most of the highly-determinate shades, seem to be learned items. So there's a close connection between concepts and memory and learning. And since the colors in question are indistinguishable across a memory delay, that puts pressure on the idea that conscious contents are exclusively conceptual.
In the paper I sketch a model that can allow the conceptualist to explain away the data. The picture accompanying this post gives a flow-charty gist of the explanation. The main elements involve positing nonconscious nonconceptual inputs whose down stream effects are solely conceptual. Crucially, solely conceptual contents in conscious experience plus solely conceptual contents in a short-term memory buffer give rise to the relevant discrimination judgments (or lack thereof).
Anyway, comments welcome!