Source
Wu & Silva (2024; v3 2026), LIT-1032, read in NOTE-724. The paper’s §6 and Fig. 8 present successful coalitions as showing collaborative and adversarial actions together throughout their lives. NOTE-724 records it as claim C6, “lasting artworks persist under sustained overwriting, not in its absence”, with weak support.
What is asserted
That on a contested canvas, lasting is a matter of continued repair under attack, not of being left alone: as a population, the artworks that survive to the end are the ones whose coalitions kept repainting them while they were overwritten.
What the record holds toward it
THEORY-216, which is Active, holds what the log measured: early survival is barely predictable from four coalition features, coalitions consolidate, and some lasting artworks were held under sustained overwriting. The last is shown in ten chosen lifecycles. No population measure, such as the share of lasting artworks with sustained repair, is given.
What this does not say
- That persistence under repair is the only way to last. Falmagne et al. (THEORY-214) find some compositions safe through neglect on the 2022 canvas.
- That the recovered coalitions are the real ones. They are inferred from co-placement.
- Anything about why coalitions consolidate.
What would confirm or reject it
The promote_when. QUESTION-169 asks it together with whether interaction features restore early predictability. If it holds, CLAIM-280 has its case of persistence by contest and repair as a population fact, not an illustration.