13.11 Residual failure and declaration scope

The admitted schema does not solve every confirmation game. The three unseen failures are all salt-shaker tasks; the three seen failures require finding a soap bottle, box, or newspaper. In each case the bottleneck is long target search rather than invalid manipulation. The accommodation has exposed a new obstruction one level down: \(\mathsf{LocateTarget}\) is too weakly structured.

This is how a persistent theory should grow. The failure does not motivate loosening the admission threshold after the fact. It motivates a new declared candidate family: receptacle priors learned from observations, an information- seeking search option, or a relational object–source model with its own counter-witnesses. The existing macro remains admitted where its contract holds; the localized stage becomes the target of a subsequent theory revision.