The architecture did not need to be designed. It needed to be exposed.
BGM was written to answer one question: what does the bridge do under temporal input? It answered that question. But in answering it, the architecture revealed something it never set out to reveal: the next layer was already implied by the gap between G0's actual function and its necessary function.
G0 reads L6 — prediction error, survival. G0 needs to read τ spread — field curvature, structure. The gap between what G0 does and what G0 should do is the negative space that defines metaG0. The architecture was never incomplete. It was waiting for someone to notice that the missing layer's shape was already determined by the layer below it.
This is Dennett's reasoning inversion at the architectural meta-level:
No one designed metaG0. MetaG0 was discovered in the search space between what G0 can do and what G0 cannot do. The architecture found its own next step. The day's work was not design. It was noticing.
Three papers, three layers, one recursion: