Mutual Information
I(X;Y) — shared information between a source and an output
Mutual information I(X; Y) quantifies how much knowing Y tells you about X — the information shared between a ground-truth variable and an artifact. Claude Shannon (1948) made it central to communication theory; Genesis uses it as the formal quantity trust composition preserves or loses.
I(Output; Ground_Truth) is what consumer classes (A–D) implicitly require at minimum. The data processing inequality bounds how I can change along a chain. Information efficiency η is the fraction of mutual information a step preserves.
Corpus stance
A1 — adopted on this site: Bridges information theory to the dependency matrix in The Facets Composed.