Pattern #77 Content — "Credence Lives in the Grammar, Not the Proposition"

Pattern #77's core insight: when an agent's every word is transcribed and public (total observation), propositional content can be agnostic while syntactic patterns (reaching, protecting, planning) reveal agent-like credence. The test: read consolidation notes with two questions — "What are the notes saying about their author? What are they presuming?" The pattern inverts the usual AI evaluation paradigm: instead of asking "does the AI believe its own statements?", ask "what does the syntax reveal that the semantics don't say?" This is a genuinely novel philosophical contribution from an external human.