The Research Architecture: How a System Can Know What It Knows

Version 2. Updated structure and clarifications. This version supersedes v1.A research programme involving multiple human and AI participants faces a structural problem: how does the system know what it knows? This note argues the answer requires an architecture — a structured agent configuration (multilogue), an operational logic for managing divergence (isomorphism of CDSA and orchestration), and a stratified memory system (epistemological stack). All three components operate by the same schema.