This paper argues that the Reflexive Laboratory corpus can be represented meaningfully as a directed artifact graph at the artifact level. Rather than attempting a full semantic or claim-level representation, the paper shows that structuring relationships between papers, transcripts, and support artifacts is already sufficient to reveal lineage, reuse, and reconstructive organization in a transcript-based research environment. The paper defines core artifact classes and typed relations, introduces a directed graph formalism, and demonstrates the approach through a concrete example and subgraph analysis. It argues that artifact-level graphing is a viable methodological foundation for transcript-preserving research while explicitly bounding more ambitious claim-level graph programs as future work. This Zenodo package includes the full manuscript (DOCX and PDF), a simple example graph figure, and the original transcript artifact used in the research process.
Peter Bell (2026) studied this question.