workbench.02 — directional influence forensics
Under The Influence
An evidence-based scholarly workbench that scans, scores, aligns, and maps intellectual influence hypotheses across texts, chronology, and library contact paths.
Active Projects
No active projects loaded in storage.
Appendix A: Dynamic Scoring Settings
Evidence Channels (Positive Weights)
Direct Text Reuse25%
Long rare phrase contiguous n-gram matching overlaps.
Semantic Similarity (TF-IDF Cosine)15%
Concept word space similarity regardless of phrasing.
Explicit Citation / Mention15%
Mentions source author names or work titles in target text.
Access & Historical Plausibility15%
Access events: letters, borrowing records, shared schools.
Temporal Plausibility10%
Directional chronology: source was written/published first.
Other Modalities
Concept Similarity10%
Background Specificity5%
Epistemic Penalties
Common-Source Risk Penalty-20%
Lowers score if a third earlier text C explains both A and B.
Genericness / Commonplace Penalty-15%
Penalizes shared proverbs, biblical boilerplate, or clichés.
Composite Formulation Check: 100% Summed Capacity
Note: Penalties will subtract directly from the clamped positive score.