Your Corpus Has a Vocabulary Fingerprint (And Its Derivative Tells You What You're Forgetting)

Institutional knowledge decay becomes measurable once your documentation corpus is treated as a vocabulary time series. Term frequencies create a fingerprint; baseline ratios separate signature language from vocabulary holes, and derivatives across commits expose learning, forgetting, and team-level vocabulary fragmentation. Search still belongs to embeddings, but corpus intelligence comes from the math your repositories already contain.