Kat

Potential change points compare rolling three-year thematic distributions with Jensen-Shannon divergence. They are model-derived boundaries, not authoritative divisions in the author’s intellectual biography.

Primary-theme assignments are moderately concentrated. The most recurrent primary themes are Law and civil rights, Transgender identity and history, Family and relationships.

Lexical statistics are descriptive properties of the available texts. They do not diagnose writing quality, intelligence, personality, or intent. Distinctive terms are compared with a sampled Collective baseline using smoothed log-odds.

Source families are inferred from link domains and URL patterns. They describe the ecology of linked evidence and references, not the reliability or ideological position of an individual source.

Co-authorship, corpus similarity, citation relationships, and publication overlap are shown as distinct relationship types. A corpus relationship does not imply a personal, institutional, or political association.

Reception measures documented circulation and reuse across the evidence sources available to the Collective. They do not establish quality, correctness, cultural worth, or the author’s social standing. Provider totals can overlap.

Comparisons use the rest of the registered Collective as the default baseline. Percentage-point differences and standardized effects describe corpus contrasts; they are not rankings or evaluations of authors.

Coverage measures recurring places, institutions, historical periods, scopes, and population terms in the registered texts. It describes what is present in the corpus and does not judge omissions or infer the author’s identity or attitudes.

These categories describe structural positions in the registered corpus. They do not establish scholarly quality, correctness, or cultural value.