
Start by naming recurring entities like Person, Project, Source, Concept, and Claim. Then pick relationships that truly matter, such as authored, cites, refutes, supports, or depends on. Keep classes gentle: labels that guide rather than gatekeep. Avoid exhaustive taxonomies. Your ontology should earn trust by answering sharper questions, not by expanding spreadsheets. Let real research tasks determine what receives structure and what stays freeform.

Plain links connect pages; properties add direction. With a simple subject‑predicate‑object pattern, a note about a study supports a claim, and a person authored that study. You can express this as readable sentences, short properties, or lightweight templates. The goal is not perfect semantics, but enough structure to run useful queries, draw clear diagrams, and assemble arguments you can defend and revisit confidently.

Bring the ontology to life with repeatable patterns. A project collects tasks, questions, and sources. Each source links to people and key claims. Claims are supported or refuted by evidence. These patterns scale from solo research to collaborative writing. They also enable integrity checks, like spotting orphan claims without sources, and help you move from scattered highlights to coherent narratives ready for publication or presentation.