This project is public because its uncertainty should be visible too.

Mechanisms, conditions, and limits.

The project studies how information, behaviour, culture, and products move through people and systems. It asks what motivates transmission, how attention and networks create opportunity, how friction and incentives shape action, and why the same mechanism can succeed in one context and fail in another.

Each evidence card preserves the source identity, a precise locator, evidence type, claim scope, results, limitations, and provenance. It is an intermediate record, not a conclusion by itself.

Causal evidence is labelled only where the source design supports causal language. Observational evidence remains associational, even when its pattern is suggestive. Practitioner frameworks and anecdotes can be useful for questions and illustration, but are not treated as independent confirmation.

Platform-specific guidance is dated. Contrary results, null findings, and boundary conditions are retained—not edited away to make a cleaner story.

The public site links only to official, DOI, author, or lawful open-access locations. It does not host local PDFs, extracted passages, private acquisition details, internal notes, or agent outputs.

One real example of repair: a vector-plot artifact in a source text layer was found during ingestion. The extraction normalizer was narrowed, the batch regenerated, and the quality gates re-run. The repair is part of the evidence record; it is not a result about virality.