TACITUS NOTES
Research notes from the team. Working ideas rather than finished arguments — the kind of writing you usually only see in private notebooks. Published as they are, so readers can spot the problems early.
PUBLISHED · 5
newest first
A schema tells you how to store data. An ontology tells you how to argue about it. Conflating the two is how you end up with yet another JSON standard and no shared grammar.
Three structural failures: time, causality, provenance. Each one a property of transformer architecture, not a tuning problem.
Claims are what a party says. Commitments are what they are on the hook for. Conflating these is how careful negotiations become careless ones.
Fisher and Ury split positions from interests in 1981. Forty years later, almost no software implements the distinction. Here is what happens when you do.
We label our products experimental because they are, not as a legal shield. Here is what we mean by it and what we ask of early users.
IN DRAFT · 7
shipping soon
Events do not happen in a straight line. They branch, overlap, and causally relate across branches. A timeline is the wrong data structure.
Power shows up in every dispute, and most conflict-resolution literature treats it as an embarrassment. The ontology does not have that option.
Disputes change shape over time — not because facts change but because the story around the facts does. Tracking drift is its own problem.
A good contradiction detector is not a better string matcher. It has to model what each actor was committed to at which point in time.
Inter-annotator agreement on Interest is always going to be messy. Here is how we deal with that honestly.
Mediation is a human act. Software that pretends otherwise is making a category error. What TACITUS does is not mediation — it is legibility.
A single dispute, walked end-to-end: ingest, extract, structure, query. Every claim in the output traced to its source span.
Seen something that should be a note? Written one yourself?
hello@tacitus.me