TL;DR
- 01Most ontologies assume an authoritative voice. Medical, legal, telecom. Policy and political work has none — the file is the disagreement.
- 02The TACITUS graph holds disagreement structurally: claim/counter-claim, evidence/counter-evidence, commitment/broken-commitment, narrative drift as a primitive.
- 03The system shapes the dispute. Humans decide what to do with it.
Medical ontologies assume an authoritative voice. Legal ontologies eventually do. Policy and political work has none — the file is the disagreement. The system's job is to shape it, not to dissolve it.
Why authoritative ontologies fail here
- ▸Medical: assumes one truth about a diagnosis.
- ▸Legal: a contract has one binding interpretation, eventually.
- ▸Telecom: a network configuration is or is not valid.
- ▸Conflict: the dispute is the disagreement. Forcing resolution at the schema layer is a category error.
How the kernel holds disagreement
Claim and counter-claim are paired objects. Evidence and counter-evidence are typed edges. Commitment and broken-commitment are bi-temporal events. Narrative drift is a primitive in its own right.
WikiContradict and ConflictBank operationalised the problem at the LLM layer. We operationalise it at the graph layer.
For practitioners
The system does not say "Sam was right" or "Alex was right." It shows: Sam asserted on Monday; Alex acknowledged ambiguously the same morning; Alex denied scope on Thursday — three turns, three bi-temporal stamps, one contestation edge. The mediator decides. The shape is the product.
SOURCES
