WHY CONFLICT · WHY NOW
Parties in a dispute are not usually irrational. They are operating on different pictures of the same situation — different timelines, different causal stories, different readings of who committed to what. Strip the asymmetry away and most arguments dissolve; what remains is a hard decision, not a deadlock. TACITUS is the infrastructure that surfaces the structure underneath the disagreement, so the hard decision is the thing humans actually work on.
Problem
Parties see different pictures of the same situation.
Response
One shared, typed, sourceable map of the dispute.
Boundary
We do not decide who is right. Humans still do.
We are not trying to end conflict. Conflict is the default state of any system with more than one agent and more than one goal; pretending otherwise would be dishonest. Whether to fight, to compromise, to walk away — that is human agency, and it stays with humans.
What TACITUS does is more modest and more concrete: we show every party the same map. Decades of research across law, political science, organizational behaviour, mediation theory, and moral psychology converge on a shared finding — disputes have structure. Actors, claims, interests, commitments, events, narratives. That structure is learnable, and until now it has been invisible.
Current AI tools miss this entirely. Large language models are remarkable at fluent text but structurally unable to reason over time, causality, and provenance — the three things a conflict actually consists of. Ask a generic LLM 'who broke which commitment, when, and under what pressure?' and you get a confident hallucination. The failure is not a bug; it is an architectural property of transformer models.
Dialectica, the neurosymbolic engine underneath TACITUS, closes that gap. It extracts eight primitives — Actor, Claim, Interest, Constraint, Leverage, Commitment, Event, Narrative — from any dispute text, stores them in a typed knowledge graph, and lets language models reason against deterministic structure instead of inventing it. The result is conflict legibility: a shared, sourceable, time-ordered picture that every participant can see, argue with, and update.
Think of it as a better pair of glasses. TACITUS does not tell you what to do about a fight. It lets you see what the fight is actually about.
We do not make peace. Peace is a human choice. We make disagreement legible enough that humans can make it.
THE EIGHT PRIMITIVES · ACO
Every conflict — workplace, commercial, diplomatic, governance — decomposes into these eight. The vocabulary changes; the structure does not.
Who participates
Individuals, teams, institutions, states, coalitions — any agent with goals and agency in the dispute.
What is asserted
Factual, evaluative, or normative statements a party makes. The surface of the disagreement.
What lies beneath
Underlying needs and motivations. Fisher-Ury: what the party actually needs, versus the position they state.
What limits outcomes
Rules, regulations, norms, hard facts. The structural bounds on possible resolution.
What holds the power
Resources and dependencies that give a party the ability to act, withhold, or influence.
What was agreed
Promises, contracts, pledges, ceasefires. The graph-layer distinction between asserted and agreed.
What happened when
Time-stamped, provenance-carrying occurrences. The temporal DAG underneath every case.
How it is framed
The story each party tells about the dispute. Captures framing, re-framing, and narrative drift over time.
WHO IT IS FOR
HR leaders
Grievances, escalations, team disputes
Mediators & ombuds
Multi-party sessions, commitment tracking
Risk & compliance officers
Whistleblower reviews, governance breaches
Legal & ADR counsel
Pre-litigation mapping, ADR strategy
Policy analysts
Stakeholder modelling, policy reception
Research & intelligence analysts
Archive-scale document synthesis
Communications & crisis leads
Message stress-testing, reframing detection
Peacebuilding & humanitarian teams
Multi-actor case files, evolving commitments
HOW THE ENGINE THINKS
Dialectica combines two modes of reasoning that each fail alone and succeed together.
Deterministic
A typed knowledge graph, temporal DAG, and causal edges. Queries are deterministic and auditable. Provenance is preserved from source document to graph node.
Probabilistic
Language models for extraction, summarization, and narrative generation. Grounded in the graph, so fluency does not come at the cost of truth.
Probabilistic fluency. Deterministic truth. Every claim traceable to its source.
TRY THE THESIS