Section pages
WHY CONFLICT · WHY NOW
Understand a disagreement before trying to resolve it.
Research and architecture in development. This page explains our design direction. It is not a list of completed PRAXIS capabilities. See the documented product capabilities and open work.
Disagreements involve evidence, history, interests, values, and power. Making that structure visible can help people understand what is contested and why. It cannot remove the underlying differences or decide the outcome for them.
Problem
Parties see different pictures of the same situation.
Response
A reviewable record of the evidence and the points of disagreement.
Boundary
We do not decide who is right. Humans still do.
INTERACTIVE · 5-DAY DISPUTE, UNFOLDING
A fictional conversation, reviewed step by step.
This is a manually authored illustration, not live extraction or a production integration. Advance the fictional conversation to inspect proposed records and review questions. The scenarios are unmeasured possibilities.
So we're agreed: you own the Q4 launch deck content, I handle design. Lock it in by Thursday?
EXTRACTED PRIMITIVES · 3
ISSUES FLAGGED · 0
No issues yet. Issues surface as contradictions, drift, or escalation vectors appear.
ONE CASE · FLAGSHIP + ENGINE + ADJACENT
What you do next depends on where you sit.
The Sam ↔ Alex thread is the same artefact for everyone. PRAXIS is where the analysis lives. DIALECTICA develops the knowledge layer. CONCORDIA and Wind Tunnel are experiments with separate validation needs.
PRAXIS
Flagship · The flagship workbench
Try the document, capsule-review and conversation workflow with a non-sensitive source set. A manager must check any generated assessment; complete brief export remains in development.
/product/praxisDIALECTICA
Engine · The engine underneath
Knowledge-engine development for source processing, typed proposals and review. This teaching case does not prove a deployed end-to-end integration.
/dialecticaCONCORDIA
Experimental · When the room actually meets
Experimental direction: explore how a transcript might connect statements and commitments. This illustration does not validate live utterance binding or drift detection.
/research/experimentsWind Tunnel
Experimental · Stress-test before you escalate
Experimental direction: compare drafts using synthetic modeled responses. These outputs are directional scenarios, not polling or evidence of likely audience behavior.
/research/experimentsWHAT YOU ARE USING TODAY
Sticky notes, a smart chatbot, or a typed graph.
Three workflows to evaluate. Compare source fidelity, persistence and human review using the same task.
Post-its & inbox
Informal notes and correspondence
- ·Easy to start; handover needs organization
- ·Dates need explicit recording
- ·Source references depend on the note-taking process
- ·Disagreement needs to be kept explicitly
- ·Judgment depends on the reviewer
General-purpose AI workspace
Capabilities depend on tools and configuration
- ·Can summarize and generate structured records
- ·Temporal claims need checking against sources
- ·Can distinguish concepts; output needs validation
- ·Can make unsupported inferences
- ·Citation quality depends on retrieval and review
TACITUS
Proposed reviewable knowledge workflow
- ✓Source and time records are design requirements
- ✓Commitment comparison needs validation
- ✓Distinct claim and interest fields
- ✓Competing claims can be preserved explicitly
- ✓Scenario analysis remains experimental
WHO IT IS FOR
Eight kinds of work. One shared infrastructure.
Same primitives, different stakes. Each role gets the same typed graph , and a product tuned to their cadence.
Policy desks & mission staff
Briefings, SITREPs, multi-actor case files
Diplomats & Track II teams
Process memory, commitments, narrative drift
Mediators & ombuds
Live sessions, commitment tracking
Peacebuilding & humanitarian teams
Evolving commitments, contested events
Communications & crisis leads
Message stress-testing, polarization risk
Risk, compliance & legal counsel
Whistleblower reviews, ADR mapping
Conflict researchers & think-tanks
Archive-scale, cited synthesis
THE ARGUMENT, IN THREE PARAGRAPHS
We are not trying to end conflict. It is the default state of any system with more than one agent and more than one goal. We want analysts to be able to compare accounts, identify missing evidence, and explain how they reached a conclusion.
Practitioners already work with actors, claims, interests, commitments, and events. Our question is how software can preserve those distinctions across a growing record, including changes of position and competing accounts of the same event.
Language models can help read and summarize that record, but they can also miss a qualification or infer a relationship the sources do not support. DIALECTICA explores whether typed records and explicit review decisions make those errors easier to find. That benefit needs to be demonstrated through evaluation, including comparison with simpler retrieval methods.
Better evidence does not settle every disagreement. It gives people a firmer basis for deciding what to do next.
THE GRAMMAR · ACO
Eight concepts for examining a dispute.
Actor
Who participates
Individuals, teams, institutions, states, coalitions: any agent with goals and agency in the dispute.
Claim
What is asserted
Factual, evaluative, or normative statements a party makes. The surface of the disagreement.
Interest
What lies beneath
A party’s stated or inferred needs and motivations. Inferred interests must remain distinguishable from what the party has confirmed.
Constraint
What limits outcomes
Rules, regulations, norms, hard facts. The structural bounds on possible resolution.
Leverage
What holds the power
Resources and dependencies that give a party the ability to act, withhold, or influence.
Commitment
What was agreed
Promises, contracts, pledges, ceasefires. The graph-layer distinction between asserted and agreed.
Event
What happened when
Occurrences linked to sources, with separate records of when they happened and when they became known. Dates may be uncertain or disputed.
Narrative
How it is framed
The story each party tells about the dispute. Captures framing, re-framing, and narrative drift over time.