Contradiction Simulator
This experiment explores how multiple agents might assign trinary values (+1, 0, -1) to propositions and how conflicts could be reconciled. It is intentionally lightweight and text-based.
Goals
- Surface how disagreement accumulates across agents.
- Prototype conflict scores to flag when orchestrators should escalate.
- Capture resolution choices that can later be journaled via PS-SHA∞.
How to Use
- Edit
model.tsto define propositions and agents with belief vectors. - Run the TypeScript model with
ts-nodeor by compiling withtscif you have a toolchain available. - Inspect the printed conflict scores and suggested resolutions; adjust weights or strategies as desired.
Notes
- No external dependencies are required beyond the TypeScript standard library.
- This is a sandbox; it is not meant for production decision-making.