Two of the most powerful predictors of human behavior are the story someone tells about themselves and the power relationships they are embedded in. The narrative and power engine group makes both computable.

The Engines

  • Narrative Physics Engine — Models an actor's self-narrative as a structured story: protagonist, antagonist, goal, obstacle, resolution arc. Scores how strongly behavior aligns with the narrative.
  • Narrative Propagation Engine — Models belief spread: how quickly and reliably an actor's narrative is adopted by others in their network.
  • Power Dynamics Engine — Detects authority asymmetry in communication patterns: who initiates, who yields, who sets the agenda.
  • Trust Graph Engine — Maps trust relationships and trust deficits between actors. A trust deficit is when stated trust is inconsistent with observed behavior.

Code Walkthrough

// Power dynamics: compute initiative asymmetry
function scorePowerDynamics(interactions) {
  const initiationCounts = {};
  for (const interaction of interactions) {
    initiationCounts[interaction.initiator] =
      (initiationCounts[interaction.initiator] ?? 0) + 1;
  }
  const total    = interactions.length;
  const actorIds = Object.keys(initiationCounts);
  return actorIds.map(id => ({
    actorId:         id,
    initiationShare: initiationCounts[id] / total,
    powerIndicator:  initiationCounts[id] / total > 0.65
      ? "dominant-initiation-pattern"
      : "balanced-interaction-pattern",
  }));
}

What to Watch For

  • Power asymmetry scores are highly sensitive to context. A client who initiates 70% of interactions with their lawyer is not "dominant" — they are anxious. Context metadata must moderate these scores.
  • Trust deficit scores require human review before any adverse action is taken based on them.