Analytic Abduction Enhances Human-AI Coordination for Complex Problem Solving

Abductive reasoning, which involves forming explanations for observations, typically operates in two modes: synthetic (building explanations from hypotheses) and analytic (identifying latent factors causing a complex state). This research focuses on the analytic mode, developing it as a non-greedy, risk-sensitive discipline for commitment. The core of the framework is the kappa-tau apparatus, where kappa encodes the epistemic interaction among hypotheses, and tau sets a commitment threshold based on decision stakes. A central contribution is the "causal cluster," a structured object detailing participating latent factors, their weights, and interaction structures, guarded by a two-level architecture to prevent causal misattribution. Demonstrated in epidemiological crisis decomposition and cyber threat analysis, this framework provides decision-makers with competing explanatory scenarios, weighted by plausibility and paired with evidence needed for resolution. This "legibility of suspended decomposition" fosters human-AI coordination by resisting premature convergence, enabling sound action even before ambiguities are fully resolved.