A Limit State Causality Framework for Failure Investigations in Structural Engineering

Structural failure investigations require causal judgments that distinguish primary causes, contributing factors, and background conditions. Yet, existing practice often relies on expert narrative without a shared, auditable rubric for integrating mechanistic models, probabilistic limit-state reasoning, and heterogeneous evidence. This paper proposes a Limit State Causality Framework (LSCF) that evaluates candidate factors using nine dimensions: temporal and process alignment; mechanism–performance and threshold linkage; counterfactual scenario impact; competing explanations and parsimony; scale and pattern coherence; evidence diversity and independence; population context and base rates; human, organizational, and regulatory context; and robustness under uncertainty. Each dimension is accompanied by tiered scoring guidance and classification rules that gate primary-cause labeling on the mechanistic, counterfactual, explanatory, and robustness core. The proposed framework is illustrated using two case studies to demonstrate dimension-wise scoring and classification under typical forensic constraints. With time, and given demonstrated reproducibility in broader use, LSCF can provide a technical foundation for future guidance language and eventual codal-style provisions for reporting causal judgments in structural failure investigations.