Abstract Scientific explanation often begins after a domain has already acquired its basic terms, relations, practices, and explanatory expectations. These prior commitments can remain implicit even when later mechanisms, models, and empirical procedures are highly developed. This paper introduces Constitutional Reconstruction as a domain-independent method for reconstructing the pre-mechanistic architecture that scientific explanation presupposes. The method does not begin by asking whether a domain is true, false, complete, or empirically validated. It asks what must already be constitutionally in place for a phenomenon, theory, practice, or explanatory domain to become intelligible. Constitutional Reconstruction proceeds through primitive inventory, dependency graph construction, gap governance, gates, candidates, bounded reconstruction trials, outcome classification, benchmarks, and preserved provenance. Its central discipline is the refusal to collapse distinct operations into one another: scope is not inventory, reconstruction is not validation, and covered reconstruction is not admission. The paper defines the method, its units, its gates, its non-admission rule, its benchmark logic, and its provenance requirements. It argues that Constitutional Reconstruction occupies a methodological layer between conceptual clarification and mechanism discovery: it reconstructs the conditions under which later explanation can responsibly begin.
Israel Don (Thu,) studied this question.