Accounting After Truth: Structural Limits on Forensic Determinacy in Financial Records

This paper develops an epistemic theory of accounting records by identifying structural limits on forensic reconstruction that arise independently of incentives, governance, or audit effort. Treating accounting systems as observable trace systems, the analysis shows that under widely used balance-centric architectures, a class of ex post inference tasks—event attribution, estimator identification, and causal replay—is non-identifiable in principle. The core argument is formal rather than normative. When aggregation collapses causal detail, estimator identity is not persistently recorded, revision overwrites lineage, or repair processes are unbounded, multiple distinct event histories induce identical observable traces. Under these conditions, no amount of diligence, enforcement, or analytical sophistication can reconstruct what occurred, even for an ideal auditor with unlimited computational capacity and full access to records. The paper derives necessary conditions for forensic identifiability as logical corollaries of these non-identifiability results. These conditions are not proposed as accounting reforms or best practices; they specify the minimal structural properties any record system must satisfy if it is to support ex post causal inference at all. Where such properties are absent, reconstruction failure reflects an epistemic ceiling rather than a contingent breakdown. By clarifying what accounting records can and cannot make knowable, the paper reframes persistent restatements, attribution disputes, and unresolved enforcement cases as manifestations of architectural limits rather than institutional failure. The contribution complements existing work in accounting, auditing, and information systems by formalizing the relationship between record structure and inferential capacity.