This paper introduces epistemic lensing: the measurable deflection of scientific anomalies by a community's shared judgment standards before a paradigm shift occurs. Against passive models of filtering or suppression, the paper argues that evidentiary preferences, refusal thresholds, and methodological norms actively bend anomalous findings toward paradigm-compatible interpretations. The paper formalises this using patchwise epistemic geometry and a collective epistemic tensor proxy, and proposes a five-step empirical method combining SPECTER document embeddings, SciBERT citation-context NLP, and factor-analytic extraction of judgment-standard axes. Pilot case: Helicobacter pylori / peptic ulcer aetiology 1982–1995.
Ben Cassie (Tue,) studied this question.