The Deuterium Assumption: An Evidential Audit of the Free-Nucleon Baseline in Deep Inelastic Scattering

No experiment has ever directly measured the structure function of a free neutron. Every published determination relies on deuterium, either as target or baseline, with nuclear corrections modelled and subtracted. This paper audits the assumption underlying that procedure—that the deuteron is, to calculable approximation, a proton plus a neutron—against three independent lines of published evidence. First, the systematic discrepancy between neutron structure extracted via deuterium subtraction (BCDMS) and via tritium-helium-3 mirror symmetry (JLab MARATHON) reaches ~20% at high Bjorken x, larger than the spread of leading nuclear correction models. Second, Griffioen et al. (2015) measure a non-zero EMC slope in deuterium itself, dR_EMC^d/dx = −0.10 ± 0.05 over x = 0.35–0.7, establishing that deuterium exhibits genuine nuclear modification at the 2σ level. Third, the universal anti-shadowing enhancement at x ≈ 0.05–0.15, present in every nucleus measured, has resisted consensus theoretical explanation for four decades despite multiple modelling attempts. These three results, taken together, raise a question that the DIS programme has not yet formally addressed: is the deuterium baseline valid? This paper does not propose an alternative model. It identifies the evidence, examines what the conventional framework predicts versus what is observed, and poses five data-grounded questions for the experimental programme.