Doctrine, Data, and the Death of DuPont
Thomas A. Reichert, Doctrine, Data, and the Death of DuPont, 36 Fordham Intell. Prop. Media & Ent. L.J. 678 (2026).
88 Pages Posted: 2 Dec 2025 Last revised: 17 Jun 2026
Date Written: December 01, 2025
Abstract
For fifty years, trademark opinions have claimed to apply a comprehensive thirteen-factor test for trademark confusion. They are deeply mistaken. Using AI powered analysis of over 4,000 TTAB inter partes decisions (2000-2025), this Article proves what practitioners have long suspected: in Section 2(d) adjudication, the test has collapsed to just two factors.
A simple categorical rule predicting confusion if and only if both mark similarity (Factor 1) and goods/services relatedness (Factor 2) favor confusion achieves 99.55% accuracy across 4,651 comparisons. Cross-validated logistic regression confirms the pattern: a two-factor model achieves 99.46% accuracy, while adding the remaining eleven factors actually makes prediction worse (99.18%), a textbook overfitting result. The thirteen-factor framework does not refine the two factor signal; it adds noise.
These findings reveal concrete doctrinal harms: practitioners brief eleven factors that do not matter, the Board analyzes them at length in every opinion, and the resulting complexity obscures what is actually a binary inquiry. The Article proposes reforms that center the two determinative factors and confine secondary considerations to narrow tiebreakers in genuinely ambiguous cases. Finally, it advances a broader "multifactor collapse" hypothesis and outlines a research agenda for testing whether other legal balancing frameworks exhibit similar patterns where doctrinal complexity masks simpler underlying decision-making.
Keywords: Trademark, Likelihood of confusion, DuPont, Mark similarity, confusion, Trademark infringement, Multifactor test, Trademark Trial and Appeal Board, TTAB, USPTO, Trademark registration, Opposition proceedings, Cancellation proceedings, Inter partes proceedings, Section 2(d), Empirical legal studies, Large language models, Machine learning, Natural language processing, Computational law, Legal analytics, Predictive modeling, Logistic regression, Content analysis, Multifactor collapse, Doctrinal gap, Law in action, Judicial decision-making, Cognitive heuristics, Legal realism, Doctrinal reform, Access to justice, Legal legitimacy, Lanham Act, Heuristics
JEL Classification: K11, K41, C45, D83
Suggested Citation: Suggested Citation