Expert annotated argument trees improve legal passage classification
Mining Legal Arguments in U.S. Corporate Case Law
Computation and Language
Summary
Legal documents often contain complex arguments that help judges decide cases. The authors created a detailed dataset of U.S. federal tax opinions, labeling different parts of the legal argument like rules, analyses, and conclusions. They showed that these labeled parts can be learned by computers to better classify and retrieve important legal passages. However, the system works best within the same case and struggles to generalize across different cases. This dataset provides a new tool for understanding structured legal arguments specifically in corporate tax law.
What this means in practice
- •For legal technology developers: Build better tools to classify and retrieve relevant legal passages in federal tax cases using expert-labeled argument trees.$Commercial implications: Creates advanced legal research software for tax law practitioners by enabling structured argument analysis and search.
- •For natural language processing engineers: Train models on the annotated corpus to improve sentence-level classification and argument structure recognition in legal text.
Authors
Luis Brena, William Jurayj, Gregory Deyesu, Zaid Al-Huneidi, Andrew Blair-Stanek, Benjamin Van Durme
Abstract
Legal argument mining supports passage classification, retrieval, and argument completion. This work introduces an expert-annotated dataset of 42 U.S. federal tax opinions on corporate reorganizations under I.R.C. §368. To our knowledge, it is the first expert-annotated, tree-structured argument corpus for this domain. Explicit spans receive one of five functional labels: Rule, Analysis, Conclusion, Background Facts, and Procedural History. Rule, Analysis, and Conclusion spans can be linked into directed support trees, while Background Facts and Procedural History serve a contextual function. The corpus provides span-based, sentence-based, flat, and tree-structured representations. Agreement analysis shows that functional node labels are more reliable than directed support edges and implicit intermediate conclusions. Directed-path agreement is stronger than direct-edge agreement, which indicates that broad reachability is more stable than exact local decomposition. Classification experiments show that functional labels are learnable under case-disjoint evaluation. Retrieval experiments show that supervised fine-tuning improves within-case retrieval. However, cross-case generalization remains weak. The dataset supports legal passage classification and provides a conservative benchmark for structured argument mining in U.S. federal tax case law.