Can LLMs help lawyers? Argument analysis in legal texts

Karla Salas-Jimenez

Keywords: Argument Mining, Legal Reasoning, Defeasible Deontic Logic, Answer Set Programming

TL;DR: This work proposes a modular framework that combines argument extraction, NL-Logic translation, and argument strength evaluation, using prompting strategies and argument schemes.

Abstract: Legal reasoning fundamentally depends on the the ability of the legal professionals to identify, formalize, and evaluate arguments in legal texts. However, current research relies on a limited set of annotated resources and rarely integrates formal logic into the full analysis pipeline. This work proposes a modular framework that combines argument extraction, natural language to logic (NL-Logic) translation, and argument strength evaluation, using prompting strategies and argument schemes. The goal is to advance computational legal reasoning by offering tools that improve transparency and consistency in legal decision-making, creating new and diverse annotated corpora to support stronger models, and contributing to AI systems that help lawyers construct and interpret well-founded legal arguments through Logic-based evaluation.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 13