Information Extraction from Legal Documents: A Study in the Context of Common Law Court Judgements

Information Extraction from Legal Documents: A Study in the Context of Common Law Court Judgements Meladel Mistica♦* Geordie Z. Zhang♥* Hui Chia♣* Kabir Manandhar Shrestha♥ Rohit Kumar Gupta♥ Saket Khandelwal♥ Jeannie Marie Paterson♣ Timothy Baldwin♦ Daniel Beck♦ ♦ School of Computing and Information Systems ♥ Melbourne Data Analytics Platform ♣ Melbourne Law School The University of Melbourne, Australia {misticam, geordie.zhang, chia.h}@unimelb.edu.au {kmanandharsh, rohitkumarg, saketk}@student.unimelb.edu.au {jeanniep, tbaldwin, d.beck}@unimelb.edu.au Abstract ‘Common Law’ judicial systems follow the doctrine of precedent, which means the legal principles articulated in court judgements are binding in subsequent cases in lower courts. For this reason, lawyers must search prior judgements for the legal principles that are rel- evant to their case. The difficulty for those within the legal profession is that the infor- mation that they are looking for may be con- tained within a few paragraphs or sentences, but those few paragraphs may be buried within a hundred-page document. In this study, we create a schema based on the relevant informa- tion that legal professionals seek within judge- ments and perform text classification based on it, with the aim of not only assisting lawyers in researching cases, but eventually enabling large-scale analysis of legal judgements to find trends in court outcomes over time. 1 Introduction The law is reason free from passion1 — but you’ll have to dig through hundreds of pages to find it. In common law countries such as Australia, a core legal principle is the doctrine of precedent — every court judgement contains legal rulings that are binding upon subsequent cases in lower courts, though how legal rulings apply in subsequent cases is dependent on the facts of the case. When prepar- ing to give a legal opinion or argue a case, lawyers spend many long hours reading lengthy judgements to identify therein the precedents that are salient to the case at hand. This time-consuming manual pro- cess has formed a barrier to large-scale analysis of * Meladel Mistica, Geordie Z. Zhang, and Hui Chia con- tributed equally to this paper. 1Aristotle, Politica (Politics By Aristotle), written 350 B.C.E, translated by Benjamin Jowett legal judgements. Even though thousands of court judgements are published in Australia every year,2 lawyers are only able to analyse small numbers of judgements, potentially missing broader trends hidden in the vast numbers of judgements that are published by the courts. There is a growing body of research at the in- tersection of Law and Natural Language Process- ing, including prediction of court opinion about a case (Chalkidis et al., 2019a; Aletras et al., 2016), classification of legal text by legal topics or is- sues (Soh et al., 2019; Chalkidis et al., 2019b), and legal entity recognition (Cardellino et al., 2017). However, our ultimate goal is to assist lawyers in identifying sections of judgements relevant to their case at hand, as well as bulk analysis of cases to identify relationships between factual patterns and decision outcomes. For this reason, we model our initial study on the sentence-by-sentence identi- fication of argumentation …

Content truncated to 500 words