Ebook: Legal Knowledge and Information Systems

This volume contains the Proceedings of the Nineteenth JURIX Conference on Legal Knowledge and Information Systems (JURIX 2006), December 7th–9th, Université Pantheon Assas – Paris II, France. This year we hoped to put two new topics on the agenda: Artificial Intelligence in police and intelligence services and the impact of AI on law, legal procedures and legal institutions. Despite the political attention paid to international crime, the war on terrorism and the concerns about the growing number of regulations, increase of administrative costs and overloading of courts, too few contributions about these issues were submitted. Nevertheless, the large number of submissions shows that our interdisciplinary community is still growing and active.

This volume again covers a broad range of topics. Argumentation is central to legal reasoning and therefore it is no surprise that researchers have focused on computational theories of argumentation. In this book four papers are dedicated to this topic. Atkinson & Bench-Capon compare disputes as dialectical trees to other abstract systems for argument representation. Bex et al. focus on causal-abductive reasoning and default-style argumentation to deal with stories, evidence and generalisations in the legal domain. Prakken & Sartor study formal models of representation of presumptions and their effects on the burden of proof. Chorley & Bench-Capon add the possibility of State interventions to improve the outcome of previous research on an argument scheme for practical reasoning.

Typical to the legal field is the use of written knowledge sources, especially legal sources. These have been subject to AI&Law research for a long time, varying from structuring and accessing legal sources to using natural language processing techniques in order to determine the semantics of language utterances. This book contains four papers on legal sources. De Maat et al. explain how parsing can provide support for combining legal content stores of different providers. Saravanan et al. propose probabilistic graphical models for automatic text summarization. Klein et al. report on methods for retrieving relevant case law within the domain of tort law. Van Opijnen describes the problems encountered in the development of a citation standard and explains the Public Index within the Dutch judiciary.

Central to AI and consequently to AI&Law are knowledge representation and ontologies. The latter especially are becoming more and more popular due to developments in SemanticWeb research. In this book we have included four papers on these topics. Roth describes an evaluation method for determining the effectiveness of deterrence. Lindahl & Odelstad analyse intermediaries by using a theory of “intervenients”. Rubino et al. present a description of an ontology that should enable semantic access to digital legal information and clarify legal theoretical concepts. Hagiwara et al. propose a theory and an implementation which detects discordance in the text of an ordinance of the Toyama prefecture in Japan.

I am not a philosopher so I am also pleased to see some practical results of our theoretical reflections. This book contains three papers on applications. Zeleznikow & Bellucci explain how notions of fairness to interests are used in their Family_Mediator system. Van Zeeland et al. explain their Personal Injury Claims Express web-application. Toyama et al. explain how natural language processing can be used to support both the compilation of a standard bilingual dictionary and unifying translation equivalents of legal technical terms. Last but not least we have four short papers on various topics.

Acknowledgments

The JURIX 2006 Conference organizers express their warm thanks to Danièle Bourcier who organized the conference together with her colleagues and the Université Pantheon Assas – Paris II. We would also like to thank the people at the Leibniz Center for Law (UvA) for providing support for the review process.

For this Conference we selected 15 full papers and 4 short papers out of 29 contributions. The authors represented fifteen countries. Papers have been reviewed by members of the programme committee:

• Kevin Ashley (University of Pittsburgh)

• Trevor Bench-Capon (University of Liverpool)

• Danièle Bourcier (CERSA-CNRS, Université Paris II)

• Paul Bourgine (CREA, Ecole Polytechnique, Paris)

• Karl Branting (BAE Systems Inc.)

• Jaime Delgado (Universitat Politècnia de Catalunya)

• Tom van Engers (University of Amsterdam)

• Aldo Gangemi (CNR-ISTC)

• Thomas F. Gordon (Fraunhofer FOKUS, Berlin)

• Eduard Hovy (University of Southern California)

• Ronald Leenes (Universiteit van Tilburg)

• Arno Lodder (Vrije Universiteit Amsterdam)

• Ronald Loui (Washington University)

• Thorne McCarty (Rutgers University)

• Henry Prakken (Utrecht University & University of Groningen)

• Giovanni Sartor (University of Bologna)

• Erich Sweighofer (University of Vienna)

• Peter Spyns (Vrije Universiteit Brussel)

• Andrew Stranieri (JustSyss Pty Ltd)

• Bart Verheij (Universiteit of Groningen)

• Maria Wimmer (University of Koblenz-Landau)

• Radboud Winkels (University of Amsterdam)

We would also like to thank all external reviewers. The conference would not have been possible without the time and effort put in by these people and so thanks are due to them all. I also take the opportunity to thank the invited speakers, Ron Loui and Burkhard Schaeffer, for their contribution. Finally, thanks are due to Jobien Sombekke and Rinke Hoekstra (University of Amsterdam) for struggling again with LATEX and succeeding in getting the final version of this book in good shape before the deadline.

I hope that both the conference and the book help the AI&Law community stay the vivid, inspiring, knowledge productive and collaborative bunch of researchers and practitioners it has been over the last two decades.

Tom M. van Engers, Program Chair

In this paper we explore and contrast different forms of dispute representation from the present and past literature. In particular, we examine the Zeno Framework, which represents disputes as dialectical trees, and compare this to other abstract systems for argument representation. We discuss the merits of each method and show how the example argument first represented in Zeno can be transformed into a more concise representation whilst retaining its desirable features.

This paper concerns the reasoning with stories, evidence and generalisations in a legal context. We will make some notions from the existing Anchored Narratives theory more clear by making use of two formal techniques from AI, namely causal-abductive reasoning and default-style argumentation. We will propose a combination of these two formalisms which solves some of the problems of the causal-abductive approach.

This paper studies the logical modelling of presumptions and their effects on the burden of proof. Presumptions are modelled as default rules and their effect on the burden of proof is defined in terms of a distinction between the burden of production, the burden of persuasion and the tactical burden of proof. These notions are logically characterised in such a way that presumptions enable a party to fulfil a burden of production or persuasion while shifting a tactical burden to the other party. Finally, it is shown how debates about what can be presumed can be modelled as debates about the backings of default rules.

In previous work we have looked at how agents reason in situations of conflict of interest, using an approach based on an argument scheme for practical reasoning and associated critical questions. In this paper we add the possibility of the State intervening to attempt to improve the outcome. We model the State as another agent in the scenario, with its own repertoire of actions, and its own interests represented as an ordering on values. Where arguments are directed towards different agents with their own different interests it is not possible to use standard means of determining the acceptability of arguments, since these methods evaluate arguments from a single perspective. We therefore adopt the approach of simulating the reasoning of the agents using a procedural version of the argument scheme and associated questions. We present our work through consideration of an extended example, draw some conclusions and identify directions for future work.

Combining legal content stores of different providers is usually time, effort and money intensive due to the usually 'hard-wired' links between different parts of the constituting sources within those stores. In practice users of legal content are confronted with a vendor lock-in situation and have to find work-arounds when they want to combine their own content with the content provided by others. In the BSN project we developed a parser that enables the creation of a referential structure on top of a legal content store. We empirically tested the parsers' effectiveness and found an over 95% accuracy even for complex references.

In this paper, we propose a novel idea for applying probabilistic graphical models for automatic text summarization task related to a legal domain. Identification of rhetorical roles present in the sentences of a legal document is the important text mining process involved in this task. A Conditional Random Field (CRF) is applied to segment a given legal document into seven labeled components and each label represents the appropriate rhetorical roles. Feature sets with varying characteristics are employed in order to provide significant improvements in CRFs performance. Our system is then enriched by the application of a term distribution model with structured domain knowledge to extract key sentences related to rhetorical categories. The final structured summary has been observed to be closest to 80% accuracy level to the ideal summary generated by experts in the area.

In the context of intelligent disclosure of case law, we report on our findings on methods for retrieving relevant case law within the domain of tort law from a repository of 68.000 court verdicts. We apply a thesaurus-based technique to find specific legal situations. It appears that statistical measures of term relevance are insufficient, but that explicit knowledge about specific formulations used in law and case law are required to distinguish relevant case law from irrelevant. In addition, we found out that retrieving legal concepts with an “interpretive” character requires a different method than retrieving concepts that do not require additional interpretation.

Case law citations are often lenghty and complex, despite the development of neutral citation standards. Within the Dutch judiciary a public index with more than 350,000 case law references is developed which makes the citing of references superfluous. This paper describes the realization of the index, the logical model and the interfaces. Some possible improvements are discussed, both in the logical model and in the way indexnumbers are assigned.

Deterrence is a relation between strategies of players involved in a game. An index of deterrence is introduced as a quantitative measure of how effective deterrence is. Using this index of deterrence, the relation between payoffs and deterrence is studied in a concrete case. As it turns out, 'more focused' deterrence here typically leads to higher payoffs for the party whose strategies are deterred, and lower payoffs for the party who can play the deterrent strategies.

Legal terms such as “owner”, “contract”, “possession”, “citizen” are “intermediaries” in the sense that they serve as vehicles of inference between statements of legal grounds, on one hand, and legal consequences, on the other. After introducing our approach to the representation of a normative system, we present a theory of “intervenients”, seen as a tool for analysing intermediaries. The paper is especially concerned with the subject-matter of open and closed intervenients as well as the related issue of negations of intervenients. Also, we introduce the idea of so-called gic-systems, where “gic” is an abbreviation of “ground-intervenient-consequence”.

In this paper we present an OWL ontology of fundamental legal concepts developed within the ESTRELLA European project. The ontology includes the basic normative components of legal knowledge: deontic modalities, obligative rights, permissive rights, liberty rights, liability rights, different kinds of legal powers, potestative rights (rights to produce legal results) and sources of law. Besides the taxonomy the ontology comprises also the semantic relations between the concepts. We hope that it may be useful for semantic access to digital legal information and for the representation of legal knowledge.

In this paper, we propose a procedure of discordance detection in an actual legal code, that is the regional ordinance of Toyama Prefecture, Japan. In this study, we expand the notion of inconsistency to the discordance including antonyms based on an ontology, and precluded the conventional negative connective. We have implemented the system that converts XML logical formats to Prolog, and has inspected the whole code.

Most negotiation support systems have focused upon the notion of meeting the disputants' interests. This neglects the vital issues of justice and power. In this article we address the issue of how to add notions of fairness to interests, through the development of the Family_Mediator system. Family_Mediator is an extension of the Family_Winner system, which advises mediators about potential trade-offs and compensation strategies for divorcing couples.

The handling of complex Personal Injury Claims is a complicated, confusing, challenging, and sometimes nasty procedure. This paper describes a project in which all relevant stakeholders collaborate to improve the procedure by establishing a Code of Conduct, the core ideas of which are implemented in a web-application, PICE — Personal Injury Claims Express. It discusses the core ideas, the application and the initial results of a pilot project.

Recently, society has expressed increased demands for translation of Japanese statutes into foreign languages. The various motivations behind these demands include social and economic globalization and the need for technical assistance to legal reform. In this paper, we describe the problem of translating Japanese statues and show how to solve it by utilizing technologies developed for natural language processing. In particular, we show how to support both the compilation of a standard bilingual dictionary and the unification of translation equivalents of legal technical terms in compliance with the dictionary by using word alignment.

This paper reports on the ongoing development of a collaborative, web-based application for argument visualization named AVER (Argument Visualization for Evidential Reasoning). It is targeted at police officers who may use it to express their reasoning about a case based on evidence. AVER provides an interface which supports the construction and visualization of argument graphs and handles more advanced argumentation concepts such as schemes. Further, it is based on a known argument ontology and has a solid theoretical underpinning in formal theories of argumentation.

At the crossroad of law and computer science, the notion of Digital Rights Management renews the research on Lex Electronica. Rights Expression Languages provide the legal semantic and syntax to be implemented by Digital Rights Management Systems. Rights Expressions are legal metadata that define the actions led on a digital text or multimedia file. This article presents some legal resources and requirements to express rights and transactions pertaining to digital documents and data, ie legally or contractually allowed usages or actions. The proposed model associate terminological resources from technology, usages, exclusive rights and public domain. Research around this article takes place in Medialex, an interdisciplinary project aiming at building a transaction ontology to license, access and reuse.

We consider a problem of causality in legal reasoning. Conditio sine qua non (c.s.q.n) is a frequently used heuristics which determines a causality in legal reasoning. We argue that a paradox of c.s.q.n. is derived from a confusion between disjunction of causes and disjunctive cause and give a logical solution to the paradox using minimal abduction.

In this article we present the structure and outline of the proposed content of the LKIF-Core ontology. LKIF is an interchange format for legal knowledge, under development in Estrella, a 6th framework European project. One of the layers of expressiveness in LKIF is to consist of a combination of the OWL-DL and SWRL knowledge representation languages. LKIF-Core adds an ontology containing definitions of basic legal concepts. A number of these concepts are used to define frameworks that capture their use in legal reasoning. LKIF-Core is intended to enable re-use in the construction of legal knowledge bases and as semantic grounding for the translation of existing knowledge formats into LKIF.