Automated Argument Assistance for Lawyers

Bart Verheij

Abstract

Automated argument assistance should be distinguished from automated reasoning: while automated reasoning systems replace the reasoning of the users, argument-assistance systems do not reason themselves, but are tools assisting the users' reasoning.

ArguMed is the successor of the Argue!-system. Argue!'s graphical interface was considered too unfamiliar for the intended users, and its underlying argumentation theory was not sufficiently transparent.

The system as described here is the second version of the ArguMed-system. Two drawbacks of the previous version have been solved. First, though the arguments were presented graphically, argument attack was not. It was graphically shown that an argument was defeated by an attacking argument, but not by which argument. Second, in the argumentation theory underlying ArguMed's first version, it was not possible to put at issue that a particular statement was a reason for another statement, or that a statement was an exception.

Solving the first of these two drawbacks has led to a new graphical representation of the arguments, in which argument attacks are shown, and to a change in the argumentation theory, viz. the introduction of a novel notion of an argument, viz. that of a dialectical argument. Briefly, a dialectical argument is an argument in which attacks (and counterattacks) are incorporated. Solving the second drawback has led to the introduction of step warrants and undercutter warrants into the argumentation theory. The resulting notion of a warranted dialectical argument is the analog for defeasible argumentation of the notion of a (Hilbert-style) proof of classical logic.

The present version of the ArguMed-system is put in context by a brief comparison with selected other systems, viz. Loui's Room 5, Gordon and Karacapilidis' Zeno, ArguMed's precursor Argue! and the previous version of ArguMed.

Reference:
Verheij, Bart (1999). Automated Argument Assistance for Lawyers. The Seventh International Conference on Artificial Intelligence and Law. Proceedings of the Conference, pp. 43-52. ACM, New York.

Download manuscript (in PDF-format)

Note: In the version of the paper as published by ACM, some of the logical symbols are printed wrongly (due to a font problem). The version of the paper that can be downloaded here shows the paper as it was intended to be. The ACM version uses for and for . The logical symbols only occur on the pages 45 and 46.