web site
Research
Home
Contact information
Research
(e.g., NWO Forensic Science)
Publications (or in blog-style)
Teaching
People
Miscellany
---
eNotes (@WordPress)
Vitae (@LinkedIn)
Author profile (@GoogleScholar)
Navigation
University of Groningen
Faculty of Mathematics and Natural Sciences
Department of Artificial Intelligence
ALICE research institute
Multi-agent systems research group
Tools
Time and date
Wolfram|Alpha
Google Scholar
Wikipedia
|
A semi-formal impression of my perspective on AI & law is available in this 2005 interview (in Dutch; pdf) or this 2007 coffeehouse conversation. My views on semi-formal argumentation are explained in this 2009 text where I trace Toulmin's footprints in AI. My 2005 book Virtual arguments contains a lot of information about my work on the design and logical underpinning of argumentation software (ArguMed based on DefLog is the software version to focus on; it has been assessed by a protocolled qualitative user evaluation). The book has more than the 2003 Artificial intelligence journal paper. A 2007 paper gives an update, and a perspective on increasing the usefulness of argumentation software for professionals. That paper also contains an accessible version of my view on entangled dialectical arguments (and associated diagrams), of which the logical formalization is studied in the 2003 paper on DefLog. A deviant strand of argumentation software, focusing more on content and less on argument diagramming, is the ArguGuide tool developed with Maaike Schweers, Stijn Colen and Fokie Cnossen (see ICAIL 2009 for a very short description of the statistically significant evidence that we have showing that the ArguGuide design improves performance on a case solving task).
I have always paid a lot of attention to the naturalness of the argument/reasoning models I developed (although this may not always be obvious). Reason-Based Logic (with Jaap Hage) is a good example of this (see e.g. my 1996 dissertation). Also noteworthy in this respect are this 1997 attempt (with Jaap Hage and Arno Lodder) to faithfully model a part of Dutch tort law, and a second improved attempt (with Jaap Hage and Gerrit van Maanen; only available in Dutch). A very sobering, hence fruitful, experience in this respect has been the development of teaching material for students of Dutch law to improve their argumentative skills. It is in its third printing (2008). Even my DefLog formalism aims to provide a natural conceptualisation of argumentation. (That is why I e.g. prefer to speak of arguments justifying prima facie conclusions and not of justified arguments.) My 2005 interpretation and extension of Toulmin's argument model/diagram in terms of DefLog is an indication of its naturalness (for more on Toulmin see this 2006 volume coedited with David Hitchcock and this 2009 paper on the reception of Toulmin's ideas in AI). Another such indication is my 2003 treatment of Walton's argumentation schemes, where I treat argumentation schemes as a kind of semi-formal rules of inference with exceptions that can be found and systematized using a knowledge engineering approach. An early incarnation of that work was presented at ICAIL 2001, formally worked out in the full report. The idea of argumentation schemes as contextual, defeasible, semi-formal rules of inference was already used in my work on Reason-Based Logic (with Jaap Hage; e.g., chapter 2, section 6 of my 1996 dissertation). The associated 'philosophy of logic' is explained in my 1999 paper 'Logic, context and valid inference. Or: Can there be a logic of law?', a personal favorite. For me, argumentation schemes are a continuation of Toulmin's ideas on warrants: contextual, defeasible, semi-formal. This perspective is explained and embedded in a research tradition in my 2009 book chapter on how semi-formal, defeasible argumentation schemes creep into logic. My interest in reasoning with evidence and legal proof started in Maastricht (see the 2000 text on the connections between argumentation and stories that I wrote for Hans Crombag). In Groningen this work really got off the ground by Floris Bex's PhD research (2005-2009), in a joint project with Henry Prakken and Peter van Koppen (and also Susan van den Braak, Gerard Vreeswijk and Herre van Oostendorp; see an early overview paper, this 2007 paper and the project's page). The 2009 volume edited with Hendrik Kaptein and Henry Prakken is also on this theme. It includes a chapter on reconstructing the anchored narratives theory using argumentation schemes. In 1996 I proposed two additional semantics for abstract argumentation in the sense of Dung 1995: stage extensions and admissible stage extensions. The latter now go by the name of semi-stable extensions (see Caminada's 2006 work). I continued this line of work in a more expressive language (see the 2003 paper on DefLog), in order to incorporate pros and cons, and in an attempt to close the gap between argumentation formalisms and logic. The DefLog language not only allows the expression of support and attack, but also of reasoning about support and attack. The latter was called entanglement by Bram Roth. ArguMed based on DefLog is an implementation of DefLog (computing its Dung-faithful version of stable semantics) and the associated argument diagrams. For abstract argumentation I have also implemented a software tool that computes small admissible sets and the grounded, preferred, stable and semi-stable semantics for abstract argumentation (see this 2007 publication). In 1996 I received my doctoral degree by defending my dissertation, entitled 'Rules, Reasons, Arguments. Formal studies of argumentation and defeat'. I have contributed to Jaap Hage's Reason-Based Logic, a model of rules and reasons, have developed CumulA, a model of argumentation with arguments and counterarguments in stages, have provided an abstract model of the law in terms of states of affairs, events, and rules (with Jaap Hage), and have implemented Argue! and ArguMed, two systems for automated argument assistance. I have also developed the logical system DefLog, focusing on the interpretation of prima facie justified assumptions. In collaboration with Bram Roth I have studied case-based reasoning in terms of the comparison of the dialectical arguments in cases. This resulted in Roth's dissertation Case-based reasoning in the law. A formal theory of reasoning by case comparison (2003), written under my supervision. My research about argument assistance systems resulted in a publication in the Artificial Intelligence journal and a book, entitled Virtual Arguments On the Design of Argument Assistants for Lawyers and Other Arguers .
Key words: legal reasoning, defeasible reasoning, dialectical argumentation, computational dialectics, nonmonotonic logics, argument defeat, artificial intelligence and law, legal ontology, rules and principles, argument mediation, argument assistance. |