Vienna.
The next session at EDEM 2009 starts with Dan Cartwright, whose interest is in argumentation processes in e-democracy projects. Decision-making through public consultation is a key part of e-democracy, of course, and there are various systems to engage citizens in such processes online; many of these are limited in their effectiveness, however.
One such approach are e-petitions, as introduced for example in the UK; typically, sites allow users to create and 'sign' e-petitions, but this provides no information on which part of the petition a particular signatory may agree with if multiple justifications for the petition are provided. One way to overcome this problem is the implementation of argument visualisation sytems such as Araucaria and decision support systems such as Zeno, which convert textual argument into a visual representation of the argument logic; however, these are difficult to use for the lay user.
Argumentation schemes, by contrast, represent stereotypical patterns of reasoning; they provide argument structure that is easy to understand. They also provide a set of critical questions for each argumentation pattern (e.g., for argumentation from expert opinion, a typical critical question is whether the expert is biased); however, for such questions, yes or no answers are not always available - instead, questions are more often answered by further arguments. This requires the development of models for argument scheme interaction, then, which enable users to develop a well-supported argument or critique the evidence provided in support of a given argument.
One such tool is Parmenides, which aims to provide structure to an online debate while remaining easy to use. The debate creator enters the details, and the site automatically creates the backend and debate critique interface; it also offers tools for examining the major patterns of debate in any given case. The tool provides for the use of a variety of argumentation schemes (and Dan now takes us through an example for this), and once created, an argument can be queried and evaluated by other users who can register their agreement or disagreement with specific points made. (So, this is a consultation rather than a debating tool - users are not able to introduce their own evidence or counter-argument.)
This generates richer information on why users disagree with particular statements - in particular, whether disagreements are about factual information or subjective evaluation of evidence. The system will be further developed and evaluated, and tested with a larger group of users.