Adversary Arguments and the Logic of Personal Attacks
Lawrence Birnbaum Yale University
Engaging in an argument is a task of considerable complexity, involving the coordination of many different abilities and knowledge sources. In this chapter, we present a partial theory of the processes an arguer might use to understand utterances in an argument and to produce adequate responses. Parts of this theory have been implemented in a computer program, ABDUL/ILANA, which models either an Arab or an Israeli arguing about the Middle East. Before examining how arguments proceed, however, we should first ask why anyone would even bother to participate in an argument, that is, what the goals of someone in an argument might be.
Arguments can basically be divided into two classes, depending on the goals and expectations of the participants. The first class consists of arguments in which the participants are motivated to reach a common agreement, for example in order to solve some problem. These are called persuasion arguments, because the participants are both willing to be persuaded as well as trying to persuade. In the second class, adversary arguments, neither participant expects to persuade or be persuaded: The participants intend to remain adversaries, and present their arguments for the judgment of an audience (which may or may not actually be present). In these arguments, an arguer's aim is to make his side look good while making the opponent's look bad. Our research (and hence this chapter) has been primarily concerned with adversary arguments.