This paper focuses on eristic in political debate of the forensic, or confrontational, type. First, some findings on the enactment and persuasiveness of hostility in a series of Danish TV-debates 1975–85 are presented, including a list of the clearly hostile debater's characteristics and a subdivision of conspiracy arguments. This presentation serves to illustrate that hostility is less persuasive than argumentation practitioners and theorists tend to assume. Next, the widespread notion of debate as a genre half-way between the quarrel and the critical discussion is challenged in a discussion of Douglas N. Walton's distinction between types of dialogue. It is maintained that the normative model of confrontational debate excludes the quarrel and that debate should not be perceived as second-rate critical discussion.