The following is a defense of Jurgen Habermas' argumentational consensual procedure against Keith Lehrer and Carl Wagner's weighted averaging consensual procedure (and, I tentatively claim, against any weighted averaging consensual procedure). The argument is twofold: if Lehrer and Wagner intend, implicity, to replace what is for Habermas the metatheoretical stage of a discussion with the aggregation of judgments of respect, then their procedure fails to make use of all available information and the participants are not committed to the weighted average position on these grounds; if, on the other hand, they do not intend to replace metatheoretical discussion by aggregation, then the conditions under which the discussion could conceivably have come to a halt are such as to provide no support for the claim that it is rational to aggregate, rather than to consider the discussion unresolved until more information is available.