Deliberation and Disagreement in Deliberative Mini-publics

How much disagreement is needed for deliberative mini-publics? How much before deliberation (call that T1)? How much afterwards (call that T2)? Begin with T1. The answer depends on what the deliberative mini-public is for. I will sketch various uses for deliberative mini-publics as a method of instilling public will formation into the policy process and in that way improving democratic practice. For those uses, there is a clear answer to our question. For deliberative mini-publics that purport to be relevant to public policy, my answer for T1: the amount of disagreement in the population that is to be represented, no more, no less. My argument is that the appropriate use of mini-publics is to give substance to the hypothetical claim: this is what the population would think if it could engage in considering the issue in depth under good conditions. To make that inference plausible, the mini-public must be representative at T1 of the population in its viewpoints and in its demographics. Hence, I will argue for institutional designs that permit evaluation of the representativeness before deliberation as well as opinion change afterwards. Further, to defend this hypothetical claim, the design must also be protected from distortions such as domination by the more advantaged or polarization (Sunstein 2002)--distortions that would suggest group dynamics are undermining deliberation on the merits, or what Habermas (1996) called “the forceless force of the better argument.” The paper develops the case for the hypothetical claim of deliberation and discusses entry points in more conventional political processes where the conclusions of such deliberations can have an impact. Deliberative Polling is discussed as a model designed to empirically investigate this hypothetical claim. For T2, the question is the degree to which successful deliberations need to produce consensus, or rationally motivated agreement. On some views, mini-publics must be designed to aim for consensus. My argument will be that if they do, they are likely to produce distortions with results motivated by group pressures and social conformity. Implications for the design of mini-publics are discussed.