EfficiencyThere is a price to pay for the pursuit of this dialectical forum. That price is inefficiency. Consensual will-formation takes a long time. When a decision has to be made, stamina sometimes is as important as argument. Although the ideal of consensus guided all policy-making at the Seneca Peace Camp, “in reality decisions were often made on the basis of who could survive the

longest in meetings that stretched on for hours or days.”33 As many of the organizers had anticipated, consensual decision-making was a difficult, drawn-out process, frustrating those who were more interested in acting than talking. If a relatively small group, united by a common goal, found discourse an unwieldy tool of decision-making, can we even begin to envi­sion the wider discourse to which Habermas alludes? A consensually steered society implies a discourse that includes all its members. Is this plausible? I want to suggest that the implausibility of the Seneca model for the society at large does not point to the implausibility of practical discourse as such. It points instead to implausibility of discourse as a deci­sion procedure within democratic politics.

Discourse is constraint free. This means that no one may force closure. The conversation continues until (ideally) every single participant is in full agreement. The larger and more diverse the group the more difficult and drawn out the process. Now clearly this is not a realistic model for all the decisions we associate with democratic government. The question then becomes which decisions should be made discursively and which by more efficient means? One answer is that the more the issue is a foundational one dealing with the legitimacy of the rules the more we are under an obligation to include all citizens.

But, in what sense is a face-to-face conversation between all citizens a feasible model of democratic legitimation? We cannot all sit in a circle facing each other as did the activists at Seneca. Nor can we be expected to devote the kind of time necessary to such an undertaking. Do we imagine a series of participatory face-to-face constituent assemblies? In large modern democracies this does not seem plausible. The problem here is that we are imagining practical discourse as a decision procedure with a determinate outcome. Thinking of discourse in this way will always bring us back to small, manageable groups like the women at Seneca Peace Camp. A deci­sion procedure implies a set of rules which govern closure. These rules tell us when the process is over—what counts is a fair decision that can be acted upon. Now, as a decision rule, discourse stipulates that full, rational agreement under the ideal conditions of discourse of all affected by a norm constitutes the point of closure. However, when translated into the real world of politics it turns out that this point can never be definitively reached. Because real agreements can never be perfectly universal they never settle a question once and for all. Through the idea of an ideal communication community we can imagine the conditions of a perfectly rational consensus and therefore the criteria of universal validity. But, as we can never attain the ideal in the real world, the process is one of degrees of approximation. Discourse is not a contract where there is a privileged moment of promising which is then binding on all parties for perpetuity.

Discourses must be understood as open ended and fallible. This means that discourse is ongoing and conclusions and agreements reached by means of discourse are always open to revision.

Once we move beyond small groups, the notion of consensual will- formation cannot be understood as the outcome of one conversation, but must be seen as the cumulative product of many crisscrossing conversations over time. The point is simply to highlight the diffuse nature a real practical discourse must have if it is to underwrite a legitimate social norm. On this reading, then, practical discourse is a long-term process through which citi­zens construct common understandings, not a decision procedure. Unconstrained discourse is highly inefficient. The closer our conversations come to embodying this ideal, the more inefficient they are. The more general the norm under discussion, the more diffuse, fragmented, and complicated the web of discourse and the longer the process is likely to take. In this light, it becomes difficult even to talk about a decision being taken in discourse; instead, we must visualize discourse as a process through which collective interpretations are constructed. If we cannot come to decisions in discourse, then what role does discourse play in demo­cratic politics? In conclusion I would like to briefly sketch the idea of a discursively formed public opinion as an answer to this question.