We said that a decision-making processes should make it practical to include input from as many stake-holders as possible; to ensure that all are heard; and to ensure that arguments will not be ignored just because they are complex or require specialized knowledge. Let us explain further where we expect such a tool to help.
The primary goal is to reach a decision; and to reach it in such a way that as much input as possible has been taken into account. There are two main issues to contend with: First, from the viewpoint of collecting opinions, as the number of stakeholders increase, the number of issues, proposals and arguments increases. If we want everybody's input to be taken into account, the task becomes staggering: who can even read, much less synthesize, that much information? Second, we are trying to reach a decision using this process: in most cases, diversity of expectations, beliefs and interests make it almost certain that consensus does not exist, and very likely that it cannot be reached; how then can people agree on course of action?
Let us first focus on the synthesis issue. To structure positions into an intelligible synthesis, the traditional approach has been to limit time, space and people. The people who can contribute information are a small number of representatives (including not only politicians, but also policy experts, lobbyists, etc.); the time and space was conspicuously the session of deliberative assembly, but also included debates in the mass media, and of course any policy advice within governmental offices. Also, within this time and space limit, procedures such as rules of order constrained the time that each person was allowed to contribute. In general, those three constraints can be relaxed using online deliberation: of course we still need a deadline for a decision, but anyone can consult and enrich the ongoing debate anywhere in the meantime. We have to turn to new organizing principles.
The first has to do with the structure of interventions: we need interventions to be structured, so they can be woven into a global structure. Ideally, interventions should be analyzed into propositions, projection of impact and evaluation. Propositions should be further analyzed into components so that they can be matched to similar propositions of other contributors, and projections and evaluations can be carried over if appropriate.
However, this highly formal analysis is difficult, and runs the risk of shutting out many potential contributors. It is important to allow both formal and informal contributions, and a way to progressively incorporate an informal contribution into the formal argument network. Who will do this, and why? I suggest that, insofar as formal contributors will benefit from support for their argument, they will have an incentive to show how supporting informal contributions tie into their argument. Of course, the original author of the informal contribution would have to approve of this specific analysis.
This ties in with the second principle: not only the structure, but the content of contributions would be evaluated. It is not enough to oppose a proposal to vote it down; any opposition has to be justified, by an analysis of impact. Such analysis can be evaluated according to likelihood and desirability, in a way similar to that used in the Delphi method.
The further advantage of this procedure is that it leaves a trail of projection, and contributors can be evaluated on their track record. This is a very delicate matter, that has been analyzed in depth by Tetlock [http://press.princeton.edu/titles/7959.html]Philip E. Tetlock , Expert Political Judgment: How Good Is It? How Can We Know?, Princeton University Press, 2006, 352pp.. This, in turn, allows a new form of expertise to arise, based on the capacity to evaluate impact in a verifiable way. I suggest that in such a context, we can re-introduce representation without curtailing the quality of the debate, in the form of vote delegation. So the third principle is argumentative accountability. It can take other forms: Accountability comes from commitment, not only to opinion but also to action. Contributors should be encouraged to promise contributing resources (effort, finances or otherwise) should their favorite option be chosen, to express that they are willing to support them more than intellectually.
So, to sum up: The prime characteristic of online deliberation is openness to new ideas; this requires an ability to synthesize contributions that, I argue, is best served by structured deliberation techniques. However, these formal techniques should be applied progressively to informal input. Options can then be discussed with substantive arguments, as opposed to simple votes. Delegated votes can be based on a reputation system and tracking commitment of proponents.