Part 1 is here.
Extension of self-representation
Like many other authors discussing sortition (Dahl, Leib, Zakaras, Fishkin, and others), Stone and Dowlen choose, then, to drastically downgrade sortition from a tool of radical democratic reform (as presented by C&P, or earlier by C.L.R. James) to an add-on to the electoral system. Such a retreat is certainly not warranted by the theoretical considerations discussed in the first part of the article. The claim that sortition can be expected to produce good government can be put on a much more solid theoretical foundation than the faulty intuitive argument provides. An alternative argument works by employing the properties of sampling in order to extend self-representation of the decision-making group into representation of the entire population. It goes as follows:
- A small group of people, under reasonably favorable conditions, is able to represent its own interests. This claim is not directly associated with sortition, but is rather a claim about the political dynamics of small groups of people in general. The claim is that when a small group of people, meeting on an a-priori egalitarian basis, has the opportunity to make collective decisions that would promote the interests of the members as they perceive them, then it will tend to do so. This is a situation which most people would be familiar with – group decision making in the family, within a group of friends or with colleagues. “A small group” is taken to be a group in which all-to-all communication is possible. The upper size limit of such a group would depend on the circumstances, but even under the most favorable circumstances a few hundred people seems like the most that would fit the description.
- Policy that promotes the interests of a small group of people which are selected as a sample of a larger group will tend to promote the interests of the larger group as well. Since the interests of a group selected as a sample of a larger group are typical of those of the entire group, policy that promotes the interests of the sample would tend to promote the interests of the group. In particular, if a certain policy promotes the interests of a majority of the members of the sample then that policy is likely to promote the interests of a majority in the population. There would be some obvious exceptions to this extension from sample to population. Policy that applies directly to the members of the sample in their role as members – their salaries for example – affects interests for which the sample members are very atypical. In a government by sortition such exceptions would have to be treated separately.
The extension-of-self-representation argument avoids the most problematic point of the mirroring argument by avoiding an appeal to the hypothetical “what the entire population would decide” construct. By avoiding this standard, it also avoids the need to make assertions about similarity of policies. The extension from the small group to the large group occurs on the basis of similarity of interests, not policy.
The validity of the extension argument can be tested empirically. If the system is working as the argument describes, most citizens would feel, upon examination of the workings and decisions of the decision-making body, that the policies enacted represent their interests. To measure whether this is the case, a separate body, also made up by sortition, can be put in charge of monitoring the decision-making body. This monitoring body can then decide whether the policy made by the decision-making body represents the interests of the population. The sentiment of the monitoring body should reflect the interests of the population.
The issues of preference aggregation and the supposed intractability results claimed by social choice theory apply, if at all, at the first point of the argument. While some sort of a procedure for aggregation of preferences needs to be employed within the decision-making group, the characteristics of any such procedure, including any drawbacks, are small-group issues that potentially affect its ability to represent its own interests. But if problems of self-representation exist then they affect any small group making decisions – they are not problems scale. Whether those problems are severe is something that every person should be able to assess based on their own personal experiences in small group decision making.
Finally, the argument also sheds light on the issue of exclusions from the sampling pool, which Stone mentions as a problem for the mirroring argument. Sortition proposals, including that of C&P, often specifically exclude certain groups of people from the set of those eligible for selection to the decision making group. The largest group, and the one most consistently proposed for exclusion is children, but sometimes other groups, such as criminals or the insane are also targeted. The extension of self representation argument indicates that there could be two grounds for exclusion. The first is when the interests of those excluded are considered illegitimate and therefore should not be represented at all – this is presumably the case for those suggesting to exclude criminals. The second ground is when the group is considered as deserving representation, but at the same time it is considered as being unable to represent itself – this is the case for children and the insane.
Arguments which aim to diminish democratic expectations are part of the arsenal of opponents of democracy – both elitist opponents that wish to justify existing inequalities of political power, and anarchist opponents that claim to aspire to a society without any form of large scale powerful organization. Those arguments are not a-priori without merit. On the contrary: modern democratic dogma notwithstanding, large scale democracy is not a naturally occurring phenomenon – it exists nowhere today and existed (as far as I am aware) only in some Greek poleis in the 5th and 4th centuries B.C., if at all.
In view of that, an attitude of skepticism toward claims that a certain technique provides a basis for a democratic government is sensible. But this background also motivates breaking with the established institutions and looking for a radical alternative, rather than attempting to patch those institutions in various ways. Sortition provides such a radical alternative. The discussion above indicates that this alternative is more than an idle hope – it is a concrete goal that can be reasonably expected to result in democratic government.