The number 1,000 seems to have some kind of charm when it comes to allotted bodies. There is of course the G1000 – “a Belgian platform for democratic innovation” backed by the renown of David Van Reybrouck. But more generally, there is somehow the notion that 1,000 is a good size for an allotted body. Supposedly, 1,000 is how big a body has to be in order to be “representative”. This intuition may be to some extent reinforced by the fact that opinion polls often use (or claim to use) samples of a similar size. There is also the fact that when one is surrounded by 1,000 people there is a feeling of being in the presence of a crowd and one becomes an anonymous, insignificant point in that crowd – and maybe that seems to reflect what membership in a mass community is about.

In fact, the number 1,000 is completely arbitrary. Its use in opinion polling is rather coincidental, and there is certainly no reason to use it when allotting political bodies. Indeed, the feeling of being lost in a crowd of 1,000 people is a strong indication that 1,000 is too many.

As is generally the case when considering the design of allotted bodies (and when thinking about sortition on the whole) it is most fruitful to consider the issue of body size via the model of extending self-representation. For the decision-making body to make policy that represents the interests of the people, two things have to happen:

  1. The body has to be internally democratic. That is, there has to be an equality of political power within the body.
  2. The membership of body has to reflect the population in the sense that its values and world view match those of the population.

Those two conditions generate two conflicting considerations: since large groups of people tend to generate spontaneous inequalities within the group, the first condition implies that the size cannot be too large. The second condition implies that the makeup of the body has to be statistically representative, so that it should not be “too small”.

How those two conflicting considerations are resolved may change depending on the circumstances. However, the first condition can only hold when the body is small enough to have an all-to-all communication mode. The question of how large a group can be and still allow for all-to-all communication within it is a question that has been discussed in the anthropological literature and the maximum size of such a group has been given the name Dunbar’s number. Of course, the number varies depending on the circumstances, but even under ideal conditions, this Dunbar’s number is no more than a few, very few, hundred people. This thus serves as an absolute upper bound for the size of the an internally democratic body. Bodies that are larger than that would inevitably have influential elites within them that would be able to manipulate decision making according to their interests and values, making the question of statistical representativity of the group meaningless. Even at this upper limit of a group of two or three hundred people, conditions have to be conducive to communication and to equality, and members should be fully invested in the process for it to function well.

As for statistical representativity: The notion that a body of, say, 200 people is “not representative enough” is essentially meaningless. There is no “standard level of statistical representativity” that needs to be reached. Calculations can be made of the chance that a minority of certain proportion will obtain a majority in a body of a certain size, but such calculations should not be taken too seriously because the political situation in any real decision-making body is much more complex than they allow. For example, it is true that with an allotted body of size N, any decision that is made with a majority of the square root of N is essentially too close to call. So a 90-110 split in a body of 200 is a marginal split that could turn out the other way if the sample for the body was redrawn.

But what of it? A decision has to be made one way or another, and to the extent that the close split does indeed reflect options that seem to the people at large almost as good, then it really matters little which decision is taken. If things turn out poorly, a subsequent allotment of the body, or even the same allotment with additional information, will reconsider and change course. The notion that good policy is about single decisions in high-stakes settings is part of the existing electoralist system and ideology, but it is in fact anti-democratic, because democracy puts the people in permanent control, so policy formation and re-formation is an ongoing process.

Thus powerful allotted bodies, like a parliament, should push against the upper bound of Dunbar’s number and have, say, 250 members give or take a few dozen. Less powerful bodies should be smaller. There is nothing “statistically wrong” with an allotted body with 30 members in the right circumstances. Such a smaller body could be useful for managing a media organization, for example, or for supervising a police force. Of course, sizes of allotted bodies, like other design parameters of those bodies, and like the entire political institutional system should be dynamically adjusted through the democratic political process rather than be set in stone once and for all – self-adjustment is an essential component of any democratic system.

6 Responses

  1. Size does matter… but it all depends on the role a body plays among the many tasks in developing and adopting policy.

    We can dispense with the Dunbar number, however. While the basis for this theoretical limit of social connections has been disputed (and some say refuted), it is not a relevant concept for a deliberative body. Dunbar’s theory only relates to the number of stable and significant relationships a person might have, not the number of people one can meaningfully interact with. If the number did apply, we would need to first subtract all of the family, friends and enemies a member already had before joining an assembly. There is no basis for linking the theoretical (and likely invalid) Dunbar number to a deliberative body size at all.

    Then we come to Yoram’s assertion that in large bodies there will inevitably develop elites and domination. Yoram asserts this is inevitable in a large body but can be avoided in a smaller body that allows face to face interaction. What is the basis of this claim? Don’t many small groups of friends have elites within them?… Where one member directs group activities and others placidly follow? Why can’t a well-structured large group use rules and procedures that prevent the development of elites (limits on speaking time, rotation of functions, etc.)? Size does not determine whether there are internal elites… process does.

    Size IS important however due to the issue of attention and rational ignorance. If a group is too large, many members will engage in cognitive free riding, and feel it isn’t worth the effort to think for themselves. Again, good process can help (members might hear the pro and con arguments from witnesses, but not know how any other member of the body is leaning, promoting self decision-making.)

    I agree with Yoram that smaller bodies that are less precisely representative, but that rotate and are replaced by subsequent bodies can make and change decisions that waggle around the middle but will ultimately regress to the mean. Thus, sortition can be a self-correcting democracy over time. Only monumental or irreversible decisions (to go to war, etc.) are deeply problematic for small mini-publics. These decisions may need larger mini-publics to trust their representativeness.

    One solution is to have a larger body that will hear pro and con arguments and vote yes or no on proposal washout debate among members. By not engaging in advocacy or argument themselves, any concern about elites within body, or “Dunbar relationships,” are moot. Smaller give-and-take deliberative bodies that craft and amend proposals might be smaller to allow intense face to face deliberation, but without final authority, their less accurate representativeness is less problematic. The key is to create different bodies that are suited to distinct tasks to protect and strengthen the democracy.

    Liked by 1 person

  2. Hi Terry,

    [I am not sure why you adopted the third person form of address.]

    > Dunbar’s theory only relates to the number of stable and significant relationships a person might have, not the number of people one can meaningfully interact with.

    I am not sure what the difference is. Collaborating with other people on creating policy should imply understanding their worldviews, values and interests. That to me means having “a stable and significant relationship” with them.

    > Size does not determine whether there are internal elites… process does.

    Process is important – in small groups, but no process can allow for an all-to-all communication in a large group. Nothing meaningful would come out of 1,000 5 minute speaking slots or a one-day rotation at chairing meetings. This is due to a cognitive constraint which is captured by the Dunbar number.

    > rational ignorance … cognitive free riding

    Disengagement may very well occur (although these standard economist names are misleading) but it is of secondary importance because even if everybody was perfectly attentive, there would still be no way for 1,000 people to share equally in the decision making process, and elites would form no matter what institutional arrangements were made.

    > One solution is to have a larger body that will hear pro and con arguments and vote yes or no on proposal […]. By not engaging in advocacy or argument themselves, any concern about elites within body, or “Dunbar relationships,” are moot

    The notion of having a process with multiple bodies (either symmetrical or with differentiated powers) is a separate issue with its own difficulties that needs to be analyzed on its own.

    In any case, having a separate body with an up-or-down voting power does not, of course, address the issue of potential elite control of the body which draws up the proposal and the pro-and-con arguments. If the policy agenda is controlled by an elite body then the up-or-down vote is largely meaningless.


  3. There are many possible procedures to prevent the rise of elite domination in a very large body. The concern arises because we are used to thinking of bodies that elect a leader, or an executive committee, etc. That is not needed. A prior random body can spend a lot of time designing an optimal set of anti-elite procedures. What follows is not a concrete “proposal,” but rather an example of one of many ways one piece of the process could be done.

    Any member who wished to address the body about some agenda topic, would submit it in writing. This would go to a smallish random assortment of other members who would vote it up or down, as being worth other members’ time to read. Because we don’t want to suppress minority views it only takes a small minority of yes votes to advance the piece. A huge number of these random groups would be constantly getting new comments simultaneously. The size of these filter groups, and the number of yes votes needed to advance an item would be adjusted depending on whether the full body agreed they were getting too many, or too few comments being advanced to them. Experience shows that only a handful of members will WANT to submit a comment, so there is never a problem of ALL members wanting to speak to all other members on all topics. However, to further avoid “elite” over-talking, it might be appropriate to require members who haven’t yet spoken to submit a comment now and again (if it is worthless, only the tiny filter committee will read it). But introverted members will have ideas pulled out of them.

    Elite domination is a result of process, not size.


  4. > Experience shows that only a handful of members will WANT to submit a comment, so there is never a problem of ALL members wanting to speak to all other members on all topics

    You are here in effect assuming a-priori that there would be an elite who would be setting the political agenda while most people would be passive listeners. Thus what you present as a solution to elite rule is in fact a presumption of elite rule.

    Furthermore, the notion that what determines power is the right to formally address a mass audience is obviously mistaken as well. This comment is formally addressing a mass audience – anyone on the planet can read it. The same is true for an editorial in the New York Times. Obviously, the two have very different impact in terms of political power.

    Fundamentally, your proposal rests on the mistaken notion that formal equality, or formal political symmetry, translates easily and automatically to democracy – i.e., equality in political rights. This is a fundamental error. Such a translation is fairly straightforward (but not trivial) in a small group but is far from straightforward in a large group. In fact, only by breaking the symmetry through sortition and concentrating decision-making power in the hands of a small group, can the formal political symmetry in a large group be translated to democracy.


  5. Since I mentioned Dunbar’s number, and since Terry claimed that it is both irrelevant and “likely invalid”, here is a reference to a 2011 paper by Dunbar which discusses the 150 number constraint and its relevance to social institutions:

    Constraints on the evolution of social institutions and their implications
    for information flow
    R. I. M. DUNBAR
    Journal of Institutional Economics (2011), 7: 3, 345–371

    Excerpt from the conclusions:

    [T]here may be optimal sizes and structures for institutions. If the organization is below about 150 individuals, then informal management structures may work very effectively. Of particular importance in this respect is the sense of community, of ‘belongingness’, that such institutions will have: as a result, individuals will feel a greater sense of
    obligation towards each other and may cooperate more effectively in achieving the institution’s goals. Shared cultural markers of various kinds may be important both in identifying other community members and in engendering a psychological predisposition to cooperate (Nettle and Dunbar, 1997; McElreath et al., 2003). However, once an institution’s size exceeds ∼150, then formal management structures are needed to solve two key problems associated, respectively, with different kinds of information bottleneck (I will refer to these as casual and deliberate bottlenecks). One is the fact that, in large organizations, information simply does not flow where it should, even if everyone is keen for it to do so. This is especially important in the case of the kinds of casual conversations that are often responsible for imparting crucial but quite unexpected items of information. Such casual information bottlenecks may be quite unintended, but they arise simply because of lack of opportunity for face-to-face interaction. This is the problem of having too small a watercooler: not enough people can gather around it. The second problem is that rivalries are more likely to build up because individuals owe loyalty to their immediate group (the 150) and this automatically creates what amounts to in-group/out-group effects whenever sections of an institution come into conflict. Indeed, irrespective of any conflict or rivalry that might exist, they might arise simply because, in very large organizations, a natural in-group/out-group division emerges between those who encounter each other often (the 150) and those who are, effectively, strangers (the rest). Unintended in-group/out-group effects of this kind will unavoidably give rise to reduced cooperation and the development of deliberate (as opposed to casual) information bottlenecks.


  6. As a rule of thumb, the magic number of 150 may (or may not) be a reasonable size limit… But there is no legitimate evidence supporting this popular-media exaggerated “Dunbar’s number.” It is essentially a “Just So” story, without scientific support. The number was only a mid-point of a conjecture by Robin Dunbar, a primatologist and anthropologist who suggested that maybe one could extrapolate optimal group size among primates based on neocortex size. Even if this mere conjecture were valid (which many scientists reject, noting that human culture and language make the extrapolation a nonsensical exercise), better data has shown that 150 is simply not a magical number. Also, if it WERE a good number, you would need to assure that all members of the organization had no family or friends (because it is these close relationships and mutual grooming that Dunbar’s number is in reference to – not organization size), if you wanted to use 150 as a significant cut off point for an organization’s size.

    The problems of larger organizations in terms of bottlenecks and information flows, etc. mentioned in the quote above may be valid, or nonsense, but hanging one’s hat on Dunbar’s number displays a fundamental lack of understanding about its derivation and status among researchers today. Malcolm Gladwell who popularized many things (including recently the use of lotteries!) inserted Dunbar’s number into popular awareness in his “Tipping Point” book, and that error has percolated along ever since.

    As two researchers who showed errors in Dunbar’s method recently concluded in an academic paper:

    “It is our hope, though perhaps futile, that this study will put an end to the use of ‘Dunbar’s number’ within science and in popular media. ‘Dunbar’s number’ is a concept with limited theoretical foundation lacking empirical support.”

    “‘Dunbar’s number’ deconstructed”
    Patrik Lindenfors , Andreas Wartel and Johan Lind
    Published:05 May 2021https://doi.org/10.1098/rsbl.2021.0158

    Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: