Athenian Democracy Reincarnate

Athenian democracy involved a combination of sortition (boule, juries and most magistracies) and direct democracy (ecclesia). Sortition fell into disuse in large modern states and direct democracy was replaced by representative elections. There has been a flurry of proposals recently for the reintroduction of sortition, but it is unclear how — or indeed if — this can be reconciled with mass democracy, as the latter is elitist, populist, undeliberative and frequently hijacked by rich and powerful elite interests, leading to sharp exchanges on this list. However a recent debate suggests an acceptable compromise, which I outline in this post. There would be three stages to the legislative process, with the relevant Athenian institution in parentheses:

1) Policy initiation (boule)

Terry Bouricius has suggested a multitude of randomly-selected forums for the development of policy proposals. In Athenian democracy there was a reasonable chance that an individual might be selected for the boule (hence the 2-term lifetime limit), but large modern states would require a multiplicity of forums in order to even approximate this. As policy forums would fulfil an advocacy role, the sortition would be on a voluntary basis, and would largely attract activists and interest groups. Each of (say) 50 fora would contain several hundred members and at the end of the deliberative period each would perform an internal votation (ranking of preference by secret ballot) to determine the group’s favoured policy proposal out of all of those discussed.

2) Public votation (ecclesia)

The individual policy preferences of each group would be offered to the public, who would choose (and rank) their selection in a public ballot (an aspect of Swiss democracy known as votation). As this stage would be subject to considerable influence from lobbyists there would be very strict limits on campaign advertising, and public broadcast slots would be allocated equally, a process that works well in the UK. The (say) top ten policy preferences would be selected for the final stage:

3) Deliberative scrutiny (nomothetai)

Random selection for the final deliberative body would be mandatory, as in jury service, and it would be considered a great honour and civic responsibility to be selected. Policy proposals would be introduced by their original (stage 1) advocates and would be opposed by members of the house of advocates, composed of members of a wide range of civic society and interest groups. Membership of the house would be selected randomly from a shortlist of candidates put forward by these groups, who would then choose which debates to participate in, depending on their particular skills and interests. This model of voluntary advocacy has been demonstrate to work reasonably well in the UK House of Lords, where the debate is better-informed and less partisan than in the Commons.

The (temporal) space between the three stages would be filled, as per Condorcet’s model, with public deliberation, and would presuppose the break-up of media monopolies.

This proposal is based on a strict separation of legislative and executive powers, with government ministers appointed on merit. Items of secondary legislation (statutary instruments) would go straight to stage three, and the randomly-selected house would also approve ministerial appointments (and remove ministers via censure motions).

90 Responses

  1. Is the third chamber elected at random from the citizenship or from a shortlist of voluntary advocates? The text seems to be a bit confusing…



  2. I will answer Yoram’s questions of me from the Sortition in Spain thread here. Here are his first related questions…

    Yoram wrote:
    “In general, why would you want to limit the power of the final decision making allotted chamber to an up-or-down vote on a given set of proposals? Having put in the effort of learning the ins-and-outs of a particular subject, why shouldn’t they introduce changes into the proposals as they see fit? Why would you assume that some self-selected groups would do a better job in designing those proposal?”

    The ultimate allotted deciding group should be as fully descriptively representative of the population as possible. The notion is that the task should be semi-mandatory, rather than relying on stratified sampling. They are asked to exercise common sense, and not spend large amounts of time as a group to become experts (which moves them steadily away from being representative). Any debate and give-and-take between members of this group also would risk the proven psychological dynamics of polarization and group-think (people defer to the more dominant). Surowiecki’s _The Wisdom of Crowds_ explains how optimal decision making can occur by secret voting among diverse opinions. This garners the most privately held information, and maximizes the group intelligence. This potential benefit of sortition would be lost if the group began internal debate.

    On the other hand, if we forgo the give-and-take of internal deliberation we also lose a different benefit of sortition, which is the discovery of win-win synthesis opportunities by hearing from diverse members. Therefor, I propose that this refinement of proposals (amending, mixing and matching elements) be done by a separate sortition group (I called them a Review Panel). It would be unreasonable to expect enough average individuals, who were mandated to serve, to become de facto experts. That is why I proposed the Review Panels be made of randomly selected volunteers, who are willing to become experts on the issue.

    Finally, the raw material (proposals) for the Review Panels would be drafted by advocates, experts, self-interested, and loudmouths volunteering for these Interest Panels. Here is where anybody who cares to, can participate…The only role sortition plays here is mixing individuals to assure diversity, rather than in limiting who gets to join. The goal is to allow any citizen to contribute to the democratic process, and maximize their contribution by letting them focus on the issue they know about. I think this will produce more useful raw material for the higher bodies to digest.

    Thus I do not see the Review Panels using a popularity voting system to pick which proposal advances, but rather for them to amend and shape a potential finished proposal from all of the Interest Panel proposals. Likewise, I don’t think Keith’s referenda process can fulfill this function either. Why revert at this mid-point in the process to uniformed popularity. If there is a felt need to involve the full citizenry (beyond those who are willing to join an Interest Panel), then I would suggest that either we allow initiative referenda to START the process of legislation or come at the END… the AC could adopt laws directly by a super-majority, and refer closely voted measures to referenda.


  3. Jorge: The third stage is selected at random from the citizenry and is, as Terry confirms, semi-mandatory

    Terry: I was trying to follow as exactly as possible the three stages of Athenian democracy, in which the popular vote comes in the middle. As we can’t fit 60-300 million people into the agora we have to use votation. The problem with Review Panels (similarly to Kevin O’Leary’s proposal) is that they would be a very easy target for lobbyists to corrupt. Why not just have each of the loudmouth panels select their chosen policy and then leave the final choice to the people at large?


  4. Keith,

    Just for the record, let me point out that – as is your habit – you are misrepresenting here other people’s positions. The “acceptable compromise” you present is your own creation and is acceptable to you alone.


    The way you put things now, it seems that the final decision making body would be weak – being deprived of the benefit of discussion, and discouraged from forming more than a layperson’s understanding of the matter at hand. It would be little more than the electorate writ small, with many of the associated shortcomings.

    The “review panel” – which is actually empowered to phrase proposals – would wield real power, but whether this power is wielded by the panel members themselves and is shared equally among them (or by external elites and elites within the panel) would depend on the details of the office – scope, term, authority and resources.

    The interest panels would probably simply be a “vox populi” backdrop that provides convenient material that can be used during the review process to justify various decisions but can be safely ignored or misrepresented.

    (You have not specified how the process would be initiated, but that would also be a point of application of political power.)


  5. I’ve certainly compromised my position considerably, and it would be acceptable to anyone who was prepared to acknowledge that democracy has always required the assent of the people. I sincerely hope that I’m not the only democrat in this forum.


  6. Keith,

    I’m intrigued by your proposal, but confused about your three stages of the legislative process. Here are my questions:

    Policy initiation:

    1. Would your 50 fora, of several hundred members each, divide themselves into committees for the purpose of drafting proposed legislation? The work of designing and writing proposed laws is extremely difficult to do in a large group.
    2. At the end of this stage, what would be the end result?
    3. Would each forum choose only one proposal?
    4. What is “votation,” and how is it different from “voting?”

    Public votation:

    You write that “The individual policy preferences of each group would be offered to the public, who would choose (and rank) their selection in a public ballot (an aspect of Swiss democracy known as votation).”

    5. What does “individual policy preferences” mean? One proposed law from each forum?
    6. What does “choose and rank their selection” mean? Would members of the public be ranking all the proposed laws in order of preference, from the ones they wanted most to the ones they wanted least?

    Deliberative scrutiny:

    7. What would be the end result of this stage? Would it be “yea or nay” votes on each proposed piece of legislation?


  7. 1) This is Terry Bouricius’s proposal, so I will leave it to him to respond, but I think the idea is that these would only be outline proposals (akin to public e-petitions), rather than draft bills.

    2) & 3) Terry’s suggestion is that they would then go to a Review Panel for winnowing down, my suggestion is that each forum should choose (say) one proposal (by secret vote) which would then go onto the ballot for a public votation.

    4) Votation is a contemporary Swiss term whereby “the citizenry choose between several legislative options that are either proposed by the government or put forward by popular initiative” (Gil Delannoi, in Delannoi and Dowland, Sortition: Theory and Practice, 2010, p.15). I’m less convinced that government bills (secondary legislation and statutory instruments) should go through this route (assuming that the executive role is no longer a political one).

    5) Yes

    6) Yes, and presumably only the top “x” would get through to the deliberative scrutiny stage, depending on the time frame involved.

    7) Yes

    Many thanks for the chance to clarify these issues (would you like to identify yourself?). I look forward to Terry’s response to 1, 2 & 3 — to my mind for a system to describe itself as democratic, there has to be at least one element of public acclamation. To leave it all to sortition would be klerotocracy, not democracy.



  8. Yoram,

    As I imagine the Review Panels, they would be roughly equivalent to the allotted chamber you favor. They would engage in joint fact-finding, debating, amending, and drafting of legislation. They would, however have their agenda set by a separate process, such as referenda, polls, or a sortition panel (not by political parties, elites, nor by themselves). As you say, this body has the bulk of power. They could use the raw material developed by Interest Panels (and remember these would be randomly diverse groups, so any proposals that emerge from them already have gone through lots of give and take), or not as they wish. They would have professional staff, fact-checkers, and access to experts (perhaps selected by a different sortition administration and oversight panel).

    The Final sortition body, is merely a final check to make sure the Review Panel has not been corrupted, or gone off-track through a process of group-think. Any protections against lobbyists or corruption that you would favor for your ideal Sortition legislature would be applied at this stage. The main difference between the Final Deciding Assembly and a popular referendum, (other than the obvious issue of sheer numbers of participants) is that the assembly members would hear structured pro and con presentations and devote more than aday to consideration (rather than the 20 seconds spent by most voters in a referendum).


  9. Keith and anonymous,

    I imagine the Interest Panels would do far more than come up with broad concepts (e-petition-like). Instead, I would expect these panels to have experts, advocates, and deeply engaged individuals (not representative of the population – but of that segment of the population that are equivalent to the Ho Boulomenos of Athens. These bodies, however would be diverse, including opposing perspectives. Many such “fora” would fail to find a joint compromise to advance to a Review Panel, but presumably some would.


  10. Terry: [Review Panels] would, however have their agenda set by a separate process, such as [a] referenda, [b] polls, or a [c] sortition panel (not by political parties, elites, nor by themselves).

    There is a marked difference in the democratic status of a/b and c. Note also that the powers of the review panel would be so extensive that it’s hard to imagine exactly how it could be immunised from lobbying and corruption (this would not be the case if the winnowing process was done by votation). It would be prudent therefore to opt for a self-correcting mechanism, rather than one that necessitates an important role for the police.


  11. Terry, if (as you argue) the Interest Panels would be self-selecting, then why not allow the public to choose which options it wants, rather than leaving it to a derivative of these fora? Do we really want to replace one constitutional settlement that “totally excludes the people in their collective capacity from any share in governing” with another that does exactly the same thing?


  12. Here is my underlying thought process…
    1. Elections of representatives distorts the agenda and proposed legislation in the interests of adversarial gamesmanship, and the elite.
    2. Referenda ask an inevitably ill-informed public their off-the cuff opinions as shaped by elite interests through advertising.
    3. The only way to enable informed decision making in the public interest that has at least the possibility of being free from the corrupting influence of elections and elite is through sortition bodies. (Yes, there will be attempts to lobby/corrupt them, but good institutional design MIGHT protect them, whereas this is impossible in adversarial elections situations.)
    4. Mandatory sortition bodies would have too many people who were not willing or able to engage in law drafting.
    5. Law drafting can be left to those who volunteer, so long as these volunteers do not also have the power to adopt the laws.
    6. The only roles I see for the general population are: a) legitimizing the existance of a sortition process in the first place through a popular vote and b) referenda for special cases where the sortition body has already perfected a proposal, but believes the law is only advisable if the public will support it, and the public can decide to trust them or not.


  13. OK, but that puts you (along with John Burnheim and Yoram Gat) in the company of 2,500 years of reactionary thinkers who wish to totally exclude the people in their collective capacity from any share in government. Personally I would prefer a solution that did not require chucking out the baby (democracy) with the bathwater. I can see that you, like John, have been wounded at the front line of democratic practice, but I think you are overreacting. I would like to think that we the public are capable of choosing sensibly between the various options thrown up by the Interest Panels, and even if we exercise our choice poorly, this will be corrected by the final stage of allotted scrutiny. Just in the same way that institutional design could protect allotted chambers from elite influence, the integrity of the public votation process could also be protected by strict limits on campaign spending (this works reasonably well in the UK).

    I also think that the chances of successfully introducing yet another scheme to totally exclude the people in their collective capacity from the political process has 0% chance of being implemented.


  14. Keith and Terry,

    My name is David Schecter, and I’m the “Anonymous” who posted a question in this thread last month (that was a mistake – I had meant to enter my name). I’m very interested in sortition and deliberation, and I’m on the advisory board of a project (the “Our Democracy” initiative of the Greenlining Institute in Berkeley, California) that’s looking at possible reforms to the California ballot initiative process, including something like the Citizen Initiative Review recently enacted in Oregon.

    I’m very interested in your proposals, and want to fully understand them. It will help me if I can ask you about one aspect at a time –the stages, then the actors and roles, and then the processes. So, I’d first like to get clear about the stages. Here’s my understanding of the stages that you two have proposed:

    Policy initiation (developing proposals – or is it bills?)
    • Deciding which issues to write proposals for
    • Formulating proposals
    • Writing bills (or does this happen later?)
    • Reducing the list of proposals (or is it bills?)

    Public votation (choosing which bills should be voted on)
    • Advocacy
    • Deliberation
    • Votation (deciding which bills are worth voting on)

    Deliberative scrutiny (voting on bills)
    • Advocacy
    • Deliberation
    • Voting (enacting or rejecting bills)

    To what extent is this correct? Also, at what points is there review and revision of proposals?


  15. Terry and myself have slightly different positions on this (I argue for a combination of sortition and direct/elective democracy, whereas Terry argues for a hierarchy of sortive bodies):

    1) Policy Initiation
    I think we both agree that the initial stage is for policy proposals rather than bills. If the source is multiple sortive chambers, then each chamber would be authorised to put forward a limited number to the next stage.

    2) Reducing the List of Proposals
    This is where Terry and myself have agreed to differ. Terry argues for the decision to be taken by a second-level sortive body whereas I would prefer a public votation. If the latter then advocacy might be limited to a short website statement as in the UK Government e-petitions system. Deliberation would take place in the public arena. The proposals that receive the most public votes would then move to stage 2a.

    2a) Preparation of Parliamentary Bills
    The original (stage 1) advocates would prepare the bills along with officials from the relevant government department. This would be the primary stage for review and revision, and for calculation of the fiscal entailments by the Treasury (who, in my proposal, would be legally bound to a balanced budget over the financial cycle).

    3) Deliberative Scrutiny
    Allotted chamber for deliberation and voting. In my proposal advocacy for (and against) the bills under consideration would be provided by an independent house of advocates, whereas Terry would prefer allotted representatives to play a more active role. As well as passing or rejecting the bill, the allotted chamber could vote to return the proposal to 2a for revision.

    Hope that helps; no doubt Terry will respond in due course.


  16. Keith,

    Thank you – that does help. Here’s a revised list of the stages.

    Policy initiation (developing proposals)
    • Deciding which issues to write proposals for
    • Designing and writing proposals

    Votation (reducing the list of proposals)
    • Advocacy
    • Deliberation
    • Votation (deciding which bills are worth voting on)

    Writing and revising bills
    • Drafting legal language for the selected bills
    • Review of the selected bills
    • Revision of bills as needed
    • Calculating fiscal consequences

    Deliberative scrutiny (voting on bills)
    • Advocacy
    • Deliberation
    • Voting

    Please let me know if this is accurate. After Terry responds, I want to add some of my own thoughts, and then ask both of you about the actors in each stage, and their proposed roles.

    Also, I’m working on a table that summarizes legislative stages, actors, roles, and processes. Can you tell me how I can incorporate a table in a post on this blog?


  17. Certainly works for me, await Terry’s response. Need to check with Yoram if it’s possible to incorporate tables in comments.


  18. David,

    I should first say that my design is not yet fully “thought through.” It is a work in progress, and I offer it merely as a possible plan…

    My notion is that the initiation might happen this way…

    A sortition meta-legislative body can review polling results, submissions of petitions, and listen to different experts, and then decide some particular issue needs to be tackled, and mandate the formation of a sortition body to deal with it. (This is in lieu of the votation process Keith advocates).

    Then, anybody who wants to participate, can volunteer to join a randomly mixed group of citizens who will try to hammer out a specific bill. If it is a hot issue, there might be many separate such groups, each trying to draft a bill. My notion is that we can’t expect drafted citizens to draft bills…this is a task for volunteers and paid professional staff. Because they are volunteer, they will likely deviate from ideal descriptive representation of the public. This is okay, because they are not ADOPTING any bills.

    These mixed drafting bodies will likely include experts and die-hard advocates for conflicting policies. But their work can only advance to the sortition deciding body if they can agree (perhaps by some super-majority) on a draft bill. They also have a built-in interest in compromising enough to craft a bill that has a chance of ultimate adoption.

    I next envision what I call a Review Panel …a sortition body that is descriptively representative, that may revise bills that have been submitted. This is a traditional deliberative body. Here we want more accurate representativeness, to assure all segments of society can weigh in. This would function much like any legislative body.

    Finally, I agree with Keith that the ultimate decision to pass or reject a bill should be handled like a trial, by another sortition chamber where pro and con arguments are delivered to a descriptively representative body that does not engage in debate (to avoid group-think and other distortions, such as corruption), and simply votes by secret ballot.


  19. Terry,

    If I understand it correctly, your stages go something like this (again, without specifying the actors).

    1. Deciding which issues to write bills for
    • Taking polls, submitting petitions
    • Expert testimony
    • Deliberation
    • Deciding issues

    2. Writing and revising bills
    • Designing bills
    • Writing alternative bills
    • (If there are too many alternatives, selecting a smaller number)
    • Reviewing bills
    • Revising bills

    3. Deciding which bills to enact
    • Advocacy
    • Deliberation
    • Voting

    Is that correct?


  20. Basically… except that the word “deliberation” has two meanings.
    Deliberation as in debating goes in step two as part of o0ffering amendments to bills and arguing the pros and cons of the amendments.
    Deliberation as in the sense of internal weighing of pro and con arguments by each individual before casting a secret vote occurs in stage 3.


  21. Thanks again for the clarification.

    If I understand correctly, both you and Keith are making what I think is a very useful distinction between “advocacy” (arguing for and against proposals or bills, as lawyers do in trials) and “deliberation” (the decision makers discussing the pro and con arguments before deciding, as juries do in trials).

    What gets a little confusing is to use the word “deliberation” both for individual weighing of pro and con arguments (as you describe), and also for shared weighing of pro and con arguments through conversation (as most of the “deliberative democracy” people advocate).

    As soon as I work out how to upload tables, I want to post a table that has the stages in one column, possible actors and roles in another, and possible processes in another. Hopefully that will make it easier for people to get their minds around all the details.


  22. Yes, the two uses of the word deliberation are really confusing. This is because there are two distinct political languages in use — the Anglo-American and the Germanic. In the former the derivation is from the Latin ‘liber’ (weight), hence the emphasis on silent inner deliberation in (for example) Rousseau, Fishkin’s Deliberative Polling experiments, and the trial jury (jury room debate is only necessitated by the legal requirement privileging consensus). The Germanic tradition derives from ‘deliberative stimme’ (deliberative voice), hence the emphasis on discourse and consultation in the German-derived literature on deliberative democracy (Sintomer, 2010a, p. 36). It’s important to recognize that we are dealing here with two distinct deliberative traditions, one (Germanic) emphasizing the importance of good discussion and the other (Latinate) the decision of a collective body (ibid, p. 47). Deliberation in the Anglophone literature usually conflates both meanings.

    Habermasian deliberative democracy relies on rational discourse to arrive at the (epistemically) ‘best’ outcome and is not particularly concerned with representativity per se. By separating out advocacy and judgment Terry and myself are attempting (albeit in different ways) to retain the best of both traditions. To my mind the advocacy element should be supplied by conventional elective or direct democratic mechanisms (including votation), whereas Terry (having undergone years of bitter experience of the dysfunctional nature of elective institutions) argues for a hierarchy of sortive bodies. We are both agreed the final ‘deliberation’ (in the judgment sense) should be taken by an allotted assembly, where representativity is privileged. We do need some help in resolving this issue and I’m delighted that you have joined in the debate.


  23. Keith,

    Thank you! I am delighted to be part of this conversation, and glad to help in any way I can.

    What you write about the two different origins of “deliberation” is fascinating – I had no idea!

    It sounds like you and Terry are both proposing a distinction of roles between “advocates” (whose job it is to present the strongest possible case for or against proposals, but who don’t make decisions – analogous to lawyers in trials) and “decision makers” (analogous to jurors – they hear the arguments, and then either discuss the arguments together before voting, or else proceed to voting with only “internal” deliberation).


  24. Yes indeed, Terry and I agree on this. The distinction is derived from Federalist 10: “No man is allowed to be a judge in his own cause, because his interest would certainly bias his judgment, and, not improbably, corrupt his integrity. With equal, nay with greater reason, a body of men are unfit to be both judges and parties at the same time.” Parties (to Madison) equate to factional interests. The other parallel with the founders is that the descriptive representation afforded by sortition (the judges) fits well with the Antifederalist argument that the legislature should be a portrait in miniature of the whole nation.

    Where we differ is on how the advocates should be selected, Terry argues for sortition whereas I take Madison’s use of the word “party” literally — the agenda should be set via elections or some kind of direct-democratic initiative (petition, votatation etc).


  25. Encouraging an open discussion and active inquiry is essential to getting to informed and rational decisions. Sticking decision-makers in a passive situation in which they are dependent on mis-information provided by interested parties is a recipe for manipulation and poor results.


  26. Yes that’s true. The problem is (as Madison understood so well) that the influence of interests is very hard to eliminate. The liberal (Madisonian) remedy is to seek an artificial balance of pros and antis within elite advocacy; whereas the republican (Machiavellian) remedy is to eliminate interests by restricting debate to a random sample of the plebs. The problems with the first approach are exactly as you catalogue them; the problem with the latter is it’s 1) hard to reconcile with democratic norms (Pitkin, 1967, p.145); and 2) impossible to eliminate the intervention of sinister interests as the lottery effect only sanitises ex ante.

    The current scholarly consensus is that the American founders were more republican than liberal, yet the republican reliance on disinterested virtue failed immediately. I’ve been re-watching the HBO John Adams series and noted that factions were emerging even before the first presidential election (Hamilton allegedly fixed the votes in the electoral college to give Washington a clear majority). One can argue that this was a consequence of the reliance on election, but an all-sortive alternative would be as hard to sanitise for the above reasons, hence my preference for a hybrid solution. The hybrid solution also has the practical advantage of being less threatening to existing power holders.


  27. > The problems with the first approach are exactly as you catalogue them; the problem with the latter is it’s 1) hard to reconcile with democratic norms (Pitkin, 1967, p.145); and 2) impossible to eliminate the intervention of sinister interests as the lottery effect only sanitises ex ante.

    Well, I am looking at Pitkin’s book, p. 145, as I am writing this and I have no idea what you think you are referring to regarding “democratic norms”.

    As for eliminating intervention: It is obviously true that interested parties will always try to exert undue influence over decisions. It is also obviously true that minimizing undue influence is a difficult, ongoing task in which sortition is but one tool. But it is also just as obviously true that having such undue influence built into the system through elections, “expert advocacy”, or any such elitist mechanism is completely antithetical to the democratic objective.

    Madison was, of course, openly antagonistic to democracy and therefore his choice of mechanism – elections – was understandable. But for those who claim to have democratic objectives in mind to advocate elitist governance mechanisms is incoherent.


  28. The Pitkin citation was just a shorthand reference to the many posts that I’ve already made distinguishing between aggregate functions (voting) and individual functions (introducing and advocating interests) and Pitkin’s argument that descriptive representation only applies to the former.

    Regarding the “incoherence” of including (in a controlled way) elite interests, this has always been an important aspect of liberal democratic theory. Although you may find Mandeville’s argument on the link between private vice and public good ethically unattractive it certainly isn’t incoherent. No doubt the principle behind variolation, inoculation as a method of purposefully infecting a person with smallpox (Variola) in a controlled manner so as to induce immunity against further infection is equally “incoherent”, but it works rather well.

    Whilst republicans privilege civic virtue over interests, nevertheless republicanism always involves a mixture of aristocratic and democratic forces, so you are casting the incoherence net rather widely. There are no examples of pure democracy in large states, and the coherency claims of implementing this model by sortition alone are undermined by Pitkin’s argument. But we’ve been over all this before, and it always seems to generate a lot more heat than light.


  29. Yoram,

    My (tentative) preference for having the final decision made by a near-mandatory service, fully descriptive allotted body that does not amend or debate, but rather listens to pro/con arguments (and asks clarifying questions) is somewhat different than Keith’s. I am not relying on Pitkin, but instead on these two facts…

    1. We can’t expect “drafted” citizens to put in the effort and time needed to PERFECT legislation. But we can ask them to weigh it in relatively quick order. The writing, amending, and perfecting will inevitably fall to those who are willing to do the work (volunteers), and this group will inevitably be less representative – though I still favor sortition in selecting these volunteers (rather than Keith’s notion of election, or chamber of party advocacy.)

    2. Credible research has shown that once a group engages in discussion, various psychological dynamics distort decision making. While Fishkin and other deliberative democracy advocates focus on the benefits of this deliberation, it can also have negative effects… through group-think, the desire to “fit in” with the apparent majority, deference to elite members, etc. If the most articulate, charismatic, handsome, tall male in the group favors policy X, then policy X is far more likely to prevail. It is better for the decision making process if people use their own judgment and nobody knows which way this “elite” member is going to vote.



  30. Terry,

    > 1. We can’t expect “drafted” citizens to put in the effort and time needed to PERFECT legislation. But we can ask them to weigh it in relatively quick order.

    Selecting between pre-defined options is not inherently easier than – and could be much more difficult than – designing new, improved options. In general, there is no way around requiring a representative sample to put significant time and effort into studying a problem in detail. To whatever degree such spending of time and effort is avoided, the decisions made consequently would be less informed and less considered, and thus of lower quality.

    > 2. Credible research has shown that once a group engages in discussion, various psychological dynamics distort decision making.

    Discussion is an indispensable part of high quality decision making. Of course, it is not perfect and prone to various problems, but the alternative – decision without discussion – is obviously inferior by far.

    As for the disproportionate influence of elites – in a situation of open discussion in the appropriate settings (e.g., sufficient resources and support staff for all members, equal standing and equal speaking time), such disproportion can be minimized. In a no-discussion setup, on the other hand, elite influence – through control of the agenda, of the options provided, and of the information provided – is a virtual certainty.


  31. I’m sure Yoram’s last claim is open to empirical verification. To take one obvious counter-example, in the case of the Chinese DP, the public chose not to follow the clear preferences of the party elite. If this is the case in a country with a single political elite, one would hope that in postmodern liberal societies, where there is no homogeneous elite, a sortive assembly would do even better. It would make for an interesting political science project for someone to study the 20-year history of the DP to see whether or not Yoram’s claim is justified.

    Of course there would need to be some preliminary work in theoretical sociology, defining exactly what constitutes an elite, as the term has been used rather loosely on this forum. My preferred starting point would be Harrington’s cognitive definition — elites are better at articulating interests discursively (and rhetorically) than the general public. I think this would provide a better fit with modern societies — where there is a weaker fit between cognitive elites and economic derminants than in classical models derived from Marx, Mosca, Pareto or Mills.


  32. Keith,

    you wrote: “in the case of the Chinese DP, the public chose not to follow the clear preferences of the party elite”

    I read Fishkin’s article on this some time ago, so I may not remember it right…but I thought he said all of the options presented to the allotted group were okay with the party, and that Fishkin’s group rather than the party ran the meetings, so that any subtle (let alone “clear”) preference by the party elite would not be evident to the participants.

    Do you have a citation for this?


  33. An arrangement in which the elites set the agenda, set the options and determine the information provided to the allotted chamber is presented as a counter-example to elite control – truly an argument worthy of Keith Sutherland.


  34. The reference is to Fishkin’s book When the People Speak: “While [local party leaders] expressed surprise at the public’s priorities, the top twelve projects selected by the people were constructed” (p.109). The party leaders were surprised because their own preference was for grandstanding projects like a “fancy town square” (p.108) whereas the DP opted for sewage works, nevertheless party leaders implemented the decisions of the DP in full. (Note that I’m not suggesting that party leaders did not approve of the outcome, only that they were surprised, as they had different priorities.) This leads Fishkin to suggest that the Chinese DP showed the “first glimmerings of another model . . . an example for public consultation in many settings around the world” (pp.155-6). The “different model” that Fishkin is alluding to is a Harringtonian combination of “deliberative democracy by the people themselves and elite deliberation by an elected body” (p.156).

    Personally I think it would be possible to secure such a positive outcome within the liberal democratic framework, rather than the one-party state. The policy options would be decided by competitive elections (or petition/initiative followed by votation), the final deliberation and choice being left to an allotted chamber. The Chinese leaders were pleased with the process, and elected politicians would be unlikely to see this as a serious threat. No doubt those who argue for sortition-only solutions would not like to suggest anything that pleases the political class, but there is something to be said for proposals that have a chance of being peacefully implemented, without having to recourse to the barricades.


  35. I’m starting to write up a separate post that first attempts to lay out a basic organizing framework of questions about the design of a legislative system that incorporates sortition (for example: what are the necessary activities? what are the possible actors for each activity? what are the possible processes for each activity? etc.), and then the main answers to those questions that have been proposed in conversations like this one. My clarifying questions to Keith and Terry about stages of lawmaking were a first step in that direction.

    Before I do that, I’d like to read what people have already said on this subject in this forum, besides this conversation and Yoram’s piece on design parameters. Can anybody suggest links to other relevant conversations?


  36. David,

    We had many discussions that could be relevant to these issues. One such discussion was in relation to Alex Zakaras’s proposal: 1, 2, 3. (Some of those discussions become quite heated.)

    As you can see in those threads, my own opinion is that attempts to find ways to limit the powers of allotted bodies by only allowing them certain functions are (almost always) misguided. Giving non-representative bodies decisive authority in certain areas is simply anti-democratic.


  37. The competing “pluralist” view is that democratic representation comes in various shapes and sizes, and that the statistical representation afforded by sortition only works for aggregate functions like voting. Although we are all aware of the shortcomings of elective representation and direct-democratic procedures, nevertheless this is an essential element for non-aggregate functions such as policy initiation and advocacy. Most of the posts on this site involve some variant of this argument. Unfortunately there has been little or no progress towards bridging the gap between these two perspectives.


  38. Yoram and Keith,

    Thanks to both of you for the clarifications. I will read the threads that you suggested, Yoram – I glanced at them and they look very interesting.

    Personally, I’m undecided as yet about whether randomly selected bodies should perform all legislative functions or just some, whether there’s a useful role for elected legislatures, etc., so I won’t try to advocate. Instead, I hope to make some small contribution toward clarifying and simplifying the areas of agreement and disagreement, and the rationales behind them.


  39. > aggregate functions like voting

    Obviously, voting is not any more of an aggregate action than, say, agenda setting.

    Individuals vote and then their votes are aggregated into the vote tally. Analogously, individuals propose laws, or sections of laws, or amendments, or rewordings, and their proposals are aggregated into bills that are put up for a vote. In both cases an aggregate activity is made of individual actions through some sort of an aggregation procedure.


  40. Not so. Every vote is numerically equal — irrespective of whether it is cast by Winston Churchill or Mr. Pooter, whereas the illocutionary force (persuasive power) of their respective speech acts — such as introducing and advocating legislative proposals — is anything but, as it depends on the rhetorical and status differences of the individuals involved. (If I had a dollar for every time I’ve made this point on this forum then I’d probably qualify for membership of your elite.) Votes are digital, whereas speech acts are analogue — votes are equal in every respect, whereas speech acts are quantitively and qualitatively unequal.

    The only reason “deliberative democrats” downplay this rather obvious point is because they are no more interested than Burke in representative equality, as they only want to get the epistemically right answer, whereas I would hope those of us posting here would respect the title on the masthead of this forum.

    Democracy presupposes that the majority view constitutes the right answer and (in large political communities) this requires fidelity of representation. Statistical representation only applies in aggregate, hence the need to limit the behaviour of statistically-representative assemblies to aggregate functions (i.e. deliberation in the latinate sense of weighing and voting). Active political functions require elective or direct-democratic mechanisms.


  41. > Every vote is numerically equal […] whereas […]

    That, Keith, is a very different assertion from your previous assertion (that voting is an aggregate activity while, say, agenda setting is not).

    I know that for you such minor differences do not matter, that your positions are fluid, and that when you make various assertions it is a mistake to assume that you are seeking to make a substantive claim, but as a matter of personal habit I do tend to expect that words and meanings will generally be in correspondence.


  42. I’m sorry Yoram, but I’ve no idea what you mean. “Aggregating” I take to mean counting, summing, adding up etc. and this requires quantifiable units (such as votes). You cannot aggregate the persuasive power of ideas or arguments. This is the position that I’ve always taken, so I really don’t know what you mean by saying that my positions are fluid.


  43. If your claim is that democracy is about counting, then it is no wonder that you are attempting to obfuscate this absurdity by misusing words, but regardless of what you take it to mean, aggregation is the collection of parts into a whole.


  44. Yoram,

    If you’re defining aggregation as the collection of parts into a whole, then I’d suggest that Keith is trying to make a useful distinction between two very different types of aggregation – one in which the whole is a simple quantitative sum of “independent” parts, and the other in which the whole comes from a complex interaction of “co-dependent” parts.

    The way that a total vote on an issue is the sum of individual votes is an example of the first. The way that a new group consensus emerges out of the interactions between arguments and counter-arguments is an example of the second.

    Keith, please correct me if I misunderstood.


  45. Keith wrote:
    “Democracy presupposes that the majority view constitutes the right answer.”

    But deliberative democrats insist it must be the considered view derived from discussion, not the uninformed majority view.

    While Keith and I agree that there are some aspects of deliberation that can be garnered just from listening to thoughtful pro and con presentations, deliberative democrats, (and advocates of diversity, such as Cass Sunstein) believe the quality of decisions (the majority view) is generally superior (in the opinion of that majority) if new and different perspectives are taken into consideration. Sometimes a win-win option can emerge from a synthesis of the pro and con stances, that would be lost in a silent secret ballot.

    This is why I want to have my cake and eat it too…benefiting from BOTH kinds of deliberation among allotted bodies (two separate bodies) to get the “right” answer in the view of the majority, and to get the best possible majority view.

    As to varying quality of speech acts…I don’t see any difference between this occurring in an allotted chamber and in an elected chamber…For example, my elected representative, who I voted against, may happen to be eloquent and powerful committee chair, while another representative (from another district perhaps, that I never had the chance to vote for) with whom I agree, happens to be lazy and inarticulate. In no way is there genuinely democratic equality of representation in elected chambers.


  46. Yoram: Aggregation, in the rational choice and political science literature, is a term used to refer to preferences and votes. In the latter case it does reduce to arithmetic, that’s why deliberative “democrats” dismiss voting as mere preference aggregation, as opposed to the consensual wisdom they like to think emerges from group discussion (sceptics would dismiss this as groupthink). To my mind this is no more plausible than Burke’s notion of disinterested and virtuous deliberation and has nothing to do with representative democracy.

    David: What you are illustrating is the two forms of deliberation (derived from the latin and german traditions). Deliberation in the latter sense is qualitative and in the former quantitative. Qualitative deliberation is required for epistemic purposes (getting the right answer) but in the final analysis democracy requires counting votes, as what other way is there of determining what is the right answer?

    So yes, Yoram, democracy is about counting. This was the case in Athens and it’s the case in the modern world. Votes were counted at the Athenian assembly and this is the origin of democracy. If you don’t like that then you should use some other term — my suggestion was kleorotocracy. I’m glad that we’ve had this exchange as hopefully there will be no further need for acrimony as it’s clear we are discussing two completely different things. You are clearly not an advocate of democracy (in any sense that would be understood by a political scientist). Your views appear to be those of a deliberative theorist in the Habermasian tradition. By all means continue to advocate that position, but please don’t claim that it has anything to do with democracy.


  47. Terry, we both agree in informed decision making after participation in a deliberative forum, that’s why we’re posting on this list. Our only disagreement is to the validity of speech acts by individual members. To my mind Fishkin’s experiments provide a via media between the need for representative equality and the need for new perspectives. Jim is insistent on the importance of the small group sessions (where participants deliberate actively), so who am I to disagree, although I’m worried about the reliance on moderators as their influence may be challenged in a legislative chamber.

    Regarding your comparison between allotted and elective representation, people get to choose their representative and the one who gets the most votes wins. If she is lazy and inarticulate then you shouldn’t vote for her. If the (articulate) person you voted for didn’t win, then that’s democracy I’m afraid — the arithmetic didn’t add up in your favour. But on aggregate it does add up, the winners are the ones who garnered the most votes and lazy and inarticulate representatives are unlikely to be returned the next time.


  48. Keith,

    Two responses…
    1. First a minor historical point…According to Hansen there was no actual counting of votes in the Athenian Assembly. There was a show of hands, with designated (randomly selected) officials estimating which side had more votes — which was subject to appeal. There was, however a counting of votes in court trials (where secret votes were cast by brass tokens).

    2. In your point about the voting for an elected chamber, you state… “but on aggregate it does add up…” reveals a weakness in your argument against having allotted legislators speak. The elected legislature can only in fact be trepresentative AS A WHOLE. Just as the elected assembly AS A WHOLE is deemed to be representative (though we know it is not in fact), any single representative can, at best, be representative only of a portion of a portion of the electorate…thus individual speech acts of elected representatives are no more nor less legitimate than speech acts of allotted legislators. Some legislators (whether elected or allotted) will exert more influence than others, thanks to personal attributes. This is inevitable among humans whenever there is discussion. Unless you want to also deny elected legislators the right to debate, you have no democratic grounds for denying this opportunity to allotted legislators. If your response would be that individual elected legislators have been AUTHORIZED by their constituents, then so can all allotted legislators be authorized by the citizenry as individuals to debate, by popular referendum establishing general rules.


  49. Terry: I’m aware that the Athenians had no quick way of counting thousands of votes, so had to rely on estimates, but no-one will deny that it was a quantitative estimate that was involved.

    As for the distinction between elected and allotted representatives, in the former there are two selection processes involved: a political career tends to appeal to people with reasonable communicative skills and (apparent) intelligence (Harrington’s “natural aristocracy”) and electors privilege representatives that demonstrate these qualities. Not so with allotted representatives — there will be substantial variation in these factors that would affect their debating skills. The only area where equality will be maintained is voting.


  50. > So yes, Yoram, democracy is about counting.

    This extremely superficial idea has been promoted by the elites in the West for centuries, but I think by now most people understand that it is false. As Dahl, for example, emphasizes, equality during voting is just one requirement for a democratic government system.


  51. Aristotle’s “extremely superficial” distinction between the three ideal types of government was numerical — the one, the few and the many. This does not preclude qualitative distinctions (between, in the case of the rule of the few, aristocracy and oligarchy), nevertheless a political system that precludes the counting of the votes of the many is not a democracy. Your own preferred scheme (rule by a tiny assembly, selected by lot) would be classified under the rule of the few, irrespective of its qualitative merits. It would not have gone down very well in fifth-century Athens.


  52. Keith and Terry,

    Both of you are proposing a stage of advocacy, for and against finalized bills, before the deliberation stage, by people other than the final decision makers. I want to be sure I understand who would do this in your different models.

    Keith, I believe that in your proposal, the “advocates for” bills would be people who wrote them, and “advocates against” would be members of a “House of Advocates,” randomly selected from candidates put forward by a variety of civil society groups. Is that correct?

    Terry, who would perform this same function (advocating for and against finalized bills) in your model?


  53. Yes that’s right. The advocacy stage involves deliberation in the active (Germanic) sense (deliberative Stimme), and the decision making involves deliberation in the reactive (Latinate) sense of weighing the results (liber). The decision-makers are present during the exchange between the advocates and (in the Fishkin model) also actively exchange views in small group sessions in order to generate questions for the plenaries with the advocates. They then determine the outcome via the secret ballot, in which the advocates do not participate.


  54. David,

    My scheme is still evolving…but I imagine that allotted review panels (of volunteers) who have debated and amended the final draft bill would have some members who support and some who oppose the final draft. A sort of caucus of each (pro and con) side would be charged with overseeing the development of a presentation for the final deciding allotted body. They could bring in outside experts or advocates, do it themselves, etc. I imagine that questions of fact would need to be verified by staff (or agreed to by both sides), so that the presentations are as factually accurate as possible. I can even imagine a requirement that both the pro and con side, once drafted, be presented by the same skilled presenters, to avoid the influence of personal charisma of presenters.


  55. Terry: “I can even imagine a requirement that both the pro and con side, once drafted, be presented by the same skilled presenters, to avoid the influence of personal charisma of presenters.”

    Yes I can see the case for that but it has the danger of making politics even more of a charade than at present, so I think the need for sincerity would trump the argument for equal charisma. Given that this will be a proper deliberative exchange (rather than trading sound-bites and dog-whistling) hopefully charisma won’t be quite so important. There is a lot to be set for establishing norms for civilised rhetorical decorum, similar to the House of Lords (unlike the bear-pit of the Commons). Archaic court-room (and courtly) conventions do have their value.


  56. Terry, would your allotted decision making panel, who vote yea or nay on the final bills, be one allotted chamber that votes on all bills? Or would they be divided into smaller groups that only vote on bills within a certain subject matter, or even a single bill?

    Keith, I’m assuming that in your proposal, all members of the allotted chamber vote on all bills. Is that correct?


  57. I’m agnostic on this, although my preference is for assemblies convened on an ad hoc basis (ie a new one for each legislative bill), exactly as in jury service. This has two advantages: 1) no chance for allotted members to “go native” and 2) direct involvement of the maximum number of citizens in the political process. Of course this would only work if the AC only had decision power (policy making and advocacy being the responsibility of others). I’m sceptical of Yoram’s claim that an assembly with full powers could have a life-span of four years and continue to be statistically representative.


  58. I largely agree with Keith on this…single issue is best. To maximize descriptive representativeness we need to make sure service is not too burdensome. So, we want to keep the term of service to a matter of days or weeks — like jury service. Also we want to minimize the opportunity for vote-swapping (“I’ll vote the way you want on this bill, even though I don’t agree with it, if you vote with me on my favorite bill next week.”), and the risk of lobbyist corruption. I imagine that legislative panels would be very analogous to a jury, where jury-tampering is strictly illegal. Lobbyists an d the public would submit arguments to the entire group through carefully structured balanced presentations organized by staff. The staff themselves would be assigned by lottery from a pool of career professionals, who would risk their job if they were revealed to be biasing presentations.


  59. Sounds good. The main defence against vote-swapping is the secret ballot.


  60. Having single-issue allotted panels that are ad-hoc and time limited, instead of an all-purpose allotted chamber with a multi-year term, would greatly reduce the time commitment required, and greatly increase the number of citizens who had a chance to play this role in their lifetimes – both important advantages, I think.

    Keith, what do you mean about allotted chamber members “going native?”

    Also, what about the argument that single-issue panels would tend to have a narrow focus, and miss important connections *between* issues? (Yoram, I think I got this from you)


  61. David,

    There are multiple problems with single issue panels.

    The most obvious one is who convenes the panels, determines their scope, decides whether they overstepped their authority and supervises the implementation of their decisions. Whatever body carries out those functions wields a lot of power (probably more power than the ad-hoc panels themselves) and therefore must be representative (i.e., be selected by sortition). In effect, the body that makes has those powers becomes the real seat of authority, with the ad-hoc panels being merely its subordinate committees.

    In addition, limiting the scope and the time of a panel reduces the ability and the motivation of the panel to reach a thorough understanding of an issue and therefore reduces the quality of the decisions made.

    Relatedly, inexperienced people – as the members of the ad-hoc panels will be – can be relatively easily manipulation by various elements such as bureaucrats and lobbyists.

    In view of the above, I would say that the question of whether, when, and how single issue ad-hoc panels are convened should be considered part of the regular on-going decision-making role of a permanent allotted body, rather than an a-priori system design question.


  62. I’ll also comment that I disagree with Terry on the issue of “vote-swapping”. I don’t see vote swapping (or logrolling) as inherently bad but rather as a legitimate political activity in which people consider (as they should) public policy making as a whole rather than as disconnected multiple acts and make compromises accordingly.

    (Voting in secret is a particularly bad idea going against the general principle that public policy making should be made in the open.)


  63. David, by “going native”, I meant that people cease over time to be statistically representative on account of being removed from their normal everyday lives — this is the prime reason for allotted chambers with very short membership lives. But this would require that allotted members only have limited (voting) powers, so it would not be feasible for a chamber with the extensive powers that Yoram proposes. Unfortunately by repeating the slogan that representativity = selection by sortition he is continuing to ignore the compelling argument that sortition only provides statistical representation and that the sortive mandate is not applicable to the active functions that he is proposing. If it were, then we could consider the merits of his proposal. This presupposes, of course, that none of us wish to replace one form of oligarchy with another.


  64. I agree with Yoram, that the body that delegates law-making assignments to these short-term panels would have vast power, and should be selected by lot. I refer to this body as a “meta-legislature,” in that they pass no laws, but establish the agenda and rules of procedure, and staffing policies for the panels. This meta-legislature necessarily will need to serve for longer terms (to deal with the inter-connections between issues, etc.). Thus they w8ill likely be less-representative than the law-making panels, but still more representative than any existing legislature. A separate meta-legislature might deal solely with evaluating staff, etc.


  65. Obviously anyone selected by lot will be more typical of the general population than anyone elected (on account of the principal of distinction), nevertheless anyone who fails to be selected (ie the vast majority) will be disenfranchised, as the agenda will be set by a tiny number of individuals. Such a system would fall into the category of oligarchy according to the standard political science taxonomy and would be highly vulnerable to corruption.


  66. Keith, any system of government by representatives – whether elected or allotted – leaves most of the population with very little power in government, and I see that as a big problem to resolve. What could be done?

    As you and Terry point out, an allotted chamber with single issue panels might give a substantial number of people the chance to play a meaningful role in making political decisions at some point in their lifetime — certainly much better than the situation we have now, in terms of the people actually ruling.

    Incorporation of mass voting at some stage in lawmaking, as you and Terry both advocate, could also give some power to the people as a whole – but it would have to be a lot better than the initiative and referendum here in California!

    Another possibility would be to create a role for direct democratic lawmaking at a very small scale, like the New England town meeting (in which citizens propose, discuss, and pass laws face to face). It seems to me that somebody must have proposed ways to adapt this model to districts or neighborhoods within cities, but I haven’t found them yet.


  67. Yoram,

    In an arrangement with an allotted meta-legislature, and allotted single issue panels, I don’t think the panels would be “merely subordinate committees” to the meta-legislature. The panels would make binding decisions about laws, which is a lot more power than ordinary citizens have in most existing systems of lawmaking. The meta-legislature would have more power than any individual panel, but it would not be able to enact any laws.

    You write, “limiting the scope and the time of a panel reduces the ability and the motivation of the panel to reach a thorough understanding of an issue.” Actually, I’d imagine that limiting the scope of a panel would *increase* the amount of attention that it can devote to one issue. The members of general-purpose legislative chambers are likely to be overwhelmed by more bills than they can properly study. Elected legislators are constantly voting on bills they’ve never even read.


  68. David,

    Here I can speak from experience (as a former legislator). Members of legislatures only become semi-expert on a handful of issues (those their committees deal with). Plus… the nature and bounds of their expertise is shaped by considerations of re-election and fund-raising potential, rather than long-term consequences. Yoram is clearly hoping/expecting that an allotted chamber, without these considerations, could become genuinely expert on numerous issues. I tend to agree with David, that single-issue panels — even meeting for short duration — would become MORE expert on a given issue. The PROBLEM is the risk that they would lack appreciation of the interconectedness of issues, and lack the ability to balance relative importance. In other words, a panel dealing with health care might quickly come to view it as the most important issue for society to devote resources to, while some other panel, dealing with primary education, would decide that it was more crucial than health care, etc. Thus, my single-issue panel proposal does not deal with the challenge of overall budgeting very well.


  69. David: “Keith, any system of government by representatives – whether elected or allotted – leaves most of the population with very little power in government, and I see that as a big problem to resolve. What could be done?”

    That’s where statistical “representation by proxy” comes in, but the statistical mandate is a limited one. Let me illustrate this by a simple thought experiment (that could be very easily put to the test):

    Scenario One: Allotted assembly limited to voting powers
    Convene any number of assemblies by random sample, each one large enough to be statistically representative but small enough that members are motivated to pay attention (as their vote will make a significant difference). Each assembly to listen to the identical debate from the same competing advocacy teams. Given the secret ballot and that all votes have equal power then, if statistical theory is correct, one would anticipate that each assembly would vote in the same way; if not then the sample size would need to be increased. In this sense every citizen is represented by proxy as the assembly is a microcosm of the considered view of the entire population — hence the argument for sortition as a system of democratic representation.

    Scenario Two: Allotted assembly with full powers
    Given that full powers include policy initiation and advocacy and these are dependent on the speech acts of individual members, there is no reason to believe that every assembly would come up with an identical agenda. As such the assembly would not be a proxy for the whole nation as it would reflect the priorities of the individuals in the particular assembly and would be disproportionately influenced by members with high status and/or high persuasive powers. Such a system would be classified as a form of oligarchy rather than democracy. (And given the vulnerability of such a system to lobbying and corrupt inducements, the abuses of elite electoral systems pales into insignificance).

    As mentioned above this thought experiment would be very easy to test (given the necessary funding), but would anyone realistically expect the outcome of each assembly in scenario two to be identical (a portrait in miniature of the considered views of the whole population)? If not then some sort of democratic filter (elective, direct-democratic, votative etc) is required. We are all aware of the shortcomings of the Californian initiative system (along with the problems of partisan politics), but in my two-stage legislative model, poor-quality proposals would be rejected at the stage of the sortive assembly. The Athenians introduced the rotation-based legislative courts to overcome the failings of direct democracy and this is the model that we should seek to emulate, rather than simply abolishing democracy by adopting a machine that was designed for an entirely different purpose (rotation).

    Advocates of sortition really do need to stop sloganising (“representativity = selection by sortition”) and wake up to the fact that sortition is not a magic bullet and that it needs to function alongside other mechanisms in order to be characterised as democratic. And rather than just opting for two chambers (one elected and one sortive), much better to analyse the functions of each process (election and sortition) and assign roles accordingly.


  70. Keith,

    Could you think through with us how votation might work? Wouldn’t there be an incredibly large number of potential items for voters to choose among? How would they at all competently do this? Wouldn’t this process suffer from all of the rational voter ignorance and media manipulation problems of other elections (only more so)? Would votation only be a secondary filter for the items that have already been filtered by an elected chamber, or a petition hurdle? What about low-visibility, non-sexy, non-partisan, yet important long-term issues like infrastructure maintenance? Would they ever make it to votation, or if so pass?


  71. Terry, you wrote, “I tend to agree with David, that single-issue panels — even meeting for short duration — would become MORE expert on a given issue. The PROBLEM is the risk that they would lack appreciation of the interconectedness of issues, and lack the ability to balance relative importance. In other words, a panel dealing with health care might quickly come to view it as the most important issue for society to devote resources to, while some other panel, dealing with primary education, would decide that it was more crucial than health care, etc. Thus, my single-issue panel proposal does not deal with the challenge of overall budgeting very well.”

    One solution would be to have some panels whose area of responsibility was the interconnections between bills around a particular issue, and some responsible for the connections between issues. These panels would not decide on particular bills, and they would not be “hierarchically superior” to the panels that decide on bills.

    In participatory budgeting, it’s common to have groups of delegates who focus on the needs of neighborhoods, and other groups who focus on cross-cutting “themes” such as environment, health care, and education.


  72. Keith, you wrote, “rather than just opting for two chambers (one elected and one sortive), much better to analyse the functions of each process (election and sortition) and assign roles accordingly.”

    I’m not sure I understand. Are you saying that it’s important to decide what elected bodies are best suited for, and what allotted bodies are best suited for, before making a priori decisions about chambers? If so, I agree, and I have questions for you:

    What do you think allotted bodies are best suited for? I’m assuming that your answer is something like “voting yea or nay on bills, in a way that’s descriptively representative of the whole population.” Is that right?

    What do you think elected bodies are best suited for?

    What about all the people? (as in initiative and referendum)

    Do you think there is a role for chief executives? (they have veto power in many current systems)

    Do you think there is a role for small-scale, face to face, direct democracy, as in the New England town meeting?


  73. > Yoram is clearly hoping/expecting that an allotted chamber, without these considerations, could become genuinely expert on numerous issues.

    No – I agree that this would be unrealistic.

    The allotted chamber would delegate authority and consult with experts, just like present day elected chambers do, and if the allotted chamber would see fit to convene other allotted panels for certain purposes, that’s also fine – I am not in a position to decide that it is an inappropriate method.

    My point is that such an arrangement has some significant potential drawbacks and that forcing such an arrangement onto the allotted chamber makes no sense. We can expect the allotted chamber to be able to judge the proper decision-making procedure in any particular situation. Why, then, would you force the allotted chamber to use a particular procedure (namely, convening ad-hoc panels on every issue)?

    Furthermore, beyond producing poor results in some situations, such an arrangement would be anti-democratic as well since it would not allow the representative body to decide on the best policy but rather would force it to use a policy that was designed by some other, non-representative, body.


  74. David,

    > The members of general-purpose legislative chambers are likely to be overwhelmed by more bills than they can properly study. Elected legislators are constantly voting on bills they’ve never even read.

    Why would the members of a legislative chamber – that can arrange their work procedures in any way they see fit – arrange their work procedures in a way that is unsuitable for them? Elected legislators vote on bills they don’t read because they either don’t really care about those bills or because they are beholden to interests that favor (or oppose) those bills.

    Again, if it is an allotted chamber’s informed and considered opinion that a certain piece of work should be delegated to a certain body (allotted or not) then that is perfectly fine, but forcing such an arrangement onto an allotted chamber is both poor design and anti-democratic.


  75. Yoram, you wrote “Why would the members of a legislative chamber – that can arrange their work procedures in any way they see fit – arrange their work procedures in a way that is unsuitable for them?”

    If they have far more bills to pay attention to than time needed to do it, no re-arrangement of their work procedures will solve that problem. It seems to me that having a larger number of people with smaller scopes of responsibility could potentially solve two problems: it could reduce the overload faced by legislators, and give meaningful voice and power to more citizens.


  76. Terry,

    Assuming an initiative- or petition-based system for policy proposals, there would be a minimum threshold to get on the votation ticket — in the case of the UK Government e-petition system the threshold is 100,000 electronic signatures, but the threshold could be adjusted to ensure that the public votation was between a manageable number of alternatives. As for low-visibility issues like infrastructure maintenance this would be the province of the relevant government department — this sort of “housekeeping” legislation would go straight to the AC, without the necessity of a public votation stage.


  77. David,

    Yes I agree that the functional distinctions are a priori, and that statistically-representative bodies are suited for the decision function (ie voting yea or nay). Policy proposals would best be served by a combination of election (for accountability) and direct democracy (to ensure genuine choice, rather than weak veto power of the AC).

    The veto power of the executive would be limited to the statutory requirement for fiscal equilibrium — every proposal would be costed, and every cost increase would have to be matched by a tax increase or a corresponding budget cut. On top of this the only influence government ministers would have would be the threat to resign.


  78. Terry,

    Sorry, I missed part of your post. Yes, the votation would be subject to rational ignorance and manipulation, but (unlike as present) the final decision is in the hands of the AC. This is the reason for the 5th century innovation of the legislative courts and what I’m suggesting is its modern analogue. And yes, votation is the second filter (out of three in total).


    Ditto. I don’t have any particular views about small-scale local direct democracy, but don’t see the relevance for national government.


  79. Yoram wrote:
    “Why would the members of a legislative chamber – that can arrange their work procedures in any way they see fit – arrange their work procedures in a way that is unsuitable for them? Elected legislators vote on bills they don’t read because they either don’t really care about those bills or because they are beholden to interests that favor (or oppose) those bills.”

    There is more to it than that. Legislators actually read very few bills at all. Firstly, they rely on the member of their party who serve on the committees that took testimony on each bill for guidance. They also rely on their staff (in legislatures big enough to have staff specific to each legislator), who develop some policy expertise— at least as to how legislation will affect re-election and fund-raising.

    David is correct that there are far more bills proposed than legislators have time to read. And even if they did read them, they wouldn’t understand them nor their implications, unless they sat through the committee hearings too.

    Presumably, in an Allotted Chamber, members wouldn’t have partisan comrades to rely on for guidance, though they might be allowed to appoint personal staff, I suppose. But I think it is unlikely that most members of an Allotted Chamber would have the time (let alone the inclination) to understand the volume of legislation handled currently by elected legislatures, because they wouldn’t have the heuristic of relying on fellow-party members on each committee.


  80. Terry, I’m thinking about the problem you posed earlier, that single issue panels could develop “tunnel vision,” and not see the larger context around their particular bill (or particular piece of legislation). It seems to me that a good solution would be to have some different groups of people responsible for coordination between the panels, across dimensions such as theme (for example, environment, education, social justice) and geography (how the sum total of legislation affects particular regions). But I haven’t thought through how these bodies would be constituted (more allotted panels? staff?) or what their power would be. What do you think?


  81. David,

    I haven’t figured it out yet. this is where advocates of maintaining one elected chamber have a leg to stand on (maybe). It takes time to amass enough knowledge to have both institutional memory and policy expertise to even perceive where panel policies might have unintended consequences. I guess my leaning is towards having people appointed by an allotted body, who would serve as panel advisers, to monitor these sorts of broader impacts of pending legislation. The key is to have these appointed “experts” be accountable to some allotted evaluation bodies. An analogy might be the citizen review boards that oversee police departments in some cities.


  82. > David is correct that there are far more bills proposed than legislators have time to read. And even if they did read them, they wouldn’t understand them nor their implications, unless they sat through the committee hearings too.

    Bills are not a natural phenomena that you can have too many of. Limiting the number of bills that are presented for a vote is a simple matter of procedure. Again, the reason that today there are so many bills is that this arrangement is useful for those who made it.

    One could argue that there are too many pressing issues to address. I think it is rather clearly not the case, but even if it is, it should be up to the allotted chamber to triage those issues and decide how to deal with each matter (make the decision in the chamber, delegate it to an ad-hoc allotted chamber, or otherwise). Forcing the chamber to use a pre-determined procedure is anti-democratic, serves no useful purpose, and could easily result in poor decisions.


  83. One drawback that I have not yet explicitly mentioned of having important decisions made by a multitude of ad-hoc bodies is that spreading power too thinly makes it difficult for the public to follow and monitor the decision making process, making it easy for interested parties to influence the process.

    Having all top priority issues discussed and decided upon by a single body limits the agenda (exactly because of the cognitive limitations of the delegates that you were referring to above). This makes is possible for the public to follow those issues, discussions and decision making processes, creating a situation in which there is openness, accountability and participation.

    Having, on the other hand, a multitude of bodies making important decisions on multiple issues simultaneously would make it impossible for the public attention and discourse to track public policy. The result would be a situation of unaccountability and public confusion, reducing trust in the system even if it works well, and increasing the chance that it doesn’t.


  84. Two points…

    1. Yoram,
    You make a good point that bills are not inevitable natural phenomena, and thus can be limited through procedures. However, I suspect that the number of genuinely desirable or essentially necessary bills would still swamp the ability of a single AC to cope with, without using some heuristic device (such as party allegiance used by elected legislators) to ease their “cognitive burden.”

    2. Yoram argues that having a single AC (instead of a huge number of panels) focuses and limits the agenda, which has the added benefit of allowing the general public to follow the proceedings
    “creating a situation in which there is openness, accountability and participation.”

    “Opennness” I can agree with, but “accountability” and “participation” are tricky concepts in the case of sortition.

    I’m not sure the concept of “accountability” as it is usually used even applies to a lottery body…
    and is certainly hard to see how it could apply to any individual member (other than avoidance of bribery).

    “Participation” is also hard to figure, other than through petition or formal submission of testimony. We don’t want motivated non-allotted individuals to be non-representatively participating.

    I think participation by any who wish (regardless of whether they were picked by a lottery) is crucial, but not at the stage of yes/no decision making. That would be akin to jury tampering. Participation should be open to all at the earlier drafting, designing stage (what I called Interest Panels).


  85. Terry and Yoram, I’d like to ask you the same questions I asked Keith earlier (slightly revised).

    Are there *any* functions that you think elected bodies are better suited for than allotted bodies? If so, which ones?

    Do you see any role for “all the people?” (as in initiative and referendum)

    Do you think there should be a role for chief executives in lawmaking? (they have veto power in many current systems)

    Do you think there is a role for small-scale, face to face, direct democracy, as in the New England town meeting? For example Frank Bryan in “The Vermont Papers” argued that having the experience of making laws on a small scale would make citizens much better able to choose representatives and hold them accountable.


  86. Roles for “all the people”:
    1. The allotted chamber may choose to refer some issues to referendum simply because gauging popular acceptance for some new law may be more critical than the logic of the law itself.
    2. There should be a referendum to authorize the sortition system (and other constitutional amendments).
    3. There needs to be an accountability option by referendum to discontinue the sortition system.
    4. All the people can initiate legislation through petition.
    5. Frank Bryan’s learning function of democracy can be achieved by vast numbers of local and regional sortition panels, such that most people will participate at some time in their lives.

    Roles for elections among the general population:
    1. Could be used for symbolic head of state (without substantial power). I do not see any role for a chief executive in law-making, other than possibly a minimal veto power delaying a new law until another allotted body can reconsider it.
    2. Could be used in tiny jurisdictions, where people actually know each other, and sortition is impractical. In such cases this should be coupled with a New England Town Meeting model.


  87. David,

    Elections in my opinion are essentially worthless as a tool of representation, except in very small groups, and I see no reason to rely on them.

    Referenda are also not of much use for run-of-the-mill policy making. They may of use in special situations. Specifically, referenda may be the appropriate decision making tool to use when there is a reason to suspect that (1) any delegation would have an a conflict of interest on the matter at hand, and (2) the issue at hand is clear enough so that the inherent problems of mass politics are not too severe.

    Individual people with any kind of significant political power should be avoided whenever possible.

    As for small scale participative processes, I am very much in favor of those when appropriate. For example, see my suggestion for democratic media, where editorial decisions would be arrived at by average citizens in such a small scale setting. Beyond the direct usefulness of such activity, it may provide training for decision making in more high-powered bodies for those few who would get to be on such a body. On the other hand, I don’t think this activity would provide any useful training for a voter in an electoral system since this no amount of training can make one a good voter – it is a completely meaningless activity.


  88. An Athenian group has a website at


  89. A key design element to avoid undue influence of special interests (“rent seeking”) is to randomly select panels for limited terms, too short for them to be “got to” before their terms expire. That is why trial juries can work. A weakness of the Venetian doge was that he was selected for life. Better a term just long enough to learn the job and not so long as to become corrupted by it.


  90. Absolutely. I’m in favour of taking the jury trial analogy literally and selecting a new panel for every legislative bill. The corollary though is that continuity, consistency and accountability will suffer, so ad hoc sortition panels will require a stronger role for the executive. The relevant departmental minister would, of necessity, be the most important advocate (for or against) — although she would not have a vote in determining the outcome. Furthermore, assuming the Chancellor/Treasury Secretary was obliged to deliver a balanced budget, then the fiscal implications of every legislative proposal would be a crucial factor — cost increases in one area would have to be balanced by savings or tax increases. If it was felt that this was insufficiently democratic then a much longer term would be required for the sortition-based legislature, but such a body would be very easily “got at”.


Leave a Reply

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

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

Twitter picture

You are commenting using your Twitter 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: