On the legitimacy of citizen assemblies

Dear Kleroterians,

I am currently writing on the legitimacy that grounds sortition-based representation in general, and citizen assemblies in particular. Not the perceived legitimacy of citizen assemblies (whether people actually see them as legitimate or not), but the reasons that we might have to see the decisions of such asssemblies as binding.

I realize that you have thought about this much more than I have. And this is why I would be interested in having your opinion on the three following questions:

  1. What are the potential sources of legitimacy for citizen assemblies, besides political equality, representativeness, impartiality and ordinarity?
  2. Among these different potential sources of legitimacy, which one(s) do you see as the most important?
  3. Finally, because I am expecting many of you to highlight representativeness as the main source of legitimacy, I add a third question:

  4. Would you say that a citizen assembly of 50 to 100 participants, with optional participation, still has some legitimacy? Would your opinion be different with stratified sampling?

Thank you very much for your input! I will make sure to credit the Blog if a publication comes out of this!

59 Responses

  1. Hi P-E,

    IMO, legitimacy is much more of a retrospective, empirical characteristic of a government system than a prospective, principle-based characteristic. A system that produces good outcomes is legitimate. A system that does not, is not legitimate.

    Elections are by now widely recognized as producing poor results, but are still considered as producing better results than all alternatives, and so they sputter on, in a state of semi-legitimacy.

    As sortition-advocates we need to argue that a sortition-based system is likely to produce good outcomes (at least better than those of electoralism). Based on this likelihood, sortition would get a provisional approval and would get a chance to prove that it does indeed produce good outcomes. If it does indeed live up to its promise it will become legitimate.

    Liked by 1 person

  2. I think that an important legitimacy of citizen assemblies is that they give the citizens optimal freedom (if they are selected by lot) to participate fully in the lawmaking proces (positive freedom). Ancient Athenians used the word isegoria, the equal right to speak in the Assembly and suggest proposals.

    Producing good outcomes cannot in itself, in my opinion, be considered as a legitmation ground. Citizen assemblies can make serious errors and dictators can deliver satisfying decisions!

    Liked by 1 person

  3. Hi Pierre-Etienne,

    Besides the four attributes you mentioned, I would add something like “depth.” Under the right circumstances, randomly selected citizen bodies can (and do) spend substantially more time on an issue, and examine it in more depth, than what usually happens in an elected or appointed decision making body whose members have many other issues to be concerned with.

    Liked by 3 people

  4. Pierre-Etienne,

    I am not sure what you mean by “ordinarity.”

    I assume you are thinking of citizen juries/assemblies in the context of deciding or recommending laws.

    There is one key thing not on your list, and another that I think is not made clear in your list (if it is there at all).

    1. A requirement for any good method of deciding laws is, it seems to me, that they be decided in an informed manner. Juries are well-suited for doing this (unlike for example popular vote).

    This is a necessary condition (for good decision-making), not a sufficient condition, it seems to me.

    Poorly informed and uninformed decisions lack legitimacy, in my view.

    David’s point about “depth” is I think a similar point. “Depth” is part of what it is to be “informed.”

    2. Part of what is required for an informed and fair decision is that it happen on a level playing field. For example that the cases for and against a law be given equal time before the jurors, rather than one side having a lot more time than the other or only one side being heard.

    Juries are well-suited for providing hearings on a level playing field.

    Your mention of “impartiality” is in the spirit of level playing fields, but I think refers to a random sample not being partisan or tied to special interests (which is also important).

    Liked by 1 person

  5. Yoram,

    > As sortition-advocates we need to argue that a sortition-based system is likely to produce good outcomes (at least better than those of electoralism).

    This is clearly true, it seems to me.

    The reasons it is likely that citizen juries will produce good outcomes include that they are representative and informed, and make their decisions in the context of fair procedures and a level playing field.

    Also key re likelihood to produce good outcomes, is the independence juries have from politicians, political parties and concentrated wealth.

    Pierre-Etienne, if you’d like to see more of the reasons I think laws should be decided by legislative juries rather than by politicians and popular vote, see my June 22, 2020 article “Let Legislative Juries Decide Laws.” Available online, easily found.

    Liked by 2 people

  6. One important feature is ideological pluralism of the proposals presented to the assembly or jury. I have outlined the superminority method for doing this, particularly as it applies to the recent French Citizens’ Convention on the Climate.

    Liked by 1 person

  7. Alex raises a good point that is also essential to the list of what makes legislative juries good (legitimate) ways to decide laws.

    Any well-designed system of legislative juries must ensure pluralism in proposals. In particular, legitimacy is undermined if the proposal that most people would prefer if they became well-informed is blocked or not proposed.

    Liked by 1 person

  8. 1) ordinarity – as you inelegantly call it – which I guess comes under representativeness and isegoria (to which I give a deeper meaning than it is often given. For equality of speech shouldn’t just be equality of the right to speak, but equality in discourse in so far as it is warranted. Essentially our democracy stigmatises those without a lot of formal education with disastrous results.
    2) consideration. As I like to put it, polling gives us the opinion of the people and citizens’ juries give us their considered opinion.

    But 3) which your post has prompted me to articulate explicitly for the first time there’s a sociality and communicativity about the process in a citizens’ jury. I can identify at least three aspects of this

    A) The jurors are individuals, but they’re making their decision NOT in the privacy of a ballot box but as part of a deliberative group. That is a very different way to frame an individual’s decisions about what they think and what they want to do as a group. So it’s individual choice in a social context.
    B) I think of this as a social thing. But it’s also a cognitive thing – it’s also a communicative context.
    C) The communicative context makes it easier for the relative intensity of feelings about different options to find its way into the decision. A lot of decisions involve people who really care about something amidst lots of people who have mild preferences. It’s good to try to address that to the extent that it can be done practicably and fairly.

    Liked by 3 people

  9. Nick:> A lot of decisions involve people who really care about something amidst lots of people who have mild preferences. It’s good to try to address that to the extent that it can be done practicably and fairly.

    Are you suggesting that people who feel strongly about something should have their opinion weighted more? That seems to violate the equality constraint. It also reproduces a flaw in our current system, where special interests override the general interest.


  10. Regarding question one, there’s a big difference between equal chance of selection and the political equality of all citizens (that would only apply to a truly representative sample). I don’t see any reason to believe that a randomly selected group would be impartial (except in an aggregate sense) and I don’t see any merit in ordinarity (if you mean being comprised of “ordinary” people).

    Unsurprisingly Yoram ignores your distinction between perceived and formal legitimacy. And it’s interesting that, apart from Alex, nobody so far has claimed that representativity is the key to the legitimacy of sortition. This, of course, is my view and it would mean ruling out small, voluntary allotted groups (your question 3) on principle (there is no way that you can stratify for refusal to participate). Claims that allotted groups are “more” legitimate than elections simply conflates the different representational principles that Pitkin outlines so clearly in her book.

    Liked by 1 person

  11. Thanks Alex

    I thought I’d get that response here :)

    Actually I am suggesting what you suspect I’m suggesting. I’m sure you agree. Only the most unreasonable would not. The reasons we rule it out of court, and should mostly rule it out of court is that giving more votes to people who care more is so easy to game. We can’t observe how strongly people feel and we don’t want to open up a process by which the self-important, the dramatically inclined and the unscrupulous can simply avail themselves of more votes by claiming to strength of feeling. Gresham’s law would almost immediately exert its force and degrade the whole idea in most arrangements of formal voting.

    But if you’re in a group and someone feels strongly about something and you don’t think they’re putting it on and you disagree, but not strongly, you may well be disposed to allow them to have their way. That’s why I added the rider that this was a good thing to do “to the extent that it can be done practicably and fairly”.

    Within democracies built on the principle that voting is necessary and exclusively a competition of interests, which is very much how it’s framed in modern liberal sensibilities, there is no room for this, so it doesn’t even arise as a question.

    But human beings are better than that. We are the species that solves problems in groups. We need to keep this generous frame when we speak about democracy, even if we will also need to heed the dictates of practicality and procedural fairness.

    Liked by 2 people

  12. Thank you all for your inspiring responses! I’m surprised by the fact that many of you ground the legitimacy of citizen assemblies on the expected quality of their decisions (be it in terms of responsiveness, equal consideration of arguments, or depth). I actually share this view, but it’s not that popular in political theory. Many people believe, as Ronald points out, that it’s insufficient to ground democratic legitimacy.

    I’m also surprised by the fact that you de-emphasize representativeness, which is central in Keith’s case for sortition. But I guess it’s necessary to find other grounds because most existing citizen assemblies can’t make a credible claim to representativity, for the reason explained by Keith. (Unless we’re prepared to reject all of these innovations, of course.)

    “Ordinarity” is a category I borrowed from Dimitri Courant (https://serval.unil.ch/fr/notice/serval:BIB_315BD0030F49), who has done a great job on this issue of sortition-based legitimacy. I guess it can generate perceived legitimacy: people trust the assembly because it’s composed of ordinary citizens, not professional politicians with their partisan interests and class biases. But I’m not convinced from a normative viewpoint.

    I agree that the impartiality of the assembly cannot be assumed. But the independence from partisan ties and electoral promises contributes to its deliberative potential.

    Finally, I’m a bit skeptical about the optimal/positive freedom argument for sortition. I guess it works only with multiple bodies and frequent rotation. But used under contemporary conditions, sortition would not equalize this freedom; it would be restricted to the happy few.

    Thank you again! This was helpful!

    Liked by 1 person

  13. Interesting to see the divergence between the views of political theorists and sortition activists. I confess that I do reject recent innovations as democratically illegitimate.


  14. Nick:> I thought I’d get that response here 🙂

    I know, I know. I took the bait.

    Nick:> But if you’re in a group and someone feels strongly about something and you don’t think they’re putting it on and you disagree, but not strongly, you may well be disposed to allow them to have their way.

    I think this is a case where common experience does not generalize to the political. What you say is eminently reasonable if you are sitting at the pub with a friend trying to decide what football game to watch. It doesn’t work at all in the representative context. We are not “allow[ing] them to have their way”. We are listening to their opinion on a matter of public importance, for which we are duty-bound to be dispassionate.

    Nick:> But human beings are better than that. We are the species that solves problems in groups.

    I agree with this statement. But it does not follow that allowing those with greater feeling should be listened to more. You are assuming your conclusion that differential weighting is a better decision rule. I disagree. I disagree STRONGLY, so I win ;-)


  15. I had in mind a specific situation when I proposed the hypothetical.

    After an Adelaide citizens’ jury, a 6’2″ man commented that he wasn’t worried about safety at night. “Then being with other people: older, smaller, females, you learn that their experiences are very different”

    It’s just human decency for the man to accept that things should be done for them to address their more pressing concern against his mild negativity towards doing so.

    This is how people make decisions. Ideas like one vote one value are abstractions that emerge in the context of contest. Often it comes to that. One definition of democracy would be to say that everyone has a right to retreat to those rights whenever it comes to a vote.

    If it’s a contest, the man has the right to have his vote count as much as theirs. As a matter of problem-solving, decency and all the rest of it, he decides in the goodness of his heart that he doesn’t want it to come to that. So, against the backdrop of the guarantee democracy gives them all of the formal equality of all of their votes, they come to an understanding which arrives at an even better decision.

    And if you say you feel so strongly about this that my exemplary mildness should yield to your passion, I’m afraid I don’t believe you. And since you only get the right to count your passion more than my mildness by my grace and favour, I’m afraid I don’t grant you that favour. I’m too worried about Gresham’s Law. I’m too concerned that I’m encountering bad money – which of course always announces itself as good money :)

    Liked by 2 people

  16. Nick,

    People are certainly entitled to make up their minds however they wish, including internally overweighting those who express their opinions forcefully. I am however concerned about loss of independence. In The Wisdom of Crowds, James Surowiecki makes clear that independence is absolutely essential to producing a satisfactory result. I consider this to be a matter both of epistemic quality and representativeness, as groups that take a giant helping of social interaction diverge from the general public in highly non-linear ways.

    One possible solution is to have a jury that is composed of independent small groups. For example, if a jury has 500 members, they can be divided into 50 small groups of 10. I’m not against social interaction per se, but I am against throwing everyone into a single group. Such a configuration threatens the legitimacy of the entire project.


  17. Alex and Nick,

    It seems to me that strength of feeling here is standing as a proxy for urgency of interest. Disabled access to public buildings, for instance, is a pressing requirement for the disabled and a minor cost to the abled, and we would all agree that the urgent needs of the former should be prioritised over the convenience of the latter.


    I think the normative legitimacy of sortitional systems derives partly from the fact that they avoid other systems’ perverse incentives, allowing them to produce better outcomes, and partly from the political equality they express. Although only ‘the happy few’ may actually participate, the fact that they are chosen completely at random serves a symbolic function: it says to every other citizen that they, too, have the right to rule. The drawing of lots is therefore simply a fair means of apportioning the limited number of positions of authority available. Other systems, by basing their legitimacy on some special virtue of the rulers, express a kind of contempt for their ordinary citizens by saying they lack that special virtue; sortitional systems do the opposite.

    Liked by 3 people

  18. Anyone who has done jury service will know the extent to which a jury can be persuaded one way or the other by single persuasive jurors. In the judicial case that’s not necessarily a bad thing as the jury is tasked with finding the (epistemically) right answer — guilty or innocent — and some people may be smarter and/or more attentive than others. But a legislative jury is entirely different as randomly selected members are charged with representing the beliefs, interests, values (and prejudices) of their virtual constituents. So Alex is right to be concerned about the ability of jurors to be swayed by those with passionate convictions about the “right” course of action.


  19. The way intensity of feeling or interest is expressed in relationships is by making cross-area deals. In policy making circles this is called log-rolling and is often unjustly considered a negative phenomenon. Such practice, however, is impossible if decision making in different areas is carved out to different decision making bodies.


  20. Pierre-Etienne,

    I believe different sorts of mini-publics can have different bases for legitimacy, depending on their task. Final decision-making mini-publics enacting a law must have representativeness as a core legitimizing feature (more about this below). And as Yoram says… over time, good outcomes will do more to provide legitimacy than any a priori theoretical grounds. However, some mini-publics might handle a specific task, where mere (relative) impartiality, or mere diversity are sufficient grounds for legitimacy, such as a mini-public that receives proposed draft policies from other citizens and winnows them down and refines them into a final bill that a separate large, more fully representative, mini-public can hear pro and con arguments about and vote to adopt or reject. This relates to your final question about size. For SOME tasks a body of 50 is sufficient, but not for the task of final enactment of a law.

    On the question of representativeness, there are some inherent conundrums. We may never get a genuinely FULLY representative mini-public, because some people will simply refuse to serve. For example, Jehovah’s Witnesses won’t vote in elections nor serve on juries because they believe both are too worldly, and it is God’s role to judge people, not theirs. In some sense, ALL democracy, from Athens on, has an element of self-selection. All we can do is remove as many barriers as possible and make it a cultural expectation (quasi-mandatory), to serve if called.

    Liked by 3 people

  21. Good points, Terry! Thanks for this. I still have doubts about this idea of a final enactment by a mini-public, even if it’s large enough. But I think you’re right that we can relax the representativeness condition for other tasks.


  22. P-E,

    What doubts in particular? (I also have reservations, unless certain very tough constraints are in place).


  23. Keith,

    It’s hard for me to imagine, as a non-selected citizen, accepting the verdict of a body of representatives over which I can exercise no form of control. But maybe that comes from having been socialized in an electoral democracy. I’m willing to change my mind, but it’s kind of resisting.


  24. Assuming you do, why do you accept the verdict of juries?


  25. I do, but the decision that juries make does not apply to the whole population and the risks of misrule are lower because jurors have not their interests at stake in the decision.


  26. P-E

    I agree with your scepticism regarding the non-selected citizens, that’s why I spoke of the need for tough constraints. Briefly this means a) respecting Dahl’s criterion that the demos needs to have control of the agenda (this is where Alex’s project comes into play); and b) demonstrating — both theoretically and by experiment — that it makes no difference which empirical citizens are selected by lot, the outcome would be the same (within specified margins of error). The latter would involve a host of constraints to do with sample size, quasi-mandatory participation and an absence of speech acts (no small group face to face deliberation). We’ve been over these constraints many times on this forum and you certainly end up with an impoverished form of deliberation that would be rejected by Habermasian and other deliberative democrats but it would be a necessary sacrifice to ensure that the verdict is acceptable to the vast majority of citizens disenfranchised by the aleatory coup.


  27. There’s an interesting divide in how we think about legislative juries’ legitimacy. I’m very much with Terry and Yoram in that I think demonstration of effectiveness will do the bulk of the heavy lifting. I also think the degree of descriptive representativeness required for a jury to be seen, by the public, as descriptively representative is quite a lot less than Keith seems to think. Representation is first and foremost a relation of trust between represented and representatives, and I think it’s best to think about the various kinds of representation – descriptive, elected, etc. – as shorthand for different bases for that trust. ‘Do you trust this group of representatives with this power?’ is the primary question. The citizens might then answer, ‘Yes, because we (collectively, not individually) control their chances at reelection,’ or ‘Yes, because they’re generally like us and share our fundamental values’ – or, ‘Yes, because I met my MP and she seemed trustworthy’, or ‘Yes, because the Party helped me out when I was suffering’, or ‘Yes, because they’re not an untrustworthy elite’, or ‘Yes, because the country’s done pretty well under this system’, or ‘Yes, because the priest says God chose them’, etc.

    Now we might try and divide these kinds of legitimacy into ‘democratic’ and ‘undemocratic’, or ‘representative’ and ‘non-representative’. But in fact that’s going to be pretty difficult to do. I think this is something Yoram got right in his post about the benevolent dictatorship – that there’s no really good way to draw those distinctions. It’s much easier and more productive to ask, ‘Is that a good reason to trust these “representatives” or not?’ If that’s our operative question, the question of normative democratic legitimacy turns into a question of whether the members of the demos have good epistemic reason to trust the “representatives” with the power the system gives them. After all, any system in which the citizens do in fact trust the government with its power perforce enjoys the benefits of empirical democratic legitimacy, and vice versa. We might even then say that a ‘democracy’ in the broadest normative sense is a system in which the people have good epistemic reason to trust the government with its power.

    In short, the point I want to make here is that we need to think of legitimacy in broader terms than hitherto – so any good epistemic reason citizens have to trust their government with its powers is a bona fide form of normative legitimacy. I think if we make this move, concerns about the magic authorising power of voting or the strict exactitude of descriptive representation lose their force.

    Liked by 1 person

  28. One element of legitimacy that has not been discussed is the transparency of the jury/assembly itself. The French climate assembly did not broadcast its work, it just reported its results. This is necessary for obvious reasons, but as a result, the general population does not really participate in the discussion.

    Instead, why not publish the proposals in advance? Let the proposals sit out there for a few weeks, then convene the jury. This gives the general public a chance to chew on it themselves, and it also brings the press into a real policy debate, rather than just covering the event. Further, that few weeks will affect the jurors themselves. This would be a bad thing if this were a criminal jury; we’d say the jury pool was tainted. Not so here. Exposure to a national debate is a good thing in a legislative context.

    It might seem as though this robs the jurors of their agency, but I don’t think so. The members of the climate assembly were always in the position of a homeowner picking fixtures during a remodel, as the organizational structure and expertise came from outside the assembly. Publishing the proposals in advance brings the general public in, it doesn’t shut the jurors out.

    One last advantage: if the proposals come from the mainstream political system through, for example, the superminority method, then the public debate will also shed light on those mainstream politicians. Citizen assemblies shouldn’t just make decisions, they should also force elected politicians to show their cards.

    Liked by 1 person

  29. Oliver wrote:
    >” the question of normative democratic legitimacy turns into a question of whether the members of the demos have good epistemic reason to trust the “representatives” with the power the system gives them.”

    Perhaps, but only instrumentally (for a deeper reason). To have trust (which I agree is the key), what I want is fair-minded people without bias against people like me, who have heard sufficient pertinent information, considered options, and made a thoughtful competent decision. In other words, the group does not necessarily need to be representative, but more like neutral judges. However, having a fully representative group may be the optimal way of ACHIEVING this deeper goal I have to win my trust. For some people a representative group with representative prejudices against a minority I am a member of is not optimal (a black man on trial in a mostly white town in the Mississippi in 1950 isn’t thrilled to have a merely representative jury hand picked by the prosecutor).

    The point is, that elected representatives almost always openly admit they are not neutral and impartial, but have prejudged most policy matters (even setting aside for a moment concerns about catering to campaign donors). Random selection can generate representativeness AND impartiality, which hand in hand create trust. As with a trial jury (without the distortions of peremptory dismissals), random selection is the optimal tool for generating impartial bodies, even if not perfectly representative.

    So, I also disagree with Alex’s last post, because exposing jurors to persuasion efforts of elites (purchased media, etc.) before they convene as a mini-public IS tantamount to jury tampering.


  30. I fully share Oliver’s view that the question of legitimacy concerns essentially the epistemic reasons we have to trust our representatives (+ the general acceptability of the procedure, that political equality offers), and that independence, representativeness, deliberativeness, etc. can be epistemic reasons of this sort. My disagreement with Keith, then, I guess, would be that I value deliberation more than representativeness.

    Liked by 1 person

  31. OK, so five questions for those championing perceived legitimacy (aka “trust”):

    1. What if different samples end up generating wildly differing “verdicts” on the same topic (as with the three Texas utility DPs that we have discussed on this blog)? Which jury should we trust?

    2. What if the jury comes to a different decision from the population that it represents? I imagine in the case of the minimum age of the Irish president, participants in the referendum would have been told that the citizens convention supported lowering the age, yet the public clearly did not trust their judgment. Why should this be the exception that proves the rule?

    3. Why is there a definitional connection between trust and democracy? In theocratic regimes I imagine there is a high degree of trust in the Ayatollah, Pope or whoever.

    Oliver:> because they [citizen juries] are generally like us and share our fundamental values’

    4. Really? What about societies that are deeply riven by culture wars and the like?

    5. Those who rely on trust are claiming that Hobbesian “representation” is for real (rather than, as Pitkin concluded, a confidence trick). Are you comfortable with that?


  32. Keith,

    Thanks for the questions!

    1. It is not a question of choosing which jury to trust, unless there’s some sort of constitutional system in which multiple juries deliver verdicts and the public votes on which to follow. Democratic legitimacy on this model is a question of whether the citizens have good reason to trust the government actually in power with that power. ‘Juries deliver widely variable verdicts’ might be a reason not to trust rule by jury, of course.

    2. ‘Juries sometimes come to conclusions the public disagree with’ might be another reason not to trust rule by jury, but if the basis for your trust in the system of government is not ‘it will only do things I agree with’, that might not be a problem. Justified trust in a system or in a particular set of representatives is about their performance and goodwill in general and over time.

    (I’m kind of swerving between systems of government and particular representatives here because either or both can be objects of justified or unjustified public trust. So you can have more complicated situations where a government is only partly legitimate, or legitimate in one aspect but not another, and so on.)

    3. The connection is not between trust and democracy, but *trustworthiness* and democracy. If you want a conceptual link, you might say the demos have de facto delegated their power to the government, and whether they have good reason to trust that government with it determines whether that delegation was a bad idea or not.

    4. In a society riven by culture wars and so forth, probably not many people would trust the government on that basis. My list of examples was a list of examples people *might* give under *some* different circumstances.

    In a society riven by culture wars, actually, descriptive representation would probably be a bad source of legitimacy for just the reason you describe. A deliberative-consensual system, on the other hand, in which a descriptively representative sample of the population overcome their differences through deliberation to come to agreement on issues, might in such a situation gain trust by virtue of that positive moral example, and the system gain public trust because it produced such examples.

    5. I would take it the other way – all representation is fundamentally fictional, a kind of magical link by which one group of people ‘stands for’ another. The real question is whether the people with power over you can be trusted with it or not. So I guess I’m basically a kind of deflationist about democracy – I don’t think it’s coherent for ‘the demos’ to ‘be in power’ for the same reasons I don’t believe in ‘the people’ but only the public.


  33. Thanks Oliver, I’m at work now so can’t respond to your response (other than the obvious point in 1 that you are objecting to rule by sortition based bodies) but there’s one question that I forgot to put:

    6. I believe that SUVs are currently the best selling type of new car (by volume, not just by value). Yesterday the UK climate change citizens assembly published a 500 page report, and one of its recommendations was banning sales of SUVs. Given that citizens are likely to trust the judgment of “people like us” does this mean that sales of SUVs are now going to hit the skids?


  34. Terry,

    I look forward to your meta-response! In answer to your question 6, there is a world of difference between the public trusting some people’s judgement and those people being the government. What you’re describing is a collective action problem – people might very well accept SUVs being banned, while at the same time thinking it pointless to voluntarily abstain from buying them themselves, since they judge only a small minority of their fellows will do so. There is also a major difference in public standing between a citizens’ assembly put together by a pressure group and a constitutionally-mandated one at the head of the state.

    Liked by 1 person

  35. It was a project by the House of Commons, hardly a pressure group:


    Liked by 1 person

  36. Keith,

    My response to your 5 questions:

    1. Whichever jury met last, or if they all met a the same time, then a new jury that considered the various conclusions to make a final choice. The point is that there IS NO RIGHT ANSWER. Even the exact same jurors meeting on a different day might reach a different conclusion. The point is that a sortition system is generally representative and impartial (without electoral ambitions), and decisions of jurries will regress to the mean, wobbling around the hypothetical “right” answers without a consistent bias in favor of some special interest or elite.

    2. It is to be HOPED that a roughly representative group of citizens who devote substantial time and effort to understanding an issue will come to different decisions than the population as a whole (or they themselves) would make when having almost no information and giving almost no time to thinking about it. the legitimacy is derived from people appreciating that their off-the-cuff opinions might be very different if they studied the matter closely, and trusting that an impartial jury has done this. the Irish constitutional amendment on the minimum age of a president is a good example. In that case very few voters even KNEW that a jury had considered it and recommended the change. But this is also a good example of one of the rare cases where a referendum made sense… because what matters most on THIS topic is what the uninformed people feel comfortable with.

    3. I don’t make a definitional connection. This thread is about LEGITIMACY rather than democracy. I see democracy as a set of procedures and principles, while legitimacy can be well-founded, or the result of monopoly control of information by a dictator. The hope (and plausible case) is that a sortition democracy will indeed win legitimacy with trust.

    4. As in deeply divided and partisan electoral “democracy” the best that can be hoped for is that citizens will trust the impartiality of juries and believe that with good arguments their “side” may carry the day in a future deliberation. A key advantage of sortition democracy is that it doesn’t have a built-in motivation for legislators to exacerbate divisions, which are often driving tribalism in electoral systems. Jury members can have a motivation of seeking win-win outcomes, even if these can’t always be found, while electoral system motivate the nominal “deliberators” to demonize each other and AVOID win-win outcomes because they are not helpful for winning in the next election cycle.

    5. I do not endorse Hobbes’s notion. Trust is necessary for legitimacy, and sortition democracy has the best reasons for giving the government trust.

    Liked by 1 person

  37. Keith,

    Just realised that in a fit of absent-mindedness I called you ‘Terry’ in that last comment. My apologies!


  38. And apologies to the real Terry for taking his name in vain


  39. I agree with Keith that popular trust is not enough for democracy. For a regime to count as democratic, I’d say that people should have means to express their trust and distrust. And by this, I mean not only the freedoms of expression and association, but also institutionalized means to express their disagreement with the way they are represented. Electing dissenting representatives is one way. Recalling disappointing representatives is another. And the abrogative referendum is another yet. Recalling allotted representatives seems too harsh to me, but maybe we should consider coupling sortition with the abrogative referendum.


  40. Responding to Oliver and Terry, as per the earlier numbered points:

    1. “Juries that deliver widely variable verdicts might be a reason not to trust rule by jury.” The limited evidence we have (the Texas utilities DP being the best example) would suggest that this is indeed the case. And the majority rule in democracies has a more exacting standard than ruling out wide variability. Agree with Terry that there is no “right” answer in the epistemic sense of the word, but it’s no good saying it will converge on the mean over time. Just as we have to accept the results of each election (in which everybody gets to vote) then a sortition-based system would need to return the same verdict on every occasion, irrespective of who is included in the sample. That’s what majority rule means. The fact that a later jury can overrule the decision is not the point.

    2. I don’t see why the age of the Irish presidency is different from any other political decision. Speaking for myself I would only accept the verdict of a legislative jury where it had been demonstrated that it would have made no difference who was included in the sample (see 1, above). We all believe in well-informed decisions, but would only trust the decision given the above criterion.

    3. Neither of you seem prepared to accept P-E’s distinction between perceived and formal legitimacy.

    4. There seems to be a lot hanging on “the best that can be hoped for”, mights, maybes and positive moral examples. One is reminded of Madison’s observation that if men were angels then government would be unnecessary. Alex seems to be the only one prepared to grasp the nettle of difference and turn it into a positive virtue, and has pointed out that the polarisation being currently experienced is a product of the decision rule.

    5. Hobbes certainly claimed that representation was fictional — his principal examples being drawn from acting and writing (“authorisation” is what an author does). I would hope that we might be less deflationary in our commitment to democracy.

    Agree with P-E that democracy requires concrete means to express trust and distrust, and this means resorting to an institutional definition.


  41. Keith,

    In response to your response to our responses,

    1. Your demand for consistency in the outputs of the juries rests on the idea that, for every political problem, the opinion each person would have on it if they thought about it is somehow already determined before they do so. That’s an inference we would have to draw if juries always delivered the same responses to the same problems. And I just don’t think that’s a realistic picture of the way we think – there’s a significant amount of path-dependency to our opinions.

    2. I don’t think very many people decide legitimacy on a decision-by-decision basis. Very often in happier times we would look at a government decision and think, ‘That’s clearly a bad decision, but it’s legitimate because the system as a whole is legitimate.’ In other words, ‘the system is using its power wrong in this case, but I trust it with its power on the whole’.

    3. This isn’t the case – I’ve been talking about empirical versus normative legitimacy, but it’s the same distinction. My position, put crudely, is that empirical/perceived legitimacy is whether the public *do* trust their government, and normative/formal legitimacy is whether they *should*.

    4. Alex’s superminority proposals are fascinating and promising, but they still rely on a second chamber of some kind picking a winner when measures contradict one another. There must still be winners and losers, and a system that tends to dampen down the animus resulting from that holds a significant advantage over one that tends to gin it up.

    5. I think this comes down to what I was talking about in point 1, above – whether or not you think ‘the people’ are a coherent agent capable of rule.

    I agree that a legitimate state must have institutional means of dissent and disruption of the power of the powerful. That’s a product of the ‘good reason to trust’ model of legitimacy combined with the grim facts about how people and power operate.


  42. Oliver,

    1. Path dependency is the problem that I’m seeking to address. For example the sequence in which people get to speak (and the status and persuasiveness of the speaker) can have a strong influence in swaying the debate one way or the other. When I did jury service all my colleagues were in favour of a guilty verdict but I managed to persuade two of them to switch to innocent so the jury was dissolved and the judge ordered a new trial which swiftly returned a guilty verdict. The second outcome was the democratic one even though (IMO) the verdict was wrong.

    2. In government by political parties, voters are obliged to go for a manifesto package which bundles together many unrelated issues. Many voters, for example, might want a left-wing fiscal policy but are then obliged to accept liberal social policy. The great thing with ad hoc sortition is there is no bundling.

    3. The is/ought issue is orthogonal to perceived/formal legitimacy. Electoral legitimacy depends on the principle of choice (Manin claimed that the triumph of election was on account of the natural law theory of right). So we need an equivalent formal principle for sortition — I would claim that the principle is resemblance and that the ongoing resemblance of the sample to the target population requires some serious constraints on the design and mandate of sortition bodies.

    4. Sure. My claim is that if it can be demonstrated to the losers that it would have made no difference if they attended in person (they would still be in the minority) then they will accept the outcome.

    5. I don’t believe that “the people” are a coherent agent capable of ruling, that’s why my preference is for a mixed constitution in which sortition plays a part.


  43. Keith:> I would claim that the principle is resemblance

    I agree, and I think it’s important to put it this way. We forget that most people aren’t responding to high-minded philosophical principles, but to the visceral experience of politics.

    In that vein, I think we should focus on incrementalism. One of the great sources of illegitimacy in the legislature is its volatility, both real and perceived. Huge laws are passed by narrow majorities, and are then reversed once another party comes to power. One feature of juries coupled with the superminority system is a large increase in bandwidth. You can have many juries going at any time working on different proposals, which are easy to generate because no consensus is required. The best use of this consensus is to make smaller, more incremental laws.

    More broadly, I think part of our test for legitimacy should simply be the “cortisol” test (also known as the “Oh, s***” test). Basically, anything that causes people’s cortisol levels to shoot through the roof is illegitimate. For example, in the executive selection process, I think it’s better to select thousands of political office holders using hundreds of juries. Then it’s just normal. Choosing only a few high level officers will actually be seen as less legitimate, due to its seeming unusual.

    U.S. Senate Majority Leader Mitch McConnell has a strategy that seems counterintuitive. He doesn’t care much about passing legislation, instead he just packs the courts with right wing ideologues. This drives liberals crazy, but it passes mostly unnoticed by the general public. Why? Because it’s normal business. It just happens. The courts, in turn, can legislate incrementally, because all they have to do is pluck out the cases they want, then make rulings–all of which appears perfectly normal, even though the policies enacted are quite radical.

    Compare that to the Affordable Care Act (aka Obamacare). As a matter of substance, it was a pretty mild reform. It left in place the existing system, and filled in some gaps around it. But the process by which it passed seemed dramatic and unusual. The consensus based assembly, with its limited bandwidth and copious opportunities for shambolic protestation, turned what should have been an ordinary piece of legislation into circus.

    Liked by 1 person

  44. Thanks for your last comment Alex,

    I agree with your priorities. A great deal of the discussion on this blog is in terms of high blown concepts that barely make it out of learned journals. Some of them have their everyday equivalents, but many barely do.

    I think the issue of bandwidth is a huge one, but not one I’ve done enough thinking about.

    Liked by 1 person

  45. Alex and Keith,

    Having read your ‘Isegoria and Isonomia: Election by Lot and the Democratic Diarchy’, I’m quite taken by the distinction you draw between proposing and disposing as functions of reason, so I’m more open than I was to your preferred silent disposing juries for the legislature.

    Keith:> “The is/ought issue is orthogonal to perceived/formal legitimacy. Electoral legitimacy depends on the principle of choice (Manin claimed that the triumph of election was on account of the natural law theory of right). So we need an equivalent formal principle for sortition — I would claim that the principle is resemblance”

    In that case, I misinterpreted the perceived/formal distinction, and you were right in your initial assessment of your position. In Alex’s words, “We forget that most people aren’t responding to high-minded philosophical principles, but to the visceral experience of politics.” The implications of this are twofold – first, that formal principles of legitimation aren’t all that important to a state’s empirical legitimacy, but second, and more importantly, that theoretically inconvenient aspects of political life are often as normatively significant as those elements of it that lend themselves to grand, simple theories of legitimation. Formal principles of legitimacy, beyond their use as slogans, obscure this important truth about what normative legitimacy really is, making it harder to think clearly about it. I regard them as doing more harm than good.

    Alex:> “I think we should focus on incrementalism.”

    I agree insofar as we need more proofs of concept in order to get sortitional institutions to register and appear palatable in the public eye. But because instituting them in the height of government requires disempowering existing power-holders, there will be some threshold beyond which incremental changes becomes increasingly difficult to get and to keep. The process will have to be adversarial.

    One potential way to go about it might be through a specialised Sortition Party. This party would operate in a particular way: upon getting a representative into (let’s say) Parliament, it would hold a lottery on the electoral roll, selecting a random citizen to be the real representative. The elected rep themselves would simply relay the allotted rep’s words, dictated through an earpiece, to the chamber, vote as they dictate, and so forth, and split their salary with the allotted rep. The aim would be, first, to show the feasibility of putting ordinary people in power, and then to take the offices required to change the constitution to institute sortitional democracy. Because no members of the party would hold power themselves at any point, the corrupting influence of that power on the party’s mission would be minimised.

    An alternative, potentially more costly but epistemically better, means of operation would be for the Sortition Party to set up a citizen’s assembly parallel to to the real parliament, with this assembly delegating some of its members (not Sortition Party members!) to play the roles of the allotted reps as I’ve described. The Party might shift from the first to the second mode of operation as it grew. It might also hold deliberative events to help inform its allotted citizen representatives.


  46. *your initial assessment of my position


  47. Keith wrote:
    >”Just as we have to accept the results of each election (in which everybody gets to vote) then a sortition-based system would need to return the same verdict on every occasion, irrespective of who is included in the sample.”

    This makes no sense. An election on day 1 might elect one legislator, while the same electorate might elect a mirror image opposite legislator on another day. Popular elections AND legislative chambers are not consistent… If you think delivering the same decision is the measure of true representativeness, then all legislatures clearly fail.

    Your typical response is something like, “yes, but everybody got to VOTE for their legislator, so that automatically makes them representative, while with sortition, I probably won’t be selected to serve.”

    This argument is also deeply flawed. Even if we set aside the corruption of the electoral process, money, media or elite domination, the narrow range of choices on offer, etc. … Let’s even set aside the fact that many voters and in elections like those of the UK and America –first past the post, winner take all — it is often a majority of voters who did NOT vote for the declared winner. So… Looking at an imaginary fair and honest election… as a voter I have much LESS chance of casting a vote that matters in terms of determining the winner (it would need to either break a tie, or create a tie), than I have a chance of being selected to serve on a mini-public. The voter has far LESS likelihood of mattering in selecting representatives than a citizen in a lottery pool. It is only the CUSTOM and familiarity with voting that causes people to pretend that they have “selected” their representative, and give that selection legitimacy. If we were accustomed to using lottery, that same familiarity would apply, and we would be stunned to think that casting one vote out of thousands or millions gives any legitimacy to the winner as MY representative.


  48. Oliver,

    Various people in various countries have come up with the same idea of using sortition to direct a member of parliament from a “sortition party” how to vote, or inserting randomly selected citizens into seats won by the sortition party. This is totally unworkable. As for the insertion of a random citizen, this absolutely defeats the representativeness of these members, as it is only large numbers that create representativeness of a groups as a whole. As for use of mini-publics to direct a sortition party MP, this is also nonsensical. As a former legislator, I can attest that MOST policy work is NOT on the floor of the chamber on formal votes. Most of the important work is done behind the scenes, with no formal process at all (trading votes, cutting undisclosed deals, etc.) Even just the need for quick decision-making assures that a mini-public process can not function to guide a member of parliament. Sortition CANNOT be incorporated INTO a chamber of elected members. Sortition needs to be used to create bodies that are separate, not inter-mixed.


  49. Terry:> An election on day 1 might elect one legislator, while the same electorate might elect a mirror image opposite legislator on another day.

    This is a problem that afflicts all forms of democracy — direct, elective and aleatory. The Athenian assembly was known to decide one thing one day and then reverse the decision the next day; electoral democracy is just as fickle (as you point out) and allotted samples could return different decisions over time. But in every case the decision at that time would be deemed to be the sovereign voice of the demos. But that cannot be the case with two concurrent samples returning different verdicts as there would be no way of knowing which was the representative decision. (If I had a dollar for every time I said that I’d be up there with the hated 1%).

    >It is only the CUSTOM and familiarity with voting that causes people to pretend that they have “selected” their representative, and give that selection legitimacy.

    Not so. At the time of the birth of representative government, the prevailing systems were monarchy and aristocracy, and the historical references to democracy were to selection by lot in the ancient world. Voting was an unfamiliar practice to most citizens. Manin attributes the decision to opt for election rather than sortition to the Natural Right theory of consent (choice) which was influential at the time. Although the mathematical notion of probability goes back to Cardano, Pascal and Fermat it was not very well known and the technology was not available to put it into practice. In sparsely-populated countries like America it was much easier to elect a representative, put him on a horse and send him off to parlay on their behalf, but the key factor was that the representative was chosen by the people.

    If there was any element of custom it was the religious notion of the elect — a natural aristocracy that could be identified by ordinary voters (the American founders were much more familiar with Calvin than Locke). Nowadays we are (rightly) sceptical about this and the technology does exist to implement sortition but we still need to identify the legitimising principle (resemblance) and implement sortition in such a way as to ensure that the principle is not corrupted in practice. We all agree that the sample should “look like America”, but it should also act like America and this involves some serious constraints including (inevitably) sleeping with the electoral enemy. I’m sorry to bang on about it, but it’s all there in Pitkin — we need to stop trying to reinvent the wheel.

    We also agree on the pathologies of elections that you outline, but these are addressed head on by Alex with his notion of superminorities. In theory you can have any number of alternative agendas (the only constraint being practicality), but it still fully respects Dahl’s insistence that the demos should have complete control of the agenda — a principle undermined by your self-selecting agenda councils. And ho boulomenos in large states is even worse as it merely privileges those with the loudest voices (and deepest pockets).

    I know it would involve the pain of a second act of apostasy but there comes a time when you’re in a hole that the best solution is to stop digging and acknowledge that democracy in large states will always involve an alliance between election and sortition. I know you’ve given yourself 200 years to put your utopian plan into practice, but let’s go for something that can work straight away — as Alex puts it, all you need to do is change the decision threshold in proposing chambers and appoint a jury — both these are entirely uncontroversial and fully respect natural right theory (in the former) and mathematical probability (in the latter).

    So what’s the problem?


  50. Just a quick note: the strategy advocated by Oliver and rejected by Terry is actually on trial in the Brussels regional Parliament, where the newly created party Agora obtained one seat in recent elections. They organize citizen assemblies on a diversity of issues and the MP does nothing else in Parliament than advocating sortition and transmitting the assemblies’ recommendations.

    This strategy is highly controversial because the MP “wastes” votes by abstaining on any issue that has not been previously discussed by citizen assemblies (and it’s impossible to have an assembly on each issue). Yet it might have contributed to the even more recent implementation of sortition in “deliberative” committees of the Brussels Parliament.

    Liked by 1 person

  51. Keith wrote above:
    >”with two concurrent samples returning different verdicts as there would be no way of knowing which was the representative decision.”

    Either the two juries could confer and seek common ground, or a third (larger) mini-public could be called to settle the matter. But your focus on this is odd, since this EXACT problem regularly occurs all the time in bicameral ELECTED legislatures. The house votes for policy A, and the Senate votes for policy Z. Which is the “representative decision?” We don’t need to have differing election rules like the US House and Senate… in 49 of the U.S. states both chambers are elected using the identical one person one vote rule (since the series of Supreme Court rulings in the 1960s). So it is typically the same voters on the same election day generating two elected chambers who often are exactly opposed to each other on particular issues. Yet you only harp about a hypothetical dual mini-public disagreement.


    Keith also wrote misrepresenting my design with a randomly selected agenda council, and mini-publics that can review, reject, meld, or amend proposals from any group of citizens:
    >” your self-selecting agenda councils.”
    Because I do not propose that a mini-public agenda council have mandatory service, it is likely many citizens would decline to serve — so Keith makes the pejorative leap to calling these “self-selecting agenda councils.”


    And then Keith wrote: about my open to all proposal design…
    >”And ho boulomenos in large states is even worse as it merely privileges those with the loudest voices (and deepest pockets).”
    In my design ANY citizen can participate on a panel to try and draft proposals (ho boulomenos) as raw input for mini-publics to consider. Keith instead supports a design where political parties get to have a monopoly on the drafting of proposals for a jury to consider, and yet somehow call MY design the one that “privileges those with the loudest voices (and deepest pockets).” That is the perfect description of the design he favors that excludes ordinary citizens from policy generation and limits it to party activists who have the loudest voices and can get enough money to win elections.



  52. Terry,

    > political parties get[ting] to have a monopoly on the drafting of proposals for a jury to consider […] is the perfect description of the design he favors that excludes ordinary citizens from policy generation and limits it to party activists who have the loudest voices and can get enough money to win elections.

    Very true. But this does not contradict (in fact, it demonstrates) the general rule that any mechanism of formal equality translates to privileging the powerful.

    It is only through sortition that this characteristics of mass politics can be avoided and democratic control of the agenda can be achieved. Thus agenda setting must be done by an allotted body without tying its hands in various ways through a formally equal (and de-facto oligarchical) mass political process. (Interestingly, it appears this was indeed how the Boule functioned in Athens.)

    Liked by 1 person

  53. Terry,

    Leaving aside the Founders’ desire to frustrate simple majorities, the difference is that electors choose which persons they wish to exercise their independent judgment. The aleatory equivalent is that sortition “delegates” are chosen to exercise their judgment by the invisible hand of the LLN, but that presupposes a stochastically-determinate relationship between representatives and their constituents. If you can’t demonstrate that empirically then the judgment has no democratic legitimacy. The mistake you are making is to think that the only difference between elected and allotted representatives is the balloting mechanism.

    >Keith makes the pejorative leap to calling these “self-selecting agenda councils.”

    I acknowledge using hyperbolic language but if (like me) you believe the only democratic value of sortition is representing the preferences, beliefs (and prejudices) of the target population descriptively, anything that contravenes Dahl’s principle that the demos should have exclusive control of the agenda is anathema.

    >Keith instead supports a design where political parties get to have a monopoly on the drafting of proposals for a jury to consider.

    Absolutely. And Alex’s superminority proposal ensures that the proposals will be broadly representative of the preferences of the target population. It’s hard to envisage a democratic alternative — and we can rule out one based on principles designed for tiny direct democracies.


  54. Keith wrote:
    >”electors choose which persons they wish to exercise their independent judgment.”
    You must know that isn’t true. Why do you say it? Voters are only effectively allowed to choose from a microscopic range of choices of elite or elite-sponsored “loud mouths.” And then, many of these voters don’t even get to have the representative they CHOSE get elected. So they are “represented” by somebody they disagree with on every important issue. The notion that elected members are actually representatives is a formal myth only. The mythology that the voters CHOOSE their elected representatives is recognized as fallacious by anybody who examines the myth for more than a moment.

    Keith then wrote that letting these extremely unrepresentative partisan loud mouths have a monopoly on agenda setting somehow conforms with:
    >”Dahl’s principle that the demos should have exclusive control of the agenda…”

    Liked by 1 person

  55. > “The mythology that the voters CHOOSE their elected representatives is recognized as fallacious by anybody who examines the myth for more than a moment.”

    I think that’s too strong! There is a dimension of choice, and a dimension of constraint (as in most choices). And choosing from a range of options can already be seen as more valuable than not choosing at all, even if the range of options is too narrow and unsatisfactory on many accounts.

    Liked by 2 people

  56. Terry,

    Leaving aside for the moment the question of “perceived” legitimacy there are two distinct aspects to our exchange: 1) formal principles and 2) practical constraints.

    1. IN PRINCIPLE: We both agree that democracy in large multicultural poleis presupposes some form of representation. Electoral representation is based on the principle of choice and stochastic representation is based on the principle of similarity — a random sample being a microcosm of the target population. These principles are true in the formal (a priori) sense, irrespective of the degree that they are implemented in practice.

    2a. IN PRACTICE — ELECTIONS: We all agree that existing electoral institutions constitute a piss-poor form of choice. But, as Alex has shown, this is on account of the decision threshold (51%) in legislative institutions. If the threshold could be reduced by an order of magnitude then the choice available would be hugely enhanced, as small political parties would stand a real chance of having their agenda considered and implemented (by a randomly-selected jury). In practice an order of magnitude would probably be too much, so Alex has suggested a decision threshold of around 20%, but there is no preset rule to limit the representation of the diversity in the target population. Unless you have a principled objection to free-market mechanisms, that sounds like an effective way of ensuring that the demos has exclusive control of the agenda (and would involve only a tiny tinkering with existing arrangements).

    2b. IN PRACTICE — SORTITION: We both agree that stochastic representation requires large quasi-mandatory samples, balanced advocacy and a tight restriction on speech acts, so there is no dispute here. The problem only occurs when you try to use sortition for the agenda-setting process.


  57. Keith:> If the threshold could be reduced by an order of magnitude then the choice available would be hugely enhanced,…

    As a technical note, this quantity is easily measured. For example, take a 100 person legislature. The current norm is to get one output only, requiring 51 votes, effectively wasting 49. For two outputs, the threshold is 34, so that 68 votes contribute to the options, wasting 32 votes. Here is a short table showing this dynamic:

    Total Votes
    Wasted Votes








    With five options, we have reduced the number of wasted votes from 49 to 15, a reduction of more than two-thirds.


  58. Sorry, the table didn’t render the way I wanted. Here is a simpler version:

    options | threshold | total votes | wasted votes
    1 51 51 49
    2 34 68 32
    3 26 78 22
    4 21 84 16
    5 17 85 15
    6 15 90 10
    7 13 91 9


  59. ARG!!!! I can’t get it to render!!!!!

    options | threshold | total votes | wasted votes


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 )

Google photo

You are commenting using your Google 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: