Citizens’ assemblies are open to manipulation

Naomi has just flagged up an interesting article from Village Magazine on the Irish experiments with constitutional conventions in 2011 and 2016 by Eoin O’Malley, one of the participants. The article adds to my concerns that ‘full mandate’ allotted bodies are open to manipulation:

Some research shows that the act of deliberating with others has an impact beyond exposure to arguments or evidence. That is people given the evidence and arguments don’t move as much as those who are asked to discuss that evidence and arguments with others. This sounds like something positive for deliberative mini-publics. But it might not be.

The reason for this is because (as Condorcet demonstrated), independence is the key to getting the ‘right’ answer and this suggests that communication between jurors should not be encouraged — all that is needed is exposure to balanced arguments and evidence (as Goodin and Niemeyer discovered in their study of the Bloomfield Track citizens’ jury). This is distorted by the need to come to collaborative conclusions:

Because they are not independent the same flawed thinking or arguments can be magnified. For instance we could see the citizens in the mini-publics engage in groupthink. Some opinions might be aired, but can be effectively suppressed by the atmosphere in the room. There is significant evidence in social psychology that groups can push opinion to extremes and silence minority opinion. To prevent this great care has to be taken that all views are respected.

Fishkin’s insistence (personal communication) that small-group deliberation is an essential part of the opinion formation process might well be on account of his need to demonstrate that the DP leads to significant alterations in opinion (irrespective of whether or not such a shift would truly represent the target population). O’Malley’s experience validates the need for an adversarial and dialectical information advocacy policy to ensure well-balanced and opposing briefs:

The choice of experts was highly significant. In an earlier experiment I did with others we found that when experts agreed the members of a citizens’ jury tended to follow the expert judgement. While we did not deliberately set out to produce a result, the fact that our experts all agreed with the retention of the Irish electoral system meant there was less debate on its potential flaws than there should have been.

There would be no guarantee that leaving the choice of experts to the happenstance make-up of the randomly-selected body would ensure a diverse choice of experts, especially as political amateurs would be highly reliant on the suggestions of the administrative support staff, along with ‘suggestions’ from powerful lobby and media interests. Another problem was the pseudo-random outcome of the (voluntary) sortition process:

Another problem is that the 66 so-called ‘citizen members’ were not randomly chosen. Aside from the problem in identifying ‘random’ citizens, anyone who agreed to spend multiple weekends discussing sometimes arcane constitutional provisions had to be unusually interested in politics. There was a strong self-selection bias driving the composition of the Convention, and they may not have had representative views. Practical matters also got in the way. Few (if any) mothers of young children were there presumably because of the onerous time impositions. If the members are not representative, it is questionable what claims to legitimacy the Convention can have.

Or as one of the comments on the blog website put it the (voluntary) sample was unrepresentative because ‘it was more likely that a Dublin-dwelling Irish Times reader would agree to it than the average citizen who might have duties to their family on a Saturday.’ This would suggest the need for a larger sample, as in the Athenian legislative courts, and that attendance should be at least as mandatory as judicial jury service.


28 Responses

  1. Re manipulation

    The basic question is what are the grounds on which a community should rationally assent to a certain arrangement as “right”? Is it whatever a majority will assent to, prior to any well-informed and critically evaluated debate? Can it be assumed that people generally will have an adequate understanding of the pros and cons of various possible constitutional arrangements?

    I believe that most people are well aware that what seems plausible in such matters will very often contain serious flaws. As in any important matter, what they would hope to be offered is a thoroughly discussed set of consistent arrangements, based on arguments on which they have a genuine opportunity to comment at any point in the process.

    In a face-to-face discussion that is impossible, since the time open to any speaker must be limited. On an internet site it is possible for everybody to have a say. Very many people who lack the ability to express themselves clearly in public speaking can, especially in collaboration with others, find ways of expressing the point that they wish to make clearly and relate it to what others have said.

    In cold, considered print the psychological pressures that so easily distort viva-voce discussion can easily be minimised. It is very likely that the major problem of getting a hearing will no longer be the dominance of the forceful personalities and skilled orators or slogan-makers, but that the audience one is trying to reach is snowed under by the mass of material. There will be much repetition of any particular point by people who fail to recognise that what they are saying has already been said and answered. It needs to be emphasised that the point of the discussion is that every relevant consideration is brought out explicitly. not, at this stage, how many people ant to support that consideration, In this respect, too, since the written word remains there to be seen long after it is posted, the difficulty of timing that afflicts live discussion is minimised.

    Any practical proposal is going to involve many distinct considerations. It would be important that all the comments on each consideration be grouped together by professional editors, who might single out certain contributions as clear expressions of a certain contention, subject to the right of aggrieved contributors to object to the treatment of their contribution.

    In any community with a history behind it, unless it is divided by deep and pervasive conflicts, it is to be expected that there will be a substantial degree of agreement about what characteristics are desirable in any solution to the problem that faces them. But it is also likely that People will put different relative weight on values that they share. All may agree that they value both liberty and security, but differ about how much of the one they are prepared to sacrifice to the other. There can be no formula that can produce a right answer to such differences in general terms, but if there is to be a binding common arrangement, it is necessary to reach a generally acceptable balance between
    those values in a particular arrangement in the light of the consideration shown to be relevant to the issue under discussion.
    The only way to reach such a generally accepted decision, it seems to me, is negotiation between those with different preferences. Such negotiations must involve some accepted scheme of representation. But the important thing is that the negotiations be entirely public and in writing. Ultimately, in constitutional matters \, the result will need to be endorsed by a plebiscite. But the value of any vote is only as good as the grounds for the decision it endorses.


  2. The author is certainly right that the “citizen assembly” model (short-term, ad-hoc body) is very problematic. Many of his points are strong. This point for example is spot-on:

    Quite why gathering randomly chosen people and asking them to discuss something for a weekend is perceived as more legitimate than asking our elected representatives to make that decision is moot. It says more about the low esteem in which party politics is held than it does of the process of the Constitutional Convention.

    However, the “independence” argument is completely wrong:

    Usually we ask more than just one person to make decisions because we assume that a large number of people coming to the same conclusion are more likely to be right. That’s the logic behind juries. If 12 people independently think you’re guilty, the likelihood that you’re actually guilty is high. But independence is key. A problem with juries, including citizens’ juries such as constitutional conventions, is that the logic assumes that the people form their opinion independently whereas they actually collaborate to come to conclusions.

    Quite the opposite. At least when it comes to policy making, collaboration is exactly what is aimed at. The notion that there is some sort of a “balanced” background knowledge based on which the average citizen can construct, in isolation, an “independent” policy opinion is not only silly and obviously false, it is manipulative because it aims to give those who control the background knowledge, the prejudices, conventions and preconceptions effective control of decision making.

    Liked by 1 person

  3. John,

    I’m astonished that, given the experience of this internet site — in which no progress at all has been made on the topic as a result of the exchange of views — you still hold this at the core of your demarchy proposal. We all continue to talk past each other. Is it the lack of “professional editors” that has prevented the emergence of a consensus?


    >The notion that there is some sort of a “balanced” background knowledge

    The balance is achieved by a dialectical process. If the brief of the assembly were to come to an informed decision on Britain’s membership of the EU, the information would come from the two opposing organisations (Remain and Vote Leave). If the provision of information were left entirely to allotted members it would reflect the random prejudices of the tiny proportion of the group who happened to know something on the topic, the assembly secretariat and the swarm of lobbyists knocking on the door. The chances of balanced information and advocacy are practically non-existent.

    Two simple questions — the adversarial exchange of information is the best (only?) way we have of establishing the truth in a judicial trial; if so, then (1) why would you not use the same mechanism for the trial of legislation in the High Court of Parliament? What you are suggesting is equivalent to dispensing with the services of judge, prosecution and defence and leaving the whole trial in the hands of the jury. Most people would dismiss this as some sort of kangaroo court, yet that is exactly what you are suggesting for the trial of legislative proposals. (2) Why do you deliberately ignore the only successful template we have for this sort of process (i.e. the 4th century Athenian legislative court)? The Athenians would, rightly, dismiss your proposal as thorubos.


  4. John,
    minor point…The idea that deliberation should be written has a long pedigree. Jeremy Bentham argued that oral debate was flawed (oratory skills/emotion/psychology) and that legislative arguments should always be reduced to writing. Your suggestion that in a written system of deliberation that “comments on each consideration be grouped together by professional editors,” however, raises a concern about undue influence by such editors. I don’t know much about it, but there is a quickly growing body of work dealing with online collaboration that uses a variety of crowdsourcing techniques to handle this task without reliance on outside expert editors. It is worth looking into this.


  5. As I argue in my multi-body sortition paper, if we rely exclusively on give and take deliberation by a mini-public (as Yoram favors) we risk the dangers O’Malley discusses of groupthink, flawed information cascades, a tendency to discount our own private knowledge in favor of group “going along,” polarization etc. But if we rely exclusively on independent assessment of presentations (as Keith favors), we forfeit the benefits of give and take group deliberation that can uncover win-win possibilities, etc. My solution is to simply have two distinct (sequential) mini-publics each avoiding one of the risks while maximizing the benefits of the other approach. Another argument for this strategy is the principle that the AUTHOR of a proposal should not also act as the JUDGE of that proposal… thus the mini-public that drafts a final proposal (necessarily less representative due to the time commitment and interest needed) passes their product on to a more representative semi-mandatory mini-public (policy jury) with short duration (so young mothers can participate), which hears pro and con presentations on the proposal before voting.


  6. Terry,

    I agree with the distinction involved, merely arguing that sortition is not the best way of establishing the necessary level of cognitive diversity for the first (authorial) panel. As you know I do make the argument that both functions (isegoria and isonomia) need to be representative and that the level of voluntarism required for the authorial panel would make it less representative than a body constituted by election. Your authorial panel is quite similar to a demarchic council, the principal difference being that in John’s proposal, recommendations generated by the demarchic council need to gain the approval of the general public before being subject to scrutiny by the allotted panel. Given that the second function is primarily a veto role, if the first panel is unrepresentative (in that it will be dominated by activists) then this will skew the final outcome. I can see that this is unproblematic for those (like yourself) on the progressive wing of politics, less so for those of us with a conservative inclination who don’t like fixing things that are not broken (or are persuaded by the law of unintended consequences).


  7. Keith,

    Setting aside for now the additional level of Interest Panels in my design (which generate the raw material for mini-publics to chew on and are akin to demarchic councils in that they will likely be be populated by those MOST concerned with the policy area)… I disagree with your statement that for a deliberative, give and take mini-public (Review Council) “the level of voluntarism required for the authorial panel would make it less representative than a body constituted by election.”

    Clearly real world election systems have delivered overwhelmingly non-diverse bodies dramatically over-representing males, wealthy, professional and business owners of the majority religion and ethnicity. There is no way a random sample, even distorted by the degree of voluntarism necessary to serve for any extended period, would be any where near as UNrepresentative as an elected chamber. In addition various scientific stratification tools can be used in drawing the sample. While stratification can be done in theory with an electoral system (some jurisdictions require party lists to have a minimum percentage of women, for example), the accuracy of representativity will likely be vastly superior under either a random sample, or scientific sampling system than an electoral one.


  8. I am puzzled by the extent to which there is so little reference to the extensive body of professional knowledge and practice regarding dialogue and deliberation (e.g., which embraces dozens of methodologies through which people in groups can better hear each other and better digest and evaluate information, thus minimizing groupthink and manifesting more collective intelligence (the group being smarter together than any of its members individually). My own niche in that field is to inquire how we can not only be smarter together, but actually wiser together. I have my own thoughts about that, but I think the importance of my answers pales in comparison to the importance of the general inquiry into collective wisdom (given our dire civilizational problematique) and of practical tests of anyone’s answers. I believe discussions like those in this post and its comments would benefit by moving beyond arguments that assume groupthink and manipulation will or will not occur to exploration of how to minimize groupthink and manipulation in any deliberative activity, and who should be involved (and how) in order to maximize the potential for collective wisdom to emerge. – Tom Atlee,


  9. Terry,

    You are assuming descriptive representation (age, gender, income) as an intrinsic good, whereas I see it as a rough proxy for political beliefs. I don’t much care what my political representative looks like but I do care what she believes in. A system that encourages participation by activists and other forms of political anoraks may look a lot more like America than a bunch of rich old white lawyers, but the latter are constrained to track the interests and preferences of their electorate, whereas the former are only accountable to themselves. There is a serious danger that an all-sortition system will just be a form of symbolic representation (a portrait of America, in miniature) whereas what we want is something that acts like America. This, IMO, requires a mixture of election, direct-democracy and sortition — the first two mechanisms being suitable for policy innovation and the latter for aggregate judgment only. And the latter must be quasi-mandatory, in order to retain its descriptive legitimacy.


    Quis custodiet ipso custodes?


  10. Yoram,

    >those who control the background knowledge, the prejudices, conventions and preconceptions [have] effective control of decision making.

    Obviously a randomly-selected group will not be immune from prejudices and preconceptions, your argument (I assume) being that this is OK from a democratic perspective as the prejudices and preconceptions of the microcosm will accurately mirror those of the target population. OK fine, the problem, however, arises because some will have stronger preconceptions than others and, if gifted with high status and/or persuasive powers will be able to impose these on other members. This is particularly relevant to choosing advisers and sources of information, as most randomly-selected conscripts will know diddly-squat about the topic under consideration. This is why the only way of ensuring balanced information and advocacy is to arrogate this to the hands of the competing (elite) parties to the dispute. To leave it down to the intrinsic makeup of the grouping would be, er, random (and unrepresentative of the target population). I’m puzzled that you don’t appear to grasp this blindingly obvious point — perhaps it’s because you’re a statistician and, as such, not particularly interested in group dynamics and other aspects of social psychology.


  11. sorry, I should have said ‘this blindingly obvious point”


  12. I had to look up your Latin, Keith. “Who will watch the watchers?”

    This is always a good question – AND it is provocatively self-reflexive (Who is asking about who will watch the watchers, and why?) and recursive (Who will watch whoever is watching the watchers?… Who will watch whoever is watching the watcher-watchers?….ad infinitum…)

    I’m particularly interested in who would oversee the deliberative and facilitative methods used in carrying out citizen deliberations (especially when such panels are empowered to have significant impact on public policy). This focus comes from my sense that deliberative designs and facilitation approaches can have a profound impact on the quality, fairness, and collective wisdom of both the process and the results – AND that the general public and most officials of various sorts are profoundly ignorant of that fact. Thus, manipulation of the process can result in slanted or stupid results from a citizen deliberation that, using the processes, would have generated excellent results, but the public and officials would be none the wiser. This makes process a potential Achilles Heel in sortition.

    My inquiry on this topic (for which I don’t know the Latin) might be expressed as: “What approach to watching the process people would be least subject to bias and manipulation?” Since dialogue and deliberation professionals have an interest in what designs and facilitations are used in citizen deliberation, they (despite their expertise) would not be ideal overseers of the process. Nor would government, since it is already infested with special interests and is largely ignorant of deliberative dynamics and potentials. One answer I’ve explored (at a theoretical and networking level) is to organize an association of previous participants in sortition-based citizen deliberations. Members of such an association could be trained in (or about) various deliberative approaches, and then panels made up of such citizens randomly selected from the association could evaluate the process used in any particularly empowered or vital citizen deliberative activity. As an added check (per your inquiry), an independently convened parallel process-evaluation panels could evaluate the original evaluative panel’s conclusions (as an appeal court does with an earlier trial and ruling). The fact that these panels would be made up of randomly selected members of a pool which itself was originally randomly selected (i.e., the participants of prior sortition-based citizen deliberations) would mitigate against bias or manipulation. At least it seems to me.

    What would YOUR solution be to the inquiry you posed to me? There’s certainly nothing sacred about my solution. But there’s certainly something truly vital about your inquiry – as long as it was offered as a true inquiry rather than merely a rhetorical question.


  13. I’m curious, Keith: Do you believe there are only two sides to an issue, and that’s why you insist that “the only way of ensuring balanced information and advocacy is to arrogate this to the hands of the competing (elite) parties to the dispute”? I ask because I don’t share that assumption.

    The National Issues Forums – the largest annual citizen deliberative activity in the U.S. – creates briefing books based on “framing an issue for deliberation”, which entails presenting key arguments and evidence for 3-5 diverse approaches to a public issue, to challenge the thinking of the deliberators. They never use two “sides” because that feeds polarizing group dynamics. And they never use more than 5 views because research suggests it is hard for most people to hold more than 5 items in their minds at the same time.

    Processes like Dynamic Facilitation go even further, leaning over backwards (or I should say forwards) to ensure that every view is both well heard and tapped for its implications (concerns, specific actions, tradeoffs, etc.), preparing the way for participants’ realization that (and how) all their original views were oversimplifications of the true complexity involved and thus triggering a collective search for deeper insights and wiser solutions.

    I’ve also recommended that deliberators in citizens juries – after being briefed, hearing “expert witnesses” or “diverse stakeholders” and undergoing some collective deliberation, be divided up into small competing teams to research their issue on the web and see who can come up with the best (by vote of their peers) new information, perspective, or solution. My intention with that recommendation is to throw novelty into any over-restrictive framings or briefings they have been given, thus undermining bias and groupthink in the information dimension of their work.

    I offer these as the kind of thoughtful work being done in the fields of dialogue and deliberation which help us transcend the kinds of oppositional dynamics that characterize this and so many other such forums (as well as most of our tragic political life).


  14. Tom,

    The concept of offering information from “both sides” Keith refers to derives from the Yes/No binary choice of whether to adopt or reject the final version of a draft law. Keith does not propose any give-and-take deliberation among the randomly selected panel members (leaving that to an elected chamber). I also advocate such a final stage, but with the deliberative seeking of the wisdom or the crowd occurring in a separate mini-public, which is where all of the new science and psychology of optimal decision making you mention, in the process of refining and drafting a final bill for the jury comes in to play.


  15. Thanks for the clarification. tbouricius. I realize this relates to (among other things) Oregon’s Citizen Initiative Review where a minipublic evaluates a proposed ballot initiative on behalf of the electorate – although that is not exactly what Keith was referring to.

    Now, although I should first apologize for complexifying the discussion, I can see more governance functions than just (A) creating a draft law and (B) accepting/rejecting it.

    Starting from conceptual scratch, for example, I can imagine dividing the law-making job into functions like these:
    (a) identifying, naming and prioritizing public issues most in need of consideration (an often overlooked but extremely influential role);
    (b) deliberating on a given issue with the intention of providing broad guidance about how to address it (including examining existing proposals from various groups/experts), i.e., the creation of an overall proposed “approach” for the polity to take on that issue;
    (c) detailing how such broad guidance would translate into specific implementable actions, i.e., an overall proposed “policy”; and
    (d) creating proposed legislation that frames those actions into law (a task of legal articulation).

    Before each of one of these four activities I can imagine there being processes allowing for input from certain entities – and after each of these activities I can imagine certain entity/ies reviewing its output to accept, reject, or alter that output. I can also imagine an oversight function accompanying each process to ensure its integrity or quality.

    If all those functions were performed by specific entities, that would constitute a dozen different functions: the four basic functions (a-d) listed in the second paragraph above [times] the three functions listed in the paragraph after that – input, review for ok/not-ok, and oversight – that could be done around each of a-d.

    Various entities that could perform those twelve functions might include (for example):
    (1) elected representatives;
    (2) experts or technocrats;
    (3) partisans and stakeholders;
    (4) minipublics (sortition-based panels); and/or
    (5) the broad public (e.g., plebiscite, prediction markets, open input forums, etc.)

    There is a certain logic – i.e., there could be certain arguments – to support mandating each of 1-5 to perform each of the preceding functions. Many different systems (variously democratic, republican, polyarchic, meritocratic, etc.) could be imagined by combining them in different ways. I wonder what a dialogue would be like that would explore why one might choose to give certain roles to certain particular entities (i.e., how and why to distribute the various tasks of policy-making). In this forum, it would be appropriate to explore what role (4) sortition could/should/should not play in each of the above functions and why and in what relationship to the other entities and functions.

    All this would, of course, be in addition to – but overlap – the systems and dynamics of conflict resolution, knowledge systems, self-organization, administration/implementation, etc., all of which are extremely relevant to creating communities and societies that are collectively intelligent and wise. Which is why I like to start from the inquiry of how to make democratic systems that are more collectively intelligent and wise, dropping down into these more detailed inquiries in whatever way will serve that larger inquiry.


  16. Tom,

    Thanks for sharing some interesting and challenging ideas. The notion of collective wisdom is an attractive one (although in practice it generally morphs into fascism of the left or right, with the Fuhrer/Duce/Supreme Leader embodying the collective wisdom of the volk). Terry has summarised my position, which I would describe as democratic pragmatism, accurately. The pragmatic element is the need to work within the existing political culture, which is (broadly speaking) agonistic. Changing the institutional arrangements is hard enough, changing the political mindset is just a bridge too far. (Those who argue that the abolition of political parties would end agonism have failed to study the historical record.) The democratic element presupposes that the will of the majority should predominate and in large polities this requires some form of representation. The descriptive representation instituted by randomly selected bodies presupposes that the decision outcome is invariant across allotted samples, otherwise there is no way of knowing which decision is the representative one. Unfortunately this imposes serious constraints on the deliberative style of the assembly, whatever the cost for collective wisdom. Variable outcomes would be the same as voting for a political party and then having no idea at all what sort of policies they would introduce (whereas if you vote Republican or Democrat you have some heuristic clues as to what the policy outcomes would be, even if they have little to do with collective wisdom). If you really believe that policy making by chance is legitimate this would suggest that the case for the religious origin of lot has been prematurely dismissed. Or do you feel that the decision of each group would be perceived as legitimate as it is a) “wise” and b) produced by ordinary people (as opposed to the political elite)?


  17. Good to hear from you, Keith. I just finished Andrew Sullivan’s article “Our democracy has never been so ripe for tyranny” and I’d be curious how you’d respond to it.

    I don’t believe that abolishing political parties would end agonism. But I do believe that in the last 30-40 years we have developed more and more resources for resolving conflicts, addressing complex problems, and using diversity creatively to generate the practical wisdom that comes from integrating diverse perspectives and gifts. The fact that these resources are not widely understood and used is more a challenge and a tragedy than a determining factor in our destiny.

    Given the uncertainties of our future, I think it is vital that diverse strategies be undertaken simultaneously. So I salute your efforts to work within the system. I am more (r)evolutionarily inclined, due both to my activist background and my sense that the current systems are totally unprepared to address the converging tsunami of global catastrophes already visible on our horizon – like climate chaos (and its dozens of dangerous domino-like consequences, e.g., human and disease/pest migrations and crop failures leading to political backlash and instability), peak oil and other resources, the concentrated destructive technological power available to both governments and small groups, and dozens of other mutually reinforcing challenges. The only hope I see (albeit a slim one) is in creating significantly new systems that systematically address the shortcomings of existing ones. My awareness of rapidly evolving approaches referred to in the second paragraph above leads me to try imagining how they could be applied in alternatives to the fundamentally agnostic systems we have in place currently which make it so hard to envision and work together towards common goods (goods which one would think would become more obvious to all as the consequences of these trends begin to impact us all).

    That said, I can imagine such combinations of existing and new approaches, such as an elite legislature (Senate, House of Lords) whose policies could be overruled by 2/3rds majorities of a randomly selected lower house that has the benefit of the knowledge and methodologies noted in the second paragraph above.

    Or a citizens jury empowered to convene a group of experts to find a solution to a public issue whose recommendations are then deliberated and revised by the original citizens jury in collaboration with the experts and randomly selected legislators.

    Or experiments (which I have advocated) which run three independent concurrent citizens juries on the same topic along with three independent stakeholder deliberations on that topic. They would be videotaped and studied for factors that led to differences and similarities in their conclusions. I further imagine mixing the membership of those six groups into six new groups each containing a few members of all six of the first groups, and then putting those new groups through a Dynamically Facilitated process, with the outcomes again compared. Ideally such an experiment would be intended to seek a reliable way of coming up with a non-reductionist expression of the informed wisdom of We the People. If the above process didn’t suffice, further research would be undertaken. The point is that I agree that citizen juries in their current form cannot defend their own legitimacy except within the sortition worldview – but then neither can elected legislatures except within the questionable republican worldview. Experiments could perhaps discover approaches that had independently defensible legitimacy.

    Finally, ultimately, the “wisdom” of any decision is (in my view) dependent on the extent to which it provides long-term broad benefits. This is not something that is either known in advance (it is proven by experience) nor is it readily measurable (although approximate measures could be devised, as is being proposed in various alternative indicators to GDP). However, there are OBVIOUS ways that decisions are made that involve very narrow and/or short-term benefits (with often disastrous broad harms) which could be progressively weeded out of our decision-making and governance systems. It seems worthwhile to explore how THAT might be done.


  18. Hi Tom,

    I read the Sullivan essay a couple of weeks ago, when it was republished in the Sunday Times, Rupert Murdoch’s flagship UK broadsheet (he was on the staff for 20 years). I’m a big fan of Andrew — we published his PhD thesis (Intimations Pursued: The Voice of Practice in the Conversation of Michael Oakeshott) — and I know him slightly. The phenomenon that he discusses is not limited to the US and is widespread in Europe (Austria is on the verge of electing a far-right president). I’ve studied this trend in the UK and it’s strongly connected with the lack of agonism in politics. During the first decade of this century all mainstream UK parties supported a policy of (effectively) unlimited immigration and multiculturalism for a variety of reasons, so ordinary citizens opposed to this policy (mostly the indigenous white working class) had nowhere to go. In 2004 we published The Great Immigration Scandal and it was dismissed at the time by the media/political class (who genuinely believed their enlightened views represented the common good) as a racist and xenophobic rant, but the views contained in the book are now part of the new post-UKIP consensus. If political parties had been doing what they were designed to do (reflect the views of citizens, rather than the Washington/Westminster elite), then there wouldn’t be a need for the Tea Party, Trump, Sanders, UKIP, the Freedom Party etc. All these groups indicate a popular backlash against the political and cultural elite.

    Despite my scepticism that any political perspective could, or even should, be “obvious to all”, I applaud your efforts to ensure the representative legitimacy of citizen juries. But the videotaping, non-reductionism and dynamic facilitation is far too complicated and cerebral and is still subject to Juvenal’s quis custodiet? My own preference is to run a number of large parallel juries and only accept the unanimous verdict of them all (on a contentious issue). Terry and myself agree that the final verdict should be an up/down one and based on exogenous advocacy as opposed to an attempt to tap into emergent collective wisdom. If all juries agree, then it’s hard to see how anyone not directly involved in the process could object to the outcome, as it would have made no difference whether or not they were included in the sample. For better or worse we live in a democracy, not an epistocracy — your proposals are really just a demotic variant of Platonism.

    >a citizens jury empowered to convene a group of experts to find a solution to a public issue.

    I’m strongly opposed to this approach for the reasons given in my response to Yoram’s comments (ignored, as per usual).


  19. Hi Keith,

    You may not realize it, but I am totally in favor of agonism, if by that you mean the presence of and interaction among competing ideas. I see this as vital for both the health and evolution of human systems. But such diversity and conflict can be used more or less destructively or creatively, and is seldom appropriately reducible to only two perspectives. I also don’t believe that the outcome of such interaction must be the victory of one “side” over the other; it is quite possible and usually desirable for some perspective or solution to emerge out of the interaction that was not present before it began. This can range from negotiated compromise to deeply shared insight regarding some totally new angle or possibility.

    Furthermore, I agree that excluding people who will be affected by a decision from shaping that decision (whether intentionally or through privileged neglectfulness) is a recipe for backlash against the final decision and/or those who pushed it through. I, too, see this in the backlash against “unlimited immigration and multiculturalism”. But the principle applies MUCH more broadly, and includes those “foreigners” pressing to be included. (I find it interesting that Austria ended up electing the Green – their equivalent of Bernie Sanders – rather than their far-right candidate. The closeness of the vote, however, is a red flag for all of us seeking a healthier body politic.)

    I can understand why you (or anyone else) would prefer a simple up/down vote on a policy option, rather than “attempting to tap into emergent collective wisdom” – especially if you have not witnessed any process that dependably calls forth anything resembling “collective wisdom”. But I have and I believe it is possible that several independent juries supported in generating collective wisdom on an issue COULD “all agree” on how to address that issue, at which point it would be (as you say) “hard to see how anyone not directly involved in the process could object to the outcome”. The fact that I have indicators that that would be possible is, of course, totally insufficient to convince you, since I don’t have any real proof, just extrapolations leading to hypotheses. So it is incumbent upon me (and others so inclined) to do the research to demonstrate the possibility that we believe is present. Whether such research will ever be done (it would require considerable resources to do it well) remains to be seen. But I reiterate my assertion that we DO urgently need real wisdom to address our 21st century challenges, especially the growing number of “extinction level issues”. In the absence of such wisdom, I have trouble seeing any improvements to our business as usual democracy that would save civilization from its self destruction. So I think the research noted above would be highly desirable.


    PS: I’m not sure what you mean that my “proposals are really just a demotic variant of Platonism.” Can you clarify?


  20. Tom,

    >But the principle [all affected interests] applies MUCH more broadly, and includes those “foreigners” pressing to be included.

    Unfortunately you then end up with a demos that includes the whole world and this is not practical. As a result there is no serious alternative to traditional geographical models of citizenship, however unattractive that may be for cosmopolitans.

    >I believe it is possible that several independent juries supported in generating collective wisdom on an issue COULD “all agree” . . . So it is incumbent upon me (and others so inclined) to do the research to demonstrate the possibility that we believe is present.

    Using the Deliberative Polling model this would be quite feasible (and not hugely expensive). In fact it’s already been done (with the Texas utility DPs) and found to be wanting, in that all three DPs came to different conclusions. The cause, IMO, was the small-group deliberations (the source of “emergent wisdom”), so further experiments need to be done to see what changes are necessary to ensure consistent outcomes. Without consistency there is no way of knowing which decision is deemed to represent the target population.

    >I’m not sure what you mean that my “proposals are really just a demotic variant of Platonism.”

    Epistemic platonism is the view that there are “right” answers in a knowledge domain. The notion of emergent wisdom (as opposed to considered preferences) indicates an epistemic approach to political problems. Whereas Plato argued for the judgment of philosophers, you are content to leave it in the hands of ordinary citizens, hence the “demotic” qualifier.


  21. Keith,

    First I want to thank you and the whole Equality by Lot blog for patiently providing space for considering what I see as very important points. I am learning a lot even as I share my own views and information.

    Now to your points:

    When I spoke of “foreigners pressing to be included”, I did not mean they were pressing to (or should) be included in citizen deliberations. I meant that immigrants were pressing to be included in European societies. I was suggesting, however, that failure to include their views and interests in decision-making about immigration would create problems just as surely as failure to include the views and interests of immigration-resistant citizens would create problems.

    Regarding boundaries about who is included in citizen deliberations, I agree that national boundaries are important touchstones. But I wish to point out that many issues – including immigration – are international or global in scope, and thus that the perspectives and needs of non-nationals are often highly relevant to issue deliberations, whether or not they are represented by non-national deliberators. Failure to do so invites failures in policy.

    Several months ago I was involved with evaluating a 2015 Civic Council of randomly selected citizens in Vorarlberg, Austria dealing with asylum and refugee issues The panel happened to include a recent Middle Eastern immigrant who ended up being a vital contributor to the outcome due to his experience as a refugee. Descriptions of his presence and role are not readily available on the web, but the evaluation team noted it and considered how consciously including such “impacted people” could improve future deliberations.

    A related approach to this is “stakeholder dialogues” in which stakeholders are defined as people who will be significantly impacted by a decision or have strong opinions on the issue, as well as those with special knowledge or power related to decision-making and implementation re the issue. In the case of immigration, such a dialogue would include spokespeople for those strongly opposed to immigration as well as spokespeople for migrants and refugees (among many other interests). When agreement can be gotten among such “agonists”, it can have a powerful impact. The Consensus Council in North Dakota is a notable example of such an approach, as is Future Search

    Given that a public issue by definition involves conflict, I’ve often recommended that BOTH citizen deliberative councils AND stakeholder dialogues be convened. My research proposal involves 3 parallel, comparable and independent citizen deliberative councils AND 3 parallel parallel, comparable and independent stakeholder dialogues all convened to address the same issue. I think there are powerful arguments for both approaches and I’m not sure why we would do one without the other, especially if we are seeking collective wisdom on an issue.

    I had not heard of the Texas Utility Deliberative Polls. I looked for them on the web and found a reference to a series of eight such polls done over two years (1996-98) – but not to a series of three. Can you give me a link to your series that makes the point you make? The 8 DPs done in ’96-98 were done in different cities by different utilities (an approach similar to Planning Cells in Germany, and thus were not truly comparable in the sense of my research proposal (which would require that the random selections be done in the same manner from the same larger population). Nevertheless, I find it fascinating that, although there were regional differences among the Texas Utility Deliberative Polls, there were sufficient similarities that investigators surmised that “it might have been possible to produce the same basic results … with fewer events.” This makes me further curious about your DP example. I also want to note that the deliberations in Deliberative Polls are briefer and less rigorous than, say, deliberations in Citizen Juries, and are also not designed to generate a collective statement of recommendations. That latter mandate strongly shapes what happens in citizen deliberative councils like Citizen Juries, and supports convergence.

    Another data point: Randomly selected “Civic Councils” (like the Vorarlberg example above) are dynamically facilitated. They are modeled after “Wisdom Councils” designed by Jim Rough, the founder of Dynamic Facilitation. While teaching DF for many years to corporate and government clients, Rough had his students practice facilitating public issues that mattered to them (since they needed real issues to work with and the students all came from different organizations). Over the years he noticed that classes that chose (for example) homelessness came up with similar recommendations at the end, despite containing very different students. This led him to create the Wisdom Council as a democratic application of DF. This is one more tidbit that leads me to believe there may be more capacity to generate congruent results than is currently believed – and that the power of DF to transform conflict into breakthroughs may be one of the most useful tools in pursuit of that end. Thus my proposed research deliberations would mix-and-match participants in the 6 first-stage panels (3 citizen + 3 stakeholder) into six second-stage panels each containing several members of the 6 first-stage panels, and to have them independently dynamically facilitated to see if there is a convergence in their separate recommendations.

    Finally, I do not believe there is “one right answer” to any issue. However, I do believe that as we come to understand more of what is involved in an issue (different perspectives, systemic dynamics, deep needs, etc.) there is a tendency for false solutions to fall away and for approaches that address the issue at deeper levels to become increasingly clear. I believe there are possibilities for wise convergence in that fact and would love to rigorously test that belief. As Oren Lyons, Faithkeeper of the Turtle Clan of the Onondaga Iroquois says, “We keep talking until there is nothing left but the obvious truth.” So I wonder: How do we apply that potential to modern societies?


  22. Tom,

    Many thanks for your thoughtful comments. In her 2013 book Democratic Reason: Politics, Collective Intelligence and the Rule of the Many, Helene Landemore outlines two distinct traditions of collective intelligence, the talkers and the counters. The first tradition (the talkers), which she traces back to Aristotle’s pot-luck meal analogy, views collective wisdom as the emergent property of the exchange of reasons in small deliberative groups. Key theorists in the tradition are J.S. Mill, John Dewey, and modern epistemic and deliberative democrats. As the key focus is on reasonableness (the forceless force of the best argument), writers in this tradition have scant interest in accurate representativity of the group vis a vis the population at large. Yves Sintomer argues that the notion of deliberation used by this group derives from the German deliberativestimme (deliberative voice).

    The second tradition (the counters) is concerned with the aggregation of independent epistemic perspectives, rather than the exchange of reasons. Although having its origins in Plato’s Protagoras and practical implementation in the large randomly-selected legislative juries of 4th century Athens, it’s theoretical justification had to wait the birth of modern mathematical theory in Spinoza, Condorcet and the law of large numbers. The best modern outline of the tradition is Surowiecki’s Wisdom of Crowds, and Sintomer claims this perspective on deliberation is derived from the Latin liber (weight) in which large groups of de-liber-ators silently weigh the arguments of competing parties and then determine the outcome via majority vote. Fishkin’s work in deliberative polling is a good example of this tradition. This tradition is concerned with majority rule and it’s a zero sum game. Whilst the inclusion of perspectives from refugees and other foreigners should be part of the information/advocacy stage, nevertheless the outcome is determined by the members of the relevant political community (or their proxies in the allotted microcosm). The approach is inimical to the very notion of a Wisdom Council, and facilitation (dynamic or otherwise) is ruled out via the quis custodiet? principle.

    Your work on consensual outcomes and emergent collective wisdom is clearly in the first tradition. My concern, by contrast, is that the decision (which may or may not be a wise one) should represent the considered view of the target population, so it is a modest improvement on business as usual (rationally-ignorant decision making by mass electoral democracy). I’m sure your approach is more suited to your apocalyptic view of the problems of humankind, my concern is just to muddle along a little better. But I think my simple reforms would be more likely to be perceived as democratically legitimate by the vast mass of people who would not be included in the emergence of collective wisdom. For better or worse this is what we call “democracy”, even if it would not pass muster with the Faithkeeper of the Iroquois Turtle Clan.

    Thanks for pushing me on the Texas DPs. My original source was an article by Robert Goodin, but I can’t find it at the moment. I thought it was the Texas DPs that I discussed in a private email exchange with Jim Fishkin, but I’ve looked it up and it looks like I’ve done a Bush (i.e. misremembering) as it was the article by Farrar et al on the New Haven DP that we were discussing. I have to finish the chapter of my PhD on epistemic democracy now, but will search out the Goodin ref. as soon as I’ve finished.

    in haste



  23. Tom,

    Found it — from Bob Goodin’s 2000 essay “Democratic Deliberation Within”, reprinted in Fishkin and Laslett Debating Deliberative Democracy (2003):

    “On the face of it, [ongoing descriptive representativity of the minipopulus] seems unlikely. From everyday life we know that different conversations with different participants (or with the same participants interjecting at different points) proceed in radically different directions. Given the path dependency of conversational dynamics, and the sheer creativity of conversing agents, it beggars belief that any one group would come to exactly the same conclusions by exactly the same route as any other. (Lawyers say it is a ‘well-known secret’ that ‘no two juries and no two judges are alike.) Yet that is what strong advocates of ersatz deliberation must be claiming to be at least approximately true, in insisting that deliberation within a representative subset will genuinely mirror, and can therefore substitute for, deliberations across the whole community.” (pp. 58-59)

    [fn.28] “Cf. Fishkin’s Voice of the People, p. 220 report of deliberative polls done for three different local public utilities in Texas. There he is pleased to report that in all three cases the shift in public opinion, pre- to post-deliberation, was in the same direction. But the absolute numbers nonetheless diverged wildly. In one case, half the respondents thought post-deliberation that ‘investing in conservation’ was the ‘option to pursue first’, whereas in another case less than a sixth thought so. In one case, over a third still thought post-deliberation that ‘renewable energy’ should be the top option, whereas in another case less than a sixth thought so. Clearly, these deliberating groups ought not to be regarded as interchangeable. Neither, in consequence, does this evidence inspire confidence in the general theory of ‘ersatz deliberation’, treating smaller deliberative groups as microcosms capable of literally ‘substituting’ for deliberation across the whole community.” (p. 74)

    Those of us who (like me) privilege representativity over collective wisdom find this deeply disturbing and would seek to isolate (experimentally) and exclude those aspects of deliberation that lead to inconsistency between the samples. My hunch (it’s no more than that) is that this means excluding all intra-group deliberation, leaving jury members with little more to do than ask questions and then record their anonymous votes at the end of the exchange of reasons between advocates selected endogenously.


  24. Keith,

    I am so appreciative of your taking the time to respond so substantively. I am learning a lot and clarifying my own thinking at the same time. Some thoughts:

    1. Re Landemore’s two deliberative traditions (which have fascinating inexact parallels with Jane Mansbridge’s adversarial and unitary democracy concepts in BEYOND ADVERSARY DEMOCRACY)

    I would agree that I am in the first tradition, albeit with two qualifications: (a) The chances are high that no participant’s “best argument” will produce the final wisdom coming out of such a conversation. If it is truly “collective” and “emergent”, the wisdom will arise newly from the larger picture painted – and creative energies evoked – by the interaction among the parties involved. Its wise insight derives from it being born from a broader – collectively generated – perspective than any individual participant arrived with. AND (b) the fact that prior “writers in this tradition have scant interest in accurate representativity of the group vis a vis the population at large” is no reason to associate that disinterest with current theorists and practitioners, witness the obsession with representativeness by all the mini-public theorists involved with this sortition blog, notably including myself.

    Re the second tradition, I’ve always thought Surowiecki’s claims to the word “wisdom” are excessive, given that his prime example is prediction markets – guesses about the future – rather than the formulation of strategies and policies that will produce broad long-term benefits (a far more complex proposition). And, as you note, majority rule tends – when it is working satisfactorily – to generate gradual changes – since these arise from moderating extremes (and, in the case of legislatures, compromises and deals) rather than deep collective insight into the issue under consideration. Majoritarianism is, above all, *workable* (up to a point, depending on the agreeability of the losers and the match between the decision and the realities it was designed to address, i.e., the wisdom factor).

    Your raising the quis custodiet concern re “who will oversee the facilitators in an interactive deliberation?” has a corollary set of concerns in the “voting on the best option” forms of deliberation, i.e., “Who chooses which alternatives are presented to the deciders and who watches over them?” and “Who oversees the integrity of the voting process?” And of both approaches we can ask: “Who oversees the providers of the information used to brief the deliberators?” and “Who oversees the articulators of the questions that frame the debate in the first place?” I suggest to you that such questions should not serve to invalidate either approach, but are best used to design good oversight into whatever deliberative processes are undertaken.

    Finally, and I believe this to be my most important point here: You say “I think my simple reforms would be more likely to be perceived as democratically legitimate by the vast mass of people who would not be included in the emergence of collective wisdom.” I would fully agree with that, at least as you state it. However, I suggest that the vast mass of people CAN be included in “the emergence of collective wisdom”. In fact, that is a current focus of my own inquiry and advocacy.

    For example, here are just four of the approaches I see:

    * Millions of people vicariously experiencing a transformative deliberation among members of a small but legitimate mini-public. The best case I know of this was achieved by Canada’s Maclean’s magazine and Canadian TV in 1991, a case described in detail and analyzed on my website at Maclean’s chose 12 Canadians who represented many demographic factors and had Roger Fisher and 2 Harvard colleagues work with them for 2.5 days, producing (with some considerable dramatic conflict and emotionalism) a consensus statement. Maclean’s 40 pages of coverage (all posted on my website) notably included bios of the 12 panelists and a blow-by-blow account of their evolving interactions – and was accompanied by a 1 hour TV documentary. The combination generated massive dialogue across Canada (a fact I discovered from interviewing the editors).

    * Combining sortition-based face-to-face deliberations with mass participation in online forums and inputs, with well publicized results and further dialogues. One innovative initiative along these lines was the Citizen Parliament in Australia, about which a whole book was written

    * Minipublic deliberators taking input from the public, soliciting both written material and participation in public hearings. This was done by the British Columbia Citizens Assembly on Electoral Reform.

    * Combine both of Landemore’s traditions – the talkers and the counters by having voters vote on the findings and/or recommendations of a minipublic deliberation. This was done in the BC Citizens Assembly and is currently used in Oregon’s Citizen Initiative Review process (a form of citizens jury investigating ballot initiatives on behalf of voters)

    2. Re Fishkin and the dependability (or replicability) of mini-public deliberative judgments

    I own a hardcover copy of Fishkin’s THE VOICE OF THE PEOPLE (1995), which was published before the Texas public utility DPs were held (1996-98). Yale published a paperback reprint in 1997 “expanded to include a new Afterword” which probably contained Goodin’s footnote reference. I suspect the three DPs Goodin mentions were held right before Fishkin wrote his 1997 Afterward, and then 5 more were held subsequently.

    But the important fact here is this: The eight DPs described in the Texas public utility report I referred to in my previous entry were held by eight DIFFERENT public utilities. The Goodin essay footnote about Fishkin describes “deliberative polls done for three DIFFERENT local public utilities in Texas”. So I want to stress here that BOTH these references describe DPs that were not truly comparable: They involved people selected from different populations served by different utilities (as well as other differences). Thus they actually offer NO basis for saying that two or more COMPARABLE mini-public deliberations on the same topic would “of course” produce significantly different results. We have no such examples to compare.

    So, while I agree with Goodin that “Clearly, these deliberating groups ought not to be regarded as interchangeable”, I must point out the overreach of his subsequent conclusion that “this evidence [does not] inspire confidence in the general theory of ‘ersatz deliberation’, treating smaller deliberative groups as microcosms capable of literally ‘substituting’ for deliberation across the whole community.”

    While I can argue that certain forms of microcosm deliberation may be replicable, and you and Goodin can argue the contrary (as he does powerfully in your reference in the preceding blog comment), the fact remains that adequate experimentation has not been done to decisively support either of our claims. I suggest that scientific integrity requires us all to admit that we simply do not know. Furthermore, I must sadly agree that it is up to those of us seeking a truly legitimate and replicable “wisdom of the whole” to provide the missing evidence that our goal is in fact achievable.


  25. Tom,

    It’s great (and highly unusual — especially on this forum) when two people with very different perspectives are prepared to actually listen to each other!

    Agree that Surowiecki’s title is misleading, and also odd that the one exception to his rule is political decision making (I think this is probably on account of the conflation of the talking and counting traditions in Fishkin’s work). I believe that Mansbridge was one of Landemore’s thesis advisers at Harvard.

    I agree that the quis custodiet? problem in agenda setting is deeply problematic and hard to resolve in democratic terms (I’m not too keen on overseers!) — the longest chapter in my thesis draft is grappling with the problem of representative isegoria (without overseers) — I’ll send it to you if you want to contact me at

    Regarding your four examples combining talking with counting, I think participatory democrats are wrong to focus on absolute numbers (millions, 12 or whatever). My primary concern is always representation, i.e. reflecting the views of the vast majority who don’t choose to read Maclean’s magazine or participate in online forums and public hearings. Those who choose to participate are generally activists, busybodies, axe-grinders and political anoraks, my concern is always how best to represent the views of the silent majority.

    Thanks for the clarification on the Texas DPs, I agree that this doesn’t constitute a valid test case. Fishkin’s institute would be ideally constituted to conduct such an experiment, unfortunately he doesn’t seem to be much interested in extending his model to assemblies that have some sort of legislative function. Judging from my email exchange with him he’s so committed to combining the counting and talking traditions that I wonder if he would condone research that undermined the representativity of the former? The well-established research on information and status cascades that Goodin referred to would tend to rule out talk-based solutions a priori, so he might well view the experiment as shooting himself in the foot. But I agree we just don’t know and that the experiments should be done.

    >witness the obsession with representativeness by all the mini-public theorists involved with this sortition blog.

    I only wish that were the case!

    best wishes



  26. Keith and Tom,

    I haven’t had time to participate in this discussion, by I just wanted to add my appreciation…. it helps to shape and clarify my own views from reading such discussions. One quote I think is valuable is this one about how we establish our OWN PREFERENCES from a participant in the Australian Citizens Parliament
    “you’ve kind of just got your own, because you don’t really get a chance to question your own points of view, it’s only when you’re confronted with another point of view that you can even deliberate [in your own mind].”
    [Andrea Felicetti, John Gastil, Janette Hartz-Karp, and Lyn Carson (2012) “Collective Identity and Voice at the Australian Citizens’ Parliament,” Journal of Public Deliberation: Vol. 8: No. 1, Article 5. page 18.


  27. Hi Keith,

    I’ve taken you up on requesting your draft chapter. I’m very interested in the topics you are working on (representation and oversight).

    Here are some other thoughts and questions:


    I’m curious how you define “the silent majority”. My experience with randomly selected wisdom councils and citizen juries in the US and elsewhere is that many ordinary people who have not previously been politically involved agree to participate if they are chosen and, in fact, very few “activists, busybodies, axe-grinders and political anoraks” (great word, “anorak”! I’ve never heard it before) end up on such panels. Of course (in my experience), where online forums and public hearings – and even professionally convened “citizen participation” initiatives – are “open to anyone”, the chances are high that many if not most of those showing up will fit your categories of pushy political agents. That is one of the reasons I prefer random selection.

    While I understand your concerns about Maclean’s, I also find them a bit odd. Maclean’s has a readership of about 300,000 Canadians in a country a hundred times that size. That makes it comparable to Time magazine whose circulation of 3 million represents 1% of the population of the U.S. Both are the leading glossy newsweeklies for their respective countries (twice the size of their nearest competitors). And the Canadian CTV news program W5 (which covered the Maclean’s deliberation) is the longest-running newsmagazine program in North America and the most-watched program of its type in Canada, analogous to 60 Minutes in the US (America’s most watched news program).

    So although many people don’t read Time or Maclean’s or watch W5 or 60 Minutes, very very many do. The impact of these media is considerable, reaching far beyond their immediate viewers/readers (especially when, as in the 1991 case, BOTH media forms are synchronized). I think it is very valid to point out their potential for engaging the broad public in vicarious deliberation experiences. I should especially add that the interviews I arranged with Maclean’s editors disclosed that much of the dialogue their project stimulated took place on talk radio – listened to by many people who don’t read Maclean’s – which engaged public officials who resisted the challenge and poured oil on the bubbling waters of national conversation. (As I say on my site, if only Maclean’s had done the same exercise annually after that first proof of concept….!) Since the Maclean’s panelists were scientifically selected to represent the full diversity of Canada, and since the conversations stimulated went far beyond the direct audiences of Maclean’s and W5, it seems unfair to simply dismiss the effort because many people don’t read Maclean’s. It is a complex picture in which the readership of Maclean’s is one of many factors. (Similar arguments can be offered to invalidate “counter” approaches – in Landemore’s framework – including but not limited to elections where only a portion of eligible voters participate – raising the issue of whether the results are “really” majoritarian? While the question is a valid concern, I think it is unfair to use it to dismiss elections totally.)

    Anyway, back to who “the silent majority” is. As I meditated on that concept this morning, it seemed to me far too simplified a concept. What came to me was a kind of spectrum of political roles arranged by how visible and assertive they tend to be in public forums. Certain people fit stably within one or two roles; most people take different roles depending on the issue and engagement circumstance, and therefore show up differently in public forums at different times. Here’s what I came up with (a draft, of course):

    1. Interested party with significant emotional/material investment in the issue being considered
    2. Distorted ego “with too much time on their hands” (including busybodies, axe-grinders, and political anoraks)
    3. Active partisan (tirelessly advocating one ideological side of many issues)
    4. Generic activist (supporting various causes they believe in, but not necessarily explicitly ideological in their stance)
    5. Concerned citizen (taking citizenship seriously: voting, being informed, studying issues and candidates, discussing politics with friends, etc.)
    6. Aware citizen but too busy/distracted to do much (a political spectator who feels somewhat apathetic or disempowered; although occasionally drawn in by an issue that especially matters to them, they’re not sure what to do about it)
    7. Backseat driver/critic (very opinionated talking with friends; natters and complains but takes very little real action)
    8. Idealist/change agent/volunteer who is active outside of major political venues (education, community, business consulting, environment, etc., but nothing explicitly political)
    9. Disengaged/disenchanted/disgusted/cynical citizen; “been there, done that”; forget it!!
    10. Person who thinks politics is something that happens elsewhere or with other people (they know it is there but feel no personal connection to it)
    11. Person who remains quite unaware of the political dimension of life; pays no attention at all to public affairs
    12. Person who is throughly oblivious (of anything beyond their own immediate life situation)

    In my experience, random selection tends to draw people from categories 3-8, with an interesting few from 9. I would assume, from what you’ve said so far, that you would define “the silent majority” as folks in categories 5-12, since most people who show up for most “public engagements” are from categories 1-4. Is that true? Do you have other thoughts about this list (not to distract you from your PhD work!!)?


    I totally agree with you (perhaps for different reasons, given your orientation to the “counters” mode and mine to the “talkers” mode) that representation is a central concern. Most modern random selections are stratified according to various obvious demographics (race/ethnicity, gender, politics, location, etc., as well as often demographics relevant to the issue being discussed). (And of course there is the self-selection bias present wherever participation is not mandatory… which it seldom is… which makes both the targeted population(s) and the invitation particularly important to consider well.)

    There are, of course, vast differences in people that are not reflected in most demographic sorts (e.g., which “true” random selection (whatever that is) would presumably include a smattering of.

    My own perspective is that the more diversity that’s present, the more potential wisdom could be generated IF (and it’s a big IF) the process used is able to utilize that diversity creatively (rather than destructively, dysfunctionally, or neglectfully). Unfortunately for me as a deliberative democracy “talker”, that condition is best met with a limited number of people. When we get into hundreds of people, it is hard to have high quality deliberation. But the fewer people we have, the greater the loss of diversity. This is where I think in terms of 1-3 dozen people and, if there is a larger number, thinking how that larger group can usefully be broken into smaller groups that are then integrated in some way back into the larger group (something often done in diverse ways in many processes). Online processes can easily handle hundreds or thousands of people, although the greater the number of people the more it makes sense to switch more toward “counter” (rather than “talker”) approaches. Unless, of course, we do the kind of combined methods I described in my previous comment post.

    But back to you. What kind of “representation” are you looking for? Who is included and how?


    Your statement that “The well-established research on information and status cascades that Goodin referred to would tend to rule out talk-based solutions a priori”. I’m not familiar with that research, but I did a quick check and I THINK these refer to various dynamics that lead to conformity, herd behavior, or conservative (keep the status quo) behavior. (Please correct me if I’m wrong.) If this is true, I see all these dynamics as forms of groupthink – the tendency of a group to cohere around a non-optimum solution or worldview, rather than being grounded in a reasonable and dynamic relationship with reality. This often results from psychosocial dynamics suppressing dissenting views within themselves as individuals or within their group collectively, instead of processing them in some way that contributes to the group’s collective intelligence.

    I can see where someone would advocate against “talk-based solutions” if their primary experience (or knowledge) is of group dynamics that generate groupthink (a.k.a., co-stupidity). My own experience has been that conversation is like leadership or religion, in the sense that some forms are liberating and wisdom-generating, while other forms are oppressive and stupidity-generating. The lack of the former conversational experiences is tragic but quite correctible by wise design which, again, is a central concern of my own studies.

    Again, I can’t recommend exploration of Dynamic Facilitation highly enough. I am an eclectic student of process familiar with dozens of conversational modes and I have been awed by the power of DF to transform conflicted diversity into shared insights and possibilities, using a very evocative, non-directive approach. In fact, I see it as an extremely powerful antidote to groupthink in all its forms.

    It is in this process factor that I see hope for “emergent collective wisdom” as long as adequate diversity of people, perspectives, information, etc., is also taken into account.

    That’s it for now – again with tremendous gratitude for the spirit and quality of your engagement, which has been such a benefit to my work and thinking.



  28. Terry,

    I agree we should all be grateful to Tom for the huge amount of work that he has put in to these commentaries. On the issue of preference formation I think it’s fair to say that the comment you report would apply to both models (face to face deliberation and adjudication by jury) as the final judgment is always a mental one. E.M. Forster’s quip — “how can I know what I think until I hear what I say?” — was a product of the era of behaviourist psychology.


    On the issue of Macleans and Time magazine, it’s an interesting coincidence that we are dealing with 1% — exactly the same figure that is used by critics of “elective oligarchy”. Although this is unlikely to be the same socio-economic elite, at least in the electoral example the 99% can choose which oligarch to represent them — not so in the case of proxy or ersatz representation. People who choose not to vote can only blame themselves if the wrong representative is chosen — not so in the case of representation by proxy. That’s why the latter has to be (effectively) mandatory for those whose lot is drawn as this will ensure that the microcosm accurately reflects the distribution of interests and beliefs in the target population.

    >Since the Maclean’s panelists were scientifically selected to represent the full diversity of Canada.

    Did you not say earlier that there were only 12 of them? In which case your claim would not make sense to a statistician or opinion pollster.

    >much of the dialogue their project stimulated took place on talk radio

    That’s what I feared — the sort of people who phone in to talk radio are (by definition) not members of the silent majority.

    I think your list of 12 varieties of citizen engagement is spot on — the silent majority are most likely to be drawn from categories 6 through 12. Such persons would be less likely to take up the opportunity to participate, so the people they “describe” would be under-represented. No doubt some people would write them off as idiots, but idiotes in the Athenian democracy was not a pejorative term, it was just an Ordinary Joe.

    >But back to you. What kind of “representation” are you looking for? Who is included and how?

    Two distinct mechanisms are involved — isonomia (equal votes) and isegoria (equal speech). Only the former is open to the sort of statistical representation offered by sortition. The latter requires a variety of mechanisms, outlined in the thesis chapter I’ve just sent you. But my ultimate concern is pragmatic — that the decision of the group should be accepted by everyone who doesn’t participate themselves, and I think this is a serious problem that is largely ignored by deliberative democrats.

    Informational/reputational cascades are an aspect of group-think, but the main point is that they introduce unrepresentative randomness — the agenda is set by the order in which people happen to speak and the illocutionary force of their speech acts.


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 )

Google+ photo

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

Connecting to %s

%d bloggers like this: