What Sortition Can and Cannot Do

There is considerable disagreement regarding the political potential of sortition. Dowlen (2008) argues that sortition is not primarily a system of representation as its invention in classical time predates the discovery of probability. Fishkin (2009) has long advocated sortition as a method of deliberative polling but has not (to date) suggested that it should be incorporated permanently into the system of governance. A host of writers (including the present author) have argued that sortition should either replace or supplement the institutions of electoral democracy as part of a mixed constitutional settlement. At the opposite end of the spectrum to Fishkin a small number of brave souls (mostly active on this blog) have argued for the wholesale replacement of electoral democracy by sortition. In this post I argue that, for purely conceptual reasons, the role of sortition can only be one, albeit an essential, part of a mixed constitution and that the attempt to extend its usage beyond this role undermines any claims that it may have to be a democratic mechanism. My case is based on some partially developed arguments in Pitkin (1967).

In her book The Concept of Representation, Hannah Pitkin argues that there are a variety of aspects to representation – aesthetic, symbolic, formalistic, descriptive and active – the latter two being the most relevant to political representation. Descriptive representation involves “standing for” and requires a degree of identity between the representative and her constituency, as evidenced by contemporary demands for all-women candidate shortlists and positive discrimination on account of ethnic minorities. Random selection is the best way of achieving descriptive representation, hence James Fishkin’s choice of this method for his Deliberative Polling programme. On the other hand, Active representation requires the representative (in a similar manner to a trustee or advocate) to act in the interests of her constituents; there is no intrinsic need for the representative to in any way mirror their identity, thereby justifying electoral representation in single-member constituencies. According to Pitkin, descriptive representation does not cover what the representatives do, while active representation is indifferent to who does it.

A similar distinction, drawing on Skinner (2005), is made by the republican theorist Philip Pettit:

There are two fundamentally contrasting varieties of representation, indicative and responsive. Indicative [descriptive] representers stand for the representees in the sense of typifying or epitomizing them . . .  Responsive representers act for or speak for the representees, playing the part of an agent in relation to a principal; how they act is responsive to how the representees would want them to act. (Pettit, 2010, p.65)

The argument between the Federalist and Anti-Federalist factions at the Constitutional Convention of 1787 can also be read as a dispute between these two conflicting concepts of representation. The Anti-Federalists argued that the legislature should represent all ‘classes’ (occupations) ‘descriptively’ – ‘the farmer, merchant, mecanick and other various orders of people, ought to be represented according to their respective weight and numbers’ (Dry, p.125). The Madisonian faction won the ratification debate and descriptive representation was thrown into the dustbin of history (until now, that is).

If Pitkin’s argument is correct, and political representation requires both descriptive and active aspects (although her analysis privileges the latter), how is it possible to realize these two different principles in the design of political institutions? Fishkin’s experiments in Deliberative Polling have indicated that a randomly-selected microcosm of ordinary people can decide the outcome of a debate on political issues when assisted by balanced teams of expert advocates. The quality of the decision making would appear to be no worse than decisions taken by elected representatives, few of whom even attend parliamentary debates and only generally vote along partisan lines. Thus the Deliberative Poll design would appear to be a paradigm example of informed decision-making by a descriptively representative body.

A Deliberative Polling assembly would normally consist of 2-300 members but if an allotted chamber (AC) designed along similar lines were given statutory (as opposed to advisory) powers it may well be felt that the number should be even larger, as legislative chambers typically involve 5-600 members. And this is where the problem starts – the vote of an AC is democratic (descriptively representative) in aggregate only, whereas the maximum effective size for group deliberation is between 12 and 24 (Coote and Lenaghan, 1997). Thus any small-group deliberations fail in terms of democratic criteria, hence Fishkin’s reliance on carefully-trained monitors. If, however the AC had statutory powers the independence of the monitors would require endless scrutiny, on the principle of quis custodiet?

The problem becomes even worse if the AC were also responsible for providing advocacy for and against legislative bills. Although Pitkin only deals briefly with sortive bodies in her book, the following sentence outlines the problem:

If the contemplated action is voting, then presumably (but not obviously) it means that the representative must vote as a majority of his constituents would. But any activities other than voting are less easy to deal with. Is he really literally to deliberate as if he were several hundred thousand people? To bargain that way? To speak that way? And if not that way, then how? (Pitkin, 1967, p.144-5)

What Pitkin is saying here is that voting on binary issues (yea or nay) is a simple aggregative process – it would be possible for example for a delegate representative (as at a trade union conference) to simply embody the votes of her members. But as soon as the representative rises to her feet to speak, she can only give a particular view and immediately ceases to represent her constituents descriptively as only those constituents sharing that particular view would be represented. Anything other than aggregate behaviour by a descriptively-representative chamber is in breach of its democratic mandate.

The problem becomes even worse when one considers the generation of legislative proposals by an AC (or an allotted forum, in Martin Davis’s parlance). Although comparisons have been made with the Athenian Council of 500, proposals made by the Boule were subject to ratification by the entire citizenry at the assembly (ecclesia), before undergoing further deliberative scrutiny in randomly-selected nomothetai (legislative courts). The proposals contained in Sutherland (2008) are an attempt to provide a modern-day instantiation of these principles, in which ratification by the ecclesia is replaced by an election under universal suffrage (either for parties or for ranked legislative proposals).

Needless to say the appointment of individuals to executive or judicial office by lot fails to achieve democratic legitimacy for the same reason – the logical impossibility of descriptive representation by a single individual. The principal argument for selection of government or law officers by lot is the Aristotelian principle of rotation (ruling and being ruled in turn), but such a principle is only applicable to extremely small-scale societies. It is hard to envisage a democratic mechanism for the appointment of government officials in complex large-scale polities (although appointed officials could be easily held to account by a democratic chamber, as in Rousseau’s constitutional proposal).

In conclusion, in order to retain democratic legitimacy, a randomly-selected assembly (AC) has to be a minimum size (several hundred), and can only legitimately act in aggregate. As soon as the assembly is sub-divided, or as soon as any member assumes an active function it loses its democratic mandate.  As such, sortition can only be one element in a mixed constitution, alongside electoral representation (for advocacy and policy initiation) as well as other non-democratic mechanisms (to ensure executive competence and expert advocacy).


Coote, A., and Lenaghan, J. (1997), Citizens Juries: Theory into Pratice (London: IPPR).

Dowlen, O. (2008), The Political Potential of Sortition: A Study of the Random Selection of Citizens for Public Office, Exeter: Imprint Academic.

Dry, M. (1985), The Anti-Federalist, ed. Herbert J. Storing, selected by Murray Dry (Chicago).

Fishkin, J.A. (2009), When the People Speak: Deliberative Democracy & Public Consultation (Oxford: Oxford University Press).

Pettit, P. (2010), ‘Varieties of Public Representation’, in Ian Shapiro, Susan Stokes, E.J.Wood and Alexander Kirshner, eds, Political Representation, Cambridge University Press, pp. 61-89.

Pitkin, H. (1967), The Concept of Representation (Berkeley, CA: University of California Press).

Skinner, Q. (2005), ‘Hobbes on Representation’, European Journal of Philosophy, 13, pp.155-184.

Sutherland, K. (2008), A People’s Parliament (Exeter: Imprint Academic).

55 Responses

  1. The great conceptual problem with this otherwise excellent discussion is that it concentrates on the state, assuming that decision on matters of public goods must be consigned to states. As I argued in Is Democracy Possible? and will in some forthcoming publications, the way out of most of our problems is to abolish power-trading, which inevitably results in poor decisions. Because in almost every decision politicians have to form coalitions in which they trade support for proposals that are not important to them for support on proposals where they need it, very many decisions are determined by considerations that are irrelevant to the merits of the case. They are determined by the power strategies and tactics of politicians, the structure of the distribution of votes and the power of organised interests.
    The only solution is to entrust decisons in particular functional matters to specific authorities that are statistically representative of the legitimate interests most directly affected by their decisions.
    The obvious objection is that these authorities would want to spend too much. That can be overcome by allocating them a standard budget, but allowing them to overspend if their constituency is willing to pay a surcharge on their income tax to meet the bill.
    That would be deliberative, flexible democracy.
    It would get away from the tyranny of the budget and the dead hand of huge bureaucracies.


  2. Have you looked up the various forms of probability vs. non-probability sampling?


  3. John, I agree that “statistical representation of the legitimate interests most directly affected by their decisions” would be ideal, but I wonder if, given the global extent of many issue areas, we’re moving closer to a utopian cosmopolitan vision rather than the current reality, and whether “legitimate interests” would ever be seen as fulfilling even the debased democratic criteria that we currently have to live with (it was the argument used against the extension of the franchise beyond male property-owners). In my book I do devote a chapter to localism but confess I’ve never grappled with the other end of the spectrum. It’s also the case that once you accept the principle of legitimate interests this would leave the power to set budgets overwhelmingly in the hands of those who pay the most, which would have serious implications for anyone believing in redistributive fiscal policies.

    But I do think that sortition will put an end to power-trading. Even if political parties continue to set the agenda in general elections, each legislative bill will be debated on its own merits by the allotted chamber. Whether or not the members of the chamber represent the “legitimate interests most directly affected by their decisions” is debatable, but I do think we need to start from where we are (the empirical reality of the nation state and its corresponding demos) rather from where we would rather be. If you take for example the hostility of the British public to the EU, which does aspire to a vision of politics a little closer to your perspective, I think you must agree that the starting point is a long way from your vision.

    However it might be hoped that if the AC is constituted on an ad-hoc basis (a new sortition for every legislative bill), and is demonstrated to be effective, then in the longer run there is no reason why the constituency for the sortition should not be adjusted depending on the reach of the issue (from local to global). I know your preference is for long-term statutory bodies for each functional matter, but I don’t see how that could in any way be squared with current democratic norms (I think they also might easily be hijacked by policy wonks and activists and could lose sight of the general interest). My proposal for a mixed constitution does provide a key place for expertise, natural right and long-term interests but the final decision is always in the hands of the (national) demos, albeit that the latter is very significantly constrained by the former.

    Many thanks for your input.


    PS I was struck by how your vision might well have resonated with early-modern democratic practice. According to a 2001 paper by Mark Goldie the extent of ad hoc officeholding (usually in local bodies) in early modern England has been dramatically underestimated — “with rotation, during a decade, over half the adult male population would have held office”. The problem was this was all destroyed by centralisation, professionalisation and the growth of party politics, with all the malign consequences that you outline in your post.


  4. Keith, thanks for your attention to my points.
    Part of the reasons for my insistence on interests is to minimise “expressive” politics in which decisions are made out of resentment, racism etc. Of course, what is a legitimate interest is very debatable in many contexts.
    So I have focused on the delivery of services, where interests are pretty clear and most people have a variety of interests and need to arrive at a balance between them. I would hope that starting small we might learn how to tackle the bigger issues. That may seem too slow , given the urgency of so many of the big problems, but it is amazing how quickly opinion can change. THink back to the sex revolution.
    Fundamentally, people need to recognise that the point is to get good decisions. The experts must be critically evaluated and controlled, but on a pragmatic basis. That is, in fact, I believe what most people want, but they are frustrated by having to buy package deals sold on crude appeals to “brand image”. We have to show how good decisions might be generated, in the first place on matters that they can understand from personal experience.
    Cheers, John


  5. “Fundamentally, people need to recognise that the point is to get good decisions.”

    Agreed. Unfortunately the procedural view of democratic legitimacy is so deeply ingrained that this is a tough one to sell (compounded by the difficulty in deciding over what is a legitimate interest and the scare-quoting of expertise). This is why I derive my view of democracy from Condorcet, but many would deny that there is such a thing as a “right” answer in political affairs.


  6. There is no right answer in political affairs, but there are a lot of wrong ones, and the fact that there is no absolute best does not mean that none is any better than any other.
    As Condorcet discovered, procedures have their limitations. Even the best systems of logic generate paradoxes, and all procedures have preconditions that limit their applicability, as we all know.
    Majority voting replaced dynastic succession as the legitimate way of deciding who is to exercise power. The concrete arrangements that instantiate it are almost as arbitrary as the “rights” of monarchs (Dieu et mon Droit) or English spelling. That, in itself, doesn’t matter too much. We have to have arbitrary conventions that enable us to coordinate, communicate and act. Changing circumstances can make some conventions counterproductive, as our present political ones are in relation to the problems of our situation.
    I doubt whether any constitutional top-down change will work, or any procedural fix.
    The trouble is that it is very difficult to change conventions by agreement, except in regard to quite specific matters, like adopting the metric system. But it does happen that changes introduce in particular contexts catch on and are adapted to other contexts, suggesting new possibilities. (Compare the case of development of technologies. Free enterprise beats central planning every time, but we need to have it not just in “private” form but in collective forms.)
    Attempts to fix systems from the top-down tend to be dominated by the problematic that reflects past problems rather than present or emerging ones.
    Cheers, John


  7. I ain’t got no book learnin’, Keith. I just nail a new bit onto the model periodically and think myself clever if it goes on straight. Pray indulge my naïve hammerings, even if my parlance is somewhat less than scholarly.

    Actually I have read The Concept of Representation since you encouraged me to do so a while back. We can easily agree on Pitkin’s definition of ‘descriptive representation’ and the likely efficacy of sortition as a tool towards achieving it. However, Pitkin clearly does not have sortive bodies in mind when making her assumptions about ‘active representation’.

    The key principle of random selection is that it ensures equal opportunity of political participation for every citizen. Given that we agree on ‘descriptive representation’, the remaining philosophical requirement is ‘equality of participation’, and for that we need systems, not philosophy.

    Geographic constituencies are an anachronistic construct left over from the days when we put a man on a horse and sent him off to parliament to represent our community. Politicians perpetually bleat about their constituency duties, which for some means spending 70% of their time writing letters on behalf of constituents, more often than not to do with housing disputes, immigration issues or disputes with local authorities. The representatives use their privileged status to apply pressure in a system based upon confrontation. Allotted representatives cannot be expected to undertake this role and there’s no reason for them to do so. It can easily be accommodated by a beefed-up, independent, citizen’s advice/ombudsman service.

    Representatives need not be burdened with the notion of representing anyone else, especially not an artificial constituency. As you point out, elected representatives routinely represent their own interests and those of their political parties ahead of the interests of their constituents.

    ‘Responsive representation’ as Skinner describes it, is unnecessary and impractical in an AC if the allotment is correctly weighted to be representative of the universe. Modern technology means that it can be. This is what dictates the representative sample size, not the conventions of existing representative structures. I could give you a formula for it if I had enough spare time to do the work.

    I don’t recognise the numbers you quote as being necessary to constitute a representative AC, although I do think Coote and Lenaghan are close with their ideal size for deliberative groups.

    I have no doubt that systems designed to reduce friction in the deliberative process can ensure both equality of participation and democratic outcomes that the majority of citizens will accept as being more legitimate than those arrived at by elected representatives under present system.

    Fishkin’s reliance on carefully trained monitors is conclusive proof that his system doesn’t work. Government of the people by the people is not a simple matter of aggregating votes on binary issues either. Representatives – maybe we should call them ‘allottees’ – need to be able to contribute.

    Rather than arriving at decisions that are ‘no worse’ than those achieved by elected representatives, ordinary citizens can use mechanistic methodology to arrive at the ‘rightest’ possible decisions based upon the quality of information supplied to them.

    I agree that the legislature must be independent but otherwise it is necessary to imagine a governmental structure that is dramatically different from those that currently exist. It’s like comparing a Monopoly board with a three-dimensional chess set. The model has to work cohesively from a local level to a national level. Certainly, the Palace of Westminster and all the council buildings in Britain could not accommodate it but that’s no reason not to develop it. You have to carefully consider the way in which policy is arrived at and all of the forces that might act upon it. And you have to fully understand the factors influencing decision-making.

    As I said before, the Athenians could walk to the Assembly. We can’t do that but we don’t need to because we have communications technologies that they didn’t have. Neither is there is no point in us trying to replicate Athenian systems. We have far superior systems technologies thanks to advances in social science.

    Mixed constitutions fall down on the notion that elections confirm legitimacy. Everywhere in the Western world, turnout is falling unrelentingly for the very reason that elections are understood to be illegitimate expressions of the will of the people. They’re recognised as a confidence trick, little more than a measure of advertising effectiveness won, more often than not, by the party with the biggest war chest. Why try to maintain the myth that elections are in any way legitimate when manifestos are abandoned as soon as the results are declared?

    I appreciate John’s point about it being very difficult to fix systems from the top down but I don’t see that as a reason not to try. The only legitimate way to change the system under the current system is to win an election. It so happens that anti-political feeling is so strong in the wake of the banking crisis and the failure of politicians to deal with issues such as climate change that the prospect of winning an election on a platform of changing the system is becoming imaginable, especially in Wales where the political settlement is weak (despite the outcome of the ‘powers’ referendum in which only 34% turned out to vote).

    No system is perfect, as John says, but the system we have no longer meets the needs and aspirations of society. There’s never been a better time to try to change it by ‘democratic’ means.


  8. Martin: “The key principle of random selection is that it ensures equal opportunity of political participation for every citizen.”

    I think you’re confusing sortition with the National Lottery. Depending on your perspective the principal benefit of sortition is either 1) the elimination of corruption and faction via the mechanism of the “blind break” or 2) the best (only?) way of ensuring that the legislature should be a “portrait in miniature” of the population being sampled. Most of us on this blog are interested in the second function as a way of instituting real democracy, whereas adherents of the first position argue (Dowlen et al.) that sortition is just a way of cleaning up electoral representation.

    Whatever the minimum number required to achieve descriptive representation may be, if you subdivide it then the AC ceases to possess this quality — thereby forfeiting its only claim to democratic legitimacy. This is emphatically the case when an individual member stands up to speak, for at that point in time she is only representing herself. Similarly it can be argued that the AC as a whole has a (general) “will” but only when it votes by secret ballot. When an individual member either proposes a legislative act or acts as an advocate for a proposal, she is only expressing her own will.

    As this is the central conundrum for advocates of sortition, I think we should focus on this point alone. I would dearly love someone to prove me wrong!


  9. John, forgive me if I don’t respond here, as I’m anxious not to distract from the “central conundrum”. Why not send a post to the blog so that we can have another thread focused on your vision? Best to send it to Yoram for uploading:

    I removed my email address to prevent harvesting by spam-bots. My email address appears here: https://equalitybylot.wordpress.com/about/


  10. Keith,

    You are correct that it is only the allotted chamber as a whole that is representative of the demos, but that does not mean there is any forfeiting of democratic legitimacy when individuals actively participate. Members can legitimately advocate, learn and participate as individuals in order to contribute to the final representative decision. The individual members do not have any constituency to speak of, and don’t have individual authority. They are not subject, nor should they be, to any sort of accountability to some specific constituency or group. Each member of an allotted chamber simply acts as he/she thinks best. Each member is only representing him or herself, and his or her notion of what is good policy for the society in which the member lives.

    In terms of small group deliberation, it would certainly be possible for an allotted chamber to divide and redivide into committees for deliberation by lot. These “break-out groups” would bring back insights, facts, or analysis to the plenary group.


  11. Terry,

    Individual members of the AC only occupy their (highly privileged) position on account of the descriptive criterion which (as you acknowledge) only applies in aggregate. This is particularly problematic when the AC is given unchecked powers, as many people on this forum have advocated. The only way of reconciling active participation with democratic legitimacy would be if each member spoke for exactly the same length of time and (to adopt Austinian parlance) the “illocutionary force” of their individual speech acts were equalised (an impossible task).

    Moving from linguistic and democratic theory to practice, it would be quite possible for either one, or a group of charismatic individuals to dominate the proceedings of the AC and effectively overrule the silent majority. To believe otherwise would be to accept Yoram’s view that the emancipatory power of AC membership will magically transform every member into passionate and eloquent advocates for their chosen cause, that these causes will mirror the will of the wider citizenry, and that this transformation will occur in an entirely even-handed way.

    But what if this magical transformation doesn’t happen? If I am not allotted a place on the AC, then my will is represented by the person or persons with views like me. That works fine when it comes to voting — if my views are minority views then my will will be overruled by the AC, in which all votes are equal, irrespective of the charisma and rhetorical powers of each allotted member. But it’s entirely different when it comes to active representation — if my “representatives” happen to be quiescent and uncharismatic then they are far less likely to influence the outcome. This is not just a thought experiment as activists and others with an axe to grind are far more likely to possess the character attributes that go with persuasive speech acts than ordinary folk. This was how the Militant Tendency managed to infiltrate and then dominate constituency Labour Party groups during the 1980s in the UK, and is also how militant activists within trade unions predominate over the majority of moderate members. So it’s not just possible but highly likely that similar effects will happen to an AC that has anything other than aggregate powers.

    Regarding “break-out groups”, the role of these in the DP was restricted to formulating questions to pose to the expert advocates during the plenary sessions, and even this required careful moderation in order to ensure that they were not dominated by charismatic individuals and/or those with perceived higher status. Whatever one may think of Fishkin’s work, it’s based on two decades of careful experimentation, as opposed to (optimistic) speculation on how the group dynamics of an AC might work out in practice.

    In sum, active participation by individual members of a group who’s democratic legitimacy is only defined in aggregate is a breach of the democratic equality of all citizens. Surely the juxtaposition of “individual” and “aggregate” in the previous sentence makes its truth inescapable — “analytical a priori” in philosophical parlance? As Galen Strawson put it: “You can see that it is true just lying on your couch. You don’t have to get up off your couch and go outside and examine the way things are in the physical world. You don’t have to do any science.”

    In which case why are we still arguing about it?


  12. The ‘magical transformation’ happens through cybernetics. Without systems you have an irresolvable conundrum.


  13. I’m as big a fan of a systems approach as the next man, having argued that Harrington was the original second-order cyberneticist, but it’s not magic or alchemy.


  14. I’ve been meaning to comment on this for a while. FWIW, here’s my two cents.

    I think we need to distinguish two questions here. First, what do we want a decision-making body to do? Second, in what ways do we want a decision-making body to be representative? Presumably, our answer to the second will be critically informed by our answer to the first. But both are important. Nobody wants descriptive representative just for its own sake. The idea is that a descriptively representative body will do something that a body that isn’t descriptively representative won’t.

    Keith, I take it that your concern is that a descriptively representative body can do some things well–express an opinion on a bill presented to it–but not others–like draft legislation. I think there are two reasons for this, on your account. First, while the opinion expression function can be performed equally well by everyone–it doesn’t take much to be able to cast a vote for or against a bill–but the process of drafting legislation requires a much more active kind of work, and you cannot assume that this will be done equally for all interests in society (the way you can assume that a descriptively representative body will vote in such a way as to equally represent the interests of everyone).

    Second, only small bodies can draw up legislation, and small bodies cannot be descriptively representative, except in very crude ways. And so even if an AC selected committees, etc., descriptive representation would have failed.

    So here’s a question, Keith. It sounds to me like you don’t believe ANY body can be both descriptively representative and perform these active functions–not a randomly-selected AC, not an elected body, none. If so, why be bothered by the fact that an AC relies upon a non-descriptively representative sample to draft its legislation? It won’t be descriptively representative, but then nothing will, so we’re no worse off in that regard here than with any other proposal?

    What I guess I’m pushing for is a list of the functions that must be performed, and some criteria for telling when they’re done well. Only then can we say that such and such body will perform them well. Your evocation of Pitkin’s distinction is a good starting point, but it doesn’t give us a complete list of the things we’re looking for in the process of drafting and enacting legislation.

    Hope that makes sense.


  15. Peter: “So here’s a question, Keith. It sounds to me like you don’t believe ANY body can be both descriptively representative and perform these active functions [drafting legislation] – not a randomly-selected AC, not an elected body, none.”

    I agree with your assessment of my arguments; let’s deal with the issue of democratic representation first. The policy proposals of an elected body do possess a democratic mandate, albeit not a descriptive one. Whilst we are all concerned that elected bodies are a) elitist and b) returned on the basis of rational ignorance, nevertheless this is what passes as democratic by contemporary standards (whereas the speech acts of individual members of allotted chambers don’t). For better or worse, like it or lump it, ‘democracy’ means political parties formulating proposals and putting them up for (uninformed) judgment by the electorate. The party or parties that wins the general election has the right to turn these proposals into legislative bills, even though few people have bothered to read the manifestos in which these proposals are buried. Although we are all agreed that the sortive alternative can dramatically improve the scrutiny and judgment function, its democratic legitimacy only applies in aggregate. Opinion polls only legitimately reflect public opinion in aggregate: no pollster would ask an individual respondent what they thought about an issue and then make any claims on the basis of that individual response; similarly any speech act by an individual allotted member immediately loses any claim to democratic legitimacy. This is an analytic apriori argument, in that the very notion of descriptive representation rules out any form of active individual (or even small group) behaviour by members of the AC. As I have no formal credentials in analytic philosophy I would dearly love to be proved wrong, although I confess I can’t see how.

    The combination of preference elections and sortition that I propose in my book allows for one form of democracy (active) to be checked by the other (descriptive), so would appear to be preferable to a system that only has one form (assuming that Pitkin’s distinction is a valid one). Pluchino et al’s policy for allotted MPs would also effectively combine these two distinct forms of democratic representation: https://equalitybylot.wordpress.com/2011/03/16/more-from-pluchino-et-al/
    See also Peonidis: http://www.uclouvain.be/cps/ucl/doc/etes/documents/Peonidis.AllottedMPs.pdf and
    Marcus Schmidt’s Danish proposal (Hansen, 2005, pp. 54-5), along with other bicameral proposals in which the elected house alone has the right to introduce policy proposals, such as Zakaras: https://equalitybylot.wordpress.com/2010/09/24/the-elected-legislators-burden/
    and Lieb: http://www.psupress.org/books/titles/0-271-02363-5.html Whereas the proposals by O’Leary, Callenbach & Phillips etc. fail to make this distinction. As for McCormick’s Machiavellian Democracy, I’m not sure, as I’m only half-way through it.

    >What I guess I’m pushing for is a list of the functions that must be performed, and some criteria for telling when they’re done well. Only then can we say that such and such body will perform them well.

    This is moving into the issue of competence, as opposed to democratic legitimacy, so I’ll reply to that later, so as not to conflate the two issues.


  16. Peter: “What I guess I’m pushing for is a list of the functions that must be performed, and some criteria for telling when they’re done well. Only then can we say that such and such body will perform them well.”

    I take this to be a question regarding the effectiveness of a legislative assembly in proposing and drafting legislation, so I’m replying to it separately from the question of democratic legitimacy. Once again, most writers (ancient and modern) have argued that the operational efficiency of randomly-selected groups (the wisdom of crowds) only applies at the collective (judgment) level. Mogens Hansen pointed me to the following:

    ‘The many, of whom each individual is but an ordinary person, when they meet together may very likely be better than the few good, if regarded not individually but collectively, just as a feast to which many contribute is better than a dinner provided out of a single purse. For each individual among the many has a share of virtue and prudence, and when they meet together, they become in a manner one man, who has many feet, and hands, and senses; that is a figure of their mind and disposition. Hence the many are better judges than a single man.’ (Aristotle, Politics, 3, 11)

    For Condorcet, the relative probability of a randomly-selected jury arriving at the ‘right’ answer increases exponentially with the size of the group (‘terms and conditions apply’). So, in a group of only 399 people with an individual probability of .55, the aggregate result converges rapidly to .98 (Przeworski, 1999, p.27). From an epistemic viewpoint it’s hard to imagine a more reliable way of getting the ‘right’ answer.

    But this only applies at the aggregate level, individual jury members are significantly likely to have the ‘wrong’ answer, just as, in the ‘guess the weight of the ox’ competition referenced in Surowiecki (2004). Francis Galton was right to suspect that many of the amateur guesses were wildly inaccurate, yet was astonished to find that the average estimate was only one pound short of the actual weight (1,198 pounds). By analogy the high reliability of the aggregate view of the allotted chamber should not blind us to the fact that individual views may well be completely off the wall. (Needless to say the above analysis presupposes that there is a ‘right’ answer to political questions, as opposed to the competing ‘formal’ theory, which is only concerned about adherence to the correct democratic norms.)

    But are the legislative proposals of elected chambers any more likely to be effective? Leaving aside my own preference (that legislative proposals should, for the most part, be originated by government ministers appointed on merit alone), with elected chambers at least voters have the power to appoint legislators who appear to be competent and (more importantly) to dismiss them when they prove, through their own actions, that they have betrayed this confidence. Whilst shutting the stable door after the horse has bolted may not be entirely optimal it’s about as good as it gets if one doesn’t accept that competence criteria can be assessed in an objective fashion (or if you argue that any such an assessment will be overriden by democratic priorities).

    Once again, I must stress that this commentary is only dealing with the ‘efficiency’ of different methods of generating legislative proposals. (In the last response I argued that only individual elected representatives have the necessary democratic mandate to turn manifesto pledges into parliamentary bills.)


  17. Keith (and Peter),

    It seems to me that an Allotted Chamber (AC) will take ultimate action in aggregate by saying yea or nay to proposals placed before it. We agree that the democratic legitimacy in this case is clear. This was the primary function of both the full assembly and the allotted bodies in Athens as well. Orators made their cases, and the members judged the arguments.

    The source of the various proposals, however, need not have the same descriptive democratic credentials. A good idea might come from a single expert outside the government, for example. We simply need an AGREED UPON PROCESS for items to come before the AC. (We could use a randomly selected panel to set the agenda items, or randomly select from among proposals – assuming there are too many to look at them all, or use a thumbs-up-thumbs-down mass voting system like YouTube, or committees of the AC, etc.)

    The democratic legitimacy comes from the prior agreement to the process…not necessarily the descriptive representativeness of the source. Keith suggests an elected chamber as the source…Fine, if that is what is agreed upon. But I don’t think such a chamber has any more democratic legitimacy than committees of the AC itself (setting aside the competency issue, as you suggest we deal with it separately). The elected chamber may well be LESS representative than the poorly representative AC committee.

    Let me propose a little thought experiment…There were 200 survivors on a deserted island with no hope of rescue. They formed a democratic society in which they all could discuss and vote on all matters facing their little society. Though all are ALLOWED to participate, some were more eloquent and convincing than others, even after they imposed equal time-limits on speaking. This is still “democratic,” as absolute equality in all attributes is not necessary, merely equal opportunity and equal voting power.

    This became too time-consuming, with not enough time remaining for gathering food, etc. So they had a choice…They could elect a representative group to make all decisions (but worry that this group might become corrupt, ensconced and decide to hoard the scarce food sources for themselves, etc.), or they could randomly take turns governing, with half making decisions, while half gathered food, and then switch roles every other week. In either case, just as when ALL were making decisions, either representative body will continue to have variation in speaking ability among members, etc. The democratic legitimacy is the same, if the process is agreed to in advance.

    Keith, you are putting a far higher standard of legitimacy on the AC members, while simply accepting as given the legitimacy of individual members of the elected chamber.


  18. ‘The democratic legitimacy comes from the prior agreement to the process…not necessarily the descriptive representativeness of the source.’

    ‘We simply need an AGREED UPON PROCESS for items to come before the AC.’

    What Terry says. Exactly.


  19. Terry, many thanks as always for your very thoughtful response. I need to ponder over this carefully, but for the moment would note that your “procedural” criterion for democratic legitimacy is identical to Schumpeter’s, and it’s precisely this formal, minimalist conception of democracy that most deliberative theorists are seeking to improve on. Also to note that elections produce a different form of representation than the descriptive variant — elected representatives are legitimate in the sense that they have received more of the votes than their competitors, and this is how modern democracy is defined. I know we are all dissatisfied with this form of representation, but I don’t think you can say that it is illegitimate.

    But let me ponder a bit and then respond tomorrow.

    many thanks



  20. >I know we are all dissatisfied with this form of representation, but I don’t think you can say that it is illegitimate.> Keith

    Agreed… I said individual members of EITHER an elected or an allotted chamber can reasonably be deemed to have democratic legitimacy.

    The question is which will more closely approximate the decisions that the entire population would make if the entire population participated in a good deliberative process.

    There was actually a survey conducted on attitudes on this question in the U.S. in 1999. Check out this link:


    The purple bar chart at the end of section 2 asks about confidence in a 500-member sortition group vs. Congress, with 66% having more confidence in the allotted sample and 15% having more confidence in congress.

    “One of the more striking findings from the current poll is that a strong majority of respondents said they would have more confidence in the decisions of a random sample of Americans who were given information and a chance to deliberate on a subject than in the decisions of Congress. Respondents were given a description of what is sometimes called a “deliberative poll.” 1 By a four-to-one margin, respondents expressed more confidence in the majority decisions of such a sample than the decisions of Congress. Also, 68% said the decisions of such a sample “would be more like the wishes of the majority of the public” than the decisions of Congress” (less, 25%). Fifty-seven percent said this sample “would be more … likely to come to consensus than Congress” (less, 31%). As will be discussed below, the ability to come to consensus is seen as something positive by very strong majorities. “


  21. This report is a great find, Terry – I haven’t seen it before. I’ll create a post linking to it.


  22. Yoram,

    The fellow at the University of Maryland who was behind that poll is still doing amazing work — on an international level. His name is Steven Kull, and has several institutional outlets…

    Check out his Program for Public Consultation at
    and also WorldPublicOpinion.org
    Program on International Policy Attitudes (PIPA)

    An interesting survey on global views of how representative various governments are believed to be is at

    This last report has a fascinating result for Egypt…showing in 2008 that the public there reported feeling that their government represented the people BETTER than in most other countries. Interesting in light of recent events. Of course, I don’t know the nature of the survey (were people unwilling to be frank??)



  23. The Kull survey bears out my point, as no-one here would be surprised by the claim that the decisions of a deliberative poll were greeted with greater confidence than the decisions of an elected Congress, the crucial points being a) ‘given information’ and a chance to deliberate on b) ‘a subject’ [‘a’ and ‘b’ being derived externally]. Such decisions represent the informed view of the whole population and show what descriptively-representative groups do well, i.e. make decisions, as opposed to the active tasks performed by individual representatives (policy introduction and advocacy).

    As for the claim that an ‘agreed upon process’ is sufficient for democratic legitimacy, such formal, procedural approaches result from the hollowing out of substantive legitimation. The standard Schumpeterian view, that democracy merely enables the orderly alternation of elites is a hollowing out of the theoretical potential of elections to institute substantive representation. It would be perfectly possible for each elector to expend the considerable time and energy required to properly research the background and policies of the candidates in a general election and vote for the candidate that genuinely reflected her own interests and views. Parliamentarians could also attend all the debates and only vote on the basis of their considered judgments, ignoring all partisan pressures. Diligent voters could watch the parliamentary debates on TV and then hold the candidates to account at the next election. All this is possible and is how electoral democracy ought to function, but we know of course that it doesn’t work, hence the resort to the cynical, hollowed-out view that elections just provide an ‘agreed upon process’ to throw the rascals out every now and then and let in the next bunch.

    For an ‘agreed upon process’ to be legitimate, legislative proposals have to be arrived at democratically, and the only way we know of doing that is via elections. The elected chamber will undoubtedly be less descriptively-representative than the committee of the AC, but its manifesto proposals were approved by popular vote, whereas individual acts by members, or sub-groups of the AC have no democratic legitimacy whatsoever, as the principle of descriptive representation only applies in aggregate. This is true by definition.

    Regarding your thought experiment of a group of 200 survivors, I agree with you (and Aristotle) that governing and being governed in turn is a form of democracy that is valid in small societies, where everyone will have a turn, but large groups require representation. What I’ve tried to demonstrate in this post is that both kinds of representation – active and descriptive – are necessary for a large-scale system to be described as democratic.


  24. Terry,

    Thanks for the links.

    I was aware of the Program for Public Consultation and favorably compared the methodology of its study of the public’s budgetary priorities to the methodology of Fishkin’s DPs.

    As you point out, their international survey on governance and democracy is very interesting as well.


  25. Two points, one short and one long. First, I remain unimpressed by the claim that an AC is somehow legitimated because it is endorsed by the people in advance. By that line of argument, a monarchy could easily become legitimate. And in any event, unanimity is a pipe dream, and without that any claim that a system is legitimated by consent simply pushes the problem back one step (i.e., instead of needing to know what procedure is legitimate, we need to know which procedure for selecting procedures is legitimate).

    Second, if I understand you, Keith the problem of effectiveness (i.e., do good decisions get made) has nothing to do with the problem of democracy. You think that there is a sense in which a system that employs elected and accountable representatives is democratic, and another sense in which an AC (collectively, not individually) is democratic, and so a system that combines both is more democratic than one that features only one. Is that it? That’s how I’m interpreting this…

    “The combination of preference elections and sortition that I propose in my book allows for one form of democracy (active) to be checked by the other (descriptive), so would appear to be preferable to a system that only has one form (assuming that Pitkin’s distinction is a valid one).”

    It might be true, but I’m not sure I see the argument for it. At least I take it that you believe the combined system is more DEMOCRATIC, not simply good in some other respect.


  26. Peter: “First, I remain unimpressed by the claim that an AC is somehow legitimated because it is endorsed by the people in advance.”

    Sorry, who is supposed to be making this claim? Certainly not me. My argument is that an AC can only claim democratic legitimacy because it is descriptively representative (not because it has been endorsed in advance), and this only applies at the aggregate level. Perhaps you are referring to my claim that policy proposals that have been endorsed in an election are legitimate. If so then that’s true; furthermore I don’t know of any other democratic way of endorsing policy proposals other than by elections. What way would you suggest?

    I’m afraid I don’t understand your second point. Obviously effectiveness and democratic legitimacy are distinct characteristics, so I dealt with them in two separate responses. Regarding the argument for two forms of democracy (descriptive and active), this is all derived from Pitkin — I don’t claim any authorship for it. We’re all agreed that sortition is the best way of enabling descriptive representation; I’m just making two further points:

    1) An AC cannot fulfil the active role of political representatives because its legitimacy only applies at the aggregate level, therefore

    2) If active representation is (according to Pitkin) the most important function of political representation, then elections must continue to be part of the mix.

    All these arguments rely on the democratic criterion, I’m not referring to the separate point of effectiveness. I’m not really sure how to make this any clearer, but please tell me if there’s some element missing.


  27. Keith,

    I think I can clarify Peter’s first point…I believe he was referring to me…and our exchange in which I first raised the notion that a democratically agreed upon process for bringing items before the AC gave democratic legitimacy to a legislation drafting process. You and Peter disagree (You associated this with a Schumpeterian view of democracy).

    My notion is that an AC can legitimately delegate tasks, as long as ultimate decisions are subject to the AC. Tasks could include executive and administrative functions, drafting proposals to bring back to the AC, etc. You suggest that only elected individuals have individual democratic legitimacy to act as individuals. Why? Could not the AC, (in plenary with full democratic legitimacy) direct an individual, or a committee, to perform some task for the full AC to review? Couldn’t they adopt a policy that proposed legislation submitted by petition with 100,000 signatures would automatically be put before an allotted body to consider?
    Elections give individuals legitimacy by tradition, more than logic. With first past the post election rules typical in the UK, Canada, and the US, the elected person may actually be strongly opposed by a majority of his or her constituents.


  28. Terry, thank you — as always — for bringing your usual clarity to this issue. I’m relieved that not everybody speed-reads blog posts!

    “My notion is that an AC can legitimately delegate tasks, as long as ultimate decisions are subject to the AC. Tasks could include executive and administrative functions, drafting proposals to bring back to the AC, etc”

    Regarding the drafting of proposals, the trouble is that an AC is of necessity a small group (in order to enable deliberation, and in order to preclude the problem of rational ignorance). And in a small randomly-selected group a few vocal and opinionated individuals will inevitably dominate the proceedings. I acknowledge that they will ultimately be checked by the aggregate decision, nevertheless the range of options will of necessity be prescribed by the vocal and opinionated individuals. i just dont’ think this is democratic.

    And I don’t understand why we would want to select individual executives and administrative officers randomly. This makes sense in a small polis, where all can rule and be ruled in turn, but this is impossible in a large nation state, so we might as well select executives and administrators on the basis of competence. The AC could still hold them to account and would, in fact, be much better when the individuals are at arms length, rather than being concerned not to upset or criticise one of their own (realising that it could be their turn next and they could be just as incompetent!). And imagine how someone (ie most of us) who did not draw the golden ticket from the kleroterion would feel to know that not only legislative decision-making but also policymaking and executive/administrative functions were being performed by the members of the magic circle. “Disenfranchised” is the word that immediately comes to mind.

    “Couldn’t they adopt a policy that proposed legislation submitted by petition with 100,000 signatures would automatically be put before an allotted body to consider?”

    The only problem I have with this is the word “automatically”. It would be quite easy for a lobby group to collect 100,000 signatures. Just stand in the market square and ask passers-by to sign a petition for anything (appearing) warm and cuddly and most people will agree to sign. To my mind this should be only the first of three stages for the initiation of policy proposals:

    1) Collect (say) 100,000 signatures.
    2) All the proposals that meet this criterion are then voted on, and prioritised in a general election.
    3) The top proposals are then put to the AC for deliberative scrutiny. Those that receive a majority vote in the AC become law.

    Note that the time gaps betwen the 3 stages will be filled with debate in the public arena, spurred on by media, bloggers, pub conversations and lobbyists (for and against). This is very important for good quality legislation, for all the reasons that Condorcet has given us.

    “Elections give individuals legitimacy by tradition, more than logic. With first past the post election rules typical in the UK, Canada, and the US, the elected person may actually be strongly opposed by a majority of his or her constituents.”

    OK, but we don’t need to throw out the baby with the bathwater. The electoral process that I’m suggesting would not be FPTP, and elected persons would be directly representing the views of their “constituents” on the particular legislative issue that they are advocating.

    Elections are really the only way of ensuring a democratic mandate for a legislative proposal, whereas sortition is the only way of ensuring democratic scrutiny. We need a judicious mix of both forms of democracy (active and descriptive); as soon as one form steps into the territory of the other then democracy is the loser. As you rightly say the problem with FPTP is that elected officers are out of synch with the views of their constituents (ie elected representatives fail the descriptive criterion); the trouble with prioritising the speech acts of individual members of the AC is that they then lose their only claim to democratic legitimacy (descriptive representation). So we need elections and sortition.

    Once again, many thanks for bringing some clarity to this issue!



  29. Keith,

    A few clarifications and comments.

    1.You wrote
    >And I don’t understand why we would want to select individual executives and administrative officers randomly. It would be quite easy for a lobby group to collect 100,000 signatures.Elections are really the only way of ensuring a democratic mandate for a legislative proposal…the trouble with prioritising the speech acts of individual members of the AC is that they then lose their only claim to democratic legitimacy<

    Do I understand you to mean, that the AC should not engage in any deliberation itself, but only listen to pro/con arguments coming from elected legislators (or others) and then vote yea or nay? (I confess I haven't read your book yet) That is compact and simple, but I fear misses some of the major benefits of sortition, which includes the give and take and compromise possible OUTSIDE partisan frameworks.


  30. I used symbols like this “>” in my last post that seem to have made most of what I wrote in my post invisible…Can Yoram or an an administrator access the full text of what I typed and post it, so I don’t have to retype it. Only 10% of what i typed appears above.


  31. Terry – sorry, I think this is irrecoverably gone. Using < and > in your comment confuses the system since they are interpreted as html tag delimiters. If you wish to use those symbols, you should use “&gt;” for “>” and “&lt;” for “<“.


  32. Terry: Do I understand you to mean, that the AC should not engage in any deliberation itself, but only listen to pro/con arguments coming from elected legislators (or others) and then vote yea or nay?

    It depends what you mean by deliberation. Jim Fishkin is adamant that the word derives from “weighing” (up an argument), in which case individual speech acts by members of the AC are not required (this was also Rousseau’s view).

    On the other hand Fishkin is adamant (personal communication) that the individual small-group sessions are an essential part of the deliberative process, I imagine on the basis of “how do I know what I think until I hear what I say”. But again he is insistent that this requires very careful moderation, in order to ensure that the debate is even-handed and doesn’t privilege high-status and opinionated individuals. So I’m happy to defer to Jim on this, as he’s just about the only one of us to have moved beyond armchair speculation to doing the social science experiments. The ultimate role of small groups in the DP is to prepare questions for the expert advocates in the plenary sessions.

    But the issue of trained moderators is a difficult, if not intractable, one in a legislative assembly, and I’m extremely wary of the idea of individual members of the AC trying to sway the debate in plenary session. If there is some way of ensuring that this doesn’t disturb the aggregate balance then I suppose it would be OK, but I can’t frankly imagine how this would be possible. Much better for the AC to preserve the integrity of its descriptive mandate by limiting itself to the judgment function and to leave the debate to the opposing advocates. But I’m open to persuasion! (Note this is an entirely different issue from the introduction of legislative proposals).

    I’d love to react to the other 90% of your message and hope you can resurrect it.



  33. Keith,

    Well, here is a shorter version of what I typed and trashed with errant symbols…

    1. You wrote:
    And I don’t understand why we would want to select individual executives and administrative officers randomly.

    I respond:
    I agree…I think the AC can select executives, administrators, drafters, moderators, etc. — not from among their own numbers…but professionals. I refer to the system of municipal gobvernment in many U.S. cities…where there is no elected chief executive (mayor), but instead an appointed city manager. I think it would be best to have an allotted meta-legislature that only deals with “once-removed” issues such as this, creating a productive, corruption free framework for subsequent ACs to function optimally.

    2. You wrote:
    It would be quite easy for a lobby group to collect 100,000 signatures.

    I respond:
    I used that as a hypothetical. If the petition route were to be available, perhaps it might function similar to the way public campaign financing works in the the U.S. state of Maine. To qualify for public financing a candidate needs to raise a large number of $5.00 donations. This is earnest money. the amount needs to be large enough to stop thoughtless signing, but small enough to provide a minimal barrier to low-income citizens. Even $1.00 would probably work. this money could go into the AC operating budget.

    3. You wrote:
    Elections are really the only way of ensuring a democratic mandate for a legislative proposal

    I respond:
    Why does a proposal need any kind of mandate at all? It is just a proposal. We simply need a fair process for deciding what matters will go before the AC…A fair process could involve polling, random selection from among proposals submitted, an AC meta-legislature that only deliberates on what items will go on the agenda for future AC, petition, submissions by NGOs with more than X thousand members, etc., etc. It seems that elevating elected legislators to being the sole source of legitimate sourcing of proposals is unreasonable.


  34. My understanding of Keith’s argument is that an AC has democratic credentials by virtue of its descriptive representativeness. When it delegates power to someone else, or selects another official, the democratic credentials don’t carry over to those selected. And since many of these positions (executives, legislative committees) cannot be descriptively representative in themselves, they need other democratic credentials, and elections can provide them. Do I have this right, Keith? The only remaining piece of the puzzle is that Keith also wants to make sure that SOME piece of the system has democratic credentials from a source other than elections. (Otherwise, why not just elect the legislature?) Not sure I see the argument for the last piece, but perhaps it’s not a consideration grounded in democratic theory.


  35. Peter: “When [the AC] delegates power to someone else, or selects another official, the democratic credentials don’t carry over to those selected.”

    No. My point (derived 100% from Pitkin, p.145) is that the only thing that an AC can do without breaching its descriptively-democratic mandate is to VOTE (this is because its legitimacy only applies at the aggregate level). An AC could, in principle, vote on anything — including, inter alia, the choice of candidates for executive office and how long their refreshment breaks should be; indeed in my book I make it clear that the AC should have the final choice over which candidates (previously selected on merit alone) are appointed to executive office. What the AC does NOT have a mandate to do is to propose candidates for office, as such speech acts are made by individual persons; the same is true for the introduction of legislative proposals.

    Rather than ensuring a residual role for descriptive representation (“SOME piece of the system has democratic credentials from a source other than elections”), I want to ensure that ALL decisions (other than the day-to-day operational decisions of delegated government officials) are made by the AC.

    In sum:

    1) the judgment function (and ultimate sovereign power) in a legislature should be entirely in the hands of the SORTIIVE body. But, as the descriptively-representaitve mandate only applies in aggregate, anything that involves individual speech acts (nominating candidates for office, initiating policy proposals etc.) has to come from elsewhere.

    2) The only democratic mechanism we have for initiating policy proposals (and/or those people or parties who are advocating them) is ELECTION.

    3) As the criterion for executive office-holders is competence, candidates should be SELECTED (shortlisted) on merit, for final approval by vote in the AC.

    Hope that clears up the confusion — if not then please let me know what part of the argument is still in need of clarification.



  36. Terry, thanks for going to the trouble to resurrect your message. I’m grateful, as always, for the care you take to read other people’s work and I’ll try to respond in the same manner!

    1) On the selection of executive officers, does the two-stage system that I suggest in my response to Peter (immediately above) cover this? I think we are in agreement; I just want to focus on the distinction between a) shortlisting and b) approving candidates. It seems to me that an AC is the right body for b) but not a).

    2) I really like your idea of people having to “put their money where their mouth is”; and agree this would be an excellent way of finding out what legislative proposals people REALLY want. I only worry that this will be viewed (by silly people) as some sort of reversion to a property franchise. Don’t get me wrong, I think it’s a really excellent idea, but would be interested to hear the comments from those who argue the stronger case for absolute equality (although most citizens could afford $1-$5). Has there been any work or discussion on this sort of proposal for filtering legislative proposals?

    3) I’m sympathetic with your other suggestions for filters. My only concern is that all your suggestions (apart from polling, but including random pre-selection) will privilege activists, and it seems to me that democracy should be more about empowering the silent majority. Most people would not be inclined to propose legislation, but would like to be able to pre-select those proposals that they find favourable (and dismiss those that they don’t). In this sense polling and elections are the same thing. My preference for “election” would be an online opinion poll, rather than a vote for political representatives and parties (although they will continue to act as advocates for their policy proposals).

    I get the feeling that we are in agreement on the principles involved and only differ (slightly) over how they might be implemented. I’m particularly honoured to be debating these issues with someone with a long and distinguished history of political service, rather than the rest of us armchair philosophers and activists!

    best wishes



  37. “My preference for “election” would be an online opinion poll,”

    Sorry, I should have qualified that: “with universal suffrage, rather than proportional sampling”. The point is that voters are choosing legislative proposals, rather than political representatives.


  38. OK, let me try one more time to make sure I’m getting this. Anything that a descriptively representative body does is democratic. (So, presumably, would be anything done by direct democracy. I assume an AC is preferable on efficiency grounds–i.e., it will make better, more well-informed decisions.) But it has to do it as a collective body; otherwise, it’s just the action of a particular agent, and that’s not descriptively representative. There are at least two tasks that need to be done and yet cannot be done by descriptively representative bodies–drafting legislation and executing the laws. The former ought to be done by elected officials, because elections have some democratic legitimacy (more than sortition? less?), and the latter ought to be nominated (subject to AC approval) by some non-elected source (why non-elected?).

    I hope once I’m in Europe to participate regularly in the excellent Manchester Workshops in Political Theory. Perhaps a future workshop on democratic institutions would be a good venue for future work on this topic.


  39. Peter, so far so good; two further points that need clarification:

    “[drafting legislation] ought to be done by elected officials, because elections have some democratic legitimacy (more than sortition? less?)”

    Not just more, because sortive bodies have NO legitimacy other than in aggregate. Apart from voting, I can’t see how the descriptively-democratic mandate could be extended to anything else. To lampoon my own argument, the AC can choose between tea and coffee but cannot suggest an alternative (Tequila, Southern Comfort etc.).

    “and [executives] ought to be nominated (subject to AC approval) by some non-elected source (why non-elected?)”

    No problem in principle with an elected appointments commission, but there is a distinct probability that the executive would become politicised as what you suggest would be equivalent to a primary. Whilst the founders didn’t intend the executive to be a political office it immediately became one — an inevitable consequence of preference elections (indirect, direct or otherwise). Given that competence is the defining principle there are better ways of selecting competent executives than via an elected appointments commission that would avoid these partisan forces. I agree, of course, that with a statutory appointments commission the problem of quis custodiet applies, but I think this is easier to overcome than the problems inherent in an electoral solution.


  40. So when it comes to administration, there is no need for democratic legitimacy (although it wouldn’t hurt–i.e., you don’t object in principle to electing the people to do the selection)? That means, if I understand you right, that ability level is all-important here, in the sense that there are no competing considerations (like the need to ensure democratic legitimacy). That certainly sounds consistent with what I remember from your books. I wouldn’t say that I’m convinced, however; it seems a little naive to think that there’s some sort of completely neutral, nonpartisan way to conduct administration. Certain routine administrative tasks, maybe, but not the entire administrative process.


  41. “it seems a little naive to think that there’s some sort of completely neutral, nonpartisan way to conduct administration.”

    Agreed, but the alternative — an elected executive — is worse. And bear in mind that government ministers have to be approved by and can at any time be removed by the AC. This process will be augmented by intense media pressure, who will delight in taking as many ministerial scalps as they possibly can (acting in a partisan way being an offence that can result in a motion of censure).

    The idea of a non-partisan administration is more familiar to UK observers than a country like the US which operates an administrative spoils system. The civil service has become used to serving governments of all political persuasions and there is no reason, in principle, why the minister heading each department should not operate according to the same standards. Of course it won’t work quite like that, but then we don’t live in a perfect world. But I think having three separate principles — election, sortition and appointment — for each of the key elements of government is right, even if in practice it’s hard to live up to the ideal type.


  42. Keith,

    Two points…

    1. In your tea and coffee example, where the AC is limited to approving or rejecting choices presented, rather than making any original proposals, I think you are forfeiting one of the best potentials of the AC. Sometimes (often) the BEST course for society is one that none of the political parties are putting forward. What is more, upon deliberation, this fact is often quite apparent to average citizens. I think the AC needs a mechanism for rejecting with some sort of directive or request for an alternate proposal. This alternate proposal might be a simple splitting the difference between partisan proposals, or a completely different approach to an issue. I agree that the AC members individually (and in aggregate) are not the ideal drafters of such alternate proposals, but they should be able to give broad guidance to some more skilled drafters about what sort of proposal they are eager to consider.

    This leads me to my second point…
    2. A major weakness of sortition democracy is its lack of popular participation (unless some special provision is made). It is vital that any citizen who feels strongly about an issue have SOME fair means of trying to influence a decision. Otherwise public alienation of the most interested people will be rampant. Elections and campaigning provide a (largely symbolic) means of participation in a competitive democracy. Under sortition, activists could obviously organize rallies, write letters to the newspaper, blog, etc. to sway the general public (and presumably those who get selected for an AC). But more focused than this, I would propose that activists be able to draft proposals, write letters to the AC directly, etc. When it comes to writing letters to the AC, obviously not all memebers should be expected to read all letters. I like the idea of distributing such communications to individual AC members using a lottery system, so that no letter writer will know which AC member will get the letter…This will encourage activists to present arguments that contain justifications (an element of deliberation) that might be persuasive to anyone, (rather than arguments intended for like-thinking partisans). Activist proposals could be popularly evaluated (as a filter)…perhaps put online for thumbs up and down voting, to see what proposals have enough interest to make it to the AC.


  43. Thanks Terry, dealing with your two points in turn:

    1. All the AC needs to do is to reject the bill. It would then would be left to the next election for someone to propose a better alternative (the elections would be comparatively frequent, as one is not appointing a government). Given that the deliberations of the AC would be on public record, along with widespread media debate, it would be very evident why the AC had rejected the bill, but as soon as you leave it to the AC to propose an alternative you are, in effect, empowering individuals within the AC. However, so long as the AC-derived alternative was simply listed as another option at the next election the status of this proposal would be on a par with any other, so would not be problematic from a democratic perspective. The only advantage that the proposal would have is the publicity resulting from the debate in the AC, so what you suggest is democratic so long as it has to undergo the further hurdle of the popular vote in an election.

    2. Why do you want to further privilege the status of activists? — already (in the UK) there is the impression that overseas aid policy is being decided by Bono and Geldof and energy policy by lobbyists for renewable energy. I’m all for popular participation, but I don’t see why you would want to privilege someone simply because they feel strongly about an issue. And how do you discriminate between activists and lobbyists? In another thread Matteo opined that he would be happy with Greenpeace lobbying but not some corporate interest, but isn’t this rather subjective? (and lobbyists are pretty good at disguising corporate interests). Given that this will inevitably open the door to lobbying on a massive scale, I think it would need, as you suggest, to go through the medium of online voting before a letter was submitted to the AC. But even then, I think the lobbying would need to be addressed to the whole assembly, rather than to individual members. Besides which, a letter received by one member only would have no noticeable impact, unless you are proposing to empower individual members of the AC to seek to persuade their colleagues on behalf of the lobbying that they have received, in which case you run up against the legitimacy problem. Wouldn’t it be simpler to keep interests and judgment completely separate?


  44. Keith,

    Re: activists
    I admit to not having figured this one out yet…but my notion is that unless some means of engaging activists is incorporated within a sortition democracy, the activists will seek to overthrow the system. One version of the democratic principle, embraced by participatory democracy, is that if a decision substantially affects an individual, then that individual should have some say in that decision. Now sortition allows some random individual LIKE that individual to have a say, but not the specific person. You have offered continued elections as a means for activist participation (they will campaign hardest, run for office, etc.) and have their extra influence that way. I am trying to imagine how activists can be engaged and actually provide benefit to society in a non-electoral democracy. My concept is that activists who have given an issue lots of thought might be tapped for their ideas somehow, (rather than engaging in voter turnout /suppression campaigns and deception efforts).


  45. Yes I see the point, but surely the problem of the lack of direct participation in a sortive democracy by the vast majority of people applies to everyone, irrespective of how passionate they feel about their particular cause. And isn’t it the case that passionate conviction will always find a form of expression? (the converse also being the case, and my principal concern — the emancipation of the Average Joe).

    Although us moderns tend to see the elite in socio-economic terms, the Founders would have viewed it in quasi-religious terms (ie the elect); I’m concerned that we don’t just produce a new elite of activists and busybodies telling us all what to do. I think this is the real agenda of many deliberative theorists — the recent books by Michael Saward and Steven Elstub being particularly egregious examples. There is a danger that we overlook the second word in the phrase “deliberative democracy”.

    As for frustrated activists wanting to overthrow the system, isn’t that a matter for the police? As you can tell I’m not a big fan of activists, as my conservative prejudices incline me to the belief that there is little correlation between passionate conviction and getting it right. I suppose that’s the real reason why my arguments are ignored by most people on this blog (Peter and yourself being the exception that proves the rule!)

    Another outlet for activism might be via the house of independent advocates that I propose in my book. Although I view this more as a house of expertise, in practice it’s difficult to draw a hard line between the two concepts. For example, such an advocate house would include Greenpeace, on the basis that their membership passed a specified threshold. Take any other movement — for example anti-globalisation activists — if they wanted to be included in the debate they would need to organise along the same basis, thereby indicating that they had sufficient public support. Would that sound like a sensible balance between conviction and democratic principles? This is the paragraph from the book (p.137 where I outline (albeit in a rather vague way) how the advocacy house might be constituted:

    “Quotas would have to be specified for business and the trades unions, science and its environmentalist critics, transport advocates and countryside conservationists, religion and the military, education, law, health, finance, culture, sport and the myriad professions that might have something useful to contribute to the national debate.”

    Note that members of the advocate house get to speak in the AC direct, seeking to influence the outcome of the debate, so it seems to me that this could well include the direct influence that you are arguing for. I also argue that the advocate house should be a source of legislation analogous to Private Members Bills. I would be happy to rewrite that part of the book making it clear that the function of the advocacy house is to include not just experts but also representatives of activist movements that pass a minimum membership quota. But I think there needs to be some way of ensuring that activists are already in possession of widespread support before they get to address the AC. If strength of conviction was the only criterion then neo-Nazis and religious fanatics could also claim the right to address the AC.


  46. > my notion is that unless some means of engaging activists is incorporated within a sortition democracy, the activists will seek to overthrow the system.

    How are activists more engaged in an electoral system than in a sortition-based system? In both cases, the legitimate way to act is by attempting to inform and convince the public of certain ideas.


  47. Yoram,

    In a competitive democracy (Schumpeterian/elective) activists do not focus on persuasion. The thirst for participation by some members of society is satisfied through campaigns. Activists primarily seek to mobilize supportive voters, and suppress the vote of opponents. Persuasion is done by the media (whether Fox News, or paid ads, etc.). Even mere voting itself, provides a FEELING of participation. Voters FEEL they have had an impact (until someone points out that their vote didn’t ACTUALLY matter). This FEELING of participation lends legitimacy.

    A purely sortive democracy needs some participatory element at least as significant as elective democracy…and ideally, more real. One approach might simply be to have vast numbers of allotted bodies, focusing on discrete issues for short periods (even just a weekend), such that nearly everybody will serve on several decision making bodies at some point during their life.

    I feel that the participation question is the Achilles Heel of sortition, that needs a fix.


  48. > Activists primarily seek to mobilize supportive voters

    How is that done if not through persuasion?

    As for voting, this probably does give a (mostly illusory) sense of participation, but this is aimed generally at the average citizen rather than at the relatively narrow activist sector.

    Finally, I think that a proliferation of allotted bodies would, at best, provide a false sense of participation. I suggested participation in democratic media as a meaningful participatory activity. See here: Implementation of democratic mass media.


  49. Terry: “I feel that the participation question is the Achilles Heel of sortition, that needs a fix.”

    An interesting point, but I agree with Yoram and wonder if this is more from the activist/poliitico perspective (a tiny minority). I’m not sure if the average joe is really that bothered, so long as he feels that the government is being run competently and not against his interests. As to whether an allotted legislature would nourish that perception is an open question. And do people vote because they think they are making a difference or is this just a residual sense of civic obligation?

    As for the issue of persuasion, I’m inclined to put activists and the media in the same category. Lobbyists, activists and the media all seek to persuade/influence, and in a democracy there is no clear delineation between persuading the legislature and persuading the electorate.

    Having said that, no doubt activists and other busybodies will feel disenfranchised by sortition, but that strikes me as a thoroughly benign outcome!


  50. Yoram asked about mobilization under competitive democracy…
    “How is that done if not through persuasion?”

    I have a vast amount of experience with electoral campaigns, and can state confidently, that there is very little effort at persuasion with regards to issues in most campaigns. This is one aspect where allotment will normatively trump competitive democracy, in that it must be all about persuasion on values and policy directions. Current campaigns seek to elicit emotional reactions from voters. Campaign consultants advise candidates that in a campaign you should never seek to educate voters, but rather plug into things voters already believe, so they will be nodding their heads when they hear your ads. What little persuasion exists is about the integrity and character of individual human beings (the candidates) rather than policy.


  51. Terry, do you think the persuasion issue might be very different with an AC? If there was a decent time gap between the initial policy proposals and the meeting of the AC and if the deliberations were a) very public and b) quite extended in time, then one would hope that the temporal gap would be filled by persuasion (Condorcet’s argument). If the size of the AC was large enough then the persuasion would fill the media and the public domain (rather than private lobbying of individual members) and would provide an excellent opportunity for activists to pursue their goals without contravening democratic norms, as activists are surely in the advocacy/lobbying category and should not be offered any privileges denied to ordinary citizens (even if the latter don’t want them!)

    Given the problem of rational ignorance, I’m not at all surprised that persuasion is of very limited value in electoral campaigns for, as Wilde quipped, it takes far too many evenings.


  52. Terry,

    > there is very little effort at persuasion with regards to issues in most campaigns

    This is certainly true, and, as you write, that is a fault of electoral systems rather than an advantage. Such activity should be discouraged in a properly functioning political system. People who seek to engage in political activity should have other, more productive, avenues to do so.

    I was not referring to electoral campaign activities but rather to the activity of interest groups such as environmental groups, civil liberties groups, peace activism, etc.


  53. Yoram, why do you want to discourage persuasion by interest groups?


  54. > Yoram, why do you want to discourage persuasion by interest groups?

    I guess I wasn’t clear: persuasion is good. It’s the electoral campaign activity – that is not persuasion but manipulation – that should be discouraged.


  55. […] Martin, “A Senate Drawn by Lot” – following Alex Zakaras and linking to a post by Keith Sutherland on this blog, and “The Popular Branch” – following Ethan […]


Leave a Reply

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

WordPress.com Logo

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

Facebook photo

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

Connecting to %s

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

%d bloggers like this: