Two recent sortition advocacy pieces by Simon Threlkeld

Simon Threlkeld is a former Toronto lawyer (law degree from Osgoode Hall Law School), holds an MA in philosophy (University of Toronto), and writes about democracy. In 1998 he published an article in the academic journal Social Policy titled “A blueprint for democratic law-making: Give citizen juries the final say” whose abstract is below.

17 years later, Threlkeld is still a committed advocate for sortition, and has two recent pieces in the Canadian press advocating the use of sortition in order to democratize the Canadian government and media. In both cases Threlkeld is not proposing to use sortition to select office holders, but rather to use sortition to select committees that would appoint the office holders.

In September Threlkeld proposed in the National Post to have the Canadian Senate members appointed by randomly selected juries:

Simon Threlkeld: Select senators by jury

Athens in the 5th and 4th centuries B.C., often called the “cradle of democracy,” was largely run by juries chosen from the citizens by lottery. This kept a wide range of decision-making power firmly in the hands of the citizens. Juries can be used for the same purpose in modern societies, including for the selection of Canada’s Senate.

The Senate can be chosen by juries of randomly sampled Canadians who meet together face-to-face to make an informed decision after deliberation. Such a Senate will be independent from political parties, and chosen in a highly democratic, non-partisan and well informed way.

Now, Threlkeld is proposing in the National Observer appointing the board of the Canadian Broadcasting Corporation by randomly selected juries:

Let citizen juries bring independence and democracy to the CBC

The only way to end the patronage problem at the CBC is for the prime minister and politicians to be removed from the selection process.

Classical Athens sheds light on how this can be done in a highly democratic way.

Widely considered the birthplace of democracy, Athens was largely run by juries chosen from the citizens by lottery. Juries kept decision-making in the hands of the citizens, and can do the same for public broadcasting today.

The power the prime minister and his or her party have over the CBC can be transferred to randomly sampled juries of Canadians. These juries would meet together to make an informed decision after deliberation, and can be paid to work full-time for as many weeks as needed.

Random-sample juries engage a representative cross-section of the public, and provide the democratic ideal of informed rule by the people.

A blueprint for democratic law-making: Give citizen juries the final say

Simon Threlkeld, 1998, Social Policy, Vol. 28 Issue 4.

Abstract: The article discusses the role of juries or jury assemblies in giving the citizens a final say about laws. Juries are chosen by random selection because that is the best way to give a representative cross-section of the citizenry. Each citizen has the same chance and right to be chosen as any other. A jury is well suited for making an informed decision because the jurors can meet face to face and work full time for the days, weeks, or months needed to become fully informed about the matter at hand. Citizen groups can be allowed to bring proposed laws before a jury for a short preliminary hearing of the arguments for and against the law. After the preliminary hearing, the jury decides by majority vote whether to reject the proposed law or to refer it for full in-depth jury hearings. If a jury approves the law by majority vote after such full hearings, then the law goes into effect. Juries are suitable for deciding all types of laws at all levels of government, from municipal by-laws about smoking to national laws on the environment and taxes. Referenda are unsuitable for the informed decision-making needed for real democracy. In-depth jury hearings are an infinitely better basis for an informed decision than a referendum vote. An effective referendum campaign requires a lot of money to get the message out to the voters.

61 Responses

  1. His 1998 article sounds spot on, but I wonder what made him backpedal, so that randomly-selected juries should only make appointments rather than determining policy outcomes? It seems to me that he is misrepresenting Athenian democracy if he claims that it provides some sort of precedent for this kind of indirect democracy.


  2. Dear Keith, Thank you for commenting! Just so you know, I hold the same positions I set out in 1998 in Blueprint and elsewhere as I do now (possible minor revisions aside).

    I still support the use of sortition for direct democracy in certain cases (such as law-making, and making the decisions about public broadcasting now made by politicians).

    In1998 as now, I also also support the use of sortition for choosing some of the decision makers now typically chosen by politicians (such as the Canadian Senate and British House of Lords, and the board of the CBC).

    It is of course true that sortition was not used for indirect democracy in Athens (nor of course was it used for final decisions about laws, though of course the Council of 500 played a significant role in law-making). Although the applications of sortition in modern societies can (and in my view should) differ from those in Athens, it is of course from Athens that the sortition idea mainly originates and was most widely used.


  3. Simon,

    Good to have you on this Blog! One important correction, however, to your comment above regarding Athenian democracy…The final say about laws WAS made by randomly selected bodies, since a court could over-rule the Assembly on a law passed by the Assembly in the 5th century BCE, and as part of the reforms of the the 4th century BCE the Assembly gave its former law making authority over to randomly selected legislators (nomthetai) (while retaining the power to declare war, etc. to the Assembly)


  4. tbouricius, yes the courts can over-rule laws in the 5th century (and had a final say in that way, though I believe they did not make laws). Yes I failed to mention the very interesting law-making role of the nomthetai in the 4th century. I did realize that after posting, but saw no way to edit my post once it was posted.


  5. As tbouricius correctly points out, the first bracketed clause in the the last paragraph of my first comment in this thread is wrong. It should have been deleted.


  6. Simon,

    Thanks for the clarification. Most of the regular commentators on this forum are primarily interested in law-making by allotted bodies (it’s the subject matter of my ongoing PhD research), hence my interest in your 1998 article.

    >I still support the use of sortition for direct democracy in certain cases (such as law-making.

    This is primarily a terminological issue, but most of us agree that large randomly-selected juries entail a form of “descriptive” representation, for which we have devised the neologism “stochation”. There is some disagreement here as to whether such juries should play an active deliberative role (as in modern courtroom practice) or be restricted to listening to competing advocates and voting in secret (as with the Athenian nomothetai). I’m more disposed towards the latter, as it seems to me that individual speech acts will distort the representativity of the sample.

    >Referenda are unsuitable for the informed decision-making needed for real democracy. In-depth jury hearings are an infinitely better basis for an informed decision than a referendum vote.

    Agreed. I have a short op-ed advocating that the forthcoming UK referendum on EU membership should be replaced by a public enquiry with the outcome determined by a stochastically-representative jury:


  7. PS I can’t seem to access your Social Policy article via my university collection, can you provide any sort of link?


  8. Keith, I’m afraid I don’t know a link to it. Also in 1998 I wrote a short piece called “Let juries legislate” in The Next City (at least that’s my recollection of the title they gave it on the cover of the magazine), and an oped in the Hamilton Spectator on jury law-making (as well as two other articles about juries in contexts other than law-making).

    Here’s a link to The Next City piece (you have to scroll down a bit):


  9. Here is a direct link to “Let juries legislate”:


  10. If you scroll down below the said piece in The Next City you will also see me defending jury assembly law-making from objections. No attempt was made to argue against my defence. (Urban Renaissance Institute published The Next City magazine.)


  11. Keith, your article “Yes Lord Mandelson, we need a Lottery!” seems excellent to me.


  12. My 1998 oped in Hamilton Spectator about jury assembly law-making.
    “Referendums not ideal for public input on laws, taxes”: [Final Edition]
    Threlkeld, Simon. The Spectator [Hamilton, Ont] 16 Dec 1998: A11.

    Archive link to it (not sure the link will work):


  13. Here’s a direct link to my said defence of my position on jury assembly law-making in The Next City (January 19, 1999):


  14. Simon,

    Interesting responses — I think we are 99% in agreement. My principal concern is how to ensure that the “verdict” of each jury mirrors “the decision that everybody would have made under good conditions” (Fishkin, 2009), the problem being that three Deliberative Polls performed by Jim Fishkin for Texas utility companies delivered three different outcomes, so which is the representative one? I discussed this briefly with Bob Luskin, Fishkiin’s statistics guy and suggested that this might be on account of the arbitrary nature of the small group discussions (he did not disagree). Jim told me that the small groups are an essential part of opinion formation, but I think he is just kow-towing to mainstream deliberative democrats, who view the DP as too highly constrained (and who are not interested in representation). I don’t object to face-to-face deliberation in principle, I just think consistent outcomes are the most important factor if the jury decision is going to be accepted by the millions of people disenfranchised by the sortition process. I wish I could get more public interest in the jury as a referendum replacement. The Mandelson piece was written for the Spectator, but they cut it down to a short letter, which went below the radar.

    PS recommend you read the thesis discussed in this thread as it’s very relevant.


  15. Thanks Keith! Glanced through the link you provide, and will read more carefully later.

    On the word “enfranchise,” just a friendly comment about the use of the word with which you may agree. it is fine I think to use it the way you have in discussion between us, I get what you mean, and am not going to misinterpret or misrepresent what you said.

    Anyway, I think well designed sortition enfranchises the public, and that by comparison popular vote tends to disenfranchise us. I think you agree.

    Sortition enfranchises the public by giving every citizen an equal chance and right to be chosen, and because random sampling (perhaps especially with methods of stratified random sampling) gives women and others representation proportionate to their numbers. It also enfranchises the public by being a method of informed rule by representative cross-sections of ordinary citizens who’s have the same, or more or less the same, interest as the whole. Popular vote by contrast disenfranchises the public in ways I believe you are quite familiar with.

    Lets not let language be skewed against us.


  16. “Enfranchise traces back to the Old French word enfranchiss, a combination of en-, meaning “make, put in,” and enfranchir, meaning “to set or make free.” In the 1680s, enfranchise came to mean “to admit to membership in a state,” something that usually came with the right to vote. You may know the word disenfranchised, an adjective that describes people who lack rights or liberties. To enfranchise is to give or restore rights to the disenfranchised.”


  17. As for the Spectator, their loss.


  18. >Anyway, I think well designed sortition enfranchises the public, and that by comparison popular vote tends to disenfranchise us. I think you agree.

    Yes I do, but the key qualifier is “well designed”. It seems to me that there is a need to demonstrate by experiment that the decision output of an allotted assembly would be the same regardless of which empirical individuals (including you and me) are included in the sample. Fishkin’s experiments are well designed but they demonstrate that different samples return different decisions and this will be fatal for the perceived legitimacy of democratic decision making by stochation. So we need to find out by trial and error which aspects of the DP are responsible for the variation, and my hunch is that small-group deliberation is the culprit. This would be particularly problematic if the assembly had statutory powers, because the presence of trained facilitators to moderate the small-group sessions would be ruled out on account of the quis custodiet principle. The only other possibility is that the sample is not large enough (most statisticians would argue for a minimum of 1,000 to return a 4% margin of error, but this is getting very close to the rational ignorance threshold.)


  19. The fact that different judges have different opinions on the same matter, and a different trial jury might have reached a different verdict, does not mean we should replace trial by judge, and trial by jury, with trial by referendum.

    The question is, I think, which method of decision-making is best, or is on average best.

    Yes it is worth knowing how consistent mini-publics are in their decisions, and what factors might influence this. But it seems to me that absolute consistency is not at all required in order to clearly show they are better than for example a referendum vote.


  20. Simon,

    I agree. Sutherland has been repeating this line about the “consistency” requirement (which for some reason he proclaims only applies to allotted bodies) for a while but has never been able to explain why it is of such importance.

    Beyond being an arbitrary requirement, it is also not well defined. For example, if two bodies are making budgeting decisions and one of them decides to spend $130,000 on a particular project while the other body decides to spend $135,000 on the same project, does this count as the two bodies being “inconsistent”?


  21. The fact that the nomothetai may have reached a different decision if a different group or 500 or 1,000 or whatever had been chosen, or that the nomothetai “changed their mind” about (that is, had a different view from) what the nomothetai had decided in a previous year, does not mean the decisions of the nomothetai should have been transferred to the ekklesia or the areopagus or the boule or whatever. The question is not which way of deciding is perfect or 100% consistent, but rather which is best, or best on average.

    We don’t have to demonstrate a standard of perfection to reach a conclusion about what method of deciding is best.

    Also, consistency can be over-rated. If opinions are divided right down the middle on something, any fair and democratic majority vote on which way it goes is going to be more or less a coin toss. If there is about equal support for three mutually exclusive options, and most people are fine with any of the three, the decision in any fair and democratic process is going to be a bit of a dice roll no matter what. The point is, was the decision-making method as good as it could reasonably have been, were all the relevant facts presented and taken into account, were the deciders representative, were people’s views tested in discussion with others [;o)], and so on.

    Liked by 1 person

  22. Simon,

    I have no problem with the nomothetai model — my proposal for modern-day lawmaking is a nearly exact analogue of 4th-century practice. The problem is that most modern advocates of “deliberative democracy” wrongly claim classical-era Athens as the template for their own project. As Daniela makes clear in her thesis, there was nothing “deliberative” about the workings of the Athenian jury — their job was to vote, not to talk. The advocates (who did the talking) were either those who were prepared to risk all by proposing a new law or else were elected by the assembly to defend the existing law — this was a high-risk role for those involved, in stark contrast to that of the randomly-selected jurors (who enjoyed the prerogative of the harlot).

    I am not seeking to compare an idealised model of decision-making by representative microcosm with other flawed institutions (referenda, mass assemblies, mass elections etc). The representation principle (the law of large numbers) underlying stochation only applies at the aggregate level, not at the level of the individual persons selected by lot,** so the unaccountable speech acts of randomly-selected individuals do not benefit from any kind of democratic mandate. This doesn’t bother most epistemic and deliberative democrats (including some commentators on this forum) as they scornfully dismiss such matters as kow-towing to the discredited tropes of liberal democracy. I don’t include Yoram in the above (he has no interest in epistemic outcomes) but he bases his claim for the representativity of small groups on the shaky foundation of a deductive syllogism.

    PS given the margin of error in the sampling process (4% @ 1,000?), this would suggest that a simple majority would be insufficient. At this size of jury anything less than (say) 55/45% would require a retrial with a larger jury. I’m not very good at the math but the principle would be that it makes no difference who is in the jury, the decision outcome would be the same. And I made no reference to diachronic consistency, merely that the decision taken at any one time would need to be representative of what everyone would think at that time. Just as parliament cannot bind the hands of future generations, all laws passed in such a manner would be open to repeal/modification in the future.

    ** While the group represents the target population “descriptively”, no such claim can be made for individual members, and the illocutionary force of individual speech acts is anything other than equal. This being the case, different samples could generate very different decision outcomes, depending on the convictions and persuasive abilities of the individuals who happened to be in the sample. This would be random is the pejorative sense — like as in the “postcode lottery” for NHS health-care provision (it can be better or worse, depending on where you happen to live). It’s interesting to note how a single principle such as random selection can have good or bad entailments depending on the level of analysis employed.


  23. >PS given the margin of error in the sampling process (4% @ 1,000?), this would suggest that a simple majority would be insufficient.

    The math is at the very edge of my pay-grade, but I suspect that’s actually at a 95% confidence interval. So a 95% chance of drawing a sampling with preferences on a particular bill coming within 4% of the population preferences. We should probably make a distinction between the percentage that the sampling deviates from the population and the odds of passage of errant legislation. We aren’t holding an opinion poll, so vote totals in the sampling that reflect the ideal, population-level, support are not really necessary.

    For Simon’s sake, I would like to reiterate my support for a two-vote system. Hold a vote in the sampling. If it comes out with over 50% support, wait a few months (or however long it takes for the sample to turn over completely) and hold another vote. If it passes that vote, then it is enacted into law. If the odds of passage of a bill that should NOT pass (given population-level support) are 10% in any given single sampling, they fall to only 1% in two successive samplings. Likewise, if the odds of passage of a bill that SHOULD pass (given population-level support) are 90%, they fall to 81%. Both represent the same interval around a population-support of 50%. The exact interval would depend on the size of the sampling of course. This is a perfectly respectable trade-off, in my opinion. To ensure we can still pass urgent bills we can have an arbitrary supermajority threshold which would make the second vote unnecessary. Since it’s unlikely the members of the sampling will know how their peers will vote, they will not know whether the outcome will be within a few votes of either the supermajority threshold for immediate passage or the 50% threshold which would kill the bill outright in the first vote. So there should not be a rational ignorance problem in either sampling, especially when you consider the fact we can get away with a much smaller sampling. Both supporting and opposing advocates should show their hands (rhetorically) in debate before the first sampling, and advocates will have months to come up with effective counter-arguments, so debate in the second should be more rigorous than is possible in the first (or in a single-vote system).

    While I agree that consistency is important in principle, I’m starting to doubt the practical importance. Few people have an intuitive grasp of statistics. I’ve known fellow chemists with a remarkable lack of basic knowledge in this area, which is kind of embarrassing since the very heart of chemistry is statistical mechanics.

    I’m also not sure why you care so much, given your distaste for democracy. If a deviation from fully democratic decision-making in the sampling results in serious public opposition/delegitimization, then public opposition/delegitimization will be the result no matter where in the overall system the deviation is found. At least there’s no chance of a systematic bias in a statistical margin-of-error. If a systematic bias is introduced elsewhere, then surely the public opposition/delegitimization will be far greater.


  24. Naomi,

    I think even your stochastically-challenged colleagues understand that an election is won or lost on the basis of the votes counted. If they are in the losing camp then they will (reluctantly) accept the outcome so long as the election is, formally-speaking, free and fair (i.e. no hanging chads etc). If the result is the implementation of laws they disagree with, they will nevertheless reluctantly consent, as they understand the concept of majority, plurality or whatever.

    But if, as the result of the aleocratic revolution, they lose the right to vote then surely they will demand safeguards that ensure a free and fair outcome? That all samples (from which they will almost certainly be excluded) return a consistent result? I would endorse the mechanisms that you outline to reduce the margins of error (double votes, supermajorities etc), but I think structural constraints need to be put in place to ensure accurate representativity as well. And my hunch (it’s no more than that) is that this would exclude face-to-face deliberation in the jury. That’s certainly what the Greeks believed and they did have hands-on experience in this matter. It may also have influenced Harrington in his view that any juror who spoke should immediately be taken out and hanged, but perhaps that’s a little extreme.

    >I’m also not sure why you care so much, given your distaste for democracy.

    Yes, I’ve asked myself that question. It’s not that I have a distaste for democracy, but that I believe in mixed government, that includes a role for the one (the magistrate), the few (the advocates) and the many (the demos). What I am deny is a role for randomness per se — random selection should be used for its stochastic and prophylactic benefits, not as an end in itself. Incorporating “ordinary people” into the conversation only has merit if their input is strictly in proportion to the representativity of their views — absent that then it should not be encouraged. Remember that in my (our?) scheme, anyone (including ordinary people) can make a proposal but only those who can demonstrate an adequate level of support for their views (by election or public votation) get the right to address the legislative jury.


  25. Keith,
    two points…
    1. you referenced lottery selection of legislative jurors in place of elected legislators means the citizens “lose the right to vote,” which they would view as a loss.We can’t know what perception would dominate, since many citizens would feel they had GAINED the equal right to participate in a final decision. which they never got in an electoral scheme for representatives. Clearly the majority PARTISAN segment of the population would feel aggrieved, but since MOST citizens are minimally or NON partisan and the minority partisans would be better off, my guess is that a vast majority would prefer lottery selection of legislators that could include them.

    2. You wrote:
    >” Incorporating “ordinary people” into the conversation only has merit if their input is strictly in proportion to the representativity of their views — absent that then it should not be encouraged. ”
    That is ONLY the case if the disproportionality were consistently biased in one direction. (remember we are talking about those who offer advice to the decision-makers). The principle of isegoria is not about proportionality, but that ANY citizen should be able to offer advice and offer proposals (though obviously it may be dismissed by the decision makers). And it may be that one particular citizen has one key piece of information that would make all the difference in some decision…It is the worth of the thing the person says that matters, not what segment of society they come from (already had one too many from that ethnic group). The ADVISORS should be both experts randomly selected from a qualified pool, and randomly selected non-experts, with the addition that ANY citizen should be able to submit testimony that can climb to the highest level by some sort of ranking (or thumbs up and down) algorithm.


  26. Terry

    In both cases you are choosing to ignore the problem of scale. In ancient direct democracies anyone who wanted to perform a political office would stand a high chance of being selected, and anyone could also speak to the assembly. In large modern states the chance of winning the lottery is infinitesimally small and if everyone who wanted to could speak then you would end up with the tower of Babel. Hence the need to rely on representative mechanisms.


  27. 1. Keith, are you possibly also opposed to citizens discussing the candidates in a popular election, or the issues in a referendum vote, with other citizens (were it possible to stop that), or is it only jurors you possibly don’t want to talk to each other?

    2. Do we know there was no discussion of the proposed laws by the nomothetai? Were proposed laws not posted ahead of time (weeks ahead of time) where all could see them, and would not nomothetai (who could soon sit in judgement of those proposed laws) be likely to discuss them with other nomothetai, dikasts and citizens? Did the nomenthetai not take breaks when serving during which they might well discuss the laws they were sitting in judgement of? Did they not break for a mid-day meal, and did they not take such meals in common (I’m not sure if they did, I am just asking, but I seem to recall Aristotle talking about common meals as being part of democracy)?


  28. 1. Should legislative assemblies have to have a super-majority to pass legislation? Given that the answer is no, why should legislative juries, which are far more representative of the public than any legislature, have to?

    2. I think the basic democratic idea of the equality of citizens entails that legislative juries decide by majority vote rather than by a super-majority vote.

    If super-majority voting is used it means that the minority is systematically advantaged with each of their votes counting for more than each of the votes of those in the majority (blatantly contrary to the equality of citizens). A systemic advantage is given to minorities to stop proposed laws, something which moves away from democracy and towards rule by minorities.

    If an exactly representative legislative jury would result in 52% voting in favour and 48% voting against, then with a random sample of say 500 or 1,000, the 52% majority would probably still win (maybe with 54% or 53% or 51% of the vote). It is true that though the 48% minority would probably lose, there is still a chance they could win (I’m not sure of the math but maybe with a jury of say 800 a 15% chance or something like that). Is this really a significant problem?

    Here are some reasons why I think it is not:

    1. There is a reasonable trade-off between accuracy and not spending the money to have legislative juries number in the tens of thousands. Legislative juries, even at just 500 or 1,000, will still be better than other options for deciding, including in terms of accurately representing the pubic.

    2. Majority vote makes any inaccuracy the luck of the draw, unlike a super-majority requirement which systemically builds inaccuracy and inequality into the system (it loads the dice in favour of minorities). With the luck of the draw type of inaccuracy the public are on a level playing field and citizens are treated as equals, and their is a fairness to it, unlike with the super-majority requirement where the playing field is skewed in favour of those in the minority.

    Analogy: Six soldiers are ordered to cross a field in which they know there might be mines. They correctly decide that the best way to minimize casualties is to cross in single file, so that if the first soldier is blown up those following will be safe continuing on the path past the exploded mine. How shall they decide who leads the single file? One fair way would be to number off from 1 to 6, and then let a dice (die) decide which of them goes first. But what would not be reasonable would be for example for soldier 3 to say, I have a loaded dice here which falls on 5 and 6 more than on other numbers, let’s use that dice! This would be rejected as contrary to equality between the soldiers, as unfair, and as not being a level playing field. Same problem when you load the dice or stack the deck in favour of minorities by requiring a super-majority.

    3. Why does democracy require that the 48% of the public who hold one view have to lose 100% of the time instead of just 15% of the time (or whatever exactly it is)? Why should their views have to be guaranteed to count for nothing 100% of the time when it is so close – what’s wrong with them having a 15% chance of winning, and the 52% majority having an 85% chance of winning? When it is so close that it is almost a tie anyway, why be so fussed about which way it goes – is it not enough that the 52% majority will probably win? And when it is so close none of us will likely know which side has the edge (that is, know which side would have the wafer thin majority were all the public to become well informed and vote), so what do any of us even care as we don’t know whether we stand to lose or win from a slight inaccuracy in the random sample?

    Liked by 2 people

  29. Simon,

    >1. Keith, are you possibly also opposed to citizens discussing the candidates in a popular election, or the issues in a referendum vote, with other citizens (were it possible to stop that), or is it only jurors you possibly don’t want to talk to each other?

    Only the latter, as jurors are performing a (descriptively) representative function and any potential breach of aggregate representativity has to be minimised. Citizens who only represent themselves are free to chatter as much as they wish.

    >Do we know there was no discussion of the proposed laws by the nomothetai?

    Any speech acts from the jurors were frowned on — disparaged as “thorubos” (uproar). Everyone is allowed a lunch break and people will naturally talk amongst themselves, but potential breaches of aggregate representativity (of the sample to the target population) should not be actively encouraged by an institutional design that arrogates a key role to talking.

    >I seem to recall Aristotle talking about common meals as being part of democracy?

    This is probably a reference to an article in Political Theory by Jeremy Waldron, in which he claimed that Aristotle’s concept of the “wisdom of crowds” was based on the analogy of the pot-luck supper. Daniela argues in her thesis that Waldron has mistranslated the relevant passage from Politics (this chapter was also republished in Political Theory)


  30. Simon,

    >why should legislative juries, which are far more representative of the public than any legislature, have [a supermajority]?

    Because the stochation mechanism involves a margin of error, which could mean that different samples would return different verdicts and it would not be possible to say which verdict was the representative one. The fact that (unlike a legislature) the sample would be made up of “ordinary” or “typical” persons is beside the point as elected represenatatives are chosen by voters. When I choose my solicitor I don’t choose a person who resembles me, I choose one who I think will best act in my interests, so the representative principle is a different one (“active” rather than “descriptive” in Pitkin’s terminology). Unfortunately most sortition advocates conflate these two very different representative principles.

    >If an exactly representative legislative jury . . .

    Given the nature of stochastic processes how would this be possible? Such a jury would have to include the whole target population. As for the various numbers that you crunch, my only concern is for those who have lost their vote and are not included in the sample. It’s no consolation to them to learn that the decision was taken by a sample of ordinary Joes like them, they need to know that the decision would have been the same whether or not they had participated themselves.

    The problem is best understood in Rousseauian terms — how is it possible for us all to live together and at the same time be free? Rousseau argued that it is by uncovering/discovering the general will and realising that it was also your own will. Although he accepted that this could be determined by a majority vote, he nevertheless insisted that all citizens should participate in the formation of the general will on all matters deemed worthy of law-making. But he acknowledged that this was impossible in large modern states, but I think that this is possible by stochation — but only if everyone can understand that their own presence of absence makes no difference to the determination/uncovering of the general will. And that presupposes consistency between different samples and, if necessary, statistical measures to ensure that the margin of error is such that it makes no difference to the outcome.


  31. Re your response to 1: A mini-public is a microcosm of the public, more or less, and in my view close enough (close enough in my view if it is a reasonably large random sample). If deliberation helps them reach a more informed decision, which I think it does, then they should deliberate. The point is, it seems to me, that they are to express the informed judgement of the people, meaning the judgement the whole people would reach were it possible for all to to meet together, to work full-time to become well informed, and to break into small groups for deliberation to help them become better informed, and so on.

    Just as discussion of ballot measures between citizens during an initiative campaign can help citizens cast a more informed vote (leaving aside the fact that initiative votes are generally, in my view, undemocratic methods of uninformed rule by the people), so too can discussion and deliberation between members of a mini-public.

    I am skeptical of your saying jurors perform a “representative function.” I think it is more accurate to say that they are representative (by virtue of random selection) than that they perform a “representative function” (whatever exactly that means).

    Because a mini-public is representative, those on it should represent themselves! It is exactly by each of them representing his or her self that collectively they are representative of the people. I agree with your statement that “citizens who represent only themselves should chatter as they wish [that is, discuss and deliberate, preferably in the context a process and set of procedures designed to ensure an informed result]” but perhaps unlike you believe that representing only themselves (or only his or her self) is exactly what each juror should do.


  32. Perhaps it would be helpful to take a few steps back for a minute.

    If all the parties to a conflict agree to binding arbitration, it doesn’t matter what form that arbitration takes. What matters is that the arbitration is agreed to. This is basically what politics is… policy conflicts between factions in the citizen body being resolved through an arbitration process agreed to by broad consensus. Things get a bit fuzzy, because in this case the participants in the conflict are also the arbiters. But then again, actually, that’s only in principle. In practice we actually do appoint arbiters, who we judge later. In principle, if the degree of separation between the judgement of the people and the judgement of the arbiters we have now is greater than the separation that would exist in another method, there shouldn’t be a legitimacy problem. There may still be opposition from those who feel their right to participate was taken away, but I don’t think that objection will be muted no matter how tight the results are.

    If the ultimate measure of whether a policy should be enacted into law is the *ideal* support of a majority of the people, then there will always be doubt. If you strongly support a defeated measure, that lingering doubt will exist whether the chances of your measure being defeated by a random fluke were 1% or 10%. That doubt would be greater if the chances are greater, of course, but fear and doubt are not rational things. We fear shark attacks when we swim in the ocean, but not the drive to the coast, even though the drive is much more dangerous. Same with flying. We see it in our personal lives, our family situations… doubts that have a way of nagging at us and growing beyond all reason. The same danger exists here. Does the exact percentage matter all that much when it comes to irrational doubt? Can we even make that judgment based on what we know now? What if the media makes a habit of pointing out the ambiguity everytime the results of a vote are discussed?

    The chances that the agreed arbitration rules were followed is 100%. Assuming they actually were followed, of course. There’s no ambiguity there. If the rules are really agreed to by consensus, there should be no real ambiguity to that fact and following those rules is, in a sense, democratic. If we all decide we should decide things a certain way, using a different method is in some ways non-democratic even if the different method is more ‘democratic’ in other ways. It is, perhaps, convenient to model appeals to the people and following agreed to rules as different legitimacy modes. But that’s something of a stretch, because an appeal to the people is a rule we agree to by consensus. If we did not have an ideological commitment to democracy, as we often see in Africa and other developing parts of the world, we would care much less. Anyway, the two modes can be very much complementary, with the rule-following mode making up for ambiguities in the samplings. The extent to which tightening the odds up is necessary from a legitimization standpoint perhaps depends on the extent to which rule-following alone is inadequate. Tightening them as much as possible further legitimizes the system. So it is probably very desirable, in any case.

    The possible benefits of randomness in the results were brought up by Simon a little while ago, and I have some thoughts on that. Any policy-platform that lasts would need to be supported by a sufficient supermajority to begin with. Finding common ground is thus incentivized, but not forced. Which is nice. In some ways this is like conventional Lijphartian consensus democracy, of which I am very fond. I agree that the majority should not always get its way on everything all the time, but usually I favor the negotiation and leverage that comes from giving the minority a proportional share of political power. The big downside of occasionally giving the minority a majority is that some actions are not reversible. If a minority wants to go to war, and gets lucky in the selection process, everyone is bound to the consequences of their actions. We can split the difference, and have a two-vote system with small enough juries where a large minority has a modest chance of winning in one of the votes. I suppose that would kinda sorta be a supermajority requirement. But not really, because you could still aim at 51% and prevail often enough. Finding common ground would be incentivized because it would improve your odds of passing your proposal. I could live with that.

    Let’s consider for a moment the Texas deliberative polls that Keith brought up. If the people have strong feelings on a matter, it is unlikely small-group discussions (or full assembly debate for that matter) will move opinion very far. I don’t imagine people have very well-thought-out views on energy policy. In that case there’s no well-defined popular will that needs to be respected. The sampling forms it’s own will. If the process is perfectly representative, we would expect it be consistent regardless of the sampling. However, the debate will drive the formation of a will within the sampling. If you change the debate, you change the outcome and the sampling’s will may well coalesce around some other position. To get constancy, we need not only a small margin of error in the sampling, but also consistent advocacy. That’s why the deliberative polls were inconsistent, no? Different small groups having different internal discussions resulting in different outcomes. Thus the construction of the advocacy process demands a similar degree of rigor. But that’s impossible. Different people have different perspectives and will argue differently. We have to rely on something else besides consistency in the advocacy and acknowledge, by extension, that the will that forms in a particular sampling with a particular advocacy may not necessarily reflect the will formed through a different sampling with a different (but similarly constructed) advocacy, even if you were able to get the margin-of-error in the sampling down to zero.

    That’s a lot. I type all my comments on my phone, so I apologize for the long list of errors that are surely up there.


  33. Re speech by nomothetai being frowned upon, I believe that only applies to when they were interrupting and heckling those presenting the case for or against the law they were considering, not to any speech and discussion between them at breaks, nor prior to the hearing.

    Re common meals, I mean actual common meals. The prytaneis (so I read) had lunch and supper together during the time they served. Maybe it was the same with nomethetai (I don’t know the answer, but in any case it would have been natural for them to do so informally even were it not formalized).


  34. Simon,

    >Because a mini-public is representative, those on it should represent themselves! It is exactly by each of them representing his or her self that collectively they are representative of the people.

    That presupposes that the speech acts of each person have equal illocutionary (persuasive) force. In practice this is far from the case, because some people are a lot more persuasive than others on account of differentials in perceived social status, shyness, gift with language etc. This is exacerbated by the fact that the sample body is made up (in effect) of conscripts, most of whom will have no knowledge or experience of the legislative topic under consideration (voluntarism skews the sample to an unacceptable degree). This means that all the care that is taken to ensure the representativity of the sample will be undone at a stroke. A public opinion poll is only representative at the aggregate (statistical) level, what you are suggesting is equivalent to privileging individual survey forms.

    PS nobody is disputing the epistemic benefits to be gained from face-to-face deliberation, the problem is that democratic representativity requires balanced information input and the best way to achieve this is to limit speech rights to the opposing advocates (as was the 4th century Athenian practice).


  35. Keith, I fail to see how trying to define elected legislatures as “active” representatives rather than “descriptive” representatives in any way excuses them from their failure to legislate in accord with the informed judgement of the people, or why they should not be judged on that standard just as much as minipublics should. It seems plain to me that we should judge both minipublics and elected legislatures on how well they express the informed judgement of the people, something minipublics do much better. (Thus, if you believe in supermajorities to prevent inconsistency with the informed judgement of the people in minipublics, it is hard to see why you would not believe in it for elected legislatures where the problem is far greater. I, as said, am opposed to supermajorities being required either for minipublics or elected legislatures.)

    Liked by 1 person

  36. Keith, I accept there is a real issue with “persuasive Individuals” having a disproportionate influence in discussions and deliberation. This problem also exists in popular election and referendums, where persuasive friends or relations might convince someone of something, or where (perhaps more disturbingly) large “corporate persons” like Fox News or CNN or the The Sun (U.K.) might persuade people of things.

    I would not say that “representing themselves” in a minipublic presupposes equal speech rights, but I of course see the point. Also, as just said, I think this is just as much a problem in popular election and referenda, and in fact I think it is probably worse there (Fox News and Sun being perhaps especially blatant examples of why).

    I do think the “persuasive deliberator” problem needs to be addressed, but do not think eliminating deliberation to do it is likely to be a net gain.


  37. Keith, I think that though speaking in plenary before the nomothetai was limited to the advocates, that there was likely plenty of discussion in small groups before the hearing, and on breaks during it, (and also I assume in plenary earlier in the ekklesia when the referring of laws to the nomothetai was decided, presumably in part by nomothetai themselves, in their capacity as members of the ekklesia).


  38. Simon,

    > It seems plain to me that we should judge both minipublics and elected legislatures on how well they express the informed judgement of the people, something minipublics do much better.

    Yes. That is, the standard is “how democratic is the legislature”.

    > I do think the “persuasive deliberator” problem needs to be addressed

    It seems to me that as long as the deliberator is indeed persuasive (making persuasive arguments), rather than an effective manipulator, then there is no problem. As for manipulators, I would presume that the allotted would be able to handle those both by using general procedural safeguards and by identifying problematic individuals and treating them accordingly. As you note, this is something that is much more difficult to do in the context of an electoral or a referendum-based system.


  39. I think the discussion above is wonderful. I find myself generally in agreement with Simon (though I like Naomi’s proposal of two smaller juries with some period of time between them simply because of the likelihood of it resulting in better decisions). One point where I lean to Keith’s view (though for completely different reasons than his) is the value of silent internal deliberation. To be clear, I think a first mini-public should DEFINITELY engage in active deliberation, digging up and sharing all the diffuse knowledge of the members, challenging expert witnesses, and the like. But there are SOME problems with this as the SOLE method. Simon stated “If deliberation helps them reach a more informed decision, which I think it does, then they should deliberate.” Yes… BUT. There are also serious HARMS that can arise from such active deliberation…Group-think, deference to high-status members, and information cascades (where members do NOT offer their own knowledge because it is more pleasant to go with what appears to be the majority.) As many “wisdom of the crowds” experiments have shown, in SOME situations the BETTER decision is made when each member votes independently and is NOT influenced by what that member thinks other members’ opinions are. This independence requirement not only underlies the Condorcet Theorem but has powerful psychological bases. If active deliberation seems to reveal a certain trend, many members will suspend their own cognitive efforts at assessing information and defer to some others… the same as when a hundred people are playing a game of tug-of-war researchers have shown that most members do not pull as hard as when only six people play. All this leads to my proposal published here:
    of a multi-body system where one mini-public actively deliberates, amends and perfects a proposal, and a separate mini-public only listens to pro and con presentations and votes secretly without external deliberation. Only with two separate bodies can we get the best of BOTH models.


  40. Simon,

    I am not seeking to defend electoral representation; my concern is only that our proposals for sortition should not introduce additional distortions to the representative process. It’s clearly true that elections are distorted by all kinds of persuasive forces, but every sovereign individual has the right to make up her own mind. In a stochation-based system the overwhelming majority of people will not be involved in any shape or form — they have no rights to do anything at all — so it is vital for the integrity of the system to ensure that nothing distorts the relationship between the proxy (the allotted body) and the population that it seeks to represent. I’m not against face-to-face deliberation in principle: as far as I’m concerned the dispute can be resolved by an experiment that has (say) a dozen samples of 300 randomly-selected citizens of which six might simply listen to balanced advocacy and then vote, whereas the other six could engage in face-to-face deliberation before voting. If the decision outcome of the second set shows no more variance than the first then my hypothesis is disproved. I think also historical precedent is on my side — at least that is what Daniela has concluded with her study of the Athenian courts, which did not indulge in institutionalised face-to-face deliberation. Athenian society was intensely disputational, so if they saw fit to ban talking they must have had a good reason (what people do in their lunch-break is their own business).


  41. Naomi,

    The excellence of your two-thumb typing skills is only matched by the clarity of your thought! The DP makes a categorical distinction between balanced advocacy (at the information stage) and face-to-face discussion (at the small-group stage). In the Texan utility DPs I believe that all three groups had the same formal advocacy. Ensuring the representativity of the formal advocacy is a non-trivial problem, but would be open to a (partisan) political solution — easy, for example, in the case of my proposal for a DP to replace the forthcoming EU referendum. Getting the second right is simply impossible as it’s just . . . random. Hence my hypothesis, but I would be glad to put it to the test, if someone could come up with the funding.

    I’d love it if someone chose to respond to my Rousseauian rambling. It occurred to me when I woke up in the middle of the night and that’s a bad time to come up with a coherent argument. The nice thing is that it focuses on the equal freedom of the non-participants, not just the lucky winners of Willie Wonka’s golden ticket.

    I will always defer to Terry’s understanding of the epistemic benefits of deliberation in a diverse group — these are undeniable. But his model is essentially unpolitical — on an earlier occasion he opined that the “political” problems would have to be resolved first before the problem-solving begins. I’m half way through Scott Page’s book and I think this is his view on the need to resolve “perspective”-based issues beforehand. Is that right Terry?

    Liked by 1 person

  42. Keith,

    Societal decision making will never be so neat as to separate “political” problems from problem-solving towards agreed upon goals. Some will bring hidden agendas to nominally problem-solving sessions. My goal is to facilitate the best decision-making in the general interests of a community that is possible while avoiding the consistent bias that results from electoral selection of representatives or mass referenda. I don’t know about Page’s view on this…I took an online course (MOOC) he taught and read two of his papers but NOT his book.


  43. Scott Page deals with this on pp.11-13 of The Difference (emphasis original):

    “Preference diversity can and does [create conflict]. For this reason, management books stress agreeing on a common goal — a common fundamental preference. If people disagree on what they’re trying to accomplish, they function purely as a collective. . . diverse fundamental preferences frustrate the process of making choices

    This strikes me as close to the claim you made some time ago that fundamental preference issues have to be determined before the epistemic problem-solving process begins. But of course politics is all about disputes over fundamental preferences, and this is probably why both Page and Surowiecki appear to believe that cognitive diversity is of value to everything else but democratic politics. If we want to use stochation to resolve disputes involving disagreement over fundamental preferences then the yes/no voting of the nomothetai is the model to follow, not the face-to-face conversation valorised by deliberative and epistemic democrats (that has no precedent other than Burke’s idealised model of the House of Commons). This is the take home message of Daniela’s thesis, which should be required reading for anyone seeking an Athenian precedent for their sortition model.


  44. Keith,
    I can’t guess what percentage of political decision-making is a matter of fundamental disagreement about goals, and what percentage is a disagreement about methods of achieving common goals. I contend that a lot of contemporary politics is not about EITHER, but about power of office holders (who exaggerate differences for electoral reasons). THIS portion of “politics” would not exist in a purely sortition system since all decision-makers will soon be out of power. Real public decision-making can benefit from cognitive diversity EVEN IF there is a disagreement about goals among participants…the subsets that DO agree with each other can formulate better proposals with diversity AND even from the criticisms of those with different goals.


  45. Terry,

    I guess our disagreement originates in our differing perspectives on the origin of the political — to you agonism is the artificial product of differences manufactured for electoral purposes, whereas to me voting originated as a less costly way of counting the strong right arms of opposing armies. If the outcome of the battle can be computed in advance then you might as well just vote — then the losing side can live to fight another day. Those of us who believe this to be true become anxious when faced with any attempt to replace political democracy by the epistemic variant. I think history is on our side, whereas you have to rely on an appeal to heaven (i.e. a normative, if not utopian, project). Remember again that Page and Surowiecki claim that cognitive diversity/wisdom of crowds does not pertain to political decision making.


  46. Keith, first a comment that may be more about how things are said, than any disagreement between us on substance.

    1. Re your statement “In a stochation-based system the overwhelming majority of people will not be involved in any shape or form — they have no rights to do anything at all”. I don’t think that’s a reasonable and correct way to look at it or characterize it. I’d instead say, the public have something much better with a well-designed system of jury assembly law-making (such as the one I outline in my three above mentioned 1998 articles) than with referenda and the initiative, or with law-making being a monopoly of politicians. The right to cast one vote of millions in referenda is far less important a right than what is gained with jury assemblies. With jury assemblies citizens have (and we can say have a right to, if you like), law-making grounded in the informed judgement of the people, law-making on a level playing field (unlike initiative and referendum that are heavily skewed in favour of wealth and power), law-making that embodies the equality of citizens (unlike popular vote where younger citizens and others are under-represented among those who vote), and an equal right and chance to be chosen to serve. Instead of saying “the overwhelming majority of people will not be involved in any shape or form — they have no rights to do anything at all” the reality is that the public are far better served by good arrangements of jury assembly law-making, and that’s what we should say. It is referenda that are (in comparison to jury assemblies) in general (most of the time) contrary to the public’s interests and rights.

    The fact that we have trial by jury rather than trial by referendum, is not reasonably viewed as being about the public “have[ing] no rights at all” or “have[ing] no rights to do anything at all”. but rather as the public having the right to trials being decided by a jury “of our peers” who heard the evidence and deliberated about it, rather than by an ill-informed referendum vote by people who never so much as set foot in the courthouse or heard a minute of the evidence. Far from trial by jury being a matter of the public “have[ing] no rights … at all” or “have[ing] no rights to do anything at all,” compared to trial by referendum, it is exactly the opposite – it is trial by jury that gives us the rights that matter, including the right to trial by jury, the right to a fair trial, freedom from unjust conviction, freedom from arbitrary imprisonment. (So far as I’m aware, no one has ever complained that trial by jury is problematic because “the overwhelming majority of people [unlike in a referendum vote] will not be involved in any shape or form — they have no rights to do anything at all”. Such a complaint would in any case be absurd, for the reasons mentioned, and it is the same for a similar complaint about a well-designed system of jury assembly law-making.)

    Home delivery postal service has been partly phased out by Canada’s previous Conservative government, with people having to walk to “community mailboxes” with their key. We would not say that home delivery postal service deprives people of “the right to do anything at all” (meaning the right to walk to “community mailboxes” to get their mail), but would instead say, which is what the reality is, that ending home delivery deprives people of the right and convenience of home delivery. Similarly, using only referenda and initiative deprives people of the much better arrangements for citizen law-making that jury assemblies can provide.

    2. I’ve already spoken to why I don’t share your concern about exact consistency by all the minipublics that could have hypothetically been chosen to decide on a law.

    3. I’ve already spoken to why I am skeptical of the view there was no discussion by nomothetai of the laws they considered. In addition to the considerable opportunities for discussion I mentioned, was there time for discussion after the advocates presented their cases – were nomothetai given time to ponder their votes, (during which time they were free to discuss the proposed law with other nomothetai)?

    It may well be that the decisions of nomothetai were mostly based on their individual assessments of the cases presented to them by the advocates. But, except when the advocates were talking, far from a “ban” on discussion, there was a lot of opportunity for it. (How extensive such discussion was I of course don’t know, and so far I’m not sure if anyone does. If the nomothetai took common meals while serving, which I don’t know, I would not say there was no “institutionalized” discussion. Common meals were part of the “institutionalized” discussion of the prytaneis, I think it would be fair to say.)

    Liked by 1 person

  47. Simone,

    Well said. your point that “it is trial by jury that gives us the rights that matter,” is important. As a citizen, i don’t want the BURDEN ((“right”) to have to learn about a large number of issues, or even a large number of candidates, knowing my one vote is de minimis. I want the RIGHT to have people like me (and many others with different perspectives) to dig in to each issue seriously and make a decision that the majority of citizens would probably make given time, resources, information and motivation. I don’t even care if each and every decision is “correct” or optimal as long as there isn’t a consistent bias built in. Simon’s point that the equal chance to exercise power as a member of a jury is part of the value but the bigger right is the right to have decisions made by others in an informed unbiased, less-corrupted way.

    As to the possibility of informal discussion at a legislative jury in Athens… it hardly matters whether they talked to each other, since we have learned a lot about good decision making processes since then. But as mere historical interest, I think Keith is right that the INSTITUTIONAL structure discouraged discussion and expressions of opinion by jurors. At the conclusion of the pro and con presentation, I don’t know if they hung around and chatted, but the voting was conducted in a carefully designed procedure to hide the vote of each member (the metal disks with a solid or hollow center for yes and no held between the thumb and finger and dropped into the voting urn so that nobody could see whether the yes or no disk was deposited.)


  48. tbouricius, yes, agreed on what the important rights and interests at stake are (and being one of millions of referendum voters is not one of them).

    Secret ballot seems a sensible part of Athenian legislative jury decisions.


  49. Simon,

    Your point No.1: You don’t need to convince me regarding the merits of legislative decision making by stochation, this is something that I’ve argued for in two books and it is currently the topic of my PhD! The problem is Joe and Josephine Public — although they may well feel dissatisfied with the current state of affairs they still need to understand the case for decision making by allotted proxy. And to my mind this will only work if the randomness is confined to the selection of persons, rather than decision output — i.e. it would have to make no difference which persons are selected, the outcome will still be the same (within an agreed margin or error), otherwise there is no way of knowing which decision is deemed to be the “representative” one to which everyone else is deemed to consent.

    >So far as I’m aware, no one has ever complained that trial by jury is problematic because “the overwhelming majority of people [unlike in a referendum vote] will not be involved in any shape or form.

    The difference is that trial jurors are charged with determining matters of fact (guilt or innocence) in which they have no (obvious) interests, so this is a purely epistemic process. Legislative juries, however, have to indicate their informed preference, so interests predominate — this is a political decision-making process, hence the need to ensure the ongoing representativity of the allotted sample. The mistake made by deliberative democrats is to attempt to turn politics into an epistemic decision process.

    >was there time for discussion after the advocates presented their cases – were nomothetai given time to ponder their votes, (during which time they were free to discuss the proposed law with other nomothetai)?

    No. And trials tended to only last a few hours (never more than a day), so the opportunity for meal-break discussion of a particular case would have been extremely rare (it may even have been prohibited).

    >Secret ballot seems a sensible part of Athenian legislative jury decisions.

    I believe secret voting only applied to the law courts, not the legislative panels. This is just a clarification of historical fact, as I believe (pace Mill) that secret voting would be necessary to protect the political process.

    >unlike popular vote where younger citizens and others are under-represented among those who vote.

    Again it’s interesting that the right to serve on a 4th century nomothetic panel was restricted to older citizens who had taken the dikastic oath (unlike legislative decision making in the assembly during the 5th century).


    >a decision that the majority of citizens would probably make given time, resources, information and motivation.

    That’s fine by me, but we need to demonstrate this empirically — i.e. by convening parallel samples and checking that the decision outcome was the same (within an acceptable margin of error). Otherwise this is just an article of faith or, more accurately, a hypothesis that has already been refuted (by the three Texas public utility DPs).


  50. Keith,

    1. I think all that needs to be shown is that there are democratic reforms using citizen juries that are a significant improvement on what is now in place. Yes, we of course want the procedures and arrangements for juries to be as well designed as possible, and it should also be improved on over time. My proposal on that in articles of mine published from Spring 1997 to January 1999, and today, is that juries choose a director or commission to oversee the working out the best possible design, and that they improve that design over time. The commissioner/director(s) can, and in my view should, be chosen by juries for a set term, and should be advisory, with all final decisions about procedures and arrangements being made by juries after hearing from the commission/directorate and others. The commission/directorate could have the power to have such “procedures juries” convened from time to time.

    2. Do we know that nomothetai hearings only lasted not more than a day, and how soon after the advocates finished voting took place? Are you extrapolating from dikast juries hearing trials, or is there a credible source for this proposition specifically about nomothetai hearings? Laws are of great importance, more than just a trial about one or a few individuals – maybe the hearings were therefore longer than a day and more time was given for the nomothetai to ponder their votes? I am just speculating, but in I think a reasonable way – is there clear direct evidence on point that you or anyone else can cite?


  51. Simon,

    >I think all that needs to be shown is that there are democratic reforms using citizen juries that are a significant improvement on what is now in place.

    When it comes to putting things into practice I like to think I’m just as pragmatic as the next man, but that doesn’t mean there should be no role for the sort of principled distinctions that are the bread and butter of political theorists like me. The principle of democratic equality underpinning electoral representation presupposes that any citizen can stand for office and that the vote of every citizen has exactly the same value. We know in practice that rich, articulate, educated and well-connected citizens are more likely to succeed in the competition for office and that the causal power of each individual vote is negligible, but that’s not the point — from a formal perspective the rules of equality have been respected. The same considerations have to be applied to stochation and the one that I tend to focus on is the need to preserve the democratic equality of the overwhelming majority of citizens who are excluded from the process. Not only is this important from a normative perspective, it’s also crucial in terms of the perceived legitimacy of stochation as a representative principle. It also has something to do with the meaning of words — stochation is a form of statistical representation and statistical representation only occurs at the aggregate level, so any attempt to apply the stochastic principle to individual persons is simply meaningless. For all these reasons individual speech acts by jurors can be ruled out as illegitimate a priori, irrespective of whether they might be “significantly better” than our existing arrangements, particularly when you take into account the law of unintended consequence.

    >Do we know that nomothetai hearings only lasted not more than a day, and how soon after the advocates finished voting took place?

    “The act of legislation proceeded like a trial. The author of the proposal for change comes forward as the accuser of the existing laws. After he has spoken, the five advocates chosen by the people [in the assembly] to defend the existing laws are given their turn. When both parties have spoken, the nomothetai decide by show of hands [not secret vote] . . . A meeting lasted only a single day; actually, it is likely that the nomothetai could deal with more than one proposal in the same meeting” (M.H. Hansen, The Athenian Democracy in the Age of Demosthenes, London, 1999, p. 169).

    Note that this is the only historical example that we know of for lawmaking by a randomly-selected panel of citizens. My modern proposal attempts to follow it as closely as possible, but making allowance for issues of scale (this pertains only to the procedure for selecting advocates as the stochastic principle is scaleable). As Daniela demonstrates in her thesis, attempts by deliberative and epistemic democrats to use classical Athens as a model for their modern proposals are unsupported by the historical evidence.


  52. Keith,
    1. In describing those who had an equal chance to be drawn but were not, you regularly use a variant of the phrase “overwhelming majority of citizens who are excluded from the process.” Your use of “excluded” seems intentionally loaded, and not fair. All citizens were treated equally, as they are nominally in a traditional election. In an election…. would you say all those who voted for a losing candidate were “excluded” from the decision? What about all those who voted for a candidate who already had more than enough votes to win…were they all excluded? Indeed perhaps the ONLY citizen who was not excluded is the one voter (although we can’t say which one it was) who tipped the balance and moved the winning candidate ahead of the others. Since almost none of the voters participate in the actual law writing and voting process, in a traditional electoral scheme you could also say nearly all citizens are excluded. So I propose that you find a less loaded term… All of the citizens in a lottery ARE INCLUDED int the PROCESS, in that they have an equal shot in the lottery, and far more citizens can be involved in the actual law-making tasks than under current all-purpose elected representative systems.

    2. I wonder where Hansen found reference to raising hands by nomothetai, as it is sooooo contrary to normal trial process in Athens. However, if so, it does drive home the notion that the nomothetai were seen as a representative sample of the demos if they used the same voting procedure (raising hands) as the Assembly.


  53. Terry,

    Yes, my use of the word “excluded” is deliberately provocative, in order to counter the illusion that there is something “participatory” about stochation. With electoral democracy every registered voter receives a polling card, inviting them to attend their local polling station in order to register their political will whereas, after the aleatory coup, only a tiny number will be in receipt of the golden ticket (Roald Dahl allusion intended). The vast majority of citizens will receive no communication whatsoever and will have no way of indicating their preference, considered or otherwise. I’m aware of the limited causal efficacy of individual votes in mass democracy (that’s why I’m a sortition advocate), nevertheless the polling slip is a potent symbol of isonomic equality — voting power is shared evenly between all citizens (or at least it would be were it not for anachronistic territorial constituencies). As for the fate of those who vote for the losing party, this is simply a consequence of the majoritarian principle, which is accepted by most reasonable people (not just the followers of Rousseau and Bentham).

    Given the complete lack of even token involvement of the vast majority of citizens in a kleristocracy, it is absolutely vital that everyone should understand that it makes no difference whatsoever who gets the golden ticket — the outcome would be the same, my (speculative) contention being that face-to-face deliberation would vary the outcome, thereby striking at the heart of the representative principle involved. The disenfranchised will only accept the stochation principle so long as they are fully confident that the relationship between themselves and their proxies is an invariant one. Otherwise this will result in randomness in the pejorative sense.

    >It does drive home the notion that the nomothetai were seen as a representative sample of the demos if they used the same voting procedure (raising hands) as the Assembly.

    Yes indeed, the nomothetai were a delegated subcommittee of the assembly, hence my PhD supervisor’s arguments that this particular reform was largely for administrative convenience. The fact that there is no clear dating for the reform (Hansen speculates that it was around 403/2) would suggest that it was not intended to be a major change (although it is very important for those of us seeking to construct a modern proposal for stochation-based democracy).


  54. Keith,

    1. Thank you re the nomothetai quote, that clears up some of my doubts. It of course does not mean nomothetai did not discuss and debate the proposed laws with others before the hearing, as well as on such breaks as took place during the hearing.

    Also, it is not true that deliberation was not a part of citizen juries in Athens. For example, the boule was deliberative (and also played a role in legislation). I also think that many of the juries were deliberative (though the jury courts seem not to have been), but I need to look that up a little further or be guided by someone who has.

    2. The speech acts of jurors (many jurors) are surely aggregate.

    3. I agree with Terry re “excluded.”

    4. I think we are not going to agree on terminology, though our conclusions may be the same or similar. Elected assemblies are far more accurately described as “exclusionary” than a citizen jury, it seems to me, for reasons which may be obvious or easy to extrapolate by now. Similarly, a citizen jury is far more accurately and reasonably described as “inclusive” than an elected legislature. (Just because something is exclusive in one regard does not mean it is exclusive in the regards that matter most. Just because prisoners may enjoy the freedom of not having to shop for and cook their own food, does not mean they do not lack freedom in important regards, or that they are free.)


  55. Simon,

    The nomothetai would not even know which case(s) they were going to try until immediately before the trial. As I’ve said repeatedly I’ve no problem with informal conversations, only the institutionalisation of face-to-face deliberation in a manner that fails to ensure the representativity of the speech acts wrt the target population (an impossible task). The boule (council) was not a jury, it was a collective administrative magistracy charged with preparing the agenda for the assembly. And how can you have face-to-face deliberation between 500 people? The only courts in classical-era Athens were the law courts and the legislative courts and, as Daniela’s thesis makes clear, neither were deliberative in the modern sense of the word. Deliberative and epistemic democrats who use classical-era Athens as their template are simply wrong.

    >The speech acts of jurors (many jurors) are surely aggregate.

    Only if the illocutionary (persuasive) force of each juror were identical. In practice there is enormous variation — most people will remain silent, a few people will speak and some will be more persuasive than others, on account of differentials in rhetorical powers and perceived social status. Just imagine if “your” proxies (people who shared your views and interests) were either too shy or lazy to speak or were unpersuasive bozos (remember these people are randomly selected, as opposed to elected politicians, some of whom [Tony Blair for example] had to choose between a career in politics and acting). The only thing that is equal in jurors is the power of their vote, so this is the only thing that can be aggregated in a manner that will reflect the preferences of the target population.

    >Similarly, a citizen jury is far more accurately and reasonably described as “inclusive” than an elected legislature.

    Because it includes “ordinary” people, as opposed to old male Caucasian lawyers? This claim conflates two entirely different forms of representation. If I want an advocate to represent me then I will compare the stated views of those who have nailed their colours to the mast and pick one who seems to share similar views, who is knowledgeable and intelligent and has good powers of persuasion. I don’t really care if she resembles me in any respect, and may well opt for an old male Caucasian lawyer. But a jury that decides the outcome should be “inclusive” in the sense that it represents the target population “descriptively”. These are two entirely different meanings of the word representation and there are good arguments for why they require two different “bodies of men” (Madison’s term).


  56. Keith, I’m well aware the nomothetai did not not know which laws they would decide til the day of the hearing.

    I thought it was clear that by a jury (or citizen jury, or jury assembly, or minipublic) I mean a random sample, or sortition chosen sample, of citizens who meet to make a decision. That includes the boule, and also all the other sortition chosen bodies in Athens.

    You say: “The boule (council) was not a jury [it is on my definition, as I thought was clear], it was a collective administrative magistracy charged with preparing the agenda for the assembly. And how can you have face-to-face deliberation between 500 people?”

    I think your apparent view that the boule (or sub-groups of it such as the pryteneis) prepared the agenda for the assembly without discussion (and apparently in your view also carried on all their other work without discussion), is not plausible. I’m going to guess that not a single historian who has written a book about Classical Athens would agree with your apparent view. How did the boule (or sub-groups of it) draw up and agree on the agenda for the assembly (and also do the rest of the things they did, such as sometimes pass decrees) without discussion with each other?

    Leaving aside the dikasts, I imagine that other sortition chosen decision-makers in Athens would have engaged in discussion with each other in the course of making decisions.

    Regarding the broadness of the terms “jury” and “jurors,” as you of course know, in modern usage the term “jury” is not restricted to trial juries, but is of broader application. There are for example coroner’s juries, grand juries, and juries to decide who gets book awards, and so on (and of course I am well aware that a book award jury is an “elite jury” rather than a “citizen jury.”) In any case, by “jury” I include all groups of citizens chosen by random selection or sortition who meet to make a decision, and by “juror” I mean a member of a jury.

    I don’t see how understanding that citizen juries are accurately and reasonably described as “inclusive” in important regards “conflates” the two different senses of “representation” you like to mention.

    I am aware that being representative and being a representative are different concepts.

    For me (though apparently not you) an “aggregate” does not have to consist of parts all of which are exactly 100% equal. (In any case, I think the issue of substance, rather than of definitions, is that citizens have unequal abilities and means of persuasion, which has of course been discussed above.)

    The Athenians, as of course you know, did not at all share your concern for exactly representative citizen juries (emphasis on the word “exactly”)

    Liked by 1 person

  57. Simon,

    The distinction between magistracies and juries is fundamental in the literature on classical-era Athenian democracy. The task of the former was primarily administrative or executive and any decisions taken were of an immediate, limited and ad hoc nature. Magistrates, unlike jurors, were held to account at the end of their term of office. The responsibility of the juries was either to determine the law or to apply the law to individual cases — i.e. their fundamental role was decision-making, not administrative. And the Athenians required that all such decisions be taken after silent deliberation.

    I know of no historian of the period who believes that the council was a jury, not a magistracy. Of course it had to engage in discussion in order to fulfil its task, all committees have to! But it wasn’t a law-making body, so there was no need to ensure the representativity of the discursive acts that were undertaken (the same would apply to the other collegial magistracies). And those who claim the boule as a template for “deliberative democracy” simply can’t count — the maximum size for a deliberative body is around two dozen (Coote and Lenaghan, 1997), so this clearly cannot apply to the Council of 500.

    >I am aware that being representative and being a representative are different concepts.

    Good! Most people on this forum conflate these two (descriptive and active) meanings.

    >For me (though apparently not you) an “aggregate” does not have to consist of parts all of which are exactly 100% equal.

    It depends on the use of the word “aggregate”. When you mix concrete, you have an (unequal) aggregate of sand, cement, chippings and water, whereas the aggregation process that we are referring to here is a representative one. Votes are of exactly equal value, so (under certain conditions) aggregating the votes of the microcosm represents how the target population would have voted under identical conditions. However:

    >In any case, I think the issue of substance, rather than of definitions, is that citizens have unequal abilities and means of persuasion

    Right. And that’s why a representative aggregate has to exclude speech acts, otherwise the relationship between the microcosm and the macrocosm will be random. Or are you saying we should privilege the speech acts of the more persuasive members of the sample, because this is the way it is in the wider world?

    >The Athenians, as of course you know, did not at all share your concern for exactly representative citizen juries (emphasis on the word “exactly”)

    The concept of mathematical probability was unknown to the ancient world, so their notion of representation would have been the crude notion of the cook using her spoon to sample the contents of a pot full of soup. And we should also remember that modern notions of pluralism and multiculturalism were unknown in the ancient polis (the Athenians put a great premium on homonoia [same-mindedness]), so there would have been no need for exact representativity. The quasi-Marxian perspective of Aristotle, who subdivided the citizen body into the rich and the poor, would have been sufficient to ensure that randomly-selected juries were sufficiently representative (that’s why he didn’t like them, as it meant rule by the poor). The modern appeal by the Occupy movement to this principle is sociologically anachronistic, notwithstanding its attraction to the old lefties on this forum.


  58. Keith, some interesting comments in your most recent post above.

    >The responsibility of the juries was either to determine the law or to apply the law to individual cases — i.e. their fundamental role was decision-making, not administrative.

    There is such a thing as administrative decision-making. The board of a public broadcaster for example would probably be called administrative, but nevertheless it engages in a lot of decision-making. Ditto for many other boards.

    >the maximum size for a deliberative body is around two dozen (Coote and Lenaghan, 1997), so this clearly cannot apply to the Council of 500.

    So, are you saying that for example the British House of Commons, City councils that number 25 or more, the U.S. Senate, Canada’s new cabinet (which has 30 members), are not deliberative bodies? I think that would be very much news to for example Canada’s new PM and his cabinet.

    A large group can also divide into smaller sub-groups for the purpose of deliberation. A legislative jury of say 500 can break into sub-groups of say 15 for deliberation. The boule also had smaller sub-groups, as mentioned.

    On my definition of a citizen jury, the boule and the sortition chosen magistracies of Athens are juries, as mentioned, because what I mean by a jury is a sortition chosen body (or randomly selected body) that meets to make a decision (or a number of decisions).

    I of course accept that “juries” can different functions and ways of proceeding.


  59. Simon,

    The UK House of Commons is certainly not a deliberative body — if you take a look at Bagehot on the English Constitution, by his time deliberation had long been arrogated to the Cabinet — a small committee meeting in secret. And since the Thatcher era that is also an anachronism, as serious deliberation is now confined to the party-leader’s snug. And your view of the Athenian council as a jury would not be accepted by any historian of that period that I know of.


  60. Keith,

    My view that the Athenian council was a jury or citizen jury would be accepted by every historian aware it was chosen by lottery/sortition, met, and made decisions, because, as I said, “what I mean by a jury is a sortition chosen body (or randomly selected body) that meets to make a decision (or a number of decisions).”

    The council (and its 50 member subgroups, the prytanies) made decisions, and there is no way they could make the decisions they made without discussion and deliberation. For example, as you know, the boule prepared the agenda for the assembly, and put forward draft proposals (probouleumata) to the assembly.

    I grant you that it is interesting, and possibly instructive, that the jury courts voted without any formal discussion between them.

    Perhaps you should let Britain’s MPs know that they don’t engage in deliberation at the House of Commons, as for example Jeffrey Corbyn seems to think they were having a debate and engaging in deliberations regarding Syria:


  61. Simon:

    I’m not aware of a single historian who referred to the Athenian council as a jury, and the Athenians made a categorical distinction between (accountable) office holders and jurors (who enjoyed the prerogative of the harlot). It would be more accurate to describe juries as a subset of the assembly, as this was their origin in both the judicial and legislative case. Athenian juries never indulged in discussion and deliberation, so your attempt to arrogate a modern perspective on juries to antiquity is anachronistic. Our knowledge of the procedure of the Athenian council is very patchy, but I have greater difficulty imagining the discussion between 500 randomly-selected persons than between the few who actually turn up to discuss policy in the UK House of Commons.


Leave a Reply

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

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

Twitter picture

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

Facebook photo

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

Connecting to %s

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

%d bloggers like this: