Rousseau’s general will and sortition

Along with its famous opening sentence: ‘Man is born free, and everywhere he is in chains’ Rousseau’s Social Contract is best known for its clear distinction between sovereignty and government. The latter was a delegated administrative function: the ‘Prince’ could either be single, few, or many (monarchy, aristocracy or democracy) but was a mere servant of the sovereign popular will. Although Rousseau argued that democratic government was more suitable for small states, he had no problem in principal with the notion of elected delegates administering government under the watchful eye of the sovereign people and subject to their dismissal if the delegated mandate was breached.

Although he did not rule out balloting by lot as a way of appointing magistrates (government officials), his preference was the Roman comitia over the Athenian boule, the former privileging election by choice. Nevertheless he was insistent that sovereignty remained with the assembly of the whole people, irrespective of the method employed for selecting delegated officials (election, sortition, appointment, heredity etc).

Rousseau’s insistence that popular sovereignty could not be alienated led him to deride representative democracy:

The people of England regards itself as free; but it is grossly mistaken; it is free only during the election of members of parliament. As soon as they are elected, slavery overtakes it, and it is nothing. The use it makes of the short moments of liberty it enjoys shows indeed that it deserves to lose them. (III: 15)

Unfortunately Rousseau was sketchy regarding how popular sovereignty should be maintained outside of the electoral process, giving rise to the accusation that Rousseauvian ‘sovereign’ assemblies did little more than rubber stamp the policy decisions of elected or appointed officials. Rousseau claimed that his proposals were ideally suited to small communities, such as his native Geneva or the island of Corsica, expressing a ‘presentiment that this small island will one day astonish Europe’ (II:10). How to extend the principle of indivisible popular sovereignty beyond such small communities was a serious problem, one solution being confederation, another being that the assembly should ‘move from town to town, and assemble by turn in each the Provincial Estates of the country’ (III: 13).

Regarding the inalianability of soverereignty this is because it is an expression of the general will:

I hold then that Sovereignty, being nothing less than the exercise of the general will, can never be alienated, and that the Sovereign, who is no less than a collective being, cannot be represented except by himself: the power indeed may be transmitted, but not the will. (II: 1)

Rousseau contrasts the general will with the will of all (the aggregation of individual preferences), the role of each citizen, when voting on the outcome of a legislative proposal, being to divine the general will rather than to state his own preference. This being the case it’s unclear why all citizens would (in principle) need to participate, as a microcosm would be expected to behave in a similar manner to the whole, in the same way that a holographic image is a miniature representation in full. This being the case, sortition, or balloting by lot (as a method of drawing a portrait in miniature of the whole citizenry) would be as good a method of divining the general will as the vote of every citizen. As such sortition would possibly be the only candidate for instituting a Rousseauvian political solution in a large political community, without breaching the principle of popular sovereignty as inalienable and indivisible.

Although Rousseau attributes ultimate sovereignty to the popular assembly, nevertheless day-to-day policy initiatives are the role of the elected magistracy (as is the case in representative democracy). Exceptions are drawn for fundamental constitutional proposals, assigned to the shadowy figure of the Legislator, but that is not the subject of this post. Ultimate sovereignty in Roussseauvian politics resides in the people’s free choice in electing government officers, along with the yes/no affirmation of legislative decisions. Rousseau was keen to ensure the separation of discursive debate and decision making, a distinction usually overlooked by advocates of deliberative democracy as a way of establishing the general will, although his insistence that citizens should be ‘adequately informed [but] without any communication among them’ (II:3) has been seized on by advocates of the Condorcet Jury Theorem: ‘The general will would always result from the great number of slight differences, and the resolution would always be good’ (ibid.).

Two issues remain outstanding to this consideration. Recent scholarship has suggested that Rousseau’s insistence on the inalienable nature of sovereignty may have been because the vote in the assembly constituted (rather than merely uncovering or demonstrating) the general will (Wyckoff, 2011). If there is a pre-existing ‘right’ answer in a legislative vote then it matters little which, or how many individuals discover this by voting. If, however, a Rortyan perspective on ‘right’ decision-making is taken, in which the right decision is constituted by the outcome of the vote, then issues of accurate representation are more important. Fortunately sortition is agnostic on this point: there is good evidence that Condorcet-style procedures are the best way of achieving the ‘right’ outcome from an epistemic perspective, and it’s also the case that the vote of an appropriately-sized sample will accurately mirror the judgment of the community that is being sampled.

Rousseau also wobbles a little regarding his distinction between the general will and aggregated preferences:

There is often a great deal of difference between the will of all and the general will; the latter considers only the common interest, while the former takes private interest into account, and is no more than a sum of particular wills: but take away from these same wills the pluses and minuses that cancel one another, and the general will remains as the sum of the differences. (II: 3).

Once again sortition is agnostic as to whether the vote tally indicates the general will or the aggregation of preferences (the will of all). Deliberative democrats exhort participants to vote on the basis of the former, whereas liberal theorists are content with the latter. Sortition as a tool is amenable to both interpretations.

Jason Wyckoff, ‘Rousseau’s General Will and the Condorcet Jury Theorem’, History of Political Thought, XXXII (2011), pp.49-62.

44 Responses

  1. PS. Grofman and Feld’s 1988 paper, Rousseau’s General Will: A Condorcetian Perspective, might be recruited in favour of the notion of sortition as a way of discovering the general will (via the Condorcet [Citizen] Jury Theorem). Grofman and Feld take the traditional (epistemic) view that the general will exists independently and has to be discovered, whereas Wyckoff argues that it is constituted in the vote. But the case for sortition would apply whether you are an epistemic or a Rortyan democrat.


  2. Rousseau’s notion of sovereignty involves the surrender of all one’s rights to the sovereign in the unanimity of the general will. It is intended to provide an ultimate, unchallengeable source of legitimacy to the nation state. Ideologically it has been a disaster.
    If it ha any consistent application, it could only be to humanity as a whole. No modern community can constitute itself as a purely self-regarding entity.
    Modern thinking rightly insist that there are human rights that are never validly surrendered to any claim to sovereignty.
    There are many types and degrees of authority, constituted by conventions that are constantly under review in the light of their effects on the human interests of different kinds that they purport to promote. While many of these have a legal status in some jurisdiction, they rarely owe their authority to that status, which is merely a way of defining their relations to other authorities.
    It is a fallacy that there must be some sovereign authority with power over all other authorities. All that is required is that in regard to certain unresolved and dangerous conflicts it is desirable that there be a recognised way of arbitrating certain aspects of that conflict.
    The political problem of our age is to give those affected by the decisions of authorities a more effective say in their decisions.
    At present democracy means in practice that people can kick up a fuss after the decisions have been made. It is often not too difficult to get agreement that a decision is bad. But the popular voice has little capacity to generate a better decision, mainly because people have different reasons for rejecting it and no means of negotiating their differences. They can only buy what the politicians offer as part of a package that is negotiated on the basis of the power trading among politicians that is decided in ways that are irrelevant to many substantial interests.


  3. Thanks John for an excellent summary of the dark side of Rousseau’s philosophy. But isn’t the problem that Rousseau provided no way of discovering/constituting the general will other than going to the assembly and approving/disapproving laws proposed by delegated government officials. In the Social Contract he discusses feasibility issues in the numerical context of Geneva, Corsica and the Roman comitia, none of which are of any contemporary relevance. I agree with your discomfort over surrendering all rights (life, liberty and the pursuit of happiness) to the collective sovereign (with Hobbes it was only the first of these), but a sortive approach would provide some protection by ensuring that the general will was the product of the views of an empirical mini-public, rather than a mysterious emergent phenomenon (defined, in practice, by the shadowy figure of the Legislator). If so all we would disagree on is the need for a plenitude of demarchic committees and the need for sortition to be based on those who volunteer their services.

    I suppose what I’m arguing is that sortition could provide a liberal solution to the problem of the general will (if necessary by demonstrating that there is no necessary difference between it and the will of all).


  4. Keith, this is interesting and without discussing the technicalities of Rousseau’s position I would suggest that you might be overlooking the most important parallel between Rousseau and your own thoughts on sortition. One reason that Rousseau rejected representative sovereignty was that it would introduce inequality of legislative right into the sovereign (legislative) body – that is, inequality between the represented and representatives. As the legislative body was the moral branch of the state for Rousseau (as opposed to the physical branch of the executive/government) this would violate the ‘moral and legitimate equality’ established by the contract. The only possibility Rousseau countenanced to preserve moral equality amongst the citizens was if every citizen had to assemble and have an equal vote for laws to be passed. But you might argue that sortition preserves the equality of legislative right that was so important for Rousseau without succumbing to the inequality of legislative right that, on Rousseau’s analysis at least, is inevitably introduced once representatives are permitted in the legislative body.


  5. Thanks Robin, that makes a much more compelling case, and the focus on inequality will go down well on this blog [Robin Douglass has recently (successfully) defended his thesis on Rousseau, so he has the advantage over me of knowing what he is talking about].

    Another fruitful area might be the issue of consent. Bernard Manin has argued that sortition was replaced by election on account of the natural right theory of consent that was dominant at the time: “However lot is interpreted, whatever its other properties, it cannot possibly be perceived as an expression of consent (Manin, 1997, pp.84-5). This is all based on slippery arguments that Iain Hampsher-Monk exposes in the Locke chapter of his HPT textbook, and one might well claim that sortive “consent by proxy” (Fishkin’s term) is at the very least no less plausible.

    But a much stronger line could be taken by developing Robin’s point. It borders on an abuse of language to suggest that somebody has tacitly consented to a government that she didn’t vote for, but if the sovereign (which has the role of approving/disapproving legislation formulated by the physical branch) can be accurately described by a portrait in miniature without a breach of equality then the collective judgment of the micro-sovereign could be said to embody the consent of the whole population. The crucial phrase is “without a breach of equality”; this would not be possible if allotted members took on the individual active functions that Rousseau reserved for the physical branch (proposing and advocating legislation) as those who drew the golden ticket would be a lot more equal than those who didn’t. Such an assembly would be a worse breach of equality than elected representatives, who can at least say that they were chosen by their peers (so could claim an ascriptive mandate).

    Whether the judgment of an allotted assembly could be said to indicate the general will, as opposed to an aggregation of individual judgments is a further issue. If the former is true, then consent is clearly implied (as dissent is defined as mistaking the general will). Consent in the aggregate case (the will of all) might well require additional supporting argument, but at least would be no worse than Manin’s claim for the consent entailment of liberal democracy. Jason Wyckoff is intending commenting on this thread and I’d be curious to learn if his argument that the general will is constituted in the vote of the sovereign body, would be of any relevance to the consent issue.


  6. I would want to keep the issues of equality and consent separate. If we think in terms of a distinction between equality of outcome and equality of opportunity then we might think sortition satisfies the latter more than representation (neither satisfies the former, at least if what is being equalised is legislative right). I agree that the ascriptive mandate is a good justification for elected representatives, but not because this satisfies equality. It is also worth noting that Rousseau has a far more demanding notion of moral equality than most (and I am not suggesting he is right), as it might simply be retorted that the equality of citizens only entails equality under the law and not necessarily equality in forming/ratifying the laws.

    It is well to point out as well that (historically) consent operated at two levels – at least for two stage contract theorists (pre-Leviathan Hobbes, Pufendorf and some readings of Locke). First, the consent of every individual is required to establish a body politic or a ‘people’; second, the consent, of the people (as a collective body) is required to confer the legislative and/or executive powers onto another body. Whether or not sortition is legitimate, as opposed to representation or direct democracy, is a question about the second level of consent or contract – for this reason I suspect Manin’s point might miss the mark, as natural right arguments tended to operate at the first level.


  7. Any implication that elections were preferred to sortition because of one theoretical consideration or another has its direction of causality reversed.

    Policy is not determined by theory, but the other way around: official theory is made to rationalize policy. Policy is determined by the interests of the powerful.


  8. Robin,

    My impression (and I may well be entirely wrong here, I need to re-read Manin’s book as I’m due to debate this issue with him at the Paris sortition workshop in May) is that he’s dealing with a third level — i.e. consent to legislative outcomes. Locke’s concern was consent to specific levels of taxation. However fair or necessary taxation is, ‘still it must be with his own Consent’ (1967, para 140, p.380). Admittedly he then goes on to fudge the issue by saying ‘i.e. the Consent of the Majority, giving it either by themselves, or their Representatives chosen by them’, but that’s the problem with electoral consent – if you choose the losing candidate, do you still consent to the tax? Is it true to say that one consents to a legislative outcome that one disagrees with? (as opposed to merely tolerating it as a law-abiding citizen). In the case of elective systems, one tolerates such outcomes by acknowledging that one’s own view is in a minority, but I’m not sure how one can claim that this embodies your consent – especially if it’s a case of tacit consent to a hypothetical contract (or three contracts, by Robin’s analysis).

    Fishkin’s alternative is the notion of ‘consent by proxy’ – i.e. if an allotted assembly is an accurate microcosm of the citizenry then its consent could be taken to embody the consent of the full citizen body. But that would be easier to argue using the language of the general will, as allotted proxies who were in a minority were simply wrong about the general will. I’d be interested to know if this would be undermined by Jason’s argument that the vote constitutes (rather than reveals) the general will, but note Robin’s skepticism as to whether this is an accurate reading of Rousseau. I hope Jason is wrong (!), because the traditional reading of Rousseau would support a stronger version of allotted consent as well as supporting the claim that a diverse legislative body would be better from an epistemic perspective (getting the ‘right’ answer, as opposed to merely the option that claimed democratic legitimacy).

    Thanks to Yoram for his brief tutorial in the doctrine of historical materialism.

    Locke, J. (1967), Two Treatises of Government, P. Laslett (ed), CUP.


  9. Keith, it is a while since I read Manin so my comment was based on your gloss of his argument. A charitable reading of Locke’s slip/fudge is that individual consent only operates at what I previously referred to as the first level – individuals unite together to form a community (/body politic). At this point they consent to the will of the majority determining the will of the body politic thereafter, and the majority place the legislative power in the hands of representatives as a trust (they could equally place it in the hands of a hereditary monarchy if they so willed). When the representatives pass the tax, this is indirectly legitimised by the will of the people, which is in turn legitimised by the will of every individual. But what is important here is that individuals only consent to live under a body politic where the legislative body takes a certain form of constitution; they do not (directly) consent to specific electoral outcomes, laws or policies. The right of the people is only restored if the constitution of the legislative body is changed (the trust is broken) without the consent of the majority.

    I guess I think that it is misguided to attempt to justify representation or sortition based on anything more than indirect consent – that is, that we (implicitly/ tacitly) consent to the system under which laws, policies etc. are formulated, and only indirectly to those laws and policies themselves. This is to admit that we don’t have a better theory of consent than that which Hobbes provides (a conclusion I would willingly accept), and if we don’t want to embrace Hobbes’s theory of representation then I think our best bet would be to decouple justifications of representation (or alternatives such as sortition) from consent – but then I haven’t read Fishkin so I am probably speaking somewhat prematurely here.


  10. Thanks Robin, Lockean contractualism is beginning to sound a lot like nonsense on stilts, so I can understand why you prefer the Hobbesian alternative (what prudent death-fearing men would rationally choose to do). Doesn’t the Lockean story imply some sort of historical instantiation and, given that the nature of the representative body has changed over time, contractual amendments are required? (Filmer at least had the virtue of a plausible historical narrative.) Locke, of course, was only really interested in the circumstances in which consent could rightly be withdrawn, but to found that on a Just So story does not provide a plausible foundation for the withdrawal of consent. In this sense I’m sympathetic to Yoram’s view that it’s only epiphenomenal froth.

    Manin is adamant that natural right theory rules out sortition in favour of election, and the determining factor is the presence or absence of consent. It’s not clear, however why this should be the case as he appears to be referring to how the constitutional system operates (as opposed to how it was instantiated/legitimised).


  11. The Lockean story does imply some historical instantiation whereby individuals unite together to form a community, which does seem fairly incredulous. But most of his theory seems to rest on tacit consent, which to my mind is just less well articulated Hobbesian consent with some further natural law constraints placed on that to which individuals (as God’s property) can consent.


  12. Robin, I hugely enjoyed your paper which I think is of great relevance to the discussions on this blog. Is there a version on SSRN or another repository so that others can read it? The problem, as you point out, is the theoretical and practical difficulties in establishing a ‘tolerable medium between the most austere democracy and the most perfect Hobbism’ (p.29). This leads theorists into all sorts of sleights of hand, such as the claim that representative government is democratic, as we are all free to choose the aristocrats (or thespians, in Manin’s case) we prefer. M.H. Hansen, in his HPT article, argues that we have a mixed system of government, but devotes four pages to his description of the monarchical and aristocratic elements and only one short paragraph to the democratic. Deliberative theorists like Rawls and Cohen recruit Rousseau to their cause but ignore both his preference for silent deliberation and his principled (your claim) argument against representation. Indeed it is necessary to question whether the notion of representative democracy is even defensible as a coherent idea (p.5). As you point out in your paper there are two issues, liberty and equality.

    Rousseau rules out representatives in the legislature as ‘sovereignty – being the exercise of the general will – can only be represented by itself’ (p.4). Owing to the elitist nature of elective democracy (Manin’s principle of distinction), representatives are, ex hypothesi, different from ordinary citizens, so cannot be viewed as a representation of the sovereign. In Rousseau’s Geneva citizens were only a small minority of the inhabitants of the city (p.12), so it was entirely feasible for all citizens to affirm legislative acts, thereby not alienating their own sovereign will. Not so in large political communities (at least for those of us who would rule out referenda on epistemic grounds).

    But would sortition provide a means for sovereignty to be ‘represented by itself’. In theory if the role of the assembly was restricted to voting up/down legislative proposals from delegated officials then, so long as the sampling process was accurate, the sample size large enough and the discursive input the same, any number of assemblies would return exactly the same outcome. As the assembly is a microcosm of the full sovereign body then such an outcome would involve no alienation. Note that this would be true whether or not the vote represented the general will or the will of all, and whether one took an epistemic or constitutive perspective on the status of the vote.

    You also note that Rousseau (echoing Locke) claimed that tyranny would ensue if the legislative and government powers were exercised by the same body (p.6) and that brings us to the second aspect.

    Rousseau distinguished between the physical (government) and moral (legislative) functions of the polity, arguing that the former was best achieved by elective aristocracy (pp.6-7), the term that both Manin and Hansen use to describe the constitutive principle of modern legislatures. The physical aspect, in Rousseau’s view included not just the execution of the laws but also the proposal (and active deliberation) of new laws, the role of the sovereign body being to deliberate in silence, without communication. (This distinction is based on two competing etymologies for the word deliberation, which I go into in my paper for the Belfast PSA meeting). The role of the sovereign body is, by necessity, limited, as active, individual functions (proposing and advocating legislation) would be in breach of the moral equality that Rousseau held to be fundamental to all citizens. That ‘all should have an equal right in making the law’ (p.24) is even more important when sortition (as opposed to direct democracy) is the constitutive principle for the sovereign body as the sampling process could only guarantee a consistent outcome when the assembly is considered as a whole (i.e. when voting – all other speech acts being, of necessity, ascribed to individuals who are anything but equal in their perceived status and persuasive powers). This is the reason that Rousseau restricted his discussion of the general will to the *vote* of the majority (p.19). The only way that the moral equality of the sovereign could be preserved would be by ascribing the initiation and advocacy function to another body (either elective aristocracy or other forms of direct-democratic initiative). Indeed, even if legislation were to be proposed by a professional (appointed) executive or hereditary monarch it would not affect the democratic legitimacy of a sortive sovereign assembly with monopoly rights to affirm or reject the proposals. This is the reason that Rousseau used a single term (the Prince) to refer to the executive arm of government.

    In sum, although sortition is a form of descriptive representation (rather than direct democracy), it is fully compatible with Rousseau’s strictures on the constitution and role of the sovereign legislature. In his ‘Considerations on the Government of Poland’ Rousseau did allow representatives in the sovereign body as long as the general will prevails (p.22). ‘The liberty of the individual citizen would not be violated as long as the general will prevailed’ (p.23). Sortition, would, ex hypothesi, ensure that it did.

    Although Rousseau considered (and dismissed) sortition, this was for the formation of the government, not the legislature. Sortition would not have been necessary in small citizen bodies such as Geneva, and the technology would not have been available to Rousseau to consider it as a means of establishing a sovereign legislature in a large political community.


  13. Keith, thanks for your comments. I’m currently revising the paper for AJPS and would prefer not to put a copy online (I’m never sure about all the copyright issues), but I’ll happily email a copy to anyone else upon request. Very briefly, I suspect that sortition might contravene Rousseau’s arguments about the inalienability of sovereignty but perhaps not his claims about the moral equality of the citizens. Also, notice that I argue that Rousseau’s ideas on the liberty of the people (as a collective body) and the liberty of individual citizens come apart in ways that I do not think he fully appreciated. The latter will be satisfied so long as the general will (common interest) prevails and is realised in the laws, irrespective of how these laws are formed (as in Poland) – so then the question just becomes one of which legislative process is most likely to realise the general will/common interest. In the Social Contract Rousseau answers majority voting, so long as certain other conditions obtain, but this is not (on my reading at least) because the majority vote is constitutive of the general will and, in principle at least, other methods such as sortition may realise it.


  14. I am not a scholar of Enlightenment philosophers, but I saw in Rousseau’s “A Discourse on Political Economy” that he states that “all philosophers and jurists of any repute” agree that taxes must have consent of the people “or its representative.”

    If consent is seen as affirmative agreement, it can’t logically be granted by one group (the majority) on behalf of another (the losing side). The majority may infer consent by the losing side if it relents. In modern anarchist terminology, this is declining to “block consensus.” So “consent,” as used by Locke and Rousseau, can only be tacit (implied authorization from lack of revolt). Consent can’t be actively given, but can only be TAKEN AWAY from a ruling body. The question is whether it takes a majority to withdraw consent, or if a minority can also withdraw consent – resulting in illegitimacy.


  15. Terry, I think you’re right to focus on the withdrawal of consent, especially in the case of Locke. What puzzles me then is Manin’s argument that the triumph of election was on account of the natural right theory of consent — this would only make sense if everybody boycotted the vote. However the sortive equivalent would be everybody refusing to accept their allotted place, so the balloting mechanism (election or sortition) is neutral in respect to (the withdrawal of) consent. I do think that the most fruitful place to look for a positive theory of consent is in the notion of the general will and whether it can be instantiated by a mini-public (Dahl’s term, but surely one Rousseau would have approved of). It will be interesting to push Manin on this at the Paris workshop.

    Robin, if this would still involve an alienation of sovereignty then it’s hard to imagine how sovereignty could not to be alienated in a large political community (assuming that direct democracy is ruled out for epistemic reasons). If political theorists are going to be of anything other than of historical interest then I think we need to speculate on ways in which the distinctions they made can be realised in a modern setting. Rousseau attempted this himself with his Polish constitution, so others might be legitimately tempted to follow his example.


  16. Keith, as it happens I don’t think Rousseau offers much for contemporary theorists except for a challenge; indeed, I only see his ideas on sovereignty working within a small citizen body and I think this would have been a conclusion he would have willingly accepted (Poland addresses different questions and is not really to do with sovereignty). As I mentioned in my last post, I think that the importance of consent operates at a level different from the one (I assume is being) appealed to in many contemporary debates. There is a valid debate to be had over whether election or sortition is a better form of selecting representatives, but I don’t think it should be couched in terms of consent. If it is, the understanding of consent you are going to be left with will be so weak as to lose what was important about it in the first place.


  17. Thanks Robin, all the more ammunition for beating up Bernard Manin, as this is his sole reason for explaining the “triumph of elections”.


  18. Keith,

    Its been a while since I read Manin, but didn’t he also acknowledge the republican virtue concept of the natural aristocracy as a benefit of election that the framers sought? His “principle of distinction” was understood by the framers (arguing that even if lower class citizens were granted the right to vote they would elect their betters). So universal sortition clearly couldn’t fulfill the goals of people like Hamilton or Madison. But I suppose sortition could have satisfied them if they sufficiently restricted eligibility as in the Italian City Republics…but then they wouldn’t have gained the support of the broader population for their cause. It may be simply my interpretive overlay on top of what Manin actually wrote…but I REMEMBER his book as having more than just the consent argument for election’s defeat of sortition.


  19. Terry, I’ve just re-read Manin’s chapter The Triumph of Election. It’s quite true, as you say, that authors from Harrington, Montesquieu and Rousseau through to Madison and Hamilton valued election as a means of discovering the natural aristocracy, and as a way of screening out incompetents, but this was by way of preference and cannot explain why, by the time of the Federalist Papers, lot was not even considered as a method of appointing political officers. Manin attributes this entirely to the natural law theory of consent, choosing examples from the Putney debates, the Declaration of Independence and Thouret’s draft declaration of rights (1789). Furthermore he states that he is not just referring to the sovereign decision regarding constitutional systems (Robin’s second stage of consent) but the choice that was exercised during every election:

    “Under an elective system, by contrast, the consent of the people is constantly reiterated. Not only do the people agree to the selection method – when they decide to use elections [stage 2] – but they also consent to each particular outcome – when they elect [stage 3]” (Manin, 1997, p.85).

    He then goes on to shoot himself in the foot by pointing out the ascriptive origins of electoral representation in the Roman principle of QOT, by which rulers could declare their subjects bound by the decisions of their representatives – “You consented to have representatives speak on your behalf; you must now obey what they have approved” (p.88); thus electoral ‘consent’ was of more use to the rulers than the ruled, as minorities would be deemed to be as included as the majority who had elected the representatives.

    He also points out (pace Terry’s argument) that the system of electoral representation owed more to feudalism (Gothic prudence) than the republican tradition (p.90), although the relative weight of the civic humanist and Lockean influences on the American founding is disputed.

    The slogan of the Boston Tea Party – ‘no taxation without representation’ – is taken as another indication of the natural right theory of consent (p.86), but Manin nowhere considers why representation needs to be electoral in form. I suspect this is because the traditional view of sortition is as a system of rotation, or democratic equality, not as a means of representation of the general will (or the will of all). Manin’s book, ‘The Principles of Representative Government’ charts the development of representative government out of non-representative systems such as Ancient democracy, monarchy and aristocracy. This is why I think we need to focus on the representational potential of sortition, rather than the Dowlen-Stone ‘lottery principle’.


  20. Message from Lee Ward (University of Regina) who is currently writing a paper on the role of the Legislator in Rousseau’s Social Contract:

    “I think your take on sortition is very interesting. It raises very real questions for me about how applicable Rousseau is for modern democracy. Are our very sophisticated polling/sampling techniques nowadays capable of illuminating something like a Rousseauian general will? Perhaps, but then again if the alternative to sampling is referenda or plebiscites with high rates of participation, would that have more legitimacy in Rousseau’s terms than an elected representative government claiming a ‘mandate’ based on high levels of public support in polls? I suspect Rousseau would favor the plebiscite model as a more authentic expression of the general will.”

    KS: Referenda and plebiscites certainly fulfil Rousseau’s normative requirement that sovereignty should not be alienated; the objection to mass direct-democratic initiatives is an epistemic one — given the miniscule influence that an individual vote has on the outcome, there is no reason why a rational agent should taking the trouble to educate herself on the issues. It’s even worse from the perspective of the general will as very often referenda are used to punish the incumbent administration, with voters completely ignoring the topic of the referendum. Regarding the mandate that elected governments claim, not only does this involve the afore-mentioned problem of rational ignorance, it also conflates the distinction between the physical and moral aspects of politics, and contravenes Rousseau’s dire warnings about the danger of having delegated government officers in the sovereign legislature. This is why it strikes me that sortition is the only *possible* way of realizing the general will in a mass democracy.

    Sortition would also make sense from a formal (procedural) point of view, but would in turn violate Rousseau’s distinction between the moral and physical aspects of government if the deliberation of the sovereign body were to take a more active form than that which he proposed. Active functions, such as proposing and advocating legislation would contravene the fundamental equality of all citizens which is the sine qua non of the moral case for popular sovereignty. But if the role of the sortive body was silent deliberation and voting in the Rousseauvian or Fishkinian sense then equality would not be sacrificed as it would make no difference whether or not “I” was present, as the verdict of an infinite number of mini-publics would be exactly the same. If that were not the case then it would be down to a sampling error — a technical rather than theoretical problem.


  21. I’m coming to the discussion a bit late, but I second Robin’s comment re: the two levels of social contract theory in Locke and Rousseau. Rousseau’s objective is “to find a form of association which defends and protects with the whole force of the community the person and goods of every associate, and by means of which each, uniting with all, nevertheless obeys only himself, and remains as free as before.” (SC I.6.4) This form of association must be unanimously consented to (SC I.5.3), and here, sortition is of no help. What’s at issue in stage one (for Rousseau, anyway) is the formation of a unitary public person, the Sovereign, whose will is one’s own, qua citizen–thus ensuring one’s freedom, understood as self-government, which in turn is understood as obedience to one’s own will. At this stage, one must voluntarily alienate one’s rights to the community.

    But I’m intrigued by Keith’s suggestion that once one has committed to entering into civil society, an allotted assembly could reveal the general will as effectively as general referendum could. Since what is at issue then is not consent, per se, but obedience only to one’s own will, then this position is at least not *obviously* wrong. It does no good to reply that one cannot be bound by the outcome of a vote in which one has not participated on the ground that the vote does not represent one’s will, since the same problem of bindingness arises any time one is in the minority–at least when the voters look to their own interests rather than the common good. This is precisely the problem that the general will is supposed to solve; while (arguably) my private interests can’t be represented by an allotted assembly, my will as a citizen perhaps could. By reconceiving the bindingess problem in the way he does, Rousseau at least opens up this possibility. I’ll admit, though, that I don’t have a settled opinion yet on whether this line of argument is fruitful.


  22. Thanks Jason, I’m relieved your comment finally got through to the blog — no idea why it took so long to appear! Stage 1 contract theory always strikes me as little better than a Rudyard Kipling Just So story but, as you point out it’s tangential to the issue of how the general will is subsequently revealed/constituted.

    I’m greatly encouraged that you find the notion of sortition in the second level of contract theory plausible. It’s clearly the case that consent is not at issue if one presupposes the existence of the general will, and I do believe that sortition is the best (only?) way it could be revealed/constituted in a large political community, given that empirical studies on referenda show that they rarely provide a considered verdict on the issue at hand (unless its a sharply polarised matter like abortion).

    A greater challenge is if you disavow the concept of the general will and seek to make a similar claim from a liberal aggregative perspective — where the consent of the losing minority is the issue. How my private interests could be represented by an allotted assembly? But the problem is the same for losers in an electoral or direct-democratic process, so it’s not clear how any process could have a convincing resolution of this problem. It strikes me here that Fishkin’s claim for consent-by-proxy is at least as convincing as Manin’s natural right argument for the triumph of electoral consent (although that’s damning by faint praise!). I have a paper defending consent by proxy at and Jane Mansbridge made some interesting comments on sortive approaches to consent in a symposium on Fishkin’s last book:


  23. Keith,
    Explaining how sortition can fit with Rouseau’s concept of the general will seems sort of like solving a chess problem…intellectually intriguing, but of questionable real-world value. Playing along though, Wouldn’t there be different “general wills” depending on whether the citizenry were ignorant or informed? Well executed sortition (that includes the element of education) might reveal what the general will WOULD be IF the entire citizenry was educated on some matter. But this might be in direct opposition to what the general will actually is.


  24. Terry

    I can understand if this seems like a scholastic quibble, but it does have significant real-world entailments. The sense that assembly members have a duty to get the “right” answer can largely be attributed to the concept of the general will, along with the preference for public reasons and/or consensual mechanisms that attracts many deliberative theorists. On this list Yoram has expressed the view that representation is not possible in a “corrupt” society, which he defined as one in which allotted representatives put their own and family interests before the general good. The attraction of the general will from the point of view of this blog is that sortition is the only viable candidate for expressing it in large political communities and this overcomes the problem of consent, as those who dissent have simply misunderstood what is in the general good.

    Personally I’m a little sceptical that human agents can (or even ought to) put aside interests and vote purely on the basis of a reasoned appraisal of the general good. Sortition, as a mechanism, is agnostic on this issue, but the issue of dissent remains a significant problem, even if consent-by-proxy does a better job than Manin’s natural-right argument for electoral consent.

    >”Well executed sortition (that includes the element of education) might reveal what the general will WOULD be IF the entire citizenry was educated on some matter. But this might be in direct opposition to what the general will actually is.”

    Exactly — “what everyone would think under good conditions”, as Fishkin put it when he proposed his model of consent by proxy. But the gap between the informed and uninformed general will is only problematic if the microcosm is not an accurate representation. Each randomly-selected assembly should deliver the same verdict of the general will if provided with exactly the same balanced information (“education” in your words), but this would not be the case if their role were more than deliberation in the Rousseau/Fishkin sense. Everybody should consent to the outcome if it were demonstrated that the outcome would be the same whether or not they happened to be present in the microcosm.


  25. > an allotted assembly could reveal the general will as effectively as general referendum could

    A referendum is in fact a very poor tool for revealing the “general will” (unless that term is used to refer to the uninformed and unconsidered selection of some section of the population from a menu of options constructed by a politically privileged group).


  26. BTW, Jason’s comment was stuck in the spam folder. The spam filter generally does a good job eliminating garbage comments, but it is sometimes too aggressive about it and flags legitimate comments as spam. If anyone ever posts a comment and sees that it doesn’t immediately appear, please send me an email to let me know I need to go extract it out of the spam folder.


  27. Keith,

    You wrote (and have stated previously) that
    “Each randomly-selected assembly should deliver the same verdict of the general will if provided with exactly the same balanced information”

    Surely this is hyperbole…Even a large statistical sample will (rarely, but not NEVER) generate a group that would make a different decision than other samples. Also, even the exact same sample may make different decisions simply depending on the time of day they vote (before lunch or after lunch). Likewise, if the entire population became educated and voted, even THEIR decision might change from hour to hour….especially if the vote is close using a majority decision rule. The more appropriate standard is that MOST allotted bodies would make the same decision with the same information.


  28. Terry,

    Yes, fair point, I do suffer from hyperbolemania. But I think the claim is true in principle, subject to a) ideal statistical methodology, in terms of sample size and stratification; b) 100% take-up of allotted places; and c) exactly replicated advocacy. (I’m sure that rules could also be put in place to ensure that votes were taken before or after lunch, and the cafeteria menus should all be the same!). The claim would be empirically true (and the probability could be ascertained experimentally) to the degree that actual conditions approximated the ideal (that’s why we refer to the representative principle as ‘statistical’), and the consent that it embodies would be true at (e.g.) the 98% level, depending on the consistency the experiments returned. Elective democracy (and its associated claim to consent) has a very poor record in returning consistent decision-making, as it suffers from cycling, rational ignorance and information-provision that is anything but balanced. So, by comparison, the sortive alternative would, in practice, be extremely close to the theoretical ideal (98% or whatever). The principal reason why we would never attain 100% consistency is this would involve an infinitely-large sample size and rational-ignorance constraints would appear at a much lower level.

    But my main point (Pitkin’s, that is) is that if members of the AC were to have active individual functions (policy initiation and advocacy) then we would have no reason to expect that each AC would function in the same way, as it would depend entirely on which individuals were selected — and this would be true irrespective of whether one adopted a general-will or will-of-all perspective. The decisions of such assemblies would be just as arbitrary as legislatures convened on monarchical (presidential cronyism) or aristocratic (elective) principles. As such there is no good reason to see why their outcomes could be said to embody the consent of all citizens. I think you’ve broadly accepted this argument as your own proposal situates ACs with voluntarist/activist functions at an earlier stage in the legislative process, to be judged later on by an AC which better approximates the statistical ideal. I’m broadly sympathetic to your proposal for sortive policy generation, but would suggest that an intermediate direct-democratic votation stage prior to deliberation in a statistically representative AC would be required in order to fulfil the consent principle.


  29. Terry,

    Even if the assembly would somehow “always deliver the same verdict”, how will we make sure that it always has the same proposals to choose between?

    And even if somehow we had a process that always delivered the same proposals, and we were assured that the assembly would always choose between them in the same way, why would this artificial condition be considered a goal for the political system? If that consistent verdict is one that consistently provides bad results, then the consistency is worse than useless.

    Why should we prefer bad decisions, consistent or otherwise, over good decisions, consistent or otherwise? The consistency requirement – at least in the very narrow and artificial sense that Keith is offering – is a red herring.


  30. Yoram,

    That’s a very fair point, my democratic model privileges formal equality over epistemic considerations — “good decisions” in your words. To you it’s self-evident that an allotted chamber with full policy proposal and advocacy powers would deliver better results from an epistemic point of view, however the same claim has been made by supporters of Plato’s guardians, Madisonian elective aristocracy and enlightened despotism. My concerns are with the formal requirements of democratic consent, but are leavened by the evidence that the aggregate judgment of an cognitively-diverse group leads to better outcomes from an epistemic perspective. If we are genuine democrats I think that’s about as good as it gets.


  31. > To you it’s self-evident that an allotted chamber with full policy proposal and advocacy powers would deliver better results

    No, I do not think that the value of an allotted parliament is self-evident – it should be argued for based on theoretical analysis and empirical evidence.

    > the aggregate judgment of an cognitively-diverse group leads to better outcomes

    You claim that you are in favor of the judgement of a diverse group, but you propose a system in which all political activity, except for a formal up-or-down vote, is dominated by narrow groups.


  32. Yoram,

    All theoretical analysis is grounded in anthropological and sociological assumptions — for example the existence of distinct and homogeneous classes, each with particular interests; innate differentials between individuals and groups (natural aristocracy); faculty psychology; the existence of a universal disinterested class etc. etc. You can take your pick. My preference is to bypass all this by privileging formal democratic equality, but without claiming any foundation other than that it is the normative standard that currently rules. As for empirical evidence, all we have to go on is a growing body of support for the claim that cognitive diversity is better than the judgment of elite groups, except in highly specialised, technical domains (which would not include political judgment). That doesn’t mean that the judgment of individual Ordinary Joes is better than that of individual members of the political elite; the claim is merely an aggregate statistical observation.

    I’m very aware of the risk of elite domination of policy initiation, that’s why I support Terry’s proposal for multiple, voluntary sortive bodies for this purpose. But I think it would both be very complicated and unlikely to win public support. This is why my preference is for an initiative/petition process, followed by public votation. The widespread and cost-free availability of electronic media make such a proposal more viable now (for ordinary citizens) than at any previous time in history. It would also be comparatively easy to impose tight campaign finance restraints and I also support your call for diverse and plural public media. All these will help to undermine the influence of the rich and powerful.

    As for the problem of elite advocacy, the best we can hope for is to balance pros and cons. Bear in mind that the pro-advocates would be the individuals or groups that won the proposing-AC contest (in Terry’s example) or the petition-initiative process (in my example) so would not be quite so subject to the principle of distinction as electoral elites. The rich and powerful will always seek to exert a disproportionate influence, so it’s better to acknowledge and constrain this, rather than to drive it underground as would be the case in an AC with full powers, which would be particularly vulnerable to corruption. This is why Chouard had to devote so much time in his talk to the complex web of scrutiny that would be required to police such an assembly.


  33. I’m coming to this discussion a bit late, but I must admit I’m not sure I see the problem for Manin. I agree with those who claim that the idea that voting for representatives conveys any meaningful form of “consent” to the laws enacted by those representatives just doesn’t work. But that surely wasn’t obvious the first time people took up the problem. It surely was obvious, however, that sortition never had a chance of doing the job. There’s no meaningful sense in which I have consented to the decisions of an agent that I had no role in selecting. (You might claim that I consented to the system of sortition. But then you would have to admit that if I consented to a system of elections–or to a dictator-for-life–that the resulting system would have my consent. Consent at that level of abstraction does not privilege any system.) So it seems perfectly fair to me to say that people who valued consent above all else would look to election as a means of doing it, and dismissed sortition as a means of obtaining it.

    This is primarily a historical point to me. I think we’d be better off not pretending that our political systems are grounded in consent, whether they use elections or lotteries. But I don’t see where, as a point of history, Manin goes obviously wrong. Just because the arguments he identifies don’t work doesn’t mean that the people who made them centuries ago didn’t believe them.


  34. Yes it might well be historically true for a very limited franchise and a miniscule state. Locke — who was no democrat — was principally concerned with consent to taxation and it’s conceivable that the shared interests of the political class were sufficiently homogeneous so that the consent of elected representatives could be said to embody the consent of the electorate. But you introduce another anachronism by implying the existence of competitive political parties elected on the basis of policy commitments. I believe it’s the case that prior to the Great Reform Act most parliamentary elections were uncontested, so as a consent mechanism this would be not unlike the recent (unopposed) election for the president of Yemen.

    It seems more plausible to argue that consent was based on descriptive representation (chaps like us in parliament) rather than voting on the basis of policy preferences — or even persons — in which case a sortition based on the same forty-shilling franchise would have had the same result as election. Given that attendance at parliament was deemed a chore, what difference would it have made to implied consent if the ballot mechanism was sortition rather than election?

    Moving on to the modern context, the argument for sortive representation-by-proxy as a consent mechanism is based on the claim that any number of allotted chambers would return the same decision (given balanced information). If it literally makes no difference whether I was present or not, why would I not consent to the outcome? Needless to say this presupposes a constrained DP form of deliberation, rather than an assembly with proposing and advocacy powers, in which case there would be no good reason to anticipate consistent outcomes. So even if Manin’s historical argument is true, it’s certainly no longer the case (unless you believe that Audience Democracy — in which we consent to the decisions of the best snake-oil salesman — is a variant worth having.)


  35. > Consent at that level of abstraction does not privilege any system.

    Sure – but if this is the only level at which “consent” works then it is clearly a meaningless condition.

    > Just because the arguments he identifies don’t work doesn’t mean that the people who made them centuries ago didn’t believe them.

    This is putting the bar very low – people will believe anything that is convenient.

    By the way, Xavier Marquez has a recent paper about legitimacy (=consent?) which is quite interesting.


  36. In the pre-1832 House of Commons only 15% of country seats were contested, so if consent required electoral choice then it didn’t work very well in Britain. US elections were contested but Madison deplored any sense of partisan choice, so all that leaves the residual sense of consenting to be ascriptively bound by the judgment of your chosen member of the great and the good. But it’s equally likely that the founders dismissed sortition as an option on account of religious reasons (see ), geography, (lack of) technology or simply fear of rule by the poor. I would agree with Yoram that people will believe anything that is convenient.


  37. Two responses, and so I’ll post them separately. First, Keith writes:

    “It seems more plausible to argue that consent was based on descriptive representation (chaps like us in parliament) rather than voting on the basis of policy preferences — or even persons — in which case a sortition based on the same forty-shilling franchise would have had the same result as election.”

    I really don’t know what it means to say that there’s any “consent” here. If all you mean is that people are satisfied with the outcomes of the system, then the word really has lost all possible meaning. If consent means anything at all, it means prior authorization of some kind, either direct (I said yes, I agree to that law) or indirect (I said yes to some representative, and will agree to any decisions to which he agrees). That’s what Manin has in mind by consent, it’s clearly what Locke had in mind, but I don’t think it’s what you have in mind at all.


  38. Moving on to the second point…

    “Moving on to the modern context, the argument for sortive representation-by-proxy as a consent mechanism is based on the claim that any number of allotted chambers would return the same decision (given balanced information). If it literally makes no difference whether I was present or not, why would I not consent to the outcome?”

    I agree that this is a critical claim, but I think there’s more to it than that. You have to be committed to all of the following:

    1) Direct democracy (or whatever the consent-granting mechanism should ideally be) will under the right conditions realize a unique outcome.
    2) An AC will, under the right conditions yield exactly the same outcome.
    3) No other procedure will be better than the AC at yielding this outcome. (If a dictator was incredibly good at selecting that outcome, we’d presumably prefer relying upon him to an AC.)
    4) THERE IS NO OTHER VALUE TO THE CONSENT-GRANTING SYSTEM APART FROM REALIZING THAT UNIQUE OUTCOME. To have the “consent” of the people simply IS to reach the same decision that it would reach under well-functioning direct democracy.

    Step 4 is really critical here. You really seem to think that the consultative act itself is of no importance except as a means to an end. This is a very controversial claim. IMHO, it stretches the word “consent” to the breaking point. To use an example I have used before, suppose somebody steals 20 euros out of my wallet and uses it to buy me some Chinese takeout. I did not consent to this action. It doesn’t matter that the person who stole it is very much like me (say, a white middle-aged left-wing atheist American expatriate Ph.D.). It doesn’t matter if I acquiesce after the fact, and accept the Chinese food. It doesn’t matter that, left to my own devices, I might have purchased the exact same Chinese takeout. It doesn’t even matter that the person who took the money had a mind-reading device and accurately gauged what I would want for dinner tonight. I simply did NOT consent under any meaningful use of the term. To claim otherwise is 1) to advocate a change to the English language and 2) to say that there’s no value to consent other than reaching the right outcome, and I can’t imagine anyone would endorse that conclusion.

    If I’m right about this, there simply is no meaningful sense in which I can be said to “consent” to the outcomes of an AC–even if there are people like me in it, even if it makes a decision I like, even if I acquiesce to the decision after the fact, the decision simply wasn’t mine. And that’s just what consent means. (Again, for me that’s simply an argument for denying that politics ought to be based upon consent. But clearly Locke et al. did want politics to be so based.)


  39. Ascriptive consent, of the form that you describe, was entirely for the convenience of the executive, and here I’m happy to cite Manin: “You have consented to have representatives speak on our behalf; you must now obey what they have approved”. But what if there was no choice on offer? (85% of the country seats in the UK, prior to 1832) or only one candidate in a presidential election? Although prior authorization may make sense formally, it is generally devoid of substance. The competing theory (consent by proxy) may not live up to the formal requirements of natural right theory, but has a much better substantive claim to embody the principle of consent. I won’t repeat the arguments for this as I’ve already outlined them in this thread, suffice it to say that a descriptively-representative proxy would vote in the way that you would. One would naturally consent to the vote that you cast yourself, so why would you not be said to consent to the vote cast by your proxy? What’s the difference?


  40. Peter, many thanks for your careful response to the second point (the above was a response to your first). Dealing with these issues in turn:

    P.S.>1) Direct democracy (or whatever the consent-granting mechanism should ideally be) will under the right conditions realize a unique outcome.

    It’s impossible to achieve the “right conditions” for mass direct democracy, without totalitarian control of the media. Direct democracy is not my concern, as sortition (in the modern context) is a case of (statistically) representative (indirect) democracy. I forget who it was who argued in Paris that sortition was a form of direct democracy, but they were clearly wrong.

    P.S.>2) An AC will, under the right conditions yield exactly the same outcome.

    This is (formally) true ex hypothesi; in practice the exactness of the similarity of the outcome would be easily testable (by, for example, the Stanford Institute). As electoral representation is anything but exact, then I think we might be content with “broadly consistent” rather than exactly the same outcome. I acknowledged earlier in this thread that my use of the word “exact” was a tad hyperbolic.

    P.S.>3) No other procedure will be better than the AC at yielding this outcome. (If a dictator was incredibly good at selecting that outcome, we’d presumably prefer relying upon him to an AC.)

    This thread is not concerned with the epistemic value of different decision mechanisms, only the normative. My assumption is that democratic equality is non-negotiable in this respect.

    P.S.>4) THERE IS NO OTHER VALUE TO THE CONSENT-GRANTING SYSTEM APART FROM REALIZING THAT UNIQUE OUTCOME. To have the “consent” of the people simply IS to reach the same decision that it would reach under well-functioning direct democracy.

    Yes (even though “well-functioning direct democracy” is, in practice, oxymoronic). Regarding your 20 Euro Chinese takeout, bear in mind that in my proposal the judgment of the AC is only the second stage in the legislative process; policy proposals would be generated by either direct democratic initiative or electoral representation. This would require a party/initiative/referendum/petition proposal to confiscate 20 Euros from every white middle-aged left-wing atheist American expatriate Ph.D (legislation has to be for the general case). If the policy passed the first stage and the AC decided that it would like to spend the money on a Chinese takeout then I’m afraid the appropriation would be entirely legitimate.

    But could you be said to consent to it? To adopt a Rousseauvian perspective if you didn’t consent to the confiscation then you would simply be mistaken about the general will; to adopt a liberal perspective you would be deemed to (reluctantly) “consent” to the will of the majority by not scuttling back to the Land of the Free, where such felony leads to incarceration. That’s what we mean by democratic consent — not everyone will agree to everything, but you have to accept the majority position. No doubt a just constitution would include safeguards protecting the interests of white middle-aged left-wing atheist American expatriate Ph.Ds along with other such vulnerable minorities.

    Although this exchange might appear to be facetious, it does reveal some important distinctions:

    1. The necessity of a formal separation between policy generation and outcome.
    2. The inapplicability of descriptive representation to the former. I totally agree with you that it’s no comfort at all to you to learn that the person who stole your 20 euros is your doppleganger. Nevertheless you would be deemed to consent to the appropriation if the legislation passed the above two stages. Adopting a liberal perspective such consent would be at least no weaker than the electoral approximation; adopting a Rousseauvian perspective, sortiton would appear to be the only way of instituting such consent in a large distributed political community.


  41. It seems to me that the Natural Rights crowd never imagined all INDIVIDUALS consenting, but rather a MAJORITY of “the people” as a whole, OR THEIR REPRESENTATIVES, could legitimately(?) bind the other 49%. The 20 Euro analogy may help define “consent” for an individual…but that isn’t what they were talking about. It was Locke and his ilk that transmorgrified the word “consent.”

    Like others here, I question whether ACTUAL consent is a particularly practical concern, since it can NEVER really be achieved in any large society.


  42. Terry,

    Agreed. But even if actual consent is hard to achieve in a large society, nevertheless it remains an essential normative goal, irrespective of the political system (monarchical, oligarchic or democratic). And, assuming the latter, it strikes me that sortive consent-by-proxy is more reliable than electoral approximation — that’s certainly Fishkin’s extrapolation from 20 years of experiments. It would also be the case that sortive consent could actually be achieved from a Rousseauvian perspective, but that might well be a bridge too far for liberals.

    Of course this only remains the case for an AC with a role limited to the decision function. I would certainly not consent to a tiny group of random individuals putting into law whatever whim they happen to have at the time. Does anyone seriously still think that anybody else would consent to such an arbitrary and undemocratic system? Perhaps the reason for Manin ruling out sortition from a consent perspective is the failure to separate the decision function from policy generation and advocacy. Jim Fishkin told me he has attempted to convince Manin of the political potential of the DP model but has run up against a brick wall, so it’s possible he’s not fully aware that the DP does maintain this distinction.


  43. Keith wrote:
    “I would certainly not consent to a tiny group of random individuals putting into law whatever whim they happen to have at the time. Does anyone seriously still think that anybody else would consent to such an arbitrary and undemocratic system?”

    Not to re-activate the classic Keith/Yoram debates, but yes, I think most people would support such a system IF it could be put in place AND it seemed to work well (and I would deem it to be democratic). As evidence I would mention that 1999 Public Ploicy Center survey

    I’d rather allow a descriptively representative random sample of short duration make arbitrary decisions (on a whim) on my behalf than a semi-permanent elite group. You and I might prefer that decisions NOT be arbitrary, and with good design we might be able to avoid untoward decisions… but the current electoral elite system pretty much makes arbitrary decisions right now.

    Of course, the “$64,000 question” (do you have that old game show expression in the U.K.?) is what sortition reform is realistically achievable in the near term that may pave the way for an optimal sortition system down the road.


  44. Terry,

    This is how the question in the 1999 survey was put: “This [allotted] group then met and were informed on all sides of the policy debate on a number of public policy issues”. I would also happily assent to the verdict of such a group as members of the group were not picking the issues themselves and balanced advocacy was provided externally (“were informed on all sides”). This is nothing like the unlimited mandate that Yoram has been advocating, as the design is close to that of the DP and would therefore be fully compatible with the descriptive-representative mandate.

    The terms “arbitrary” and “on a whim” refer to the policy initiation process, not the decision outcome (verdict). There is every good reason to anticipate a consistent decision outcome between different allotted groups; not so for policy initiation and advocacy as these are individual speech acts rather than a collective aggregative function (silent deliberation, followed by voting). As such the 1999 survey proposal would not breach the democratic equality of all citizens as it would MAKE NO DIFFERENCE which concrete individuals were the representers and which were the representees. This being the case why would any rational individual object or view the verdict as in any way different to that of a trial jury? By contrast if an allotted jury were drawn, and then chose who to arrest and prosecute on the basis of their own personal grudges and whims and then appointed their own judge, prosecution, defence attorneys, expert witnesses, sentencing policy (and even appointed their own executioner and prison officers, making up the law as they went along on the basis of the arbitrary whims of the more persuasive members of the assembly), the proceedings would rightly be viewed as on a par with the Salem witch trials.

    But I do agree that legitimacy would depend less on theoretical arguments and more on what was found to work well and what is realistically achievable in the near term. This is why I spend a good proportion of my waking hours arguing patiently against fellow Kleroterions who scorn valuable experiments showing the efficacy/legitimacy of advisory sortive bodies as weak and half-hearted, and propose instead a popular struggle for a sortition-only system — as there is a zero chance of implementing such a system in the near term, as turkeys tend not to vote for Christmas or Thanksgiving.


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 )

Google photo

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

Twitter picture

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

Facebook photo

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

Connecting to %s

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

%d bloggers like this: