Obligation and Consent

The last Paris sortition meeting was devoted to Bernard Manin’s argument that sortition was replaced by preference election on account of the natural right theory of consent. I challenged Manin on this with an argument based on Fishkin’s claim that the decision process of an allotted assembly modelled on a Deliberative Poll would be a proxy for the informed consent of all citizens. During the report presented to the recent Dublin meeting Peter Stone returned to this point by arguing, with Bentham, that the whole social contract theory of consent was just nonsense on stilts.

Peter recently referred me to Hanna Pitkin’s two-part paper on Obligation and Consent. The paper is hard work, but could be boiled down into just two claims:

  1. Lockean consent theory is not so much tacit as hypothetical – we are deemed to consent automatically to a form of government that accords with the laws of nature (and may legitimately overthrow one which doesn’t, such as a tyranny) – ‘the only “consent” that is relevant is the hypothetical consent imputed to hypothetical, timeless, abstract, rational men.’ (Pitkin, 1965, p. 997). Phrased like this it sounds a tad tautological and Pitkin fails to adequately address Locke’s need to ground such an argument in religious claims regarding the divine provenance of the laws of nature. Absent such shared religious convictions it’s hard to make much sense of this argument.
  2. Pitkin then shifts to a postmodern analysis of the language of consent, drawing heavily on Wittgenstein’s Philosophical Investigations. According to this approach, ‘government and authority are concepts grammatically related to obligation and obedience’, they are part of the same linguistic family. Thus ‘obligations oblige . . . is what the words mean’ (Pitkin, 1965, pp. 47-8). In The Concept of Representation Pitkin adopts an Austinian ordinary language approach to fruitful ends – for example her dissection of the meaning of the phrase ‘descriptive representation’ has important entailments for those who seek to design political institutions based on sortition, but Wittgenstinian arguments on obligation and consent have little consequence in the real world, other than telling people to shut up and put up with it (‘that’s just what the words mean’). If religion is the opium of the people, then Philosophical Investigations is the opium of the over-educated classes – leading to the deluded belief that if you change language you change reality. A more robust conclusion would be that the khazi and the rest-room both smell the same.

It strikes me that a more fruitful approach to obligation and consent would be heuristic, rather than hypothetical or soporific. Such an approach owes more to Hume than Locke as it is based on what concrete individuals would actually consent to (as opposed to how an idealised ‘rational’ person in a hypothetical situation might choose to act). Hume had little time for reason as a transcendental principle, seeking instead to ground it in emotion and interests. A heuristic approach to consent might look something like the following:

  • Citizens are more likely to consent to governments that further their own interests (‘It’s the economy, stupid’).
  • Citizens are more likely to consent to governments composed of people ‘like them’ (if I can’t decide myself, then I’d rather have someone like me decide on my behalf).
  • Decision-making by ‘people like me’ requires balanced information and advocacy. This is the procedure in a law court, where the whole population is deemed to consent to the considered verdict of their peers.

Fishkin refers to such a model as ‘consent by proxy’. Given its heuristic (rather than deductive) nature it is unlikely to appeal to political philosophers but is, perhaps a good example of Jeremy Waldron’s recent argument for a return to an Aristotelian institution-based approach to political theory. Such a division of labour would leave political philosophers at liberty to play around with linguistic hypotheticals such as (in Waldron’s pithy phrase), ‘57 varieties of luck egalitarianism’, thereby freeing political theorists to get on with the real work of designing the sort of political institutions that would be likely to reflect the informed consent of citizens in large modern polities. Such work will involve a partnership between theorists and political scientists but will not provide a congenial retirement home for political philosophers.

51 Responses

  1. Glad you found the article of interest. A few quick points.

    1) Not sure why you say “Pitkin fails to adequately address Locke’s need to ground such an argument in religious claims regarding the divine provenance of the laws of nature. Absent such shared religious convictions it’s hard to make much sense of this argument.” Sure, Locke grounded his claims about the law of nature in religious considerations. But that does nothing to disturb the claim that consent is irrelevant to his argument. You are considered to have “consented” to a government that effectively protects your rights. You cannot “consent” to one that regularly violates them. Actual “consent” seems to do no work at all here.

    2) Yes, the idea of hypothetical or reasonable consent has problems. But it seems to be the only way to keep any meaningful idea of consent in the story. Not sure why you call her argument here “postmodern” and the argument in The Concept of Representation “ordinary language.” The technique seems the same. It’s an argument that says that certain arguments about “consent” simply aren’t using the words the way we ordinarily use them. Dictionary.com defines “consent” to mean “to permit, approve, or agree; comply or yield”. There’s no meaningful sense in which I have permitted, approved, or agreed to either the U.S. Constitution or the Irish Constitution. To talk about consent (as opposed to hypothetical or reasonable consent) is just changing the meaning of the word.

    3) I don’t think it’s fair to say her position is “shut up and put up with it.” Rather, she believes that legitimacy and obligation are linked simply by the meaning of the words. It makes no more sense to ask if one is obliged to obey a legitimate government, according to Pitkin, than it makes to ask if a bachelor is unmarried. Once you have decided a government is legitimate, the obligation to obey it follows naturally because that’s part of what it means to be a legitimate government. Now, that does nothing to settle the matter of what makes a government legitimate, but it does locate the problem in the proper place, I’d say.

    4) There’s nothing wrong with using actual consent as a guide to what hypothetical consent or reasonable consent might be. Indeed, I don’t know how else it would be. But in ordinary life I think it’s meaningless to talk about people giving actual “consent” unless you can say what it would be like if they weren’t giving actual “consent.” What do I have to do to withdraw my consent from the U.S. government? Renounce my citizenship? Riot in the streets? If so, then most Italians consented to Mussolini, most Germans consented to Hitler, etc. It’s different with hypothetical/reasonable consent, where one can imagine whether or not an individual would authorize the institutions in question if they had a real choice. No real citizen ever meaningfully gets a shot at such authorization.

    5) I don’t see how “consent” adds anything to your heuristic approach. Why not just say, governments are better when they involve a broad cross section of the community, and when they take everyone’s interests into account, and when they have balanced information and advocacy? What work does “consent” do here? I’m just not seeing it.


  2. Thanks Peter, glad I didn’t completely misunderstand the piece. Dealing with your points in turn:

    1) Pitkin would agree with you that consent does no work in Locke’s theory, as adherence to the laws of nature is not a matter of the expression of a preference. That’s why I referred to it as “a tad tautological”. It made sense to Locke on account of the religious premises; not so to those of us who don’t share those premises.

    2) and 3). The argument in the paper is postmodern because language to Wittgenstein is a holistic system that may or may not have external referents. Austinians, by contrast, are interested with the effects of language in the real world, hence the emphasis on the illocutionary force of speech acts. Once concepts like “descriptive representation” have received the necessary (Austinian) clarification then we can use this as a guide as to how to construct political systems and then judge those systems by the degree to which they accurately mirror the conceptual constraints. Not so with Wittgenstein’s philosophy, where “descriptive representation” is just a token in the closed system of a language game. Austinians would seek to recast institutions to align them better with the concepts that they are supposed to instantiate, whereas followers of Wittgenstein are more disposed to re-engineer language. That’s why the latter is a postmodern approach to politics.

    5) As my proposal is heuristic, I don’t much care what we call it, consent or whatever. The Paris paper was intended only to show that an allotted assembly would make a better fist of consent than electoral representation. I can’t think offhand of a better word: “legitimacy” is generally a formal criterion, whereas (tacit/heuristic) “consent” implies that (most) citizens don’t disapprove of the arrangements. What word would you prefer to use?


  3. > Once you have decided a government is legitimate, the obligation to obey it follows naturally because that’s part of what it means to be a legitimate government.

    That puts the bar of legitimacy unreasonably high. I could think that a government is legitimate – it was formed legitimately, it uses legitimate procedures, it even acts morally (as I see it) most of the time – and yet disagree with some of the rules it tries to enforce and therefore decide not to obey them whenever I can.


  4. Either the bar is high or (more likely) it’s just an empty tautology.

    Regarding Peter’s call to abandon use of the word “consent” in political theory, the reason is that it originates in the realm of individual volitions. The term “consenting adults” means that all parties willingly partake in a sexual act. If this is problematic in the realm of sexual behaviour (as the Assange case would suggest), it’s a real can of worms when it comes to the aggregation of collective volitions that are required by democratic theory. Rousseau’s solution (the alignment of individual wills with the general will) failed to specify a workable mechanism and led to some markedly illiberal outcomes.

    The liberal notion of consent via preference elections would work were it possible for individual voters to reliably gauge the volitions of the candidates. However the metamorphoses of representative government over time have led to a situation when this is impossible as, in the age of audience democracy, candidates have assumed the role of stage actors. Consent by proxy (via sampling) is, heuristically speaking, an improvement as one would automatically consent to the volitions of one’s doppleganger. Needless to say both theories (election and sampling) presuppose an acceptance of proportionate majoritarian principles — if one’s views and volitions are statistically in a minority then one’s proxies will be extremely approximate and unlikely to prevail. But that strikes me as an inevitable consequence of living in political societies.

    But I don’t think that words like approximate, proportionate, majoritarian etc. mean that we need to give up on the notion of consent. If a group of friends are deciding where to go for dinner and the majority choose the pizza restaurant, the minority who voted for the Indian restaurant would indicate their consent to the majority verdict by tagging along and eating the pizza. Why does the same principle not apply to the larger group of friends that we refer to as political societies?


  5. Authority is a matter of convention, like property rights, kinship structures and pretty well everything else about our lives.
    We are not born free, but helpless, lacking not just the most elementary capacities to survive, but almost all means of communication and interpersonal action.

    We become free individuals by mastering the relevant conventions that enable us to interact meaningfully with others. The existing conventions in the social milieu into which we grow are second nature to us, and we accept them as such. That is the condition of our being accepted by others as participants in social life in whatever role convention assigns to us, submitting to the existing power structure.

    Conventions may be assessed for their efficiency, moral defensibility and so on POST FACTUM, but the fact that they are open to improvement does not suffice to deligitimise them. Especially where they are institutionally entrenched AND ENFORCED, and where the ramified effects of changing them are unpredictable. Most changes in conventions happen as a result of minor departures from existing norms. That almost always takes place in virtue of possibilities opened up by ambiguities in existing conventions that significant bodies of individuals find it advantageous to exploit, gradually shifting the patterns of practice.

    Where consent comes in is when there is question of CHANGING CERTAIN SPECIFIC CONVENTIONS DELIBERATELY AND COMPULSORILY. That never happens by universal consent. There are always interest groups that find it very disadvantageous (often mistakenly) to lose their present position and there are always differing views about what to change and in what direction to change existing conventions. The relevant question about the moral justifiability of imposed change is not consent but consequences. imposed changes need more robust justification.

    On the one hand, even almost unanimous consent does not justify the imposition of practices that deprive people of certain moral rights and freedoms. One is not entitled to consent to a totalitarian regime, even if one likes the idea. On the other, it can be right to impose needed restrictions on the rights of slave owners against their will, even against the majority opinion in certain jurisdictions, provided the consequences of doing so are not disproportionately damaging to those who should benefit from the change.

    My point is that the conservatives are right to emphasise that political change is NECESSARILY EVOLUTIONARY. Like organic evolution it must proceed by piecemeal modifications of existing processes and procedures to achieve particular improvements to the existing situation. The philosophical illusion that it is possible to guide choices in such matters by some set of principle derived from an analysis of an ideal situation is both ridiculous and pernicious.

    Where the conservatives are wrong is in thinking that in a complex modern social order the gradual processes of largely undirected evolutionary change is sufficient to deal with the problems that beset us. Where they are right is in seeing that increasing the powers of nation-states is neither adequate nor desirable.

    The unspoken assumption behind the consent debate is that we are stuck with the dilemma of assigning all power of decision in matters of common interest to such states or leaving them to markets and social fashion. What we need are specific specialised authorities that are governed by the specific constituencies that are relevant to their functions. The point is to get away from the situation in which
    issues are decided not on the merits of the case, but on poser trading in which the primary consideration is cobbling together alliances against other alliances in pursuit of generalised power.

    Such authorities could only grow out of existing governmental institutions by piecemeal changes, making statutory authorities
    and relevant international bodies.more independent from “politics” but more responsible to the interests they are supposed to serve. UNTIL POLITICAL THEORISTS ACCEPT THAT THIS IS THE WAY OUT OF OUR DILEMMA WE’LL GO ON PISSING IN THE WIND.


  6. JB: >My point is that the conservatives are right to emphasise that political change is NECESSARILY EVOLUTIONARY. Like organic evolution it must proceed by piecemeal modifications of existing processes and procedures to achieve particular improvements to the existing situation. The philosophical illusion that it is possible to guide choices in such matters by some set of principle derived from an analysis of an ideal situation is both ridiculous and pernicious.

    Agree 100% (and good to hear from you again John). I’m as opposed as you to the monotheistic fundamentalism that underlies the more strident posts on this site. That’s why I prefer heuristics to hypotheticals. My concern, though, is with ongoing consent to the day-to-day political process, as opposed to the legitimisation of change (constitutional or otherwise). It strikes me that “trust” is a workable heuristic for consent. I consent to the decisions that my accountant makes (on my behalf) because I have no knowledge of tax law (and no wish to master Tolley’s Guide) and trust both his professional judgment and that he is acting in my interests. If he makes the wrong decisions then I (warily) transfer my trust to another accountant, hopefully before the onset of bankruptcy.

    Can one scale this up? Both Locke and Rousseau viewed government as a revocable trust relationship (Hobbes eschewing all such moral language in favour of the geometrical requirements of the political system). According to Manin the first stage of representative government (parliamentary democracy) was government by trustee — local “notables” were elected on the basis of their perceived character and competence. Universal franchise led to the second stage (party democracy), and this was also characterised by trust — electors consented to the decisions of their chosen party as they trusted it to represent their interests. Although such representatives were delegates rather than trustees, nevertheless consent was still based on trust. However the breakup of homogeneous class-interest groups (and the end of the age of deference) led to the replacement of party democracy by audience democracy, and it’s unclear exactly what it means to trust a stage actor who is writing her autocue script in real time based on what the focus groups tell her professional advisers the audience wants to hear.

    Consent-by-proxy is an attempt to heal this breakdown of trust. We would all trust our own informed judgment (informed being the operative word), so would therefore consent to the informed judgment of people like us. That’s about as good a heuristic as one can hope for in a mass-democracy context. If Manin is right that “consent” was the reason for the shift from sortition to election, then the manifest failure of his third stage of representative democracy is a cue to to return to sortition as a better heuristic for the informed consent of citizens.

    PS the consent-by-trust relationship is not exclusive to democracy — it would also apply to the rule of Plato’s Guardians, Good Queen Bess, the Führer (yes the German people did consent to Nazi party rule) or the Ayatollah Khomeini. It’s just that we have good reason to believe that the informed judgment of a jury of our peers would lead to better epistemic outcomes than any of the above.


  7. John Burnheim wrote:
    “The philosophical illusion that it is possible to guide choices in such matters by some set of principle derived from an analysis of an ideal situation is both ridiculous and pernicious.”

    I would like to add a caveat (which he may agree with).

    While blue prints or master plans of political theorists or revolutionaries may be ridiculous or pernicious, they may still serve a useful purpose. Institutional evolution often starts in the mind, and the ideas of theorists may act as the random mutation that allows for an evolutionary change. In other words, if nobody ever imagined or proposed some new institutional form, it could never take root. Only after the idea is put “out there” is there a chance that it can catch on.

    If nobody ever heard of an alternative to elections as a means of selecting representatives, then we would never see any alternative being tried.


  8. My impression is that John is contrasting principles deduced from idealised thought experiments (Rawls’s original position being a good example) with principles extrapolated from everyday experience. My own conversion to sortition was the result of a casual (and slightly drunken) conversation at a party. In John’s case it was a reaction against radical experiments in mass student democracy during the early 1970s. And the Greek invention of democracy was the result of contingent historical and geographical factors, as opposed the attempt to put into practice an idealised vision of political equality handed down from Platonic heaven. According to Finlay Aristotle merely ‘adumbrated’ the concept of democracy from the messy business of practical politics. The approach that John and I both favour tends to be along the lines of if something works in one context, then why not try another? Idealised thought experiments tend to be a useful source of normative ideals but can be distinctly unhelpful when it comes to practical politics as the tail generally ends up wagging the dog.


  9. Keith. I heartily agree with your comments about trust, and I would like to pursue the matter further.
    There are three sorts of trust that are relevant here: 1. trust in an individual person, 2. trust in a type of person (eg a professional) and 3. trust in an institution or process.
    All three are deeply dependent on the conventions of a culture and belief in their grip on all concerned. So we trust professionals to the extent that we believe that if they are in good standing in their profession they are (a) likely to be competent, to value their professional esteem, including their self-esteem, much more highly than any competing motivation, and (b) that there is a mutual understanding of what is expected of each of us in our mutual dealings, eg that the professional will give personal attention to our needs irrespective of profitability and that we will give them a certain loyalty and respect.
    The sort of trust we need in our relations with people who make important decisions on our behalf depends principally on our confidence in them as professionals, which in turn depends on our confidence in the institutions that educate, supervise and reward them for sound performance.
    It is very difficult to maintains such trust where the pursuit of power and of money are the most insistent motivations that are the focus of day to day activities in the relevant institutions. In such a situation, even though the avowed norms are admirable, the suspicion that htey are honoured only hypocritically is unavoidable, and trust becomes mistrust.
    That is the case with our electoral systems, because they are dependent in all their day to day workings on power trading and on money for lobbying and advertising. The hope of sortition is that it will eliminate these distortions and restore the power of conventions of responsibility and service.
    However, that is not enough. We also need competence. One of the problems about centralised government is that politicians can have little knowledge of or concern for most of the issues on which they are expected to vote. So their vote is more likely to be determined by opportunistic deals between factions or ideological prejudices than by the merits of the case. Even if you put non-professionals in the same situation, they will almost inevitably fall into the same pattern of decision-making. They will fall for the smart operators.
    I persist in thinking that in the long run we need to recognise that issues need to be disentangled from each other and from the quest for supreme power. The major say on any issue must be that of the people who are most directly and materially affected by it. They have reason to get to grips with it and bear the consequences of the inevitable compromises that have to be made.
    Of course, i can’t say in advance just what structural changes would work in this way, it is a matter of changing existing institutions, practices and opinions in piecemeal ways.


  10. > We also need competence. One of the problems about centralised government is that politicians can have little knowledge of or concern for most of the issues on which they are expected to vote.

    I don’t think competence is an issue. Almost any reasonably sized group of people given the incentive and resources to solve a problem would follow a course of action which would be deemed effective by most people sharing their values. I don’t think elected politicians are an exception.

    When politicians follow a course of action that is not effective for serving the public it is not because they are incompetent, but because they have no incentive to serve the public.

    > Of course, i can’t say in advance just what structural changes would work in this way, it is a matter of changing existing institutions, practices and opinions in piecemeal ways.

    This won’t do. With all the limitations of theory, giving up on theory altogether leaves the public wandering aimlessly in the space of policy, following haphazardly political slogans and fads whose ineffectiveness in confronting entrenched and organized interests leads to more frustration and distrust. The ineffectiveness the OWS movement is a demonstration of this phenomenon.


  11. This is promising as I find myself agreeing with both John and Yoram.


    Pitkin’s age of trustee representation (“parliamentary democracy” in Manin’s terminology) roughly equates to the period between Burke and Gladstone/Disraeli. The reason that I prefer Pitkin’s earlier (Austinian) approach to ordinary-language philosophy, rather than the postmodern (Wittgenstinian) approach is that it is grounded in concrete practices. The notion of trustee representation derives from law — the officials who manage a trust fund on behalf of minors or others who are not deemed competent to make their own decisions. Trustee representation in politics presupposes incompetents voting for trusted ‘notables’ — selected on the basis of ‘aristocratic’ characteristics such as proven competence, honesty, integrity and leadership charisma — to make these decisions on their behalf. Aristocratic characteristics were at the time associated with wealth as this was a prerequisite for education, leisure and (according to Burke) the ability to decide in a disinterested way. Madison went further and associated disinterested judgment with the professional classes (rather than the landed interest), and there may well have been an implicit Calvinist notion that those who showed the outward signs of religious election were more worthy of political election. Unsurprisingly this led to a parliament or congress that was largely comprised of wealthy professionals (or male white lawyers).

    The notion of trustee representation came under attack during the 19th century. After Marx, Freud and Durkheim it was hard to believe in the notion of disinterested judgment and, as a consequence, trustee representation gave way to the representation of interests (Pitkin) or party democracy (Manin). That worked OK for nearly a century (although the focus on interests arguably meant competence being reduced to the role of second fiddle). But then the homogeneity of class interests started to fall apart and the hollowed-out political parties were taken over by carpetbaggers and refashioned to suit the covert interests of their leaders (or, more charitably, simply in order to tell voters what the focus groups told them they want to hear). This period came after Pitkin’s book, but has been theorised by Manin as the age of “audience democracy”. Manin argues (but fails to convince) that this a perfectly valid metamorphosis of representational government, but qualities like trust and competence are now in the hands of spin doctors and image manipulators who are only concerned with the next day’s newspaper headlines. This can only be understood using the tools of postmodern philosophy — where all that matters is the language — Austin is of no use at all.

    So how do we reclaim a) trust and b) competence? I can’t see us ever returning to Burke and Madison’s concept of trustee representation, and it’s impossible to trust political parties that are driven by opportunism and focus groups (Labour presenting itself yesterday as more eurosceptic than the Tories is a case in point). Which brings me back to the notion of consent by proxy. We’re all educated now and most of us would trust our own INFORMED judgment. So if I can’t judge myself then I’d rather have someone like me making an informed judgment on my behalf. If most proxies are like me then I win; if not then I lose — that’s what we call democracy.

    I would certainly consent to such a procedure, even though the connection between my own choices and the final outcome is only a virtual one. In fact this ties in well with Burke’s notion of virtual representation — the citizens of Manchester (who had no MP) were represented “virtually” by the MP for Bristol and other similar conurbations. The only difference between Burke’s deliberative assembly and a modern allotted chamber is that the former did not require accurate statistical representation, such was Burke’s faith in the exchange of reasons. Nowadays we are less sanguine about the power of discursive exchange and rightly insist that interests be represented proportionately.

    Proxy representation may work for the sovereign body, where judgment alone is required. Not so for executive officers, where proven ability is the selection criterion. The separation of legislative and executive powers has already benefited from over 250 years of theorising so I’ve no need to argue the case that government is a delegated function that should be subject to recall by the sovereign body. Government officers should be appointed, just as in any other organisation, and should be held to account by an allotted assembly. Government was viewed by Locke and Rousseau as a delegated trust function (where the trustees could be removed and replaced according to the will of the trust beneficiaries) and this is the domain in which the 18th century notion of trust (as a legal function) is still relevant. It just doesn’t work any more for (electoral) representation.

    Where I disagree with John is that I think his notion of demarchic committees conflates judgment and executive action (sovereignty and government) in a particularly British way. And I think we ALL have a direct and material interest in most things, as politics is largely a matter of taxation and spending, so I’m sceptical regarding the role of volunteering.


    I agree with your focus on interests and incentives, although I do think competence is an independent variable. And I’m glad that you now view political theory as a causal factor (rather than an epiphenomenon of underlying interests). The argument John and I are having with Terry is over whether theory should be top-down or bottom-up; idealised thought experiments or adumbrating from existing experience.


  12. I find much of this very sensible, but at the risk of sounding like a broken record, I think very little of it is really about consent. I totally agree that we should be creating institutions that make us say, “Well, yes, let’s do that, that sounds great.” If the political system really worked well, I’d expect most people would say that about the institutions governing them. One might even say, “I WOULD consent to that,” to use Keith’s phrase, although it is important to note that this is most definitely hypothetical consent. But that’s as far as “consent” goes.

    One of the reasons I harp on this point is that there are many situations where actual consent really does matter. It’s wrong to take money from my wallet, or enter Keith’s home, or make sexual advances on a person, without their consent–not their hypothetical consent, not their reasonable consent, but their actual consent. It really doesn’t matter if you think the reasons someone doesn’t want to have sex with you are bad ones–if they didn’t consent, that’s the end of the story. I fear that if one pretends that any form of actual consent takes place on a regular basis with our political institutions–or that it could do so, in any meaningful sense, even under the best institutions–then it cheapens the currency of consent. And in some situations, it’s too valuable a currency to cheapen.


  13. Peter, surely the examples you choose (pertaining to “me”, “Keith”, “Julian Assange’s sexual partners” or whatever) are not the domain of the sovereign legislature. The concern of the legislature is with the general conditions of what makes an act theft, rape or some other crime, not the particular instances. So the legislature would decide that it’s wrong for anyone to take money from other people’s wallets unless invited them to (but OK to sequester 20-50% of their salary in taxation for the common good), wrong for anyone to enter one’s home without invitation (but OK if in possession of a search warrant), and wrong to make unwanted sexual advances at work (but OK to to try it on in bars and nightclubs). The function of the legislature is the drawing up of these universal rules, and consent-by-proxy is a perfectly legitimate way of making these decisions. This maps directly onto Rousseau’s distinction between the sovereign legislature and the executive magistrate, the latter’s job being to apply the universal rules to particular instances. All the examples you provide — both here and in the Dublin paper (renting out your flat without your permission) — relate to the latter. Or have I missed something?

    The reason I prefer heuristic to hypothetical consent is because it involves decisions made by actual human agents (proxies), who act on the basis of a combination of reason, interests, gut feeling and prejudice. Hypothetical consent, on the other hand, relies on the notion of the the idealised “rational” agent and fails to recognise the complex and hybrid nature of human volition. Political philosophy has come up with far too many of these artificial creatures and that’s why I prefer a bottom-up to a top-down approach. The reason consent-by-proxy is not hypothetical is because your proxy actually stands for you. If you were there yourself you would consent (or dissent) in person, so your actual consent in the proxy situation is dependent on the presence or absence of people like you in the assembly. The only precondition is a commitment to majoritarian decision making — as I argued earlier the friend who preferred Indian food actually consents to the majority verdict by tagging along and eating the pizza. The same would be true if a proxy were involved. There’s nothing hypothetical about this and to describe it as tacit is a misunderstanding of the nature of the proxy relationship.


  14. One key element of John’s demarchy model is the establishment of numerous allotted bodies, each tackling a narrowly defined policy area, rather than one super-legislature, which handles everything and becomes a power center (prompting ambitious power-seekers to fill the chamber). I agree with this, not only as a way of stemming the flow of megalomaniacs into decision-making positions, but also as a way of improving effective competency. By avoiding vote swapping (“I’ll vote for your new road if you vote for my farm subsidy bill.”) and the practical impossibility of members of an all-purpose legislature becoming well versed in the myriad issues that arise daily, numerous separate bodies allow people of average ability to delve into narrowly defined matters deeply enough to genuinely gain an understanding. “Competence,” within existing elected legislatures is 90% about the skills of public relations and political maneuvering, rather than policy.

    The one place where I (and others on this Blog) part with John is his notion that we should have those with close connection to a matter dominate that decision making body. It is often better to have disinterested and impartial outsiders involved. I think random selection among the entire population makes sense unless the issue is uniquely encapsulated (e.g. the members of the Water District).


  15. One thing I’d be interested to hear from John is how the individual demarchic committees deal with overlapping policy areas and, more importantly, how the fiscal requirements are addressed. If I remember correctly from his book it involves junking the general taxation powers of the state in favour of a Henry George-style land tax. Unfortunately this would take us well into the realm of utopian studies, as opposed to the cautious piecemeal pragmatism he has been advocating here.


  16. My book was deliberately utopian, in the hope of inspiring people to think that by going in the direction I advocate we might arrive at a polity that is very much preferable to what we have. But in the end, of course it is a matter of keeping the boat afloat while we try to fix certain problems that threaten to overwhelm us.

    I distinguish sharply questions of responsiveness in the delivery of the public services we want from questions of justice and efficiency in the allocation of power to manage specific services. The latter must express the moral and economic wisdom of the community, the former its needs and preferences. I propose that the former would be quasi-judicial in style and would be staffed by people who in the opinion of those who have worked with them have the necessary intellectual and moral qualities to do that sort of work. Among their tasks would be adjudicating boundary disputes and coordination failures that the administrative bodies failed to deal with and to hear representations about the need for restructuring them. In all sophisticated bureaucracies there are such bodies, though they are often fragmented, inadequately resourced, excessively politicised and more concerned with defending the turf of that bureaucracy than with making it more equitable and efficient.

    I would hope that we could get away from the syndrome of excessive regulation that large bureaucracies have recourse to in the hope of avoiding possible abuses on the basis that we could trust the sort of supervision I envisage to build up a culture in which good sense would be respected and expected.

    There will always be need for collaboration between authorities, and the normal practice for assuring it is information and negotiation, but, failing that arbitration by the judicial authority.

    All of these matters could be addressed within the framework of the existing state if only there was sufficient public pressure on politicians. The crucial problem is finance, pipers and tunes. But in fact if a program is popular, it will get financed, and the level of finance is mainly a matter of precedent. However, one could envisage options in which certain authorities might offer a higher level of service than the prevailing average and have the Tax office attach a corresponding surcharge to the taxes of those who can access those services.
    The fundamental problem is consumerism. People are used to buying on the basis of advertising and just taking their custom to the competition when they are disappointed. Here in Sydney there is a foundation called New Democracy that attempts to move things in the direction of Demarchy. It attracted support from two respected former state premiers. One of them, having been appointed to review Sydney’s transport system, proposed setting up a citizen jury to deliberate on the problems. The popular reaction was that he was just fudsging his responsibility. “That’s what we are paying you to do”.


  17. If one reads Locke more carefully one finds that:

    1. While the basis for consent to the social contract can be an explicit agreement among adults, the usual basis for consent is entering onto or remaining on the territory of the society (which when combined with an exclusive territory becomes a “state”) past the age of consent. When you as a adult step across the border into a state you consent to its “constitutional” laws, and cease to be subject to them when you step out of its territory.

    2. The primary and default process for consenting to the social contract is being inducted into it as a child by parents or guardians, a process called “socialization” by human development scientists. It starts with a “filial contract” (“You be a good child and I’ll be a good parent.”) and graduates into the social contract by the transitivity of the social contract relationship (“If A is in the social contract with B and B is in the social contract with C, then A is in the social contract with C.” “You, child, are in the social contract with me, your parent, and I am in the social contract with our family, neighbors, and countrymen, therefore, so are you.”) The transitivity is not always perfect, and may psychologically weaken with social distance, but that is the logic.

    As for consent to representation, that is based on consent to the law that defines how the representatives are selected, which could be by sortition, not to the particular instance of the selection, as long as the procedure is scrupulously followed.

    Votation as a procedure for selecting representatives invariably magnifies the weight of majorities at the expense of minorities, compared to what the results would be if every issue were decided by votation, a result more closely approximated by selecting representatives by sortition.

    A challenge for advocates of sortition is to provide procedures to weight the selection of representatives in favor of those with greater virtue, while avoiding favoring special interest groups. The promise of votation advocates is that it offers some weighting for virtue, even if that tends to also weight for special interests. The traditional solution has been to conduct a multi-phase process in which random selection alternates with screening for virtue, usually by having the lottery selectees elect the candidates for the next random round. They have an incentive to elect the most virtuous when they can’t elect a crony, because they can’t be sure the crony will win, so electing the most virtuous is a better bet.


  18. > They have an incentive to elect the most virtuous when they can’t elect a crony, because they can’t be sure the crony will win, so electing the most virtuous is a better bet.

    Once the elites control the process (having gone through the electoral filter), then “virtue” would be defined by its interpretation by the elites. At this point it would be very surprising if the “virtuous” would be serving the interests of the population at large.


  19. Jon, thanks for the clarification on Locke. It’s worth noting that the implicit consent mechanism that you describe would apply equally to any political system — monarchical, elective, sortive etc, so there is no reason to accept Manin’s argument that election as a mechanism instantiates consent in a way that sortition cannot possibly do. But I don’t think consent is a concept that we can do without, and I’m using consent in the stronger sense (consent to legislative outcomes) rather than just the legitimacy of the political institutions. In this sense consent-by-proxy does a better job than the very approximate consent instituted by electing political representatives.

    I think we are perhaps using votation in different ways (you probably mean just plain voting). My usage is derived from Yves Sintomer’s chapter in the Delannoi and Dowland collection. According to Yves, votation is a unique aspect of Swiss direct democracy whereby voters indicate their preferential ranking of a number of citizen initiatives. The top x proposals would become parliamentary/congressional bills, for judgment by an allotted jury.

    I’m a little alarmed by your quest for virtue in political representatives. This can give rise to all sorts of misguided policy initiatives — Blair certainly viewed the Iraq war as an act of righteous intervention and believed he had divine guidance (as, supposedly, did GWB). If the role of allotted representatives is to sit as a jury, voting up/down legislative proposals, there is a good case to be made for an accurate cross-section of the citizen body, warts and all. Virtuous people frequently believe that others are motivated in the same way as themselves and this has led to all sorts of unintended consequences in (for example) the field of welfare payments. If there is a role for virtue, it would be in appointed government officers rather than politicians — what we refer to as the “public service ethos”. Public servants should be motivated by higher considerations than just pay and benefits.


  20. Just been reading the draft of Dario Castiglione’s chapter on the origin of civil government in James Harris (ed) British 18th Century Philosophy (OUP, forthcoming), which confirms Jon’s interpretation of the natural law tradition of consent. Castiglione also views Hume as a turning point from speculation on the acts of a hypothetical rational individual to a natural history of civil government and its utility. This all makes more puzzling Manin’s original claim for the triumph of election, as social contract theorists were more interested in consent to the institutions of government rather than to particular outcomes. A shift of emphasis to utility does suggest an increasing focus on outcomes and (I would argue) consent-by-proxy is a much better way of indicating approval of outcomes than preference elections. Given that, as Hume and, later, Adam Smith argued, allegiance is as important as utility we could never do without the notion of consent as a government could be formally legitimate without it benefiting from the allegiance of citizens. According to Smith allegiance used to be on account of a deference to those who show “natural” leadership qualities, but in a demotic age like our own it would be better directed to an assembly comprised of folk like us (or the winners of the X Factor).


  21. Keith, I may be continuing to sound like a broken record, but to the extent that the subject is consent, the responses just strike me as non sequiturs. I give examples where you agree that actual consent (not hypothetical consent, reasonable consent, or consent by proxy) is required–taking money from my wallet, etc. You say that this is not relevant because those are decisions of a specific nature, not the general nature that characterizes laws. And so what? How does that change the definition of the word “consent?” If you need my “consent” in the take-money-from-the-wallet case AND the law-making case, how is it that you can get away with one type of “consent” in the one case and not the other, without simply changing the meaning of the terms? What is your highest level understanding of the meaning of the word “consent” that lets you count both the need to ask my consent before taking money from my wallet and government by consent of the governed as two examples of the same phenomenon? And if they are both forms of “consent,” why? What makes that so?

    When it comes to consent-by-proxy, you actually say a bunch of different things that don’t fit together in any way I can see. You say “The reason consent-by-proxy is not hypothetical is because your proxy actually stands for you.” That sounds like another non sequitur–so what? He stands for me? What does that mean? You then immediately say “If you were there yourself you would consent (or dissent) in person, so your actual consent in the proxy situation is dependent on the presence or absence of people like you in the assembly.” Wait–so hypothetically, I would have consented in person, therefore I really gave my ACTUAL consent so long as there are people like me in the legislature? How do you know I would have actually consented if I was there? Just because there was somebody like me in the legislature? How does that prove anything? And how does this hypothetical consent magically become actual consent without simply changing the meaning of the word.

    Now, the presence of someone like me in the legislature might ensure that my interests are represented, by having someone with similar interests present. But that story is independent of consent; my interests can be represented without my consent, and I can surely consent to policies that are not in my interests.

    And then we get “The only precondition is a commitment to majoritarian decision making — as I argued earlier the friend who preferred Indian food actually consents to the majority verdict by tagging along and eating the pizza. The same would be true if a proxy were involved. There’s nothing hypothetical about this and to describe it as tacit is a misunderstanding of the nature of the proxy relationship.” Two big problems here. First, this negates everything you just said. If consenting to the decision-making procedure is enough to generate consent to the system, why does consent-by-proxy matter? If all you need is my consent to the system, then surely ANY system receiving my consent (monarchy, demarchy, whatever) enjoys my consent, regardless of whether there’s any descriptive representation or not. Second, you then add the caveat that I’ve made a “commitment to majoritarian decision making” as long as I somehow tag along and enjoy the benefits of the system. But in response to Jon, you correctly point out that “the implicit consent mechanism that you describe would apply equally to any political system.” Every tolerably stable system enjoys “consent” in this incredibly tacit manner, and so the concept does no work here.

    Perhaps I am missing something here, Keith, but you seem to understand consent in at least three different ways. There is actual consent, as is needed in my wallet example–which no political system ever enjoys. There is tacit consent a la Locke–which every stable political system (Maoist China, Czarist Russia, Nazi Germany) enjoys. And then there is a grab bag of ideas like representation of interests, ideas, perspectives, etc., which doesn’t seem to depend on consent in any meaningful sense. These three things seem to me to be so totally different that it does more harm than good to try and lump them together. And once you separate them properly, the concept of consent can quite properly be eliminated from discussion without harm.


  22. Peter,

    The trouble with the word consent is that it is derived from individual acts of will (as in the example of consenting adults that I referred to earlier). But if you remove it from the “grab bag” you run the danger of a political system that may be formally legitimate but does not embody the approval of its citizens. This is why I disagree with your claim that “the concept of consent can quite properly be eliminated from discussion without harm”. What word would you prefer to use instead (assuming you disapprove of political systems that rely on coercion, indoctrination or bread and circuses)? I can’t think of a better word, so the challenge is to take a word that was designed for individual acts of will and apply it at the collective level — this is, as you rightly point out, a serious challenge. The reason that I argue for a heuristic (rather than hypothetical) solution is that this is the only way I can conceive of resolving the problem that doesn’t reduce it to a single idealized function (rationality, prudential self-interest, divinely-inspired laws of nature or whatever). Consent-by-proxy is a combination of informed judgment along with interests, tastes and prejudices — such being the mixture of faculties that individuals bring to bear on any decision-making process. An allotted assembly deciding the circumstances in which it is lawful to remove money from citizens’s wallets (via taxation) would arrive at their decision based on a combination of the above. Can we all be deemed to consent to the outcome? Yes, albeit tacitly. Note that in the Lockean case tacit consent is limited to the system of government (monarchy, democracy, lottery, aristocracy etc.), rather than legislative outcomes, so consent-by-proxy does a lot more work than Lockean consent.

    To take the example of the group of diners, minority consent is not limited to the decision-making process but extends to the actual outcome — ie eating the pizza. Scaling the process up, if an allotted assembly were to decide that, in order to encourage community cohesion, all citizens should eat the same food one day a week, then if the votes of my proxies (who preferred Indian food) were in the minority then I would indeed consent to the outcome by eating pizza every Friday (although fish might be a better example in your adopted homeland). You might seek to argue that the consent mechanism is tacit, but this would not be the case if you can swap me and my proxy any amount of times and still end up with the same outcome. I can even imagine Rousseau agreeing that this is a valid method to arrive at the general will via a process of sampling, and the general will must embody the informed consent of all citizens (even those who voted against the majority).

    What’s the difference between agreeing/dissenting directly and agreeing/dissenting indirectly via a proxy? — the outcome is exactly the same. I appreciate this adds another assumption — that there is an affinity relationship between “me” and “people like me” that is statistically robust over a period of time. Attitudes, beliefs, interests etc do not vacillate randomly but can be represented with some degree of consistency — if not then public opinion polls would be entirely meaningless.

    Of course if you think individuals decide everything on an arbitrary whim, without any consistency or predictability, then it matters little what decision rule you implement. But if you don’t, then someone “like me” will, by definition, behave like me, so her speech acts can be taken as a reliable proxy for my own. I’m not sure why you find that problematic in any sense other than the tautological.


  23. P.S. I should make it clear that the notion of consent-by-proxy is not my own (it was lifted straight out of Fishkin’s last book), just as my understanding of the limited nature of the descriptive mandate is taken straight from Hanna Pitkin’s Concept of Representation. Needless to say I agree with both of them and I don’t think I’ve misrepresented their arguments. Fishkin and Pitkin are both far more eminent political theorists than me (I’m just a student). I just wanted to clear that up in case it appears as if I’m taking some sort of idiosyncratic position.


  24. Peter: “Wait–so hypothetically, I would have consented in person, therefore I really gave my ACTUAL consent so long as there are people like me in the legislature? How do you know I would have actually consented if I was there? Just because there was somebody like me in the legislature? How does that prove anything? And how does this hypothetical consent magically become actual consent without simply changing the meaning of the word.”

    Hypothetical is not really the right word, as (in this context) it is reliant on an idealised model of human behaviour. Social contract theorists like Hobbes and Rawls have a particular model of how a hypothetical rational risk-averse agent might choose to act. By contrast the argument for consent-by-proxy is tautological — the fact that my proxy acts like me is true by definition. So the consent is actual, in so far as it makes no difference at all whether I, or one of my proxies are present — the result would be exactly the same. Apart from this tautology, the argument rests on three conditions:

    1) That there is an affinity relationship between me and people-like-me that is robust over time (i.e. that decisions are deliberative — in Bob Goodin’s sense of deliberation within — rather than just the outcome of some random whim or chance event). Existentialism makes for interesting leisure reading but is of little value in the forum.

    2) The actions of the allotted assembly are limited to aggregate functions such as voting and are not distorted by individual speech acts. If individual speech acts were permitted, then the interchangeable proxy argument would fail. This being the case advocacy (arguments in favour of pizza or Indian dining) would need to be externally sourced and consistently balanced.

    3) Simple majoritarian decision making, resulting from the secret vote.

    Given the above three assumptions (and a sufficiently large sample) then the decision outcome would be exactly the same every time. So you’d better get used to eating pizza (although culinary fashions do change over time, and Indian food may well become popular again).

    So let’s restrict the word “hypothetical” to Lockean consent; consent-by-proxy is better described as “tautological” or “heuristic” (given that it results from a grab-bag of factors, rather than an idealised model of human behaviour). If human beings evolved into the state when self-interested risk-averse calculators were in a minority this would rule out Hobbesian or Rawlsian consent but would have no effect whatsoever on consent-by-proxy as no hypothetical model of human behaviour is involved.


  25. This discussion is reaching for the concept of the ancient “law of agency” that underlies almost all legal systems. A principal delegates limited powers to an agent, and as long as that agent acts within those limits the principal is legally deemed to have consented to them and is liable for the consequences. If the agent acts ultra vires, beyond his delegated powers, then that is no longer with the consent or liability of the principal, so long as the principal intervenes and reins in or removes the agent as soon as he becomes aware of the usurpation. If he is aware and does not intervene, that is legally deemed consent to the usurpation. If he is unaware or negligent in his supervision, when he should have known, he remains liable, because he is deemed to have consented.

    These concepts are more clearly exhibited in military justice systems, where officers, who are both agents and principals over other agents, may be prosecuted for things like “dereliction of duty” and “failure to supervise”. These are among the “high crimes and misdemeanors” for which officials may be impeached and removed under the U.S. Constitution.

    So, for example, if a property owner appoints a leasing agent to lease and manage a house, he consents to any lease agreement that is made within the bounds of the agent’s authority. But if the agent lets a buddy stay in the house for free, the buddy may thereby acquire a temporary right of possession until the landlord can evict him, and if the manager agent orders the house be torn down instead of painted, the landlord may nevertheless be held liable for the cost of the demolition, and may have as his only recourse a suit for damages against the (now former) agent.

    The relevant use of the term “consent” is as a legal term of art, with a long history going back to before we have records.

    By the way, if I had article posting privileges and not just comment posting privileges I could contribute much of interest to the members of this forum.


  26. > By the way, if I had article posting privileges and not just comment posting privileges I could contribute much of interest to the members of this forum.

    Happy to set you up – just sent you an invitation to become a contributor. Let me know if you encounter any issues.


  27. Jon,

    The law of agency certainly applies to elected representatives, but I don’t see how it can possibly apply to statistical representation as there is no principal-agent relationship involved in the latter. Indeed this is the most persuasive argument in favour of the retention of elections, as the principal-agent relationship applies (irrespective of whether you adopt a trustee or delegate perspective). However in the case of allotted representation a small sample of the principals take it in turns to decide in person, so the law of agency does not apply as there is no mechanism to hold them to account. There are no agents, only principals.

    Allotted representation is the paradigm example of getting the government we deserve as it relies on the wisdom of crowds and this can cut both ways. That’s why I’ve argued all along for a mixed system of government that includes an appropriate role for professional and elite contributions and these would continue to be held to account by legal and electoral mechanisms.


  28. It is consent to the process that establishes agency, not direct appointment of the agent by each principal. Otherwise elected officials would not be agents of those who didn’t vote for him.

    A randomly selected trial jury certainly consists of public agents, who may be penalized if they fail to attend and deliberate. They may not, of course, be penalized for the verdict they reach. See Bushell’s Case.


  29. A new article on Rawls by Mark Bevir casts interesting light on the contrast between hypothetical and heuristic approaches. Bevir views Rawls’s project as a modernist attempt to find a new foundation for ethics, given the absence of religious belief (his wartime experiences leading to an erosion of his earlier Christian faith) — this is of interest in the light of the recent scholarly emphasis on Locke’s need to ground his theory of consent in divinely-inspired laws of nature. Modernism involved the attempt to find a replacement for such objective foundations and most philosophers (including Rawls) turned to logical positivism for an analogy between scientific and ethical reasoning: the very use of the terms “thought experiment” and “hypothetical” in philosophy indicates the influence of logical positivism. Rawls’s thought experiment (the Original Position) led to a social contract based on the abstract and ahistorical claim of what a hypothetical rational agent would consent to if deciding from behind a veil of ignorance. Rawls was later influenced by Wittgenstein, Quine and other post-analytical philosophers but this could almost be seen as a search for a new form of foundationalism — reducing ethics to the study of syntactic and semantic networks. Once moral philosophy has been reduced to something else (the will of god, scientific facts, the meaning or words etc) then there’s not a lot more to say on the subject. Heuristic consent-by-proxy, by contrast, is truly non-reductive and anti-foundational as there is no attempt to explain human choices in terms of anything other than the preferences of concrete human agents.

    M. Bevir and A. Galisanaka, ‘John Rawls in Historical Context’, History of Political Thought, XXXIII (2012), pp.701-725.


  30. I agree with “constitutionalism” that social contract theory focuses on consenting to the process rather than the outcome. My claim is that consent-by-proxy is different in that it is a real-time way of indicating consent to particular legislative decisions. In this respect it might be helpful to view it in terms of informed direct democracy (“what we would have chosen had everyone deliberated”, as Fishkin puts it) rather than as representation conceived in principal-agent terms. This is why it’s so puzzling that Manin views the triumph of election as an outcome of the natural law theory of consent. Perhaps this is because he views it in terms of its historical use — rotation/impartiality rather than statistical representation. The latter is a modern innovation and it’s unfortunate that the debate has been coloured by archaic considerations.


  31. Let me put in my two cents with regards to the discussion between Peter and Keith.

    First, if we have a randomly-selected legislature, we might ask what actual citizens would think of it. Keith you seem to be under the impression that a move to such a body would increase the approval that citizens would have of government institutions. I suppose one thing we could say here is that there would be an increase in tacit consent of our institutions among citizens. I leave in the middle here if we would want to invoke the idea of tacit consent, or perhaps some more general notion such as that citizens would simply have a more positive attitude towards their institutions.

    Keith, IF your empirical claim is true, then I would agree that this gives some kind of value to a randomly-selected body over an elected body. It could be intrinsically valuable, but it might also be valuable because we would expect less riots and protests against government or some other types of consequences for our social order.

    Yet, whether this empirical claim is true is really pure speculation to my mind. Keith you note that you would be much more approving of such a body, but it is anyone’s guess what citizens might think of it in the short and the long run.

    If the empirical claim is true, someone who values only actual consent might think that this is the end of the discussion with whether to have a randomly selected legislature over an elected body. But I would say it is only one possible value for a citizen legislature.

    A second thing we might ask is what value a randomly selected legislature might have independent of what actual citizens think of it. And it is here that notions of hypothetical consent might come in to play.

    Keith and Peter, I think some of your discussion might be resolved in the following way:

    Keith you cite Fishkin’s consent by proxy. Fishkin clearly emphasizes that consent by proxy has value because the decisions made by such a body are what the whole public informed and deliberating together WOULD have decided. As Peter notes, there is clearly a sense in which this is hypothetical. Just because a randomly selected body has agreed on something, no one of the actual public has actually consented to anything.

    Consent by proxy is not hypothetical, however, in that Fishkin doesn’t make any assumptions about all citizens being rational or impartiality motivated (though I would presume he believes that most people are to some extent). That stands in contrast to Rawls and Scanlon and others whom we usually think of with regards to theories of hypothetical consent. So in this sense Fishkin’s consent by proxy is not hypothetical in the same way.

    It is important to note that even if a randomly-selected body carries out what the people WOULD have done, whether ACTUAL citizens would increase with respect to their tacit approval with respect to institutions (or more generally if their positive attitude towards our institutions would increase) is still anybody’s guess. These issues are tangential.

    My big question on all of this: Why should I care what the citizen body as a whole would have consented to? Fishkin operates on the implicit assumption that the be all and end it all of democracy is the aggregation of preferences. He just thinks those preferences should be informed, rather than knee-jerk reactions. He is, therefore, not all that different from Arrow and others in the social choice tradition (at least, that’s how I interpret him).

    But for someone like me who thinks that a main justification for political institutions is getting the right answer against some objective standard, why should I care what the public WOULD have decided on?

    I can see some value to actual tacit consent. I really don’t care much for consent by proxy, unless it can be shown that randomly selected bodies would get us closer to the truth. I think you can make such an argument, but most certainly not by including ALL of society’s voices. I don’t really care for contributions from psychopaths in deliberation.

    Fishkin’s claim that a randomly selected body would have made the same decision as the entire public is also seriously questionable. For one, any sensible body will exclude certain categories of people (convicted murderers, children, etc.). Second, the only body descriptively representative of the public is the entire public. No selection procedure is going to guarantee the decisions by the entire public and some portion of it will be the same. Fishkin might argue that randomly selected bodies get us closer to what the public would want than anything else. But as my comments in the previous paragraph indicate, do we really want that? Deliberation without psychopaths surely leads to better results.

    A final note on Locke. I think I’m in line with Peter’s comments here. If Pitkin’s interpretation is correct, there is a distinct way in which Locke conceives of rational men coming out of the state of nature. It is obvious he would believe the laws of nature that these rational men would consider when forming such an agreement are divinely inspired. But, Keith, even if you don’t think there are such divinely inspired laws, you might still think that men meeting certain conditions of rationality would come to particular types of agreements. So the fact that there might not be divinely inspired natural laws does not seem to imply that the hypothetical consent approach fails more generally.

    As Scanlon I think convincingly shows, for a hypothetical consent approach to justify particular institutions we need not believe that all of those affected share the same beliefs and convictions. They only need to be reasonable and impartially motivated. We might therefore ask whether reasonable people would agree to a randomly selected legislature over an elected one. The possibility that actual human beings would have more tacit consent of such institutions (or perhaps just have a more positive attitude towards them) is one possible relevant consideration of course. Though there are surely others.


  32. Thanks to Jan, as always, for the clarity that he brings to these issues. I think there is some ambiguity in the way we use concepts like “consent” and “hypothetical”, so let me try and make some distinctions.

    Consent 1: When Jan refers to the approval that citizens have of government institutions, this is, indeed, what social contract theorists refer to as government by consent. I would prefer, however, to call this “perceived legitimacy” or weak consent. In his last book Fishkin points to tentative indications that the DP may well lead to a form of decision making that is perceived by citizens as more legitimate than our existing electoral arrangements. In the Good Society symposium on Fishkin’s book Jane Mansbridge and Sandford Levinson outline two preconditions for perceived legitimacy: greater public understanding of proportionality and further evidence that randomly-selected forums are capable of generating epistemic outcomes that are (at least) no worse than elite or expert bodies. The perceived legitimacy of political institutions can, of course, be measured by a variety of tools available to political scientists.

    Consent 2: My concern is with strong consent — ie consent to individual policy proposals and this is a far more demanding test than that required by social contract theorists, or, for that matter, procedural theorists of democracy. Here I would like to echo Fishkin’s (strong) claim, namely that the DP outcome constitutes “informed consent”: what everyone would think if properly informed. Peter and Jan argue that such consent is “hypothetical”, so I think we need to distinguish two meanings of the latter concept:

    Hypothetical 1: “What everyone would think if properly informed” is ia testable social-science hypothesis. The hypothesis is that it would make little difference to the decision outcome which concrete individuals were selected by the allotment process and this is indeed testable by experiment. Select x number of forums, comprised of x randomly-selected individuals, furnish them with identical policy choices, information and advocacy, restrict their role to the secret vote and see if the hypothesis is confirmed. Consent-by-proxy is hypothetical in this sense.

    Hypothetical 2: A better word for Pitkin’s description of Lockean [and Rawlsian] consent as hypothetical would be “idealized” in that it depends on an abstract “rational” choice-maker — an ideal-type (homo rationalis — a distant cousin of homo economicus) rather than concrete individuals. This character is “hypothetical” in the sense that it has no existence other than in the mind of the theorist. Whereas Hypothetical-1 can be tested by a real-life social-science experiment, Hypothetical-2 of necessity relies on a philosophical thought experiment. Whether or not such approaches derive from a religious worldview isn’t important (although it is of interest in the case of Locke and Rawls).

    Jan also introduces the issue of epistemic outcomes. I agree with him that it’s important to get the “right” answer, but that’s not the subject of this thread — my concern here is with the intrinsic properties of various systems of decision-making as opposed to the quality of the decision outcome. I agree that restrictions may well need to be made regarding those eligible for selection by lot — it would be easy to make an argument for the exclusion of felons, minors and those with serious mental illness, although most would balk at my earlier suggestions for tests for mental competence and (even more controversially), the exclusion of all those who take the King’s Shilling. Although we do want to know what everyone would think if properly informed there is no reason why this should not be tempered with pragmatic constraints designed to encourage an acceptable epistemic outcome (by excluding, primarily, those who are unable to benefit from being “properly informed”). Note that these pragmatic constraints apply to any form of democratic decision making, so it would be unfair to subject Fishkin’s proposal to any more exacting standards than other procedures that seeks to bring power to the people.

    It’s also the case that the DP methodology (which is implicit in all my posts) makes no distinction as to whether participants are motivated by reasons or interests. Deliberative democrats (and the likes of Scanlon) would privilege the former, whereas their more traditional political science colleagues are satisfied with the latter. I’m not disposed to view reasons and interests as mutually exclusive, but that’s the subject for another post, and I would encourage Jan to submit it.


  33. Sorry, I forgot to add that In the case of Hypothetical 1, the case for consent-by-proxy is that it makes NO DIFFERENCE whether I attend in person or via my proxy. If I attend in person but am in the minority then I am deemed to consent (assuming the majoritarian principle), so the same would be true if my proxy attends. Most rational people would prefer to have their proxy attend, leaving them free to go to the pub, play cricket, or watch the paint dry on the back door, but this makes no difference at all to their consent to the legislative outcome. To reject this argument is to reject the majoritarian principle which is certainly implicit in Consent 1 (when applied to democratic decision-making). But if you accept the majoritarian principle then you would need to accept Hypothetical 1 (depending, of course, on the outcome of the experiment). If the experiment shows that the decision outcome is not identical (within an acceptable margin of error) then the case for representation-by-proxy fails and this is the main reason why I’m opposed to descriptively-representative assemblies doing anything other than voting — individual speech acts would be likely to cause the decision outcome to be less consistent between different samples. This usually raises Habermasian hackles, but it’s the price you have to pay for accurate representation-by-proxy.


  34. I am largely in agreement with a number of the points your raise here, but am somewhat skeptical on one.

    Consent 1: I can agree with everything Keith says here in it fits well with how political scientists generally understand the concept of legitimacy: “If people hold the belief that existing institutions are appropriate or morally proper, then those institutions are legitimate”. (Dogan, Mattei. 2004. Conceptions of legitimacy. Pp. 116–127 in Mary Hawkesworth and Maurice Kogan (ed.), Encyclopedia of Government and Politics, Volume 1. Routledge: London. The quote is from p. 116).

    With respect to the quality of outcomes: I suppose the contention here is then that having more perspectives increases the value of our institutions in some way, independent of the implication for the quality of decision-making. Keeping the outcome quality constant, more views expressed is better in some way. Maybe even making some sacrifice in the quality of outcomes can be justified if it means more perspectives are present. We might also say something like more types of people participating has some kind of value, independent of outcomes. That has some plausibility.

    With respect to interests and reasons: My empirical hunch like yours is that people are motivated both by their own interests, as well as the force of the better argument. And I suspect that both would make an appearance in any deliberative body. So I am in agreement with your position here as well.

    There is only remark within your comments of which I am skeptical: the idea that ” ‘What everyone would think if properly informed’ is a testable social-science hypothesis”.

    First, I must confess that I am skeptical whether the outcomes would be the same in many of these randomly-selected groups if they look at the same issue. Do we have any evidence on this from experiments with mini-publics? Particularly if there are multiple positions such a group could take on an issue, I am skeptical that we would always see a very strong convergence on issues.

    Second, and perhaps more importantly: even if a whole bunch of randomly-selected bodies came to the same conclusion independently on some issue and supposing that there was not too much dissent on the reached conclusion in these groups, what would this actually show?

    It would certainly indicate that a randomly-selected citizen if she were in such a body exposed to deliberation would LIKELY come to a similar conclusion as most of the citizens that actually participated (at least if we did not have any information about her that would indicate the contrary). With an even higher degree of confidence, it shows that if we had enough funds so that every citizen could get a turn in a randomly-selected body in the near future, that it is LIKELY that the majority of citizens would hold the opinion that we have already seen prevail in the previously conducted citizen bodies.

    And we might think that all this has some kind of value.

    But does this show anything about what the citizen body as whole would conclude if deliberating together? Or perhaps more accurately what the citizen body as a whole would be LIKELY to conclude if deliberating together?

    I’m not convinced that we can actually make Fishkin’s step here, because we cannot actually replicate the conditions that exist for quality deliberation in small groups among the citizen body as a whole. So it strikes me that this is not really in any way a testable hypothesis, though someone better steeped in the empirical social sciences might be in a better position to comment on this.

    I also wonder if this step about what the citizens deliberating as a whole would think by Fishkin is really relevant in any way. It strikes me that there is value in the fact that a randomly-selected citizen if she were in a randomly-selected body exposed to deliberation would LIKELY come to a similar conclusion as most of the citizens that actually participated. Is there any need for making claims about all citizens ‘deliberating’ together would be likely to conclude?


  35. JB: >First, I must confess that I am skeptical whether the outcomes would be the same in many of these randomly-selected groups if they look at the same issue. Do we have any evidence on this from experiments with mini-publics? Particularly if there are multiple positions such a group could take on an issue, I am skeptical that we would always see a very strong convergence on issues.

    The scanty evidence available goes against my hypothesis. There is some research by one of Fishkin’s colleagues (reported in one of Bob Goodin’s paper — probably the one on deliberation within) that shows a lack of consistency for a travelling DP roadshow. I think it was multiple DPs on energy options in Texas. My hunch is that the variation is on account of the very limited speech acts that Fishkin does still permit — ie the small-group deliberations and the resulting plenary questions. Fishkin is adamant that active small-group deliberation is an essential component of opinion change so he did not respond favourably to my suggestion that he sacrifice this aspect in order to ensure greater consistency between fora. But Fishkin’s DPs are only advisory; if one is seeking to create a legislative assembly that is truly representative then it seems to me that these sacrifices have to be made — even if the resulting deliberation is a very meagre affair in Habermasian terms. The DP small-group deliberations rely heavily on trained moderators to level out inequalities in speech acts but this would not be possible in a body with statutory powers on account of Quis Custodiet concerns.

    >I’m not convinced that we can actually make Fishkin’s step here, because we cannot actually replicate the conditions that exist for quality deliberation in small groups among the citizen body as a whole. So it strikes me that this is not really in any way a testable hypothesis, though someone better steeped in the empirical social sciences might be in a better position to comment on this.

    I share your scepticism, that’s why my experiment would proscribe quality small-group deliberation. Terry Bouricius accepts this point and has argued convincingly that such deliberation should happen earlier on, as part of the agenda-setting process — the modern equivalent of the Athenian boule. We agree that this should not be conflated with final legislative decision making, where the role of the jury should be limited to listening to the arguments of the advocates and then voting up/down the outcome. The corresponding Athenian institution would be the nomothetai. Where Terry and I differ is on the need for a modern analogue of the ecclesia, my argument being that all citizens need to choose between the legislative proposals on offer. Otherwise the agenda is entirely down to the whims of a random oligarchy.

    If there were real concerns for consistency then the allotted sample would simply need to be larger — the only ceiling being the rational ignorance threshold (if allotted members feel that their vote has little effect on the outcome they may well snooze off during the deliberations). It’s interesting that the Athenians specified larger juries for important cases — I’m not sure whether this was in order to prevent corruption or on account of concerns over accurate representation. The fact that juries sometimes ran into 1,000s might suggest the latter (even if the Greeks had no mathematical concept for proportionality).


  36. I have two unrelated, and relatively minor observations about some things JBurgers psoted above:

    1. JBurgers wrote:
    “Fishkin might argue that randomly selected bodies get us closer to what the public would want than anything else. But as my comments in the previous paragraph indicate, do we really want that? Deliberation without psychopaths surely leads to better results.”

    Firstly, how is it to be decided who is not suited for deliberation? Once it was agreed (by men, at least) that women were not suited. In the U.S. some states (like my own) allow current prisoners to vote, while others ban both current prisoners and even those who have completed their sentences.

    My argument is that we should allow all people to participate in the lottery. The “bad apples,” however defined, will not dominate. If it is just a vote, as Keith advocates, then we can trust that the psychopaths will be out-voted. If it is a free-wheeling assembly as Yoram advocates, then we leave it to the assessment of the members selected to decide which other members they should ignore. But the fact is, having the perspective of a current mental patient (about the healthcare system, for example) may inform the body and prove crucial to making an epistemically “right” decision. I would even say that children should be allowed to participate if they wish, to further enrich the available information…but I expect many would say I am going too far, though some jurisdictions have voting ages as low as 16.

    2. On the matter of whether deliberation will reflect interests or reasons, I want to draw your attention to a point made by Bryan Caplan in his book, “The Myth of the Rational Voter: Why Democracies Choose Bad Policies.” He argues that voters often vote ideologically in exact opposition to their own objective self interests. In some cases these voters may simply be mistaken about their own self-interest, but Caplan offers a different explanation. He suggests that voters intuitively understand that their single vote won’t turn the election, and that they get a greater psychological payoff (feeling like a good citizen) from voting as they think is best for society (as filtered through their ideology or world view), rather than voting selfishly, for their personal interests. He contrasts this with how the same individual might “vote” in making a major personal purchase, where their one vote is DECISIVE. In this case the individual will give more weight to his own self-interest, and rationalize away concerns about “the common good.”

    So while Mansbridge and others (including me) might think a randomly selected deliberative body will give greater attention to the “common good” than to mere “interests,” than an elected legislature does, there are countervailing forces to consider. By making the randomly selected members more decisive we prompt them to pay attention and learn, but perhaps also to vote more selfishly.


  37. The final part of my previous post raises another problem …It may be meaningless to imagine comparing how “all people WOULD vote if All could deliberate” with how a sub-sample votes, because the individuals might CHANGE their vote based on how many other people they believed were participating.


  38. Terry’s argument about the bad apples being factored out is persuasive when it comes to an assembly with voting powers, less so with an assembly that mandates individual speech acts. Adolf Hitler was a persuasive orator. As for his second post, it might make sense to focus on the interchangeability of deliberators rather than the hypothetical case of all deliberating, for the reasons Terry provides. The key issue regarding individual consent is does it makes any difference to the outcome whether I attend myself or leave it to my proxy? If not then I don’t see the difference between consenting in person and consenting by proxy.


  39. Keith: “But Fishkin’s DPs are only advisory; if one is seeking to create a legislative assembly that is truly representative then it seems to me that these sacrifices have to be made — even if the resulting deliberation is a very meagre affair in Habermasian terms.”
    There is another way that decisions could be ‘representative’, accepting the majoritarian principle: namely by forcing the entire citizenry to vote on every proposal coming out of the government or bureaucracy. That might face some practical obstacles, but that is not the only reason we would be skeptical of such a suggestion: the conditions under which those votes were cast are far from ideal, as people are generally uninformed. I would be similarly skeptical of seeing any value to representation if now we confine mini-publics only to voting. Sure they might reach similar results if restricted in this way, but so what? The most valuable thing about a mini-public is that people become informed and actually change their minds. And much of that occurs I suspect via actual deliberation and exchange of viewpoints, not merely watching a few speeches by experts and then voting on what one thinks is best after “deliberation within”.

    Or to express it in another way, we could get all these mini-publics to agree to the same thing if we held a gun to their heads. But so what? What matters are the conditions under which these mini-publics would come to particular conclusions.

    Incidentally, following your train of thought Keith, why is it that we need all randomly-selected bodies to converge on their thinking? What if a simple majority of randomly-selected bodies converged on their thinking? Is there some way to say something about representation and consent in this way?


  40. Terry:

    You might be making two different arguments here: 1) that more diversity is always a good thing for getting the right answer; 2) that, even if diversity is not always a good thing for getting the right answer, we can’t select whose perspectives would be good and as a rule of thumb more diversity is better.

    Both of these positions defy common experience. With regards to how to build a bridge, my perspective is pretty much worthless because I don’t have any clue as to how to build bridges. Nor is that something I might helpful with, even with some more information acquired within a reasonable time and listening to the perspectives of experts. We might think that representation is somewhat of a different kind of task, and I would agree. But why do we automatically have to think that everyone’s perspective is all of a sudden relevant to getting the right answer? Or, even if diversity is not always strictly beneficial to the right answer but can also have costs, that more diversity is generally better than less?

    Literature on organizations I think well shows the folly in this kind of position (See, e.g., Katherine Williams and Charles O’Reilly, “Demography and diversity in organizations: a review of 40 years of research”, Research in Organizational Behavior, 20 (1998), 77–140). Though diversity can improve performance of groups, it can also be detrimental to that performance if we allow people in who are incompetent or people with particular kinds of motivations (e.g., racists). There is little empirical evidence to support the idea that including everyone will somehow always be better for decision-making.

    Regarding the question then of who to include. It strikes me that there a few obvious categories we can eliminate: people who can’t read; people who can’t speak the language of the assembly; people who believe that the muslim brotherhood is secretly taking over their government and as a consequence commit terrorist acts; people who have cheated their clients out of millions of their life savings.

    Just because any such selection system is going to be very imperfect, doesn’t mean we cannot eliminate some obvious candidates.

    Incidentally, it strikes me as suggestions to include everyone in a citizen assembly that precisely makes such proposals so implausible. Just because there might be value to having more diversity versus elected assemblies, does not mean that the value of diversity is not subject to constraints.


  41. Sorry I wasn’t logged in, but both of those comments were mine.


  42. Jan >The most valuable thing about a mini-public is that people become informed and actually change their minds. And much of that occurs I suspect via actual deliberation and exchange of viewpoints, not merely watching a few speeches by experts and then voting on what one thinks is best after “deliberation within”.

    Although I’m not sure that changing your mind is of any inherent merit, informed decision-making is clearly the goal. There must be some quantitive evidence on the information gain in non-participatory fora — for example there have been quite a few lecture-theatre and TV debates where the audience is polled before and after the discussion. Recent examples in the UK are the Intelligence Squared series, so I think your suspicion regarding the importance of exchange of viewpoints may be unfounded. If it is over a significant issue and (more importantly) if individual votes have some serious causal power (rather than just an opinion poll) why is it so important to generate hot air? Harrington and Rousseau believed this to be entirely counter-productive, so why are we so insistent now that deliberation has to involve verbal exchanges? (other than just deferring to Habermasian norms on communicative action).

    As for the requirement for convergence, bear in mind that I’m proposing a political settlement in which all decisions are taken by a small randomly-selected jury, period. If you don’t want to give up on the notion of democratic consent then this presupposes that the jury will come to the same decision (within a reasonable margin of error) irrespective of which concrete individuals are involved. So long as all members are properly informed and motivated to stay awake what does it matter what the conditions are? Our concern should be with the decision outcome and that this should be valid from both an intrinsic and instrumental perspective (your chosen terms — my preference being normative and epistemic). Arguably the lack of direct participation may not lead to the indirect benefits from civic education that Mill saw as a consequence of political participation. But this is a small price to pay for democratic consent and my proposal for audience democracy would certainly be a lot more wholesome than Manin’s.


  43. Jan Burgers,

    I disagree with your assessment of the state of research on the value of diversity in group problem solving and decision making. I will look up the paper you mentioned. I would offer:

    Helene Landemore’s paper on Cognitive Diversity, “Deliberation, cognitive diversity, and democratic
    inclusiveness: an epistemic argument for the random
    selection of representatives,”

    Scott Page’s Diversity Trumps Ability Theorem described in his book “The Difference. How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies,”

    Asher Koriat’s, “When Are Two Heads Better Than One and Why.” Science. Vol. 336 no. 6079.

    And most famously, James Surowiecki’s book “The Wisdom of Crowds. Why the Many are Smarter than the Few and How Collective Wisdom Shapes Business, Economies, Societies, and Nations.”

    In your bridge analogy…a legislative chamber can have all the engineers as consultants they need…but the underlying question is “shall we build a bridge here, at this cost?” and the bridge engineers can’t answer that for society.


  44. JBurgers,

    > But why do we automatically have to think that everyone’s perspective is all of a sudden relevant to getting the right answer?

    The “right” answer is defined, in a democracy, as the sum of all perspectives. Excluding some perspectives from the sum, by definition, results in a wrong answer.

    > It strikes me that there a few obvious categories we can eliminate: people who can’t read; people who can’t speak the language of the assembly;

    And so, who will take care of the interests of the illiterate and immigrants? Those same people who set up a system where these people are marginalized?

    > people who believe that the muslim brotherhood is secretly taking over their government and as a consequence commit terrorist acts;

    And what about people who believe Wall Street is taking over their government? Should they be excluded as well?

    > people who have cheated their clients out of millions of their life savings.

    Such people are such a tiny minority that their exclusion will change nothing. But exclusions of such people will inevitably lead to demands for excluding prisoners and felons, resulting in the disenfranchisement of large segments of the population.

    For a general discussion of exclusions – pre-selection, at selection and post-selection – see a recent post of mine.


  45. Terry and Yoram nicely indicate the very different epistemic and intrinsic arguments for sortition, the trick being to design sortive institutions in such a way as to harness both benefits. This is a non-trivial problem as the “sum of all perspectives” is based on an arithmetic function that presupposes the common identity of the units being summed. Whilst it’s possible to sum votes, I’m not sure what kind of mathematics would be required to sum ideas. Perhaps this is why many people working in this field tend to privilege one aspect over the other. When you try to reconcile both perspectives you end up either with byzantine complexity (Terry) or a meagre form of deliberation (me).


  46. I think that this discussion is useful. Keith, the distinctions you are trying to draw between different forms of consent, and further between different forms of hypothetical consent, seem important. Perhaps another way of putting it is that we can distinguish between the hypothetical consent of actual persons and the hypothetical consent of hypothetical persons. Rawls, Harsanyi, and Scanlon clearly have the latter in mind, whereas Fishkin and you (apparently) have the former in mind. But it’s important to remember that there is a spectrum between the two, and the line between them is not as clear as you make it sound. After all, Fishkin wants an AC to reach a decision that is the same, not as what the entire population actually would say if asked, but actually would say if asked AND given proper information, time to think about it carefully, etc. And that’s making the situation even more hypothetical. And if you introduce deliberation, whereby AC members talk to each other, the relationship between the AC and the general population becomes even more hypothetical.

    But there’s also a huge fallacy of division being committed here. Maybe you could convince me that an AC will reliably reach, via majority rule, the same decision that the entire nation would have reached via majority rule if asked. But that is NOT the same as saying that there is some member of the AC who will reliably vote the same way that I would have voted if asked. Keith, your examples run these two possibilities together. You could, I suppose, say that I have hypothetically (in one sense) consented to a decision if 1) somebody else makes the decision and 2) that somebody reliably makes the same decision I would have made had I been there. It’s arguable, but let’s let that slide. Maybe you could even say that I have hypothetically consented to a decision if 1) an AC makes the decision via majority rule and 2) that AC reliably makes the same decisions via majority rule that I would have made had it been up to me alone.

    But your examples are not like that. They involve one legislator who is sufficiently like me that s/he reliably votes the same way I would have voted. There are two things wrong with it. One, it’s incredibly implausible that there will be any such person in an AC for ever single American citizen. Do you really think that there would be somebody in even a thousand-member AC who would reliably make the same decision as Peter Stone would? Two, even if it was true that some such individual existed in the AC for every single American (which would require, for a thousand-member AC, that there are 200,000 Americans who would always decide the same way), the group is deciding by majority rule. What about the minority that got outvoted? If the vote of their members is supposed to reflect their consent, then did they not consent at all?

    I suppose somebody’s going to bring up consent to the decision-making process itself here. It does seem to keep coming up. But it still makes absolutely no sense to me. For one thing, if consent to the process is what matters, then all this stuff about consent-by-proxy is just beside the point. For another thing, it remains utterly ludicrous to imagine that I or most anyone else “consent” to our decision-making processes in any meaningful sense, either now or in the future. I repeat–when the hell did I consent to the decision-making processes of the U.S. government? They are still normally regarded as binding me whether I vote or not, or whether I obey the law or not, or whether I accept any specific government benefit or not. Hell, I no longer live in the United States, and I’m STILL subject to U.S. law. (As a U.S. citizen, if my foreign income exceeds a certain amount I am liable to U.S. income tax on it.) I would literally have both to leave the U.S. AND renounce my citizenship to escape the jurisdiction of U.S. laws. This would be true even if I never voted once in my entire life. The idea that I have tacitly consented to the government of the U.S. remains ridiculous beyond words to me. And I cannot imagine that even the best sortition-based system we can devise will be any different in this regard.

    There’s more I could say, but I’ll leave it at that for now. I may try to come back to this in a future post, as it still seems rather plain to me that the entire discussion would proceed better if the word “consent” simply disappeared from it. It just gets in the way.


  47. Peter,

    I apologise for using the word “proxy” in the singular — I confess it never occurred to me that anyone would think I was suggesting we would all have a one-to-one doppleganger in an allotted assembly. I’ve consistently argued that descriptively-representative assemblies have to be theorised at the aggregate (statistical) level, as opposed to the individuals that they are comprised of. In this sense “my proxy” is an idealised construct for the degree to which my own individual characteristics (understanding, prejudices and interests) are shared in part by members of the allotted assembly. If more people share more of my characteristics in the whole political community, then the microcosm will reflect this and my will (as modified by the effects of deliberation) will prevail. No rational person could fail to consent to such an outcome. If my characteristics are in the minority then I still consent in so far as I am bound by the majoritarian principle. We do all need to stop thinking about statistical representation in terms of the (singular) “allotted representative” (as with an MP or Congressman), so I’m sorry if my loose talk of an (individual) “proxy” has put this process back.

    I agree with you regarding the absurdity of the social-contract argument for consent to the decision-making process. Most citizens have no clear idea how their political systems function — Bagehot’s distinction between the “dignified” and “efficient” constitution was based on the observation that most Victorians believed they were ruled by the Queen (the dignified explanation), whereas efficient power resided in secret cabinet committee — an institution that wasn’t even formally recognised at the time and failed to make an appearance in Mill’s competing essay. How can we consent to smoke and mirrors? Given the absurdity of social contract explanations I can understand your frustration with the use of the word “consent” in the political context. My (or rather Fishkin’s) argument is that consent-by-proxy is unique in that its concern is with outcomes rather than constitutional niceties. If you want to call this the hypothetical consent of actual persons then I’m very happy to live with that. As my “interchangeability hypothesis” is open to empirical confirmation hopefully the status of consent-by-proxy will at some point in the future cease to be hypothetical.

    Given that most of us would prefer to live according to our own reflective judgment (as opposed to snap decisions) I think we would happily live with the gap between the deliberative judgment of our proxies and the rational ignorance of the majority of us who do not draw the golden ticket (the argument being that our own views would have been modified in a similar way had we been chosen to deliberate). The relevant analogy is our own consent to a sex act resulting from a drunken fumble (and regretted the following morning) as opposed to a measured decision to start a new relationship. Given the choice I think most people would prefer to consent to the considered judgment that they would have made rather than their uninformed off-the-cuff whim on a subject that they knew very little about. I know all the “woulds” make it all a bit tacit/implicit/hypothetical, that’s why I’m hoping someone will come up with the money to confirm or refute the interchangeability hypothesis.


    PS: to avoid any confusion, I’m using the word “deliberation” in the Rousseau-Fishkin, not the Habermasian, sense. Active (Habermasian) deliberation would refute the interchangeability hypothesis (that any number of allotted assemblies would arrive at the same outcome [given a small margin of error], if subject to identical information and advocacy).


  48. >The relevant analogy is our own consent to a sex act resulting from a drunken fumble (and regretted the following morning) as opposed to a measured decision to start a new relationship.

    Anticipating the response that you can’t have sex by proxy, have you never heard of Second Life? Furthermore, to extend the analogy, everyone have multiple Second Life avatars (alts).


  49. >Fishkin is adamant that active small-group deliberation is an essential component of opinion change

    There is some evidence that Fishkin might be mistaken. In his 2008 book Bob Goodin reviews the Bloomfield Track (Cairns, Oz) citizen jury and makes the following observations:

    “No one thought the “group discussion” was the most important factor in changing their opinion; three-quarters of them thought it was the least important factor (p.51) . . . The simple process of jurors seeing the site for themselves, focusing their minds on the issues, and listening to what experts had to say did all the work in changing jurors’ attitudes. Talking among themselves, as a jury, did virtually none of it.’ (p.58).

    In other words it is the information phase, not the group deliberation, that matters most. This fits well with Goodin’s notion of “deliberation within” and causes Andrew Dobson to argue for a new emphasis on listening rather than talking as the most important aspect of communicative rationality:

    “It might appear ironic that discursive democracy could be improved by having less discussion, but it only appears so because we have got so used to thinking of discursive democracy in terms of speaking.” (2012, p.857)


    Goodin, R. (2008), Innovating Democracy: Democratic Theory and Practice After the Deliberative Turn (Oxford: OUP)

    Dobson, A. (2012),’ Listening: The New Democratic Deficit’, Political Studies, 60, pp.843-859,


  50. Goodin is discussing only one kind of deliberative assembly, a jury judging the credibility of evidence. Most assemblies of interest to us are legislative in character, introducing proposals and debating them. There might be evidentiary hearings but it makes a difference when the subjects are provided by the members of the body.


  51. Yes, that’s true, but the argument of this thread is that the only way that the deliberative judgment of an allotted assembly can be said to embody the informed consent of the whole political community is if the mandate excludes the introduction of legislative proposals. The mandate of a regular (elected) legislative assembly includes speech acts and members can (at the very least) claim the consent of those who voted for them; not so in the case of an allotted assembly. Deliberative polls work to an externally-provided agenda, but Fishkin has argued that small-group deliberation is an important element in opinion change. However Goodin’s evidence would appear to contradict this. This is important if the deliberative judgment of the assembly is to embody the informed consent of everyone, as the “interchangeability hypothesis” (see above) presupposes listening and deliberation within, as opposed to the speech acts that normally characterise legislative assemblies.


Leave a Reply

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

WordPress.com Logo

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

Facebook photo

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

Connecting to %s

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

%d bloggers like this: