Estlund Reading Group Chapter 1

Welcome to the first instalment of our virtual reading group on David Estlund’s Democratic Authority: A Philosophical Framework. Today’s post will focus on chapter 1, which offers a synopsis of the book’s central arguments and conclusions.


Estlund begins by noting the apparent tension between democracy and political quality. The masses seem ill-suited to making the best political decisions, and thus many democratic theorists see their primary task to be one of explaining why democracy is valuable or desirable despite the risk it poses to making good decisions. Estlund takes a different approach. He claims that it is in fact democracy’s tendency to produce good decisions that explains democracy’s legitimacy and authority. Legitimacy, according to Estlund, refers to the moral permissibility of the state issuing and enforcing commands due to the process by which they were produced, whereas authority refers to the power of one agent to morally require or forbid actions by others through commands (p. 2). Estlund’s central thesis is that ‘democratically produced laws are legitimate and authoritative because they are produced by a procedure with a tendency to make correct decisions’ (p. 8).

Perhaps the most obvious objection to this thesis is that it doesn’t conform with how we make other kinds of decisions. Democratically taken decisions about medical matters are not likely to yield correct results – we go to the experts in medical matters. Why should politics be any different? But, Estlund notes, there are important disanalogies between the medical and political examples: doctors are recognized medical experts, and this is why we usually consent to their decisions. Doctors’ authority to make medical decisions for us derives from our consent, and cannot be derived from the mere fact of their expertise. Moving directly from expertise to authority is what Estlund calls the expert/boss fallacy: just because you know more doesn’t make you boss.

Furthermore, the problem we face in the political arena is that there is no one individual or group to whom we are willing to cede authority on the grounds that they know best. There are no recognized political experts. But Estlund also warns against going too far in the other direction. Just because we recognize no political experts among us does not mean that our political decisions must be acceptable to everyone – we cannot make our political decisions hostage to the clearly irrational or immoral. He concludes instead that political decisions need to be justified in a way that is acceptable to suitably qualified individuals, with ‘suitably qualified’ requiring further specification.

There are two opposing worries about this Rawlsian idea of requiring political decisions to be acceptable to some idealized constituency. The first is that the constituency is not inclusive enough, but as Estlund insists, how can it be wrong to exclude Nazis or other clearly unjust people from the constituency of acceptability? The opposite worry is that it is too inclusive, especially if some of the qualified points of view turn out to be false. Why not simply insist that political decisions must be true? Estlund suggests that the idea that political decisions should be acceptable to qualified persons or points of view need not conflict with the aim to embrace only the truth, for it may be true, at least when it comes to political decisions, that we have sound moral reasons to follow the ‘acceptable to qualified points of view’ condition.

The next objection Estlund confronts is that democracy’s value inheres not in its tendency to produce correct decisions, but rather in its procedural fairness. Democracy gives everyone an equal chance to influence the outcome, and this is why democratic decisions have legitimacy and authority. Estlund’s response is to point out that such purely fairness-based arguments have no plausible way of explaining why we should prefer a system of deliberation and voting over a randomized decision procedure such as a coin toss. If you believe it would be wrong to toss a coin, Estlund suggests, it must be because there are epistemic reasons that count against doing so.

Though some might think epistemic concerns justify an epistocracy, rule of the experts, the problem we face is lack of agreement regarding who the experts, if any, are. So how can we justify democratic procedures on epistemic grounds, especially when we know they will surely get the wrong result some of the time? Here Estlund makes extensive use of an analogy with the jury system. The jury system is an exemplar of what Rawls calls imperfect procedural justice. We have an independent standard of correctness – convicting all and only the guilty defendants – and the jury system is valuable insofar as it tends to regularly deliver correct results. If it regularly convicted the innocent and set the guilty free we would not grant that system the authority and legitimacy that we do. Estlund argues democracy’s authority and legitimacy has a similar structure. Democratic procedures, he claims, are more likely to deliver correct results than random decisions, and they represent the epistemically best decision-procedure among the acceptable alternatives. This defence of democracy is called epistemic proceduralism.

But what about consent? Traditional theories hold that there can be no authority without actual consent. Given that few people actually consent to democratic states, how can they then have authority? Estlund points out that even consent theories will specify the conditions under which consent becomes void (e.g. duress). But then shouldn’t there also be conditions under which non-consent becomes void, that is, conditions under which it would be illegitimate or impermissible to refuse to consent to someone else’s authority? Estlund suggests that if it would be morally wrong to refuse to consent to someone’s authority then any non-consent is void. In such cases there is what Estlund calls normative consent.

The relevant question is then whether it would be morally wrong to refuse to consent to democratic authority? Estlund claims the answer to this question is yes, largely because of democracy’s tendency to produce substantially just laws and policies, as well as the fact that there is no non-democratic system of political decision-making that could be endorsed by all qualified points of view as doing a better job of making just decisions.

Estlund recognizes that if democracy is going to deliver the epistemic goods we will need a conception of democracy that makes fairly significant demands of citizens – citizens will need to be engaged, informed, and not purely self-interested. One worry is thus that Estlund’s theory is hopelessly utopian. He responds to this worry with a counter-charge, namely, that political philosophers should be wary of what he calls utopophobia: ‘the fear of normative standards for politics that are unlikely to be met’ (p. 14). Estlund claims that just because people won’t act in the required way doesn’t show that they couldn’t do so, and it’s only when the latter is true that our normative theory has become problematically utopian.

Estlund goes on to distance his theory from two other views about democracy’s authority. The first is Condorcet’s Jury Theorem. Estlund denies this theorem provides much evidence of real democracies’ epistemic quality for two main reasons. First, the jury theorem works best for binary choices whereas real democracies rarely face only two choices. Second, the jury theorem famously requires people to have a better than random chance of getting the correct answer whereas Estlund doubts this will regularly be true of real people who are known to have many cognitive biases.

The second view Estlund wants to reject is the idea that democratic procedures get their authority because they are somehow analogous to the models of contractualism popular in contemporary moral and political philosophy. There are two critical disanalogies which undermine this view: (a) real democracies don’t, and shouldn’t, offer individuals a veto, but an individual veto is an essential part of most contractual theories, and (b) parties in contractual theories don’t address the question of justice or rightness directly – they are usually motivated by some more partial concern, whereas in real democracies we will need citizens to directly address the question of what justice requires.

The chapter concludes by noting that the epistemic proceduralist approach doesn’t need democracy to have a very high probability of delivering correct decisions – democratic procedures just need to do better than random and be the best acceptable mechanism in epistemic terms. Estlund also reminds us that his theory may have very different institutional implications than theories which claim there is some intrinsic value in modelling our real world democratic institutions on ideal procedures. Because Estlund sees the value of democratic procedures as instrumental rather than intrinsic, if non-ideal conditions require very different institutions to deliver correct decisions, then we may be justified in radically departing from those procedures that would track (or constitute) the truth under ideal conditions.


There’s an awful lot one could say about this first chapter since it provides an outline of all the book’s main arguments, but I’ll restrict myself to just a couple of simple points. I know that all these issues receive more detailed treatment in later chapters, and I’m wary of pre-empting the discussion of those later chapters (plus I haven’t read them all yet!). I’m hoping, however, that there’s no real harm in flagging some issues now, which we can always return to in later weeks.

One obvious worry for any epistemic theory is the required premise that there are always correct and incorrect, or at least better and worse, political decisions. Don’t we sometimes make democratic decisions where there is no preference-independent standard of correctness? Suppose our city has an annual cultural/sports budget, and the question arises as to how that budget should be spent this year. We can either spend the money on an art exhibition or we can spend it on upgrading the local football facilities. If we assume that everyone already has their just share of resources, does it make sense to talk about an objectively correct decision in this case? Can’t citizens simply vote their preference? But if there is no epistemic value, where does the authority or legitimacy of the decision come from?

Of course Estlund could respond by saying that, in examples like the one above, the correct decision is the one that tracks the majority’s preferences, thus democracy will be an ideal way of tracking correctness. But this invites a further worry, namely, that we need to have some sense of what constitutes ‘correct’ (or better and worse) decisions. We can be confident about the jury system because we have our independent standard in hand: convicting all and only the guilty. But, as Waldron and others have pointed out, isn’t one of the problems we face in politics the fact that we cannot agree on what would constitute a correct decision? Utilitarians will favour one view, right-libertarians another, Rawlsians another, and so on. Without some agreement on what constitutes correctness, how can all qualified points of view accept that democracy is the most reliable epistemic procedure? This looks especially difficult for Estlund to address given his rejection of the democracy/contractualism analogy, part of whose power is its alleged ability to design procedures that will yield just results even when we don’t know, or can’t agree, on what justice requires.

A third concern is whether epistemic proceduralism suffers from the same kind of rule-worship objection as standard forms of rule-consequentialism. Estlund readily admits that democracies will not always produce correct decisions, they only tend to do so more than the alternatives. But when we are very confident that a democratic decision has produced an incorrect result, why should we accord that decision any legitimacy or authority? If what we really care about is correct decisions, surely we have no epistemic reason to treat an incorrect decision as binding just because it’s the result of a generally sound decision-making procedure?

Fourth, recall Estlund’s claim that fairness-based theories of democratic authority and legitimacy cannot explain why we should prefer deliberation and voting rather than a randomized decision procedure between policy options like a coin toss. Estlund notes that such randomized procedures also give everyone an equal chance to influence the outcome since they give everyone no chance (p. 6). But what if we supposed that what matters is giving each person the greatest chance possible, consistent with a similarly large chance for others: a greatest equal chances principle. Here a coin toss between policy options or political parties wouldn’t do the trick since this applies the greatest equal chances principle to policy options or parties rather than to persons. The only randomized procedure (i.e. not voting) that could satisfy the greatest equal chances principle with regard to persons would be a lottery system which selected a citizen or group of citizens to make each political decision. This procedure seems, at least to me, to have a greater claim to authority and legitimacy than the coin toss between proposals.

Fifth, Estlund says that in order for the charge of utopianism to be troubling the critic must show that the normative standard under discussion is not only something that will never happen, they must show ‘it is not something people could do (or at least not without more effort or sacrifice than it’s appropriate to require)’ (p. 14). Though Estlund intends this passage about utopophobia to serve as a reply to those who think his theory might be too utopian, I was left wondering whether a theory can ever be too utopian. For starters, if a theory asks people to do more than it is appropriate or reasonable to require of someone, then this theory is not utopian, it’s just an incorrect moral theory. If it’s too much to ask me to sacrifice my life to save someone from paralysis, then this requirement is not utopian, it’s just mistaken. This leaves normative requirements that are physically impossible – demands that contradict laws of nature or human nature. But even here we might think there’s nothing troubling about a normative ideal that cannot be realized. Suppose we believe that justice requires there to be no inequalities between persons caused by brute luck. It may be impossible to achieve this ideal given the way natural talents and dispositions are distributed amongst persons. But this impossible ideal nevertheless gives us something to aim at, and especially when we cannot be sure what is impossible and what merely looks impossible, it may be very important to begin with normative ideals that are not constrained by any ‘facts’ about the human and natural world. If this is right, then the definition of utopophobia could be more inclusive than Estlund’s – it could be the unfounded fear of normative standards that are unlikely to be met or even impossible to meet.

Finally, I’m not sure how Estlund’s epistemic approach coheres with his idea that political institutions need to be justified in a manner acceptable from all qualified points of view. Suppose that there are in fact ‘political experts’ among us, the problem is simply that all qualified persons cannot agree on who these experts are. But let’s also suppose the experts know they are expert, and also possess the means to seize power and implement their decisions. Estlund’s theory tells us this epistocracy would lack legitimacy because it cannot be justified to all qualified persons. The ‘all qualified points of view’ test thus represents a constraint on the pursuit of epistemic value, and it’s natural to wonder what justifies this constraint given the great weight epistemic considerations carry in Estlund’s theory. Is it respect for persons that motivates this constraint? And if it is something like respect for persons, why doesn’t this play a more direct role in accounting for democracy’s authority and legitimacy, that is, why aren’t democratic procedures generally binding because the process of deliberation and voting manifests a respect for persons as agents or practical reasoners, each of whom has their own views about justice and the good?

Be Sociable, Share!

About Jonathan Quong

Jonathan Quong is a lecturer in political philosophy at the University of Manchester
This entry was posted in Posts, Reading Group. Bookmark the permalink.

20 Responses to Estlund Reading Group Chapter 1

  1. Ben Saunders says:

    Wow, something of a tour de force indeed. As you say though, many of these issues will recur in later chapters, so I want to restrict myself to a more general worry.
    Personally, I’d assume that whether a system is democratic or not is essentially a descriptive question and then, within a wide range of democratic theories, you could meaningfully debate which were better or worse (perhaps democratically or by some other standard). Now Estlund doesn’t seem to distinguish between what makes something a democracy and what makes it a good democracy. Is his theory supposed to justify all democracy or simply its best form?
    I’ll have more to say on the coin-flipping when I write about chapter 4 (I hope), but to illustrate what I mean let’s assume that democracy is simply about procedural fairness. In this case, coin-flipping would count as democratic. Now, majority rule could also count as democratic, if procedurally fair, and may have epistemic advantages. So does this mean Estlund isn’t defending all democracy, but only saying that democracy is justified, when it is justified, by epistemic values? And that when it has no epistemic value, it isn’t justified?
    Finally, a quick comment on something at the end of your summary Jonathan. You say "Estlund sees the value of democratic procedures as instrumental rather than intrinsic". Maybe I’d have to re-read the chapter, or the book, but is ‘rather than’ strictly correct? Estlund certainly stresses instrumental advantages, because these differentiate him from many others, such as the procedural fairness people. However, isn’t it true that he at least could (whether or not he does) say that democracy is both intrinsically and instrumentally valuable?

  2. Hi Ben, thanks for your comment. On your first point, I would say, though I haven’t finished reading the book, that Estlund is using the term ‘democracy’ in its common descriptive sense, that is, a system of free and fair elections etc…I take it that his theory of epistemic proceduralism then does at least two things. First, it shows us how democratic procedures, in this descriptive sense, also have normative characteristics (authority and legitimacy) that non-democratic systems lack. But second, epistemic proceduralism can also serve as a standard of evaluation between democratic systems (e.g. some democracies will tend to produce better decisions because the political system is less distorted by inequalities of wealth than other systems). This means his theory provides some guidance (though Estlund is hesitant about institutional implications) as to how we should structure our democratic procedures and political behaviour in order to increase the tendency to produce good decisions.
    On your second point – you’re right I was a bit sloppy with my formulation. Estlund does not explicitly rule out (at least in this chapter) the idea that democracy may have some intrinsic value related to the fairness of its procedures. What I should have said is that in terms of authority and legitimacy (which depend on the tendency to produce good decisions), the procedures are instrumentally valuable, though the procedures may be intrinsically valuable in other ways not related to authority and legitimacy.

  3. Jonathan’s summary of chapter one is, as Ben says, a tour de force.  Since, as both of you point out, the chapter is intended as a roadmap of what is to come, it might be premature to press hard for details at this stage.That said, let me forward a comment from one of my students, Joe Adams:

    Estlund sets up normative consent as the standard which epistemic proceduralism will have to meet in order to justify democratic authority.  So Estlund is granting that epistemic proceduralism can never establish actual consent.  As he says “we’re not assuming that consent would be required to legitimate rule” (3-4).  Since the standard has been changed, the question whether receipt of benefits can establish normative consent needs to be asked.  If epistemic proceduralism cannot establish true consent but can establish normative consent; and if receipt of benefits cannot establish true consent, perhaps receipt of benefits can establish normative consent.  Perhaps we could establish authority by appealing to normative consent and receipt of benefits without bringing the complex machinery of epistemic proceduralism to bear.  Indeed, it seems that anything that has been seriously entertained as the basis for actual consent, should be re-entertained as a possible basis for normative consent.  

    I think Joe’s point is a good one.  Chapter one alludes to consent accounts of political obligation, but doesn’t refer to some of the other lines that have been tried, e.g., the principle of fairness.  Accuracy (and even modest improvements in accuracy) in matters of common concern sounds like a nonoptional public good of the type George Klosko has called "presumptively beneficial."  The suspicion arises that Estlund’s epistemic proceduralism could turn out to be, in essence, a sophisticated variant of some other type of theory.  Estlund says:

    “Democratically produced laws are legitimate and authoritative because they are produced by a procedure with a tendency to make correct decisions.  It is not an infallible procedure, and there might be even more accurate procedures.  But democracy is better than random and is
    epistemically the best among those [procedures] that are generally acceptable in the way that political legitimacy requires.”  (p. 8, my emphasis) 

    Presumably, a generally acceptable procedure will be a fair procedure.  Joe’s point is that  the “generally acceptable” qualification may turn out to be carrying a lot more of the load than the “tendency to make correct decisions,” especially given the Expert/Boss Fallacy and the No Qualified Objections Condition.

  4. Hi Bill and Joe, thanks for your comments. The idea that receipt of benefits, or a fair-play approach could also generate normative consent is an intriguing idea, though I don’t think it’s one Estlund entertains in the book. I suppose this would really just be the fair play theory with a different label attached as the conclusion – instead of saying receipt of benefits generates a duty to obey, we would say it generates normative consent.
    I’ll have to think more carefully about whether increased accuracy in decision-making itself could be defined as the kind of presumptive benefit Klosko has in mind, but in the meantime, I’ll point out one way such a fairness theory seems different than what Estlund is offering. Estlund’s standard of correctness seems to be justice, that is, democratic decisions are more correct the more they accurately map what justice requires. Epistemic proceduralism is thus, I think, parasitic on a kind of natural duty account of political obligation. We ought to obey democratic procedures because they are the most accurate permissible mechanism of telling us what justice requires, and we ought to do what justice requires. So, at least to my mind, Estlund’s theory has more in common with natural duty approaches than fair-play or receipt of benefit approaches. But this conclusion still fits with your more general point that Estlund’s theory may be a more sophisticated variant on an already existing account of authority/obligation.
    On your final point – I’m not sure whether a generally acceptable procedure must necessarily be a fair procedure in Estlund’s theory if by fair we mean equal. Consider what Estlund says at the very end of the chapter:
    Equality in political matters is also not some natural right, even if a certain kind of equal regard is. Political equality depends on, and finds its limits in, what sorts of arrangements will allow the promotion of justice and the common good in a way that can be justified to the broad range of points of view that are owed acceptable justifications for the coercive political arrangements under which they live. Inequality of various kinds is bound to pass this test, but I have argued that the overall system seems bound to be recognizably democratic in its procedures for making law and policy (p. 20).
    To me this passage seems ambiguous as to whether a system that contained political inequalities could be acceptable to all qualified points of view. I could imagine a proponent of epistemic proceduralism claiming that the main reason political equality in voting rights is a necessary feature of the system is because we cannot agree on who the political experts among us are. If there was universal agreement amongst qualified persons regarding who the political experts were, maybe epistemic proceduralism would the dictate that votes be distributed unequally? It seems like this could be justified to all qualified points of view once we grant the premise that the experts are identifiable.

  5. Thanks for your summary and critical remarks, Jonathan. I had some questions about the chapter too.

    A. Your final point touches on a concern I had. I should say that I approach questions of political legitimacy (and much else) from the viewpoint that the compatibility or acceptability of a political system with or in terms of the (reasonable, sincere, etc.) moral attitudes, beliefs, and volitions of citizens, is no real criterion of legitimacy. What matters are the characteristics of a political system itself, and not also its justifiability from within the moral perspective of each (reasonable, sincere, etc.) citizen. A sort of externalism rather than internalism, if you like. So I am not a subscriber to, e.g., Rawls’s liberal principle of legitimacy.

    In the context of rejecting the "expertise -> authority" implication, Estlund very explicitly raises the banner of what I call internalism:

    “The problem isn’t exactly that you haven’t consented, and we’re not assuming that consent would be required to legitimate rule. It’s about what you believe; you do not believe that they are experts.” (3-4).

    A person’s (qualified, reasonable, sincere, etc.) belief that X is not an expert (or something similar) prevents X from having authority over them.1

    But this just seems far too quick for me. Granted, the pope does not have authority over me, and would not even if he were an expert on moral matters. But there are a variety of ways in which the "e -> a" implication can be blocked. Most obviously, the pope hasn’t won a democratic election. Perhaps even if he had won an election, authority could be blocked on other grounds too: perhaps he is likely to violate my rights where rights are necessary conditions of legitimacy. Perhaps the kind of view he (correctly) espouses doesn’t have enough empirical confirmation to be authoritative, where empirical confirmation of X may be distinct from X being (reasonably, sincerely, etc.) morally acceptable to each (qualified) viewpoint. And so on.

    Now, I know that Estlund can build in all sorts of criteria of legitimate authority such as the ones above, so I’m not advancing them as alternatives to Estlund’s total set of criteria. Rather, I’m confused why we also need this commitment to acceptability to all qualified points of view as itself a criterion (and one that may do a lot of work in the rest of the book). The need to block "e -> a" is insufficient, and I can’t really make out anything else. I take it the discussion of normative consent (esp. the claim on p. 11 that there must be some or other  link to the person’s will for authority to be justified) is based on the earlier discussion of qualified acceptability, so consent theory as such wouldn’t be the basis. In short, why go so (sort of) Rawlsian so (sort of) soon?

    B. Something different: about your first concern re: correctness of outcome as the standard of assessment of a procedure, would anything change in the logic of an epistemic approach if we made it, instead, something like goodness where a good decision is a minimally acceptable one (i.e. not tyrannical or in violation of basic rights, etc.)? Here coin flips may still be problematic since they need not track good decisions.  But we wouldn’t have to worry about subsidising football v. opera because, hey, either way it’s all good. This won’t entirely eliminate the second Waldronian concern since we disagree about goodness/minimal acceptability too. But it may make it a bit more manageable, since people may be able to agree that democratic procedures are pretty good at getting pretty decent decisions, even if they disagree about just what good or pretty decent or minimally acceptable decisions look like. There could be a rough overlapping consensus on that, even if there were no consensus on correctness. This seems to me to be similar to the logic of Rawls’s later "family of political conceptions" position on stability. It would weaken the epistemic nature of theory a little,  but Estlund already does that with his modest epistemic proceduralism. So instead of having a modest epistemic proceduralism towards correct decisions, we could instead have a less-modest epistemic proceduralism towards good/decent/OK decisions.

    This may begin to merge with what the point that Bill has relayed from Joe Adams in #3 above, if receipt of enough benefits is sufficient for an outcome to be good enough.

    1 Similarly, Estlund writes in his review of Waldron that some cases of disagreement defeat claims to political authority, i.e. the wrong sort of (reasonable) moral controversy about the wrong sort of (reasonable) moral question can itself scupper claims to authority (“Jeremy Waldron on Law and Disagreement,” Philosophical Studies, 99 (1), 2000: 111-28, p. 111). He’s writing about Waldron’s views there, but if I remember it’s a sympathetic comment, and it coheres with what he says here. I’m not sure if Waldron has to be interpreted in a way that implies his endorsement of this sort of view, especially as Estlund is right that it would create all sorts of problems for him. But that’s another matter.

  6. I should add that if anyone thinks they might like to participate in discussion, but have not yet registered, please do so today, i.e. Tuesday, as over the next few days registration approval will be very slow. If you have tried to register but have not received an email with your password then please email me directly at simonmay at vt dot edu. 

  7. Hi Simon, thanks for the comment. On your first point, I’m more sympathetic to the Rawlsian move Estlund makes, so let me try and provide some sort of reply to your worry.
    You use the example of the pope and say he would not have authority over you even if he were in fact an expert on moral matters. But you don’t think we need Estlund’s idea that it is because you have a qualified belief the pope is not an expert in order to block the move from expertise to authority. In favouring a kind of externalism about authority, I take it you want to say that your beliefs about who is an expert shouldn’t really come into the equation. You go on to suggest there are more straightforward ways of blocking the inference from expert ? authority, but I wasn’t so sure. You suggest, ‘Most obviously, the pope hasn’t won a democratic election’. But doesn’t this just beg the question? We want to know why someone needs to win an election to have authority. You go on to say that maybe the pope’s views, though correct, lack a sufficient degree of empirical support to be authoritative, and that this could be distinct from acceptability to qualified points of view. But can we cash out the idea that a claim requires level n of empirical support in order to be authoritative without relying on something like the qualified points of view test? I think this could just be like saying ‘until a claim has level n of empirical support, it won’t seem sound or acceptable to everyone – people could reasonably reject it and propositions can’t be authoritative if people can reasonably reject them’.
    Of course you are also raising a more general objection to contractual approaches, namely, the claim that they are vulnerable to the Euthyphro dilemma. The appeal to what certain people endorse or accept is unnecessary since we can just cut straight to the reasons why they endorse or accept whatever it is. I think this isn’t the place for me to try and defend contractualism from this sort of objection (mainly because I don’t think I’m up to it!), but I’ll make one small comment in defence of the idea of qualified or reasonable agreement. These appeals to idealized agreement don’t have to be construed (thought they sometimes are) as somehow adding to the moral reasons that already exist. They can instead be, as Rawls says, a device of representation – a helpful way of systematizing our moral reasons and making them more vivid. Of course some people think they aren’t helpful, that they just obscure what’s really doing the work, but this might just be a matter of what we’re philosophically most comfortable with. I happen to find contractual approaches helpful, particularly when  trying to cash out the anti-aggregative idea that Rawls calls the separateness of persons.
    A smaller interpretive point – I don’t think that Estlund’s discussion of normative consent is necessarily based on the qualified points of view test. Normative consent, as I understand it, just refers to those conditions where it would be morally wrong to refuse to consent to someone’s authority. It seems to me someone could believe such situations exist (e.g. when there’s a medical emergency on the street and a passing doctor starts issuing commands to nearby people) without accepting anything that Estlund has to say about authority needing to be acceptable to qualified points of view.
    On your second point – I like your suggestion and it seems as if it would make the justification of epistemic proceduralism easier, so it will be interesting to see what David has to say!

  8. Hi Jonathan

    I don’t want to beg any questions about democracy of course : ) — Sure, there would have to be a theory about why egalitarian democratic processes generated legitimacy and authority. My point is really just that it’s not implausible to think that one could provide such a theory without having to appeal to reasonable/qualified acceptability. It may be that we can get a good theory of why certain people have the entitlement (or the justified claim) to make and enforce decisions over others by appealing to (a) the need for some or other decision procedure, (b) something uniquely fair about democratic procedures, and maybe (c) something neat about how democratic procedures tend to achieve the right sort of results.  None of this *has* to appeal to the idea that legitimate authority depends on reasonable/qualified acceptability. Of course, we could defend the view that such acceptability is indeed a necessary condition of legitimate authority, but it’s surely not true just because ~(e->a).

    Re: publicity, if one did buy into the reasonable acceptablity criterion, then it would surely follow that ceteris paribus, political arrangements needed support from public evidence. But I don’t think the converse is true. E.g. in the case of trials, we limit our conclusions to only certain kinds of evidence. There are many reasons we do this — for instance the general propensity of public evidence to lead us to right outcomes, and the propensity of privileged information to lead to incorrect outcomes (even if the privileged information happens to be right in a certain case). These aren’t the same as saying that public evidence is needed because only then could people be reasonably expected to accept the process/outcome. I think there is a hell of a lot more to say about the function of publicity in a democracy than just this, of course, but the point I want to make is the small one that it doesn’t all have to be grounded in reasonable/qualified acceptability to be compelling.

    I don’t really want to commit myself to any particular theory right here, or any particular reason why ~(e->a) … Rather, I just want to put in a marker that ~[~(e->a) -> reasonable acceptability is a criterion of legitimate authority]. Or at least, the gap needs to be filled.

    Re: normative consent, I agree with what you say about how the notion could be used without any commitment to reasonable acceptability. But I guess the question I had is why do we need to talk about *any* sort of consent at all? DE wants to keep *some* connection to the person’s will, but why is this? I took it that reasonable acceptability could provide a reason for keeping some or other form of consent in the theory, even if not the original actual consent theory enchilada. Or maybe I am reading too much into that section, and the point of the normative consent discussion is simply to deflate actual consent.

    (I find contractualist approaches actually very amenable in the manner and for the reasons you cite — particularly as a way to avoid aggregation and respect the separateness of persons. But I don’t think a contractualist device needs to be cashed out in terms of justification to each person relative to their reasonable beliefs. We could, instead, just specify a range of interests that each person in a contractual situation is motivated to protect. I know if its contractualism, it’s the black sheep of the family. But I think this is moving a bit beyond democratic theory in particular.)

  9. Just a quick note, since it was mentioned that I’d be delighted to participate in this discussion. The comments so far are thoughtful and constructive and it’s tempting to chime in after each one. But I think it might be best if I (mainly) just say some things toward the end of each week’s discussion (which can, of course, continue after that). So, if I can restrain myself until then, I’ll plan to address some of the points this weekend or so. This first week is a little unusual, since most of the themes are treated in detail later in the book, but there’s still plenty in the comments to talk about. Please let me know if some other plan would suit you all better.

  10. Simon: 
    I agree with your point that the reasons we have to restrict the kind of considerations that can be used in criminal trials may have nothing to do with what reasonable/qualified people could accept. But I’m not sure that shows that we can justify the authority of these procedures without reference to something like the reasonable/qualified acceptability condition.  Unlike you, I’m less certain we can block the move from expertise to authority without appealing to something like reasonable or qualified acceptability, but maybe we’ll have to agree to disagree for now as to how necessary this feature is to a persuasive version of epistemic proceduralism.
    On normative consent, I do think (though perhaps David will clear this up for us) that the point of this idea is simply to deflate the more extravagant claims of voluntarists who think consent is always necessary for legitimacy or authority.
    David: It’s great to have you as a participant in the group, and you should feel free to join in the discussion whenever you like. If you feel the best way to do it is by posting a comment towards the end of each week that’s fine, but don’t hold back if you want to respond or comment on something right away!

  11. Hi David — let me second what Jonathan said about joining the discussion. Also, if you see discussion that is tending to proceed on the basis of a misinterpretation of Democratic Authority that no-one is clearing up, that may be a good reason to jump in sooner than you would to simply respond to criticisms.

    Jonathan — right, I think we should see how the epistemic proceduralism unfolds in the book to see whether qualified acceptability is a compelling part of the theory as a whole.

  12. Ben asks: "isn’t it true that he at least could (whether or not he does) say that democracy is both intrinsically and instrumentally valuable?" It would be nice to hear David’s answer to this question. Is he committed to the stronger claim that the value of democracy is only instrumental-epistemic or is he committed to the weaker claim that at least part of the value of democracy is instrumental-epistemic? If the latter, would David recognize that democracy also has intrinsic value? If so, what would this value be, and what, if any, would be its role and weight in a philosophical defense of democracy?
    It seems to me that those who defend democracy by appealing to the idea of "giving every (adult) person an equal chance to influence the outcome of the decision"  (p.6) are not only concerned with procedural fairness as described by David (though they do care about that). Perhaps they think that an important intrinsic value of democracy is that it recognizes the status of citizens as equal agents who ought to have the opportunity to affect the outcome of the political decision making process through their actual participation. On this view, people do not only care about what decisions are made, but also about who makes them. Imagine, for the sake of argument, that democracy has an intrinsic value understood in this way. If David accepts this (would he?), what weight would this value have? Would it be part of the explanation why we have reason to accept as legitimate some decisions made in a democratic way even if they are substantively unjust? Would it be part of the reason why we would be uncomfortable with "epistocracy" even if perfect epistocrats could be found? Might there be reasonable tradeoffs between considerations tracking intrinsic value and considerations tracking instrumental value?

  13. David: You probably answer these kinds of questions later on in the book. So far I have only read ch.1…

  14. I have a question about the utopophobia point (Jonathan’s 5th comment above).  I’ll just summarize the debate so far:Estlund points out that in moral philosophy the fact that people often lie is not taken as a reason to revise our views that lying is wrong.  In political philosophy, however, people often think that if a moral standard is too hard to satisfy, we should lower our sights.  If citizens are too unlikely to vote for what is right or good for all, for example, it may be said that we should not hold citizens and our institutions to the demanding standard of epistemic proceduralism.  Why the asymmetry, Estlund asks – why not simply say our existing practices are not fully legitimate, and that citizens ought to do what they tend not to do, just like they ought not lie even though they do tend to lie?  In both moral and political philosophy, we can’t have an obligation to do the impossible, but there is big gap between "unlikely" and "impossible".Jonathan adds two further points.  First, if a candidate moral standard is morally too demanding, it’s not unrealisitic, but just wrong.  Second, even if we can’t have an obligation to do the impossible, it is useful to have identified pure, fact-free ideals, because we are generally uncertain about what is possible and what is not.  I guess the fear is that if we don’t start with pure, fact-free ideals, we may label as ‘simply right’ practices that are only right given that fact X holds, and so we may be less than fully critical about whether X is really an inevitable fact, rather than a local or changeable fact.Let me then make the case for viewing moral and political philosophy differently.  Moral philosophy is about what is right – full stop.  Political philosophy is about the right way to handle disagreements about what is right, and conflicts of interests where there may be no clearly right resolution.  Any moral standard for assessing political practices must therefore not be so ideal that if satisfied it would eliminate the need for authoritative collective decision-making procedures in the face of serious disagreement. There is a parallel here, of course, with debates about the circumstances of justice.  Joe Heath makes the point in a recent defence of Rawls on global justice:  "We cannot argue about property rights, without the recognition that people often refuse to share with others, even when they have more than they need. In an ideal world all of these problems would not exist, and it is always “possible” that they could be eliminated. But if we assume them all away, we simply eliminate from our purview all of the interesting questions that a theory of justice might be expected to address. In fact, there is good reason to think that in an ideal world, we would not need justice at all."  ( don’ t think the demand that voters deliberate about justice and the common good is utopian in this sense, because even if voters were public-minded in the voting booth there would still be disagreement.  But it struck me in reading pp.12-15 that there was more to be said in favour of there being a difference between moral and political philosophy, in terms of utopianism and facts about feasibility.   

  15. Sorry about the lack of paragraphs  – I’m not sure why it didn’t like my carriage returns.

  16. Hi Andrew,
    I’m interested in your proposed way of distinguishing between moral and political philosophy, but I’m a little unsure about it. You suggest that maybe ‘Moral philosophy is about what is right – full stop.  Political philosophy is about the right way to handle disagreements about what is right, and conflicts of interests where there may be no clearly right resolution’.  This implies that ‘any moral standard for assessing political practices must therefore not be so ideal that if satisfied it would eliminate the need for authoritative collective decision-making procedures in the face of serious disagreement’.
    But I’m not sure this maps on to how most people make the distinction between the moral and the political. I think most people’s intuitive distinction between the moral and the political probably doesn’t make reference to the fact of disagreement, but rather distinguishes between those things we see as more ‘personal’ matters (e.g. lying to one’s friends) versus those political/macro matters that where the proposed principles govern the conduct of society as a whole (e.g. the right distribution of income in society). Some people might also make the distinction by appealing to justice or coercion: moral philosophy is the broad domain of rightness, and political philosophy is the sub-field of rightness where we think questions of justice or permissible coercion are applicable. Neither of these two ways of making the distinction refers to the fact of disagreement or conflict where rightness is non-existent.
    If one of my proposed distinctions is plausible, then we have no reason (thus far) to think political philosophy should be any more fact-sensitive (to use Cohen’s term) than moral philosophy. Political philosophy addresses specific questions within the broader terrain of normative ethics, and people may or may not disagree about these questions just as they might about any other normative questions. Consider the question: what is the just distribution of burdens and benefits in society? Surely I’m doing political philosophy if I address this question directly and try to work out the answer? Just to drive this point home – imagine a group of people who (in the way they always do in political philosophy) find themselves washed up on a deserted island. The question then arises, what is the just way of distributing the island’s resources? The shipwrecked people don’t disagree about the answer, they just don’t know what the answer is. They are genuinely puzzled and would like to know the answer, but they lack the philosophical ingenuity or confidence to some up with competing positions. Surely the question they’re puzzling over is one that we would call a question of political philosophy, even though there is no disagreement or conflict?
    Of course people in our world do disagree about the answer to this question, and so there is then a further interesting question regarding which distributions the state can legitimately impose given the fact people disagree about the first issue. But I think this latter question isn’t definitive of political philosophy, it’s just one topic within political philosophy. It’s also worth pointing out that there’s the possibility of infinite regress here. We can always ask the yet further question: what can the state legitimately do given that people disagree about what the state can legitimately do? And so on.
    I don’t think we should define political philosophy purely in terms of addressing these further questions about what to do in light of the fact of disagreement about some other question, particularly since it seems to me that our answers to the disagreement question must be informed by our beliefs about ‘first-order’ questions that do not incorporate the fact of disagreement. That is, in order to decide what we can legitimately do in light of the fact that people disagree about distributive justice, surely I must directly address the following question: when can one person or group legitimately exercise power over another? I need some first-order views about legitimacy before I can address problems of disagreement. What do you think?
    **Also, just a quick note on the quote from Heath. I’m a little puzzled by his suggestion that we can’t think or argue about distributive justice without acknowledging that people sometimes don’t share with others even when others’ need is greater. He suggests that in an ideal world (presumably one where people did share with others when others’ need was greater) we would have eliminated all the interesting questions about justice – that we would not need justice at all.

    But Heath’s assumptions about what an ideal world would look like don’t seem to me to eliminate the topic or the question of justice, rather they posit a world where everyone behaves as justice requires. It sounds like Heath is saying, what justice ideally requires is something like distributing resources to people with the greatest need. A world where everyone followed that principle isn’t a world where debates about justice don’t make sense or have been eliminated (we would want to know how to define neediness for a start), it’s rather a world where justice, according to this conception, has been realized. So I don’t think I accept the idea that too much idealization of people’s behaviour somehow eliminates the philosophical question of justice – Heath’s example shows we need to think about the question in order to know what would count as idealized behaviour.

  17. Piers Turner says:

    Thank you to Jonathan Quong for his nice summary of Estlund’s first chapter. Thanks also to Estlund himself for his work – which I think is some of the best stuff out there. I’d like to raise a question about the expert/boss fallacy and Estlund’s not going epistemic whole-hog.

    I’ll start with a simple summary, as I see it [but am not entirely confident of]:

    1. Assume (for now) something like the ‘liberal principle of legitimacy’ (LPL), such that a political procedure must be generally acceptable to qualified points of view.
    2. Then ask: “Why democracy?”
    3. It can’t be mere fairness, because a coin-flip would then suffice.
    4. Rather, we care (also) about the correctness of outcomes. This introduces an epistemic (truth-seeking) dimension.
    5. Democracy has epistemic value (better than random choice)
    (a) not because of Condorcet
    (b) not because of democracy/contractualism analogy in which deliberation/agreement constitutes moral facts
    (c) but because it approaches “an ideal epistemic situation” for “deliberators for whom there are independent facts about what ought to be done” (18).
    6. But given epistemic dimension, why not epistocracy?
    7. Answer: because it can’t pass LPL
    8. Of those procedures that can pass LPL, democracy has the greatest epistemic value.
    1-8 seems to establish democracy’s legitimacy.
    9. Given 8, democracy also enjoys normative consent (because it would be morally wrong to refuse consent)
    10. Normative consent establishes democracy’s authority (because “there must be some link to the obligated person’s will”)
    1-10 establish democracy’s legitimacy and authority.
    Despite the name “epistemic proceduralism,” Estlund denies that authority and legitimacy are purely epistemic matters. To believe they are is to succumb to the “expert/boss fallacy,” but (as he writes) “authority does not simply follow from expertise”. As with doctors, expertise gives us reason to consent to some authority, but it is neither the expertise nor the consent alone that grants authority. So with democracy: “democracy… is epistemically the best among those that are generally acceptable in the way that political legitimacy requires” (8).
    Estlund seems to offer two necessary conditions on political legitimacy. A legitimate political procedure must be generally acceptable to qualified points of view and it must have epistemic (better than random) value. The latter may aid in its being generally acceptable. The truly authoritative and legitimate procedure will be the one that has the greatest epistemic value of those that have passed the general acceptability test, because it would be “morally wrong to refuse to consent” to this procedure. This procedure thus enjoys “normative consent”. And this procedure is democracy.
    Now, as Simon has emphasized, what we get in Ch. 1 is some discussion of the epistemic value condition, but little discussion of the general acceptability condition. This is presumably because the general acceptability condition is so generally accepted. But given the focus on quality of outcomes in Estlund, we might wonder why he doesn’t simply go epistemic whole-hog. The general acceptability condition does work in Ch. 1 in Estlund’s rejection of ‘epistocracy’. But we wouldn’t want that to be the only reason it is there. So we’ll wait on that discussion.
    But consider the possibility of going epistemic whole-hog. Estlund says this would be an instance of the expert/boss fallacy. Consider his case: “Suppose your religious point of view is not the true one. If you think the knowers should rule – if expertise entails authority – then you must think that those with the true religious perspective, whoever they might be, should rule even over people like you who mistakenly doubt that they are the knowers.” But there is an ambiguity here. Estlund seems to regard ‘knowers’ as those with merely true belief (there is a pedigree for this way of talking, in some of Goldman’s work). I might be misreading him here. But, continuing the thought, I don’t think the epistemic whole-hoggist is committed to the view that authority is to rest with whoever happens to have true belief. This is implausible. Rather, authority on epistemic grounds has to do with a tendency, roughly, to be responsive to the available evidence. Roughly, authority would track rational belief, not true belief – even though what is rational would be cashed out ultimately in terms of conduciveness to truth. Estlund worries that epistemic reasons couldn’t provide legitimacy in the absence of correctness. But can’t there be epistemic legitimacy even in the absence of correctness, on analogy with epistemically justified belief that is false? He writes: “We should not assume that authority and legitimacy lapse whenever the procedure gets a wrong answer.” (7) But the epistemic argument isn’t committed to that, I want to suggest.
    I guess, then, I don’t quite see why I should agree that “someone’s knowledge about what should be done leaves *completely open* what should be done about who is to rule” (p. 3). This might be obvious if knowledge and expertise is merely a matter of having true belief. But if expertise is a matter of being the most likely to make a correct judgment – based, say, on past experience – then expertise might be an important guide for who should rule. I think Mill was onto this in On Liberty.
    The response, presumably, is that you can reasonably doubt expertise in this sense, too – you can reasonably doubt that some agent is most likely to make the best judgment – and, as in the other case, “this doubt appears to block the inference from their expertise to their authority” (p. 3). But why? I’m tempted to say that if some person or procedure is most likely to make a correct judgment on some matter, she/he/it has authority whether I recognize it or not. If democracy is, as some think (e.g. Cheryl Misak), the epistemically best system – why wouldn’t that be enough to make it authoritative? Estlund himself notes the epistemic concern with being “intelligent” (p. 4), but intelligence is less a matter of having true belief than of holding beliefs rationally. (So I may be misreading Estlund on knowledge as true belief.)
    One worry it is tempting to read into Estlund is that a purely epistemic argument for democracy may not yield moral authority because it involves a kind naturalistic fallacy. Estlund wants to show “how [democracy] might have some epistemic value in a way that could account for the degree of authority we think it should have… epistemic reasons are not what we need. The hope is to show how democracy yields moral reasons to obey the law and a moral permission to enforce it” (7) Authority and legitimacy are moral notions. As it happens, the concern for correctness of outcomes will matter to all qualified points of view, and this agreement ultimately generates moral legitimacy and authority. And so it seems that the expert/boss fallacy is (for Estlund) analogous to the naturalistic fallacy because it tries to turn epistemic meat directly into moral sausage, when we need a moral-sausage machine (LPL and normative consent) to turn that trick. (And this accounts for Estlund’s “completely open” comment above – it is like the ‘open question argument’.)
    Do we need a moral-sausage machine like LPL? Perhaps not. Consider Buchanan’s work on social moral epistemology. There, a generic concern about moral risk (the risk of performing morally objectionable acts) as a result of having false (factual) beliefs that distort our moral powers (like sympathy), gives epistemic institutions moral import. But to make that move from epistemic to moral, we didn’t need to move through something like LPL (even if Buchanan himself wants to do that.) On consequentialist grounds, one could argue that good epistemic institutions are important contributors to improved moral and social conditions. I’d have thought democratic institutions would be a part of this moral-epistemic picture, pace Estlund’s worry about democracy as the best overall epistemic procedure.

  18. I agree that the shipwrecked people face a question of political philosophy, when they are trying to figure out what the just distribution is, prior to any disagreement.  If there is a question about the just distribution, however, it must be because there is scarcity relative to wants, and thus a conflict of interests (unlike in Plato’s "healthy city" i.e. the "city of pigs" from Bk 2 of the Republic), or scarcity relative to conflicting ends, based on disagreement.  If people were perfectly moral, maybe there wouldn’t be a problem of distribution to solve.  But then I suppose you would say (Jonathan) that the true principles of justice would be satisfied by default.  And in any case, these principles would not depend for their justification on the factual question of whether or not a society needs to discover and apply them.   (paragraph break, I hope)  When we move to a discussion of democracy, we have to presuppose that people disagree.  We’re looking for a collective decision procedure because people have different views about what we should do, and we have to aggregate these views to generate a decision that people ought to regard as binding even if they disagree with it.  So our standard of democratic legitimacy can’t be so demanding that, if satisfied, we wouldn’t need a decision procedure at all. Of course, asking voters not to vote selfishly doesn’t preclude their disagreeing, so this isn’t really a problem for Estlund.  I was just struck by the parallel between moral and political philosophy, on p.12, and trying to figure out why there seemed to me to be an important difference.

  19. Andrew: OK, I certainly agree with you that once we move into the realm of democratic theory, we have to assume disagreement as one of the ‘facts’ which motivates our theory – no disputing that.  I just wanted to defend the view that political philosophy as a whole shouldn’t be defined by reference to addressing problems of disagreement, even if this forms a big part of the field.

  20. I have just posted some responses to the excellent discussion of this chapter so far. Simon preferred that it be a "post" rather than a comment. 

Leave a Reply