This article addresses the philosophical foundation for the legitimacy of the democratic procedure of majority rule by relating it to a debate that has recently been attracting a lot of attention within the field of democratic theory: whether it is possible to justify democratic procedures by relying exclusively on a set of “purely procedural” values (Dahl 1998; Estlund 1997; Fabienne 2007; Habermas 1994; Rawls 1993; Shapiro 2003; Urbinati and Saffon 2013). For the purposes of this analysis the latter are defined as values pertaining only to the procedure through which collective decisions are generated. The counter-concept is therefore that of “substantive” values, which permit an evaluation of political outcomes independently of the procedure through which they are generated (Estlund 2008).
Several issues are at stake in this debate. The first concerns the nature of the reasons for supporting democracy in the first place. If majority rule can be justified based on a set of purely procedural values, then democracy can be considered intrinsically valuable in the sense that it expresses a set of presupposed normative commitments (Urbinati and Saffon 2013). Conversely, if the philosophical justification of majority rule requires the appeal to substantive values, democracy is only instrumentally valuable because it is simply a means to an independent set of ends (Anderson 2009). This, in turn, poses the question of whether democratic procedures are to be considered normatively desirable even if (or when) they fail to achieve a specific set of substantive outcomes—such as social justice or welfare maximization (Landemore 2013; Schwartzberg 2015).
Furthermore, the debate over whether democracy can be justified in purely procedural terms also of course touches on whether there are any substantive standards of political truth or correctness at all—that is, whether there is an objective criterion of justice or social welfare (Arendt 1967; Elkins and Norris 2012; Estlund 1993; Misak 2002; Rorty 1997; Zerilli 2006). Not having to make any definitive commitments in that regard is surely an attractive feature of purely procedural justifications of democracy (Rawls 1993). To be sure, the latter do need to presuppose the normative validity of the procedural values they employ, so neither purely procedural nor substantive justifications can do without prior normative presuppositions (Dworkin 2011). Nonetheless, it seems plausible to assume that there is likely to be a much wider degree of disagreement over the normative acceptability of political outcomes, compared to the procedures for generating them. Thus, in the final analysis the question of whether it is possible to provide a purely procedural justification of majority rule affects the degree of normative pluralism with which a philosophically sustainable defense of democracy is compatible (Galston 2000).
All of these issues are posed in a particularly poignant way by a specific strand of contemporary democratic theory, known as “epistemic” theories of democracy (for an overview see Schwartzberg 2015). In his book on Democratic Authority (2008), for instance, David Estlund argues that purely procedural theories of democracy fail to establish a moral duty to obey collective decisions because such a duty can only stem from the presumption that political outcomes correspond to an objective standard of normative “truth.” Similarly, in her book Democratic Reason (2013) Hélène Landemore has sought to show that democratic procedures have a tendency to produce “good” decisions based on the assumption that a purely procedural justification of majority rule would be “insufficient.” Other authors who have also recently advanced analogous claims include Elizabeth Anderson (2006), Susan Stokes (2011), Josiah Ober (2012), and Adrian Vermeule (2012).1
The idea of providing an epistemic justification of democracy therefore seems to be attracting widespread interest to the extent that some commentators have even begun to speak of an “epistemic turn” within the field of contemporary democratic theory (Jorke 2010; Landemore forthcoming; Palumbo 2012). To be sure, many authors have also expressed concerns about the project of reintroducing an orientation toward “truth” within the field of democratic theory (Elkins and Norris 2012; Invernizzi-Accetti 2015; Urbinati 2014; Urbinati and Saffon 2013; Zerilli 2006). As yet, however, there exists no systematic attempt to respond, from a critical perspective, to one of the key contentions of this strand of contemporary political theory, namely that democratic theory needs to make reference to a substantive criterion of normative “truth” because it is impossible to provide a purely procedural justification of majority rule. This is precisely the claim I intend to take up in this article, showing that a purely procedural justification of majority rule is not only possible but also philosophically preferable to the epistemic ones advanced by the authors mentioned above.
The discussion is divided in two parts. In the first I show that a purely procedural justification of majority rule is possible by addressing the criticisms raised against it by epistemic theorists of democracy and developing a version of such a justification that is not susceptible to them. In the second part I show that this purely procedural justification of majority rule is preferable to the substantive ones proposed by epistemic theorists of democracy, pointing out a number of theoretical disadvantages inherent in the project of reintroducing a normative standard of truth within the framework of democratic theory.
Part I: Can Majority Rule Be Justified in Purely Procedural Terms?
Let me begin by considering the objection that epistemic theorists of democracy raise against the very possibility of providing a purely procedural justification of democratic decisions. This is usually framed as a criticism of the idea that procedural fairness can provide an adequate basis for a philosophically self-sufficient justification of majority rule. The key argument advanced is that the latter is too broad and, therefore, potentially compatible with a wide variety of collective decision-making mechanisms, not all of which are necessarily democratic.
For example, in his book on Democratic Authority David Estlund defines “procedural fairness” as the idea that all participants in a collective decision-making procedure ought to be treated “impartially” in the sense that each should be given an equal chance of affecting the final outcome, independently of who they are or any other distinctive characteristics they might have (Estlund 2008: 69–73). On this basis Estlund first concedes that majority rule can indeed be considered procedurally fair because it supposes an impartial treatment of individual votes. However, he also immediately adds that the notion of procedural fairness is insufficient for providing a “complete” justification of majority rule because the latter is not the only way of making collective decisions that satisfies the requirement of impartiality thus defined: “A procedure that holds a vote and chooses one person’s vote randomly,” he writes, “has all the procedural fairness of majority rule and more, since it is blind to all features, not just some. [Thus], if procedural fairness were the justification of majority rule, why not go all the way? … Why not flip a coin amongst possible decisions?” (Estlund 2008: 82).
Elizabeth Anderson, in her article “The Epistemology of Democracy,” advances a very similar argument, as she writes that “there is a longstanding tension in democratic theory between accounts of success that are internal and external to the democratic decision-making process. Internalists, or proceduralists, hold that, to vindicate a decision-making process, one need only show that it is procedurally fair.… [However], if we decide that a problem, such as air pollution, is of public interest and that dealing with it requires joint action under the law, we don’t just flip a coin to decide what pollution laws to enact, even though this would be procedurally fair. Rather, we will judge the success of democratic institutions according to criteria that are (partially) external to the decision-making process” (Anderson 2006: 10).
The key contention on which these arguments rely is that, from the point of view of procedural fairness, majority rule and random selection turn out to be morally equivalent. This is supposed to function as a sort of reductio ad absurdum of procedural fairness itself. What should be noted in this respect, however, is that while the first proposition of this argument is certainly valid, there are a number of reasons why its capacity to sustain the conclusion it is supposed to establish can be called into question.
First of all, even on its own terms, the idea that procedural fairness is incapable of establishing the legitimacy of majority rule does not follow logically from the premise that random selection can also be considered legitimate from the point of view of procedural fairness. For the original question was not whether procedural fairness is capable of establishing that majority rule is the only legitimate decision-making procedure but rather whether procedural fairness is capable of establishing the legitimacy of majority rule as such. There is simply no logical connection between the procedural fairness of majority rule as such and the fact that random selection is also procedurally fair according to the same criterion. At most, what follows from this is that both procedures are legitimate from the point of view of procedural fairness. Note, however, that this still implies that majority rule is legitimate from the point of view of procedural fairness.
In order to make their critique of procedural fairness work, epistemic theorists of democracy are therefore forced to rely on a further premise that has not been stated explicitly up to this point: that making collective decisions through random selection amongst votes is somehow unacceptable or morally absurd. Estlund, for example, alludes to something of this sort shortly after the passage quoted above where, after having established that, from the point of view of procedural fairness, majority rule is at most morally equivalent to random selection, he asserts that “this is an absurd proposal in most political contexts, of course” (Estlund 2008: 82).
The problem is that throughout the whole body of literature that employs this argument as the basis for a critique of procedural fairness, no real reason is ever provided for concluding that random selection is indeed “absurd.” This is simply assumed as a matter of course but never actually justified. Moreover, it doesn’t follow logically from the way in which the notion of procedural fairness has been defined; on the contrary, it runs counter to a long and established tradition in the history of democratic thought that asserts that randomness is a legitimate democratic mechanism for making collective decisions precisely because it is consistent with the requirements of procedural fairness (Delannoi and Dowlen 2010; Manin 1997; Sintomer 2011). In absence of any further discussion of why epistemic theorists of democracy arbitrarily dismiss this whole body of thought without even taking the time to consider its merits, their refutation of procedural fairness as the grounds for majority rule’s legitimacy must therefore be considered unsuccessful or, at least, incomplete.
Finally, when attempting to refute purely procedural justifications of majority rule, epistemic theorists of democracy are too quick to assume that procedural fairness is the only purely procedural ground on which the legitimacy of majority rule could potentially be established. As is clear from the passage by Anderson I quoted above, they simply take it for granted that “proceduralists hold that, to vindicate a decision-making process, one need only show that it is procedurally fair” (Anderson 2006: 10). This, however, is not necessarily the case because there are also a number of other purely procedural values, apart from the notion of procedural fairness, on which the legitimacy of majority rule could be established.
In particular, I show in the following section of this article that a purely procedural justification of majority rule can be provided on the basis of the value of consent, defined as the idea that collective decisions ought to be considered legitimate if the subjects for whom they are binding have consented to them, presumably through a vote. This conception of consent can be considered a purely procedural value according to the definition provided above because it only concerns the procedure through which collective decisions are taken, without making any claims to the substantive conformity of the outcome with any set of procedure-independent criteria of normativity. However, it is striking that epistemic theorists of democracy generally devote very little attention to it, as if their purported refutation of procedural fairness were somehow equivalent to a refutation of all forms of proceduralism.2 To show that the value of consent can provide the basis for a purely procedural justification of majority rule therefore constitutes a way of advancing a counter-example to the epistemic critique of proceduralism that escapes the objections raised against it.
In this section I show that a purely procedural justification of majority rule can be provided on the basis of the notion of consent. This argument relies on an extrapolation from insights already advanced by the great Austrian jurist and political theorist Hans Kelsen in his seminal treatise The Essence and Value of Democracy (1929), a text widely regarded as foundational in continental European democratic theory but is still relatively understudied in Anglo-American contexts (on this point see Urbinati and Invernizzi-Accetti 2013). Within the overall framework of this article the purpose of this argument is to supplement what has already been shown above concerning the connection between procedural fairness and majority rule by suggesting that the notion of consent can also provide the basis for a normative preference of majority rule over other procedurally fair mechanisms. It is therefore primarily to this argument that I shall be referring in the second part of this article, where I show that, as well as being possible, a purely procedural justification of majority rule is also preferable to the substantive ones offered by epistemic theorists of democracy.
The obvious difficulty that needs to be overcome is that the idea of making collective decisions through consent may initially appear inconsistent with majority rule inasmuch as the latter implies that a collective decision may potentially be taken without the consent of all the individuals for whom it is binding. I contend, however, that this apparent difficulty actually rests on a confusion, as the relevant subject of consent to a collective decision cannot be the sum of all the individuals on whom it applies but must rather be the collectivity itself. The value of Kelsen’s text in this regard lies precisely in paying attention to the conditions for collective consent, demonstrating that this cannot require the consent of all the individuals for whom it is binding but rather must take the form of an exercise of majority rule.
The key argument relies on two complementary claims. First, requiring the consent of all individuals for whom a decision is supposed to be binding would deprive the notion of a collective decision of its essential meaning. The second is that consent to a collective decision can be more plausibly assumed to require the consent of a majority of the individuals for whom it is binding. Let me look at each of these claims in turn. The reason Kelsen suggests the first claim is that a unanimous vote is effectively no different from a series of simultaneous individual decisions to do the same thing; what is missing is precisely the element of collectivity itself—that is, the fact that it is a collective entity, and not just a series of individuals, making a decision that is binding for its members.
This point is expressed in terms of the notion of a “social order,” which for the purposes of this discussion can be treated as functionally equivalent to that of a collective decision: “Such an order,” Kelsen writes, “is by its very nature possible only if its validity is objective, i.e. distinguishable from the will of those subject to it. … In the extreme case, where the ‘you ought’ of the social order depends on the ‘when and whatever you will’ of every individual it is addressed to, the order no longer has any social meaning. If society and the state are to be possible, then one must be able to differentiate between the content of the order and the will of the individuals subject to it” (Kelsen 1929: 6–7).
Building on this insight, I establish that consent to a collective decision can be more plausibly interpreted as requiring the consent of a majority of the individuals for whom it is binding. The reason Kelsen provides is that majority rule maximizes the number of individuals consenting to collective decisions. To be sure, this may appear counter-intuitive because if the link between individual and collective consent is conceptualized in terms of maximization, it may still appear that unanimity guarantees that more individuals consent to the collective decisions in force compared to majority rule. That, however, is not the case because unanimity effectively means that, in principle, even a single individual can prevent a collective decision from being taken. If not taking a decision is seen as a decision in itself, this implies that, under a regime of unanimity, collective decisions can be in force that have not been consented to by more individuals than those who have. Majority rule, however, guarantees that the totality of decisions actually being taken (i.e., both decisions and nondecisions) have been consented to by more individuals than those who have not.
A couple of points of clarification may be in order before moving on to draw the implications that follow from this argument. First, Kelsen formulates the key claim in terms of a requirement of correspondence between the content of the social order and the will of the individuals for whom it is binding. This may seem to lapse back into a substantive conception of the grounds for the legitimacy of collective decisions in the light of Estlund’s critique of the procedural nature of consent (see footnote 2). That, however, is not necessarily the case because Kelsen’s argument can be easily reformulated in terms of the purely procedural conception of consent I have proposed without losing any of its validity. This simply requires introducing the idea that nondecisions are also to be treated as decisions as a substitute for what Kelsen says concerning the conditions required for changing decisions that have already been taken.
The only sensible premise for the principle of the majority is the idea that, if not all, then at least as many individuals as possible should be free. This means that the number of individual wills that are in conflict with the general will of the social order should be minimized. … Under these circumstances, the fewer wills one’s own has to agree with in order to effect a change in the will of the state, the easier it is to achieve a concordance between the individual will and the will of the state. Here, then, an absolute majority does in fact constitute the upper limit. Anything less would mean that the will of the state could from its very inception conflict with more wills than it agrees with. Anything more would make it possible for a minority, rather than the majority, to determine the will of the state by preventing an alteration of that will.(Kelsen 1929: 9–10)3
The key point thereby becomes that majority rule maximizes the number of individuals who have actually consented to the totality of decisions (and nondecisions) in force through the mechanism of voting and, therefore, independently of whether these outcomes actually correspond to the will of the individuals for whom they are binding. Whether that is the case is indeed a substantive question, but it is important to note that this need not be appealed to in order to legitimate the outcome because assuming that the right procedural conditions hold for the way the vote is conducted (i.e., that nobody is coerced and that individuals have meaningful options of which they are adequately informed), this can be interpreted as implying that the collectivity as a whole has thereby consented to the decision in question.
The second important clarification concerns the fact that, although the argument above is formulated primarily in terms of the notion of “freedom” (which I have reinterpreted in terms of a procedural requirement of ‘consent’ to collectively binding decisions), it does not abstract entirely from the idea of “procedural fairness” discussed in the previous section. For, the claim that majority-rule establishes the consent of the collectivity because it “maximizes” the number of individual members that need to consent to a decision for it to be passed still depends on the implicit assumption that the consent of each individual member is of equal moral worth to all others. Thus, Kelsen’s argument for majority rule presupposes that all individual members of the collectivity are treated “impartially” in Estlund’s sense, while adding the further requirement of ‘consent’ to collectively binding decisions. This yields the outcome that majority-rule is the only procedurally legitimate form of collective decision-making, thereby overcoming any residual problem that may have remained from the recognition—conceded in the previous section– that “procedural fairness” does not yield an argument uniquely for majority rule.
In the light of these clarifications I conclude that the claim according to which it is impossible to provide a philosophically self-sufficient and purely procedural justification of majority rule is false. On the contrary, what I have shown in this section is that such a justification can be provided by relying on an analysis of the normative implications of the purely procedural value of consent (and procedural fairness). This lays the ground for the claim I will substantiate in the next part of this article—namely, that as well as being philosophically sustainable, the purely procedural justification of majority rule I have proposed here is also preferable to the alternative ones provided by epistemic theorists of democracy.
Part II: The Philosophical “Disadvantages” of Truth
After taking themselves to have established (illegitimately, as I have shown) that majority rule cannot be justified exclusively with reference to a set of purely procedural values, epistemic theorists of democracy generally proceed to suggest that this key democratic institution can nonetheless be justified by relying on a substantive standard of normative truth. The several authors I have discussed do not supply the same arguments to this effect, and, indeed, a large part of the conversation within the field of epistemic theories of democracy revolves around the question of which is the most appropriate way to establish that democratic institutions do have a tendency to generate epistemically “correct” outcomes.
Estlund, for example, relies on a complex readaptation of the Millian idea that democratic deliberation constitutes a way of “weeding out” arguments that appear either incorrect or incoherent, thereby progressively approximating an objective standard of normative truth (Estlund 2008: 159–183). Conversely, Landemore relies on a combination of the Condorcet Jury Theorem and the idea that inclusive deliberation promotes “cognitive diversity” amongst the participants to show that majority decisions taken after open deliberation may be considered “predictive” of the epistemically correct outcomes (Landemore 2013: 145–173). Finally, Anderson makes use of the Deweyan notion of “practical intelligence” to model democratic deliberation as an experimental process seeking workable solutions to given problems in a way that can be considered analogous to the scientific method in the empirical sciences (Anderson 2006: 13–15).
In this article I will not get entangled in a discussion of the relative merits of these competing accounts. For the sake of argument I will instead simply assume that at least one of these accounts—or perhaps a combination of the three put together—succeeds in showing that, if there is any such thing as an objective standard of normative truth, then a set of democratic institutions characterized by open deliberation and majority rule may indeed be an effective way of discovering it and putting it into effect. The argument I do want to advance, however, is that even assuming this is the case, such an epistemic account of the legitimacy of majority rule is neither the most effective nor the most economical way of justifying democratic institutions. For once it has been admitted that a purely procedural justification of majority rule is possible, even assuming that an alternative epistemic account also holds, the question of which is to be preferred must be framed in terms of their respective theoretical advantages and disadvantages.
By this I mean that, between two positions that are equally tenable on logical grounds, there may still be valid political or philosophical reasons for preferring one to the other. Such reasons may have to do with their respective implications from the point of view of the goals they are intended to achieve, their coherence with other normative values that might also be considered important, or their practical applicability. My overall contention in this respect is that, once the question is framed in these terms, the purely procedural justification of democracy proposed here emerges as clearly preferable to these alternative epistemic accounts.
To make this case I will focus on three theoretical disadvantages of attempting to ground democratic institutions’ legitimacy on a substantive conception of normative truth.
The first troubling aspect of epistemic justifications of democracy stems from the conception of politics on which they are implicitly predicated. For the idea that the legitimacy of democratic institutions lies in the fact that they are the best available means for discovering and implementing an objective standard of normative truth seems to be predicated on an impoverished conception of politics as the search for the “right” solutions to a set of previously given problems, effectively reducing it to an exercise in problem solving akin to any other technical or cognitive endeavor.
Anderson states this explicitly in her article “The Epistemology of Democracy” when she writes, “The epistemic needs and powers of any institution should be assessed relative to the problems it needs to solve. … These questions are of particular importance when the problems we need to solve demand the utilization of information that is highly dispersed across society. … Different institutions can therefore be evaluated according to their ability to mobilize and respond to the required information” (Anderson 2006: 8–9).
The same underlying conception of the political is, however, also implicit in the texts of the other authors discussed above. Landemore, for example, begins her book Democratic Reason by posing the question of whether the electorate of her home country (France) made the “right” decision in rejecting the draft treaty for a European constitution submitted for ratification by referendum in 2005. Thus, for her, the political issue in this case was defined by the fact that there was a problem to be solved—whether to ratify the constitutional treaty—and what was at stake was whether submitting this problem to democratic deliberation was indeed the most appropriate way of finding the “correct” solution (Landemore 2013: xv).
The reason the conception of politics that underscores these various passages—and, indeed, the epistemic approach to political justification in general—appears problematic is that the notion of problem solving leaves open the question of establishing what are the problems that need to be solved in the first place, as if this were somehow a matter of fact or evidence and not a normatively laden issue in itself. That, however, is clearly not the case. On the contrary, the way in which political problems are identified and framed is an essential dimension of collective self-government as well.
For example, in the specific case of the referendum over the European constitutional treaty mentioned by Landemore, it was both a recurrent argument employed by the partisans of the “no” vote and a point that has been picked up by several commentators since then that one of the main reasons for the final result was that the French electorate rejected the form of the question itself and, in particular, the way in which it was posed to them (Lelart 2006). This suggests that, beyond the question of establishing the most appropriate solutions to previously defined problems, an important dimension of politics is that of determining the ultimate ends of political action itself and, on that basis, the specific set of problems that actually need to be solved in the first place.
By construing politics exclusively as a way of finding the “correct” solutions to problems that are assumed to be already given in advance, epistemic theories of democracy efface this whole dimension of politics, arbitrarily reducing it to a merely cognitive or technical enterprise, when in fact most of the assumptions that make such a conception of the political possible in the first place remain yet to be accounted for. Moreover, this arbitrary mutilation of the domain of the political appears particularly troubling from a democratic point of view because it effectively undercuts the scope for any meaningful exercise of autonomy within the political domain. For autonomy literally means to pose oneself one’s own laws. But if collectively binding laws are assumed to correspond to the “right” solutions to a set of previously given problems, autonomy thus conceived becomes impossible because the “right” solution to a given problem is already implicit in the nature of the problem itself. Thus, it cannot be up to the solver to determine it for him- or herself.
In fact, it is striking to note that the notion of autonomy itself—and, indeed, the very concept of freedom it is closely related to—is almost never mentioned in epistemic accounts of the grounds for democratic institutions’ legitimacy. The underlying reason for this, I contend, is that the way in which epistemic theorists of democracy construe the notion of “truth” as the foundation for the legitimacy of political institutions is incompatible with a normative commitment to the idea of freedom as autonomy, as the notion of an objective normative truth is by definition independent of the will, or at least the consent, of the subjects concerned. Thus, in order to ground the legitimacy of democratic institutions on such a notion of truth, epistemic theorists of democracy are forced to sever the conceptual link between democracy and freedom as autonomy.
Now, of course, this is only a problem if we assume that some conception of autonomy, at least as government by consent, is indeed an essential component of the way in which the notion of democracy should be understood. However, this is undoubtedly the way in which the notion has historically been understood and defined (e.g., as reflected by the fact that the word “demos” was commonly used in ancient Greek to refer in a generic sense to the object of political rule, suggesting that democracy implies some sort of identity between the subject and the object of political decisions). Thus, it seems that the burden of proof for providing an argument to dissociate the idea of democracy from that of autonomy would seem to lie with epistemic theorists of democracy themselves.
In absence of any principled reason for doing so, the attempt to reconfigure the normative defense of democracy in terms of the notion of truth—as opposed to freedom or autonomy—appears unwarranted. The purely procedural defense of majority rule I have put forward in the previous section of this article, however, does not face this problem because it does not assume that the set of problems defining the sphere of the political are already given objectively in advance and can, therefore, make sense of the connection between democracy and autonomy in a much more meaningful way by supposing that the ground for democratic decisions’ legitimacy lies precisely in the fact that at least a majority of those for whom they are binding must have consented to them.
The second notable disadvantage of epistemic theories of democracy, compared to the purely procedural justification of majority rule I offer here, lies in the fact that the former might actually give more grounds for justifying what Estlund calls “epistocracy”—that is, essentially, rule by a competent elite of enlightened “experts”—rather than democracy itself. The reason is that, if the legitimacy of political outcomes is supposed to stem from the expectation that they correspond to (or at least approximate) an objective standard of normative truth, it would seem to make more sense to entrust the responsibility for making collectively binding decisions to the set of individuals who are likely to be most competent in discovering such a truth rather than to the people as a whole.
This is of course a classic argument that critics of democracy, at least since Plato, have long employed. However, Hans Kelsen also mentions it in the treatise I referred to above, writing that “the assumption that knowledge of absolute truth and insight into absolute values are possible confronts democracy with a hopeless situation. For what else could there be in the face of the towering authority of the absolute Good, but the obedience of those for whom it is their salvation? There could only be unconditional and grateful obedience to the one who possesses—i.e. knows and wills—this absolute Good” (Kelsen 1929: 100).
To be sure, differently from what was shown with respect to the conception of politics on which they implicitly rely, epistemic theorists of democracy prove to be sensitive to this potential line of objection and attempt to provide some preemptive responses to it in their writings. My contention, however, is that the arguments put forward to this effect are unsatisfactory and ultimately end up reinforcing the very concern they are meant to dispel. To show this, I will briefly engage with two such responses in particular.
In his book Democratic Authority Estlund contends that even assuming that an objective normative truth exists and that some individuals have a better understanding of what it is, it still does not necessarily follow that these individuals should rule over the others because it might be the case that it is not immediately evident to everyone who these individuals are. What is needed in addition, for him, is therefore “a procedure for identifying the knowers” (Estlund 2008: 35–36). In this respect Estlund contends that such a procedure cannot itself be based on “epistocratic” principles because that would suppose a shared knowledge that is ex hypothesi absent (i.e., knowledge about who the knowers are). Thus, he concludes that the assumption of the existence of both normative truth and knowledge about it is not only compatible but actually requires reliance on nonepistocratic procedures in order to function as the foundation for political legitimacy, given certain assumptions concerning the kind of knowledge generally available (Estlund 2008: 3–4).
The reason I consider this argument problematic is that it implicitly relies on the introduction of a further criterion of political legitimacy alongside the substantive conformity of outcomes to a standard of normative truth: that the subjects to whom the collective decision applies must be capable of recognizing this conformity and, therefore, of accepting the legitimacy of the outcome for that reason. Indeed, Estlund himself makes this clear when he suggests that the notion of “epistemic validity” can be supplemented with a further criterion of “general acceptability” of political outcomes and still function as a basis for political legitimacy: “Even if there are true standards of better and worse outcomes,” he writes, “there may also be a true general acceptability criterion that brackets the use of true standards. This would be part of the nature of justified political authority” (Estlund 2008: 33).
This move appears problematic from the point of view of Estlund’s own purposes because introducing a “general acceptability” criterion effectively falls back on a procedural conception of the grounds for political legitimacy. For, as Estlund himself concedes, such a criterion “brackets the use of true standards” in that the legitimacy of the outcome ceases to depend on its conformity to a substantive conception of normative truth and ultimately depends on the procedure through which it was generated. Thus, even if his argument were to hold, Estlund paradoxically appears to be able to escape the conclusion that epistemic standards of legitimacy imply a commitment to epistocracy over democracy only by relying on a procedural conception of the grounds for political legitimacy. Far from supporting his overall point, this actually reaffirms the idea that, on its own terms, relying on epistemic standards of legitimacy does tend to justify epistocratic procedures.
Acknowledging this potential weakness in Estlund’s attempted refutation of epistocracy, Landemore, in her book Democratic Reason, pursues a different strategy, attempting to show that democratic institutions are actually better than epistocratic ones in discovering and implementing normative truths. As I mentioned above, her basis for making this claim lies in a combination of the Condorcet Jury Theorem and the idea that an open and inclusive process of deliberation promotes greater “cognitive diversity” among participants (Landemore 2013: 145–173). The problem, however, is that neither of these two arguments are strictly arguments for the democratic principle of “universal” inclusion within the political process, and therefore, both can provide grounds for a defense of a form of epistocracy based on the principle of “cognitive diversity” instead of competence.
The Condorcet Jury Theorem, for instance, posits that as the number of participants in a majority vote tends to infinity, the probability that the outcome will be correct tends to one by relying on the assumption that every individual vote’s chance of being correct is better than half. Whether this assumption holds in practice, it does suggest that, even in principle, the Condorcet Jury Theorem can provide grounds for excluding all the individuals who fall below that threshold from the decision-making process either because they aren’t “smart” or “educated” enough: a point that Estlund himself in fact raises in his discussion of the Condorcet Jury Theorem, where he suggests that even if we were to grant the many simplifying assumptions required to apply this mathematical theorem to real-world political situations, all that could ever be obtained by these means would in fact be an argument for an “epistocracy of the educated” (Estlund 2008: 206–236).
Similarly, Landemore’s argument for “cognitive diversity” is not really an argument for universal inclusion within the process of political decision making but only for the inclusion of the widest possible variety of different views and opinions. In a situation in which multiple individuals were to hold identical or overlapping views, this could lead to the idea that it isn’t necessary for everyone to participate. Moreover—and perhaps more worryingly—in a situation in which only a relatively narrow variety of views and opinions were to actually be represented within a given polity, it could also lead to the idea that additional views that are in fact not already present within the social body should be artificially introduced in the process of political decision making in order to maximize the polity’s “cognitive diversity.”
What shows is that the democratic criterion of universal inclusion and that of maximizing the participants’ cognitive diversity are not necessarily coextensive, implying that the latter leaves open the possibility of justifying certain forms of epistocracy based on the principle of cognitive diversity instead of competence. More broadly, however, this example illustrates that because epistemic theories of democracy do not treat political participation as a normative value in itself but only as an instrumental means for producing epistemically correct outcomes, their commitment to democratic institutions is ultimately only contingent and, therefore, necessarily compatible with the idea that nondemocratic institutions might in some circumstances be superior to democratic ones, according to their own criterion.
Conversely, the purely procedural justification of majority rule I have provided above does not face this problem because it is explicitly predicated on the idea that collective decisions can only be considered legitimate if the whole collectivity to whom they are supposed to apply has actually consented to them. As I have attempted to show, this implies that collective decisions can only be considered legitimate if at least a majority of the individuals for whom they are binding has approved, through a vote that is open to all the individuals in the collectivity, irrespective of any specific views or characteristics they might have. Thus, the prospect of giving grounds for a defense of epistocracy is undercut at the root by rejecting—or at least bracketing—the very idea that political legitimacy must stem from the substantive conformity of outcomes to an objective criterion of normative truth.
Finally, the third significant disadvantage of epistemic theories of democracy relates to the issue of “pluralism,” understood as the fact that, independently of whether or not an objective normative truth exists, individuals may have competing and perhaps even irreconcilable conceptions as to its content (Berlin 1958; Galston 2000; Rawls 1993). This constitutes a problem for epistemic theories of democracy because, if it exists, the truth is by definition one. Thus, from the perspective of truth, the existence of any differences in views or opinions must necessarily be read as the sign of an error or, in any case, a problem that needs to be solved. In this sense epistemic theories of democracy appear normatively averse to pluralism.
To be sure, epistemic theorists of democracy are likely to respond that this is far from being the case because their theories are in fact predicated on a recognition of pluralism as a background condition for politics. Estlund, for example, maintains that differences of opinion are why procedures for making collective decisions are needed in the first place (Estlund 2008: 14). Furthermore, as we saw above, Landemore goes as far as to build the notion of cognitive diversity into the core of her argument for the epistemic value of democratic procedures (Landemore 2013: 145–173). In this sense their theories appear to be perfectly compatible with—and, indeed, to presuppose—a recognition of what Rawls would call the “fact” of pluralism (Rawls 1993).
What needs to be clarified in this respect, however, is that there is a difference between treating pluralism as a fact and displaying a normative commitment to it as a value. This can be cast as a difference between the legitimacy that is accorded to pluralism at the “input” and “output” levels of political decision making. Epistemic theories of democracy can indeed recognize the legitimacy of pluralism at the input level in the sense that they treat it as a background condition for political decision making and, in some versions, even maintain it as instrumental in accounting for the epistemic value of democratic institutions. However, to the extent that the purpose of political deliberation is assumed to be that of discovering an objective standard of normative truth, differences of views and opinions cannot be considered legitimate on the output side of political decision making. Thus, even if pluralism is recognized as a fact, the overarching goal of epistemic theories of democracy must remain that of “overcoming” pluralism. And it is in this sense that I contend that such theories are normatively averse to it.
This appears problematic not just from the perspective of the assumption that a measure of value for pluralism is likely to be an ineradicable feature of politics (Berlin 1958; Galston 2000; Weber 1919). More fundamentally, it also points to a deeper tension within epistemic theories of democracy themselves if there are good reasons to believe that the actual functioning of democratic institutions is itself likely to foster and, therefore, exacerbate such pluralism, as was, for example, suggested by John Rawls in his treatise Political Liberalism (1993). For this would imply that epistemic theories of democracy are ultimately at odds with themselves: on one hand, striving to overcome the fact of pluralism through the discovery of a conception of truth that would put an end to political disagreements and, on the other hand, seeking to defend a set of institutions that is likely to foster and therefore exacerbate such disagreements in the first place.
The purely procedural justification of majority rule I have provided in the previous part of this article is instead compatible with a recognition of the legitimacy of pluralism not only as a background condition for political decision making but also as a feature of political outcomes because it is not based on the assumption that such outcomes must necessarily correspond to or even approximate an objective standard of normative “truth.” From the perspective of such a justification of democracy, the outcome of a majority decision is understood merely as the contingent result of the fact that opinions and preferences happen to be distributed in such a way that a majority of votes could be gathered for a specific proposal. Thus, individuals in the minority are not expected to conform their views and preferences to those of the existing majority but merely to respect the decisions in place to the extent that they have been made through legitimate means for as long as they are not able to modify them through the same procedures.
This implies a conception of the political process as an open-ended struggle between a plurality of different views and opinions that can only achieve dominance over each other contingently, revisably, and, therefore, temporarily instead of as a teleological process defined by the goal of progressively overcoming all forms of political disagreement in pursuit of a more or less mythical idea of normative truth.
In the light of the analysis presented above I conclude that the reintroduction of a normative standard of truth within the framework of democratic theory is not only unnecessary but also undesirable. For what emerges from the points just raised is that a purely procedural justification of majority rule is not only possible but also preferable to the epistemic ones considered in this article due to at least three separate reasons: first, grounding democracy on consent instead of truth makes better sense of the idea of autonomy that seems to be implicit in the notion of democracy itself; second, it avoids the risk of providing more grounds to justify a form of epistocracy; and third, it is compatible with a recognition of pluralism not just as a fact but also as a value.
In order to point to some broader implications of this conclusion but also note some of its limitations, I will end by taking up one of the issues mentioned at the start of this article to justify the salience of the debate it addresses. One of the key issues at stake in the question of whether democracy can be justified in purely procedural terms is whether democratic procedures can be considered normatively desirable even if (or when) they fail to produce normatively desirable outcomes. This is, of course, a very pressing political issue at a time when several presumptively democratic votes have produced outcomes (e.g., the so-called Brexit decision in the UK and the election of Donald Trump in the latest US presidential election) that are widely regarded as substantively catastrophic, even if procedurally unimpeachable. Independently of whether one agrees with these specific political judgments, does not the conceptual possibility they allude to provide a powerful argument against the purely procedural justification of democracy I have sought to defend in this article?
The first thing to note in this respect is that the idea that democratic procedures can produce substantively catastrophic but procedurally unimpeachable outcomes depends on the prior assumption of a substantive criterion of normative validity, the availability of which is in part precisely what is at stake in evaluating the relative merits of purely procedural and epistemic justifications of democracy. But even assuming that such a criterion is available—that is, that it is possible to evaluate political outcomes independently of the procedures through which they are obtained—it is not clear whether the recognition that democratic procedures can sometimes lead to catastrophic political outcomes is more damning for purely procedural or epistemic justifications of democracy.
For although in the case of the former this recognition may at most lead to a moral dilemma that gives reason to relativize the normative value of democratic procedures (and, therefore, potentially recommend complementing them with other kinds of procedures), from the point of view of epistemic justifications of democracy, the recognition that democratic procedures may sometimes lead to catastrophic outcomes strikes at the core of the normative justification for them and may, therefore, ultimately lead to the conclusion that they ought to be abandoned entirely. Far from underscoring the need for an epistemic justification of democracy, the recognition that democracy may sometimes lead to normatively undesirable outcomes therefore confirms the need for a purely procedural justification if we want to hold on to our normative commitment to democracy even in such hard cases.
The above, however, also points to a limitation of the argument I have advanced inasmuch as it lays bare an issue I have not tackled in this article: that of the relative priority—or hierarchy—between purely procedural and substantive values, if we assume both exist, in cases of conflict between them. All I claim to have shown is that a purely procedural justification of the democratic procedure of majority rule is possible and, indeed, preferable to substantive or epistemic ones that are currently in vogue (and, furthermore, that this conclusion is not infringed—but rather strengthened—by the recognition that in some cases democratic procedures may lead to normatively undesirable outcomes). How to deal with such hard cases from the point of view of a purely procedural justification of democracy is another matter, to which I am tempted to respond by referring to a point made by Alexis de Tocqueville in his treatise Democracy in America: that although democracy may often make very “bad” decisions, it has the unique privilege of being able to “correct” its own mistakes (Tocqueville 1835: 263). Exploring the full significance of this intuition, however, is the topic for a different article.
As Melissa Schwartzberg has shown in her insightful discussion of the conceptual origins of this strand of contemporary democratic theory, the idea of advancing an epistemic justification of democracy has its roots in the attempt by certain Anglo-American political theorists writing in the last few decades of the past century to respond to a challenge posed for the classical Rousseauian justification of democracy by the “paradoxes of social choice” (Schwartzberg 2015). In a highly influential book originally published in 1982, for instance, William Riker demonstrated that, given a set of plausible assumptions concerning the nature and distribution of individual preferences, standard democratic procedures—such as, most notably, majority rule—might not yield coherent or even determinate outcomes. On this basis Riker argued that what Rousseau had originally called the “general will” might not exist and, therefore, democratic procedures are better justified on the basis of more “minimal” premises, such as the fact that they offer a mechanism to enable the peaceful removal of officials from office (Riker 1982; for other versions of the same argument see also Przeworski 1999 and Schumpeter 1942). At a conference titled Explanation and Justification in Social Theory held at the California Institute of Technology in 1985, Jules Coleman and John Ferejohn responded to Riker, arguing that the fact that democratic procedures do not necessarily yield coherent or even determinate outcomes does not necessarily show that Rousseau’s “general will” does not exist because these procedures can be understood as “fallible means” for “discovering” the general will that is “out there” (Coleman and Ferejohn 1986). These arguments were then taken up and developed further by Joshua Cohen in an article entitled “An Epistemic Justification of Democracy” (1986), which still explicitly targets Riker’s critique. Initially, therefore, epistemic theories of democracy were intended to salvage an idea of collective self-government from the thrust of the critique advanced on the basis of social choice theory. Over time, however, the fact of separating the “general will” from the procedures used to “discover” it led it to be increasingly perceived as an objective standard of political truth or correctness that need not necessarily depend on what citizens actually want or how they behave. The focus therefore shifted from the idea of collective self-government to that of obtaining a specific set of (“good”) results. This can already be seen in Robert Goodin and Christian List’s (2001) discussion of the merits of the Condorcet Jury Theorem as a possible conceptual foundation for the epistemic value of democratic procedures. It becomes explicit, however, in the latest generation of epistemic theorists of democracy, who—as we have seen—explicitly claim that a purely procedural justification of majority rule would be “insufficient” (Anderson 2006; Estlund 2008; Landemore 2013). It is therefore primarily on this latest generation of epistemic theories of democracy that I will be focusing in this article.
Before moving on to provide such a justification, it may be worthwhile to first clear one more bit of conceptual ground because there is admittedly one exception to the general neglect of the notion of consent displayed by epistemic theorists of democracy overall. In his book Democratic Authority Estlund devotes a whole chapter to a discussion of this notion, which is worth considering in some detail here because this will also serve to further clarify the terms of the demonstration to be provided. Basically, Estlund attempts to show that the notion of consent cannot be treated as a purely procedural value because the idea that consent establishes the legitimacy of collective decisions depends on a prior set of assumptions concerning the substantive conformity between the outcome of these decisions and the “will” of the subjects for whom they are supposed to be binding. For example, Estlund contends that if a person is commanded to do something under threat of violence or death, he or she may formally consent. However, it is plausible to assume that this consent isn’t “genuine” from a moral point of view and, therefore, that it cannot establish the legitimacy of the outcome (Estlund 2008: 119–122). From this Estlund therefore deduces that attempts to ground legitimacy on consent cannot be considered purely procedural because they turn out to rely on a set of substantive moral premises after all that establish what can be considered genuine consent in the first place. My contention, however, is that this is not necessarily the case because the notion of consent can be plausibly defined in a way that accounts for the problematic cases Estlund points out without needing to rely on the further assumption that it is supposed to be a marker for a substantive conformity between the outcomes and the real will or preferences of the subjects to whom they apply. For example, it seems perfectly legitimate to propose a definition of consent according to which such a thing exists only if consent is formally expressed under conditions that do not involve coercion, which at the minimum implies that the consenting subject must have a meaningful set of noncatastrophic options, of which he or she is adequately informed. Such a definition of consent would enable us to explain why in some cases, even though a subject may appear to express consent, the relevant moral consequences are not assumed to follow because it provides grounds for suggesting that in fact the conditions for establishing whether consent exists have not been met. Indeed, one could tinker with the procedural definition of consent even further in order to better capture the moral intuition that lies behind it, but the main point would still stand because this does not imply that consent must necessarily rely on a set of independent moral standards in order to function as a ground for legitimacy; rather, these moral standards can be assumed to constitute the notion of consent, from which it follows that a justification of political legitimacy in terms of consent, thus defined, can be purely procedural in the sense in which this notion has been defined here: it is only a matter of specifying the concrete procedure through which the existence of consent is to be verified.
One important aspect of this passage may deserve further comment in order to fully bring out its relevance to the argument I am making. The fact that Kelsen formulates the point in terms of a requirement of correspondence between the content of the “social order” and the “will” of the individuals for whom it is binding may seem to lapse back into a substantive conception of the grounds for the legitimacy of collective decisions in the light of Estlund’s critique of the procedural nature of consent discussed above. That, however, is not necessarily the case because Kelsen’s argument can be easily reformulated in terms of the purely procedural conception of consent I have proposed without losing any of its validity. This simply requires introducing the idea that nondecisions are to be treated as decisions too as a substitute for what Kelsen says concerns the conditions required for changing decisions that have already been taken. The key point thereby becomes that majority rule maximizes the number of individuals who have actually consented to the totality of decisions (and nondecisions) in force through the mechanism of voting and, therefore, independently of whether these outcomes actually correspond to the will of the individuals for whom they are binding. Whether that is the case or not is indeed a substantive question, but it is important to note that this need not be appealed to in order to legitimate the outcome because assuming that the right procedural conditions hold for the way the vote is conducted (i.e., that nobody is coerced and that individuals have meaningful options of which they are adequately informed), this can be interpreted as implying that the collectivity as a whole has thereby consented to the decision in question.
AndersonElizabeth. 2009. “Democracy: Instrumental vs Non-Instrumental Value.” In Contemporary Debates in Political Philosophy ed. Thomas Christiano and John Christman213–228. Oxford: Wiley-Blackwell.
EstlundDavid. 1993. “Making Truth Safe for Democracy.” Pp. 71–100 in The Idea of DemocracyDavid CoppJean Hampton and John Roemer ed. Cambridge: Cambridge University Press.
EstlundDavid. 1997. “Beyond Fairness and Deliberation: The Epistemic Dimension of Democratic Authority.” In Deliberative Democracy ed. James Bohman and William Rehg173–204. Cambridge: MIT Press.
GoodinRobert and Christian List. 2001. “Epistemic Democracy: Generalizing the Condorcet Jury Theorem.” Journal of Political Philosophy 3 (1): 277–306.
JorkeDirk. 2010. “The Epistemic Turn of Critical Theory: Implications for Deliberative Politics and Policy-Making.” Critical Policy Studies 3 (4): 440–446.
LandemoreHélène. Forthcoming. “Beyond the Fact of Disagreement? The Epistemic Turn in Deliberative Democracy.” Journal of Social Epistemology.
LelartMichel. 2006. “Pourquoi la France a-t-elle rejeté la Constitution européenne?” Working article no. 38 Laboratoire d’Economie d’Orléanshttp://www.univorleans.fr/leo/images/espace_ perso/lelart/WP_ 2569.pdf.
OberJosiah. 2012. “Epistemic Democracy in Classical Athens: Sophistication, Diversity and Innovation.” Pp. 118–147 In Collective Wisdom: Principles and Genealogy ed. Jon Elster and Hélène Landemore. Cambridge: Cambridge University Press.
PalumboAntonino. 2012. “Epistemic Turn or Democratic U-Turn? On the Tension between Philosophical Reasoning and Political Action in Deliberative Democracy.” Unpublished manuscript available at http://www.academia.edu/1371787/Epistemic_Turn.
PrzeworskiAdam. 1999. “Minimalist Conception of Democracy: A Defense.” Pp. 23–55 in Democracy’s Value ed. Ian Shapiro and Casiano Hacker-Cordon. Cambridge: Cambridge University Press.
StokesSusan. 2011. “A Rational Theory of Epistemic Democracy.”Particle presented at the Yale-Oslo Conference on Epistemic Democracy in PracticeYale UniversityOctober 20–22.
UrbinatiNadia and Carlo Invernizzi-Accetti. 2013. “Introduction.” In The Essence and Value of Democracy ed. Hans Kelsen1–24. Lanham, MD: Rowman and Littlefield.
VermeuleAdrian. 2012. “Collective Wisdom and Institutional Design.” Pp. 338–367 In Collective Wisdom: Principles and Genealogy ed. Jon Elster and Hélène LandemoreCambridge: Cambridge University Press.