Economic theories of democratic legitimacy and the normative role of an ideal consensus

1. Introduction

At least since the work of Joseph Schumpeter, there has been an experiment among some political philosophers, legal theorists, and economists to determine the extent to which the adoption of democracy can be justified in economic terms (for example, according to its efficiency in generating widely accepted outcomes). ¹ An economic theory of democracy, Jon Elster (1999: 26) writes, ‘rests on the idea that the form [of democracy] should be like the market in its purpose as well as in its mode of functioning’. ² Markets efficiently aggregate individual decisions with generally beneficial results and, some believe, correct themselves when things go wrong. Why not leverage these advantages on behalf of democratic politics? By comparison, alternate normative models like deliberative democracy seem onerous and pernicious. Good deliberation requires specialized and not widely distributed abilities and circumstances. Facts about persons (for example, political ignorance and irrationality) and society (for example, moral and religious pluralism) undermine the conditions for its possibility. ³ Markets, on the other hand, can benefit the most foolhardy among us under a wide range of social conditions.

Though the experiment has found intelligent contemporary articulation in the works of Judge Richard Posner (2004a, 2004b), Ilya Somin (1998, 2004), and Bryan Caplan, ⁴ it remains unclear how successful it has been or how widely its principles are accepted. Yet, according to Judge Posner, the economic theory also promises to produce a criterion of democratic legitimacy – hence, a standard for judgments about when outcomes are legitimate and when they are not. ⁵ Since the economic model makes no appeal to standards independent of citizens’ voluntary acceptance, we may call it a ‘minimalist’ theory of democratic legitimacy and formulate its view of legitimacy in the following way. ⁶

Minimalist Legitimacy Principle (MLP): Political institutions, practices, and outcomes are legitimate if people widely and voluntarily accept social undertakings rather than submit to coerced obedience. ⁷

This article argues, first, that as a theory of legitimacy minimalism is self-defeating. Second, it suggests that the minimalist view of normative political theories and the constraints they must observe is misguided. There are deliberative theories of democratic legitimacy that both appeal to ideal conditions as a proposed basis for legitimacy and could count as practically possible given persons as they are in a pluralist democratic society. In Sections 2 and 3, I define the problem of democratic legitimacy. I then show both why minimalism cannot avoid addressing this problem and why its consequent response is self-defeating. The self-defeat of minimalist legitimacy does not alone show how a deliberative ideal of democracy could accommodate the relevant facts about persons and society. In particular, it does not show how deliberative democracy handles the plurality of ‘ultimate values’ that create deep and irreconcilable disputes in democratic societies. Sections 4, 5, and 6 collectively show that political reasoning grounded in a normative consensus does not require an appeal to ‘ultimate values’ in the first place. Deliberative democracy is a practicable possibility even if there is a plurality of ultimate values. Section 7 then offers some remarks about norms for normative theories of democratic legitimacy and judgments about permissible coercion. Methodologically, I suggest, we are not bound to begin with the aforementioned facts about persons and society, even if a theory should show proper respect for them in the end.

2. The hard problem of democracy legitimacy

The following entailment of the idea of political legitimacy seems fairly uncontroversial.

Political Legitimacy Entailment (PLE): If X is a legitimate political authority, political legitimacy entails that X has a moral right to coerce Y in enforcing whatever duties it imposes, for example obedience to law Z, even if Y does not believe Z is good or correct, and in some cases actively resists it.

However, even if we have settled upon a way of specifying the entities in question, a puzzle (often discussed as the problem of the minority voter ¹⁰ ) arises for democracy. Democracy could refer in a very broad way to the substantive values of freedom and equality or it could be considered more narrowly in terms of specific rules and procedures (for example, rule of the majority). Indeed, many problems in the theory of democracy seem to arise from the effort to reconcile the demands of the broader, substantive account with facts about the narrow, procedural one – to show, for instance, how specific democratic procedures manifest proper regard for citizens by being equality-recognizing or freedom-preserving. Certainly, when Rousseau asks how it is possible for one to be free and to obey a law not chosen by one’s self, he is asking how it is possible for the freedom of the minority voter (Y) (the broad, substantive view) to be preserved under conditions in which the majority (X) determine the legitimate outcome (the narrow, procedural view). ¹¹ The hard problem for democratic legitimacy, then, is not precisely how to cause the voluntary acceptance of citizens, particularly those in the minority. Rather, the hard problem is exemplified in explaining how to justify the outcome of a democratic procedure to the minority when the minority regards it as morally bad or incorrect. To the extent that we take the autonomy of citizens seriously, we also must understand the hard problem to be a problem of justification. Citizens (qualified ones at any rate) are owed reasons that they can all accept. Only then can the outcome be regarded as legitimate.

From this perspective, items 1, 2, and 3 above do not appear to suggest workable models of democratic legitimacy. Alternative 1 seems inadequate for democratic legitimacy since it entails that the majority has the right to coerce the minority because the majority is correct. That the majority qua majority is correct seems intuitively implausible. Moreover, it would be hard to reconcile this view with the freedom of the minority. In item 2, the problem is much the same. Indeed, it suggests that in cases in which the majority is incorrect, the minority cannot be coerced. Such circumstances invite instability. In item 3, the correctness of outcomes according to some standard or other is commonly in dispute. The mere fact it is correct, however, is not decisive with respect to its legitimacy. Some articulation of item 4, on the other hand, suggests a view that, in general, tracks with liberal theories of democracy. In this case, the procedure itself (given its recognition of the freedom and equality of citizens) is at least in part responsible for the legitimacy of the outcome.

Each citizen, then, has a reason consistent with her character as a democratic citizen to accept an outcome even though she believes it to be incorrect. In this light, the PLE can be reformulated more specifically as a Democratic Legitimacy Entailment (DLE).

Democratic Legitimacy Entailment (DLE): Democratic legitimacy entails the right of the majority to coerce the minority given proper procedural requirements and equal recognition of the substantive rights and liberties of citizens or persons.

It is worth noting that minimalism also fits, to some extent, the pattern of item 4 above. However, minimalism tends to understand the limits on, say, state power in terms of organizationally and constitutionally specified powers and roles; and the limits of democracy as established by the restrictive facts about persons and society. Indeed, the MLP suggests that minimalism rejects the formulation of the hard problem of legitimacy as a problem for actual democracies at all because it rejects the notion that persons or citizens are ‘rational’, ‘free’, or ‘equal’ as illusory. ¹³ In this way, the MLP does not understand citizens or persons as being owed reasons for accepting outcomes even when they are incorrect, though it may be useful to give reasons if it creates or sustains voluntary acceptance. ¹⁴ Indeed, minimalism does not regard an ideal consensus as available to us as a political norm. Either there are no reasons that all citizens accept or no objects upon which they will all converge. Legitimacy is brought about in the degree that citizens can be caused voluntarily to accept political authority, not the degree to which the use of power has been publicly justified. Causing citizens voluntarily to accept political authority, however, is a use of power. Thus, the use of power in minimalism seems to be justified only by itself.

Yet, minimalism might respond that even if a citizen disputes the correctness of an outcome, she may still voluntarily accept it, and only her voluntary acceptance (without further specification) is relevant to its legitimacy. The status of the outcome independent of her acceptance is not relevant to legitimacy. Voluntary acceptance (hence legitimacy) can be produced independent of whether an outcome is good or correct according to an independent standard. Provided it is not generally the result of coercion, the form of acceptance does not much matter. In this way, minimalism requires no appeal to reasons that all citizens can accept as a basis for legitimacy – hence for the permissibility of coercion. So, a political outcome or authority that is not widely accepted would not be legitimate no matter its moral or epistemic credentials.

Minimalism suggests that democratic outcomes are legitimate (or not) independent of what one might say about their correspondence with values such as freedom or equality. Indeed, the hard problem of democratic legitimacy seems to arise only if we take seriously the idea that persons have a unique moral status either as persons or as democratic citizens, and that this status is the ultimate source of democratic value. Thus, minimalism attributes little importance to the idea that persons or citizens have this moral status in democracy or that the moral and epistemic value of outcomes and authority is not a primary consideration in problems of political legitimacy. Justified coercion requires only the discovery of the efficient means by which to produce voluntary acceptance – a task for which the natural and human sciences, or perhaps the rhetorical arts, are equipped, but for which philosophy generally speaking is not.

The MLP, while compelling on these counts, fails. In the following sections, I will demonstrate that it does and explain why. I begin by explaining the difference between a justification of democracy and democratic legitimacy. ¹⁵

3. The justification of democracy versus democratic legitimacy

Let us assume there is a justification for adopting a political procedure if there is a reason to do so. The adoption of some procedure, then, is justified (or at least justifiable) when the reasons for doing so are good ones. So if a group of citizens is justified in adopting democratic procedures, their reasons for doing so are good ones. ¹⁶ Evidently, there could be many good reasons to adopt democracy, so a variety of grounds upon which it could be justified. A moral justification holds that democracy ought to be adopted because only democratic procedures are just or because humans (qua humans) have a basic right to it. ¹⁷ An epistemic justification, on the other hand, argues on behalf of the merits of democracy in making wise decisions according to some non-procedural standard. ¹⁸

As a justification of democracy, minimalism holds that democracy is an efficient way to install acceptable candidates into specified political offices regardless of the moral or epistemic quality of its results. Indeed, it emphasizes the pointlessness of moral and epistemic justifications of democracy given the deep and presumably irreconcilable moral or religious disagreements among citizens and facts about the political knowledge or rationality of voters. So, minimalism recommends restricting the use of democracy to a role as an electoral procedure, not one that governs complex political affairs for which the knowledge and experience of a specialist is irreplaceable. Democracy practiced minimally is ‘rule by officials who are … chosen by the people and who if they don’t perform to expectations are fired by the people at the end of a short fixed or limited term of office’ (Posner, 2003: 183). As a justification of democracy, then, minimalism focuses on electoral aspects of actual political arrangements rather than on the making of law or policy, which it believes should not be turned over to democracy (in any direct way at least). Moreover, as a normative and interpretive model, it holds that we should view these aspects in market terms. Yet, the justification of a political method vis a vis good reasons to adopt it does not show why coercion by a particular state or its agents is legitimate.

The minimalist view of democratic legitimacy, on the other hand, holds that the voluntary acceptance of outcomes or institutions by citizens is sufficient for their legitimacy (Posner, 2003: 207). ¹⁹ So for the minimalist, coercion in enforcing outcomes, even if they are only electoral outcomes, is permissible on that basis, whether or not the procedure expresses justice, consistently produces good outcomes according to a procedure-independent standard, or reflects any kind of unanimous acceptance among citizens.

The hard problem of democratic legitimacy has to do with the conditions under which coercion is permissible though citizens dispute the correctness of outcomes or even the presence of political authority. While it is true, of course, that citizens might accept outcomes despite the fact that they are good or correct, minimalism establishes no moral requirement that they do so. It only suggests the means by which they will tend to do so. But if this is the case, citizens might as easily have good reasons not to accept outcomes or to acquiesce. Consequently, minimalism does not really solve the hard problem of legitimacy. It simply eliminates the substantive (and moral) concerns as being a concern. There is no dilemma if one horn of it has been cut off.

In this case, however, it is not clear how minimalism could establish the PLE. It could not appeal to a broad moral or epistemic justification. Moreover, the attitude that would permit coercion (namely, voluntary acceptance) is the same attitude that obviates its necessity. Why coerce those who already voluntarily accept the outcome? Consequently, as a criterion of democratic legitimacy, the MLP is self-defeating. It counts as cases of permissible coercion only those cases for which coercion would not generally be required. Thus, minimalism sidesteps the hard problem of democratic legitimacy. Moreover, it does so illegitimately. This is because the problem of legitimacy is about the permissibility of coercion under conditions of reasonable disagreement and even reasonable nonacceptance, not under conditions of widespread voluntary acceptance. At best, then, minimalism counts as a theory showing how to minimize coercion in the service of political stability – not surprisingly, by creating the conditions for widespread, voluntary acceptance. But it does not show what would justify coercion in cases of deep moral disagreement. It is not, in this sense, a theory of legitimacy at all.

The failure of the minimalist criterion of legitimacy seems like a reason to provisionally reinstate moral and epistemic constraints on voluntary acceptance as part of a correct account of democratic legitimacy. But facts about political ignorance and moral and religious pluralism are compelling as limitations on philosophical theories of politics. ²⁰ It is hard to imagine, for example, how there could be a consensus upon ‘ultimate values’ given these facts. If there really is disagreement about ‘ultimate values’, what is the object of deliberation? Why should citizens deliberate? Even if minimalism is self-defeating as a proposed criterion of democratic legitimacy, alternative theories should be able to defend their accounts of the norms for the justification of coercion on other grounds. In the following sections, I examine the possibility of a normative consensus as a basis of deliberative democratic legitimacy given the compelling facts about persons and the moral and religious pluralism that prevails in actual modern democracies.

4. Deliberative theories of democracy in a minimalist world

No one, perhaps, would dispute the claim that political stability arises if citizens in sufficient numbers voluntarily accept political outcomes. At any rate, I will assume it is true. As discussed, political theorists divide over whether voluntary acceptance is also sufficient for political legitimacy. For deliberative democrats, in general, it is not. This is because voluntary acceptance alone does not satisfy other moral or epistemic norms; or at least, the idea of ‘voluntariness’ can seem objectionably ambiguous unless clarified along these lines. So, deliberative theories of democracy constitute one of the alternatives to which the theorist might turn once convinced that the MLP fails to explain democratic legitimacy.

In contrast to minimalism, the following items typify deliberative theories of democratic legitimacy:

insistence that legitimacy requires at least the opportunity for citizens generally to participate in public affairs,

One way to regulate such disagreements would be by appealing to a consensus on ‘ultimate values’. Minimalism suggests that appealing to a consensus of this sort is objectionable for at least three reasons:

Citizens, in general, are gravely limited in moral and cognitive capabilities (Posner, 2003: 113).

The minimalist argument can be framed the following way:

For any deliberative theory of democracy, legitimacy requires ‘rational agreement about ultimate values’.

5. Objects of consensus and their relations

The first task will be to elaborate various possible objects of consensus and their relations. We can assume that a failure to achieve consensus in each case represents a disagreement of one sort or another. But we need not assume that ‘rational consensus about ultimate values’ refers to just any sort of political agreement. Rather, this sort of agreement is about the standards that would adjudicate between positions in various first-order disputes. Call it ‘second-order agreement’. So, disagreements about this standard would be second-order disagreements. Here are several alternative objects of unanimous rational consent:

Consensus 1 (C1): A consensus that there is a single good that is the object of politics. ²⁶

Minimalism reasonably denies that C3 ever actually occurs in democracy about complex matters of law or policy. Consequent disputes likely have underlying sources in moral and religious disagreement of the type endemic to pluralist, constitutional democracy. Because of this conflict, it is not hard to see why traditional democratic theory might appeal to a consensus of types C1 or C2 to justify or regulate it. If there is a rational consensus about ultimate values, rational agreement and rational disagreement about political outcomes seems possible. So, for instance, if there is a consensus of type C2 specifying freedom as the good of politics and unanimous rational consent about ultimate values is required for legitimacy, legitimate political outcomes will be those that preserve or promote freedom. Likewise, if that good is equality, legitimate outcomes will in some way be equality-recognizing. Disagreements are about whether outcomes appropriately manifest freedom, equality, or whatever ultimate value is in question, not whether a given value is ultimately valuable. Among the possible kinds of consensus noted, then, C1 and C2 seem to be the most obvious types explaining the meaning of the phrase ‘rational agreement over ultimate values’. So, minimalism claims that either C1 or C2, or both, are not possible given the moral and epistemic capacity of citizens.

Thus, minimalism does not simply deny there is a consensus of type C3. It denies that there could be a rational consensus of types C1 or C2 – second-order rational agreement and disagreement. Consequently, political disputes could not be precisely rational so long as they appeal to these possible types of consensus. These disputes are particularly menacing for large, deeply pluralistic democracies in which disagreement about ultimate values may be widespread and entrenched. Moreover, as Posner suggests, citizens cannot be prevented from drawing their political conclusions from such values. What minimalism clearly rejects is the idea that political conflict involving such values can be regulated by appealing to a consensus of types C1 or C2 – to what it regards as ‘ultimate values’. ²⁷ So in denying the possibility of C1 and C2, minimalism denies there could be a normative consensus fashioned from a reason-independent moral or religious standard. Any effort to formulate one represents an idealization of an undesirable kind – an illusion of sorts.

Consequently, the classical project of justifying democratic authority and rationalizing political disagreement may seem to be dead in the water. If there is no way to reconcile or to commensurate value to constitute a class of ‘ultimate value’, the more specific disagreements of democratic citizens cannot be resolved through the ongoing deliberation of those same citizens. Minimalism does not propose to reinvigorate this type of theorizing. Rather, it proposes to abandon it altogether. Instead, theorists should aspire to specify ways of containing unavoidable disagreements that result from disputes arising from the failure of C3. Since acceptance of complex matters of law and policy cannot be induced by rational means, given disagreement about ultimate values and the cognitive and moral limitations that may lead to it, democracy should be applied only as an electoral procedure. Disputes that cannot be adjudicated, and hence threaten legitimacy and stability, will at least be less frequent if they are not given the opportunity to occur.

However, this does not yet say how to treat C4 as a possible type of consensus. In this case, the ‘object’ of the rational consensus is ‘constitutional essentials’, not ultimate values unavailable to the reason of all citizens. We might imagine that among these essentials is a democratic procedure as the means by which political outcomes (for example, laws) will be determined. A consensus of this sort is validated not simply because citizens voluntarily accept a democratic procedure or its outcomes.

According to John Rawls, it is valid as long as the procedure is, among other things, consistent with ‘common human reason’. Rawls (2005: 137) writes:

our exercise of political power is fully proper only when it is exercised in accordance with a constitution the essentials of which all citizens as free and equal may reasonably be expected to endorse in the light of principles and ideals acceptable to their common human reason. ²⁸

The entailments of C4 in relation to other forms of consensus can now be explained. Since the ‘object’ of C4 is a constitutional essential (in this case, a particular procedure for decision-making), it would be more difficult to regard it as a reason-independent ‘ultimate value’. Democratic procedures do not constitute the end of politics so much as they offer means to these ends. ³⁰ So, C4 does not entail C1. Additionally, since a plurality of values (for example, freedom and equality) may be consistent with the reason of all citizens, it does not require a consensus of type C2 – a consensus about the substance of the good. Finally, C4 is not a consensus about the specific outcomes of a democratic procedure. So, it does not entail C3. In these respects, C4 is autonomous with respect to other types of consensus discussed. It can exist whether they do or not. Thus, even in minimalist terms, the sort of consensus specified in C4 would be practically possible. It would not, at least, count as a consensus upon ‘ultimate values’.

Of course, even if there is a constitutional consensus upon democratic procedures, disputes about the correctness of outcomes are liable to ensue. C4 is not productive of C3, though it may condition the possibility of rational disagreement. In the following section, then, I will examine whether or not this sort of consensus could provide a workable normative basis for legitimate outcomes that are both deliberative and democratic given the fact of moral disagreement. The point is not to defend deliberative democracy comprehensively or to lay bare its entire normative structure. The point is to see how a kind of idealized unanimous consent that does not appeal to reason-independent ultimate values could legitimately regulate deliberation and its results. In the course of this discussion, at least a couple of important questions will have to be addressed, however briefly. First, in what way are values such as ‘freedom’ and ‘equality’ determinant of legitimate democratic outcomes? Second, why (and how) could other values (for example, religious ones such as ‘salvation’) be legitimately excluded from democratic deliberation?

6. Two models of practical reason

In general, the theory of practical reason concerns determinations about what should be done along with general problems of value and what is good. Practical judgments, then, are claims about what should be done and what is good or of value. A practical judgment could serve in an argument as practical evidence or as a practical conclusion. ³¹

Let PJ_e stand for a practical judgment that is evidence or a source of evidence in a political argument. In addition, let PJ_c stand for a practical judgment that is the conclusion of a political argument. Some believe that the objective evaluation of practical judgments as the conclusions of practical arguments depends on there being a known good accessible by perception or intuition antecedent to and independent of that judgment. Since the goodness of any particular item, X, depends on whether or not it instantiates the independently known good, the correctness of the judgment about particular items would be determined in part by a reason-independent standard. So, ‘heroin is good’ (PJ_c) (for example, instrumentally) if it is true that (a) pleasure is good (PJ_e) and (b) heroin produces pleasure. ³²

In this case, the judgment ‘pleasure is good’ serves as practical evidence in the argument. Thus, some of the evidence for the conclusion appeals to a reason-independent value (that is, pleasure). What reasons one has to use heroin, then, are determined in part by the judgment that pleasure is good. So the fact that heroin is good is a reason (though not in the calculus of pleasure necessarily decisive) to choose it. The determination of what is good ultimately in cases like this does not depend upon any property or quality of reason, even if the determination of particular instances of goodness does. Thus, achieving objectively valid results in practical reasoning, on this view, requires a reason-independent ultimate standard as a source of practical evidence. Consider this view to outline a nonconstructivist theory of practical reason.

It is not difficult to see what this view means if we conceive of the vital democratic values of ‘freedom’ and ‘equality’ as reason-independent and ‘ultimate’. In this case, the judgment that freedom is the ultimate political good operates as practical evidence. A practical argument could be constructed in this way by Citizen A (a libertarian).

Freedom is the end of politics (because it is of ultimate value).

These conclusions appear to be correct as long as we are viewing matters through the lens of a nonconstructivist account of practical reason. Indeed, matters are complicated if citizens introduce values (for example, religious values such as ‘salvation’) into practical evidence in the course of making public political arguments. According to Posner, nothing prevents them from doing so. Consequently, under conditions of nonrational disagreement, it would be pragmatically sensible to restrict the political role of democracy so that nonrational moral disagreement and the political disputes that arise from it may at least be contained.

Posner (and minimalists in general, I assume) means that citizens derive particular practical judgments (PJ_c) about what should be done politically from different and incompatible premises (PJ_e) about reason-independent ultimate values. Yet, not all accounts of practical reason are in accord with this view. Consequently, we may suspect that some deliberative theories of democracy will be insulated from the minimalist criticism about the ‘rational consensus about ultimate values’ given their view of practical reason. This can be made clear if we show how rational disagreement is possible under the same conditions minimalism claims make rational agreement impossible – disagreement about ultimate values given facts about persons and society. Consider a view of practical reason supporting this possibility to be ‘constructivist’.

The view of practical reason set out in political liberalism appears to be constructivist in these ways. Rawls writes:

political constructivism … may see correct judgments as reasonable: that is, as supported by the preponderance of reasons specified by the principles of right and justice issuing from a procedure that correctly formulates the principles of practical reason in union with appropriate conceptions of society and person. (2005: 111, emphasis added) ³³

The constructivist model makes no appeal to a consensus about reason-independent ultimate values as a condition for democratic legitimacy. Consequently, the determination of ‘correct’ outcomes is not modeled precisely as it might be in rational choice theory. The legitimacy of outcomes (qua reasonable) does not depend upon the possibility of being able to commensurate values such as freedom, equality, or whatever else is introduced as ends of democracy. So, the determination of ‘correct’ outcomes does not require the identification of a final end that would serve as evidence in the premise of a practical argument. Indeed, the effort rationally to commensurate values would not make sense in this case. Since freedom and equality are both regarded as attributes of democratic citizens, to assign ordinal or cardinal priority to freedom or equality would make freedom and equality unworkable as the basis of legitimacy – pitting each citizen against herself. Practical arguments that treat freedom and equality as abstract and prior consensual objects of politics, then, invariably fail to regard democratic citizens with the appropriate respect that is due to persons as citizens. ³⁵ Thus, salvation (among other possible ends) does not count as a political value in this way because political judgments (PJ_c) made upon its basis alone are not consistent with the freedom and equality of all citizens qualified to make such judgments.

The underpinning of a deliberative (qua constructivist) view of democracy, then, is not a particular metaphysics of the person. Instead, it is an appeal to the status of persons as democratic citizens as free and equal. The problem of a theory of political justice is the problem of identifying a conception of justice capable of reconciling the freedom of one citizen with the freedom of others. In any case, the basis of legitimacy is not an ‘ultimate value’ qua reason-independent standard. Freedom and equality do not function in this conception as sources of ultimate value; thus, they could not count in practical arguments as a PJ_e. The freedom and equality of democratic citizens function in political constructivism as epistemic constraints – as constraints upon valid practical evidence, not as sources of practical evidence themselves. Yet, this does not necessarily mean that constructivism regards practical judgments about ultimate values such as salvation as false. The constraint means either that certain claims cannot be entered into evidence for the sake of public argument or, to the extent they can be entered, that arguments that depend upon them cannot make legitimate claims to the use of coercive public power. Either way, the legitimacy of deliberative outcomes would not depend upon unanimous rational consensus focused upon ‘ultimate values’.

Deliberative democracy conceived in this way, then, does not require citizens to possess rarified moral characters or epistemic talents. It does not wait for a political expert. It allows for a plurality of value, but it does not require unanimous rational consensus about ‘ultimate values’ in order legitimately to determine outcomes. Indeed, practical reason retains its flexibility because reasonable judgments do not wait upon a final determination of the correct order of reason-independent value. It says more modestly that legitimate political outcomes are those that can be justified to all qualified citizens (for example, as rational or as reasonable) even when they dispute the correctness of outcomes according to an independent moral or religious standard. Moral disagreements and political disputes are not necessarily irrational. This is because they do not necessarily appeal to an unavailable ‘rational consensus about ultimate values’ in the first place. They are the products of ‘common human reason’ exercised under various constraints.

Consequently, the minimalist theory of democracy as a sort of economic theory cannot rule out the possibility of a unanimous rational consensus about constitutional essentials (C4) based only upon its skepticism about the possibility of C1, C2, or C3. We can avoid the results of irrational first-order disputes in regarding the freedom and equality of citizens as basic to democracy. ³⁶ Though this standard is idealized in certain respects, it is not an ideal that holds out for the possibility of a rational commensuration of value under conditions of reasonable pluralism. With respect to democratic deliberation, it functions more modestly as a constraint upon practical evidence (that is, an epistemic constraint). This should be enough to show at least that there is a normative consensus capable of regulating deliberative democratic practices without resulting necessarily in irrational and unresolvable political disputation.

7. Normative implications for normative theories of democratic legitimacy

One lesson that can be drawn from the foregoing discussion is this: the kinds of facts about human nature and society introduced as restrictions on the possibility of an ideal normative theory are not themselves sufficient to show that deliberative theories of legitimacy that appeal to an ideal normative consensus are incoherent or practically impossible. It depends upon the nature and function of the consensus. In this section, I will expand upon this point in order to consider norms for normative theory itself. Because minimalism is motivated in part by a concern with the supposedly pernicious effects of ideal theories on actual democratic practices, the particular interest is the way in which the relation between normative theories and actual practices should be understood.

Minimalism does not simply reflect a standard liberal skepticism about democracy qua populism. ³⁷ It is more broadly skeptical of a number of theoretical enterprises (both historical and contemporary) and their usefulness to actual, modern democratic societies. In this respect, minimalism may be understood as a type of pragmatist response to idealist theoretical agendas. It regards itself as superior to deliberative theories of democracy, for instance, because it does not ‘regard democracy as a creation of political theory and so as an apt candidate for improvement by it’ (Posner, 2003: 144). This outlook suggests that if actual democracies were to be improved, it would not be by philosophical theories of democracy, but perhaps by the panoply of empirical disciplines along with effective rhetorical practices. So, the constraints imposed by the supposed realities about persons and democratic societies represent constraints upon theories and the degree or sort of idealization in which they can sensibly indulge.

This view indicates a couple of assumptions about the relation of philosophical theories (understood here as ideal theories) to actual practices (in this case, democratic ones).

Assumption 1 (a₁). The object of (unconstrained) ideal theory is a mere creation of that theory.

In general, a normative theory aims to justify the claim that, for example, a practice or an agent should be or should function in a certain way, whatever it actually does. If the claim were justified, it would be possible to understand when, for example, a practice has broken down, to explain why this is so, and perhaps to improve it. There is little disagreement that political agents should be legitimate even if there is disagreement about what makes them so. A legitimacy claim, in general, is a claim that authority X has a right to employ power coercively when it does so. So, normative theories of democratic legitimacy aim to identify for a democracy those conditions under which the use of coercive power would be permissible – if there are any and whether they are actually observed or not. Coercion is justified if the conditions identified are satisfied, provided those conditions are correct and sufficient.

The minimalist objection to deliberative norms of democratic legitimacy is centrally that democracy conceived in deliberative terms would not be a practical possibility (not one at least that counts as an improvement of democratic practices) given facts about persons and society. ³⁸ There could be no valid normative consensus since a consensus on ultimate values is practically impossible in democracy. The minimalist ideal is not one that depends upon the widespread distribution of whatever capacities or objects are required in order to deliberate well. So it is developed within a framework that makes no such demands. What makes a theory practically possible for the minimalist, then, depends upon whether or not it is developed within this framework – presumably one with a proper regard for human and social limitations. However, as discussed earlier, deliberative democracy asserts more robust moral and epistemic principles of legitimacy, and it too seemed a practical possibility under the very same proposed limitations.

As discussed, the criterion of legitimacy specified in minimalism does not actually show us what the conditions for the possibility of justified coercion are, or even that there are not any. In this respect, it does not help us to understand how coercive practices in democracy ought to be understood, even if it describes how they often are. ³⁹ So, it does not enable the possibility of making good judgments about when coercion is permissible and when it is not in particular cases.

This result suggests at least a modest constraint on normative theories of democratic practices. They ought to aim first at improving our understanding of how, for example, a practice ought to be, not to improve it directly. How a practice might be improved is an entirely different matter. No doubt, it could include all kinds of empirical (and nonidealized) inputs about the thing in question. This suggests that the minimalist criticism of the idealistic elements of deliberative democracy is misguided. The fact that a deliberative ideal of legitimacy idealizes its agents to some extent is not (at least within certain limits) a fault of the theory. This is shown by the fact that an ideal deliberative theory, understood simply as the determination of its basic principles, can accommodate the limiting facts about persons and society. Nor is it a fault that its idea of legitimacy would not directly improve democratic practices. In this light, how would an ideal condition once determined serve as a norm for actual practices?

Consider the appeal to ideal conditions in another, perhaps more familiar, circumstance: the determination of the standards of evidence for jury trials. Eyewitness testimony is often, and sometimes controversially, used as evidence in making a case. However, the ideal conditions (given circumstances) under which testimony could carry considerable weight in establishing the facts of a case are fairly well understood – unobstructed perceptions, no deviant or self-interested motivations, and the like. Moreover, the ideal conditions in this case reveal a specific regulative function. The ideal enables us to discount the epistemic value of testimony in specific cases. For instance, the value of testimony is discounted if the witness reveals a pattern of lying, has a special interest in the outcome, was inebriated at the time of the witnessed event, or witnessed it under other obscured perceptual conditions. The assigned discount value of the actual testimony is the value of the testimony given a determination of how far it deviates from the ideal.

This means that the fact there may be a set of ideal conditions under which the witness could have countenanced what happened does not mean that testimony under less than ideal conditions is not valuable, or that it is illegitimate and groundless in establishing matters of fact. The ideal condition does not offer a standard for eliminating the value of testimony that is less than ideal. It is simply a standard according to which actual conditions and results can be evaluated and appropriately discounted within a ‘total system of evidence’. So, the fact that the ideal condition (for example, an ideal consensus of citizens who are free and equal) is not achievable is not an effective argument against the ideal case as a political norm. In sum, the fact an ideal condition has not been achieved in any given practice at a given time is not a reason to refrain from formulating the ideal condition in theory or from appealing to it in practice. Whether or not the ideal condition is ever actualized it can have specific regulative functions, as in this case, regulating the evidential value of testimony. In short, there is little reason to think that an ideal normative consensus is inefficacious or useless in practice given its ideal status. Indeed, its efficacy as a norm seems to depend upon its not having been achieved.

To reject the normative role of unanimous rational consent in the theory of legitimacy is to reject the fundamental status of citizens in democracy as free and equal as having a proper role in how we understand and evaluate actual cases. While it is not clear that a political theory directly improves a practice, one theory can improve upon another theory simply by correctly identifying what the proper ideal is, or if there is one at all. To this extent, the primary function of ideal theory is to improve the understanding of the proper norms of a practice, whether or not it improves the actual practice itself. Improvement of actual practices is a more complicated matter in many respects. In the case of the minimalist theory of democratic legitimacy, the mere widespread, voluntary acceptance of citizens could not count as a norm for democratic legitimacy even if it is a norm for stability of some sort. It would not, in any case, represent an appropriate ideal for democracy.

8. Conclusion

The argument has shown not that an idea of deliberative democratic legitimacy defeats the minimalist theory of democratic legitimacy, but that the minimalist theory defeats itself. The self-defeat of minimalism does not relieve deliberative democrats from the task of justifying its claim about its ideals (and idealization) given the facts about citizens and society. But as long as it makes no appeal to ‘ultimate values’ as the basis for legitimacy in the sense attributed by minimalism, there is no reason to think it will fail to produce at least an appropriately aspirational theory of democratic legitimacy. In theorizing about democratic legitimacy, we might keep the following items in mind.

The primary point of ideal theory is to enhance the understanding of correct principles for the purposes of sound judgment, not to improve the object of the theory directly (which is a separate and complex task involving many nonideal determinations).