Thin Versus Thick Disagreement

Under td, citizens hold different and conflicting views about how entitlements to social goods should be assigned within society. For example, some believe that justice requires significant redistributive taxation, others that it forbids it; some believe that affirmative action policies are a requirement of justice, others that they are a violation of it. Despite these disagreements, there is broad consensus on what conditions would have to be satisfied for a claim about justice to be true or false (i.e., on truth conditions). For instance, all citizens agree that average utilitarianism is true, and thus that policies are just only so long as they contribute to maximizing average utility. Their disagreements rest on the empirical question of whether redistributive taxation and affirmative action promote or hinder the pursuit of the utilitarian goal.

Under these circumstances, disagreements about justice are on a par with most disagreements in the natural or social sciences. Take the case of medicine. You and I might disagree about whether Bob has a regular flu or is affected by mononucleosis, even though we both agree on what would have to be the case for either claim to be true (i.e., a particular virus would have to be present in Bob's blood). Since, however, our medical knowledge and diagnostic equipment is limited, to settle our disagreement, we are well advised to consult a doctor.Footnote 26 The relative uncontroversiality of truth conditions in the medical domain is what allows us to identify, and agree on, medical expertise. There are facts about people's health, doctors have studied them in detail, and hence they are most likely to offer accurate diagnoses.

Similarly, consider a linguistic disagreement between a well-educated native speaker of English, and a foreigner who has only just started to learn the language. There clearly are (social) facts which determine what linguistic expressions count as correct or incorrect, and we agree that they depend on common usage and convention. A good strategy to ascertain what these facts are, and to settle disagreements about them, would be to consult a dictionary or grammar book. But suppose there are none available, and the disagreement needs to be resolved quickly. In these circumstances, if we want to get to the truth, we should follow the native speaker's instinct, by virtue of her greater linguistic expertise. Having grown up in an English-speaking environment, we can trust a native speaker to have greater knowledge of the relevant facts than a foreigner.

More examples could be given, but the general point should be clear. When there is thin disagreement about justice, people disagree about the policies required by justice in particular circumstances, but agree about the conditions that must be satisfied for those policies to count as just. That is, they advance conflicting claims about justice, but agree about the truth conditions of those claims. Their disagreement may simply be traced to unclear evidence, partial information, some reasoning error or a combination of these factors. When disagreement is thin in this way – i.e., when it does not affect the truth conditions of claims about justice – we can identify experts about justice: namely, those who have greater familiarity with the relevant facts.

When disagreement is thick, substantive disagreements about justice cannot simply be traced to inconclusive evidence, ignorance or bad reasoning. Instead, disagreements rest, at least partly, on the lack of a commonly agreed account of what would make a claim about justice true or false. Citizens who thickly disagree, for instance, about whether the state should support religious institutions, disagree not only about the normative appropriateness of state subsidies for religious institutions, but also about what would make any such policy normatively appropriate (i.e., a requirement of justice) in the first place.

Some, for instance, might believe that whether a particular policy is just depends on what God himself commands. Consequently, they may also believe that religious ministers are best placed to settle such policy issues, due to their greater familiarity with the word of God.Footnote 28 Others, by contrast, might think that we ought to accept a particular policy only if doing so maximizes overall utility. On this view, decisions about policies should be taken by those who are best placed to detect their impact on overall utility. Others still may hold that whether a particular policy is just or not depends on its compatibility with principles selected in an ideal decision procedure, such as Rawls's original position. Proponents of this view would consequently regard Rawlsian political philosophers as the relevant experts in matters of justice.

When disagreements about justice are thick in this way, i.e., when they concern the truth conditions of statements about justice, the identification of experts becomes impossible. Since the nature of the facts that determine the correctness of claims about justice is disputed, different people have different understandings of expertise. For Catholic believers, priests and bishops are much more familiar with the relevant facts than philosophers; for Buddhist believers, monks are probably the experts, and so forth.Footnote 29

In short, under circumstances of thick disagreement about justice there is no uncontroversial account of the truth conditions of justice claims, and hence no generally acceptable view of expertise.

Reasonable versus Unreasonable Disagreement

Judgements about reasonableness are, to a good extent, normative in kind, and therefore subject to controversy. That is, whether a claim is reasonable or not – i.e., whether it counts as not obviously wrong – depends on the perspective from which it is assessed. Since this article is situated in the liberal tradition, I assume that disagreements about justice are reasonable when they are broadly consistent with the liberal commitment to equal respect, and not based on evident empirical falsehoods.Footnote 30 For instance, citizens disagree about the particular tax policies that should be implemented within society: some favour proportional taxation on grounds of liberty, others campaign for progressive taxation on grounds of equality. Since neither view obviously violates equal respect, they are both reasonable, and hence worthy of consideration.

Or else, citizens disagree over whether abortion is morally justified, yet often neither party to this disagreement can be shown to be clearly mistaken. While they all agree that the permissibility of legalizing abortion hinges on whether the foetus is a person, they disagree about what qualifies as a person. For those who think that a person is created at the moment of conception, abortion ought to be prohibited, for those who think that persons must possess certain cognitive and emotional abilities, which foetuses lack, abortion should be legalized. To the extent that neither view strikes us as obviously implausible, the disagreement in question is reasonable.

Of course, an anti-abortionist might find the views of a pro-choice activist obviously mistaken, but the question is whether she has grounds for so thinking. The issue is hotly contested, and it seems dogmatic, from a liberal perspective, to deny that there is a range of reasonable positions on the matter – in the same way in which it is dogmatic to deny that there is a range of reasonable interpretations of, say, a novel or a poem, even if we endorse one in particular.

Notice that reasonable disagreement occurs not merely in the realm of morality, but also in the sciences. Scientists might disagree, for instance, about whether certain minerals could be found on a distant planet, because the available evidence is inconclusive. In such circumstances, their disagreement qualifies as reasonable. Doctors might disagree about the illness affecting a particular patient, because (like the principle of equal respect) her symptoms may be interpreted in a variety of different ways. To the extent that this is true, their disagreement counts as reasonable.

In short, when disagreement is reasonable, none of the parties involved can be accused of being irrational or obviously mistaken. To that extent, their points of view merit being taken seriously.

Although there are deep controversies about justice, from the liberal perspective adopted in this article, some positions are straightforwardly wrong, in so far as they could never count as expressions of equal respect. If, for example, someone were to argue that it is permissible to torture children for fun, or that slavery is a morally commendable practice, liberals would consider their views unreasonable. Whatever the truth conditions of statements about justice are, a liberal can safely assume that if anything is unjust, slavery and torture of the innocent are. By the same token, the claim that a just society may deny its citizens rights to free movement, thought and education, would also count as unreasonable. How can a society be just, namely express equal respect for its citizens, if it denies their most basic rights?

Once again, unreasonable disagreement is not confined to moral matters, but extends to the natural and social sciences. For instance, if nowadays someone were to defend the view that the Earth is flat, most of us would count his disagreement as unreasonable, because – from a scientifically-minded perspective – there is overwhelming evidence to the contrary. Or else, if someone were to defend geocentrism on theological grounds, we would have to discount her view as absurd, since all the evidence at our disposal points towards heliocentrism. In short, when disagreement is unreasonable, some of the views defended can be discounted as straightforwardly irrational or implausible (from the relevant perspective).

Having distinguished between these different kinds of disagreements, let us now consider how they combine with one another, giving rise to a fourfold logical space, as illustrated in Table 1. In the remainder of the article, I shall examine the justification of democracy under circumstances of, respectively, thin and thick reasonable disagreement about justice. I discount unreasonable disagreement in so far as this falls outside the liberal commitment to equal respect. Liberals are committed to equal respect qua mutual justification, but their justificatory audience does not include those who, from a liberal perspective, hold unreasonable views. If, for example, someone objects to a particular institutional arrangement on the grounds that it does not confer absolute power on him or that it leaves no room for slavery, liberals need not take his disagreement seriously. The views he proposes are clearly unreasonable. Not every objection carries normative force, only reasonable ones do.Footnote 31

Table 1

Comparison of Different Kinds of Disagreement

Before proceeding further, let me anticipate how my analysis of the value of democracy will differ from two prominent views in the literature, which also emphasize the normative role of disagreement. Both Thomas Christiano and Jeremy Waldron have suggested that the intrinsic value of democracy depends on the existence of reasonable disagreements about rights and justice.Footnote 32 In the presence of such disagreements, they argue, each should have an equal say in political decision making. As will become apparent in what follows, on my view, the presence of reasonable disagreements about justice is a necessary, yet not a sufficient, condition for thinking of democracy as intrinsically valuable. Democracy can only be defended as an intrinsic requirement of justice when disagreements are reasonable as well as thick. Footnote 33

Moreover, while my view is explicitly grounded in the principle of equal respect qua mutual justifiability, understood as the bedrock of liberal justice, this principle is not central to the works of Waldron and Christiano. The former thinks of democracy as intrinsically valuable independently of justice.Footnote 34 The latter defends democracy on the basis of a particular account of justice, but that account differs from the one I propose. For Christiano, justice is not about mutual justifiability, but about the ‘public realization of equal advancement of interests’ (for more details on how my view relates to Waldron's and Christiano's, see fnn. 32–5).Footnote 35

No Democracy

A first possibility is to think that, under trd, democracy should play no role in relation to justice. Although people reasonably disagree about justice, so the argument goes, we can plausibly identify different levels of expertise among them. The distribution of power within society should then mirror that of justice expertise. Consider the following analogy. You have had dinner with friends, and the moment comes when you have to split the bill. Each of you does the calculations and comes up with a different (yet plausible) figure. What should you do in these circumstances?Footnote 36 Suppose one of you, Jacopo, has an outstanding track record in arithmetic. If so, it makes sense to defer to his judgement.

Of course, another possibility may be to deliberate and try to reach a consensus. But assume that there is no time for that. The calculations are complicated (it is a long bill!) and you want to go to the movies. Either you pay now, or you miss the cinema. The rational thing to do, under these circumstances, is to accept Jacopo's verdict as authoritative. There is a truth about what each person's fair share is, and the procedure that best tracks that truth, under the circumstances at hand, is one that gives Jacopo the final word on the matter.

Note that this conferral of authority on Jacopo does not violate the mutual justifiability constraint. Since the goal of the group is to discover the truth about how much each should pay, and Jacopo qualifies as an expert on the matter, they all have reason to defer to his judgement. If they want to get to the truth, and they are rational, they must recognize that Jacopo is the way to go.

Similarly, assume that we could regard political philosophers (or any other professional category) as the experts on what justice requires. It would then make little sense for anyone to insist that society should be governed democratically. The outcome of democratic procedures would in all likelihood be less just than that which the philosophers could establish. More generally, if we can identify experts about political morality whose views can be trusted to reflect the truth, we are naturally drawn towards what David Estlund calls epistocracy: a form of government in which those who know best hold power.Footnote 37

Some might question the analogy between splitting a bill and deciding who should hold political authority. First, it might be argued that splitting a bill is a technical (i.e., arithmetical) problem, whereas determining who should hold political authority is susceptible to interpretation, based on moral beliefs.Footnote 38 This objection fails to acknowledge that, under trd, the question of justice is ex hypothesi just as technical as that of splitting a bill. In the same way in which there are experts about mathematics, with the technical knowledge authoritatively to solve disagreements about bill-splitting, so too there are experts about justice, with the technical knowledge to solve disagreements about who has a right to what within society.

Secondly, it might be objected that whatever reasoning is appropriate in the bill-splitting case need not transfer onto the political one because in the latter the stakes are much higher than in the former. But why should it be so? If our aim is to realize justice, we should do whatever maximizes our chances of attaining this goal. If there are experts who are more likely to identify what justice demands, the rational thing to do is to let them decide, no matter how trivial or important the decision in question is.

Thirdly, it may be argued that, unlike in the bill-splitting case, not being allowed to take part in political decision making is equal to being stigmatized as inferior (hence disrespected) within the relevant community, because political decisions are forced upon us. Even if it is rational to trust a doctor's (i.e., an expert's) medical advice, it is typically up to us to decide whether to accept medical treatment. The reason for this is respect for persons qua rational and autonomous agents: nobody is entitled to force others to lead their lives in ways they do not endorse. If so, would giving a panel of experts the power to impose the demands of justice on us also constitute a violation of equal respect?Footnote 39

It would not, because there is an important disanalogy between the medical case and that of justice. What one does with one's health is a private matter, but whether or not we act in line with justice affects what others can do with their own lives. In a just system, all are respected qua autonomous agents; in an unjust one, some are not. Norms of justice can thus be enforced without violating equal respect because their very point is to realize equal respect. One is not treated disrespectfully if one is forced to give others what they are owed on equal respect grounds. On the assumption that experts about justice enforce what justice demands, their imposition of justice on the rest of the citizenry does not violate equal respect, but is necessary to secure it.

Of course, one might still resist ‘the rule of the experts’ by doubting the effectiveness of leaving justice in the hands of a few (supposedly) enlightened individuals. After all, how can we trust the expert kings to behave as justice requires once they are placed in a position of power? These worries about power abuses lead us to the second answer to the question of why we should care about democracy under trd if at all.

Implementation Democracy

We might think that, by distributing power roughly equally across the citizenry, democracy is more likely to realize the demands of justice stably than any other political system. Following this line of argument, democracy is justified as a second best. Ideally, a society of expert kings would be better, but since in our non-ideal world we cannot trust them (or anyone else) to hold so much power without abusing it, we organize society such that power is sufficiently dispersed, namely democratically. In a democracy, political leaders and public officials must be sensitive to the interests and the demands of the electorate in order to remain in power. Moreover, it is often said that democratic institutions, with their participatory and egalitarian ethos, have the capacity to generate the social solidarity required to implement the demands of justice.

On this view, democracy is not an intrinsic requirement of justice; it is only an instrument for its implementation. As Richard Arneson says, ‘[s]ystems of governance should be assessed by their consequences; any individual has a moral right to exercise political power just to the extent that the granting of this right is productive of best consequences overall.’Footnote 40 In a similar vein, even if we can plausibly identify experts on justice, we may still want to distribute political power roughly equally across the citizenry in order to prevent flagrant abuses of it or to encourage social solidarity. If this is what we believe, then our defence of democracy is purely instrumental, solely grounded in concerns about the implementation of justice.

Epistemic Democracy

Alternatively, under trd, we may want to defend democracy by appeal to its virtues as a truth-tracking device.Footnote 41 If, instead of being concentrated in the hands of a few wise individuals, expertise about justice were equally distributed across society, democracy might indeed be the best epistemic procedure to discover what justice demands. As famously observed by the Marquis de Condorcet, if each voter has more than a 50 per cent chance of getting the answer right, and voters’ judgements are independent, a majority is more likely to be correct than each person, and the likelihood increases the more voters there are.Footnote 42

Otherwise, we may think that a deliberative form of democratic politics would offer a fruitful approach to ascertaining what the just course of action is with respect to specific political dilemmas. By exchanging reasons and sharing information, so the argument goes, citizens are more likely to discover the truth about justice.Footnote 43

Moreover, deliberation and aggregation need not be mutually exclusive. In the real world, deliberation alone is unlikely to suffice to establish political outcomes. In many cases, disagreement is bound to persist even after deliberation. Given certain facts about the distribution of expertise, it is quite plausible to argue for a combination of deliberative and majoritarian processes as the best truth-tracking strategy.Footnote 44 Since we cannot deliberate ad infinitum, or until we reach a consensus, we can think of deliberation and majority rule as working in tandem, as part of a reasonably feasible and epistemically reliable political system.

To sum up, under trd, our commitment to democracy is entirely dependent on facts about the distribution of expertise and good will. If expertise is confined to a few trustworthy people, then their views should be authoritative. If, however, experts are likely to abuse their power, we might prefer democracy as an implementation device. Otherwise, if expertise about political morality is evenly distributed within society, democratic decision procedures might be chosen as epistemically best.

In all of these cases, democracy is defended on instrumental, rather than intrinsic, grounds. The only way to defend democracy as an intrinsic requirement of justice under trd would be to stipulate that it is. What is worse, making such a stipulation would lead us to develop a potentially self-undermining account of justice, according to which justice requires democracy even though democracy is likely to generate unjust outcomes. In the light of this, if we, citizens of existing liberal democracies, were under circumstances of thin reasonable disagreement, our commitment to democracy (if at all justified) would have to be instrumental, not intrinsic. But can we plausibly claim that these are the circumstances under which we live? Perhaps not.

When it comes to morality, including political morality, citizens reasonably disagree not only about whether certain laws or policies are just, but also about the truth conditions of claims about justice. Some, for instance, believe that just policies are those which maximize overall utility within the constraints of fundamental rights; others that they are those which maximize average utility; others still think that laws are just in virtue of their conformity with Kant's Categorical Imperative or Rawls's principles of justice; some religious citizens hold that justice requires laws and policies to reflect our status as God's creatures.

As many have remarked, in these circumstances, asking experts to settle the issue will not do. We can easily point to experts in physics, mathematics, astronomy, medicine and so forth, but when it comes to morals, there is no undisputed, publicly justifiable, criterion for identifying expertise.Footnote 45 Is the Pope a moral expert? Or perhaps the Dalai Lama? Are political philosophers the true experts? What about political activists, politicians, judges or free thinkers? It seems impossible to give a non-controversial answer to these questions.Footnote 46 In short, many of the disagreements which characterize our political world are not thin, but thick, concerning the very truth conditions of statements about justice.

Epistemic Democracy

Democratic decision procedures typically involve deliberation as well as majority rule. Can both dimensions of democracy be defended on epistemic grounds under TRD? Let us consider deliberation first. It seems that under TRD there could be room for an epistemic defence of deliberation. Citizens whose views diverge (no matter how ‘thickly’), but who are also aware of their own fallibility, have an interest in exchanging reasons and confronting each other in discussion. As John Stuart Mill famously argued, by deliberating with others we are more likely to make epistemic progress. Moreover, even when we argue with opponents who strike us as deeply mistaken, by trying to persuade them, we remind ourselves of the reasons in support of our views, and avoid the risk of holding them in a purely dogmatic fashion.Footnote 47

Although deliberation can be defended on epistemic grounds under TRD, deliberation itself is insufficient for a viable democracy and needs to be supplemented by majority rule. Can we defend majority rule on epistemic grounds under TRD? It would seem not.

In the absence of a shared view of what counts as expertise about justice, we can no longer invoke Condorcet-type reasons in support of majoritarian democratic procedures. Recall that majority rule only gains privileged epistemic status when each voter is ‘competent’, i.e., when she has more than a 50 per cent chance of selecting the right answer. But under TRD, there is no unproblematic notion of expertise on the basis of which to decide whether the ‘competence’ assumption holds. Catholic believers, for example, may think that priests are the experts. Protestant believers, by contrast, may think that each individual is equally well placed to come to the truth and so forth. More examples could be given, but the general point should be clear. Under TRD, universal suffrage and majority rule cannot be justified to all rational persons on epistemic grounds.Footnote 48

This conclusion contrasts with an influential view proposed by David Estlund: epistemic proceduralism.Footnote 49 On Estlund's account, democracy (which presumably includes majority rule) is the epistemically best decision-making system among those which can be justified to all qualified points of view (to all ‘rational/reasonable’ persons). It is unclear, though, how this claim can be supported under TRD. A defence of the truth-tracking properties of majority rule presupposes an account of the nature of the truth about justice allowing us to make judgements about people's competence. But a generally accepted account of the nature of the truth about justice is precisely what we are missing under TRD. How can one argue that majority rule is better at tracking the truth without knowing what the truth conditions of claims about justice are?Footnote 50

Estlund wishes to avoid this difficulty by assuming a deflationary understanding of truth. On his view, we can assume that there is a truth about justice without offering an account of what the truth conditions of statements about justice are. By truth, Estlund means ‘the following very minimal thing: if gender discrimination is unjust, then it is true that gender discrimination is unjust.’Footnote 51 But this understanding of the truth is too empty to do the work Estlund wants it to do. Whether a particular procedure is epistemically good or bad depends on the nature of the object the procedure is trying to ascertain. For example, a blood test seems to be a good epistemic procedure to establish whether someone is affected by the human immunodeficiency virus (HIV), because the truth or falsehood of the statement ‘The patient is affected by HIV’ depends on facts about what viruses are present in (or absent from) her blood. It is because we agree about the truth conditions of this statement – i.e., facts about the blood – that we can defend a blood test as a good epistemic device. But without an account of the nature of the truth conditions of statements about justice – other than a mere assertion that some such conditions exist – we will have a hard time defending any decision procedure on epistemic grounds, including majority rule.Footnote 52

In the light of this, I conclude that, while under TRD we may have epistemic reasons for defending deliberation (as a way to keep our own fallibility in check and to make progress in understanding), we have no generally acceptable epistemic reason to defend majority rule, therefore little reason to defend democracy in its full sense.

Implementation Democracy

Under TRD, we might still want to defend democracy instrumentally, as a way to ensure against tyranny. Since democracy presupposes an equal allocation of political power across citizens, and arguably tends to foster trust and fellow-feelings, it is unlikely to degenerate into forms of government that violate the basic constitutional constraints which are part of any plausible interpretation of justice. To the extent that this is true, we may have implementation-related reasons to defend democracy under TRD.

Intrinsic Democracy

Finally, we may think that, under TRD, democracy is a justificatory device, a way of moving the process of inter-subjective justification from philosophical theory to real-world political practice. As I argued earlier, there are some guarantees that any political arrangement must provide for its citizens if it is to be justified to them. From a liberal perspective, a state that did not protect its citizens’ freedom of movement, life, bodily integrity or minimal subsistence would certainly be unjust, it would fail to respect them. Indeed, rational agents concerned with furthering their life plans could never unanimously agree to this kind of political set-up.

Apart from ruling out obviously unjust social systems, the standard of equal respect qua mutual justifiability remains inconclusive about many aspects of social organization, including redistributive taxation, school curricula, abortion laws and much else. How, then, can a state settle such matters in a way that best captures the ideal of equal respect for persons as rational and autonomous agents? On the intrinsic view, the answer is: democratically. Democratic procedures – including deliberation and majority rule – are as close as we can get, from a practical, real-world, point of view, to the ideal of mutual justification. To respect all persons’ status as rational agents under TRD, so the argument goes, is to allow each of them to contribute to collective decision making on an equal footing.Footnote 53

This way of conceptualizing the relationship between justice and democracy sheds light on the apparent inconsistency in the intrinsic account discussed in Section I. The worry took the following form: How can a theory of justice contain democratic rights to vote against what the theory indicates as requirements of justice? In other words, how can a theory of justice contain rights which can lead to violations of other people's rights? For instance, if we can unproblematically assume that justice requires implementing the difference principle, how can we also say that there is a justice-based democratic right to vote for tax reforms that would prevent the difference principle from being realized?Footnote 54

Looking at justice under TRD allows us to make sense of why this is more an inevitable tension than a genuine logical inconsistency. Consider the tax reform example. For those who advocate the difference principle on grounds of justice, citizens are treated respectfully only if the distribution of income and wealth benefits the worst off. But under circumstances of thick reasonable disagreement, we cannot unproblematically assume that this is what equal respect for persons requires. Some reasonably hold this view, but others equally reasonably believe that respect for persons has different distributive implications.

Under such circumstances, a state cannot claim to show equal respect for its citizens if it simply imposes one, reasonably contestable, view of justice on them. To do so would be to fail to recognize their equal status as rational and autonomous agents. That said, we cannot suspend judgement and refrain from taking decisions about social distributions until full agreement on matters of justice has been reached, as this would lead to social paralysis. In this scenario, justice requires that we address reasonable disagreements and come to select particular social outcomes in a way that reflects citizens’ status as autonomous agents and practical reasoners. This is what democracy, via deliberation and majority rule, allows us to achieve. In short, on this view:

In particular, by deliberating and listening to one another's reasons, we express respect for each other as rational persons. Moreover, as reasoners who disagree, we may hope through argument to make progress in understanding one another, and converge on a single answer we all regard as compelling. This would allow us fully to realize the ideal of mutual justification at the heart of the liberal understanding of justice adopted here. This ideal of complete mutual justifiability is one we should aspire to, but are unlikely ever fully to achieve. If disagreement is indeed central to politics, hoping for universal agreement is somewhat utopian.Footnote 55

Since decisions have to be taken, deliberation is not enough. The deliberative phase has to be followed by some aggregative process (most likely majoritarian) allowing us to establish which view is to prevail. This may look like a less-than-perfect solution, in that it inevitably results in the imposition of what a majority, however qualified, considers the appropriate interpretation of justice, when we know, ex hypothesi, that the minority's view could also be correct (because the disagreement is reasonable). Given the need to take decisions, this is the best we can hope for under TRD. Under these circumstances, a democratic system is the one that best expresses equal respect for persons as rational and autonomous agents. In other words, under TRD, basic rights and democracy are sine-qua-non, lexically prior, demands of justice. Other reasonably contestable claims about justice are to be seen as legitimate inputs to democratic decision making. Were they implemented undemocratically, they would not be consistent with equal respect qua mutual justification, and hence with justice itself.Footnote 56

Before proceeding, let me further clarify how my view differs from another justice-based intrinsic defence of democracy, offered by Thomas Christiano. As I have mentioned earlier in the text, Christiano's notion of justice is not articulated in terms of mutual justifiability, but is instead based on the idea of public equal advancement of interests.Footnote 57 Such a reference to the advancement of interests gives Christiano's argument for democracy a rather instrumental, rather than intrinsic, flavour. As he puts it, by stemming from ‘the requirement to advance the well-being of persons in the main social and political institutions’, his view ‘shares something with classical utilitarianism’, namely ‘respect for the empirically discernable [sic] and multifaceted conditions under which well-being can be advanced’.Footnote 58 From this perspective, democracy, understood as a combination of deliberation and majority rule, is not a real-world exercise in mutual justification, but the best mechanism for publically realizing citizens’ interests equally, thereby showing equal respect for them.Footnote 59

For Christiano, the value of deliberation largely rests on its ability to help diminish misunderstandings and to correct cognitive biases which are likely to undermine the equal advancement of citizens’ interests. In his words: ‘Public deliberation [conducted on an egalitarian basis] has instrumental value in a democratic society since it leads to the development of an informed, rational, and morally sensitive citizenry,’ which is a precondition for the pursuit of the common good.Footnote 60 Similarly, the value of majority rule largely rests on its ability publicly to advance citizens’ interests equally. Although majority rule almost inevitably generates winners and losers, it gives citizens equal control over the outcomes of collective decision making. Equal control is, in turn, seen as a reliable means of publicly promoting citizens’ interests equally, against the background of reasonable disagreement.Footnote 61

In sum, although Christiano aims to offer a justice-based intrinsic defence of democracy, his understanding of justice qua public equal advancement of interests leads him to appeal to instrumental considerations which are not equally integral to a defence of democracy based on justice qua mutual justifiability.Footnote 62

The Lottery Objection

This objection targets my claim that a democratic system is the one that best expresses equal respect for rational and autonomous agents under TRD. In particular, it says that, under TRD, we have no more reason to adopt deliberation-cum-majority-rule, than we have to adopt decision-by-lottery. Democracy and decision-by-lottery, so the argument goes, can both be justified in the eyes of rational and autonomous agents. Is this really the case? I believe not.

Rational and autonomous agents are committed to justifying their claims to one another, and mutual justification can only occur through deliberative reason-giving, not through lotteries. Imagine a Catholic and an atheist being told that the legal permissibility of abortion will be decided by tossing a coin. Surely both could reasonably object to this proposal on the grounds that it fails to express respect for their status as rational agents. Respect for this status requires their reasons (in favour or against abortion) to be heard. Adopting lottery-based procedures would amount to moving from reason to randomness.

The supporter of lotteries may accept that deliberation uniquely satisfies equal respect, and reformulate her objection more locally, suggesting that lotteries could, in principle, replace majority rule. On this view, lotteries would be employed to decide which of the views that have survived deliberation should prevail. Indeed, is not a lottery just as fair, just as respectful as majority rule is? The answer is: No. Recall that equal respect for persons requires mutual justifiability. As I mentioned earlier, the ideal of mutual justifiability is one we should aspire to, but will probably never be able fully to achieve, at least as long as there is pluralism. In the light of this, the best we can hope for is to approximate this ideal as much as possible, and no feasible decision procedure seems to be as well placed to do this as majority rule.

Majority rule ensures that reasonable political outcomes are accepted by as large a number of the populace as possible. In so doing, majority rule offers the best approximation of mutual justifiability under TRD. A minority outvoted in an election has reason to abide by the majority decision not because that decision is most likely to be correct, but because it is the most widely justified. A lottery, by contrast, may very well pick out the outcome preferred by a minority, which is less broadly shared and less widely justified. Indeed, even a weighted lottery – i.e., a lottery where the outcome preferred by the majority is given greater probability to be selected – would not ensure the maximum possible justifiability as compared to majority rule, in so far as minority-preferred outcomes could still in principle be selected (no matter how low their probability).Footnote 63 In short, under TRD, deliberation cum majority rule can be shown to be superior to lotteries solely by appeal to justice-based considerations (although instrumental considerations could also lend further support to majority rule over lotteries).Footnote 64

At this point, readers might worry about the assumption (implicit in my argument) that, after deliberation, only two options are left on which to vote. This need not be so. More than two alternative views might remain, and in such cases, majority rule – in the form of pairwise majority voting – is known to lead to cycles. To avoid cycles, we might want to adopt slightly modified, ‘broadly majoritarian’, procedures. Such broadly majoritarian procedures, however, are known to be susceptible to strategic manipulation.Footnote 65 When strategic manipulation is a live possibility, it becomes much harder to defend the claim that democratic voting is the all-things-considered best way to instantiate equal respect for persons. When decisions have to be taken between more than two options, those who have greater information about others’ preferences, or are better at strategic thought, have an unfair advantage. In the light of this, lotteries might actually turn out to be better instantiations of equal respect than conventional, broadly majoritarian, voting.

There are two possible responses to this worry. First, as argued by John Dryzek and Christian List, there is reason to believe that, in the real world, group deliberation diminishes participants’ incentives to adopt strategic behaviour. On the one hand, deliberation diminishes participants’ incentives to lie, since false statements are likely to conflict with the evidence presented by other participants in the deliberative process, and hence to be exposed as incorrect. On the other hand, deliberation helps create conditions for co-operation, by giving participants a strong sense that the problem on which they need to take a decision is a genuinely common one. It moves the debate from an ‘I-frame’ to a ‘we-frame’.Footnote 66 In so doing, deliberation minimizes the likelihood of strategic manipulation in those cases in which strategic manipulation is an option (i.e., when more than two alternatives survive deliberation, and decisions are taken through voting).

Secondly, even if, in the real world, deliberation does not always succeed in diminishing incentives to strategic behaviour, strategic behaviour itself is a moral pathology of real-world politics, and my defence of democracy in the present article abstracts away from such pathologies. Some readers might find this – and other – idealizations problematic, which leads me to the next objection.

The Idealization Objection

Some might complain that my account is implausibly idealized. After all, my intrinsic defence of democracy only works if we assume that citizens are well informed, prepared to give reasons, reluctant to adopt strategic behaviour, committed to equal respect and so forth. But this thoroughly optimistic picture is very different from what we find in real-world societies. Existing democracies are far more imperfect than those envisaged in this article.Footnote 67 Should we therefore conclude that my version of the intrinsic account is implausibly idealized? In answer to this objection, I agree that my account contains significant idealizations, but I deny that they are implausible ones.

My aim is to consider whether democracy should be part of a larger theory of justice, and any account of justice must rely on some idealizations.Footnote 68 Whether these idealizations are warranted or not depends on whether they assume away those persistent (perhaps immutable) features of human nature which give rise to the need for justice and politics in the first place. My account would therefore be implausibly idealized if it dispensed with moderate resource scarcity, assumed that human beings were angelically altruistic, and denied the existence of reasonable disagreement about justice. These are clearly persistent features of human nature without which the question of justice, and the need for politics as we know it, would cease to exist.

By contrast, a disposition to exchange reasons, an effective and transparent information system, and a commitment to justice abstractly conceived are not beyond human reach (if they were, then why would we worry about justice in the first place?) They do not presuppose a denial of the circumstances which generate the need for politics. Instead, they assume away what might be called ‘the pathologies’ of real-world politics. Of course existing societies are far from the ideal I am sketching, but this is no critique of that ideal. If anything, the ideal would be suspicious if it offered an a-critical defence of the status quo. So long as my idealizations are not self-defeating, my defence of the intrinsic value of democracy survives.

That said, I do agree that a crucial task for political philosophy is to ask what justice requires under the non-ideal circumstances of real-world politics (such as circumstances of unreasonable disagreement). This, however, is an investigation that I leave for future work.

The Asymmetry Objection

The asymmetry objection points to what looks like an inconsistency in my argument. On the one hand, I place great emphasis on the circumstances of thick reasonable disagreement. On the other, my whole argument assumes a commitment to equal respect qua justifiability to rational and autonomous agents. But where does that commitment come from? Can we say that it is a true demand of justice? Could not someone reasonably disagree with it?

I can think of three ways of answering this challenge, which I simply flag, without committing to any one in particular. Readers should choose whichever they find most convincing.

First, it might be responded that we do in fact have sufficient evidence to regard the principle of equal respect as true, in so far as all main moral codes incorporate it in one form or another, and those which do not are typically based on incorrect factual claims – for example, that people of a certain race are genetically less intelligent than others.Footnote 69 Following this line of argument, although the ideal of equal respect qualifies as a truth about justice, its implications are unclear. Responding to this fact, in a way consistent with equal respect, is the task of democracy.Footnote 70

Secondly, we might argue that a commitment to equal respect qua justifiability to rational agents is not of a substantive but of a methodological kind. On a Kantian, public, understanding of reason, a normatively valid claim must be justifiable to all rational persons. If others are rational, use their powers of reason properly and yet still disagree with us (i.e., if there is reasonable disagreement), this meta-principle tells us that our views do not have the required validity to qualify as correct beyond reasonable doubt, hence to be genuinely normative for them.Footnote 71

Thirdly, and finally, we might simply acknowledge that we, Western liberals, have such a thick commitment to mutual justifiability to rational persons that it would be impossible for us to theorize about justice prescinding from that commitment. Although we cannot conclusively establish whether it is true or not, neither can we avoid appealing to it when we think about justice. In normative theorizing we have to start from somewhere, and there seems to be no place other than our most deeply held convictions.Footnote 72