1 The principle of reasonableness is employed and defended by other democratic thinkers as well, most prominently by Gutmann, Amy and Thompson, Dennis in Democracy and Disagreement (Princeton: Princeton University Press 1996), 52Google Scholar and Why Deliberative Democracy? (Princeton: Princeton University Press 2005), 3 as well as Freeman, Samuel in ‘Deliberative Democracy: A Sympathetic Account,’ Philosophy and Public Affairs 29, 4 (2000) 371–418.CrossRefGoogle Scholar In this essay, I focus on Joshua Cohen’s arguments for the narrow conception of deliberative democracy.

2 See Cohen, Joshua ‘Reflections on Habermas on Democracy,’ Ratio Juris 12, 4 (1999) 385–416CrossRefGoogle Scholar, at 396.

3 Another interpretation of the idea of a model is that the ideal deliberative procedure is a kind of contractualist procedure in which laws and policies are justified. To the extent that these proposals are justified in such a procedure, the question is to figure out how to bring about these justified proposals in the conditions of the actual world. Cohen links the proposals justified in the ideal deliberative procedure to actual democratic deliberation by arguing that such proposals are more likely to emerge from democratic political institutions in which citizens attempt to resolve disagreement by giving reasons that each can reasonably accept. The argument is that the more actual democratic processes resemble the justificatory process embodied in the ideal deliberative procedure, the more the outcomes of the actual democratic process will resemble the outcomes of the ideal deliberative procedure. Briefly, Cohen argues that only if citizens aim at deliberative justification in the process of actual democratic deliberation can they hope to bring about outcomes that are justified in the ideal procedure. Furthermore, actual deliberation fosters mutual respect among participants and so helps foster the kind of discussion that is had in the ideal procedure. In addition, actual deliberation enables citizens to master political principles because that is the currency of discussion and finally, deliberation tends to reduce the extent of disagreement in the population.

Needless to say, the inference from the justification of the outcomes of the hypothetical procedure to the justification of the outcomes of an actual procedure that approximates the ideal procedure is highly problematic. The inference is clearly subject to the worries expressed in the theory of the second best. Just because the properties of the actual procedure resemble the properties of the hypothetical procedure, it does not follow that the properties of the outcomes of the actual procedure will resemble to the same extent (whatever that means) the properties of the outcomes of the hypothetical procedure. The four arguments mentioned above do not seem to me to ensure that the outcomes of an actual non-deliberative democratic institution will more closely resemble the outcomes of the ideal procedure than the outcomes of an actual highly deliberative democratic society. It may even turn out that the outcomes of a non-democratic but benevolent dictatorship will be closer to those of the ideal procedure than those of an actual attempt to realize deliberative democracy. We need a much more powerful set of instrumentalist arguments for the claim that an actual deliberative democracy will achieve defensible outcomes. But I will leave these issues to the side in this paper.

4 As long as there is agreement on the list of considerations, the remaining disagreements on the relative weights of the considerations and their exact interpretation can be legitimately resolved by majority rule, Cohen says. Citizens can reasonably accept terms of association when they agree on the list of considerations in support of those terms. See Cohen, ‘Procedure and Substance in Deliberative Democracy,’ reprinted in Deliberative Democracy, Bohman, James and Rehg, William eds. (Cambridge, MA: MIT Press 1998) 407–37, esp. 414.Google Scholar

5 Here there seems to be a slight addition to the account of the ideal deliberative procedure. In ‘Deliberation and Democratic Legitimacy’ (reprinted in Bohman and Rehg, 67-91, esp. 74), Cohen contrasts the reasoned aspect of the procedure to the mere aggregation of preferences or interests and insists that only the force of the better argument is exercised; in ‘Reflections on Habermas on Democracy,’ Cohen adds that the reasons have to be of a particular sort, namely, ones that can be accepted by other reasonable citizens. This may prove to be a major change.

6 See my ‘The Significance of Public Deliberation,’ in Bohman and Rehg, for a defense of some of these claims.

7 See ‘Reflections on Habermas on Democracy,’ 396.

8 Ibid.

9 Ibid., 398.

10 As Cohen and Rawls use these terms, to say that A can reasonably accept terms of association T is to say that A cannot reasonably reject those terms. And to say that A cannot reasonably accept the terms T is to say that A can reasonably reject those terms. So, given a particular reasonable comprehensive doctrine, reasonable rejectability and reasonable acceptability are exclusive. Of course, it is possible for terms T to be both reasonably rejectable by one person and reasonably acceptable to another when those persons have different reasonable comprehensive doctrines. And of course one person might change from reasonably rejecting terms to reasonably accepting those terms because of a change in reasonable comprehensive view.

11 See Cohen, ‘Moral Pluralism and Political Consensus,’ in The Idea of Democracy, Copp, David Roemer, John and Hampton, Jean eds. (Cambridge: Cambridge University Press 1993) 270–91, esp. 281-2.Google Scholar

12 See Cohen, ‘Procedure and Substance in Deliberative Democracy,’ 407-37, esp. 414.

13 Even with all these qualifications and caveats on the kind of consensus sought after, the consensus does not have to be entirely actual, even among reasonable persons. For some may wrongly believe that the terms proposed are incompatible with their reasonable doctrines. Some such false beliefs ought not to count against saying that they can reasonably accept the terms. But others presumably do count against saying that they can reasonably accept the terms. Presumably, reasonable acceptance is not closed under implication. But what the exact boundaries are is uncertain. I do not think that these difficult issues need to be settled for the purposes of this paper. See Gerald, Gaus Justificatory Liberalism (Oxford: Oxford University Press 1996)Google Scholar for a nuanced treatment of this issue in the context of this kind of theory.

14 See ‘Reflections on Habermas on Democracy,’ 405.

15 Cohen understands treating others as equals this way in ‘Democracy and Liberty,’ in Deliberative Democracy, Jon Elster, ed. (Cambridge: Cambridge University Press 1998), 192.

16 See my ‘Knowledge and Power in the Justification of Democracy,’ Australasian Journal of Philosophy 79, 2 (2001) 197-215, esp. 212-15, for a defense of this claim.

17 The democratic argument for the principle of reasonableness might seem to work if we accept Cohen's epistemological argument (which I will discuss and argue against in detail in the next section of the paper). For, given the epistemological argument, if another reasonably disagrees with me on the basis of a different comprehensive view from my own, then I appeal only to my belief in justifying forcing him to live in accordance with my views. Then I treat the other as an inferior to the extent that I make laws on the basis of my beliefs but I do not permit the other to base laws on his own belief. (See Cohen, ‘Democracy and Liberty,’ 206 for this epistemic consideration within the democratic argument.) I have given myself a right that I do not similarly offer to others and, consequently, I have treated the other as an unequal.

I doubt that even this implies a genuine inequality. The above argument fails to take into account that the proposals are made to a democratic assembly. For suppose we live together with the dissenters in a democratic society wherein each has the right to advance his views in the public forum and wherein only some can get their way. It would seem that the democratic way of making decisions treats everyone as equals in this context.

But once we reject the epistemological argument (as I will urge below that we must), disagreement on comprehensive views in no way suggests that one treats the other as an inferior when one proposes terms of cooperation on the bases of considerations they cannot reasonably accept.

18 This argument is bolstered by the last part of this paper where I argue in favor of what I call the ‘deliberative impasse’ and the impossibility of a neutral baseline which itself needs no justification and from which departures require justification.

19 One commentator has objected that ‘the devout Catholic who is told that his Catholic views are inferior may find it psychologically very hard to distinguish this from a claim that he is inferior.’ I am afraid that I do not see this. Though there may be some who might have this reaction, I do not see how, in a society where many deeply different views flourish, this would be a general reaction. Of course, when someone's deeply held views are rejected by the democratic assembly as a basis for association, that person is apt to feel somewhat alienated from his political world. But this is likely to happen more or less to anyone in a democratic society where there is substantial disagreement about how it ought to be organized. One way of seeing the point of democratic institutions is as a way of distributing this inevitable feeling of alienation in an egalitarian way.

20 This kind of case must be distinguished from the case of majority tyranny where the majority simply mistreats the minority on the basis of crazy views about justice or hatred. In the case of permanent minorities, the majority may well be treating the minority as equals in accordance with its own distinctive conception of the minority's interests and its own norms of equality.

21 See Political Liberalism, 62 and 226 (the same passage recurs).

22 See my The Constitution of Equality: Democratic Authority and Its Limits (Oxford: Oxford University Press 2008), 288-99.

23 To use Robert Dahl's felicitous phrase, democracy is in fact polyarchy or ‘rule by minorities.’ For these expressions and the idea of democracy that lies behind them, see Dahl, Robert A Preface to Democratic Theory (Chicago: University of Chicago Press 1956).Google Scholar

24 For discussions of examples of some of these kinds of arrangements, see Lijphart, Arend Democracy in Plural Societies (New Haven, CT: Yale University Press 1977),Google Scholar and Horowitz, Donald L. A Democratic South Africa? Constitutional Engineering in a Divided Society (Oxford: Oxford University Press 1991).CrossRefGoogle Scholar

25 See my The Constitution of Equality throughout for an extended defense of this thesis.

26 See Rawls, John Political Liberalism (New York: Columbia University Press 1993), 61.Google Scholar

27 Cohen, Joshua ‘Moral Pluralism and Political Consensus,’ in The Idea of Democracy, Copp, David Roemer, John and Hampton, Jean eds. (Cambridge: Cambridge University Press 1993) 270–91,Google Scholar esp. 284.

28 Ibid.

29 See ibid., 283.

30 ‘Moral Pluralism and Political Consensus,’ 284. Rawls suggests this same idea when he speaks of a plurality of conflicting and incommensurable reasonable views. As I understand Rawls, he is suggesting that the merits of different comprehensive views cannot be compared to one another. But I am not sure of this interpretation. See Rawls, Political Liberalism, 135.

31 See Cohen, 282.

32 A view can be coherent and yet still employ problematic modes of reasoning, e.g. most people would regard predicting the future on the basis of the reading of tea leaves to be a highly problematic mode of reasoning about the future. But most people would not say that the mode of reasoning is incoherent.

33 Perhaps Cohen could say that only well-developed exercises of reason can be reasonable. This might help with limiting the absurdity of the requirement of justification but it does so at the cost of limiting the set of doctrines that can be described as reasonable. Presumably Cohen's criterion of coherent, intelligible, and surviving critical reflection is meant to cast a very wide net. Furthermore, there are still degrees of development of views that seem to vitiate the requirement to respect all these views.

34 See Keith, Lehrer Theory of Knowledge (Boulder, CO: Westview Press 2000), 171Google Scholar for these ideas of subjective or personal justification and defeated justification. See also Gaus, Gerald Justificatory Liberalism (Oxford: Oxford University Press 1996), 31,Google Scholar for the idea of ‘open justification,’ or the idea that a person can have reason to believe a set of ideas once they subject their beliefs fully and without error to the standards they actually are committed to. Subjective justification, as I am using the term, is what Gaus calls ‘closed justification.’

35 It must be conceded that the coherentism that I have outlined can allow for divergences among the evaluative sets of persons and thus can allow for divergences in open or undefeated justifications among these persons. But these divergences are likely to be few and far between and will not explain the great majority of disagreements we see in politics.

36 Some of this argument is similar in some respects Joseph Raz's critique of Nagel in his ‘Facing Diversity,’ in Ethics in the Public Domain (Oxford: Oxford University Press 1993) 60-96, esp. 88-94.

37 See Rawls, Political Liberalism (New York: Columbia University Press 1993), 62–3.Google Scholar

38 Here I disagree with Wenar's claim that the ideas of the burdens of judgment and their implications imply skepticism, although I do share his worry that Cohen and Rawls are dangerously close to that view. Persons can conscientiously, intelligibly, and coherently believe different things when examining the same evidence, even though some of the beliefs are more reasonable than others. See Wenar, Leif ‘Political Liberalism: An Internal Critique,’ Ethics 106, 1 (1995) 32–62.CrossRefGoogle Scholar

39 Joshua Cohen, ‘Moral Pluralism and Political Consensus,’ in Copp, Roemer, and Hampton, 270-91, esp. 285.

40 The idea of respect for reason seems to be what Charles Larmore has in mind in his ‘The Moral Basis of Political Liberalism,’ Journal of Philosophy 96, 12 (1999) 602, by the notion of respect for persons. The difference in expression suggests that Larmore adheres to a rationalist conception of the person.

41 See, again, Rawls, Political Liberalism, 62-3, for the idea that the account of the reasonable cannot presuppose controversial doctrines.

42 I reject Larmore's claim that the respect for persons, as it functions in the argument for political liberalism does not imply significant part of a comprehensive doctrine. See Larmore, 623.

43 It is important to note that nothing I have said so far implies that the Kantian comprehensive liberalism is incoherent in itself. All that is being said is that this particular grounding of liberalism is incompatible with the Rawlsian aim of finding a mutually acceptable basis for social cooperation.

44 Recall here that we are talking about the need for agreement on the list of considerations. Cohen thinks, as we noted above, that all that is needed to respect others is that one bases decisions on a list of considerations everyone can reasonably accept. For resolving disagreements on the relative weights of these considerations, Cohen thinks majority rule is legitimate.

45 It should be noted here that different theorists disagree as to what the content of the principles that are the focus of reasonable agreement. Samuel Freeman seems to be inclined to think that Rawls's justice as fairness is the object of reasonable agreement. See his ‘Deliberative Democracy: A Sympathetic Account,’ Philosophy and Public Affairs 29, 4 (2000) 371-418, esp. 411. Rawls has recently broadened his account of what reasonable people could agree to. See his ‘The Idea of Public Reason Revisited,’ in The Law of Peoples (Cambridge, MA: Harvard University Press 1999), 143. He includes a family of liberal political conceptions of justice in the domain of conceptions reasonable people can agree to. This family is much smaller, of course, than the set of principles that are actively debated in modern democracies.

46 See Rawls, John ‘The Idea of Public Reason Revisited,’ 141.Google Scholar

47 See Miller, David The Principles of Social Justice (Cambridge, MA: Harvard University Press 1999),Google Scholar chs. 7-9 for a defense of the fundamental importance of desert claims. Obviously, examples of disagreement on principles of basic justice abound. People disagree on the importance of considerations of utility, self-ownership, community and many others to issues of the basic justice of society.

48 An overlapping consensus is a consensus among reasonable persons once the controversial elements of their doctrines have been excluded. It is the consensus reasonable persons are supposed to have on Cohen's view and on the basis of which terms of association are justified.

49 This kind of baseline is used in Gerald Gaus's Justificatory Liberalism (Oxford: Oxford University Press 1996), 162-6.

50 See Cohen, G. A. ‘Capitalism, Freedom and the Proletariat,’ in The Idea of Freedom, Ryan, Alan ed. (Oxford: Oxford University Press 1969), 9–25Google Scholar for an effective dismantling of this argument.

51 Gerald Gaus has proposed to me in discussion that there may be a baseline in politics that is analogous to the case of belief. Some think that in the case of belief, if we do not have good reasons for belief then we should suspend judgment. Suspending judgment is a kind of neutral baseline for belief. But I think that this is precisely what is disanalogous to politics. Political commitments cannot be suspended in the way belief can be, the society one lives in will satisfy some political commitment or another.

52 Of course, none of this problem need arise in a social world in which everyone agrees on all the basic principles of justice. As I understand Rawls, an overlapping consensus does not require full agreement on all the principles. It merely requires that there be agreement on some principles and disagreement on others. Then each person must restrain himself in trying to implement those principles on which there is disagreement. If so, then the problem I have outlined is a serious one. But if Rawls refers to the first kind of consensus as the overlapping consensus, then I contend that this is quite impossible in the world we live in.

53 I want to raise a final worry about the connection between the principle of reasonableness and democracy. In my view, we have reason to think that the principle of reasonableness is inegalitarian, Cohen's position notwithstanding. There are two worries here. First, the view that Cohen espouses excludes religious reasons and religiously based policies as well as other comprehensive moral views such as Kantianism, Utilitarianism, Communitarianism and a whole host of other positions that are commonplace in the politics of democratic societies. Second, Cohen’s account of democracy does not seem to assign any fundamental value to the inclusion of unreasonable people. Recall that the ideal deliberative procedure includes only the reasonable in its characterization of the democratic process. What reason is there then for the inclusion of the unreasonable? To be sure, the view does not directly entail the legal exclusion of these comprehensive views and the exclusion of views is only from the formal political forum of discussion that occurs in electoral campaigns and legislative debates, to name the most important. Still the inclusion of unreasonable views seems to be a concession to reality in the sense that it may be difficult to exclude unreasonable views without also excluding some reasonable views. The worry is confirmed by the very limited role Rawls assigns to citizens in a liberal democracy. He says: ‘Citizens fulfill their duty of civility and support the idea of public reason by doing what they can to hold government officials to it.’ See ‘The Idea of Public Reason Revisited,’ 136. This gives citizens something dangerously close to a merely instrumental role in the democratic process and if it does not literally do so, it is not clear why Rawls would be opposed to such a limited role.

54 I would like to thank John Christman, Joshua Cohen, David Estlund, Gerald Gaus, Andrew Williams, and the anonymous referees and editors of the Canadian Journal of Philosophy for their helpful comments on previous drafts of this paper.