Click to increase image sizeClick to decrease image size Acknowledgements I would like to thank participants at the ALSP conference 2003 where this paper was presented. I am very grateful to Peter Jones and Simon Caney for their helpful comments and suggestions. I also wish to thank the British Academy for the support of a Postdoctoral Fellowship during the last few months. Notes This essay makes use of ideas very similar to those in Christopher Bennett's essay in this volume: the liberal commitment to justification, the notion of conclusive justification, and the problem posed for liberalism by difficulties in public justification. However, his understanding of what conclusive justification is and of how it would be obtained in moral matters is rather different from mine, as I shall indicate below. For reasons of space, I shall use the notion of imposition or enactment or coercion loosely, but I accept that more work needs to be done to outline the relation between these more exactly. Of course, as I go on to say, one good reason for insisting on justification is that, in these circumstances, adopting a policy ceases to be imposition, at least on some accounts of justification. Fred D'Agostino (1996 D'Agostino F 1996 Free Public Reason: Making it Up as We Go Along New York: Oxford University Press [Google Scholar]) provides a list of six ‘stories’ that can be told about the need for public justification. The kinds of reason I talk about are very similar, though my notion of ‘respect for persons’ covers categories that D'Agostino would maintain are distinct. Rawls, for example, insists that a ‘public and shared basis of justification’ would be required ‘to mark the difference between comprehensive beliefs as such and true comprehensive beliefs’. It is unreasonable, on Rawls' account, for people to ‘insist … on what they take to be true but others do not’ (Rawls 1993 1993 Political Liberalism New York: Columbia University Press [Google Scholar]: 61). When we look at arguments about the instrumental or intrinsic value of justification, the two will become blurred. Because both arguments can and should be advanced simultaneously, I do not consider this a serious problem. I am thinking here of one of David Wong's arguments in particular. Wong, looking at the works of J.S. Mill, argues that the justification principle is a way of protecting what Mill would conceive as a fundamental component of well being: individual autonomy. ‘When an individual's freedom is restricted in such a way that it cannot be justified to him or her, the kind of growth that Mill thought essential to well‐being is frustrated. It is reasonable to lay down the justification principle to protect the general welfare’ (Wong 1986 Wong DB 1986 Moral Relativity Berkeley: California University Press [Google Scholar]: 184.) I am unwilling to rely on illustrative examples here, because any case I chose could be controversial, depending on how the justification principle was specified. Nevertheless, one case I have in mind is that in which liberals weigh up whether to interfere with a particular culture's harmful practices. My work here bears some similarity to Catriona McKinnon's discussion in Liberalism and the Defence of Political Constructivism. McKinnon notes the importance of debates about the internal/external character of justification and discusses how conclusive justification ought to be characterised (McKinnon 2002 McKinnon C 2002 Liberalism and the Defence of Political Constructivism Basingstoke: Palgrave Macmillan [Crossref] , [Google Scholar]: 14–24). McKinnon's aim is to defend a self‐respect‐based account of constructivism. I move more quickly in this essay than McKinnon and my objectives are more modest, aiming only to indicate the problem for liberalism arising from the importance and difficulty of justification. Depending on what is involved in justification, I remain unsure that the move to self‐respect solves this problem. Bergmann, for example, argues that we can distinguish internalism/externalism about justification from internalism/externalism about warrant (Bergmann 1997 Bergmann, M. 1997. ‘Internalism, externalism, and the no‐defeater condition’. Synthese, 110: 399–417. [Crossref], [Web of Science ®] , [Google Scholar]: 401–406). This rejection need not deny that there are objective facts. It just holds that reasoning about those facts, rather than the bare facts themselves, does the work. David Wong, for example, takes this strategy by arguing for relativism as the best explanation of people's moral experience (Wong 1986 Wong DB 1986 Moral Relativity Berkeley: California University Press [Google Scholar]). In the case of Nagel, my view (in outline) is that his discussion of disagreement is undermined by his unwillingness to show the same kind of scepticism about cases of apparent agreement. For Nagel, these latter cases offer unambiguous evidence that there are universally good moral reasons, whereas the former are largely dismissed (Nagel 1986 Nagel T 1986 The View From Nowhere Oxford: Oxford University Press [Google Scholar]: 147–9). With this in mind, I have some sympathy with Bergmann's view that many externalists do not really present externalism as a position about justification (Bergmann 1997 Bergmann, M. 1997. ‘Internalism, externalism, and the no‐defeater condition’. Synthese, 110: 399–417. [Crossref], [Web of Science ®] , [Google Scholar]: 403). I am not sure, however, that the universality of reasons is indeed something we are committed to. Consistent relativists, whatever else they might be accused of, do not affirm this aspect of ‘our’ moral experience. Given the (oft‐bemoaned) prevalence of ‘freshman relativism’, it might seem there are a number of moral reasoners who would deny Scanlon's claim. Korsgaard, for example, opts for an ‘internalism requirement’, which holds that proposals have to be justified to others were they rational (Korsgaard 1996 Korsgaard CM 1996 Creating the Kingdom of Ends Cambridge: Cambridge University Press [Crossref] , [Google Scholar]: 320). Scanlon, it seems to me, reaches a similar compromise (Scanlon 1998 Scanlon T 1998 What We Owe to Each Other Cambridge, MA: Harvard University Press [Google Scholar]). Christopher Bennett seems to invoke the idea of accessibility to explain conclusive justification – the idea that the reasons in question can be shared or appreciated by others. On my account, this is a necessary condition for a compelling justification, but not a sufficient one. We can see the relevance of someone's reasons, but still reject them as inconclusive in respect of ourselves. Adding up several less than compelling justifications does not, to my mind, make one compelling one. As Gaus notes, this kind of argument can be run in reverse, so that those who affirm internalist accounts of justification ought to be liberals (Gaus 1986 1996 Justificatory Liberalism New York: Oxford University Press [Google Scholar]: 263). D'Agostino, for example, picks out seven respects in which liberal accounts of public justification can differ: consensual/convergent, positive/negative, volitional/cognitional, direct/indirect, economic/political, prior/posterior, interactive/non‐interactive (e.g. D'Agostino 1996 D'Agostino F 1996 Free Public Reason: Making it Up as We Go Along New York: Oxford University Press [Google Scholar]: 33). Christopher Bennett's essay offers an alternative approach of mutual engagement as a way of thinking about public debate in liberal societies. This takes the form of a frank exchange of the best justification each side in the debate thinks it can muster. However, in a society characterised by diverse schemes of value, and taking the account of justification I have outlined here, demonstrating the truth or falsity of someone's position is problematised. Bennett suggests that Donna in his case clings to the belief that her reasoning is compelling, and hence that everyone else must accept it. This seems to me to illustrate neatly why the foundation of political liberalism is non‐neutral, for the liberal has to believe that Donna is wrong. For political liberalism to ‘work’, the disagreement must be reasonable disagreement, and has to be accepted as such by the parties involved. Rorty seems to think that the liberal philosopher's preoccupation with justification is unprofitable, because ‘such a philosopher is not thereby justifying these institutions by reference to more fundamental premises but the reverse; he or she is putting politics first and tailoring philosophy to suit’ (Rorty 1991 Rorty R 1991 Objectivity, Relativism and Truth Cambridge: Cambridge University Press [Google Scholar]: 178). For principled compromise, see Macedo 1991 Macedo S 1991 Liberal Virtues: Citizenship, Virtue and Community in Liberal Constitutionalism Oxford: Clarendon Press [Google Scholar]. For appeals to an impartial judge, see Gaus 1996 1996 Justificatory Liberalism New York: Oxford University Press [Google Scholar]. Both of these approaches will get us only so far by themselves, because they rest on agreement of all parties, and thus on the justification offered for such measures. People will have to be concerned enough about disagreement to implement such a solution, but not concerned in a way that makes this solution seem unattractive. We might think, for example, that the more fundamental the subject at issue, the less rational it is to subordinate our beliefs to the adjudicator, or to compromise. The result might be a split between public and private conscience that would threaten to undermine rather than maintain stability. More generally, people might question the impartiality of the decision procedure or of the judge. I use the term conscientious disobedience here as an umbrella term to include civil disobedience, conscientious objection and the like. By a content‐neutral criterion I mean one that does not involve passing judgement on the content of the disagreement or cause involved.