FORMS OF REASONABLENESS

It is useful, in considering the question of the reasonable in politics, to distinguish two kinds of reasonableness. The first is what I call reasonableness as competence. A judgement is reasonable in this sense when it is the product of competent reasoning about the matter at hand – when it gives expression to a competent assessment of the force of the available evidence, for example. This is the kind of reasonableness at issue when people speak of a reasonable conclusion. Frequently there is an implication that other judgements could be competently reasoned as well, that a competent reasoner could reach a different conclusion. This presents us with a fundamental question about reasonableness as competence. Is competent reasoning, on the basis of adequate evidence, compatible with reaching an incorrect conclusion?

Here we need a further distinction between two different ways of understanding reasonableness as competence. On the first, reasonableness admits of degrees in a positive direction. There is a continuum of competence in the exercise of human rational capacities stretching from gross incompetence to maximal competence, and judgements are understood to become reasonable at some point on this continuum short of maximal competence. That is, there is a threshold of competence in the exercise of human rational capacities at which a judgement can be said to be reasonable, and beyond this threshold judgements can be described as more reasonable and ultimately as most reasonable. If we accept this way of understanding reasonableness as competence, it will be possible for a number of judgements that disagree all to be characterized as reasonable, even if there is a single correct answer to the question being addressed, with the result that at most one of the judgements can be correct. Further, if there are public criteria for minimal reasonableness – for attaining the threshold – it will be possible for the parties to a disagreement to agree that all the positions taken are reasonable, even though at most one can be correct.

The alternative is to suppose that competence in reasoning is to be understood as full competence. A judgement is reasonable when the mental capacities generating it are functioning optimally in the circumstances. On this way of understanding reasonableness as competence, there will still be degrees of reasonableness, but they will be degrees in a negative direction, degrees of unreasonableness. If we adopt this view, there will be a place for reasonable disagreement only if different, fully competent reasoners can extract different conclusions from the available evidence. Further, since for those who have reached one conclusion the reasoning supporting opposing conclusions will usually seem defective in some respect, reasonable disagreement will generally not be perceived as such.

For the purposes of political philosophy, however, it is necessary to distinguish a further kind of reasonableness, different from reasonableness as competence: reasonableness in the concession sense. This kind of reasonableness is restricted to cooperative contexts. It marks the fact that the pattern of concessions displayed by a cooperative venture – the concessions each participant has made, or will make, from his or her most preferred way of organizing the venture – is appropriate. The standard term for characterizing this feature of a cooperative venture is ‘fair’. Fairness is a moral value, and a cooperative venture displays this value when its organization reflects an appropriate pattern of concessions by the participants. But we also refer to people who are prepared to make appropriate concessions as reasonable. For example, we say to someone who we think is holding out for a larger than appropriate concession from the other participants, ‘Be reasonable’. So we can characterize agreement or disagreement about the appropriateness of a particular pattern of concessions in a cooperative venture as agreement or disagreement about the reasonableness of the concessions, and thus of the pattern.

The two forms of reasonableness can be combined. A judgement about what would be reasonable in the concession sense might also be reasonable in the competence sense. It might give expression to competent reasoning about appropriate concession. Similarly, there can be disagreement about what would be reasonable in the concession sense, and this disagreement might be reasonable in the competence sense. The conflicting judgements might all be competently reasoned (on a given way of interpreting this idea).

It is worth noting that reasonableness in the concession sense can play a role wherever a group is attempting to reason to a shared conclusion. Each contribution to the discussion will either be, or fail to be, a competently reasoned response to the other contributions. But when a response that accepts a point made by another party to the discussion involves the abandonment of a position previously taken, it can have the character of a concession. So if competent reasoning would dictate acceptance of the point, the failure to accept it – to abandon the previously held position – can be characterized as unreasonable in the concession, as well as the competence, sense. Similarly, a position taken in a discussion can be characterized as reasonable in both senses if, given the reasons or evidence that have been presented, the person adopting it is justified in doing so, and thus justified in declining to make the concession that would be involved in giving it up. Since we would not speak of fairness or unfairness in such contexts, the fact that the concept of reasonableness in the concession sense can be employed in this way shows that it is actually a broader notion than fairness.

We also speak of the reasonable when we want to ascribe an intermediate level of satisfactoriness to an item we have encountered. Thus, if I say, ‘This is a reasonably good piece of blueberry pie’, what I mean is that it is better than minimally acceptable, but not completely outstanding. The word ‘reasonably’ can be used in this way in conjunction with any evaluative term. Other examples are ‘reasonably good-looking’, ‘reasonably pure’ and ‘reasonably intelligent’. This usage is sometimes found in a truncated form. Thus a ‘reasonable chance’ is a reasonably good chance, a ‘reasonable fit’ (of a curve to a set of data points) is a reasonably accurate fit, and a ‘reasonable sampling’ is a reasonably comprehensive sampling. To have a label for this sense of the reasonable, I shall term it the ‘blueberry-pie’ sense. The blueberry-pie sense of the reasonable is related to the concession sense. There is an implication that taking what one has been presented with involves making a concession to reality, but one that is acceptable given the limits that reality places on achieving perfection of any kind.

RAWLS AND THE REASONABLE

We can now turn to Rawls's political liberalism. Rawls says: ‘Comprehensive doctrines play only a restricted role in liberal democratic politics. Questions of constitutional essentials and matters of basic justice are to be settled by a public political conception of justice and its public reason.’Footnote ³ When Rawls speaks of different kinds of reason, he has in mind different sets of normative concepts that can be used to formulate reasons. In the case of public reason, the reasons are those that citizens who hold different comprehensive doctrines ‘may reasonably give one another when fundamental political questions are at stake’.Footnote ⁴ Both senses of the reasonable appear to be in play here. The reasons we can reasonably give one another are reasons supporting concessions that can competently be judged appropriate. For Rawls, a political conception of justice is a repository of such reasons.

Rawls's use of the term ‘settled’ suggests that in a liberal polity of the sort he means to describe, everyone will accept the same political conception of justice. But as has been mentioned, his ultimate picture makes use of the idea of a family of liberal conceptions of justice. He characterizes the ‘fact’ of public reason as

The possibility of reaching agreement on the basis of a family of political conceptions is spelled out more fully in the following passage.

The idea, apparently, is that because all the members can agree that each of the conceptions is reasonable, they can be understood as having reached agreement despite the fact that they hold mutually irreconcilable comprehensive doctrines. But more must be said about how this reaching of agreement is to be understood.

It is not obvious that satisfaction of the criterion of reciprocity, as formulated in the quoted passage, would result in agreement on at least minimal reasonableness. Arguably anyone participating in good faith in a discussion aimed at establishing the terms of political cooperation will think it reasonable for others to accept what he proposes. But often what the proposer thinks reasonable will not be thought reasonable by those to whom the proposal is being made. Rawls appears to address this problem in another passage. He says:

If to reasonably think that others might reasonably accept our reasons is to competently expect this, and competence implies that the expectation is well-founded, the result of general reasonableness within the polity should be agreement on the at least minimal reasonableness of what each is proposing.

It is a bit of a mystery how public reasoning would proceed when the conceptions of justice on which the parties draw constitute a family of the sort Rawls describes. Reasoning grounded in a conception of justice held by another member of the polity would be recognized as having a legitimate place in the public discussion that determines how political cooperation should be organized, but how would it be integrated with one's own reasoning? If one regards another conception as barely reasonable, can reasons grounded in it carry any weight against reasons grounded in one's own conception, which one presumably regards as fully reasonable?

Rawls's remarks about the resolution of disagreement by voting are relevant here. He concedes that ‘[r]easonable political conceptions of justice do not always lead to the same conclusion; nor do citizens holding the same conception always agree on particular issues.’Footnote ⁸ He says that in this case the decision must be made by a process that all can endorse, and that ‘[f]rom the standpoint of public reason, citizens must vote for the ordering of political values they sincerely think the most reasonable. Otherwise they fail to exercise political power in ways that satisfy the criterion of reciprocity.’Footnote ⁹ If satisfaction of the criterion of reciprocity is understood as entailing agreement on the at least minimal reasonableness of what each votes for, the understanding of justice that is actually adopted – the majority view – will be regarded as at least minimally reasonable by everyone in the polity. This means that everyone will be able to regard as at least minimally reasonable the basis on which political cooperation is actually organized.

The foregoing discussion suggests a more precise answer to the question of what Rawls has in mind when he says that appeal to a family of conceptions of the kind that he describes will enable the members of a liberal society to ‘reach agreement’. If the criterion of reciprocity is satisfied, whatever conception is selected for implementation by the operative decision procedure will be regarded by everyone as reasonable, if barely so. Thus all will be able to agree that the way political cooperation is in fact organized is reasonable.

When Rawls speaks of differences about which conception is most reasonable, and of agreement that all conceptions are reasonable, although some may be found only barely so, is he employing the concept of the reasonable in the competence sense or the concession sense? Are conceptions that are barely reasonable supported by reasoning that is barely competent? Or are they conceptions that articulate barely acceptable patterns of concession?

Rawls's distinction between the reasonable and the rational, and the connection he makes between the concept of reasonableness and the concepts of fairness and justice, suggests that he understands reasonableness as a matter of appropriate concession. The distinction that he makes between the reasonable and the true, by contrast, suggests that he understands reasonableness as competence in reasoning and judgement. Of course, in formulating the criterion of reciprocity, he may be employing the concept of the reasonable simultaneously in both ways, to refer to competently reasoned conclusions about appropriate concession within a political community.

We are exploring whether the concept of the reasonable can play the role that Rawls intends. I indicated earlier that I do not believe it can. To see why, we must consider separately the notions of reasonableness as competence and reasonableness as appropriate concession.Footnote ¹⁰

COMPREHENSIVE DOCTRINES

It will be useful to begin by considering the disagreement that Rawls envisages among comprehensive doctrines, religious and secular. He says: ‘such doctrines are reasonable provided they recognize the essentials of a liberal democratic regime and exhibit a reasoned ordering of the many values of life (whether religious or nonreligious) in a coherent and consistent manner’.Footnote ¹¹ Different comprehensive doctrines provide different ways of understanding what constitutes a good human life.Footnote ¹² Thus a religious and a secular comprehensive doctrine might disagree about the moral acceptability of pre-marital sex. But their reasonableness does not consist simply in their recognition of the essentials of a liberal democratic regime, as set out by Rawls's three criteria: providing each member of the polity with certain basic rights and liberties, establishing the priority of these freedoms, and securing for each the means of making effective use of them. The doctrines are also reasonable in the sense of being competently reasoned. The disagreement among them thus constitutes reasonable disagreement in the sense of disagreement among competently reasoned judgements. This raises the question whether Rawls understands the reasoning that generates reasonable comprehensive doctrines as fully competent, or merely as attaining some threshold of competence which can be surpassed.

Rawls characterizes the diversity of comprehensive doctrines as the normal result of the exercise of human reason under conditions of freedom.Footnote ¹³ The diversity is specifically attributable to the burdens of judgement, which he describes as ‘hazards involved in the correct (and conscientious) exercise of our powers of reason and judgment in the ordinary course of political life’.Footnote ¹⁴ Some of the hazards mentioned by Rawls could play a role in shaping the reasoning of a Robinson Crusoe, but they would not in that context give rise to reasonable disagreement. The hazards include: the vagueness of concepts, the difficulty of assessing evidence because of its complexity, differences in the weight that different reasoners assign to relevant considerations, the fact that the considerations on both sides of an issue may be equally forceful, the fact that a given set of institutions may not be able to accommodate everything we regard as valuable, and the fact that the way we assess evidence and weigh values is shaped by our total life experiences, which inevitably differ.Footnote ¹⁵

Rawls speaks of the correct and conscientious exercise of our powers of reason and judgement. But he also says: ‘The account of these burdens must be such that it is fully compatible with, and so does not impugn, the reasonableness of those who disagree. What, then, goes wrong?’Footnote ¹⁶ Here ‘reasonableness’ is evidently being used in the competence sense. But if our powers of reason are being correctly exercised, how can something have gone wrong?

The diversity of comprehensive doctrines is ascribed to the exercise of reason under conditions of freedom, so it may help to focus on this idea. Rawls mentions different kinds of comprehensive doctrines.Footnote ¹⁷ One is ‘pluralistic’. Doctrines of this kind combine values associated with a political conception of justice with a large set of non-political values, producing a complex of values which must be balanced against one another. What role does reason play in connection with these pluralistic doctrines?

It does not seem that people acquire the values associated with such doctrines by reasoning that derives conclusions from premises. The values are acquired in the course of being educated to the use of certain evaluative terms. This gives us a ‘default’ understanding of the acquisition of a comprehensive doctrine. Simply learning the use of the principal evaluative terms employed in one's speech community gives one the basis of a comprehensive doctrine. The burdens of judgement can then be invoked to explain why, starting from this common body of evaluative material, people come to possess different doctrines. Because of the burdens, people extract from the common body different answers to questions about how they should live their lives. If we take this view of the matter, we can make a place for Rawls's idea that there is a connection between freedom and the diversity of comprehensive doctrines by contrasting ‘free’ learning, which allows scope for the operation of the burdens, with indoctrination, which does not.

Presumably the other, non-pluralistic, comprehensive doctrines that Rawls envisages, such as religious doctrines, can be understood in the same way. Different traditions of thought within the wider culture of a polity may be organized around discrete subsets of the evaluative concepts available in the polity. The education of some members of the community to one or the other of these traditions could be expected to produce more sharply divergent answers to questions like that concerning how to incorporate sexual motivation into a good life. The doctrines associated with such traditions might also display more of the precise articulation that Rawls says is characteristic of fully comprehensive doctrines.

This picture yields an interpretation of competent reasoning in connection with a comprehensive doctrine. When things ‘go wrong’ only because of the burdens, a certain level of competence is attained. Talk of things ‘going wrong’ might suggest that higher levels of competence are possible, at which things ‘go right’. The burdens could then be understood as marking a threshold of competence short of full competence. But Rawls also speaks of the correct exercise of our powers of reason and judgement, which suggests that he does not envisage a higher level. The result of the operation of the burdens is thus a set of comprehensive doctrines all of which can be regarded as correct in that they are the product of practical reasoning that is fully competent in the circumstances. This fits Rawls's liberal theory. The idea of a single true comprehensive doctrine plays no role in that theory. We still face the question of the sense in which things ‘go wrong’,’ but Rawls may be using this expression simply to note that the convergence on a single doctrine that might be thought desirable is not to be had. Fully competent reasoning about the good life by the members of a political community will not produce a single answer.

The understanding of reasonableness as competence that lies behind Rawls's use of the burdens of judgement to explain reasonable disagreement seems, then, to be a ‘non-threshold’ understanding. The different judgements about what would constitute a good life made within the polity are reasonable when, in addition to accommodating liberal principles, they are the products of reasoning that is fully competent in the circumstances. Our primary concern, however, is the family of liberal conceptions of justice of which Rawls speaks. All the conceptions are acknowledged by everyone in the polity to be reasonable, but there is disagreement about how reasonable they are. Here Rawls clearly envisages some kind of threshold of reasonableness. Some conceptions are seen to be less reasonable and some more reasonable.

If the forgoing observations are correct, it does not seem that the burdens of judgement can explain what Rawls has in mind here. The burdens go with a picture according to which the reasoning supporting the different judgements is fully competent, although opposing judgements will not usually be seen as possessing such support. The burdens manifest themselves in different ways in the reasoning of different members of a liberal polity. So for a member in whose reasoning the burdens are manifested in one way, reasoning in which the burdens are manifested in a different way will seem defective. This is why the operation of the burdens results in disagreement.Footnote ¹⁸ Applied to the reasoning supporting conceptions of justice, this gives us the result that opposing conceptions will be perceived by the members of a polity as unreasonable to varying degrees. Consequently, the Rawlsian criterion of reciprocity will not be satisfied. It will not be possible for society to be organized on the basis of an understanding of justice regarded as reasonable, if barely so, by everyone.

REASONABLENESS AS APPROPRIATE CONCESSION

Can we do better if we focus on reasonableness as appropriate concession? Rawls speaks of a family of liberal conceptions of justice. He also says that each member of a liberal polity of the sort he envisages will recognize all the conceptions in the family found there as reasonable, if only barely so. The disagreement among these conceptions is not, then, about whether the particular ways of organizing political cooperation being advocated are reasonable or unreasonable, but about how reasonable they are – very, or only a little. We have seen that this idea is hard to understand if reasonableness in the competence sense is at issue, and the disagreement is explained by the burdens of judgement. Does focusing instead on the concession sense yield a better understanding of the idea? Does reasonableness of that sort admit of degrees in a positive direction?

Reasonableness in the concession sense is a substantive normative concept with moral force.Footnote ¹⁹ There are substantive normative concepts that work in the requisite way. One is funny. All the members of a group can agree that a particular joke is funny, but disagree about how funny it is. Or they may agree that all the jokes that have been told are funny, but disagree about which is the funniest, with some holding that certain jokes, while funny, are ‘barely so’. But it does not seem possible to understand reasonableness in the concession sense as possessing this structure. Reasonableness of this kind does not appear to admit of a threshold beyond which there can be greater reasonableness.

This can be seen by noting the connection, evidently accepted by Rawls, between reasonableness in the concession sense and fairness, which can also be understood as appropriate concession in the context of a cooperative endeavour. Fairness is not a threshold concept that admits of degrees in a positive direction. As we employ this concept, to characterize some cooperative arrangement as fair is to characterize it as fully acceptable from the standpoint of fairness. When we say that arrangement A is fairer than arrangement B, we imply that B is unfair to some extent. A may be unfair as well, but less so. The concept of fairness thus admits of degrees in a negative direction. But there do not appear to be degrees of fairness in a positive direction beyond some threshold.

These remarks may seem merely to constitute linguistic legislation. But someone who thinks so has the task of constructing a theory of fairness that makes a place for a threshold of the required kind, a point of transition between unfairness and fairness that allows for further degrees of fairness in a positive direction. It is also worth mentioning that Rawls connects reasonableness with justice. When we characterize a way of organizing political cooperation as just, we mean that it is fully satisfactory from the standpoint of justice, not that it has attained a threshold beyond which arrangements can be characterized as more just. To say that one arrangement is more just than another is to imply that the second is unjust to a certain degree.

It might be conceded that the concepts of justice and fairness work in the way described, but denied that reasonableness has this logic. An anonymous referee has proposed that as employed by Rawls, the concept of the reasonable is not identical with the concepts of fairness and justice, but rather supervenes on them. This suggestion does not alter the conclusions that have been reached, however.

It is a feature of a supervenience relation that there can be no difference at the ‘upper’ level without a difference at the ‘lower’ level. So given the supervenience of reasonableness, in the concession sense, on fairness or justice, a political conception can be less reasonable in the concession sense than another only if it is less just or fair. And it can be more reasonable than another only if it is more just or fair. But we have seen that the only degrees of fairness and justice are degrees in a negative direction – degrees of unfairness or injustice. One political conception can exceed another in fairness or justice only if the second is unfair or unjust to some degree. The first might be unfair or unjust too, but less so. This means that one political conception can be more reasonable, in the concession sense, than another only if the second is unfair or unjust to some degree. But supervenience is usually understood as a relation of determination, and it is normatively implausible that the unfairness or injustice of a political conception could determine it to be reasonable, in contrast to unreasonable, in the concession sense. It follows that one political conception can be more reasonable than another only if the second is unreasonable to some degree.Footnote ²⁰

It remains the case, then, if reasonableness in the concession sense supervenes on fairness or justice, that it has a different logic from funniness. There is no place for a threshold of reasonableness that two different political conceptions can exceed to different degrees. It is relevant to this conclusion that unlike reasonable, funny is a response-dependent concept grounded in a positive, scalar response: amusement.Footnote ²¹

Unreasonableness in the concession sense admits of degrees. Arrangements can be marginally unreasonable or they can be grotesquely unreasonable. But once they cease to be unreasonable, there is no room for further movement in a positive direction. This means that if we are talking about reasonableness as appropriate concession, there is no room for a distinction between arrangements that are barely reasonable, to use Rawls's language, and arrangements that are maximally reasonable. Maximal reasonableness in the concession sense is the absence of unreasonableness.

Applying these points to Rawls's criterion of reciprocity, we get the result that there cannot be cases in which all the ways of organizing political cooperation proposed within a group are judged by the members to be reasonable in the concession sense, but the members differ about how reasonable in this sense they are. One arrangement can be more reasonable than another only if the second is unreasonable to some degree. In a way, every arrangement that is reasonable in the concession sense is barely so. The only room for change is in a negative direction, in which case the arrangement will be unreasonable, although perhaps only slightly so.Footnote ²² It follows that disagreement among the members of a polity about the reasonableness, in the concession sense, of the conceptions of justice held within the polity must be understood as disagreement about how unreasonable they are. But then the condition for reciprocity will not be satisfied. It will not be possible for everyone to view all the conceptions in justice in play – and in particular, the conception selected for implementation by the operative decision procedure – as reasonable, if barely so.

REASONABLENESS AND LIBERALISM

Rawls speaks of a reasonable pluralism of comprehensive doctrines, and part of what makes the doctrines reasonable is that they incorporate conceptions of justice that satisfy Rawls's three criteria for a liberal conception. It might be suggested that this gives us a further way of providing for the threshold of reasonableness that Rawlsian reciprocity requires. The competing conceptions of justice in the family Rawls envisages can be said to be reasonable in the same way that competing comprehensive doctrines are said to be reasonable. The members of the family can be said to be reasonable because they all satisfy the three criteria.

On this way of providing for a threshold, it is not, directly at least, a threshold of competence in reasoning, but of appropriate concession. Any liberal view, it is suggested, attains a threshold level of appropriate concession. The different conceptions of justice contained within the family can then be seen as constituting different interpretations of the fundamental understanding of appropriate concession codified by the three criteria. The partisans of each conception will presumably regard it as providing a better interpretation of the fundamental understanding than the competing conceptions, but they will still be able to view the opposing conceptions as reasonable by virtue of the fact that they satisfy the three criteria. So the partisans of each conception will be able to judge opposing liberal conceptions less reasonable than their own without judging them unreasonable.Footnote ²³

I do not believe, however, that this proposal is successful. It creates two different possibilities. Neither yields the desired result by itself, and they cannot be combined.

On the first, the members of the polity develop different interpretations of liberalism – different conceptions of justice satisfying the three criteria – and the partisans of each conception judge it to be more reasonable than the opposing conceptions. But given the conclusions reached in the previous section, it must be the case that in judging their own interpretation more reasonable than the others that have been advanced, the partisans of a particular conception judge these others to be unreasonable to some degree. Non-liberal views will be judged more unreasonable. But there will be no bright line of ‘bare’ reasonableness separating the rejected liberal views from non-liberal views. So although this possibility gives us disagreement within the polity about the correct way to interpret the three criteria, it does not allow the members to judge reasonable, in contrast to unreasonable, the interpretations they reject. Consequently, it does not allow them to judge the rejected interpretations barely reasonable.

If we want to avoid this result, we can hold that all liberal conceptions, all conceptions that satisfy the three criteria, are reasonable in the concession sense. This will mean that the three criteria establish a threshold of reasonableness in the concession sense. But (again) as we saw in the previous section, reasonableness in the concession sense does not admit of degrees in a positive direction. To say that a conception of justice is reasonable in this sense is to say that it is free of all unreasonableness. It follows that if all liberal conceptions are reasonable in the concession sense, they must be equally reasonable – equally free of unreasonableness. This means, however, that it will not be possible for a competently reasoning member of a liberal polity to judge her conception more reasonable than the others that have been advanced.

It might be argued that this does not matter. Rawlsian reciprocity will still obtain. The particular conception of justice chosen for implementation, whichever it is, will be accepted as reasonable by everyone. But the resulting picture will lack important features of a Rawlsian liberal polity. It will be a matter of indifference to the parties, in so far as they are concerned with reasonableness, which conception of justice is chosen for implementation. Further, there will be no place for public reasoning, understood as reasoning undertaken on the assumption that some conceptions of justice are more worthy of acceptance than others. Or to be more precise, there will be a place for such reasoning only in connection with the identification and rejection as unreasonable of non-liberal conceptions of justice, conceptions that may appear to satisfy the three criteria, but actually do not.

This argument has been presented in an abstract way, so it may help to make the issues more concrete. In On Liberty, one of the founding texts of modern liberalism, John Stuart Mill says: ‘In politics, again, it is almost a commonplace, that a party of order or stability, and a party of progress or reform, are both necessary elements of a healthy state of political life.’Footnote ²⁴ We can speak here of progressive and conservative parties. Will we find this distinction in a Rawlsian polity?

Mill continues: ‘until one or the other shall have so enlarged its mental grasp as to be a party equally of order and of progress, knowing and distinguishing what is fit to be preserved from what ought to be swept away’. But it is doubtful that in a Rawlsian polity the progressive/conservative dichotomy would be transcended in this fashion. Rawls says that there are many liberalisms, and that each ‘endorses the underlying ideas of citizens as free and equal persons and of society as a fair system of cooperation over time’.Footnote ²⁵ It seems that conservatives and progressives could both claim that their conceptions of justice satisfy this description. If this is correct, however, it is unlikely that the conditions for Rawlsian reciprocity will obtain. It can be expected that the two parties will still disagree, especially about what is required to ensure that all are provided with means for making effective use of their freedoms. And in disagreeing, each will judge the patterns of concession proposed by the other to be unreasonable to some extent. That is, the first of the two possibilities distinguished above will be realized.

There is one other interpretation of Rawlsian reciprocity that we should consider. Earlier I described what I called the ‘blueberry-pie’ sense of reasonableness. Rawls often employs the notion of the reasonable in this sense. The following passage contains two of these constructions: ‘What would a reasonably just constitutional democracy be like under reasonably favorable historical conditions . . .?’Footnote ²⁶ Rawls also speaks of a reasonably just government, characterizing it as one that is effectively under the people's political and electoral control.Footnote ²⁷ He contrasts a reasonably just government with a fully just government, the point apparently being that the people may not always use their control in a way that is optimal from the standpoint of justice. On my interpretation of such statements, in characterizing a government as reasonably just, Rawls is saying that although it is marked by some injustice, this is acceptable, provisionally at least, given the limits that reality places on achieving perfection of any kind.

Used in this way, the concept of the reasonable makes room for variation in a positive direction. We might be able to construct institutions that are more accommodating of our concern to live in just conditions than those already deemed reasonably just. Similar points can be made about reasonableness as competence. The reasoning supporting a particular judgement might be characterized as reasonably good, but reasoning that is reasonably good can still be improved upon.

It does not appear, however, that in formulating his concept of reciprocity, Rawls meant to employ the notion of the reasonable in the blueberry pie sense. As was mentioned earlier, the blueberry pie sense of the reasonable is tied to the concession sense. It marks a concession that is being made to reality, given the limits that reality places on achieving perfection of any kind. But when Rawls speaks of the barely reasonable in connection with conceptions of justice, he doesn't seem to be envisaging a maximally acceptable concession to reality – a maximally acceptable level of injustice. He seems to be envisaging the instantiation, to a minimal degree, of a positive normative property of some kind.

If the appropriateness of accepting a conception of justice as a basis for political cooperation is to be understood as the appropriateness of making the concessions that acceptance requires, it is more natural to think of these concessions as being made not to reality, but to the other members of the polity. This, however, points us towards an understanding of the role of the reasonable in establishing the terms of political cooperation that is different from the one we have been exploring, an understanding that does not require the members of the polity to judge reasonable, to some degree, each of the conceptions of justice that could be selected for implementation.

THE HIDDEN ROLE OF CONCESSION

If the argument so far is correct, the idea that conceptions of justice can be characterized as barely, more or most reasonable is problematic no matter which kind of reasonableness is involved, reasonableness as competence or reasonableness as appropriate concession. But many people, including Rawls, seem to be comfortable speaking this way. How might this be explained? The most plausible explanation, I believe, involves a confusion of the competence and concession senses of the reasonable.

In Political Liberalism, Rawls makes the following observations about the objectivity of judgements of political morality:

Here Rawls clearly envisages the operation of the burdens of judgement in connection with reasoning about justice. The burdens may prevent agreement, yet are compatible with a narrowing of differences. But how can a narrowing of differences produce a result that strikes all the members of a polity as just or fair? The differences presumably concern what is just or fair. So opposing views will still be seen as unjust or unfair to some degree.

Presumably what happens when differences are narrowed by shared deliberation is that opposing views seem less mistaken about justice, or the reasoning supporting them less defective, than was the case initially. Under these circumstances, reasonableness in the concession sense would normally dictate convergence on a particular view. Since opposing views are still competently judged unjust or unfair to some degree, it would not be unreasonable in the competence sense for each member to insist that cooperation be organized on the basis of his or her own view. But if the flaws perceived in the opposing views are not great, it would be unreasonable in the concession sense for the members to insist on their own views. Agreement is necessary if political cooperation is to be instituted, and if everyone is reasonable in the concession sense, agreement can be secured. The particular conception on which the members converge will not be thought reasonable in either sense by every member of the polity, but convergence will take place because resolving disagreement about what justice requires by mutual concession at a higher level is the reasonable course in the circumstances.

In a situation of this sort, a failure to distinguish reasonableness as competence from reasonableness as appropriate concession could create the impression that a choice was being made among conceptions of justice all of which were reasonable in the competence sense – all of which could be supported by reasoning exceeding some threshold of competence. Only if this were so, it might be thought, could a member deem acceptance of a conception other than her own compatible with her status as a reasonable person (in the competence sense). Perhaps Rawls was thinking in this way when he formulated his criterion of reciprocity. But once it is understood that the higher-order concessions associated with the making of such a choice can be reasonable in the concession sense even for someone who competently judges the chosen conception unjust or unfair to some degree, the need to construe the situation in this way disappears.

This interpretation of Rawls's remarks is compatible with his talk of ‘honorable or decent’ relations.Footnote ²⁹ But a group's proceeding in the manner described would not support the conviction that there are reasons of the sort Rawls characterizes as objective, reasons sufficient to convince all reasonable persons that a particular political conviction is reasonable. Rather the opposite. Resort to reasonableness as appropriate concession is necessary precisely because, in the case at hand, there are no reasons of that kind. In a large-group setting such as a polity, the basis of cooperation would normally be established by a procedure, such as voting by the method of majority rule, widely regarded in the group as providing a reasonable, in the concession sense, way of resolving disagreement.

COHEN ON TRUTH AND PUBLIC REASON

Joshua Cohen has discussed the place of truth in Rawls's political liberalism.Footnote ³⁰ He argues that a distinctive, political conception of truth, which retains the contrast between truth and warrant, has a legitimate place in the public reasoning that will be found in a Rawlsian liberal polity. His point is that the members of such a polity should seek to identify conceptions of justice that are not merely reasonable, but true.

In the passage concerning objectivity quoted above, Rawls speaks of reasons sufficient to convince all reasonable persons that a political conviction is reasonable. It seems likely that he has in mind both reasonableness as competence and reasonableness as appropriate concession. A reasonable political conviction is a view of appropriate concession that can be supported by competent reasoning about appropriate concession. Cohen appears to be advocating that we instead understand objectivity, in so far as it characterizes public reasoning, in terms of reasons sufficient to convince all reasonable persons that a political conviction is true. In deliberating about justice, we should aim not simply to identify a conception that can be supported by competent reasoning, but one that is true.

If reasonableness in the competence sense is understood as admitting of degrees in a positive direction beyond some threshold, Cohen's claim that it requires supplementation with a separate notion of ‘getting it right’ has considerable plausibility. Deliberation about reasonableness in the competence sense, on the threshold interpretation, must presumably be concerned with whether the threshold has been attained or surpassed, and if surpassed, by how much. It is unclear how such deliberation would proceed. If Cohen's aim is to avoid this problem, however, ‘getting it right’ does not have to be understood in terms of truth. It can be understood in terms of reasonableness on the ‘full competence’ interpretation. Deliberation undertaken to identify a political conviction that can be supported by fully competent reasoning would proceed in the same way as deliberation undertaken to determine truth. It would proceed by careful assessment of the force of all the available reasons and evidence. The main difference would be that in labelling the result reasonable rather than true, we would signal a greater openness to the possibility of reasonable disagreement.

Cohen argues that a focus on truth need be no more divisive than a focus on the reasonable. Opposing views will be seen as deviating from the truth and thus as false. But the deviation will admit of degrees, and may be perceived as relatively minor. Opposing views may be seen as ‘close to the truth’ or ‘approximately true’.Footnote ³¹ The idea appears to be that under these circumstances, the convergence required for political cooperation will be possible.

But this convergence cannot be understood as dictated by competent reasoning about justice. Pragmatic considerations aside, competent reasoning cannot take as its basis a proposition that is judged to be approximately true when a proposition that is judged to be (simply) true is available. Applying this point to political judgement, we get the result that a general perception that opposing conceptions of justice are approximately true will preserve divisions.

It might be replied that pragmatic considerations will in fact be operative when a group has to generate a conception of justice capable of providing a basis for political cooperation. But these will be considerations of reasonableness as appropriate concession. A conception of justice is a conception of appropriate concession in a political community. But when the members of a polity regard opposing conceptions as approximately true, and agreement on a conception is required if the benefits of political cooperation are to be achieved, concession that brings with it mutual accommodation can be instituted at a higher level. If the differences truly are small, the work done by reasonableness in the concession sense may not be noticed, with the result that the parties have the erroneous impression that they have reached an agreement about how to organize political cooperation on the basis of reasoning about truth.

This way of securing political agreement is not, however, limited to situations in which opposing conceptions of justice are seen as approximately true. Reasonableness in the concession sense can dictate cooperation on the basis of a conception that appears to be seriously defective if attempts to reason to agreement about justice have been unsuccessful, and a general refusal to make concessions would result in a failure to realize the benefits of political cooperation. To say this is simply to say that political cooperation under realistic conditions, where both conservative and progressive conceptions of justice will be in play, can be reasonable in the concession sense. There will still be limits on the concessions that reasonableness can require of the members of a polity. The concessions necessary to implement certain political programmes may exceed what reasonableness in the concession sense can demand. To take an extreme example, a programme may fail to give expression to a conception of justice of any kind, as would be the case if it envisaged enslaving or exterminating part of the population. But reasonableness in the concession sense is compatible with – indeed dictates – a fair amount of flexibility.

Cohen emphasizes the singularity of truth. Public reasoning about justice should aim to identify a single correct conception of justice. But if the burdens of judgement give rise to reasonable disagreement among comprehensive doctrines, it seems likely that they would produce the same result in the context of deliberation about justice. Fully competent reasoning about justice by the members of a polity would not converge on a single conception of justice. Rather it would yield a number of different, incompatible conceptions. A general awareness that judgements of justice are susceptible to the burdens could, however, facilitate the process of second-order concession that I have suggested will normally be required to establish a basis for political cooperation.