1

In my estimation there has been a lot of unnecessary confusion about where the issues in the dispute between Rawls and Habermas lie.Footnote ¹ That confusion extends both to the relevant secondary literature and to the protagonists themselves in their respective contributions to the debate.Footnote ² A considerable part of the actual debate, and of its interpretation and analysis by commentators in the scholarly literature, was derailed by the assumption that the salient point of comparison between the respective political theories of Rawls and Habermas lies in their different ways of operationalizing ‘the moral point of view’, or in their respective ‘devices of representation’, namely the original position and principle (U).Footnote ³ But that is not where the interesting points of the debate lie, and little is to be gained from taking these as the relevant points of contrast. The interesting points of dispute concern their respective accounts of the relation between morality and legitimacy, and arbitrating that dispute can throw light on what is still a very important question in political and legal theory.

That the existing literature has been preoccupied, to the neglect of other issues, with the comparison between the original position and principle (U) is beyond question. This is probably because Habermas kicks off the debate by levying three criticisms against Rawls’s idea of the original position:

1. (i) The rational egoists in the original position are not able to stand in for, and model, the interests of ‘autonomous citizens’ and to understand ‘the meaning of the deontological principles they are seeking’ (RPUR 28).

2. (ii) Rights ‘cannot be assimilated to primary goods without forfeiting their deontological meaning’, hence the device of the original position does not represent an adequate procedure for the articulation of principles of justice (RPUR 29).

3. (iii) Rawls smuggles substantive considerations into the original position and does not develop it in a ‘strictly procedural manner’. The theorist must already anticipate some of the information that is meant to flow in at the later stages ‘of framing the constitution, of legislation, and applying the law’, which defeats the point of the veil of ignorance (RPUR 33).

Let me explain why I think these objections miss the point. To begin, the first two objections show that Habermas tends to construe A Theory of Justice (hereafter Theory) as a general moral theory, vying with Kant’s deontological ethics and utilitarianism as a general theory of right conduct. This in fact is how Lawrence Kohlberg reads Theory and Kohlberg’s work on the psychology of moral development exerted a significant influence on Habermas as he was developing discourse ethics.Footnote ⁴ Habermas’s talk here of ‘deontological principles’ and ‘deontological meaning’ indicates that he construes A Theory of Justice as a general moral theory. This interpretation is confirmed by his presentation of Rawls’s device of choosing principles in the original position, behind a veil of ignorance, as a rival conception of the moral standpoint to discourse ethics. No doubt Rawls himself was partly responsible for ‘Justice as Fairness’ being so interpreted. In his earlier work he was keen to present his view as Kantian, implying that the original position is the operationalization of the moral point of view, like the Formula of the Universal Law, and that ‘Justice as Fairness’ can be interpreted as a general theory of right conduct.Footnote ⁵ As he later admitted: ‘In Theory a moral doctrine of justice general in scope is not distinguished from a strictly political conception of justice’ (PL xvii).Footnote ⁶

The later Rawls dispels the idea that he is advancing a ‘moral doctrine of justice general in scope’ and rejects the view that this is an appropriate task for political philosophy. By contrast, the phrase ‘moral doctrine of justice general in scope’ perfectly describes Habermas’s discourse ethics, which is a moral theory in the sense that it is a theory of right conduct which sets out to rationally reconstruct (and to justify) the moral standpoint, or principle (U). Accordingly, in Habermas’s work ‘justice’ is equivalent to moral rightness: it is the central normative concept of the discourse theory of morality and the central phenomenon to be studied.Footnote ⁷ Principle (U), the moral principle, has universal scope, and covers the whole range of interpersonal conflicts of action that can be appropriately regulated by appeal to valid moral norms. Unlike Rawls’s two principles of justice, (U) is not tailored only to questions concerning the basic structure of society.

Now, thinking of the original position and (U) as the central points of contrast of the Habermas–Rawls debate would make interpretative sense only if the debate were one between Theory construed – in my view misconstrued – as a doctrine of justice general in scope, and discourse ethics understood – correctly in my view – as a deontological moral theory. That, however, gets the whole context of the debate wrong, and so deflects attention from the real issues.

This is the proper context. The Habermas–Rawls debate took place in 1995, three years after the publication of Faktizität und Geltung (Habermas Reference Habermas1992; Engl. trans. 1996) and two years after the publication of Political Liberalism.Footnote ⁸ The former is a theory of democratic legitimacy and the rule of law, and the latter is a defence of justice as a political conception. The full title of the former in English is Between Facts and Norms: Contributions to a Theory of Law and Democracy. The focus of the 1995 exchange is evident from the titles of their respective contributions: Habermas’s ‘Reconciliation through the Public Use of Reason: Remarks on John Rawls’s Political Liberalism’ and Rawls’s ‘Political Liberalism: Reply to Habermas’. Rawls emphasizes that the theory of political liberalism, which the debate was about, falls entirely in the domain of the political (RH 47). Habermas, though he has a different conception of the political to Rawls, accepts these terms. All of which is not a knock-down argument against those interpreters who construe the dispute as one between two moral theories or ‘two procedures of moral thinking’, but it does constitute telling evidence against it.

The second thing to note about Habermas’s opening objections is that by straightaway contrasting the Original Position with principle (U) he appears to imply that these contrasting ideas lie at the heart of the dispute.Footnote ⁹ But it should strike one as quite puzzling why Habermas would initiate a discussion of Rawls’s Political Liberalism with criticisms aimed at the Original Position, which is not a central component of that theory. I myself think he does this because, at the time of writing – as he later makes clear – he did not fully appreciate the differences between Political Liberalism and Theory (Habermas Reference Habermas2011: 284).Footnote ¹⁰

Interestingly, in his Reply to Habermas Rawls describes the difference between their respective ‘devices of representation’ by implication as the second and least fundamental difference between them. The first, more fundamental difference – which, he claims, ‘frames the second’ – is in Rawls’s words that Habermas’s theory ‘is comprehensive while mine is an account of the political and is limited to that’ (RH 47). Rawls is right not to see the difference between the Original Position and principle (U) as fundamental. As we have seen, Habermas’s principle (U) is the central principle of discourse ethics, and not the centrepiece of his political theory. That honour goes to what he calls the Principle of Democracy, which states:

Furthermore, it is fundamental to the design of Habermas’s political and legal theory that the ‘principle of democracy’ does not depend on, or derive from, the moral principle (U), nor indeed from any moral principle, moral theory, or any actually existing morality.

The central argument of Between Facts and Norms is that the principle of democracy ‘derives from the interpenetration of principle (D) and the legal form’.Footnote ¹² Principle (D) states that ‘only those norms are valid that could meet with the assent of all those potentially affected, insofar as they participate in rational discourse’.Footnote ¹³ It is held to be a rule of practical argumentation in general, i.e. one that applies to all types of practical discourse (MCCA 66). It is supposedly anchored below the threshold of morality (and politics), in the deep structure of social action, and as such expresses a necessary condition of the validity of action-norms in general – political, legal and moral norms.Footnote ¹⁴ It is, Habermas claims, ‘neutral’ or ‘initially indifferent vis à vis morality and law’ (BFN 107, 121). This is what Habermas means when he later claims – with a nod towards Political Liberalism – that the principle of democracy is ‘morally freestanding’ (BNR 80).

The legal form, for its part, is a matter of institutional and empirical fact. So the structure of Habermas’s theory, the relation between the three principles, reflects Habermas’s view that in modern societies political legitimacy does not derive from moral norms. The Principle of Democracy is supposed to rationally reconstruct ‘the democratic procedure for the production of law’, which is the sole source of legitimacy in modern Western societies.Footnote ¹⁵ In other words, according to Habermas, the normative force (and authority) of law is sui generis; it is not borrowed from the normative force and authority of morality. That, he argues, is the mistake that thinkers such as Kant, natural law theorists, and K.-O. Apel make (BFN 103–5, 104–5; BNR 78). Rather the normative force of law is borrowed from ‘the validity basis of action oriented to reaching understanding’ (BFN 297). This is not just an abstract theoretical point, namely a feature of Habermas’s reconstruction of democratically legitimate law. It is also, Habermas maintains, a feature of the actual systems of legitimate law that his theory reconstructs.

Rawls’s central idea in Political Liberalism is to formulate and to present as justified a political conception of justice, where what makes the conception political is that it is freestanding from any comprehensive moral or philosophical doctrine.Footnote ¹⁶ Thus in Political Liberalism Rawls presents and reconstrues ‘Justice as Fairness’ as a ‘module, an essential constituent part that fits into and can be supported by various reasonable comprehensive doctrines’ but that, importantly, is not dependent on any single comprehensive doctrine or any particular combination thereof (PL 12). A political conception also has to meet two further criteria: scope specificity – it only addresses issues of justice that apply to the basic structure – and publicity – it should also derive its content from the fundamental ideas contained in the public political culture (PL 11, 450–2). The latter condition is not fully met by the conception’s being part of an ‘overlapping consensus’ of reasonable comprehensive doctrines, at least not if the overlapping consensus is thought of as a serendipitous overlap, like the intersection area of a Venn diagram. For such an overlap may contain coincidentally shared preferences, or values whose universality arises, say, from hard-wired aspects of human nature, and it might not be permissible for a liberal state to legislate on the basis of such preferences or values, merely because they are shared by all citizens. For example, it might lead to overly invasive and needlessly prescriptive laws. Which is only to say that not everything in the overlapping consensus so conceived will be a political value – although no value that is not part of the overlapping consensus can count as a political value. Rather the values that are properly political, and meet the publicity condition, will be those values that form part of the overlapping consensus of reasonable doctrines that are ‘realized in and characterize political institutions’ and which ‘others, as free and equal citizens, might also be reasonably expected reasonably to endorse’ (PL 453, 450).

The salient differences from Theory stem from the fact that the focus of Political Liberalism is not on justice per se, but on justice as a political conception, which meets the conditions stated above and is thus apt to play a role in the determination of political legitimacy for a democratic constitutional liberal state. This is because the chief aim of the later work is to make a case for constitutional liberal democracy, i.e. for a régime in which the legitimacy of laws and policies flows from its having a liberal constitution and a democratic form of government, rather than for liberalism or democracy as such.Footnote ¹⁷ Crucially, both the theory provided by Rawls and the particular form of government it sets out to explain and defend have to operate, and to be made acceptable, under conditions of reasonable pluralism, which means that they can be welcomed as reasonable by free and equal citizens, whatever their comprehensive doctrine.

One consequence of his making the theory of political liberalism itself (not just the political régime it is a theory of) responsive to the fact of reasonable pluralism is that Rawls has to allow that ‘Justice as Fairness’ is only one among several possible political conceptions of justice (PL 450).

All of this dislodges the original position from the central place it occupied in Theory.Footnote ¹⁸ In Political Liberalism it is merely one device (among others) for showing that the principles of justice may be endorsed by free and equal citizens in the light of principles and ideals acceptable to their common human reason, and thus that ‘Justice as Fairness’ is a political conception.Footnote ¹⁹ To sum up, just as principle (U) is not a central organizing idea of Between Facts and Norms, so the original position is not a central organizing idea of Political Liberalism, and consequently discussions of the Habermas–Rawls debate which take this as the salient contrast are not engaging the real issues. In spite of what many commentators have assumed, the heart of the dispute lies elsewhere.Footnote ²⁰

2

I maintain that the Habermas–Rawls debate is in essence a dispute about the nature of legitimacy and the production of legitimate law in liberal constitutional democracies.Footnote ²¹ The central question they both raise is how a democratic liberal state can produce legitimate laws and policies which, to put it in Habermas’s terms, can elicit enough quasi-voluntary compliance to law to maintain social integration, or, to put it in Rawls’s terms, can evince ‘stability for the right reasons’ (PL xlii). But to identify the issues that divide them we need to examine the different answers they respectively give to this question, paying particular attention to areas where one theorist denies something significant that the other affirms. Doing so will allow us to zero in on two important and related sets of issues that divide them.

The first and I believe most significant is, to put it briefly and rather too crudely, the question of how the moral stands to the political. In other words, what role does morality (suitably specified) play in the determination of legitimate law and policy in a liberal constitutional democracy? The second set of issues concerns how the arguments of political philosophy stand in respect to the arguments by the representatives of the state, or by citizens, for a liberal constitutional democratic régime. In other words how do the reasons adduced by the political theorist in support of her theory stand with respect to the reasons offered by the state to its citizens, and by citizens to one another, in support of its power to coerce them, and of other related matters such as the extent of its reach, the appropriateness of its objects, and so forth?

These two sets of issues are broached in the second group of criticisms that Habermas makes in RPUR, and they gradually come into focus in Rawls’s response to them in RH and in Habermas’s subsequent reflection on that response in RTMW. In section 2 of RPUR Habermas makes the following three criticisms: Rawls collapses the distinction between justified acceptability and actual acceptance; he thereby fails to do justice to the ‘epistemic meaning’ of his own theory; and he reduces practical to instrumental reason by subordinating the political conception to the aim of securing stability (RPUR 36–7). Some of these objections are clearly based on misunderstandings.Footnote ²² Be that as it may, it is these objections that prompt Rawls’s remark that the chief difference between his approach and Habermas’s is that ‘his is comprehensive, while mine is an account of the political and is limited to that’ (PL 47).

What exactly Rawls means by this remark is hard to discern, because he uses ‘comprehensive doctrine’ in at least three different senses, to mean: (a) worldviews, religious and secular; (b) actually existing moralities or conceptions of the good; (c) philosophical theories of one kind or another, including moral theories such as utilitarianism or Kantianism. Further, all of these three types of comprehensive doctrine have one feature in common, which is that they are all subject to reasonable doubt, because there are viable alternatives that reasonable people might endorse. Consequently, Rawls’s counterargument can be developed in at least three different ways. The way Rawls actually goes on to develop it in RH shows that the objection to Habermas he has in mind is that discourse theory takes too many theoretical hostages to fortune – i.e. it is ‘comprehensive’ in sense (c) – whilst Rawls’s theory ‘leaves philosophy as it is’ (RH 49–51; cf. 48).

Rather than deny the accusation that his theory is ‘comprehensive’, Habermas defends himself by arguing that political theories have to take a stand on some philosophical questions which, while beyond the domain of the political, nevertheless are not ‘metaphysical’ in any objectionable sense.Footnote ²³ Habermas considers his own discourse theory to be ‘post-metaphysical’ in the sense that it has been purged of extravagant ontological assumptions, that it consists in defeasible hypotheses, that it begins from empirical assumptions and relates to empirical inquiry through the reconstructive method.Footnote ²⁴ He takes this tack presumably because he is cognizant of the fact that his democratic and legal theory in BFN does indeed presuppose the truth of discourse ethics, and some related aspects of the Theory of Communicative Action.

There is another reason too. Habermas believes it better to be upfront with his philosophical commitments, because he does not think that the political philosopher has to respond to the fact of reasonable philosophical pluralism, that is, to the fact that his theory is open to reasonable doubt, in the same way as the legislator and the citizen do. In the political domain reasonable pluralism is a fact, and the default assumption is that it will always be so. Hence, for strategic reasons, legislators and citizens had better not saddle their political justifications of a coercible law or policy with controversial assumptions. But in the realm of political philosophy, Habermas assumes, the unforced force of the better argument will eventually prevail, and consequently his theory will either prove itself or be overturned.Footnote ²⁵ In which case, Habermas thinks, there is no reason to claim, as Rawls does, that the ‘citizens, not the philosophers, are to have the final word’ on the truth or correctness of a political theory (RTMW 108). That would only be a relevant consideration if the task of a political theorist was to draw up a blueprint of the good society in the manner of Fichte or Fourier. But by Habermas’s lights, political philosophy has long since given up the illusion that its task is to work out matters of substance like that. So it is not a worry for a ‘post-metaphysical’ theory that confines itself to the task of ‘clarifying the implications of the legal institutionalization of procedures of democratic legislation’ (RTMW 108). The work of the political philosopher in the theoretical domain, the work of explication, clarification and rational reconstruction, is distinct from the democratic deliberation of citizens in the political domain, and it matters not if the theorist makes use of ideas that are not political values and ideas in Rawls’s narrow sense as set out in section 1 above.

Is this a good response? Yes and no. Habermas is right to claim that controversial theoretical commitments do not present the same kind of problem to the philosopher as the religious or otherwise doctrinaire commitments of a reasonable citizen of a liberal constitutional democratic state engaging in political justification in the context of reasonable pluralism. And the onus here is surely on Rawls to show that political theories must answer to the same constraints as first order political discourses. That said, theoretical hostages to fortune are both philosophically and practically unwelcome, since they tend to deflect attention from the main issue. And Habermas has some very hefty ones, for example his pragmatic theory of meaning and truth, not to mention his theory of social evolution.Footnote ²⁶

So far the dispute has broached only what looks like meta-theoretical questions, questions of how political philosophy is best done. Actually the term ‘meta-theoretical’ does not properly capture what is at issue, because one of Rawls’s controversial views is that the justification of political theory is itself a practical political matter: it is the work of citizens in providing the reasons they owe to one another, and of the state in providing the reasons it owes to its citizens, in order to justify the laws and policies with which it demands compliance. But this political practice, Rawls claims, the aim of which is to determine the legitimacy of a political system (and its basic structure), and to evince stability for the right reasons, is impeded when political philosophy reaches for materials beyond the domain of political values (in the sense defined in section 1 above) and makes use of controversial, non-political ideas. So on this view, which Habermas rejects, meta-theoretical controversies are also practical political ones.

That said, there is a further dimension to Rawls’s objection to the comprehensiveness of Habermas’s theory which touches upon an even more fundamental issue dividing them, namely the relation of morality to legitimate law, and the question of the moral content and status of legitimate law. In this case what is at issue is the extent to which, and the way in which, political legitimacy can be ‘freestanding’ from morality. This is a significant area of dispute. However, the point at issue is hard to pin down because the relation between morality and law in Habermas’s theory is complex and somewhat diffuse.

Habermas’s main argument in Between Facts and Norms is that principle (D) assumes ‘the shape of a principle of democracy only by way of legal institutionalization’, which is to say that the principle of democracy ‘derives from the interpenetration of principle (D) and the legal form’ (BFN 122–3). As we saw above, the principle of democracy is supposed to rationally reconstruct ‘the democratic procedure for the production of law’, which in turn is the sole source of legitimacy (BFN 448). (Recall that Habermas rejects the Kantian and the natural law view according to which the laws borrow their legitimacy from the authority of morality: BFN 65, 84, 87, 98, 448–9.)

To understand Habermas’s theory here we need to know more about what he means by the legal form.Footnote ²⁷ Legitimate laws are:

1. i) properly enacted according to a legally constituted process;

2. ii) coercible by force that is itself legitimate.

(These features of the form of law give citizens two different but conjoined reasons for complying with the law: that it is indeed the law, and that violation of the law leads to sanctions of some kind. Habermas thinks of these as the positivity of law.)

But there is still more to law than that. Legitimate laws are in addition:

1. (iii) possessed of an extra-legal rationale.

This extra-legal rationale consists in the various reasons and kinds of reasons – moral, ethical and pragmatic – why the law was made and why it remains on the statutes. This rationale is essential to the law’s legitimacy. It is because the decisions of the political system (parliaments or assemblies) are sensitive to the input of moral, ethical and pragmatic reasons that flow into the democratic-decision making body from ‘civil society’ – Habermas’s term for the informal public sphere in which discourse flows freely, and in which citizens form their political opinions and will – that their outputs (laws and policies) in turn have a rationale that resonates with the background beliefs of the citizens, and elicits widespread quasi-voluntary compliance.Footnote ²⁸ It is this feature of law that attunes it to the private autonomy of citizens, by allowing them to act lawfully out of insight into its rationale, rather than out of sheer obedience to the fact that it is the law, or out of fear of sanctions they will meet if they break the law.

The point I want to focus on is that this rationale of legitimate law is made up in part by reasons of morality.Footnote ²⁹ Although he allows that ethical reasons – reasons that have to do with conceptions of the good life and associated values – and pragmatic reasons play a role in politics, Habermas is clear that moral reasons have priority (BFN 103, 108, 113; FG 64–6; TIO 42–3; JA 13). Ethical reasons only have validity (and thus can play a role in the determination of the legitimacy of law) within the bounds of moral permissibility. In this sense there is what Habermas calls an ‘absolute priority of the right over the good’; and note that by ‘the right’ here he means moral rightness.Footnote ³⁰ Note also that this priority does not just hold in the political domain. The priority of the right (of valid moral norms) is not something that only applies to laws with which people may be forced to comply. It is not a constraint that operates just within the political domain by dint of the fact that reasonable citizens are profoundly divided by their religious, ethical and metaphysical worldviews, and yet must all abide by the same laws. Rather it applies also outside the political domain of legitimate law, in its recesses, in civil society at large, and in the moral realm more broadly. In other words the public reasons that are channelled from civil society into the law-making procedure come already stamped with the priority of moral rightness, and the resulting laws bear the hallmarks of this input (alongside whatever features they get from the form of law itself) as features of their legitimacy.

All this leads Habermas to endorse a thesis, the full significance of which, I believe, has yet to be appreciated. I shall call it the moral permissibility constraint, or MPC for short. Habermas often observes that legitimate laws must ‘harmonize with the universal principles of justice and solidarity’ (BFN 99, 155). He also says that ‘a legal order can be legitimate only if it does not contradict basic moral principles’.Footnote ³¹

Actually, there appear to be two claims here, a requirement of coherence and a prohibition on inconsistency. The former is stronger than the latter, and sits well with Habermas’s functional account whereby public reasons from arenas of discourse in civil society flow into the political system through various sluices and channels, which suggests that the claim is about the moral content of law. (The stronger claim is also implied by Habermas’s criticism that Rawls’s idea of a political conception lacks ‘deontological meaning’, insofar as it is based only on a serendipitous overlapping consensus, for that criticism suggests that by contrast Habermas’s conception of legitimacy better captures the ‘deontological meaning’ of legitimate law.) However, understood as a coherence requirement MPC would seem to be much too strong, because there are too many legitimate laws – for example, those governing which side of the road citizens should drive on – which are morally speaking neither here nor there. In such cases, unlike, say, the cases of laws prohibiting murder or theft, coherence with morality is not in question. Hence we should interpret MPC negatively as being about mere consistency of law with morality: MPC impugns the legitimacy of any law that violates a valid moral norm. Recall that on Habermas’s view legitimate laws elicit quasi-voluntary compliance, and enable the exercise of autonomy, but once set against morality laws cannot do either of these things. Laws that are not consistent with the demands of morality give rise to ‘cognitive dissonance’ and practical dilemmas (BFN 99, 350).

To be clear, MPC is not a just a theoretical requirement, and certainly not a requirement that moral philosophy’s demands be satisfied by law. Rather it is a multi-dimensional requirement of consistency that has to be met within the entire system of legislation, adjudication and administration. This is because modern societies are complex and functionally differentiated entities in which the political system is ‘just one action system among others’ which has to be able ‘to communicate through the medium of law with all the other legitimately ordered spheres of action’ (BFN 302).Footnote ³²

Now in addition to the direct constraint of moral permissibility, that no legitimate law may violate any valid moral norm, Habermas states that legitimate law ‘must also harmonize with . . . ethical principles’. This is actually quite an eye-popping claim for him to make. Habermas is a pluralist about conceptions of the good. Obviously this ethical harmonization requirement cannot imply that if someone’s – or some group’s – ethical self-understanding conflicts with any law, then that law is not legitimate. That would make legitimacy a very scarce resource, almost impossible to achieve.

Moreover, Habermas explicitly equates conceptions of the good with religious and metaphysical worldviews, so such a move would be incompatible with his methodological presupposition that his theory conforms to the precepts of post-metaphysical thinking. So Habermas would be more or less contradicting himself if he held democratic legitimacy to be dependent on any particular ethical conception or values. In this case, I believe that what Habermas must mean, and what we should take him to be saying, is just that legitimate laws should aspire to garner as much support from ethical values as possible. True, this raises tricky questions about whose ethics the legislators should reach out to, and whose not, and raises worries about the tyranny of the majority culture. In BFN Habermas does not go into much detail about the role of ethical reasons in the determination of legitimate law. However, some of the more obvious worries are allayed by the fact that he subscribes to the priority of the right, which in his view implies that all ethical conceptions and all ethical reasons are already operative within the bounds of moral permissibility.Footnote ³³ Clearly then, whatever role they play in the determination of legitimate law, ethical reasons are subject to the absolute priority of morality. Recall that Habermas is not a pluralist about morality but an unabashed universalist: there is no scope restriction on the moral principle (or on the moral norms it validates), and the fact of pluralism (reasonable or not) in his view does not apply to morality. Even so, he holds that discourse ethics, along with its central principle (U), meets all the criteria for a post-metaphysical theory, which means that MPC is still in tune with his methodological self-understanding.

With MPC Habermas claims it is a necessary condition of any legitimate law that it not violate any valid moral norm. Habermas does not claim that moral permissibility is sufficient for legitimacy, in fact he denies this (BFN 111, 452–3, 457). He sums up his position – presumably referring to the priority of moral rightness and to the MPC – thus: ‘law owes its legitimacy in essence to its moral contents’ (BNR 279).

Now one advantage of building this feature into his procedural account of democratic legitimacy is that it explains not only why citizens are guaranteed to be treated as free and equal by law, but also why they are (and should be) protected from certain kinds of moral harm or injury.Footnote ³⁴ But at the same time it raises some very tricky questions for Habermas’s overall theory.

First, how does it square with Habermas’s insistence that the political stands free from the moral? Habermas appears to think that such a question does not arise because (D) ‘remains neutral with respect to morality and law’ and because the principle of democracy is supposed to be derived from (D), together with the form of law. (This is the basis of his critique of Karl-Otto Apel, Kant and natural law: BFN 80, 89.) However, if it is a requirement – whether social, functional, theoretical or all three – that legitimate law not violate any valid moral norm, then democratic legitimacy is subordinate to morality, in the sense that moral requirements exert constraints on the production of legitimate law. This holds regardless of how things stand with theoretical (reconstructive) relations between (D), the form of law, the principle of democracy and what Habermas calls the ‘logical genesis of rights’.

It seems to me that MPC and the priority of moral rightness, as Habermas understands them, smudge the sharp line that Habermas wishes to draw between the domain of legitimate law (and the supposedly sui generis sources of its authority) and the normativity of morality. In short, if Habermas is serious about MPC and the priority of the right, then he must accept that the political system in its production of legitimate law does not stand entirely free from morality. Rather, regardless of the relation that is supposed to hold between the principle of democracy and principle (D), the moral principle (U) and moral discourse (along with its outputs, namely, valid moral norms) constrain democratic legitimacy from within and without. To my mind this is the most important respect in which Habermas’s theory is, in the final analysis, just as Rawls claims it to be, comprehensive. Yet this was not the way in which Rawls understood and developed his objection in the debate with Habermas, focused as he was largely on meta-theoretical and methodological differences. So this very significant dimension of the dispute was not addressed.

Habermas’s only recourse here would be to argue that the discourse theory of law and the conception of democratic legitimacy he puts forward in BFN are freestanding in the sense that they are not dependent on any comprehensive moral doctrine or comprehensive morality, even if, as he admits, ‘law owes its legitimacy in essence to its moral contents’ (BNR 269)

Many of Habermas’s defenders seem to think that such an argument can be made. I can see three possible lines of defence.

The first is the one we have already discussed, namely, the contention that discourse ethics is not a comprehensive moral doctrine because it is a post-metaphysical theory. But as we saw in section 2 above, that means primarily that Habermas’s theory abstains from making certain kinds of metaphysical and ontological claims, that his theory is defeasible, and that it proceeds from empirical assumptions by way of the reconstructive method. None of that shows that it is not a ‘comprehensive moral doctrine’ in any of the three senses distinguished above: that it appeals to worldviews; to actually existing moralities; or to controversial philosophical theories. On the contrary, as we have seen Larmore charge, Habermas’s theory presupposes the truth of a secular worldview by advancing a general theory about the obsolescence of religious and metaphysical narratives.Footnote ³⁵ Second, discourse ethics presupposes an actually existing post-conventional morality, of the kind advanced by Kohlberg and refined by Habermas himself. (No doubt, it purports to be a universal moral theory that applies across all domains. But this does not put it beyond controversy; it is another thing that makes it controversial.) Third and finally, it is itself a controversial philosophical theory that is open to reasonable doubt by reasonable people, and it presupposes other theories that are also open to reasonable doubt. Consequently, whether Habermas’s discourse ethics is post-metaphysical in his sense or not, there are still several respects in which it should be deemed a comprehensive theory, not a ‘political’ one in Rawls’s sense.

The second possible defence of Habermas is that discourse ethics is not a comprehensive moral doctrine because it is procedural and not substantive. Against this it can be and has been argued – to my mind, convincingly – that Habermas smuggles a lot of moral substance into the procedures he reconstructs, something that is borne out both by principle (D) and by the pragmatic presuppositions of argumentation. Since this line of argument has been well developed by others, I shall not pursue it here.Footnote ³⁶ A more direct response is to say that, whether or not discourse ethics is procedural (in Habermas’s sense) is beside the point. Habermas contends that no legitimate law may contradict ‘moral principles’ and the plural here indicates that he is referring to valid moral norms, rather than to principle (U). So the MPC is not an appeal to a procedure, but to substantive moral norms.

Now it is true that Habermas’s theory prescinds (allegedly) from stating what valid moral norms there are. The discourse theory of morality is procedural, Habermas claims, in that it leaves the determination of moral norms to participants themselves in actual discourses. It is not substantive in the sense that it does not offer a repertoire of norms and values. That said, as we have just seen, Habermas assumes that there are some valid moral norms (JA 90–2). He is not an error theorist but a success theorist: he holds that some candidate moral norms successfully pass the test of universalization contained in principle (U). His position is that whatever substantive moral norms there are, and there are some, no legitimate law may contradict them.

What needs to be shown, if this line of defence is to be successful, is that by virtue of being procedural (in Habermas’s sense) the discourse theory of morality is therefore not a ‘comprehensive moral doctrine’ in any of the senses in which Rawls levels the objection. Rawls, as we have seen, uses the term ‘comprehensive doctrine’ in contrast to ‘political conception’. Political conceptions are scope restricted and open to justification by public reason alone. Discourse ethics, though, is not scope restricted. Principle (U) and the norms that it validates are universal and apply to actions in all domains. So discourse ethics is a comprehensive moral doctrine in the sense that it is not tailored only or primarily to the basic structure.

Political conceptions, according to Rawls, have also to meet the publicity condition. Now on Habermas’s account it looks like all valid moral norms will already meet the publicity condition. That follows from their being universally valid. (Note that Rawls abstains on this point, because he, unlike Habermas, allows that reasonable people may disagree about morality and moral theory, as well as about which moral norms are universally valid.) What counts for Rawls is that a moral norm should also be a ‘political’ value, i.e. it should form part of an overlapping consensus of reasonable doctrines, and in addition, it should be the case that all free and equal citizens can be reasonably expected to endorse it as a political value. The difference here may seem small but is important. For Habermas the acceptability (and acceptance/acceptedness) of valid moral norms is carried over intact from the moral domain into the political and legal sphere, and holds both within and without. So insofar as moral norms – on Habermas’s account of morality – meet the publicity condition, they do so in the wrong way, so far as Rawls is concerned. They meet it qua universally valid moral norms – and in that sense discourse ethics is a comprehensive moral doctrine in the second of our three senses. For Rawls, by contrast, in the political domain all bets are off: all political values, whether moral or not, have to win their reasonable acceptance anew, as political values, in the political sphere from free and equal citizens.

Finally, it might be argued that in the process of its Verschränkung (interpenetration) with the form of law, somehow the comprehensive elements of morality are screened out, so that one is left with only something like what Lon Fuller has called the ‘inner morality of law’ (Fuller Reference Fuller1969: 42). The objection rests entirely on the aptness of Fuller’s phrase (the ‘inner morality of law’) to capture the normative feature of what Habermas is after with MPC. However, Fuller’s idea is quite different. For one thing, what Habermas understands by ‘moral principles’ and ‘valid moral norms’, and thus by the MPC, is something very different to what Fuller means by the ‘inner morality of law’. What Fuller has in mind is that laws take the form of rules, that these are clear and coherent, that they are promulgated, that they are not made retrospectively, etc. (Fuller Reference Fuller1969: 39). Most of what Fuller calls ‘the inner morality of law’ is in fact rendered by what Habermas calls ‘the form of law’ and by the requirement of the principle of democracy that laws be properly enacted according to a legally recognized process. It is certainly not the case, on the account Habermas’s gives in BFN, that in the process of the Verschränkung of principle (D) with the legal form, the universal scope of moral norms is screened out along with the controversial and thus dubitable features of his moral theory. So the appeal to Fuller’s theory is not appropriate and cannot be used to save Habermas from the objection.

It turns out, then, that Habermas’s theory is indeed a ‘comprehensive moral and philosophical doctrine’ in at least the second and the third of the senses set out above, and that consequently it is a doctrine of the kind that Rawls argues political theory should eschew. Furthermore, if Habermas is serious about endorsing MPC and the priority of the right (as he construes it, namely as the priority of the moral over the ethical) then he ends up endorsing a position uncomfortably close to the one that he attributes to Apel, Kant and natural law theorists, and which he rejects on the grounds that they make the legitimacy of the law depend on the authority of the moral.

This does not in itself resolve the debate between Habermas and Rawls in the latter’s favour, however, although it does uncover a deep problem in Habermas’s theory. If Habermas was to prevail, without radically altering his own theories, he would need to convince Rawls that discourse ethics (or something similar) is true, that there are some universally valid moral norms, and that such norms as there are, by virtue of being universally valid, count as political values. I doubt that Rawls would be minded to concede these points. Nonetheless, rather than insisting that his own theory is ‘morally freestanding’ and that the source of legitimacy is not borrowed from the moral, Habermas would have been better off pursuing a different line of argument, and countering that Rawls conceives the boundaries of the political too narrowly and too impermeably. Against the charge that his theory was comprehensive in the third of our senses, as involving appeal to controversial philosophical theories, including moral theories such as utilitarianism or Kantianism, Habermas, as we saw, was content to reply that Rawls has weightier philosophical commitments than he is prepared to admit: ‘Political theory’, as Habermas puts it, ‘cannot move entirely in the domain of the political and steer clear of . . . philosophical controversies’ (MW 93). Habermas could make, and in my view probably should have made, a similar argument against the charge that discourse theory is comprehensive in the second sense, namely, as involving appeal to actually existing moralities. He could and should have argued that the domain of the political and of democratically legitimate law is inevitably entangled in morality, understood as the actually existing moral practice that discourse ethics claims to reconstruct, and hence cannot, as Rawls claims, be freestanding from every ‘comprehensive moral doctrine’.Footnote ³⁷