I. Uncoupling democracy from statehood?

We know the traditional argument which goes: the only, or at least the main standard of political legitimacy is democratic in its nature (whether we construe democracy as direct, or representative, or deliberative, or a combination of these different conceptions), and democracy can only flourish within the nation state. Hence, there is either no need or no possibility to democratise transnational governance. These days, this argument is mainly honoured by its rejection. Critics of the traditional democracy/statehood nexus attack this ‘methodological nationalism’ by decoupling democracy and the nation state. They either identify (in the empirical version of the counterargument) or postulate (in the normative version) democracy in the supranational sphere. But this path is not successful, and not even very promising.

The usual number one exhibit in the empirical version of the counterargument is the European Union (EU),^{Footnote 8} but there are two problems with the EU qua an exemplar of supranational democracy. First, the EU is not typical of supranational polities; if anything, it is very atypical among contemporary sites of supranational authorities. This may explain a visible EU-orientedness (some would say, obsession) among scholars of post-national constitutionalism. It has long departed from the space occupied by classical supranational organisations – to a much higher degree than any other supranational entity in the modern world. It has much broader and more pervasive competences than ‘regular’ international organisations; its decision-making is now based on majority vote and not consensus; it affects the legal status of individuals in Member States directly; it possesses common currency (in much of its territory) and common citizenship; its Member States surrender the right to conclude international treaties in the areas belonging to the exclusive competences of the EU, etc. The second problem with the EU as an alleged exemplar of supranational democracy is that democratic procedures within the EU institutions are at best embryonic. At the level of the EU, democracy is more a matter of aspiration than reality, and the new Article 10 of the TEU which proclaims that ‘the functioning of the Union shall be founded on representative democracy’ is only a promise. The words ‘shall be’ literally have to be read as the future tense, rather than in the imperative meaning. The only truly representative body (the European Parliament [EP]) is relatively weak, and has only an ancillary role in legislation, being incapable of initiating legislation; the law-making is divided strangely between the technocratic (Commission), the intergovernmental (Council) and the parliamentary (EP) bodies in a way which hardly resembles any established principles of separation of powers; the outcomes of the direct elections (to the EP) have only indirect effect upon the composition of the main executive body (the Commission), etc.^{Footnote 9}

The second version of the counterargument against the democracy/statehood nexus is normative. It is based on an observation that a simple extension or extrapolation of democratic schemes from nation state to global constitutionalism is neither possible nor needed, and that, in the process, the very meaning and nature of democracy must be fundamentally transformed.^{Footnote 10} But the nature of the transformation has never been convincingly explicated. The claim is that when conceptualising supranational democracy, we should try to forge conceptions of democracy not based on the template of municipal democracy, and accept that international or supranational democracy will look quite different from the way it operates at the national level.^{Footnote 11} This is what may be called a ‘transformative’ conception, and is what theorists of cosmopolitan democracy invite us to adopt. They usually draw a picture of a multilayered scheme of governance reliant on a thorough dispersion of decision-making powers. But then, quite apart from the vague, indeterminate, and possibly utopian character of such visions, the question arises whether such a dispersed pattern of authority would be still ‘democratic’ in a meaningful sense of the word, given its significant reliance on contestatory functions performed by various civil society actors, and on ‘unconstrained and uncoerced communication’ within the realm of ‘transnational civil society’.^{Footnote 12} As James Bohman observes, commenting upon John Dryzek’s vision of transnational governance based on informal networks, communication and contestation, ‘why should we call such governance “democratic” even in some minimal sense? Does it provide for the basic possibility of democratisation, so that those who suffer injustice in the current democratic system have a real basis on which they may effectively make claims to justice?’^{Footnote 13} Bohman’s rhetorical question indicates that supranational democratic governance along cosmopolitan lines may actually be retrogressive compared to state democracies in terms of some important democratic values, such as minority protection or even representativeness of the decision-makers.

This is not to say that the goal of striving for more democracy, whenever possible in transnational settings, should be neglected or abandoned right at the outset. Attempts to democratise supranational institutions by bringing them closer to the model of parliamentary representation, tripartite separation of powers and aspects of direct democracy is an eminently worthwhile task. In various domains of transnational governance, any space for injecting it with democratic representation and accountability should be seized. This is the main point of an already cited article by Gráinne de Búrca who makes a plea for a ‘democratic-striving approach’ to governance beyond the state, the approach which ‘acknowledges the difficulty and complexity of democratizing transnational governance yet insists on its necessity, and identifies the act of continuous striving itself as the source of legitimation and accountability’.^{Footnote 14} The two parts of that sentence separated by the conjunction ‘yet’, identify the difference between the actual practice and the aspiration. The aspiration is noble but the ‘difficulty and complexity’ of bringing it about seem enormous, other than in some narrowly confined locations of transnational governance. While de Búrca provides convincing examples of such specific sites which lend themselves to increased democratic practices, albeit in rather embryonic forms – such as anti-poverty policies of the international financial institutions: the World Bank and the IMF – she wisely stops short of making a claim that they may provide a blueprint for a generalised participatory-representative transnational democracy.^{Footnote 15}

One should be of course careful to avoid comparing the actual supranational sphere, marked as it is by fundamental democratic deficits wherever we look, with an ideal of nation-state democracy (as constitutional lawyers are sometimes tempted to do, mistaking textual constitutional rules for the political reality). Rather, we should compare it to real-life democratic practices. We know that in most countries the actual functioning of democracies is increasingly characterised by the decline of the powers of parliaments (not only under pressure from globalisation), concentration of power of party leaderships, deep rifts along lines of class, race, religion, etc, declining participation in elections, and the growth of various largely unaccountable bodies, such as constitutional courts or central banks. But even by a comparison of these imperfect and often defective democratic practices, the supranational sphere, as currently constituted, is glaringly non-democratic.

II. Public reason in supranational sphere

The upshot of the previous section is that uncoupling democracy from the nation state is not a promising move, at least for the time being. My proposed strategy is different. Within the triangle of nation state, democracy and legitimacy, rather than uncoupling democracy from the nation state I suggest that we should uncouple legitimacy from democracy. We may come to the conclusion – eminently plausible – that tying up legitimacy with democracy (the latter being strongly connected to national statehood) may be improper because it disregards the concern for all stakeholders of decisions made by states. Their effects often transcend the boundaries of a particular state (think of environmental pollution, overfishing of the seas or compelled migration), and yet statehood-bound democracy is not responsive to those externalities. If the level at which a decision is made, or a norm generated, does not correspond adequately to the scope of its effects, then the legitimacy of the norm or decision is fundamentally flawed, even if democratic mechanisms are fully respected at the state level. This is because the constituency of the democracy does not correspond to the range of those who are affected adversely by the norm or decision. Indeed, one of the founding rationales for the GATT system, which then paved the way for the WTO, was the perception that (protectionist) policies made in one country may have negative effects upon people in a different country, the latter being denied a forum for influencing those decisions.

The point of disentangling democracy and legitimacy is not to disparage or downplay democracy, but rather to put it in its proper place. In the world of plural sites of constitutionalism we need to be careful not to insist on democratic conditions of legitimacy everywhere. But what is the proper place of democracy? As a first approximation, we may perhaps say that democracy is a crucial element of legitimacy only in the decisions and norms which are ‘choice-sensitive’, i.e. where it is important for the decisions to reflect the actual distribution of preferences and views in the society, no matter what the modes of formation and qualities of these views are. In the choice-sensitive area, the right authoritative decisions (norms, standards, etc) are those which mirror, as accurately as possible, the actual distribution of preferences, i.e. of choices that people actually made.^{Footnote 16} But if we look at the proper scope of democracy in this way, then we will probably realise that the range of decisions, for which a proper standard is only ‘choice sensitivity’, is in reality rather narrow. The textbook examples of choice-sensitive decisions (whether to build a new stadium or a new road system, from the available public funds) indicate that such stark alternatives where either choice is equally prima facie reasonable and the tiebreaker is about the degree of social support are extremely rare. For in most cases, what matters for legitimacy of decisions, democratic or otherwise, is the sort of reasons that were or that could have been provided for them. And often the mere fact of a particular distribution of preferences is insufficient (or even irrelevant) for the legitimacy of decisions which mirror this distribution, unless we are dealing with decisions which are very simple or very trivial – something that rarely occurs in politics.

A note about the concept of legitimacy used here is in order. For me, it is a standard which is located halfway along a spectrum between legal validity and justice: stronger than that of validity because it confers some moral authority upon a norm, but weaker than that of justice because reasonable people can disagree about the justness of a particular norm, while agreeing upon its legitimacy. Hence, it is a typically liberal standard concerning the grounds for respecting a rule, the merits of which we do not necessarily agree upon; in fact, a good test for the legitimacy of a rule is to ask ourselves whether we should still respect it despite disagreeing with its substance.^{Footnote 17}

Two distinctions concerning the concept of legitimacy are important: (1) between weak and strong legitimacy, and (2) between the legitimacy of authority and the legitimacy of authoritative decisions. Some people believe that legitimacy is equivalent to a categorical duty to obey a legitimate rule – that is, it is sufficient to identify a legitimate authority or a legitimate decision in order to conclude that we (people within this authority’s jurisdiction) have a duty to obey.^{Footnote 18} This is a stronger understanding of legitimacy – but it is by no means the only or even the dominant understanding of legitimacy in legal philosophy. There is also a weak understanding of the concept, where to state that an authority is legitimate is to say only that it has a right to issue authoritative directives. On the side of the citizens, the only categorical duty is to consider these directives with a degree of respect, but not necessarily to obey. If the authority’s legitimacy is based exclusively on a certain form of justification of its decisions, as in Joseph Raz’s ‘service conception of authority’ (the nucleus of which is, to paraphrase Raz, that citizens recognise an authority as legitimate when they believe that their actual goals will be best achieved if they allow themselves to be guided by the directives of the authority, rather than acting on their own preferences directly and figuring out the ways of best achievement of them on their own),^{Footnote 19} then we need some extra arguments to support our duty to obey. No one has a duty to obey a decision merely by virtue of the fact that it is justified: it may be prudent or even wise to do so, but no one has a duty to be prudent or wise.^{Footnote 20} To close the gap between these conceptions of legitimacy (in the sense of respect for authoritative decisions and the duty to obey them) we need some extra arguments, such as those about actual commitment, consent, contract or fair play. These arguments do not, of themselves, figure in arguments for legitimacy, weakly understood.^{Footnote 21} So I take legitimacy as an attribute related to the justification of decisions – in the weak sense of legitimacy.

Second, legitimacy may be seen as an attribute either of authorities (polity legitimacy) or of their specific decisions. At first blush this may seem to be a pedantic distinction. We may be tempted to believe that, by definition, legitimate authorities issue legitimate decisions, and decisions are legitimate by virtue of being issued by legitimate authorities^{Footnote 22} – but it is not so. Sometimes illegitimate authorities may issue legitimate decisions (e.g., some laws under the apartheid regime), and sometimes legitimate authorities may issue illegitimate laws (e.g., racial segregation in the US). Nothing of conceptual clarity is achieved by connecting these two things by definitional fiat. There may be decisions made by by-and-large legitimate authorities that breach the principles of legitimacy, for instance due to procedural defects or violation of rights which the regime otherwise respects. As Philip Pettit observes: ‘there is … room for claiming in the same sense of the term that while a regime is generally legitimate, certain laws or appointments … are illegitimate: they happen to breach conditions of legitimacy that the regime generally respects’.^{Footnote 23} And I prefer in this article to focus on the legitimacy of decisions rather than of authorities or polities. This dimension seems better suited to supranational spheres, in which we may identify without much uncertainty various rules, decisions, instructions or guidelines (meant to be) binding upon states, but where the existence of supranational ‘authorities’, not to mention ‘polities’, is much more question-begging.

So to sum up the conceptual argument about legitimacy so far: we adopt here a notion of legitimacy which is weak and which is about specific decisions. As I have already foreshadowed, legitimacy of political decisions crucially depends upon the sort of reasons provided for these decisions – not only those ‘provided’, but also demonstrable (as the most likely purposes pursued by a decision-maker). This is, generally speaking, a ‘public reason’ conception of legitimacy: publicly admissible reasons confer legitimacy upon decisions.^{Footnote 24} It is a reason-constraining conception of legitimacy: certain motives which trigger the law, and which are not reasonably acceptable to all – for instance because they are sectarian, or based on prejudice, or hatred, or self-interest of the rule-makers – taint the law as illegitimate. We have a right to a justification^{Footnote 25} for those authoritative decisions which apply to us, and in particular which restrain our freedom – and some justifications simply will not do, if we have no good reasons to accept them.

This last point identifies what is specific about public reason within a broader category of ‘public reasoning’ or ‘public justification’: not every publicly offered justification for a decision appeals to such reasons for action which are reasonably acceptable to all, or at least reasonably non-rejectable by all to whom a decision applies. A liberal concept of public reason, best elucidated by John Rawls and his followers, urges us to advance and advocate only such uses of coercive powers towards our fellow citizens which rest upon the types of arguments which they cannot reasonably reject. Perhaps the best articulation of public reason (very much in line with Rawls’s idea) was given by Charles Larmore, who stated the fundamental directive of political liberalism by saying that ‘basic political principles should be suitably acceptable to those whom they are to bind’.^{Footnote 26} The implication of this is clear: some arguments, even if actually present in the minds of legislators or policymakers, are not qualified to figure in the public defence of a law. The law must be defensible in terms that belong to a forum of principle rather than an arena of political bargains, or power plays of naked interest, or competition between sectarian ideologies.

Stated in this way, a natural home for the idea of public reason (in a strict sense, narrower than any public justification) is in the nation-state setting. Indeed, for Rawls and his interpreters, the confinement of public reason to a national setting had been always a tacit assumption for the conception. Significantly, when Rawls ventured beyond the nation-state context, even though he referred occasionally to ‘public reason of the Society of Peoples’,^{Footnote 27} the concept underwent an important transformation, eroding its capacity to act as a compelling legitimating device. The fundamental difference is that Rawls’s international public reason (as the public reason in the ‘Society of Peoples’ may be called), concerns a horizontal realm, i.e. the relationship between the states, or the Peoples while supranational public reason as used in this article is fundamentally about vertical relations, i.e. between supranational authorities (courts, regulatory agencies, executives bodies, etc) and states and other entities. That Rawls’s concept applies only to horizontal relations is obvious; as he says, it concerns ‘the ideals and principles of the foreign policy’,^{Footnote 28} the ‘mutual relations’ of liberal peoples ‘as peoples’,^{Footnote 29} ‘a foreign policy and affairs of state that involve other societies’,^{Footnote 30} and ‘the political relations among peoples’.^{Footnote 31} As one can see, the locus of Rawlsian international public reason is international law and international relations traditionally conceived, as they concern relations between the states. There is no room in this idea of public reason to regulate vertical relationships between supranational authorities and states and other entities, such as firms, organisations and individuals. In fact, this horizontality of Rawls’s conception is bolstered by the whole idea of the Law of Peoples, which is the proper space for Rawlsian international public reason, and concerns the relationships traditionally controlled by international law and international relations between sovereign states. Indeed in the first sentence of the book, he declares that by Law of Peoples he understands ‘a particular political conception of right and justice that applies to the principles and norms of international law and practice’.^{Footnote 32}

For these reasons, the idea of supranational public reason developed here builds upon the original Rawlsian public reason, rather than on its transformation (and, as I claim, fundamental erosion) in The Law of Peoples. This idea of public reason as a legitimating device^{Footnote 33} gives effect to the values of (1) universality (because it aims at a consensus on the underlying reasons which may be pursued in the public sphere), (2) reciprocity (because it postulates that we should propose and advocate only such laws that are based on grounds which can also be endorsable by those who disagree with us on the merits of a particular arrangement), and (3) openness to cultural contexts (because the actual substance of public reason will change from society to society and from culture to culture, and for instance the ECtHR’s idea of the margin of appreciation reflects this idea). As is clear, this approach to legitimacy transcends another conventional distinction: that between ‘input’ and ‘output’ legitimacy. The idea of public reason is both about input – because it is about the factors which, so to speak, precede the decision (about the factors which trigger a decision), and output – because those reasons imbue the decision with a particular substance; they are inseparable from the content and the effect of the decision. But it is not about ‘input’ in a purely democratic sense (e.g., in terms of the electoral mandate of the authorities) and not about the ‘output’ in terms of efficiency (that is, saying that authorities are legitimate when they deliver the goods expected by their constituencies and promised, explicitly or implicitly, by the authorities).

How much mileage can we get from this idea of public reason, as a legitimating device in new post-national or supranational norm-producing authorities? Quite a lot. Supranationalism provides a space in which there is a degree of synchronisation of national and international actors in pursuit of the legitimacy of political authority, based on standards of justification of authoritative directives by appeal to public reasons, i.e. reasons which it is unreasonable for the parties concerned to reject. Scrutiny of exercises of authority conducted in accordance with the ideal of public reasons may be seen by those parties as legitimate by virtue of its reasonableness, which resides in the patterns and the substance of justification. If those parties, and in particular individual citizens (but also associations, business entities, NGOs etc), can contest the decisions taken at a national level, and demand scrutiny at a supranational level of claims that improper or insufficient reasons have been provided by the state and its institutions, supranational legitimacy may be produced even without any concern for the democratic (or otherwise) pedigree of those supranational scrutineers. But supranational law-making and adjudication must meet general requirements of epistemological value, such as accuracy of statements of facts when they are crucial for a judgment, openness to diverse points of view in order to search for good reasons for decisions, deliberate screening off of prejudice, hostility and self-interest, in providing public justification for authoritative decisions and norms, etc.^{Footnote 34} Others are more specific to supranational authorities: epistemic access to broad comparative material and to an enlarged sample of views, the capacity to harmonise and coordinate norms and actions at a large scale, etc. More detail, supplementing these necessarily vague requirements, will hopefully become apparent in the discussion of case studies below.

I will provide two examples to show public reason at work in the supranational sphere, as a provider of legitimacy for two quasi-constitutional supranational systems of adjudication: the ECtHR and the dispute settlement system of the WTO. They are two quite different instances of deliberative, supranational constitutionalism, which claim legitimate authority based not necessarily on the consent of the states concerned, and certainly not on participatory/electoral democracy, but rather on a commitment to shared principles taken today to be fundamental. All three characterisations in the label ‘supranational, deliberative constitutionalism’ are important. The adjudicatory institutions in this category are: (1) supranational, in that they operate in the international sphere but ‘pierce the veil of the state’ rather than dealing with states as unitary entities; (2) deliberative, because the force of their determinations relies crucially upon the strength of the arguments provided, and (3) constitutional, in that they may censure (though not necessarily immediately invalidate) certain national laws and regulations, duly adopted, as incompatible with higher norms. The selection of international adjudicatory bodies as case examples may be seen as making the argument for ‘supranational public reason’ more obvious – after all, courts (or quasi-judicial bodies) are precisely the type of institutions that rely on the strength of the reasons they provide to confer legitimacy upon their decisions. The supposed contrast between adjudicatory and other (including legislative or regulatory) institutions – according to which reasons-giving is allegedly characteristic of the former but not the latter bodies – is almost certainly overdrawn. In any case, the point is that supranational bodies in general (adjudicatory or otherwise) cannot benefit from reliance on a broader national institutional context which, as a whole, benefits from electoral, ‘input’ type of legitimacy. Hence, an argument about public reason at work in supranational adjudicatory bodies can give some weight to a broader argument about public reason in the supranational sphere. In this context I would like to use the concept of ‘justificatory discipline’ coined usefully by Jürgen Neyer.^{Footnote 35} It has, for our purposes, two main interrelated aspects. The first is that individuals or states (when they act as agents or trustees for their citizens) have a right to demand and receive justification whenever their liberty is constrained. Supranationalism operates by exercising supra-state control over the sort of reasons states provide to their citizens. Second, compliance of states with supranational determinations is voluntary but disciplined, based on the strength of reasons provided. As Neyer put it: ‘Supranationalism is about the representation of arguments, and not about power and preferences. Under conditions of anarchy, states bargain. In an ideal supranational structure, states deliberate.’^{Footnote 36}

Such legitimacy of supranational quasi-constitutional entities stands in complex relation to the authority of states. But in any event, legitimacy of international entities is not a simple extension of the political authority of states founded on their consent as international actors. One reason for this is that those supranational adjudicatory or regulatory bodies engage not with the states as single unitary entities but pierce the veil of the state, ‘penetrate the surface of the state’, and interact with particular actors within the state or on a sub-state level.^{Footnote 37} Various actors within states may have diverse and sometimes mutually conflicting interests and incentives in engaging with supranational authorities. Hence, there is the possibility of synergy between supranational and national sites of authority. This is so even though the symbiosis is always fragile and tentative, resting as it does on an alliance between supranational authority with only some state institutions and actors. Precisely for this reason, there is always scope for tension, conflict and friction between different sites of authority which are not easily captured by the traditional notion of sovereignty.

III. Pursuit of legitimacy by the European Court of Human Rights

My first example is that of the ECtHR: the court initially set up as an extraordinary, super-appellate judicial body but which has become in recent decades a quasi-constitutional court for Europe. Its fundamental purpose is to scrutinise the reasons provided by states to justify laws and decisions which putatively infringe upon the European Convention on Human Rights. While an inquiry into the reasons proffered for individual decisions of lower courts is a typical approach for any appellate court, a scrutiny of reasons provided for laws is a par excellence constitutional function. This became most visible in the ECtHR decisions which deemed national laws (and not merely individual judgments) ‘un-Conventional’, i.e., to be inconsistent with the European Convention. The delivery by the Court of so-called ‘pilot judgments’^{Footnote 38} – a practice the Court itself invented – reaches the limits of a traditional court’s legitimacy, because it collides in the most obvious way with national authorities’ right to adopt legislation in conformity with local values.

But this procedure runs the Court into a fundamental legitimacy problem. The Court protects its legitimacy to issue such pilot judgments by: (a) relying upon the political branches of the Council of Europe (CoE), and in particular, the Committee of Ministers of CoE where states’ consent is most operative; (b) paying at least lip-service to the ‘margin of appreciation’ which is a doctrine requiring a high degree of deference by the Court to the Member State summoned before it, when there is no consensus in Europe on whether a particular right is recognised by the Convention; (c) careful and explicit argumentation along the proportionality analysis, with its qualities of transparency and reviewability; and (d) allying itself with national constitutional courts whenever possible, by forming a sort of coalition against recalcitrant legislatures. I will discuss, below, only the last two points, as the most directly relevant to the theme of this article.

Of particular importance is point (c). As Mattias Kumm observed ‘The proportionality test … provides a structure for the justification of an act in terms of public reason’.^{Footnote 39} This is clear: proportionality serves to discern the aim of a restriction of a constitutional right, and by scrutinising the relationship between the legislative means and the aims (in terms of suitability and necessity) it will check whether the asserted aim is likely to have been the real one. An important, potentially legitimating virtue of the proportionality analysis is that it has become a quasi-universal idiom of the contextual and purposive analysis of legal rights. The judgments handed down as a result of well-conducted proportionality analysis can be well recognised and understood by their addressees even in the absence of agreement on the results: the template of proportionality creates a common epistemic ground for different layers of authority. This diffusion acquires, naturally, a normative character. The use of the analysis becomes seen as best practice, signifying belonging to a community marked by its knowledge and enthusiasm for this framework of judicial inquiry.

It is easy to understand why this has happened. Proportionality may create a legitimacy-enhancing illusion of strict, quasi-‘mathematical’ reasoning that leads in a value-free way to a determinate outcome: the judge may be seen as merely filling a checklist, assigning particular weights to various steps of the calculus. The illusion is of course deceptive, but at the rhetorical level of weighing and balancing allegedly quantifiable and mutually commensurate values, it may strengthen the appearance of a decision being purely legal, strictly constrained, and untainted by policy choices or controversial value judgments. At the same time, its formulaic and canonical character adds to its transportability from one legal culture to another, and increases its transparency. The judgments, following a canonical pattern and observing a standard sequence of steps and tiers, will produce recognisable outcomes which lend themselves to relatively easy scrutiny in different legal cultures. This is so even if legal and cultural diversity will result in the application of different weights to different stages of the weighing and balancing process.^{Footnote 40} At the same time, proportionality analysis has an edge over a categorical approach to rights adjudication, in that it moderates the victim’s sense of loss. As Mary Ann Glendon put it many years ago, praising the Strasbourg Court’s methods of reasoning, ‘[t]he more searching and tentative style of the European Court … gives winners fewer grounds for gloating and leaves the losers less reason to feel angry and alienated’.^{Footnote 41} Especially in the supranational context, this factor of moderating the losses, where ‘losers’ may be the states whose laws have been found in breach of the Convention, is an important legitimacy-enhancing asset.

Concerning point (d) in the list of argumentative strategies employed by the Court to enhance its legitimacy: when the ECtHR, as a supranational court, manages to enlist a national court in its ruling that a particular national law breaches the Convention, and supply a convincing argument regarding a specific case which is parallel to a national constitutional court’s abstract argument about the law, concerns about legitimacy are best allayed, and may be put to rest. This is an illustration of a broader phenomenon: international adjudicative bodies sometimes form alliances with national courts, often against recalcitrant national executives and legislatures, controlled by executives.^{Footnote 42} The greater relative legitimacy of national courts contributes to the legitimacy of an international tribunal. It also strengthens the international court’s review capacity, because the international tribunal can find support in a national court’s activism. National courts, in turn, also benefit from stronger international tribunals. International tribunals can bring resources to the table that may prove valuable to national courts, and their own legitimacy may be enhanced (for instance, through the domestication of the European Court by a national tribunal, when it refers to it favourably in judgments, or is praised in extrajudicial pronouncements by municipal judges and academic lawyers, etc). Shai Dothan gives the example of the ECtHR’s ruling against the UK concerning the detention of foreign nationals suspected of terrorist activities who could not be deported because they might be tortured in their states of origin. As Dothan observes: ‘The ECtHR could issue such a controversial judgment without damaging its legitimacy because it followed all the legal decisions made by the House of Lords in a judgment issued on the same case.’^{Footnote 43}

The ECtHR is an example of supranational public reason at work, inasmuch as the Court’s authority relies primarily upon the strength of its arguments about the reasons which can be provided for laws that putatively breach the Convention’s rights. It is true that the CoE Member States endowed the Court by mutual consent with the legal grounds for its rulings (the Convention, including its successive Protocols. However, by its interpretation of the Convention, the Court has imposed upon the states more specific obligations than they had anticipated, and to which they would not have necessarily consented: they tied their hands in ways beyond their control. The Court has established and consolidated this authority by a number of argumentative moves. In particular, it has achieved this through an inventive use of proportionality analysis (as discussed above), and by aligning itself argumentatively with national courts, often against national governments. In addition, one should mention the doctrine of effectiveness (effet utile) – an important device in the Court’s arsenal of argumentative strategies, which requires that the provisions of the Convention be interpreted so as to safeguard rights that are ‘practical and effective’ rather than ‘theoretical and illusory’.^{Footnote 44} This justified an expansive interpretation of the Convention rights, including the recognition of broad positive obligations on the states that were not necessarily immediately detectable in the abstract articulation of the relevant rights themselves.^{Footnote 45}

Another argumentative device serving the expansion of protection of rights has been the doctrine of ‘autonomous meaning’ of the terms used in the Convention. The Convention meaning is independent of the meaning of those (or equivalent) terms used in the domestic law of the Member States. Thus, for instance, the Court has applied the Article 6 guarantee of fair trial (which applies to the ‘determination of … civil rights and obligations or of any criminal charge’) despite the fact that a municipal law may characterise a given dispute as belonging to ‘administrative law’, or describe an offence as ‘disciplinary’.^{Footnote 46} This means, in practice, that the Court protects European citizens’ rights despite what their Member States have actually committed themselves to. The ‘concepts’ serve as shells for the ‘conceptions’ developed by the Court, regardless of and often contrary to the ‘conceptions’ recognised at a state level.^{Footnote 47}

Between them (and with the addition of the doctrine of evolutive rather than static interpretation, by which the Convention is seen as ‘a living instrument which must be interpreted in the light of present-day conditions’),^{Footnote 48} these doctrines have served as the grounds for expanding, or even inventing new rights (such as the right not to be exposed to pollutants, the rights of aliens to reside with their spouses, to receive financial support, and not to be removed if they are at risk of being exposed to inhuman treatment),^{Footnote 49} well beyond what could have been anticipated by the contracting states.

In the end, the legitimacy of the ECtHR in the national order of Member States of the CoE is not primarily a result of its formal embeddedness in the national orders. The only formal obligation that Member States have undertaken is to abide by the judgments of the Court to which they are parties.^{Footnote 50} And even then the effect is not immediate, but depends on the political will of a state, with ‘naming and shaming’ basically the only available sanction against a recalcitrant state. The Court’s position vis-à-vis Member States is not similar to that enjoyed by the Court of Justice of the EU. There is no equivalent of the supremacy/direct effect or preliminary reference procedure which would give national courts direct access to the European tribunal.^{Footnote 51} Hence, whatever compliance effect it has acquired – and it is high – must be due to its argumentative resources, combined with its strategic alliance with national municipal actors (especially constitutional courts) which themselves have incentives to align with the European Court. Those incentives, however, are qualified, and do not translate into a strong obligation; municipal courts can and occasionally do depart from the case law and interpretation of the ECtHR and, having ‘taken into account’ the ECtHR, go their own way.^{Footnote 52} Paradoxically, this latitude increases rather than detracts from the legitimacy of the ECtHR because in those numerically dominant cases in which it does prevail, the acceptance of its supremacy is voluntary. The national courts act as willing and autonomous gatekeepers, helping align the law of their state with European legal standards of human rights. As a prerequisite for a successful coalition, there must be sufficient argumentative resources in the judgments of the European Court to facilitate the emergence of synergy, and the effective influence of the ECtHR’s judgments on the conduct of national decision-makers.

IV. WTO and ‘political obiter dicta’

Another example of legitimation through public reason at a supranational level is the dispute settlement system (DSS) within the WTO. Just like in the case of the ECtHR, there has been, over the past decade or so, a growing tendency to characterise the WTO as (quasi-) constitutional.^{Footnote 53} For our purposes, i.e. the discussion of public-reason based legitimacy, the dimension of justification of rulings is of particular importance. The evidence shows that WTO adjudicators have been increasingly engaged in a deliberate quest for legitimacy amongst diverse audiences. This is especially observable in so-called ‘trade and …’ cases, when they have to resolve tensions between trade liberalisation and other values, such as consumer protection, public morals, labour rights, environmental concerns, etc.

In such cases, it is natural that legitimacy concerns intensify, because WTO adjudicators exercise scrutiny of national choices on matters going beyond the original brief of the Organisation – that is, beyond narrow trade considerations – and so intrude upon the sphere traditionally thought to be the domain of national democratic choices.^{Footnote 54} When the WTO DSS concerns non-trade interests, its legitimacy is at its weakest – reflected by the extra care that the Appellate Body (AB) takes to explain its reasons, and also to justify its authority reaching beyond the traditional goal of trade liberalisation. The evidence shows that various legitimacy-enhancing rhetorical and argumentative strategies are pursued by the WTO AB, especially in those rulings of the ‘trade and …’ category when the AB was finding against the states. When one reads the AB’s reports in such cases, for instance the US–Clove Cigarettes case of 2012, one is struck by the careful explanatory, justificatory and defensive tone of the Report, even though the AB has overturned an important national choice (an anti-tobacco regulation prohibiting certain flavoured cigarettes).^{Footnote 55}

As WTO scholars and experts observe, there has been a steady move away from strict textual interpretation to more purposive and contextual methods, to provide adjudicators with more discretion to place the argument in a broader set of societal values.^{Footnote 56} In the context of AB rulings this is expressed, for instance, in the appearance of separate explanatory paragraphs which aim to show sensitivity to national values and choices regarding non-trade values – and to emphasise the limited political implications of the rulings. Some scholars have even dubbed these explanatory paragraphs ‘political obiter dicta’.^{Footnote 57} It has been also noted that WTO Panels refer more often to other segments of international law, which produces a kind of horizontal cross-fertilisation with the effect of enhancing the legitimacy of their rulings: ‘Reference to interpretative norms of general public international law enhances the legitimacy of the dispute-settlement organs in adjudicating competing values, as they are not specific to a regime that has traditionally privileged a single value – that of free trade.’^{Footnote 58}

It is important also to note that, as Sivan Shlomo-Agon emphasises, the WTO addresses multiple audiences. It speaks not only to governments but also to NGOs, domestic interest groups, the general public, other international institutions (such as the WHO), the media, the public at large, etc.^{Footnote 59} CA Thomas constructs the helpful notion of a ‘community of legitimation’ (or ‘a legitimating community’). He includes amongst the WTO legitimating community in addition to ‘the trade policy insiders’, ‘human rights activists, officials from developing countries, other intergovernmental organisations, anti-globalisation protestors and environmental groups’, etc and adds that ‘[t]he proliferation of these legitimating communities requires a careful rethinking of the grounds for the WTO’s moral legitimacy to accommodate its new realities’.^{Footnote 60} There have been also important external pressures (by other supranational authorities) upon the WTO to adopt a broad, context-sensitive approach in these types of cases. In the 2006 EC–Biotech case, where the considerations at play intersected biological diversity and biosafety, the EU argued that ‘[t]he issues faced by the Panel have to be taken in their broader context. … [A] failure by the Panel to have regard to this broader context will risk undermining the legitimacy of the WTO system’.^{Footnote 61}

Just as we observed in the context of the ECtHR, teleological and context-oriented reasoning characteristic of the proportionality analysis ‘favors a debate among alternative normative preferences in the interpretation of a rule, [while] a simple appeal to text would hide those alternatives and preclude a debate’.^{Footnote 62} Miguel Maduro continues (not specifically in the context of the WTO): ‘Teleological reasoning fosters the conditions necessary for such a debate in which the plurality of actors of the community of judicial discourse can participate.’^{Footnote 63} Maduro’s general thesis that legal and constitutional pluralism requires an expansion of the scope of legal arguments, beyond traditional textual approaches and upon systemic and teleological ones, is well confirmed by the WTO panels’ embracing of purposeful, contextual reasoning in their rulings. By elaborating teleological, substantive arguments, the arbitrators could pave the way for a discussion on merits; the incorporation of a balancing discourse allows the panel to weigh and balance the goals of trade liberalisation with the states’ pursuit of other social values.

V. Between the statehood and supranational sphere

What do the examples produced in the two preceding sections of the article tell us about legitimacy, in the new international architecture of the world? It would be unfortunate if the conclusion to draw from these remarks were that supranationalism renders democracy redundant, because legitimacy can be supported by justificatory discipline without democracy. Rather than justification-based legitimacy eclipsing democracy, there is a potential for important synergy: statehood-level democracy supports supranational reason-giving (justificatory discipline) because in multi-level governance, states still retain their meaningful constitutional status as important, though no longer exclusive, sites of legitimacy.

Democracy at the level of statehood supports supranational justificatory legitimacy because often, at the end of the day, it is the state which ensures compliance with supranational norms. Whether (as the rationalist strand of international relations theory suggests) state compliance in circumstances of thin or no supranational enforcement can be best explained by the rational self-interest of states (based on reputational values as capital for future cooperation, and mutual expectations in long-term, iterated collaboration) or (as the ‘normative’ strand would have it) on the persuasive pull of norms affecting states’ incentives for compliance,^{Footnote 64} states may often be the best guardians of efficiency and legitimacy of supranational norm-producers.

And vice versa, supranational public reason supports and helps improve national democracy. This is perhaps best seen in the field of international human rights at the state level. As Joshua Cohen and Charles Sabel note: ‘human rights claims can be presented as elements of a global standard – a global public reason, itself part of global politics – that sets out conditions of acceptable treatment, requiring in particular that political societies assure conditions of membership for those who live in their territory’.^{Footnote 65} In engaging with international human rights, legal and constitutional actors at state level commit themselves to transposing those ‘global standard[s]’, dubbed by Cohen and Sabel as ‘global public reason’, to their state constitutional structures. For instance, an argument can be made, and has been made,^{Footnote 66} that when national constitutions make reference to international law and international standards, and in particular to international human rights law, they tacitly empower – confer legitimacy upon – transnational adjudicatory bodies to scrutinise national laws and decisions, because those bodies are better positioned to ascertain and interpret transnational standards. The ECtHR, for instance, by virtue of its position and expertise, is well placed to state what are the European standards, whether they are sufficiently close to being consensual in order to restrict or exclude a state’s margin of appreciation on the matter, and in this way, engage directly with the domestic constitutional process. National constitutional references to international human rights law therefore ‘invite’ supranational human rights bodies to assume authority on a matter, with an effect on the municipal system. This happens particularly when there is a constitutional directive to interpret constitutional rights in accordance with international treaties:^{Footnote 67} such ‘coordinated interpretation’^{Footnote 68} assures restriction on the interpretive freedom of the national constitutional court which, at the least, must give strong reasons why it will not follow the international tribunal’s interpretation.

In fact, rather than a rivalry between the legitimacy of supranational authorities and those at the level of statehood, we see a superficially paradoxical situation: the rise of legitimate supranational authorities does not detract from, but often enhances democratic statehood, and hence, the democratic legitimacy of states. This occurs, first, in consolidated democracies, when for instance supranational adjudicatory authorities provide guidance about democratic best practices to ‘established’ but inevitably defective democracies.^{Footnote 69}Second, this occurs in ‘transitional’ democracies, when bodies such as the Venice Commission supply recommendations about democratic constitutionalism, or bodies such as the European Commission issue reports about the progress of candidate states within so-called political conditionality.^{Footnote 70} And third, perhaps most obviously, it occurs in the cases of so-called ‘fragile’ or ‘failed’ states, when a number of supranational entities of different character, but often with great clout, are engaged in helping democratic capacity-building. In particular, in the latter category, the ‘constitutional intervention’^{Footnote 71} may be seen (when it works well) as a happy case of introducing constitutional essentials to a troubled state, in accordance with transnational standards, whether foreign, international or supranational (as was the case of Bosnia Herzegovina or Timor-Leste). When such an intervention respects the conditions of non-self-interest and non-bias of foreign interveners, and at the same time there is an awareness on the ground that no successful constitution-making would have occurred in the absence of intervention, the national constitution-making process and its product acquire domestic legitimacy not despite but largely thanks to its alignment with supranational standards.

More generally, international and supranational standards often provide a backup for democracy at the state level: not only in the instances of repairing specific national dysfunctionalities, as indicated in the paragraph above, but by supplying an extra dimension of limiting domestic powers and contributing to checks and balances. As Kumm noted, ‘international legality … has the tendency to limit the options of the executive branch to claim foreign affairs prerogatives and thereby shift the power of the executive branch in a way that endangers and potentially destabilizes democracy on the national level’.^{Footnote 72} This is perhaps a hypothesis which would need to be tested by specific case studies, but as a general proposition it is eminently plausible: international structures of norms and institutions provide an extra factor of visibility, accountability and responsibility for state actions. Even if international law often leaves the specific procedures of ratification or implementation of international agreements to national constitutional arrangements, those actions are necessarily catapulted into the international sphere with the effect of enhancing national separation of powers, federalism, checks and balances, and other constitutional guarantees. Abuses of power by domestic actors become less invisible and lose much of their impunity when they become a matter of participation in an international process, and when other domestic actors and supranational entities become partners in the process.

We should therefore resist the temptation to perceive the supranational/national dynamic as a zero-sum game:^{Footnote 73} states are not displaced or sidelined by a new supranational authority, and their democratic qualities are not compromised. There is often mutual enhancement, and the challenge for constitutional legal scholarship is to re-conceptualise this coexistence. In many contexts and for many purposes, states constitute an optimal unit within which people solve their collective problems, produce public goods, and commit to accept sacrifices related to those public goods. States often constitute frameworks for an optimal equilibrium between the capacity of individuals to feel an emotional sense of togetherness and solidarity on the one hand, and requirements of political institutions to solve problems at a sufficiently high level on the other. This may mean that in a predictable future, states as we know them will remain privileged fora in which preconditions for democracy (especially electoral democracy) prevail. But it does not mean that legitimacy in the new-world constitutional architecture has to remain solely at the mercy of democracy, often questionable and defective, in nation states.

VI. International and constitutional law

Focusing on legitimacy in the supranational sphere, and uncoupling legitimacy from the preoccupation with democracy of an electoral kind, helps us liberate ourselves from two problems which have been haunting international law scholarship: the ‘existential’ problem (‘Does international law exist? Is it a real law’),^{Footnote 74} and the related problem of the relationship between international law (insofar as it is a real law) and municipal law (monism versus dualism/pluralism, etc). Both these problems boil down to the allegedly derivative, parasitic nature of international law (resting, as it has been predominantly believed, upon states’ voluntary commitments and customary practices) and the allegedly primary status of the constitutional law of nation states. The image has been based upon a Westphalian conception of a system of independent and sovereign states, with international law deriving its validity and binding force from states’ consent, supplemented by customary international law, as detected in actual state practices. The irony of the matter is that, of course, at the outset of the Westphalian model (going back to the 1648 Treaty of Westphalia) the states making up the overall international system were anything but electoral democracies. But with time, the separation of international and constitutional law, with the privileging of the latter, acquired added force by an appeal to democratic arguments: since there are so few traces of democracy in the international system or supranational governance, the argument goes, constitutional law must be primary, in the way just outlined, in order to ensure the primacy of democratic self-government at any level of authority.

But the idea of the primacy of constitutional over international law (an idea, let me repeat, favoured by insistence on democratic pedigree as the privileged source of legitimate authority in a multi-level system of governance) is deeply flawed, for at least four reasons. First, it presupposes that there is a clear cleavage between constitutional and international law, which is plain wrong. Rather than a break between the two systems, there is continuity and there are multiple points of intersection. For one thing, there is a significant and growing overlap between the subject matter of international law and constitutional law. The former has taken up much of the agenda traditionally thought reserved for the law of the states (human rights, criminal matters, food safety and agriculture, environment, banking, labour relations, etc). In fact, the expansion of the scope of international law has been so significant that it is hard to think of exclusive domains of municipal constitutional law which would under all circumstances be invisible to international law. Most strikingly, democracy itself, traditionally seen as the ideal whose locus is firmly based in state practice, is increasingly seen as a demand of international law, with the ‘emerging [international-law based] right to democratic governance’.^{Footnote 75}

Second, the idea of the primacy of constitutional law is based on the proposition that the validity of international law is essentially based in state consent or state practice: an idea contradicted by a proliferation of supranational authorities whose norm-making, norm-interpretation, and norm-enforcement is not embedded in state consent. The phenomenon of the ‘autonomisation’ of various international organisations which produce norms binding on the states and yet which are not underwritten by prior state consent (as discussed earlier in this article) is the most telling symptom of this fact. Various international organisations arrogate to themselves, with no or very little effective protest by member states, powers that had not been originally conceived for them, and effectively enforce their authority over member states. The consent-centred picture of international law is hard to reconcile with either the idea of jus cogens – peremptory, non-derogable rules which apply to all actors regardless of their consent to them; or with the idea of the international law of human rights – premised as it is on the idea of universality, regardless and, if necessary, despite a state’s will.

Third, the idea that international law may be binding merely by virtue of state consent (expressed in the ways prescribed by the constitutional law of the states) sounds odd. The proposition that state consent is binding on the consenting state must be in itself based on a prior normative proposition that expressions of consent produce binding effects on the consenters, and if its source is in the state consent (for instance, in the states’ endorsement of the pacta sunt servanda rule), we face a problem of infinite regress. In a classical work, Thomas Franck had argued precisely that: ‘the obligation of pacta sunt servanda itself cannot be derived from state consent’ and therefore we should search for ‘another basis of common obligation, another “source” of legitimate authority’.^{Footnote 76} And this other source must be other than the state’s consent because we normally hold a state to be obligated under an international legal norm, regardless of whether it had adhered to a treaty stating that commitments must be honoured, such as the Vienna Convention on Treaties.^{Footnote 77} In a posthumously published article on philosophy of international law by Ronald Dworkin, the same point is expressed in opposition to a positivist conception of law: it will simply not do to say that states should ‘follow the principle that what they regard as law is law’ (which is, we may add, an equivalent of the consent theory) because they ‘need some other standard to decide what they should regard as law’ – and if we try to solve the dilemma by saying that the states ‘accept the principle that what other nations accept as law is law’ (as, we may add, in the custom-based theory), then ‘the other nations that each nation treats as making law for it need a test of what to treat as law for themselves’.^{Footnote 78} An appeal to legitimacy at the international level as a starting point in the chain of legitimation helps us avoid this conundrum: a way of breaking out of the circle of ‘law is what is regarded by the states as law’. The circularity necessarily involved in a consent theory of international law (what are the grounds for consent-based obligations, other than an unhelpful construction of second-order consent?) can be only overcome by an appeal to non-consent based principles.

Fourth, and most importantly from the point of view of this article, there is absolutely no reason, if we adopt a legitimacy perspective, to begin our journey of legitimation from the level of nation state, and only once we validate its norms as legitimate, ‘move up’ to the level of supranational authorities in order to confirm, derivatively, their legitimacy, rather than proceeding the other way round. We might as well, when we pursue a normative argument about legitimacy in multi-level governance, begin from the international level, and derive normative sources of state legitimacy from international legal norms. After all, a great number of norms regarding the legitimacy of states – concerning, for instance, legal criteria of statehood, including criteria of self-determination, permanence, and sovereignty, standards for eligibility for membership in the UN which is these days an ultimate ‘validation’ of a state qua state, or of the continuity of states and state succession, as well as standards of recognition of states by other states – are all norms of international law. There has been a lot of academic discourse lately about international law (including the UN Charter) constituting a sort of constitutional law for the international community,^{Footnote 79} but this is the case in a literal rather than metaphorical sense for precisely this reason: international law may be seen to constitute the legal bases for national statehood, including supplying a rule of recognition about what should count as a validly legal source of international law, such as the state’s consent. To the extent that the state’s consent is a source of international law obligations (as discussed just above), in itself it derives its force from a secondary rule of international law. In the normative inference we may therefore proceed top-down rather than bottom-up: we may see the legitimacy of states (their institutions, norms and practices) as dependent on legitimate norms of international law, rather than vice versa.

The fixed point from which we start in the chain of the normative argument about legitimacy, and the resulting direction of the argument, either top-down or bottom-up, is a matter of normative choice, not of reflecting an empirically knowable reality. In fact, the ‘choice’ is not really there, no more than in the chicken-or-egg dilemma. An inter-level normative discourse about legitimacy is reflexive rather than a one-way street. We go to and fro between different levels of governance in which we identify different traces of legitimacy: electoral, deliberative (public-reason related), and other. No single direction, no particular starting point in this ‘reflective equilibrium’ of worldwide legitimacy discourse should be a priori privileged, and particular instances (rather than at the level of a general theory of relationship between international and municipal law) will be context-dependent, as a function of the nature of the issue we face.^{Footnote 80} The temptation to adopt democratic legitimacy in and of national statehood as a strong fixed point, which informs all other legitimations of authority at any level of governance, is understandable. The value of democracy is great, and rightly so. Additionally, many still see the international-law based ‘right to democratic governance’ as embryonic. But the temptation should be resisted because it drives us into anachronistic, unhelpful and sometimes incoherent conclusions about relationships between different levels of authority, such as the four problems of the alleged primacy of constitutional over international law, just depicted.