When the editors of this journal asked me to write a comment on Jaye Ellis's paper, I was immediately fascinated by the prospect of engaging in a rigorous debate, perhaps one as controversial as that between Hans Kelsen and Eugen Ehrlich on legal scholarship and legal sociology about 100 years ago – although I am unable to reach the brilliance of those thinkers, and probably also their level of polemics.Footnote ¹ Now that I have read Jaye's paper, my expectations have somewhat changed. I realize that Jaye and I agree on many important points – points that would have brought large parts of the ‘invisible college’ up in arms a few decades ago. In particular, we agree that legal scholarship should find ways of dealing with the multiplicity of soft-law instruments that look like law because they consist of sometimes quite precise rules; that function like law because they effectively guide the behaviour of states, international organizations, and private entities; and that are therefore not adequately described as ethical, political, or moral rules. We also agree that the traditional concept of international law, which considers state consent as the one and only source of legitimacy, has come under stress since the advent of global governance, particularly because other actors like international bureaucracies or non-state entities play increasingly important, indispensable roles.

While some of the grand themes of past disputes seem to be settled among us, new themes arise. One new theme on which I would like to focus in this comment is the question of the adequate perspective for a legal conceptualization of global governance. I believe that Jaye and I disagree on this question in several respects. First, although Jaye and I would both uphold a conceptual distinction between law and politics, Jaye advocates sticking to the law/non-law divide, while I propose the concept of international public authority as an overarching framework for all kinds of governance instruments. I try to explain the usefulness of this concept with the example of non-state actors (section 1). Second, I maintain that, under the umbrella of the concept of public authority, the concept of soft law still has some normative value in that it allows distinguishing different modes of authority that might require different legal frameworks in order to be legitimate (section 2). Third, Jaye and I seem to follow different ideas of legitimacy. Jaye has sympathy for the proposals by Kingsbury and Teubner, while I favour discourse theory. While Jaye rightly elaborates the problems of global democracy, I would like to explicate how I think that some of them might be solved from the point of view of discourse theory (section 3).

1. International public authority and non-state actors

Jaye points out that rule-making by non-state actors constitutes probably the greatest challenge to the concept of international law. Compared to Jaye's, my own paper admittedly leaves non-state actors in the shadow, focusing instead on international organizations and other inter-governmental fora. This has drawn some criticism,Footnote ² and I think the critics are right. The concept of international public authority, which claims to provide an adequate legal response to global governance, would be inchoate if it did not pay attention to non-state actors. I would therefore like to take this opportunity to elaborate on what the concept of international public authority might contribute to the legal conceptualization of soft law agreed by non-state actors.

My argument is that we should also ‘cut off the head of the king’ when we approach soft law produced by non-state actors. In my view, the decisive question is not whether such instruments are law or not, but whether they are forms of public or private authority. Focusing on the legal nature of these instruments alone would not bring us anywhere near a solution regarding issues of legitimacy. Probably, most of these instruments, like the standards of the Forest Stewardship Council that Jaye mentions, can readily be qualified as law, in whatever way one wants to define it: they generate normative (not cognitive) expectations for those who adhere to them, because violations are likely to result in reputational losses or other forms of ‘soft’ enforcement like the withdrawal of benefits;Footnote ³ they are the result of a broadly formalized procedure (unlike moral norms); and they contain specific rules (unlike merely political documents). Therefore, unless one insists that only rules that can be enforced by armed forces or policemen are law, and Jaye elaborates on why we should not do so, one would have to qualify such instruments as law.

But what would be the benefit of qualifying the standards of the Forest Stewardship Council as law? Would it mean that violations of those standards would give a state or another third party the right to impose sanctions, such as reprisals? Certainly not. Would it imply applying the same principles of public law like transparency, participation, representation, and review that apply, for example, to domestic administrative regulations? Not necessarily. In general, private actors are at liberty to regulate their affairs by consent. Every association may adopt its by-laws. Legitimacy becomes an issue only if non-state actors affect the liberty of those not parties to that act, be it the state, a group of citizens, or the international community as a whole. The concept of international public authority is supposed to define this point.

Under which conditions should the standards developed by the Forest Stewardship Council be considered acts of international public authority? To make things more difficult, this question needs to be answered from an ex ante perspective:Footnote ⁴ only if the Forest Stewardship Council has the chance to know whether it is exercising public authority before it adopts a standard can it respect public-law principles and ensure, for example, that all relevant stakeholder groups participate in the decision-making process. In my article, I call this the internal perspective. In the first place, one should therefore discuss the criteria that might ex ante be indicative of an act of public authority. Possible criteria include the nature and composition of the entity adopting the standard, the nature of the problem to be tackled (e.g., whether the standards are meant to protect a public or common good), the role those standards are supposed to assume in the political process (e.g., whether they are based on a delegation or pre-empt or anticipate government regulation), and the presence of (soft) enforcement mechanisms.Footnote ⁵ Taking all these criteria into consideration, one might perhaps come to the conclusion that the standards of the Forest Stewardship Council constitute a form of public authority, their private-law basis notwithstanding. By contrast, any recommendations adopted by a non-governmental organization (NGO), such as Amnesty International, will normally not amount to acts of public authority. Although they might address public or common goods, Amnesty might not be influential enough, and it probably does not aim at replacing or pre-empting government regulation, but at triggering political processes that will put in place such regulation. The approaches suggested by Kingsbury and Teubner, as I understand them, do not allow for such a distinction.

2. Is soft law a meaningful concept?

Even though I argue that the concept of public authority should be the primary filter for a legal approach to soft law, I see some merit in the distinction between hard and soft law. Jaye rejects this distinction, arguing that the term ‘soft law’ had too many different meanings. I agree that soft law should not be (and actually is less and less) used to designate vague provisions. Since the meaning of words is determined by the context of their practical use,Footnote ⁶ a certain degree of vagueness is inherent in every legal rule. But, apart from this, I actually see some benefit in applying the distinction between hard and soft law as an additional filter in order to further differentiate the myriad of acts of international public authority. Hard law has a particular function that is worth singling out, all else being equal with soft law: only hard law provides a cause of action before the International Court of Justice (ICJ), and only violations of hard law may give rise to damages. True, not all rules of hard law fulfil these criteria, such as some organizational rules or rules of procedure, but those rules support the rules to which the above description of hard law applies.

Having said that, I would like to emphasize the need for further distinctions beyond the soft/hard law divide in order to determine how exactly a particular instrument affects liberty. This would allow putting in place principles of public law addressing issues of competence, procedure, transparency, or review, which are commensurate with the specific legitimacy challenge constituted by this instrument. This is what the concept of ‘standard instruments’ suggested at the end of my paper is supposed to achieve. For example, soft-law instruments that specify the details of a reasonably specific, universally ratified treaty might raise fewer concerns than soft law adopted in the absence of any international treaty on the subject, and maybe even of any shared understanding on the subject matter.

By contrast, I do not think it would be appropriate to consider soft-law instruments as ‘legal facts’. This would blur the distinction between soft law and entirely non-legal instruments, such as indicators, statistics, or other forms of information. These instruments give rise to cognitive, not normative, expectations and therefore cannot possibly be considered legal if the concept of law is to retain any meaningful significance (and although some intermediate forms might be difficult to qualify). Nevertheless, such instruments play an important role in contemporary governance. They contribute to the formation of common understandings, they raise awareness and facilitate agenda setting, and they are closely intertwined with hard and soft legal instruments. Thus, many non-legal instruments might have to be qualified as international public authority. This is another reason why I do not share Jaye's sympathy for the approaches of Gunther Teubner and Benedict Kingsbury. To make it clear: I do agree with both Teubner and Kingsbury on many issues. Their theories greatly expand our focus on phenomena that remain below the radar in traditional approaches, and I share their concerns for justice and legitimacy. However, as fundamentally different as their approaches may be, their focus is limited on legal instruments. Their concepts for the identification and legitimation of authoritative acts are premised on the ability to qualify those acts as legal acts. It seems impossible to me to apply this strategy to entirely non-legal acts like indicators. At this point, the idea to use a rule of recognition (Kingsbury) or to use the ability of the legal system to connect with its social environment (Teubner) in order to discipline or legitimize such acts breaks down. Non-legal instruments thereby remain in a grey area that is inaccessible to public law. Not so under the concept of international public authority.

3. International public authority and global democracy

Jaye raises doubts as to the feasibility of a democratization of the international order. This is an important point that I neglect in my paper. I should therefore set out why I recommend discourse theory as a model for legitimating international public authority, and how it might work.

In my opinion, discourse theory has a number of advantages over other theoretical approaches. First, in contrast to the approaches suggested by Kingsbury, Klabbers, Toope, and Brunnée, it stands in a positivist tradition by making a clear conceptual distinction between law and morality, law and legitimacy, and law and politics. Although this distinction does not fully immunize the law against any unchecked power, it represents in my view a good compromise between the need for effective decision-making and for allowing the legitimacy of those decisions to be questioned: discourse theory does not silence questions of legitimacy, but relegates them to separate discourses about the preconditions for the exercise of authority.

Second, discourse theory offers a promising account of the relationship between legal authority and legitimacy: one may follow the law because it is enforced, be it by courts, armed forces, reputational sanctions, or group pressure. In the alternative, one may follow the law because it may be presumed to be rationally acceptable because it results from fair deliberations in representative institutions that respect fundamental rights guarantees for minorities.Footnote ⁷ Law in this understanding is more than enforceable rules, but still an important step away from morality. Joseph Raz's idea that law claims to be legitimate may be extended over all forms of public authority.Footnote ⁸

I do not think that systems theory offers a comparable theory of legitimacy. Gunther Teubner has made a remarkable contribution in that respect in a recent article.Footnote ⁹ I do not see, however, how a legal system that is understood as operationally closed should be able to produce some kind of universal, integrative justice,Footnote ¹⁰ the alleged impossibility of which being the reason why Teubner rejects Habermas's and Rawls's theories of justice.Footnote ¹¹ In any case, the end result, the postulate that justice might follow from the capacity of the idea of justice to transcend the otherwise closed operations of the legal system,Footnote ¹² seems to come very close to Habermas's idea of discursive justice, that of the closure and reopening of fundamental contestations in the course of a more hierarchical model of law-making.Footnote ¹³

The crucial issue is, however, whether discourse theory is workable on a global scale. Optimally, discourse theory would require a worldwide public sphere, a common political culture for opinion formation, and representative central institutions that would transform societal discourse into political decisions. Habermas has suggested a reformed UN General Assembly representing both states and citizens as the centre of a new institutional architecture on a global scale. As regards the problem of a worldwide public, Habermas argues that non-governmental organizations might facilitate opinion formation in society, but not to a sufficient degree. He therefore emphasizes the existence of moral consensus on the undesirability of war and the ability of international institutions to improve the economic and social circumstances of people under their authority.Footnote ¹⁴

I have some doubts about this line of argument. First, I find it difficult to imagine a General Assembly that would be representative of seven billion people and about 200 states. Also, Amartya Sen has recently emphasized the weaknesses of representative government. The preferences of the electorate are too diverse to find satisfactory expression in one single vote.Footnote ¹⁵ Further, the Arab Revolution demonstrated once again how controversial military interventions might be. I therefore consider it necessary to think of additional ways of legitimizing international public authority. One of them is accountability. Buchanan and Keohane distinguish narrow, institutionalized accountability mechanisms according to predefined standards, and broad accountability, which emphasizes transparency and general responsiveness to public concerns.Footnote ¹⁶ Narrow accountability might mitigate the lack of representative global institutions, while broad accountability would strengthen public debate. This understanding of accountability is similar to the idea of public reasoning that Sen, drawing on Rawls, identifies as the core of democracy.Footnote ¹⁷ Certainly, this model is still a far cry from the reality of some international institutions and, as a theory, this lacks the beauty of grand design. I am therefore inclined to consider this attempt to extend discourse theory to a global level as the worst solution – except for all the others.