My thanks go out to Yuval Shany, Vaughan Lowe and Irini Papanicolopulu for their comments on my book. It is truly a pleasure and a privilege to engage them in this discussion. Let me begin by responding to some of the points made by Vaughan. I fully agree that the rights set out in human rights treaties could perhaps be reconceptualised as pledges within the framework given by Lea Brilmayer in her BYBIL article; they are not simply reciprocal bargains between states. And I certainly agree that the treaties could – like domestic constitutions – be seen as limiting the powers of governments on the basis of fundamental principles. But that reconceptualization does not necessarily entail that these principles are territorially unbound. After all, issues that mirror the extraterritorial application of human rights treaties have also arisen with respect to the extraterritorial application of domestic bills of rights. In the final analysis, the scope of all these instruments depends on underlying ideological or value judgments – e.g. should citizenship matter in determining whether a state could take an individual’s life or deprive him of liberty on a preventive basis, a debate of great relevance in the United States today.
Turning now to Yuval’s comments, he and I are in basic agreement as to the causes of the confusion and conflicts in the case law, founded as they are in the underlying tension between universality and effectiveness. But even if he agrees with the diagnosis, Yuval takes issue with my prescription – the model with distinguishes between positive and negative obligations, and applies a territorial control requirement to the former but not to the latter, which it treats as territorially unlimited. Yuval argues – quite persuasively – that my model would also lead to some arbitrary results, as in the Ecuador v. Colombia example, where Colombia would under my model not have the duty to ensure the human rights of the people of Ecuador endangered by transboundary harm emanating from the activities of private persons operating from Colombian territory. He opts instead for a ‘a single concept of jurisdiction, applicable both to negative and positive obligations, which centers on the strength of the governmental power that is being applied or can be applied vis-à-vis the individuals in question’, a flexible, functional criterion which would in essence mean that the state would have a particular obligation as soon as it gained the ability to comply with it or violate it. (Note, of course, how in the Colombia example Colombia may have some power over the private perpetrators of human rights violations, but has not exerted any power over their victims – and it’s the victims who have to be subject to its jurisdiction).
In replying to Yuval, let me start off by conceding (as I do in the book) that my proposed model is not perfect. It is not ‘perfect’ in the sense that it would still lead to some outcomes that would appear to be morally arbitrary.Yuval has produced one, but there may well be others. From a universality standpoint my model would accordingly also be flawed – perhaps not as flawed as its competitors, but flawed nonetheless. The only completely non-arbitrary model from the standpoint of universality would be the one that Yuval offers – states would have obligations whenever they could actually fulfil them. But that lack of arbitrariness comes at a significant price: the total collapse of any threshold even for positive obligations. I fail to see how Yuval’s ‘points of contact’ analysis imposes any meaningful constraints on the scope of state obligations; as far as I can tell, it does collapse into nothingness. And that makes perfect sense. It is almost by definition impossible to construct a threshold which is not underinclusive. The only way to completely dispense with arbitrariness is to dispense with the threshold altogether and look at the merits of each particular case, as I indeed precisely want us to do with regard to negative obligations.
So why not apply the same reasoning to positive obligations as well? I could give several more or less formal reasons – for example, that a requirement written in the treaty can’t just be interpreted away completely, that some treaties, like the CAT, which refer exclusively to jurisdiction over territory as a precondition for various positive obligations, cannot be reasonably read in the way that Yuval suggests, or that other parts of international law – e.g. environmental law – would be far better suited to deal with any aberrant cases that slip through the cracks. But such formal objections are not what concerns me right now. I think that my model is superior to Yuval’s not because it better takes into account considerations of universality – it doesn’t – but because it does better take into account considerations of effectiveness, and it is only this mix that can make a model successful. How does it do that?
First, in order to be able to comply with the positive obligation to secure or ensure human rights even against violations by private actors in the overwhelming majority of situations the state will need to have control over the territory in which the victim of the violation is located. I do not dispute that there will be some cases in which such control might prove to be unnecessary, but by and large it will be. This is why it makes sense to set such control as a threshold, particularly because the meaning of ‘jurisdiction’ as control over territory is well entrenched in international practice. The threshold would thus serve an important purpose of legal clarity and predictability – states and courts would know with a fairly high degree of certainty when such positive obligations would apply and would be able to adjust their behaviour accordingly. The need for clarity should not be underestimated, and it seems to me that Yuval’s model is significantly lacking in that regard.
Second, the distinction between the positive obligation to secure or ensure human rights, which would require territorial control, and procedural and prophylactic positive obligations, which are tied to at least a substantiated allegation of a violation of a state’s negative obligations, which would be territorially unlimited, is also grounded in considerations of effectiveness. Assuming arguendo that the UK did not possess effective control under Loizidou over Southern Iraq, and hence the positive obligation to secure human rights would not have applied, it would still have been far, far more practicable for the UK to investigate alleged human rights violations by its own troops, rather than say random private murders or petty crime in Basra. Again, I do not deny that by setting the threshold some cases in which the UK could act would fall through the cracks, but that is a price worth paying, I think, for making the model clearer and more predictable – and more palatable to states. Similarly, while the Bankovic insistence on an ‘all or nothing’ application of the ECHR makes little sense with regard to negative and procedural positive obligations, which essentially require the state to regulate its own behaviour, there is a powerful argument that preserving the integrity of the human rights regime is necessary with respect to the onerous and all-encompassing obligation to institute law and order even among private actors living in a certain territory.
In sum, I am certainly not enamoured with my proposed model, but still think it combines the best of both worlds, both universality and effectiveness. But I certainly may be proven to be wrong!
Finally, in response to Irini’s comment, I have one simple reason for denying that the formal entitlement of a state to exercise jurisdiction over a territory or indeed an object should ever be relevant to the territorial application of human rights treaties. Accepting that such formalities should matter when the result would be human rights-friendly would undermine the argument that they should not matter in other circumstances where states exercise power only de facto. Sure, I could live with the result of Ilascu. But I still think that the Court’s giving of some relevance to naked title over territory alone has the potential of damaging the principled – and I think practically far more important – Loizidou holding that it is control alone, lawful or unlawful, which determines the applicability of human rights treaties. I understand how this can seem quite rigid and absolutist, but it should also be understood that there can be real danger in making exceptions. And I would add to that – so what if there are rules of international law, e.g. the UNCLOS, which oblige states to exercise their jurisdiction! If they fail to do so, these are the rules that have been violated, and a framework is in place to deal with that. This does not automatically translate into a human rights problem, nor should it, in my view, have bearing on the scope of application of human rights treaties. That said, I again accept that the model I propose would not lead to a result that we find intuitively appealing in each and every case – but I also think that no such model could really exist.