Models of Extraterritorial Application
Models of Extraterritorial Application
Abstract and Keywords
This chapter concludes the interpretative inquiry into the meaning of the jurisdiction clauses, and elaborates on the several possible models of extraterritorial application of human rights treaties. These include the spatial model of jurisdiction, which grounds the application of human rights treaties in state effective control over territory; the personal model of jurisdiction, based on various forms of state authority and control over individuals; and a mixed model which is based on the distinction between positive and negative obligations under the treaties.
1. Basic Models and Scenarios
Having examined the policy considerations which courts take into account when deciding on the extraterritorial applicability of human rights instruments, whether in an international or in a domestic context, I will now turn back to the human rights treaties themselves. Recall that in Chapter II of this study, we examined the concept of state jurisdiction in the treaties’ jurisdiction clauses, which ultimately governs their territorial scope of application. As I have endeavoured to explain, that concept of state jurisdiction refers to state control over territory, and perhaps also individuals, not to the prescription or enforcement of municipal law, as the European Court thought in Bankovic.
If the word ‘jurisdiction’ in Article 1 ECHR and similar jurisdiction clauses of other human rights treaties basically means ‘effective overall control of an area’, as the European Court held in Loizidou, then it of course still needs to be determined what degree of control suffices, and how exactly to geographically or spatially define an ‘area’.1 The rule about the territorial scope of application of human rights treaties would then be relatively simple—they apply whenever a state has effective overall control of an area, regardless of whether it has title over that area or not, and of whether its control over the area was obtained lawfully or unlawfully. Conversely, however, the treaties would not apply when such control is lacking. In my view, contra Ilascu, this would hold even for territory over which the state actually does have title. As I have argued above, there should be no difference in principle between intra- and extraterritorial application of a human rights treaty, in the absence of the requisite degree of state control.2
Yet, even if this territorial concept of jurisdiction was generally accepted, there would still be numerous cases where a state acts extraterritorially and thereby violates an individual’s rights, while not actually controlling the territory in which it does so. The Bankovic scenario provides one example—aerial bombardment (p.119) without occupation—but as we shall see there are many others. If all ‘jurisdiction’ meant was control over territory, then persons in such scenarios would simply not fall within it. Alternative strands of Strasbourg jurisprudence, the case law of other human rights bodies, and several authors thus advocate another conception of state jurisdiction, as authority and control over individuals. Such a personal, rather than territorial notion of jurisdiction would greatly expand the scope of human rights treaties. However, it remains to be seen whether it is tenable.
This chapter of the study will examine the possible models of extraterritorial application, the two prime candidates being jurisdiction as control over territory3 and jurisdiction as control over individuals.4 In my view, neither of these is entirely satisfactory. The former simply does not go far enough. If applied too strictly, it is apologetic of state power, and allows many human rights abuses to slip through the cracks. The latter runs against the language of some treaties, which explicitly conceive of state jurisdiction in territorial terms. More importantly, it cannot be limited on any principled basis and loses any meaning as a threshold, since it simply collapses into the position that a state has human rights obligations whenever it can actually violate the rights of the individuals concerned.
I therefore prefer a third model, where state jurisdiction is conceived of only territorially, but where that threshold criterion applies only to the positive obligation of states to secure or ensure human rights, because it is only when states possess a sufficient degree of control over territory that these obligations can be realistically kept.5 When it comes to the negative obligation to respect human rights, no threshold criterion should apply because states can always control the actions of their organs or agents. In my view, this model fits best with the object and purpose of human rights treaties and is superior in reconciling the various policy considerations relevant to their extraterritorial application. It is, however, not free of all weaknesses, and is lacking in explicit textual support at least with regard to some treaties.
Besides these three basic models which concern treaties that possess dedicated jurisdiction clauses, I will also be examining treaties without such clauses,6 as well as the ICCPR, with its perhaps more restrictive territoriality requirement.7 Before moving on to do so, however, I wish to outline several factual scenarios in which the issue of extraterritorial application may present itself. Having criticized the jurisprudence of the European Court for casuistry, I have no wish to engage in it myself. Rather, my purpose in doing so is to be able to test the various models of extraterritorial applications on the facts of these scenarios. This will simply allow us to see more clearly what is at stake, as a matter of policy as much as a matter of law, in opting for one model over another.
Most of the scenarios that I will now present are not actually hypothetical but have already occurred in practice, whether they have been adjudicated on or not. I will group them with regard to the nature of the state act which would arguably (p.120) amount to a human rights violation, for example deprivation of life or liberty. Where the scenarios within each group differ from one another is in the various contextual elements, for example in whether the killing by a state occurred in a territory under the state’s control, and it is precisely these contextual elements which matter the most for extraterritorial application. A caveat: there certainly is a degree of overlap between some of these scenarios, and I am not saying that they are all necessarily distinguishable. Nor am I saying that this list of scenarios is in any way exhaustive.
B. Extraterritorial deprivation of life
1. Killing in a territory within the state’s control
There is no greater exercise of state power than a deprivation of life—and there is no issue which attracts greater controversy. Legally, perhaps the easiest scenario is when a state kills—or generally violates the rights of individuals—within a territory under its effective overall control, for instance during a belligerent occupation. The jurisprudence of human rights bodies, as well as the ICJ, is clear that in such situations the human rights treaty would apply. How it would apply is a different matter.8
As we have seen above, whether a particular treaty is regional or universal should have no bearing on the application of the effective overall control of an area test. Also irrelevant for the threshold issue of extraterritorial application is whether the killing takes place during an armed conflict, whether within or outside a territory under a state’s control. This of course may have an impact on the substantive application of the treaty.
2. Killing in a territory outside the state’s control
More interesting is the issue of a killing by state agents in a territory not under the state’s control. This killing can take place during a large-scale use of armed force, for example in an aerial bombardment as in Bankovic, or on the battlefield, or for instance during patrol by troops in an area over which their control is arguably tenuous, as in Al-Skeini.
But again, the existence vel non of large-scale violence seems irrelevant for the threshold question of extraterritorial application. Killing can also take place in more limited, ‘targeted’ circumstances, as with the US use of drones in Pakistan to target Taliban or Al-Qaeda elements.9 Then of course there is the classical assassination (p.121) scenario, as with the 2010 killing of a Hamas military leader in Dubai by suspected agents of the Mossad.10 What all of these cases have in common is that they would fall outside a human rights treaty’s scope of application under a strict effective overall control of an area model of state jurisdiction.
3. Killing in a territory outside the state’s control, but within the territory of a state party to the human rights treaty in question (espace juridique)
But what if a Bankovic-type scenario were to occur in a territory of a state party to the treaty, for example, when Russia bombed parts of Georgia in 2008? Similarly, consider the assassination of Alexander Litvinenko in London, and assume that it was done by Russian agents. Was Litvinenko within Russia’s jurisdiction? Did Russia have an ECHR obligation not to kill him, or the obligation to investigate his death?
Litvinenko’s family has lodged an application with the European Court, so we will eventually see what the Court will say. To my mind, the only thing distinguishing the Litvinenko-type scenario from Bankovic is that the killing took place within the ECHR’s espace juridique. It is irrelevant that the killing was done by radioactive sushi, rather than by a missile from a drone or a bomb from an airplane. However, while I am sure that Litvinenko’s family’s counsel will precisely try to distinguish Bankovic on espace juridique grounds, as we have seen that notion should also be irrelevant for the question of extraterritorial application. In other words, if Bankovic governs state action outside Europe, then it should also govern it inside Europe.
4. Killing by third parties
Finally, so far we have only examined killing by state organs or agents, which engages the state’s negative obligation to respect human rights. However, killings by third parties can engage the state’s positive obligation to do all it reasonably can to prevent such killings, and the obligation to investigate them. A good example would be insurgent violence in occupied Iraq. The issue, of course, is when such positive obligations arise. One could also envisage various complicity scenarios, dealing with the state’s responsibility connected to acts by third parties. Would, for example, states selling weapons to Sudan that they know or have reason to suspect Sudan will use against civilians in Darfur have duties under human rights treaties towards the inhabitants of Darfur?11
(p.122) C. Extraterritorial deprivation of liberty
1. Detention in a territory within the state’s control
When it comes to the deprivation of liberty, as with deprivation of life, the easiest scenario would be detention by state agents in a territory under the state’s effective overall control. Perhaps the most common example would be the detention of civilians or combatants during belligerent occupation. Of course, there are complexities in defining what actually amounts to effective overall control over territory, and even more interestingly what counts as a territory. This brings me to my next example.
2. Detention in a place within the state’s control
One can easily imagine detention by a state which does not take place in a territory under its control, but in something more approaching a place. Take as one example a full-fledged military base, such as Guantanamo or Bagram. For my present purposes, perhaps the most intriguing example is offered by the so-called ‘black sites’ that the CIA operated in the war on terror, in order to render, detain, and interrogate high-value detainees, until they were shut down by the Obama administration. Though many specifics are still unknown, credible sources have asserted that the CIA operated such sites in Lithuania, Poland, and Romania among others.12
This example brings the territorial conception of state jurisdiction as the Loizidou ‘effective overall control of an area’ to its breaking point. Does a former riding school in the suburbs of Vilnius—one of such black sites—qualify as an ‘area’ under US jurisdiction, even though it is located in a territory under the control of Lithuania? Forget enormous military bases such as Guantanamo—what about lone buildings, or apartments or basements within such buildings? Do they fall within the ‘jurisdiction’ of the detaining state?
Or consider various abduction scenarios, as for instance with the capture of Adolf Eichmann by Israeli agents in Argentina in 1960. To briefly recount the facts, on 11 May 1960, as he was returning home from work, the disguised Eichmann was intercepted by Israeli agents who knocked him unconscious, put him in a car, and took him to a safe house in Buenos Aires. He was kept there until 21 May 1960 when he was smuggled out of Argentina on an El Al flight to Israel.13
(p.123) Now, because the Israeli agents operated in Argentina without its knowledge or consent, they violated Argentinean sovereignty.14 But that aspect of the legality, vel non, of Eichmann’s capture is for our present purposes beside the point. The question I am interested in, on the facts of this scenario, is at what point Eichmann would have had enforceable human rights vis-à-vis Israel. Would it be at the moment of capture, when he was as an individual within the Israeli agents’ control? Or when he was brought to the safe house, arguably a place under Israel’s control, and thus within its jurisdiction? Or was it when he was put on the El Al plane, or rather when that plane entered Israeli airspace or touched down on the tarmac? This brings me to my next example.
3. Detention on a ship or aircraft
What about persons detained aboard a ship or aircraft? Does it matter, for example, that Eichmann was drugged, disguised as flight crew, and brought to Israel on an El Al plane, which flies Israel’s flag, since general international law recognizes the flag jurisdiction of states? Or would the result be any different if Eichmann was taken on a British Airways plane? Is it, in other words, the flag that matters, or is it the actual control over the individual by the state?
In more contemporary terms, the United States has held a number of prisoners aboard ships in its ‘war on terror’.15 But where this particular issue is of the greatest practical relevance is probably with respect to piracy off the coast of Somalia. When do, for example, captured pirates enjoy the non-refoulement protections of the ECHR or the ICCPR? And when do refugees or economic migrants aboard ships trying to reach a better life on European or American shores?
4. Killing, torture, or ill-treatment during extraterritorial detention
I have so far talked about detention pure and simple. Yet, as we well know, deprivation of liberty is frequently followed by various forms of violence to the person. This, after all, was what the US programme of interrogation of high-value detainees, such as Khalid Sheikh Mohammed, the mastermind of 9/11, was all about. The political importance of such scenarios is not in doubt. However, the threshold question of extraterritorial application probably does not and should not depend on the particular act of violence concerned. What does is the fact that such violence takes place in a situation where the individual is under state control, or is in an area or place under state control. In that regard, the violence as such is inseparable from the deprivation of liberty—but what if the torturer is an agent of a different state than the one holding the detainee? And what of the situations where one state is merely complicit in a human rights violation by another?
(p.124) D. Complicity scenarios
1. Territorial complicity
There are two basic complicity scenarios, the first being when the territorial state acquiesces or is complicit in an extraterritorial human rights violation by another state. Thus, for example, if Lithuania allowed the United States to operate the CIA black site near Vilnius, knowing that it would be used for the incommunicado detention and coercive interrogation of suspected terrorists, it may be responsible under the relevant human rights treaties, such as the ECHR, because the individuals concerned were within Lithuanian jurisdiction. The same might go, for example, for states knowingly allowing rendition flights to overfly their territories, or to use their airports for refuelling.
2. Extraterritorial complicity
But complicity itself can also be extraterritorial. Consider, for example, the actions of various intelligence services in the ‘war on terror’, such as the involvement of Canadian agents in the rendition to Syria of Maher Arar,16 or in the case of Omar Khadr, or that of UK agents in the case of Binyam Mohamed, both of whom were detained and interrogated by the US in Guantanamo, with the interrogations being conducted with the connivance, presence, or participation of Canadian and UK intelligence services.17 This was of course not an isolated practice—there have been numerous allegations of UK complicity in torture in Pakistan, Egypt, and Uzbekistan, in addition to Guantanamo.18
What distinguishes these cases from territorial complicity is that the individual concerned is not located within a territory or even a place under the complicit state’s control, but is rather under the control of the principal. The actions of the principal state may themselves be committed extraterritorially, as with the US in Guantanamo. The issue, of course, is whether complicity in such circumstances is actually prohibited. Did, for example, Binyam Mohamed have ECHR rights vis-à-vis the UK while he was being held and mistreated by the US in Guantanamo, with UK collusion?
3. Distinguishing between primary and secondary rule complicity
In that regard, a distinction must be made between complicity based on primary rules of international law, such as those in human rights treaties, and complicity (p.125) based on the general, secondary rules of state responsibility. The content of the latter has been codified by the ILC in Article 16 ASR:
A State which aids or assists another State in the commission of an internationally wrongful act by the latter is internationally responsible for doing so if:
(a) that State does so with knowledge of the circumstances of the internationally wrongful act; and
(b) the act would be internationally wrongful if committed by that State.
Note that Article 16 requires that both the principal state and the complicit state be bound by the same legal obligation. As the ILC puts it:
The third condition limits article 16 to aid or assistance in the breach of obligations by which the aiding or assisting State is itself bound. An aiding or assisting State may not deliberately procure the breach by another State of an obligation by which both States are bound; a State cannot do by another what it cannot do by itself. On the other hand, a State is not bound by obligations of another State vis-à-vis third States. This basic principle is also embodied in articles 34 and 35 of the Vienna Convention on the Law of Treaties. Correspondingly, a State is free to act for itself in a way which is inconsistent with obligations of another State vis-à-vis third States. Any question of responsibility in such cases will be a matter for the State to whom assistance is provided vis-à-vis the injured State. Thus it is a necessary requirement for the responsibility of an assisting State that the conduct in question, if attributable to the assisting State, would have constituted a breach of its own international obligations.19
This creates a two-fold problem. Let us take the ill-treatment of Binyam Mohamed in Guantanamo, in which the UK was allegedly complicit, as an example. Assume that Binyam Mohamed lodges an application against the UK with the European Court, alleging that the UK’s complicity was a violation of Article 3 ECHR. The first problem is that, unlike say with the CAT, the US and the UK were not bound by the same legal obligation, since the US is not a party to the ECHR.20 Secondly, even if the first problem were to be resolved, it would be far from clear that the UK was bound to respect or secure Mohamed’s rights as he was not located in a territory under its jurisdiction. Therefore, even if the case was looked at under the CAT, the two states might not have had the same obligations towards the same individual.
The requirement for parallel state obligations thus poses a great difficulty in relying on Article 16 ASR to deal with complicity scenarios.21 Rather, recourse must be had to complicity derived from primary rules, i.e. from the positive and negative state obligations under the human rights treaties themselves. In other words, it is only if Article 3 ECHR is interpreted as prohibiting state party involvement in torture or ill-treatment by non-parties that complicity would in (p.126) fact be unlawful. This, of course, still leaves the question of extraterritorial application. If, per Bankovic, killing a person in a territory outside the state’s control would not engage the ECHR, why would helping a third state torture a prisoner do so? And even if the notion of state jurisdiction was not solely spatial, but was also personal in nature, i.e. if the jurisdiction threshold would be satisfied if a state had authority and control over an individual, would the mere provision of information or questions to third-party torturers qualify as such control?
E. Extraterritorial law enforcement
This brings me to my next category of scenarios, which I will put under the general rubric of extraterritorial law enforcement. We have already seen one example in Hape and Verdugo-Urquidez—extraterritorial searches and seizures for the purposes of criminal investigations.22 Such acts can be done under colour of law, for instance under a warrant and/or with the consent of the authorities of the territorial state, but they can also be done without such legal authority. Likewise, abductions, which we have already examined, could also have a place in a criminal law context as the means of obtaining custody over a fugitive, as was indeed the case with Eichmann.23 The question is whether this law enforcement context is of any relevance.
Consider also two further examples which are far from infrequent in practice. First, in absentia trials of persons who are not located in the territory of the prosecuting state, either because extradition has been refused or because the fugitive is simply very good at running. Would a person subjected to such a trial, as is possible in most civil law systems, actually enjoy the right to a fair trial under human rights instruments? After all, he is not within the jurisdiction, if jurisdiction is conceived of territorially. Or, rather, is the fact that an in absentia trial is a legal process dispositive for some sort of personal notion of jurisdiction?
Secondly, what about the seizure of domestic assets, such as bank accounts or immovable property, of a person located outside the state’s territory, for instance because of tax evasion, or because he is suspected of involvement in terrorism? Would that person, even though he is outside the jurisdiction, be entitled to due process, or the protection of his property rights, say under Article 1 of Protocol No. 1 to the ECHR? Should the fact that the seizure is a legal process matter? What if, say, the individual’s house was simply burnt down by state authorities as a punitive measure, without any legal process? The intuition of most people would probably be that human rights guarantees should apply in such circumstances. But why? What (if anything) distinguishes these acts from, say, an extraterritorial search or seizure, or even a killing à la Bankovic?
F. Transboundary environmental harm
Finally, let us consider a scenario of transboundary environmental harm. This is of course something that international law has increasingly dealt with, but it has (p.127) mostly done so either by considering such harm to violate the sovereignty of the state on which it encroaches, a general principle such as sic utere tuo ut alienum non laedas, or specific instruments of international environmental law.24 But what about the individuals affected by such harm?
Currently pending before the ICJ is the Aerial Herbicide Spraying case between Ecuador and Colombia. Ecuador alleges that Colombia has sprayed toxic herbicides within its borders, but near Ecuador, in its campaign to suppress coca cultivation, and that these toxins have caused significant transboundary harm in Ecuador itself. According to Ecuador’s application, this harm has not been confined to the environment in the abstract:
During and after each of Colombia’s spraying campaigns, for instance, Ecuador’s population in the northern boundary areas has reported serious adverse health reactions including burning, itching eyes, skin sores, intestinal bleeding and even death. Because of the non-discriminating nature of the herbicide used by Colombian authorities, there has also been serious and widespread damage to non-target plant species, including key local crops such as yucca, plantains, rice, coffee, hay and others. The consequences of the crop damage have been serious in the context of the subsistence farming needs of the local population.25
Assuming that the facts as alleged are true, would the Ecuadorians affected by the Colombian campaign have rights such as the right to life, bodily integrity, health, or food vis-à-vis Colombia under, say, the ICCPR, the ICESCR, or the ACHR?26
2. The Spatial Model: Jurisdiction as Control of an Area
Having now examined some of many possible scenarios of extraterritorial application, let us move to the first possible model of such application, that would conceive of state jurisdiction in spatial terms, as effective overall control of an area. This is undoubtedly the model with the most textual support, and as we have seen in Chapter II, such a model also fits with the general treaty practice of states, which use the term ‘jurisdiction’ to denote control over territory. Likewise, several human rights treaties, such as the CAT, explicitly refer to jurisdiction in spatial terms—e.g. ‘[e]ach State Party shall take effective legislative, administrative, judicial or other measures to prevent acts of torture in any territory under its jurisdiction’, in Article 2(1) thereof.
This model also fits best with the current state of jurisprudence. The golden standard was of course set in Loizidou:
Bearing in mind the object and purpose of the Convention, the responsibility of a Contracting Party may also arise when as a consequence of military action—whether lawful (p.128) or unlawful—it exercises effective control of an area outside its national territory. The obligation to secure, in such an area, the rights and freedoms set out in the Convention, derives from the fact of such control whether it be exercised directly, through its armed forces, or through a subordinate local administration.27
The spatial model of jurisdiction has likewise been adopted by other human rights bodies,28 as well as by the ICJ in the Wall and Congo v. Uganda cases.29 Its primary benefit, from a judicial perspective, is that it requires very little overruling of prior case law. Even Bankovic can stand on that count, if we forget its misguided reliance on the other ordinary meaning of the word ‘jurisdiction’ in international law, that which delimits the municipal legal orders of states. Likewise, the spatial model seems to reconcile the normative demands of universality and the factual demands of effectiveness, as extraterritorial application would happen when it is realistically possible, in the circumstances of state control over territory.30
The principal problem with the spatial model, however, is that on a deeper look it does not reconcile universality and effectiveness all that well. Adhering to it strictly would lead to numerous morally intolerable situations—intolerable from the standpoint of universality—in which a state acts extraterritorially but the relevant human rights treaty would not apply, as with most of the scenarios that (p.129) I have just outlined above. Even if it is better than just saying that human rights treaties cannot apply extraterritorially at all, it is still simply far too rigid. As we will see, this has invariably led either to the rejection of the spatial model in favour of other approaches by the more adventurous human rights bodies, or to its attenuation and the carving out of relatively unprincipled exceptions by the more faint of heart.
I will now first examine the object of the spatial test—what actually counts as an ‘area’ susceptible to being subjected to a state’s jurisdiction. Secondly, I will address the requirement of state control over an area. Thirdly, I will move on to control over places and the special problems posed by embassies and consulates, as well as by ships and aircraft. Finally, I will assess the viability of the spatial model more generally, and by reference to the scenarios discussed above.
B. What is an ‘area’?
The Loizidou test of effective overall control of an area makes perfect sense on the facts of Loizidou itself. After all, Turkey had occupied a third of the island of Cyprus, and it is such large chunks of land that we generally consider to be ‘areas’ or ‘territories’. But problems arise as the area in question spatially diminishes. Is a city an ‘area’ that can be subjected to state jurisdiction? How about a mere village? Or a military base? Or a prison? Or just any building, or an apartment within that building?
The more an area diminishes in size, the more artificial it becomes to our ears to call it an area. Obviously, a mere square foot cannot qualify as an area for the purpose of the Loizidou test. But why not? What exactly distinguishes a building, or an apartment within that building, or a room within that apartment, from your northern Cypruses or southern Iraqs? Why should size alone matter, and to what extent should it matter?
The limits of the spatial conception of jurisdiction certainly do matter, as we have seen above from the examples of CIA black sites and extraterritorial abductions. Indeed, the United States under the Bush administration denied the applicability of human rights treaties such as the CAT to the black sites precisely on the grounds that they were places located in a territory controlled by a third state. Recall that it was the Bush administration’s general position that human rights treaties, including the CAT, did not apply extraterritorially at all, thus rejecting even the effective overall control of an area notion of jurisdiction.31 However, though it had some support in the text and the travaux of the ICCPR, the language of the jurisdiction clauses of the CAT was different. In 2005 the CIA asked the Office of Legal Counsel (OLC), the division of the US Department of Justice which provides authoritative legal advice for the executive branch of government, to advise (p.130) on whether the Article 16 CAT prohibition on cruel, inhuman, or degrading treatment protected persons detained by the CIA in the so-called black sites.32 Under Article 16(1):
Each State Party shall undertake to prevent in any territory under its jurisdiction other acts of cruel, inhuman or degrading treatment or punishment which do not amount to torture as defined in article 1, when such acts are committed by or at the instigation of or with the consent or acquiescence of a public official or other person acting in an official capacity. In particular, the obligations contained in articles 10, 11, 12 and 13 shall apply with the substitution for references to torture or references to other forms of cruel, inhuman or degrading treatment or punishment. (emphasis added)
The ‘any territory under its jurisdiction’ clause clearly precludes an interpretation that would confine its application only to a territory over which the state has title, as the OLC realized:
By its terms, Article 16 is limited to conduct within “territory under [United States] jurisdiction.” We conclude that territory under United States jurisdiction includes, at most, areas over which the United States exercises at least de facto authority as the government. Based on CIA assurances, we understand that the interrogations do not take place in any such areas. We therefore conclude that Article 16 is inapplicable to the CIA's interrogation practices and that those practices thus cannot violate Article 16.33
So, the OLC accepts that Article 16 is not confined merely to those territories over which the state has title, but also to those over which the state has de facto authority. However, says the OLC, the CIA has informed it that it is not conducting the interrogations of high-value detainees in any such territory. Therefore, Article 16 CAT does not apply.34
Let me translate that for you: because we are holding these people in some undisclosed secret prison—be it someplace in Afghanistan, Poland, Lithuania, or wherever—in a place, but not in a territory over which we have effective overall control, the treaty does not apply on its own terms. A prison is not a territory. Like it or not, that’s not a bad textual argument, morally repugnant though it might be, and it brings the spatial conception of state jurisdiction to a breaking point.
There are only two ways around this argument, neither of which is immediately apparent as correct, at least on the text of the CAT. Either the spatial conception of jurisdiction can extend to something as small as a place under state control, or the negative obligation of state agents not to engage in ill-treatment does not depend on (p.131) a territorial jurisdiction threshold at all, even if the positive obligation to prevent ill-treatment does. I personally do not find it easy to accept that a riding school in Vilnius can be a territory under US jurisdiction, as Article 16 CAT requires, but this is not impossible, considering in particular the nature of the acts that the CAT prohibits, and that they invariably take place in a custodial setting.
Indeed, we should take note of Article 2 of the European Convention for the Prevention of Torture, which seems to have been drafted precisely to take such situations into account:
Each Party shall permit visits, in accordance with this Convention, to any place within its jurisdiction where persons are deprived of their liberty by a public authority.
And also Article 4(1) of the Optional Protocol to the CAT, providing that
[e]ach State Party shall allow visits, in accordance with the present Protocol, by the mechanisms referred to in articles 2 and 3 to any place under its jurisdiction and control where persons are or may be deprived of their liberty, either by virtue of an order given by a public authority or at its instigation or with its consent or acquiescence (hereinafter referred to as places of detention). These visits shall be undertaken with a view to strengthening, if necessary, the protection of these persons against torture and other cruel, inhuman or degrading treatment or punishment. (emphasis added)
This now brings me to Al-Saadoon, a case that ran its course through English courts35 and the European Court at a most hectic pace, that we have already mentioned above in the context of norm conflict.36 The applicants in that case were individuals detained by UK troops in Iraq. They were initially held as security detainees on the basis of Resolution 1546. While they were detained on this basis, per Al-Jedda, the Article 5 ECHR prohibition on preventive detention could be deemed to have been overridden by virtue of Article 103 of the Charter.37 But then the legal basis for the applicants’ detention changed. They were no longer held as security detainees, but were charged with specific crimes by the Iraqi judiciary.38 Their detention by British forces continued, but now with the consent and at the behest of the Iraqi government, in effect in a sort of pre-trial detention that is in principle permitted by Article 5(1)(c) ECHR.
As the expiry of their UN mandate and the withdrawal of UK forces from Iraq drew near, the Iraqi authorities requested the transfer of the applicants to Iraqi custody. They challenged their impending transfer, first before the UK domestic courts and then before the European Court, arguing, inter alia, that there was a serious risk that they would be subjected to the death penalty by hanging if they (p.132) were transferred to Iraqi custody, and that this triggered the UK’s non-refoulement obligations under the ECHR.
To make this argument, the applicants of course first had to establish that the ECHR applied to them extraterritorially. In Al-Skeini 39 the UK government had conceded that the ECHR applies to persons detained by UK forces in Iraq.40 It thus had to come up with something new to deny the extraterritorial applicability of the ECHR in Al-Saadoon, and that it did. It argued that the applicants were held by UK forces at the order of an Iraqi court, and that the UK had a legal obligation to abide by the ruling of that court and transfer them to Iraqi authorities since UK forces were present in Iraq with Iraq’s consent. Because it had no independent legal authority to detain the Al-Saadoon applicants, so argued the government, they were not within the UK’s jurisdiction within the meaning of Article 1 ECHR.
In its decision, the Divisional Court did not accept this argument, ruling that the applicants were indeed within the UK’s jurisdiction.41 It held, however, that the ECHR non-refoulement principle had to be qualified because the UK had a legal obligation to transfer the applicants to Iraq, pursuant to a Court of Appeal ruling42 that the Divisional Court thought was wrongly decided, but had to follow anyway.43 On appeal, the Court of Appeal agreed entirely with the government’s argument, finding that the ECHR did not apply because the UK had no independent legal authority to detain the applicants.44 The applicants promptly moved to the European Court, and obtained a provisional measures order prohibiting their transfer to Iraqi authorities. And, for the first time in many years, the UK government decided to disobey such an order by the European Court, and transferred the applicants to Iraqi custody.45
Al-Saadoon presents one of the best examples of unresolvable norm conflict, specifically between the UK’s obligations towards Iraq on one side, and its obligations under the ECHR on the other. I will now however deal only with the issue of extraterritorial applicability, which the Court of Appeal denied so that it could avoid the norm conflict. Note the UK government’s argument, which the Court of Appeal accepted, that Article 1 ECHR jurisdiction requires the exercise of a legal authority, something that the UK in Iraq supposedly lacked. As we have seen in Chapter II above, this argument results from the confusion, particularly in Bankovic, between the various concepts of state jurisdiction in international law: the one (p.133) that we are dealing with here most certainly does not require the exercise of a legal authority over a territory or an area. This Bankovic-induced confusion was to an extent also reflected in the admissibility decision in Al-Saadoon by a Chamber of the European Court,46 which ultimately quite rightly rejected the UK’s argument against extraterritorial application:
During the first months of the applicants’ detention, the United Kingdom was an occupying power in Iraq. The two British-run detention facilities in which the applicants were held were established on Iraqi territory through the exercise of military force. The United Kingdom exercised control and authority over the individuals detained in them initially solely as a result of the use or threat of military force. Subsequently, the United Kingdom’s de facto control over these premises was reflected in law. In particular, on 24 June 2004, CPA Order No. 17 (Revised) (see paragraph 13 above) provided that all premises currently used by the MNF should be inviolable and subject to the exclusive control and authority of the MNF. This provision remained in force until midnight on 31 December 2008 (see paragraphs 20–21 above).
The Court considers that, given the total and exclusive de facto, and subsequently also de jure, control exercised by the United Kingdom authorities over the premises in question, the individuals detained there, including the applicants, were within the United Kingdom’s jurisdiction (see Hess v. the United Kingdom, no. 6231/73, Commission decision of 28 May 1975, Decisions & Reports vol. 2, p. 72). This conclusion is, moreover, consistent with the dicta of the House of Lords in Al-Skeini and the position adopted by the Government in that case before the Court of Appeal and House of Lords (see paragraph 62 above).
In the Court’s view, the applicants remained within the United Kingdom’s jurisdiction until their physical transfer to the custody of the Iraqi authorities on 31 December 2008. The questions whether the United Kingdom was under a legal obligation to transfer the applicants to Iraqi custody and whether, if there was such an obligation, it modified or displaced any obligation owed to the applicants under the Convention, are not material to the preliminary issue of jurisdiction (see, mutatis mutandis, Bosphorus, cited above, § 138) and must instead be considered in relation to the merits of the applicants’ complaints.47
Note the several moves that the Court makes. First, though it refers, if not in so many words, to the general theory of Article 1 jurisdiction offered by the applicants as de facto authority and control over individuals, it does not explicitly accept it. Secondly, what it does in the second paragraph quoted above is to posit jurisdiction as control not over a territory or a wider geographical area, but control over places or premises, here the detention facility in which the applicants were kept. This is of course exactly the issue raised before with regard to the CIA black sites. Thirdly, like the High Court, the Chamber expressly disagrees with the Court of Appeal’s position that Article 1 jurisdiction requires an exercise of legal authority, and finds that the question of what impact the UK’s legal obligation to surrender the applicants to Iraqi authorities had on its ECHR obligations, if any, belongs to the merits.48
(p.134) So, what are we to make of this? If we conceive of state jurisdiction in human rights treaties in spatial terms, we can observe that space or area to which it refers on a continuum from something that we would broadly call a ‘territory,’ such as Northern Cyprus, to what we would generally call a ‘place,’ such as a UK-run prison in Iraq or that riding school in Vilnius. The question is whether that continuum extends even further, to even smaller areas or places. I, for one, cannot discern a clear cut-off one way or the other. What is certainly true is that there is a degree of artificiality to this approach, and that the artificiality increases as the size of the area decreases. For example, I would personally find it artificial in the extreme to argue that Israel had human rights obligations towards Eichmann because it controlled the apartment in which Eichmann was held before he was transferred to Israel, rather than Eichmann himself. In other words, the spatial concept of state jurisdiction as control over an area tends to collapse into the personal model of jurisdiction as control over individuals, or indeed into the absence of any threshold at all.
How then can we define an ‘area’ in a principled way? The only possible definition is in my view a functional one: only something over which state can exercise a sufficient degree of control can count as an area. The obvious drawback of (p.135) this definition—if it can even be called a definition—is that we must establish what this sufficient degree of control is. This is what I will proceed to do next.
C. What amounts to ‘control’?
1. Lawful or unlawful
We may remind ourselves once more of the European Court’s holding in Loizidou that it is the fact of state control over territory which is the basis of state jurisdiction, whether obtained lawfully or unlawfully. As I have explained in Chapter II, it might at first glance seem a bit odd to say that a word as familiar as ‘jurisdiction’ actually refers to a factual concept that does not require the exercise of legal competence. Yet this is precisely how the word is used in the treaty practice of states, as a somewhat overdone way of saying that a state has de facto authority or control over a given piece of territory. It may or may not have the right to exercise such control, but it is the fact of control that matters.
This result is of course not simply warranted by rules of treaty interpretation, but by sound considerations of policy, since any other conception of jurisdiction would allow ample room for abuse, as we have already seen in a number of cases. A state can obtain control over the territory of another state in only two basic scenarios—with or without the territorial state’s consent. In the former scenario, so long as the consent of the territorial state remains valid, and so long as the acts of the foreign state remain within the limits of that consent, the foreign state will have acted lawfully. The consent may be expressed in a treaty, e.g. a lease, or a foreign intervention may be invited somewhat less formally. All that matters, however, is that the consent is given. In the latter scenario, the control over another state’s territory may or may not be lawful. As it involves the use of force, the lawfulness of obtaining such control is assessed on the basis of the relevant branch of international law, the jus ad bellum. Yet, just as with the law of occupation and the jus in bello, whose applicability is quite deliberately not dependent on the lawfulness vel non of the use of force as a matter of the jus ad bellum,49 so should the applicability of human rights treaties be independent of such lawfulness.50 Not only does this avoid frequently quite controversial and under-determinate preliminary questions regarding the use of force, but whether a state obtained control over a territory lawfully or not its ability to affect and protect the human rights of its inhabitants is the same.
However, we have also seen that post-Bankovic the Strasbourg jurisprudence has become somewhat more ambiguous. Note that even in Al-Saadoon, the Chamber considered that ‘given the total and exclusive de facto, and subsequently also de jure, control exercised by the United Kingdom authorities over the premises in question, (p.136) the individuals detained there, including the applicants, were within the United Kingdom’s jurisdiction’.51 It is the ‘subsequently also de jure’ part which is troubling, because it is not apparent why it should matter at all. The Chamber probably mentioned it to pre-empt any Bankovic-inspired assault on its decision. This would not make such a limitation any less arbitrary, particularly when we keep in mind that the jus referred to is nothing more than the law promulgated by the occupying powers, which they could switch on or off at will.
In short, any attempt to demand that the state’s control over territory be lawful or anything other than purely factual should be resisted.52 This, of course, still leaves the question of what the quality of the control should be as a matter of fact, to which I turn next.
2. Effective overall control over territory
In the preliminary objections stage of Loizidou, the European Court described Turkey’s control over northern Cyprus only as ‘effective’. At the merits stage, the Court added the word ‘overall’ to the test:
It is not necessary to determine whether, as the applicant and the Government of Cyprus have suggested, Turkey actually exercises detailed control over the policies and actions of the authorities of the ‘TRNC’. It is obvious from the large number of troops engaged in active duties in northern Cyprus that her army exercises effective overall control over that part of the island. Such control, according to the relevant test and in the circumstances of the case, entails her responsibility for the policies and actions of the ‘TRNC’. Those affected by such policies or actions therefore come within the ‘jurisdiction'—of Turkey for the purposes of Article 1 of the Convention. Her obligation to secure to the applicant the rights and freedoms set out in the Convention therefore extends to the northern part of Cyprus.53
The Court reiterated the ‘effective overall control’ formula in Cyprus v. Turkey 54 and subsequent cases. Readers will recall from my discussion of state responsibility issues in Chapter II that the Court’s rulings in Loizidou and Cyprus v. Turkey can reasonably be interpreted in two ways: either the Court thought that the acts of the TRNC were attributable to Turkey, or it thought it unnecessary to deal with this question, finding instead that Turkey had a positive obligation to secure the human rights of the inhabits of northern Cyprus by virtue of its control over the territory. As I explained above, I prefer the latter option, because it reconciles the European Court with the ILC’s and the ICJ’s approach to state responsibility. Hence, while this Article 1 ECHR ‘effective overall control’ test bears resemblance to the ICJ’s effective control test in Nicaragua, the two are conceptually distinct—the former (p.137) refers to state control over territory for the purpose of establishing whether the state has jurisdiction over the territory, the latter to state control over actors and their specific acts for the purpose of attributing these acts to the state. Of course, the control by a state over territory must be exercised by its agents, i.e. persons whose acts are attributable to it.55
In these terms, the ‘overall’ prong of the Loizidou test serves an important purpose—applicants do not need to show that the state controlling a territory ‘exercise[d] detailed control over the policies and actions’ of the (possibly non-state) actor whose conduct directly violated their rights. As the Court put it in Cyprus v. Turkey, ‘[h]aving effective overall control over northern Cyprus, [Turkey’s] responsibility cannot be confined to the acts of its own soldiers or officials in northern Cyprus but must also be engaged by virtue of the acts of the local administration which survives by virtue of Turkish military and other support.’56
But what exactly does it mean that the state’s control needs to be ‘effective’? In the most general terms, the state needs to have enough power over the territory and its inhabitants to broadly do as it pleases. That said, control over territory is a fluid thing, and is not limitless even under the best of conditions. To move from the abstract to the concrete we would need to examine specific cases, and we would then see that the threshold of control required by courts has generally been high. In the northern Cyprus cases, Turkey had thousands of troops on the ground, and the TRNC administration was at least initially little more than its puppet. In the ICJ Wall and Congo v. Uganda cases, the situation was again one of belligerent occupation. And in Bankovic, of course, the Court affirmed that the threshold of control should be strict, holding that air strikes alone were unable of satisfying it, and that
its recognition of the exercise of extra-territorial jurisdiction by a Contracting State is exceptional: it has done so when the respondent State, through the effective control of the relevant territory and its inhabitants abroad as a consequence of military occupation or through the consent, invitation or acquiescence of the Government of that territory, exercises all or some of the public powers normally to be exercised by that Government.57
The ‘public powers’ requirement in particular appears to set the bar very high, as it seems to require the establishment of a structured (military) administration in the territory.58 On the other hand, the Court backtracked from this stringent requirement of territorial control in Issa, a Chamber case, and in Ilascu, decided by a Grand Chamber. Issa dealt with an incursion by Turkish troops into northern Iraq, and the possible killing by these troops of certain Iraqi nationals. In one paragraph of the judgment, the Chamber seemed to have allowed for conceiving jurisdiction in personal terms, as authority and control over individuals,59 while it said the (p.138) following with regard to Turkey’s territorial jurisdiction over the area of northern Iraq:
In this connection, the Court notes that it is undisputed between the parties that the Turkish armed forces carried out military operations in northern Iraq over a six-week period between 19 March and 16 April 1995 (see paragraphs 58 and 63 above). It transpires from the parties’ submissions and the documentary evidence contained in the case-file that the cross-border operation conducted at that time was extensive and was aimed at pursuing and eliminating terrorists who were seeking shelter in northern Iraq (see paragraphs 36, 43, 45, 58 and 63).
The Court does not exclude the possibility that, as a consequence of this military action, the respondent State could be considered to have exercised, temporarily, effective overall control of a particular portion of the territory of northern Iraq. Accordingly, if there is a sufficient factual basis for holding that, at the relevant time, the victims were within that specific area, it would follow logically that they were within the jurisdiction of Turkey (and not that of Iraq, which is not a Contracting State and clearly does not fall within the legal space (espace juridique) of the Contracting States (see the above-cited Banković decision, § 80).
However, notwithstanding the large number of troops involved in the aforementioned military operations, it does not appear that Turkey exercised effective overall control of the entire area of northern Iraq. This situation is therefore in contrast to the one which obtained in northern Cyprus in the Loizidou v. Turkey and Cyprus v. Turkey cases (both cited above). In the latter cases, the Court found that the respondent Government's armed forces totalled more than 30,000 personnel (which is, admittedly, no less than the number alleged by the applicants in the instant case—see § 63 above—but with the difference that the troops in northern Cyprus were present over a very much longer period of time) and were stationed throughout the whole of the territory of northern Cyprus. Moreover, that area was constantly patrolled and had check points on all main lines of communication between the northern and southern parts of the island. The essential question to be examined in the instant case is whether at the relevant time Turkish troops conducted operations in the area where the killings took place.60
The Court then found that the evidence for such operations was lacking. However, the Court allowed for the possibility that Turkey exercised its jurisdiction in Iraq: (1) over a much more limited area than the whole of northern Iraq—indeed, the immediate area of a village; (2) over a much shorter period of time—some six weeks in all—than in the northern Cyprus cases; and (3) consequently with control with significantly less permanence, stability or exercise of any ‘public powers’.
While Issa relaxes the Loizidou threshold somewhat, Ilascu perhaps goes so far as to redefine it. In addition to the question of Russia’s control or jurisdiction over the territory of Transdniestria in Moldova, that case was plagued with severe problems regarding the Convention’s temporal scope of application, and the Court’s methodological approach is generally hard to decipher. Thus, with regard to the period before the entry of the Convention into force with respect to Russia, the Court considered that
(p.139) …the Russian Federation’s responsibility is engaged in respect of the unlawful acts committed by the Transdniestrian separatists, regard being had to the military and political support it gave them to help them set up the separatist regime and the participation of its military personnel in the fighting. In acting thus the authorities of the Russian Federation contributed both militarily and politically to the creation of a separatist regime in the region of Transdniestria, which is part of the territory of the Republic of Moldova.
The Court next notes that even after the ceasefire agreement of 21 July 1992 the Russian Federation continued to provide military, political and economic support to the separatist regime (see paragraphs 111 to 161 above), thus enabling it to survive by strengthening itself and by acquiring a certain amount of autonomy vis-à-vis Moldova.61
It was at this time that the applicants in the case were arrested by Russian troops in Transdniestria, who then transferred them to the Transdniestrian separatist authorities.62 According to the Court,
…on account of the above events the applicants came within the jurisdiction of the Russian Federation within the meaning of Article 1 of the Convention, although at the time when they occurred the Convention was not in force with regard to the Russian Federation.
This is because the events which gave rise to the responsibility of the Russian Federation must be considered to include not only the acts in which the agents of that State participated, like the applicants’ arrest and detention, but also their transfer into the hands of the Transdniestrian police and regime, and the subsequent ill-treatment inflicted on them by those police, since in acting in that way the agents of the Russian Federation were fully aware that they were handing them over to an illegal and unconstitutional regime.
In addition, regard being had to the acts the applicants were accused of, the agents of the Russian Government knew, or at least should have known, the fate which awaited them.
In the Court’s opinion, all of the acts committed by Russian soldiers with regard to the applicants, including their transfer into the charge of the separatist regime, in the context of the Russian authorities’ collaboration with that illegal regime, are capable of engaging responsibility for the acts of that regime.
It remains to be determined whether that responsibility remained engaged and whether it was still engaged at the time of the ratification of the Convention by the Russian Federation.63
The Court here says that Russia had jurisdiction over the territory of Transdniestria even though the Convention was, for Russia, not yet in force. There is in principle nothing wrong with that conclusion. But note the lack of clarity in the Court’s formulation that the responsibility of Russia was ‘engaged’—does the Court here think that the conduct of the Transdniestrian authorities was attributable to Russia, and if so on what grounds? Note also that whatever Russia did at the time was not wrongful under the ECHR, as the ECHR was not yet in force for Russia. The Court then moved on to the period after Russia had ratified the ECHR on 5 May 1998:
The Russian army is still stationed in Moldovan territory in breach of the undertakings to withdraw them completely given by the Russian Federation at the OSCE summits in (p.140) Istanbul (1999) and Porto (2001). Although the number of Russian troops stationed in Transdniestria has in fact fallen significantly since 1992 (see paragraph 131 above), the Court notes that the ROG’s weapons stocks are still there.
Consequently, in view of the weight of this arsenal (see paragraph 131 above), the ROG’s military importance in the region and its dissuasive influence persist.
All of the above proves that the ‘MRT’, set up in 1991–1992 with the support of the Russian Federation, vested with organs of power and its own administration, remains under the effective authority, or at the very least under the decisive influence, of the Russian Federation, and in any event that it survives by virtue of the military, economic, financial and political support given to it by the Russian Federation.
That being so, the Court considers that there is a continuous and uninterrupted link of responsibility on the part of the Russian Federation for the applicants’ fate, as the Russian Federation’s policy of support for the regime and collaboration with it continued beyond 5 May 1998, and after that date the Russian Federation made no attempt to put an end to the applicants’ situation brought about by its agents, and did not act to prevent the violations allegedly committed after 5 May 1998.
Regard being had to the foregoing, it is of little consequence that since 5 May 1998 the agents of the Russian Federation have not participated directly in the events complained of in the present application.
In conclusion, the applicants therefore come within the “jurisdiction” of the Russian Federation for the purposes of Article 1 of the Convention and its responsibility is engaged with regard to the acts complained of.64
Put aside for one moment the lack of rigour in the Court’s approach to issues of state responsibility, both with regard to attribution and with regard to the question of continuous wrongful acts and the ECHR’s temporal application. If the Court’s judgment is interpreted as holding that the actions of the MRT have at all times been attributable to Russia, then it is clear that Russia still has jurisdiction and control over the territory of Transdniestria. If, on the other hand, the acts of the MRT are not generally attributable to Russia, but Russia is being held responsible for its own failure to comply with its positive obligations, then the Court uses a pretty lax standard to establish Russia’s jurisdiction over Transdniestria—all that mattered was that Russia had ‘decisive influence’ over it. This is a far cry from requiring the exercise of ‘public powers’ by Russia or a de facto government or administration of the territory.
The Court justified its approach by referring to Russia’s involvement in the creation of the MRT, the far more extensive military control that it had in the past, and perhaps most importantly, the potential for such control that has remained. This is of course a remarkably similar situation to northern Cyprus, where Turkey has likewise gradually downgraded its military presence, while the local administration has become more independent. In other words, Russia could still be said to exercise effective overall control over Transdniestria, because if it wanted to it could easily make its power felt more overtly.65
(p.141) So what are we to make of the case law? First, the threshold of effective overall control of a territory is set relatively high. As a general matter, it requires boots on the ground. Secondly, though the threshold is set high, that level of control still need not be as high as the one that the state has over its own territory in peacetime or during normalcy. Thirdly, effective overall control is itself a spectrum, ranging from the more entrenched and visible exercise of de facto government, administration, or public powers, to the more borderline cases of less permanent or overt state control as in Issa and Ilascu.
More fundamentally, the case law aside, I think it is fair to say that the test of effective overall control of an area is also a functional one. Its stringency in the degree of effectiveness required depends foremost on the consequences that attach to the fact of such effective control. Recall the European Court’s regime integrity concern in Bankovic that the ECHR cannot be cut up into little pieces, some of which would apply extraterritorially and some not, and that accordingly the ECHR is an all or nothing package. If the obligation incumbent on states is to secure all Convention rights in an area where they exercise effective overall over that area, then the degree of control required must be such as to allow states to realistically comply with this obligation.66
On the other hand, two considerations militate against having a threshold which is overly strict. First, the state’s ability to comply with its negative obligation to respect human rights does not depend on its control over territory. Rather, the state by definition has control over its own agents. Secondly, even with respect to the state’s positive obligation to secure human rights, that obligation is not as onerous as is it sometimes made out to be. It is an obligation of due diligence, of the state doing all that it could reasonably be expected to do to protect a territory’s inhabitants even from third parties. Accepting, say, that the UK had jurisdiction over the parts of southern Iraq that it had occupied post-2003 does not entail that the UK had the obligation to turn Iraq into a simulacrum of idyllic English countryside.67
This brings me to the next issue I will address, the relationship between the effective overall control of an area threshold for establishing state jurisdiction and the threshold of control required for belligerent occupation.
3. Relationship with the threshold of belligerent occupation
In a scenario where a state obtains control over the territory of another state with that state’s consent, the resulting occupation is pacific in nature. When, however, such control is obtained through the use of force (even if that use of force need not necessarily be opposed), IHL and the law of belligerent occupation apply concurrently with human rights treaties.68 One issue that needs to be considered is whether the state jurisdiction threshold for the application of human rights treaties (p.142) is the same as or different than the threshold required for the establishment of a belligerent occupation.69
Like the concept of jurisdiction in human rights treaties, belligerent occupation also depends on purely factual criteria. As provided in Article 42 of the Hague Regulations: ‘Territory is considered occupied when it is actually placed under the authority of the hostile army. The occupation extends only to the territory where such authority has been established and can be exercised.’ This standard is generally interpreted in the literature as being one of ‘effective control’.70
But is this the same degree of control that we have examined above with regard to human rights treaties? Could there, in other words, be occupation without jurisdiction, or jurisdiction without occupation?
For the latter possibility, we might take another look at Issa. Recall that the Turkish military operation in northern Iraq lasted for a mere six weeks, and that the control of portions of Iraq by Turkish troops was fluid and intermittent. Assuming that Turkey had jurisdiction over parts of Iraq at the relevant time, was it also their belligerent occupier? On that point, IHL has always made a distinction between an invasion and an occupation.71 Occupation may follow an invasion, but it requires that a territory is solidly or firmly seized and that its sovereign has actually been displaced—a mere vacuum in authority does not suffice. Though the precise line between the invasion and the occupation phase can be hard to draw, occupation requires a degree of stability.72 As to whether the facts of Issa disclose the existence of a belligerent occupation, it is at best a borderline case.
Or take another example, that of Gaza after Israel’s 2005 unilateral disengagement. Even though Israel no longer has troops on the ground in Gaza (at least not on a regular basis), nor exercises a military administration there, it has been argued, particularly by the human rights community broadly conceived, that Gaza still remains occupied by Israel.73 Most of the arguments in favour of this position are to my mind unpersuasive. For example, control of the border crossings into a territory is not the occupation of that territory, but a siege. Nor is control over the sea and the airspace tantamount to the actual control of the territory itself. There has not been a single instance of belligerent occupation that I am aware of where a state was considered to have been the occupant of a territory without having troops on the ground. Likewise, that Israel remains the occupying power in the West Bank does not mean that it could not have terminated its occupation over Gaza, or that it could not have done so unilaterally.74
To my mind, the only good argument that Gaza remains occupied is that Israel still possesses an active potential for regaining effective control—cf. our discussion (p.143) of Ilascu above. Occupation does not necessarily mean the control of all of the territory, all of the time, at the same level of intensity. Many occupations are met with significant and active resistance. Consider, for example, the US and UK-led occupation of Iraq. The strength of the insurgency there was such that at any given time these states lacked control over, say, some parts of the Sunni Triangle or some parts of Basra. That does not mean that their occupation of those parts of Iraq ceased, because the occupants were able to re-establish their control at will, as they did. The lapse in control was merely temporary, due to the fact that the level of resistance was such that the occupants could not control all of Iraq, all of the time. Another well-known example is the German occupation of Yugoslavia during the Second World War. The partisans fighting against the Nazis were from time to time able to temporarily liberate a bit of territory here or there, but that did not mean that Germany’s occupation ceased. This was so because, as the military tribunal at Nuremberg ruled in United States v. List, ‘the Germans could at any time they desired assume physical control of any part’ of Yugoslavia.75 The ICTY in Naletilic likewise ruled that an occupation exists so long as the occupying army has the ‘capacity to send troops within a reasonable time to make the authority of the occupying power felt’.76
In any event, it may be arguable that Gaza remains occupied by Israel, but it is far from clear that this is the case, and it is certainly not so clear for this claim to be made in such forceful terms as it is by some in the human rights community.77 Why is this argument then made so emphatically? The answer to that question is simple. Under IHL, a belligerent occupant has positive obligations to ensure the well-being of the civilian population, including the provision of food and other supplies. In a state of siege without occupation, however, the party to the conflict only has negative obligations not to interfere with relief consignments and so forth, and even these can be subject to military necessity. It does not have to provide food and supplies to the civilian population of its adversary.78 Indeed, to impose such a requirement would be manifestly absurd. The problem is of course precisely that the civilian population of Gaza is heavily dependent on Israel and needs Israel not just to let humanitarian aid through, but also to provide electricity and other supplies of its own. This is why the gentler souls among us need to argue that Israel is the belligerent occupier of Gaza. It is the only legally certain way of assigning some positive obligations on Israel to provide for the civilians of Gaza—something that I agree with entirely as a matter of policy. But the certainty is unfortunately only deceptive.
(p.144) So what are we to do then, if Israel is no longer the occupying power in Gaza, as has in fact been held by the Israeli Supreme Court?79 One way of dealing with the issue would be to say that the state of dependency of Gaza on Israel created by Israel’s occupation is a consequence of Israel’s wrongful failure to abide by its duties as an occupying power, and that it has the obligation to provide full reparation for this failure. Likewise, Yuval Shany has made the intriguing argument that Israel might not have effective control over Gaza for the purposes of the law of occupation and thus has no positive obligations under IHL, but that it might have jurisdiction over Gaza under the relevant human rights treaties, such as the ICCPR, and would hence have positive obligations towards Gaza’s inhabitants.80 In other words, there could be jurisdiction without occupation, just as in Ilascu Russia had jurisdiction simply because it had ‘decisive influence’ over Transdniestria.
Consider, on the other hand, the possibility of a converse scenario of occupation without jurisdiction. Let us take a look, for example, at the insurgency in Iraq that followed the 2003 invasion, whose strength and magnitude was such that the coalition forces could not maintain total control over all of Iraq, all of the time. As we have seen above, this does not necessarily mean that the US and the UK ceased to be the occupying powers in Iraq, even if only temporarily, since they could always make their power felt. But does this mean that they had a sufficient degree of control to satisfy the jurisdiction threshold of human rights treaties?
In Al-Skeini both the Court of Appeal and the House of Lords thought that the UK lacked effective overall control over Basra for the purpose of establishing jurisdiction even though it was the occupying power in that part of Iraq.81 In the Court of Appeal Lord Justice Brooke (joined by Lord Justice Richards) thought that it was
…quite impossible to hold that the UK, although an occupying power for the purposes of the Hague Regulations and Geneva IV, was in effective control of Basrah City for the purposes of ECHR jurisprudence at the material time. If it had been, it would have been obliged, pursuant to the Bankovic judgment, to secure to everyone in Basrah City the rights and freedoms guaranteed by the ECHR. One only has to state that proposition to see how utterly unreal it is. The UK possessed no executive, legislative or judicial authority in Basrah City, other than the limited authority given to its military forces, and as an occupying power it was bound to respect the laws in force in Iraq unless absolutely prevented. It could not be equated with a civil power: it was simply there to maintain security, and to support the civil administration in Iraq in a number of different ways.
It would indeed have been contrary to the Coalition's policy to maintain a much more substantial military force in Basrah City when its over-arching policy was to encourage the Iraqis to govern themselves. To build up an alternative power base capable of delivering all the rights and performing all the obligations required of a contracting state under the ECHR at the very time when the IGC had been formed, with CPA encouragement, as a step (p.145) towards the formation by the people of Iraq of an internationally recognized representative Government, would have run right against the grain of the Coalition's policies.82
This provoked the following response from Lord Justice Sedley:
In any event, the Court's explanation in Bankovic of the legal distinction between the bombing of Belgrade and the occupation of northern Cyprus places the British occupation of the Basra region of Iraq unequivocally in the latter class—subject to the single question of how much control is, for this purpose, effective control. International human rights law in its present phase does not answer this question. On the one hand, it sits ill in the mouth of a state which has helped to displace and dismantle by force another nation's civil authority to plead that, as an occupying power, it has so little control that it cannot be responsible for securing the population's basic rights. On the other, the fact is that it cannot: the invasion brought in its wake a vacuum of civil authority which British forces were and still are unable to fill. On the evidence before the Court they were, at least between mid-2003 and mid-2004, holding a fragile line against anarchy.
No doubt it is absurd to expect occupying forces in the near-chaos of Iraq to enforce the right to marry vouchsafed by Art. 12 or the equality guarantees vouchsafed by Art. 14. But I do not think effective control involves this. If effective control in the jurisprudence of the ECtHR marches with international humanitarian law and the law of armed conflict, as it clearly seeks to do, it involves two key things: the de facto assumption of civil power by an occupying state and a concomitant obligation to do all that is possible to keep order and protect essential civil rights. It does not make the occupying power the guarantor of rights; nor therefore does it demand sufficient control for all such purposes. What it does is place an obligation on the occupier to do all it can.
If this is right, it is not an answer to say that the UK, because it is unable to guarantee everything, is required to guarantee nothing. The question is whether our armed forces’ effectiveness on the streets in 2003–4 was so exiguous that despite their assumption of power as an occupying force they lacked any real control of what happened from hour to hour in the Basra region. My own answer would be that the one thing British troops did have control over, even in the labile situation described in the evidence, was their own use of lethal force. Whether they were justified in using it in the situations they encountered, of which at least four of the cases before us are examples, is precisely the subject of the inquiry which the appellants seek. It is in such an inquiry that the low ratio of troops to civilians, the widespread availability of weapons and the prevalence of insurgency would fall to be evaluated. But, for reasons I now come to, I am not confident that this route is open in the present state of ECHR jurisprudence.83
Just like Lord Justice Brooke, the majority of the House of Lords thought that the UK did not have effective overall control and hence jurisdiction over Basra even though it had occupied it, although their primary holding was that the effective overall control of an area test could not even apply in Iraq, as it was outside the ECHR’s espace juridique.84
(p.146) Al-Skeini is instructive on several points. First, most of the British judges thought that the UK’s occupation of southern Iraq need not necessarily result in its jurisdiction over it. Secondly, their reasoning was that in their view the Article 1 ECHR positive obligation to secure human rights would be far too onerous under the circumstances. Among these circumstances were the strength of the insurgency against the occupation and the resulting ‘near-chaos’ in the area, the complicated institutional arrangements among the occupying powers, and the existence of a local administration that has been growing steadily more independent. Thirdly, the judges were worried in particular about the procedural positive obligations under Article 2 ECHR to investigate deprivations of life, which they thought the UK could not realistically fulfil, primarily because these requirements were developed by the European Court in times of normalcy, and would be difficult to depart from.85
These concerns are real, and we have already examined them above under the rubric of effectiveness. If these fears are not addressed, the extraterritorial application of human rights treaties beckons only as utopia. So let me try to address them. Most importantly, as Lord Justice Sedley well pointed out, the obligation to secure human rights is not, and should not be as onerous as the other judges made it seem.86 It is not an obligation outside the realm of the possible. In its application we may certainly take into account the law of occupation, and have regard to the extraordinary circumstances involved, if not those that are solely of the UK’s own making.87
That, for example, the UK chose to enter into complicated joint arrangements with the United States is quite frankly its own problem. So long as its military presence is such that it remains the occupant, why should it also not be expected to do all it reasonably can to secure the human rights of the population of the occupied territory? It cannot just say that it is there only to provide security, and that it is not even particularly good at that. It should also be borne in mind that IHL equally abhors ‘paper occupations’.88 A belligerent occupation without effective control, or at the very least the potential that such control can quickly be re-established, is simply no longer an occupation.
(p.147) To my mind at least, so long as the British occupation of southern Iraq was not terminated, the UK had effective overall control for Article 1 ECHR purposes over the area in question. In all but the most exceptional of cases, the two thresholds should be the same, so long as it is borne in mind that effective control implies the power of the state not only to affect the human rights of the population, but also to secure them, yet that the consequent obligation to secure or ensure human rights is a flexible one.
Conversely, turning back to the above example of Gaza, we should also resist accepting that jurisdiction may arise without occupation in situations where states obtain control over foreign territory by force. While it may seem attractive in some cases to relax the threshold of jurisdiction below that of occupation, doing so would actually expose extraterritorial application to the charge of utopia as an ideal completely divorced from practical considerations. Jurisdiction does require effectiveness, and a relatively stable presence of troops on the ground appears to be the only way of securing it, at least for the foreseeable future. Borderline cases will of course present themselves, either somewhere on the fuzzy line between invasion and occupation, as in Issa, or in situations where the state’s grip on the territory is being challenged by another force. Likewise, the regime of belligerent occupation may be transformed into something else without the immediate loss of effective control and jurisdiction, as has arguably happened in Iraq. Though it is obviously impossible to provide a one-size-fits-all formula for dealing with such cases, one thing is certain—that they must be addressed with great caution.
4. Does control need to be exclusive?
A further question that arises is whether control over territory needs to be exclusive to be considered effective. The short answer is that effective control does normally, but not necessarily, exclude the exercise of control by the territorial state as the displaced sovereign.
When control over territory is obtained through the use of force, whether lawfully or unlawfully, effective control both for the purposes of state jurisdiction and the establishment of belligerent occupation presumes that the hostile army has established its own authority over the territory and substituted it for that of the displaced sovereign, and has successfully managed to suppress the resistance of the enemy, except perhaps in isolated pockets.
This simple scenario is perhaps the most common one, as for instance with regard to northern Cyprus. However, the legal situation can get more complicated, when (1) the occupant creates a subordinate local administration, and allows it a degree of autonomy; (2) allows the vestiges of the local administration of the displaced sovereign to operate normally; (3) an insurgency develops in the occupied territory; (4) the use of force and the occupation are conducted jointly by several states; (5) the use of force and the occupation are conducted under the auspices of an international organization; or (6) the belligerent occupation regime is terminated and evolves into some type of pacific occupation.
(p.148) Scenarios (1) to (3) are perhaps the easiest to deal with. As the European Court held in Loizidou, the obligation to secure human rights arises whenever a state has effective control of an area, ‘whether it be exercised directly, through its armed forces, or through a subordinate local administration’.89 Indeed, occupants have frequently either set up their own proxy administrations or quisling governments, or have allowed some parts of the former administration to function so long as any military resistance is suppressed. This does not imply that everything these local actors do is attributable to the occupying power. However, while effective control of the occupant over the territory persists, both occupation and jurisdiction will remain even in the face of a mounting insurgency. It is of course entirely possible that the occupation forces are defeated and that they lose effective control, but it is only then that jurisdiction over the territory will also cease. With regard to the existence of a semi-autonomous local administration, particularly a legitimate one, its competences can certainly be taken into account in a due diligence analysis on the merits with respect to the state’s positive obligation to secure or ensure human rights.
Scenario (4) is more complex, if not hopelessly so. States frequently take joint military action. They can do so in a variety of ways, from operating in a loose alliance, to putting their forces under unified military command. Depending on the exact institutional arrangements that they create, they can either exercise control over an area jointly, or they can divide a larger area into smaller units, each of which would be controlled by a specific national contingent, or perhaps even both. Post-2003 Iraq provides a good example. Iraq was under a regime of belligerent occupation from the completion of the invasion, with the United States and the United Kingdom being regarded by the UN Security Council as the occupying powers in Iraq, who had set up a unified command, the Coalition Provisional Authority.90 The belligerent occupation lasted up until the end of June 2004 at the earliest, when the CPA transferred authority to the new Interim Government of Iraq—though the issue of the end of the occupation is controversial.91 Let us consider at least the uncontested period of occupation, when the UK and the US were joint occupying powers over the whole of Iraq, but when in fact the UK mostly held the southern parts of Iraq, while the US occupied the rest. During this period, did the UK have jurisdiction over the whole of Iraq, since it was one of the joint occupying powers operating under unified command, or did it merely have jurisdiction over Basra and southern Iraq? This question is certainly a difficult one, (p.149) and to answer it one would need to examine in detail the realities of the supposedly unified US/UK command. In principle, however, it is certainly possible that both the US and the UK had jurisdiction, i.e. effective overall control, over the same areas of Iraq.92
Similar to (4), but even more complex, is scenario (5), which involves joint state action though an international organization. Take as an example Kosovo after the 1999 NATO intervention against the then Federal Republic of Yugoslavia. The Security Council passed Resolution 1244 (1999), creating an international civil administration under UNMIK, a subsidiary organ of the UN, and KFOR, a NATO-led international military presence. Kosovo was not under a regime of belligerent occupation, since Yugoslavia had ultimately consented (if under some duress) to the deployment of the international forces in Kosovo, and since the Security Council had created a comprehensive regime of international administration over the territory. Though KFOR operated under unified command, Kosovo was divided up into several military districts, each of which was controlled by a particular national contingent.
The fundamental question posed by the Kosovo example is whether, say, the inhabitants of the area controlled by the French national contingent were within the jurisdiction of France within the meaning of Article 1 ECHR. Antecedent to that question is the issue whether the acts of French troops in Kosovo were, in fact, still attributable to France. In Behrami, the Grand Chamber of the European Court held that the conduct of French troops was not attributable to France, but to the United Nations, since French troops in Kosovo operated there pursuant to a UN Security Council mandate.93 The customary rules regarding attribution of conduct to international organizations are admittedly less clear than those on state responsibility. Their codification is currently in the ILC’s pipeline, which has so far produced and adopted on first reading a set of draft articles.94 Having said so, in Behrami the European Court manifestly misapplied even the few clear rules that we actually have, by equating Security Council authorization with (exclusive) attribution of conduct to the UN, even when the UN itself lacks effective control over national contingents, and by ignoring the possibility of dual or multiple attribution. The Behrami decision has therefore attracted overwhelming criticism from numerous scholars,95 including from the present author.96 Most importantly, it has been rejected by the ILC in no uncertain terms.97
(p.150) But though Behrami may have been wrongly decided, it is true that attribution matters for resolving the question of state jurisdiction. It is only if the conduct of French troops in Kosovo was attributable to France, although it may also have been attributable to NATO or even to the UN as separate legal persons, that France could have exercised jurisdiction over a part of Kosovo. Conversely, it is only if the conduct of the troops is not attributable to the troop contributing state, but only to the organization, that the possibility that the state is exercising jurisdiction can be excluded.98
In any case, as with attribution, the exercise of jurisdiction need not be exclusive. That UNMIK ran the civil administration of Kosovo, and that other parts of Kosovo were controlled by other national contingents, are both circumstances that can be taken into account when interpreting the substantive content of the positive obligation to secure human rights.
As for scenario (6), it poses unique challenges in that it involves the transformation of a regime which is relatively clear, that of belligerent occupation, into one of pacific occupation (if that),99 the contours of which are far less defined. Consider the examples of Iraq and Afghanistan. In both cases a coalition of states used force against the two countries, whether lawfully or unlawfully. In both cases a regime of belligerent occupation ensued, whether more or less formalized. As time went on, that regime was supplemented and in parts overridden by UN Security Council action, while preparations were underway for the creation of new, internationally recognized governments. Both of these states eventually gained such new governments, which then gave their consent to the presence of foreign troops. At that point at the latest, the regime of belligerent occupation must have terminated, as such occupation can by definition be only non-consensual.100
From that moment onwards, the conduct of foreign troops was governed by the limits of the territorial state’s consent, and by applicable Security Council resolutions. The situation on the ground, however, need not have changed dramatically, if it had changed at all. Although, for example, the nascent armed forces of the new legitimate government now started participating in joint military operations with foreign troops, the interveners could still be said to have had effective control over large swathes of territory. In my view, while such control persists, the intervening states can be said to exercise jurisdiction over the areas concerned within the meaning of the relevant human rights treaties. Likewise, the same could be said of the now independently governed territorial state, so long as it also had the requisite degree of actual control. There is no need, in principle, for the control of either the intervening states or the territorial state to be exclusive.
What is certainly true is that, after the termination of belligerent occupation, the new limits on what the intervening states may lawfully do can pose substantial (p.151) difficulties. However, the interpretation of the obligation to secure human rights can be sufficiently flexible to take into account both the dispositions of authority in Security Council mandates, and the need to respect the sovereignty of the territorial state. But, as I have discussed above, though flexible it cannot be watered down so much that it becomes completely ineffective. If, for example, the US and UK had knowledge that Iraqi government forces were engaging in human rights violations in areas that were still under US and UK effective overall control,101 the US and the UK would (or should) have been faced with a choice: either put an end to Iraqi human rights abuses, by force if necessary and in likely violation of Iraqi sovereignty, or withdraw from Iraq. The situation, in other words, would be one of norm conflict.102
In sum, the scenarios that we have just examined show us that while jurisdiction over territory is normally exclusive, it need not necessarily be so. And while such cases can pose significant difficulties, these are not insurmountable. Further examples of joint or parallel jurisdiction can arise in cases, to which I will turn next, involving control over places or objects, rather than large areas of territory.
D. Control over places and objects
1. A general theory?
Readers will recall my discussion above of what constitutes an ‘area’ susceptible to state control for the purposes of establishing its jurisdiction. That discussion has ultimately proven to be incapable of providing any firm answers, except that the more an ‘area’ diminishes in size, the more it becomes a ‘place’, the more artificial it seems to apply the spatial concept of state jurisdiction to it. And of course the line between a ‘place’ proper, such as a building, or parts of that building or even a square foot or two within it, is also difficult to draw.
But assume that we can actually draw some reasonable lines that would enable us to prevent the collapse of the control over a place criterion into something resembling control over a mere speck or an atom of land. Does such control over a place present a general theory or grounds for establishing state jurisdiction for the purposes of human rights treaties? Or is it rather the special nature of some places, such as embassies, that brings them and the individuals located in them within the extraterritorial state’s jurisdiction?
The first proposition is certainly the more intriguing one, as it has the potential of providing a principled approach rather than relying on some unavoidably arbitrary criteria of speciality. Explicit textual support for this proposition can be (p.152) found in two jurisdiction clauses, Article 2 CPT103 and Article 4(1) of the Optional Protocol to the CAT.104 As for the CAT itself, the Committee Against Torture has in its General Comment No. 2 interpreted its jurisdiction clauses which refer to ‘any territory under [a state’s] jurisdiction’ as also encompassing places under the state’s control:
The Committee also understands that the concept of ‘any territory under its jurisdiction,’ linked as it is with the principle of non-derogability, includes any territory or facilities and must be applied to protect any person, citizen or non-citizen without discrimination subject to the de jure or de facto control of a State party. The Committee emphasizes that the State’s obligation to prevent torture also applies to all persons who act, de jure or de facto, in the name of, in conjunction with, or at the behest of the State party.105
Thus, in the Committee’s view, the concept of ‘territory’ includes ‘facilities’. The Committee further added that
[a]rticle 2, paragraph 1, requires that each State party shall take effective measures to prevent acts of torture not only in its sovereign territory but also ‘in any territory under its jurisdiction.’ The Committee has recognized that ‘any territory’ includes all areas where the State party exercises, directly or indirectly, in whole or in part, de jure or de facto effective control, in accordance with international law. The reference to ‘any territory’ in article 2, like that in articles 5, 11, 12, 13 and 16, refers to prohibited acts committed not only on board a ship or aircraft registered by a State party, but also during military occupation or peacekeeping operations and in such places as embassies, military bases, detention facilities, or other areas over which a State exercises factual or effective control. The Committee notes that this interpretation reinforces article 5, paragraph 1 (b), which requires that a State party must take measures to exercise jurisdiction ‘when the alleged offender is a national of the State.’ The Committee considers that the scope of ‘territory’ under article 2 must also include situations where a State party exercises, directly or indirectly, de facto or de jure control over persons in detention.106
Note, however, how the Committee’s approach tends to collapse to the notion of jurisdiction as control over individuals—in a textually problematic way, it must be said, with regard to the provisions of the CAT. As for the European Court, the case which came the closest to endorsing a general theory is of course Al-Saadoon:
The Court considers that, given the total and exclusive de facto, and subsequently also de jure, control exercised by the United Kingdom authorities over the premises in question, the individuals detained there, including the applicants, were within the United Kingdom’s jurisdiction. This conclusion is, moreover, consistent with the dicta of the House of Lords in (p.153) Al-Skeini and the position adopted by the Government in that case before the Court of Appeal and House of Lords.107
The Chamber in Al-Saadoon exhibited a high level of caution, aware as it was of contradictory case law all around it. Note, for example, how it treats the ‘dicta’ of the House of Lords in Al-Skeini, agreeing with the conclusion that Baha Mousa, who was detained by UK troops, was within the UK’s jurisdiction, yet without endorsing the reasoning of the House of Lords. That reasoning, as we have seen, was precisely that Baha Mousa was within the UK’s jurisdiction for a ‘special’ reason: that a UK military prison abroad was essentially analogous to a UK embassy.108
I, for one, find these special or exceptional explanations endorsed by the House of Lords to be entirely unsatisfactory. Not only does the analogy between a military prison and an embassy defy common sense, as there is simply no such thing as a ‘special status’ of a military prison in international law,109 it introduces distinctions which are entirely arbitrary. For example, a person detained in an embassy or in a military prison might be entitled to protection under human rights treaties, but what about a person detained in a CIA ‘black site’, such as that riding school in Lithuania? Is that also sufficiently like an embassy or not, and if not, why not?
The only thing common both to a prison and to an embassy are that they operate on the basis of the territorial state’s consent.110 But if this was the basis for the extraterritorial state’s jurisdiction, no justification is given for this position, aside from the fallacy that the application of human rights treaties would as such infringe on the territorial state’s sovereignty.111 Indeed, one would imagine that non-consensual interventions would generally be more likely to adversely affect the human rights of the population, while consent itself can be given to many things, such as to the presence of foreign forces in general, or to rendition overflights, or the operation of a CIA ‘black site’.
If, on the other hand, we were to adopt the general theory of control over places as a basis for state jurisdiction, then many of the same considerations that we have already examined with regard to control over territory would apply accordingly, with one major difference. A place is by definition located within a territory under another state’s jurisdiction, and control over such places can by its nature be much more fleeting than effective overall control over large areas of territory would normally be. In all but the most exceptional of circumstances, such control can be maintained with any stability only with the consent of the territorial state. In other words, without Lithuania’s consent, the US could probably never have operated the black site on the outskirts of Vilnius. Its control over that place was only viable because the state controlling the territory acquiesced in it. Accordingly, (p.154) this is also a scenario of concurrent, rather than exclusive jurisdiction. If we accept that the US had jurisdiction over the black site near Vilnius, this does not mean that Lithuania did not have jurisdiction—its obligation to secure the human rights of all persons detained at the black sites was undiminished. The black site most certainly did not drill a black hole in Lithuania’s own ECHR obligations.
On the other hand, if we turn back to our example of the abduction of Adolf Eichmann from Argentina, I would personally find it difficult to accept that the safe house in which the Israeli agents kept Eichmann until he was smuggled to Israel aboard an aircraft amounted to a place under Israel’s jurisdiction. Again, it is simply far too artificial to say that it is the Israeli agents’ control of a minuscule space between four walls, rather than their control of Eichmann’s person, that would entitle Eichmann to the protection of his human rights vis-à-vis Israel. In any event, though the viability of any long-term control over a place would generally require the consent of the territorial state, this is simply a question of fact, rather than a legal requirement—and new facts can always surprise us.
That is all that can be said, I think, about control over a place as a general basis for state jurisdiction. I will now turn to the supposedly special cases of embassies, vessels, and aircraft on which the House of Lords relied so prominently in Al-Skeini. The root cause of this methodological confusion can, predictably, be traced back to Bankovic.
2. Embassies and consulates
In order to limit as much as possible the scope of the ECHR’s extraterritorial application, the European Court in Bankovic held that such application can be justified only exceptionally, supposedly on grounds of general international law. One of the many problems with this holding is that it contradicted the prior case law of the Convention institutions, which the Court was now forced to ‘clarify’, rather than explicitly depart from. It thus said the following:
Additionally, the Court notes that other recognised instances of the extra-territorial exercise of jurisdiction by a State include cases involving the activities of its diplomatic or consular agents abroad and on board craft and vessels registered in, or flying the flag of, that State. In these specific situations, customary international law and treaty provisions have recognised the extra-territorial exercise of jurisdiction by the relevant State.112
Note, first, that the Court here speaks about acts of diplomatic and consular agents, not diplomatic and consular premises, as the supposed ‘embassy exception’ was interpreted by UK courts. Even so, this paragraph is doubly wrong. First, to the extent that it claims to restate the prior case law of the Strasbourg institutions, it actually provides an anachronistic reinterpretation of these cases, as we will now see. Secondly, to the extent that it argues that public international law recognizes some sort of special or exceptional exercise of state jurisdiction over embassies, ships, or aircraft, it is incorrect, and again mixes up the various meanings of the word (p.155) ‘jurisdiction’ and accordingly the various concepts of jurisdiction in international law.
Let us first take a look at the case law—the case law being that of the now defunct European Commission, since the Court itself has to my knowledge never taken up an ‘embassy’ case. Perhaps the first such case, decided in 1965, was X v. Federal Republic of Germany,113 which dealt with several complaints made by a German national residing in Morocco against the conduct of German diplomatic and consular officials there. In that case the Commission gave the briefest consideration to the issue of extraterritorial application, in the following terms:
…in certain respects, the nationals of a Contracting State are within its ‘jurisdiction’ even when domiciled or resident abroad; […] in particular, the diplomatic and consular representatives of their country of origin perform certain duties with regard to them which may, in certain circumstances, make that country liable in respect of the Convention.114
As I have argued above, the nationality of the victim of a human rights violation, on which the Commission may have relied, should have no bearing on the question of extraterritorial application.115 On the other hand, the Commission’s second sentence might be taken as regarding as relevant the supposedly special nature of diplomatic and consular agents. Ten years later, in its first decision on northern Cyprus,116 the Commission developed on its holding in X v. Germany:
The Commission further observes that nationals of a State, including registered ships and aircraft, are partly within its jurisdiction wherever they may be, and that authorised agents of a State, including diplomatic and consular agents and armed forces, not only remain under its jurisdiction when abroad but bring any other persons or property ‘within the jurisdiction’ of that State, to the extent that they exercise authority over such persons or property. Insofar as, by their acts or omissions, they affect such persons or property, the responsibility of the State is engaged.117
Again the Commission refers to the relevance of nationality, now subsuming registered ships and aircraft under that heading. More importantly, however, it says that the authorized agents of a state, including, but not limited to its diplomatic and consular agents, remain under the state’s jurisdiction, and can bring under its jurisdiction those persons over which they exercise authority. In other words, the Commission adopted a personal, rather than spatial conception of Article 1 ECHR jurisdiction, which it thought applicable to all state agents. This became the constant jurisprudence of the Commission.118
(p.156) Viewed as a whole, the Commission’s jurisprudence gave little weight to the supposedly special nature of diplomatic or consular officials, or of diplomatic and consular premises, for the purpose of establishing state jurisdiction within the meaning of Article 1. Nor does general international law, for that matter. It is true that diplomatic and consular officials are permitted to exercise certain public functions over individuals while in the territory of the receiving state—from issuing visas and entry permits, to protecting their nationals abroad. But the only reason why they can do so is because the receiving state has consented to their presence. In times past, diplomatic and consular officials in certain non-Western states had far greater powers under the rubric of consular jurisdiction, which entailed, for example, their exclusive capacity to try their own nationals for offences committed in the host state. Even then, however, this capacity was based on the host state’s consent through so-called capitulations or unequal treaties.119 In modern international law, which is premised on the sovereign equality of all states, the role of diplomatic and consular officials is far more limited.
Either way, it is entirely misleading to say that customary law recognizes the ‘special’ nature of the extraterritorial exercise of state jurisdiction by these officials. There is nothing special about it, except that it is regularized, in the sense that a defined set of customary and treaty rules, agreed upon in advance, applies to these officials once the territorial state gives its consent. But the basis for their acts is still exclusively in the consent itself. The territorial state can withdraw its consent at any time, or can provide consent which is much wider in scope, for example by inviting a foreign military intervention, or by allowing a foreign power to run a military prison. As with the case of effective control over territory generally, this consent is only relevant for assessing whether the extraterritorial act of a foreign state is lawful or not, i.e. whether it violates the sovereignty of the territorial state, but it has no bearing on the fact of the extraterritorial exercise of state power, which may affect individuals in the territorial state. Again, there is simply no valid explanation as to why consent by the territorial state should matter for the purposes of establishing Article 1 ECHR state jurisdiction.120
On the other hand, it is true that both customary and conventional international law provide for a limited immunity of diplomatic and consular officials, and (p.157) diplomatic and consular premises, from the host state’s jurisdiction.121 And, of course, the host state may well decide to grant similar immunities to other foreign agents on its territory, as is often the case, for instance, with foreign troop contingents and military bases, the deployment of which frequently requires the conclusion of a status of forces agreement that may provide for various privileges and immunities.122 Does this mean that even though they are present in the host state’s territory, or more precisely a territory under its control, these officials and premises are not actually within the host state’s jurisdiction, within the meaning of the relevant human rights treaties?
The short answer is no. To develop that, let us take a closer look at the immunity rules of diplomatic and consular law. For example, Article 22 of the Vienna Convention on Diplomatic Relations provides that the premises of a diplomatic mission shall be inviolable; that they may not be entered by the agents of the receiving state except with the consent of the head of mission; that they will be immune from search, requisition, or execution; and that the receiving state has a special duty to protect them from third parties. With respect to diplomatic agents, Article 29 of the Convention provides that their persons shall be inviolable, while Article 31 stipulates that they shall be immune from the criminal jurisdiction of the receiving state, as well as from civil jurisdiction except in certain cases, such as disputes regarding private immovable property. In any event a diplomatic agent enjoys immunity from execution.
However, pursuant to Article 41(1) of the Convention, ‘[w]ithout prejudice to their privileges and immunities, it is the duty of all persons enjoying such privileges and immunities to respect the laws and regulations of the receiving State. They also have a duty not to interfere in the internal affairs of that State.’ Article 41(3) further provides that ‘[t]he premises of the mission must not be used in any manner incompatible with the functions of the mission as laid down in the present Convention or by other rules of general international law or by any special agreements in force between the sending and the receiving State’.
So, for example, are the Serbian ambassador to the United Kingdom and the Serbian embassy in London within the UK jurisdiction for the purposes of Article 1 ECHR or the relevant jurisdiction clauses of other human rights treaties? Or are they rather within Serbian jurisdiction? Let us consider the following (purely hypothetical!) examples:
(1) The Serbian ambassador, while acting in his official capacity and under authorization from the Serbian government, kills somebody in London; was the victim of that deprivation of life within UK or Serbian jurisdiction, and could his family file an application against either the UK or Serbia with Strasbourg?
(p.158) (2) The Serbian ambassador, again while acting in his official capacity, kills someone on the premises of the Serbian embassy in London; did, for instance, the UK have the positive obligation to secure the right to life of the victim, even though he was present within the Serbian embassy?
(3) Conversely, the Serbian ambassador, now a complete innocent, is killed by some crazy person while on a stroll in Hyde Park; can his family file an application against the UK for failing to secure his right to life?
To address these scenarios, it is again important to recall the various meanings that the word ‘jurisdiction’ can have in international law. The ‘jurisdiction’ from which diplomatic agents have immunity within certain limits is the jurisdiction to enforce, the state’s authority to enforce the laws that it has previously prescribed. However, diplomatic agents do not have immunity from the receiving state’s jurisdiction to prescribe—they must obey and comply with applicable domestic law. On the other hand, our jurisdiction—the state jurisdiction referred to in human rights treaties—is neither of these. Rather, it refers to state control through its agents over territory, or perhaps a place, or maybe even an individual.123
So, to look at example (1), if the Serbian ambassador were to kill a Londoner, the victim would certainly be within the UK’s jurisdiction, as the killing took place in London, an area under which the UK has effective overall control. The victim would only be within Serbia’s jurisdiction if we consider that jurisdiction can also arise from the authority and control exercised by state agents over an individual, even if that control is exercised in an area outside the state’s control. This is of course the personal rather than spatial conception of jurisdiction, the model which I will be examining shortly in some detail. However, whatever the general validity of that model, it seems to me to be perfectly irrelevant whether the killing was committed by a diplomatic agent, such as the Serbian ambassador, or by any other authorized agent of Serbia, i.e. a person whose acts are attributable to it.
Now, if the victim of the killing was within the UK’s jurisdiction, this means that the UK had a positive obligation to secure the right to life of the victim. Normally, this obligation has two aspects: the UK would have to do all it could to prevent the killing from taking place, and if it was nonetheless unable to do so, it would have to investigate the crime and prosecute the person responsible.124 On the other hand, Article 29 of the Vienna Convention provides that ‘[t]he person of a diplomatic agent shall be inviolable. He shall not be liable to any form of arrest or detention’, while Article 31 exempts diplomatic agents from the territorial state’s criminal jurisdiction. Imagine, therefore, that under the observant gaze of UK police officer, the Serbian ambassador accosts his victim on a London street, holding a gun in his right hand, while carrying in his left hand a sign saying ‘I am the Serbian ambassador, I am acting in my official capacity, and my person is inviolable.’ Would the (p.159) police officer have the positive obligation to tackle or maybe even shoot the ambassador to prevent the killing from taking place?
This situation is one of apparent norm conflict, with the ECHR commanding the UK to prevent the killing, and the Vienna Convention prohibiting it from taking any practical measures in doing so.125 This norm conflict can perhaps be avoided. Thus, we could say either that the ECHR positive obligation to prevent the killing should be read down so that the UK does not have to act in such a way as to violate its other international obligations by infringing on the ambassador’s person, or that it is the Vienna Convention that should be read down so that a limited exception is carved out of its prohibitions, allowing the authorities of the receiving state to prevent diplomatic agents from committing imminent unlawful acts (as a matter of policy, if not the treaty text, the most reasonable solution).126 Alternatively, if such a reading down of either treaty is impossible, then we would have an unresolvable norm conflict on our hands—whatever the police officer did, the UK would violate one of its international obligations.
Turning now to example (2), if the Serbian ambassador killed someone in the Serbian embassy, could the victim now be said to be within Serbian jurisdiction? If we consider that Article 1 ECHR jurisdiction extends to control over places, and not just large areas of territory, à la Al-Saadoon, then one could say yes to that question. But again, the fact that the premises in question are an embassy to me seems to be perfectly irrelevant. If we recall for a moment the example of the CIA black site in Lithuania, it would seem entirely arbitrary to say that a person detained in the US embassy in Vilnius would be within US jurisdiction, but that a person detained at the Vilnius riding school black site a few miles away would not, just because the former is a diplomatic mission. Not only would such a position be devoid of any broader principle, but it would manifestly open the door to abuse.
But what about the positive obligations of the host country? If, say, UK police officers stationed near the Serbian embassy in London heard loud screaming coming from the embassy, with the poor victim shouting at the top of his lungs that he was being tortured inside, would the UK have a positive obligation under the ECHR to put an end to the crime, by entering the embassy by force if necessary, as it would in all other circumstances? One possible answer is that since the Serbian embassy in London is within Serbia’s jurisdiction, it would be outside the UK’s jurisdiction, and therefore that the UK would owe no positive obligations to persons in it. However, recall that jurisdiction need not necessarily be exclusive. There is to my mind nothing contradictory in saying that the Serbian embassy in London is a place under Serbia’s control, and therefore within its (p.160) jurisdiction, but that this place is located in an area under UK control, and therefore also within the UK’s jurisdiction. And if this so, then just as with example (1) we are faced with an apparent norm conflict between the UK’s obligations under the ECHR and its obligations under diplomatic law.127
As for example (3), that the Serbian ambassador to the UK is immune from UK jurisdiction in the sense that the UK may not subject him to criminal proceedings, detention, or arrest, does not mean that he is actually not within the UK’s jurisdiction for the purposes of Article 1 ECHR. If he were actually to be killed while taking a walk in London, the UK would not only have an obligation to prevent and investigate his death under the Vienna Convention—it would also have such an obligation under Articles 1 and 2 ECHR.
These scenarios could of course be considered to be quite fanciful. After all, aside from the occasional traffic or hunting accident, ambassadors do not normally go about killing or torturing people. These examples do, however, convincingly expose the sheer folly of considering either diplomatic and consular agents, or the premises from which they operate, as somehow being ‘special’ for the purposes of establishing state jurisdiction under human rights treaties.128 These agents are not distinguishable by their function from all other state agents, nor are these premises different from other places under state control. We should also not forget that states may grant similar immunities to foreign military forces or other agents. Nor, consequently, does an analogy between an embassy and any other place, such as a military prison in Al-Skeini, make any sense whatsoever. It is manifestly nothing more than a device used to narrow extraterritorial application to those situations which we substantively consider to be morally intolerable, such as the abuse and killing of a defenceless prisoner, while rejecting extraterritorial application in cases which are not so black and white.
3. Ships and aircraft
The Bankovic and Al-Skeini dicta on the supposedly special nature of ships and aircraft are similarly misguided. Again, while it is true that international law recognizes that states may exercise prescriptive jurisdiction over registered ships and aircraft, this is simply not the ‘jurisdiction’ referred to in Article 1 ECHR. Let me once more bring up the example of Article 9(1) of the UN Enforced Disappearances Convention, which provides as follows:
Each State Party shall take the necessary measures to establish its jurisdiction over the offence of enforced disappearance:
(p.161) (a) When the offence is committed in any territory under its jurisdiction or on board a ship or aircraft registered in that State;
(b) When the alleged offender is one of its nationals;
(c) When the disappeared person is one of its nationals and the State Party considers it appropriate.
The first use of the word ‘jurisdiction’ in the chapeau of this article refers to the concept of the jurisdiction to prescribe. The article as a whole requires states to make enforced disappearances crimes under their domestic law, and that the scope of application of this domestic law should be based on the territorial and active personality and if the state wishes also the passive personality heads of prescriptive jurisdiction. It is this ‘jurisdiction’ of states that general international law specifically recognizes with regard to registered ships and aircraft, by assimilating them to state territory for the purpose of the state’s authority to legislate. In other words, just as international law accepts that it is a fundamental prerogative of a state’s sovereignty to prescribe legal rules for persons and their conduct within its territory, so does it accept the state’s authority to do the same with regard to registered ships and aircraft or those flying its flag.129
But this ‘jurisdiction’ is not the jurisdiction referred to in Article 1 ECHR or in the similar clauses of other human rights jurisdiction. That ‘jurisdiction’ is the one that makes its appearance in subparagraph (a) quoted above: ‘territory under its jurisdiction’. When Article 9(1)(a) of the Disappearances Convention stipulates that a state party shall have the duty to ‘establish its jurisdiction…when the offence is committed in any territory under its jurisdiction’, it is not setting out a tautology. Rather, the meaning of the two ‘jurisdictions’ is not the same. The former refers to the authority to prescribe legal rules, the latter to effective overall control by a state over a territory. Therefore, in plain English, the article as a whole requires states to criminalize the Convention offences in all territories under their control.
The question thus is not whether the former concept of jurisdiction can apply to registered ships and aircraft—it clearly does. Rather, the issue is whether the spatial concept of jurisdiction as control over territory and places can also extend to cover ships and aircraft, as distinct man-made objects. And, if it can, what should be its parameters? Should it really matter whether the ship or aircraft is registered in a specific state, or whether it flies its flag, or should jurisdiction over such objects rather depend on state control over them?
To answer this question, we now need to take a closer look at some of the case law. With regard to ships, perhaps the European Court’s earliest case is Rigopoulos v. Spain.130 The applicants were the crew of a Panamanian-flagged ship that was on the high seas when it was boarded by a Spanish vessel on suspicion of drug trafficking, pursuant to a warrant by a Spanish court and with the consent of Panamanian authorities. They complained that they were unlawfully detained. (p.162) Though the Court declared the application inadmissible as manifestly ill-founded, it seems to have accepted that the ECHR could apply to a Panamanian ship on the high seas, though it did not explicitly consider Article 1 ECHR.
In Xhavara,131 the applicants were Albanian nationals who were attempting to get to Italy on an Albanian-registered ship, when that ship was rammed by an Italian warship on the high seas. The applicants alleged that the Italian warship’s action was intentional, and that it therefore engaged Articles 2 and 3 of the Convention. The Court declared the application inadmissible for failure to exhaust domestic remedies, but it made no mention of the fact that the applicants were located on an Albanian ship on the high seas at the time the violation occurred. In other words, it again did not seem to have considered that there was an Article 1 ECHR issue at all.
Article 1 ECHR did make an appearance in Medvedyev and Others v. France, the facts of which remarkably resembled Rigopoulos: the applicants were crew members on a merchant ship infelicitously named the Winner, flying the Cambodian flag. As part of the international effort to combat drug trafficking, the French authorities were informed that the ship might be carrying large quantities of illegal drugs. France requested, and obtained, permission from Cambodian authorities to search and seize the ship, and detain those aboard, which the French navy swiftly did. The ship was then brought to the French port of Brest, a voyage of some thirteen days, where the applicants were brought before a magistrate. In Strasbourg, they complained inter alia that their detention on board the ship was unlawful, because it had no legal basis, and because they were not promptly brought before a judge.
The case was first decided by a Chamber.132 France accepted that it had Article 1 ECHR jurisdiction, while the Court itself said the following:
…the Court notes, on the one hand, that it is not disputed that between 13 June 2002 (the date on which the Winner was intercepted) and 26 June 2002 (when it arrived in Brest harbour) the Winner and its crew were under the control of French military forces, so that even though they were outside French territory, they were within the jurisdiction of France for the purposes of Article 1 of the Convention. It further notes that the parties agree that throughout that period on board the Winner—and the subsequent police custody—the applicants were deprived of their liberty within the meaning of Article 5 of the Convention, ‘for the purpose of bringing them before the competent legal authority’ (Article 5 § 1 (c)).
That is also the opinion of the Court, which also refers to the Rigopoulos decision cited above.133
Note that the Chamber did not attach any importance to the supposedly ‘special’ nature of ships. Nor could it have done so, as the Winner was a Cambodian, and not a French-flagged ship. It therefore did not base France’s jurisdiction over the applicants on some alleged general principle of public international law, but on (p.163) the fact that the ship and the crew ‘were under the control of French military forces’.134
The case was then referred to the Grand Chamber, which delivered its judgment in 2010.135 It addressed the Article 1 issue in much more detail than the Chamber before it. First, unlike the Chamber, which did not even mention Bankovic, the Grand Chamber took great pains in restating its Bankovic position on the exceptional nature of extraterritorial application:
In keeping with the essentially territorial notion of jurisdiction, the Court has accepted only in exceptional cases that acts of the Contracting States performed, or producing effects, outside their territories can constitute an exercise of jurisdiction by them for the purposes of Article 1 of the Convention (see Banković, cited above, § 67, and Ilaşcu and Others v. Moldova and Russia [GC], no. 48787/99, § 314, ECHR 2004-VII). In its first Loizidou judgment (preliminary objections), for example, the Court found that bearing in mind the object and purpose of the Convention, the responsibility of a Contracting Party might also arise when as a consequence of military action—whether lawful or unlawful—it exercised effective control of an area outside its national territory (see Loizidou v. Turkey (preliminary objections) [GC], 23 March 1995, § 62, Series A no. 310). This excluded situations, however, where—as in the Banković case—what was at issue was an instantaneous extraterritorial act, as the provisions of Article 1 did not admit of a ‘cause-and-effect’ notion of ‘jurisdiction’ (Banković, § 75).
Additionally, the Court notes that other recognised instances of the extraterritorial exercise of jurisdiction by a State include cases involving the activities of its diplomatic or consular agents abroad and on board aircraft and ships registered in, or flying the flag of, that State. In these specific situations, customary international law and treaty provisions have clearly recognised and defined the extraterritorial exercise of jurisdiction by the relevant State (see Banković, cited above, § 73).136
Note, first, the new gloss that the Grand Chamber puts on Bankovic—that it only concerned an ‘instantaneous extraterritorial act’. With an eye on the future, this could have been meant to distinguish a killing from detention, though not on any principled ground that I can see. The Grand Chamber then reiterates the supposedly special nature of the jurisdiction exercised by states over ships registered in them or flying their flag. But, of course, even if true, this is entirely irrelevant on the facts of Medvedyev, as the ship on which the applicants were detained was not French, but Cambodian. The Grand Chamber then moved to applying (or not) these principles to the actual facts of Medvedyev:
In the instant case, the Court notes that a French warship, the frigate Lieutenant de vaisseau Le Hénaff, was specially instructed by the French naval authorities to intercept the Winner, and that the frigate sailed out of Brest harbour on that mission carrying on board the French Navy commando unit Jaubert, a special forces team specialised in boarding vessels at sea. When the Winner was spotted off Cape Verde on 13 June 2002, the frigate issued several (p.164) warnings and fired warning shots, before firing directly at the merchant ship, under orders from France's Maritime Prefect for the Atlantic. When they boarded the Winner, the French commando team were obliged to use their weapons to defend themselves, and subsequently kept the crew members under their exclusive guard and confined them to their cabins during the journey to France, where they arrived on 26 June 2002. The rerouting of the Winner to France, by decision of the French authorities, was made possible by sending a tug out of Brest harbour to tow the ship back to the French port, escorted by another warship, the frigate Commandant Bouan, all under orders from the Maritime Prefect and at the request of the Brest Public Prosecutor.
That being so, the Court considers that, as this was a case of France having exercised full and exclusive control over the Winner and its crew, at least de facto, from the time of its interception, in a continuous and uninterrupted manner until they were tried in France, the applicants were effectively within France's jurisdiction for the purposes of Article 1 of the Convention (contrast Banković, cited above).137
So, as with the Chamber, what really matters is that France had ‘full and exclusive control over the Winner and its crew, at least de facto’, not some considerations of general international law. Note again that it is unclear whether the Court applies a spatial model (control over the ship) or a personal one (control over the crew). And note how, without the slightest hint of self-irony, the Court instructs us to contrast this approach with Bankovic—as if the two were somehow related or compatible!138 Nor does it seem to have been relevant that Cambodia had consented to the seizure of the ship—should, after all, the outcome have been any different otherwise?
With regard to aircraft, the first Strasbourg case is the Commission inadmissibility decision in Freda v. Italy,139 which concerned an applicant arrested in Costa Rica and surrendered by Costa Rican authorities to Italian agents at an airport, who put him on an Italian plane back to Italy. The Commission considered that the applicant was within Italy’s jurisdiction:
Or, il est établi que le requérant a été pris en charge par des agents de la force publique italienne et privé de liberté dans un avion militaire italien. Dès lors, le requérant, à partir du moment de la remise, relevait effectivement de l’autorité de l’Italie, et donc de la ‘juridiction’ de ce pays, même si cette autorité s’est exercée en l’occurence à l’étranger.
The Commission relied, inter alia, on its prior decision in Cyprus v. Turkey (1975), where it adopted the personal conception of Article 1 jurisdiction as control over individuals by state agents, and that is exactly what it did here: ‘…le requérant, à partir du moment de la remise, relevait effectivement de l’autorité de l’Italie, et donc de la “juridiction” de ce pays’. The Commission repeated this same reasoning in Illich Sanchez Ramirez v. France,140 one of the two Strasbourg cases dealing with (p.165) the notorious terrorist Carlos the Jackal.141 The (very colourful) facts of the case were these: in the early 1990s, Carlos sought refuge in Sudan as he was being actively hunted by the intelligence agencies of several Western states. In 1994, the Sudanese authorities made a deal with French and US intelligence services to hand Carlos over. Unfortunately for Carlos, he was at the time scheduled to have minor surgery in a Sudanese hospital on a varicose vein on his scrotum. After the surgery, he was moved to a villa and assigned bodyguards. During the night of 14–15 August 1994, his own bodyguards turned against him. He was promptly tied up and tranquilized, surrendered to French agents waiting at the airport, and was immediately flown on a French military plane back to Paris, where he was later tried and convicted for various offences. He complained, inter alia, that his arrest was unlawful under Article 5 ECHR.
At what point, therefore, did Carlos come within French jurisdiction? The Commission held as follows:
According to the applicant, he was taken into the custody of French police officers and deprived of his liberty in a French military aeroplane. If this was indeed the case, from the time of being handed over to those officers, the applicant was effectively under the authority, and therefore the jurisdiction, of France, even if this authority was, in the circumstances, being exercised abroad.
Again, what mattered was that Carlos was handed over to French officers, not that he was put on a French-flagged plane.
Very similar is Öcalan v. Turkey, the case filed by Abdullah Öcalan, the leader of the Workers’ Party of Kurdistan (PKK), considered by Turkey to be the arch-terrorist. The facts of the case could be briefly summarized as follows: after having fled Turkey to several countries, in 1999 Öcalan found himself in Nairobi, Kenya. On 15 February 1999 he was arrested, and was flown to Turkey aboard a Turkish military plane. It is unclear whether he was arrested with the permission and help of (some) Kenyan authorities, or whether the Turkish forces simply engaged in an extraterritorial abduction. Before the European Court, he complained inter alia about the conditions and the lawfulness of his detention, and the lack of fairness of his subsequent trial, after which he was sentenced to the death penalty. What concerns us here, of course, is the detention—at what point was Öcalan to be considered to have been within Turkey’s jurisdiction? Was it already in Kenya upon his arrest, or was it aboard the Turkish aircraft, or was it when that aircraft entered Turkish airspace?
The Court’s admissibility decision made no mention of the issue.142 During the merits phase before the Chamber, however, Turkey raised the Court’s freshly minted Bankovic decision to disclaim that it had jurisdiction at the time of Öcalan’s arrest in Kenya. The Chamber rejected this argument as follows:
In the instant case, the applicant was arrested by members of the Turkish security forces inside an aircraft in the international zone of Nairobi Airport. Directly after he had been (p.166) handed over by the Kenyan officials to the Turkish officials the applicant was under effective Turkish authority and was therefore brought within the ‘jurisdiction’ of that State for the purposes of Article 1 of the Convention, even though in this instance Turkey exercised its authority outside its territory. The Court considers that the circumstances of the present case are distinguishable from those in the aforementioned Banković and Others case, notably in that the applicant was physically forced to return to Turkey by Turkish officials and was subject to their authority and control following his arrest and return to Turkey.143
As in the Carlos case, the Chamber clearly adopted the personal conception of jurisdiction as the exercise of ‘effective Turkish authority’ or the subjection of the applicant to its ‘authority and control’. Jurisdiction was most emphatically not based on the flag of the aircraft that transported Öcalan.
As the Chamber generally decided the case in Öcalan’s favour, it was then referred to the Grand Chamber at Turkey’s request. This time, however, Turkey did not dispute that it had jurisdiction over Öcalan at the time of his arrest, and the Grand Chamber merely reproduced some of the Chamber’s reasoning:
It is common ground that, directly after being handed over to the Turkish officials by the Kenyan officials, the applicant was effectively under Turkish authority and therefore within the ‘jurisdiction’ of that State for the purposes of Article 1 of the Convention, even though in this instance Turkey exercised its authority outside its territory. It is true that the applicant was physically forced to return to Turkey by Turkish officials and was under their authority and control following his arrest and return to Turkey (see, in this respect, the aforementioned decisions in Sánchez Ramirez and Freda, and, by converse implication, Banković and Others v. Belgium and Others (dec.) [GC], no. 52207/99, ECHR 2001-XII).144
Note how both the Chamber and the Grand Chamber in Öcalan try to make it seem that there is no contradiction between their holding that jurisdiction attached because of authority and control over an individual and the Bankovic rejection of such a conception of jurisdiction. It is unclear what exactly in the Court’s view makes Bankovic distinguishable ‘by converse implication’. Is it because ‘authority and control’ over an individual means physical custody of that individual, rather than just the ability to kill that individual, and if so, why? Or is it rather that Turkey exercised some legal power, i.e. that it was trying to arrest Öcalan on a criminal charge that rendered him within its jurisdiction, and if so, why?
In my view, at play was simply the Court’s policy consideration that extraterritorial abductions, even if somewhat extraordinary, could be effectively dealt with without much controversy or political fallout, while the use of force in an armed conflict is a wholly different game. Be that as it may, we can see that in many of the cases that we have now examined neither the state concerned, nor the Court or the Commission, which should after all have minded ex officio whether the ECHR applied and accordingly whether they had subject-matter jurisdiction, thought that there was an Article 1 ECHR issue.145 Even in cases where Article 1 ECHR was (p.167) considered, they simply did not turn on any supposedly ‘special’ nature of ship and aircraft in general international law. Rather, it was the authority and control exercised by state agents over an individual that was the basis of the state’s jurisdiction. In other words, it was not the spatial but the personal model of jurisdiction that was applied.
In sum, the European cases cited above most certainly do not support the proposition that because international law recognizes that the flag state or the state in which ships and aircraft are registered may prescribe legal rules regulating conduct occurring on them, this means that ships and aircraft somehow present a special justification for extending to them the spatial concept of jurisdiction in human rights treaties. Indeed, except perhaps for Medvedyev, the cases do not even stand in favour of a purely factual notion of jurisdiction as control over ships and aircraft, regardless of the flag or registration. Rather, it is the control over individuals than forms the basis of state jurisdiction. No amount of post-Bankovic reinterpretation or ‘clarification’ of these cases can change this—the cases say what they say.
As for the case law of other institutions, recall that the Committee Against Torture in its General Comment No. 2 stated that the ‘reference to “any territory” in article 2, like that in articles 5, 11, 12, 13 and 16, refers to prohibited acts committed not only on board a ship or aircraft registered by a State party, but also during military occupation or peacekeeping operations and in such places as embassies, military bases, detention facilities, or other areas over which a State exercises factual or effective control’.146 The Committee here endorsed a general theory of jurisdiction as control over places or objects, and explicitly referred to ships and aircraft. It had the opportunity to apply this theory in the Marine I case,147 which concerned a ship, the Marine I, which capsized in international waters with 369 immigrants from various Asian and African countries on board. In response to a distress call, a Spanish ship reached the Marine I and towed it to Mauritanian waters, where it was anchored for eight days, while negotiations between Spain and several other states were underway with the purpose of repatriating the migrants. They were then transported to a former fish plant in Mauritania, and most of them were subsequently repatriated voluntarily. The complainants refused repatriation, and complained inter alia of the inhumane conditions while on board the ship and in the repatriation centre.
Spain argued that it did not have jurisdiction over the complainants, and the Committee held as follows:
(p.168) The Committee takes note of the State party’s argument that the complainant lacks competence to represent the alleged victims because the incidents forming the substance of the complaint occurred outside Spanish territory. Nevertheless, the Committee recalls its general comment No. 2, in which it states that the jurisdiction of a State party refers to any territory in which it exercises, directly or indirectly, in whole or in part, de jure or de facto effective control, in accordance with international law. In particular, it considers that such jurisdiction must also include situations where a State party exercises, directly or indirectly, de facto or de jure control over persons in detention. This interpretation of the concept of jurisdiction is applicable in respect not only of article 2, but of all provisions of the Convention, including article 22. In the present case, the Committee observes that the State party maintained control over the persons on board the Marine I from the time the vessel was rescued and throughout the identification and repatriation process that took place at Nouadhibou. In particular, the State party exercised, by virtue of a diplomatic agreement concluded with Mauritania, constant de facto control over the alleged victims during their detention in Nouadhibou. Consequently, the Committee considers that the alleged victims are subject to Spanish jurisdiction insofar as the complaint that forms the subject of the present communication is concerned.148
Clearly, in Marine I the Committee generally conceived of jurisdiction as ‘de facto or de jure control over persons’ or ‘constant de facto control over the alleged victims’.149 It did not rely on jurisdiction as control over the ship itself. And this raises a fundamental question. If we accept control over individuals as a basis for jurisdiction, do we even need any concept of jurisdiction as control over vessels? What good would it actually do?
Perhaps the two most practically relevant types of cases which involve ships today are the concerted actions by a number of states against Somali pirates in the Gulf of Aden, and state attempts to curb illegal immigration. The former raise several potential human rights issues, from the lawfulness of the detention of any suspected pirates, to the transfer of the captured pirates to a nearby state for trial, which could implicate the non-refoulement obligations of the capturing state, for instance with regard to the risk of inhumane treatment in the receiving state or the lack of adequate fair trial guarantees.150
When it comes to illegal immigration, we have already seen one example in the Marine I case examined above. Attempts by migrants from, say, North Africa to reach Europe, or those from Cuba to reach the United States, frequently involve the use of ships and the interdiction of these ships by the state that the migrants are attempting to reach. It has been widely reported, for example, that Italy chooses to interdict migrant ships before they reach Italian territorial waters and return them to their harbour of origin, rather than assess the claims of the migrants to asylum on an individual basis.151 This again, of course, raises questions about the lawfulness of (p.169) detention, as well as non-refoulement.152 As of the time of writing, one such case, Hirsi and others v. Italy,153 is pending before the European Court.154
Turning back to the question whether the spatial conception of jurisdiction as control over ships and aircraft would have any use if the personal conception of jurisdiction as control over individuals is accepted, we could only say that such usefulness would be very limited, if it existed at all. In most cases, it would be impossible to distinguish as a matter of fact between the control over a ship and the control over individuals on it. To the extent that a personal notion of jurisdiction would govern, the spatial notion would be redundant.
If, however, the personal notion is rejected, either because it is limited artificially, or because it is at odds with the text of the relevant treaty (as is arguably the case with the CAT), or indeed, as is my view, because it is impossible to maintain a personal notion of jurisdiction which would not collapse into the absence of any jurisdictional threshold at all, then there could be some use to a concept of jurisdiction as control over objects, particularly with regard to delineating the applicability of positive obligations. What if, say, a murder occurs on a British Airways flight from London to France, either while the UK-registered aircraft was in international airspace, or indeed just after the plane had touched the tarmac at Charles de Gaulle. Which of the two states has the Article 2 ECHR obligation to investigate the crime? Similarly, what of a scenario where one shipmate kills another on a private ship registered in the UK, and the murder takes place on the high seas? Would the UK have an Article 2 obligation to investigate?
In both cases there is no direct control of a state over the ship or aircraft. All we have is the fact that the vessel is registered in a state, which certainly grants the state concerned the right in customary law to prescribe offences aboard these vessels and enforce these laws, as well as the obligation to do so under certain treaties,155 but it is not immediately apparent that this would constitute jurisdiction in the spatial sense within the meaning of human rights treaties. In the cases we have examined, that notion of jurisdiction has been predicated on factual control, rather than purely formal criteria such as registration. If we take Medvedyev as an example, it (p.170) is the fact that France controlled the individuals aboard the ship that created its jurisdiction, even if the individuals were located on a Cambodian-flagged ship.
If the killing takes place within the territorial sea or the airspace of a country (over which it maintains effective control, or is presumed to do so), whether it is the flag state or a third state, then we could indeed say that the killing took place within that state’s jurisdiction. If, on the other hand, the killing takes place on the high seas or in international airspace, then it is to me far from obvious that the mere fact of formal registration, absent any tangible degree of control, should matter at all. Imagine, for example, if the aircraft in question was registered in one state, but was leased to an airline operating in another state half the world away. Could the state of registration still be said to have an obligation to secure the human rights of all persons aboard that airplane?
If, on the other hand, it is only purely factual control over ships and aircraft that should matter—as is the case with territory and places—then we would need to define some of the contours of that control. First, it scarcely needs pointing out that control over such an object can be effected with relative ease by any state. Secondly, the simplest case of such control would be where the state’s soldiers actually board a vessel, or control it directly in a similar fashion. Thirdly, even so, control over these objects need not necessarily be so direct. If, say, a migrant ship was interdicted by a state warship, which has every ability to blow it out of the water, or if an aircraft was intercepted by military jets, this would amount to effective control just as if the ship or aircraft was actually boarded by troops.156
In short, precisely because control over such objects is relatively easy to achieve, it is the potential of such control, rather than the immediate exercise of control in any given situation, that could be said to give rise to state jurisdiction over a vessel.
Ultimately, what are we to make of the spatial model of jurisdiction? Its application is certainly the clearest when it comes to effective state control over relatively large portions or areas of territory, as in Loizidou. Though of course the facts on the ground may vary, these situations resemble most the level of control, and accordingly the capacity to either violate or protect human rights, that the state normally has over its own territory. This model also comports the best with the territorial sense of the word ‘jurisdiction’ as it is used in state practice. As a matter of policy, its adoption allows us to extend human rights protection to those extraterritorial situations where it is most likely that human rights will be violated and yet could be effectively protected, primarily belligerent occupation.
But the main drawback of the spatial model is precisely that, as a matter of policy, its rigid application will allow numerous extraterritorial cases to slip through and be beyond the scope of application of human rights treaties. There is a profound feeling of injustice and arbitrariness, as measured against the benchmark (p.171) of universality, in allowing states to avoid human rights constraints in extraterritorial situations over which they still have complete control. Why should, for example, the US be free to torture persons detained in CIA black sites located in territories under the jurisdiction of other states, when the US is still in control of the acts of its own agents and in control of the individuals subjected to its authority?
This is why the pure or default version of the spatial model, that of effective overall control of (large) areas of territory, needs to be mitigated, and there are several ways of doing so. First, by applying the spatial model to ever smaller areas, with an ever more tenuous degree of control, so that the ‘control of an area’ test ultimately covers control over places and objects. Secondly, by in some cases discarding the spatial model and opting for the personal model of jurisdiction as control over individuals. There is also a third option—that of discarding the jurisdiction threshold altogether with respect to some obligations—which I personally favour, and which I will turn to shortly.
From the Strasbourg perspective, the virtue of the first approach is that it allows the Court to avoid using the second one. This, in turn, allows it to maintain some rigour to the Article 1 jurisdiction threshold, which would be in danger of collapsing if jurisdiction meant factual control over individuals pure and simple. Even so, however, because the spatial model can only be stretched so far, the Court is, due to the normative pull of universality, in many cases unable to resist conceptualizing jurisdiction as control over individuals, thereby creating a tension with Bankovic. As we have seen, the embassy, ship, and aircraft cases that the Grand Chamber in Bankovic rationalized on the basis of some supposedly special characteristics of these objects in international law, actually turned on considerations of factual control by state agents over individuals. Even in Al-Saadoon, which can be read as embracing a general theory of jurisdiction as control over places, the Court vacillated between control over places and control over individuals as grounds of jurisdiction.157
As we have also seen above, it is impossible to clearly define the geographical space that the spatial model of jurisdiction refers to. The more we descend from large areas of territory to places and objects, the more artificial it seems to apply the spatial model at all: it simply collapses to control over individuals, rather than control over territory, however minuscule in size. And because this is so, even if stretched to the utmost extent, the spatial model of jurisdiction will allow numerous extraterritorial situations to slip through and be outside the reach of human rights protection. This, in turn, would in many cases simply be morally intolerable, unjustifiable from the standpoint of universality, and inexcusable by any apparent consideration of effectiveness.
Consider, for example, the assassination of Alexander Litvinenko in London in 2006. Assume that it was, in fact, a Russian agent who killed Mr Litvinenko—say if a KGB/FSB operative openly admitted to the crime. It would nonetheless be impossible to say under any conception of the spatial model of jurisdiction that (p.172) this killing violated Russia’s ECHR obligations, which it would have done if it had taken place on Russian soil. It would surely be absurd to argue that Russia had control over the Japanese restaurant in London where Litvinenko was fed the radioactive sushi, and that this somehow established its jurisdiction. It is only if the spatial model was discarded that we could bring Litvinenko within the ECHR’s reach. This could be done either by adopting a personal model of jurisdiction as control over Litvinenko himself, or by saying that the jurisdiction threshold should not apply at all to negative obligations, as would be my preference. Or, one could try distinguishing Litvinenko’s case from Bankovic on the less than principled ground that the killing took place within the ECHR’s espace juridique.158
The same would go, say, for the use of drones by the US in Pakistan or Afghanistan. Extraterritorial torture in a place outside the state’s control—for example, if a US agent was to interrogate a detainee in a Pakistani prison—would also slip through the cracks of the spatial model. So would the various complicity scenarios, as with the US or the UK providing intelligence and even the questions to the Pakistani interrogators.159 Various instances of extraterritorial law enforcement, ranging from searches and seizures to abductions, would likewise be out, at least until the individual concerned is brought into a place or an object under the state’s control,160 as would transboundary environmental harm, as in the example of Colombia’s aerial herbicide spraying having effects in Ecuador which we discussed above.161
This would be the price to pay were the spatial model, however tweaked and adjusted, to be applied exclusively. None of these individuals would be entitled to any human rights protection. And, of course, what makes this result morally intolerable is that the states in question in all of these cases would still have total control over the acts of their own agents. There would be nothing preventing them from complying with their human rights obligations, if such obligations were to exist. As I have argued above, considerations of effectiveness would need to be taken into account when applying these obligations, so that the possibly extraordinary circumstances of extraterritorial application do not produce unrealistic restrictions on states.162 But should effectiveness truly mean that a state is free to kill and torture outside its borders, and if so why?
(p.173) No good answer to this question can be given from the standpoint of universality. This is why the spatial model has in practice frequently been complemented by a personal one, to which I will now turn.
3. The Personal Model: Jurisdiction as Authority and Control over Individuals
Conceiving of state jurisdiction within the meaning of human rights treaties in personal, rather than spatial terms, would appear to solve most of the policy problems with the spatial model that we have seen above. This is indeed why the personal model of jurisdiction as state authority and control over individuals has been advocated in the literature,163 and why it has been adopted by various courts or quasi-judicial bodies in numerous cases, to which I will turn below (and many of which we have already seen). Saying, however, that a state has jurisdiction over an individual whenever it controls him poses its own set of problems.
First, it is entirely doubtful that the personal model is reconcilable with the text of at least some of the relevant treaties. The CAT, for instance, specifically refers to jurisdiction over territory, rather than individuals, and no amount of interpretation can make these words go away.
More fundamentally, even if the personal model presented a legitimate option in principle, what would actually count as state control over an individual? Just as we had to examine what counts as control over territory, so we would need to answer the same question here. However, it is hard to say whether it is possible to give any meaningful answer to that question. First, we could adopt the broadest possible definition of control, and consequently of jurisdiction over a person—a state would have such control whenever it had the ability to substantively violate an individual’s rights. Yet if we did so, the personal model of jurisdiction would essentially collapse. It would serve no useful purpose as a threshold for the application of a human rights treaty, since the treaty would apply whenever the state could actually infringe it.164 Secondly, to prevent this collapse we could try limiting the notion of personal control, for instance by saying that only physical custody over an individual could satisfy the threshold. As we will see, however, such a limitation is not possible by reference to any non-arbitrary criterion.
(p.174) In Bankovic, the European Court realized that the personal model cannot be limited, and this is precisely why it rejected it, at least in its more expansive form:
[The applicants] claim that the positive obligation under Article 1 extends to securing the Convention rights in a manner proportionate to the level of control exercised in any given extra-territorial situation. The Governments contend that this amounts to a ‘cause-and-effect’ notion of jurisdiction not contemplated by or appropriate to Article 1 of the Convention. The Court considers that the applicants’ submission is tantamount to arguing that anyone adversely affected by an act imputable to a Contracting State, wherever in the world that act may have been committed or its consequences felt, is thereby brought within the jurisdiction of that State for the purpose of Article 1 of the Convention.
The Court is inclined to agree with the Governments’ submission that the text of Article 1 does not accommodate such an approach to ‘jurisdiction’. Admittedly, the applicants accept that jurisdiction, and any consequent State Convention responsibility, would be limited in the circumstances to the commission and consequences of that particular act. However, the Court is of the view that the wording of Article 1 does not provide any support for the applicants’ suggestion that the positive obligation in Article 1 to secure ‘the rights and freedoms defined in Section I of this Convention’ can be divided and tailored in accordance with the particular circumstances of the extra-territorial act in question and, it considers its view in this respect supported by the text of Article 19 of the Convention. Indeed the applicants’ approach does not explain the application of the words ‘within their jurisdiction’ in Article 1 and it even goes so far as to render those words superfluous and devoid of any purpose. Had the drafters of the Convention wished to ensure jurisdiction as extensive as that advocated by the applicants, they could have adopted a text the same as or similar to the contemporaneous Articles 1 of the four Geneva Conventions of 1949 (see § 25 above).
Furthermore, the applicants’ notion of jurisdiction equates the determination of whether an individual falls within the jurisdiction of a Contracting State with the question of whether that person can be considered to be a victim of a violation of rights guaranteed by the Convention. These are separate and distinct admissibility conditions, each of which has to be satisfied in the afore-mentioned order, before an individual can invoke the Convention provisions against a Contracting State.165
Note that the Court here does not explicitly address the applicants’ reliance on prior case law which indicated that authority and control over individuals would lead to jurisdiction—presumably because it did not wish to overrule prior jurisprudence, thus maintaining the pretence that Bankovic presented no departure from it. The Court does, however, quite clearly reject the idea that anyone whose rights would substantively be violated by a state’s extraterritorial act would be within its jurisdiction.
That said, as we will see, the personal conception of jurisdiction has not remained absent from Strasbourg case law, either before or since Bankovic, even if the Court’s position on it is entirely, and one could say purposefully, unclear. Other human rights institutions, however, have been much less cautious. And in all fairness to the European Court, it must be said that this is not because these other bodies somehow have a better grasp of the whole problématique of extraterritorial application. Rather, it is because the stakes are higher in Strasbourg, because its (p.175) judgments are not only binding formally (which of course is not the case with UN treaty bodies), but, more importantly, have attained a high degree of authority and compliance in practice, that it is much harder for Strasbourg to throw caution to the wind. Let us now take a closer look at the case law.
B. Case law
1. UN treaty bodies
I will start off by examining the jurisprudence of UN treaty bodies. Although it is generally not as rich as that of Strasbourg, nor chronologically the first, it is precisely these bodies that have been the staunchest proponents of a personal conception of jurisdiction.
The Human Rights Committee first addressed the issue of the extraterritorial application of the ICCPR in two nearly identical cases, Lopez Burgos v. Uruguay 166 and Celiberti de Casariego v. Uruguay.167 The former concerned an individual resident in Argentina who was kidnapped in Argentina by Uruguayan security forces, held in incommunicado detention for a period of time in Argentina, and then transferred to Uruguay. The second applicant suffered essentially the same fate, except in Brazil rather than Argentina. Both of the applicants complained of abduction and unlawful detention. In Lopez Burgos, the Committee held as follows:
The Human Rights Committee further observes that although the arrest and initial detention and mistreatment of Lopez Burgos allegedly took place on foreign territory, the Committee is not barred either by virtue of article 1 of the Optional Protocol (‘…individuals subject to its jurisdiction…’) or by virtue of article 2(1) of the Covenant (‘…individual within its territory and subject to its jurisdiction…’) from considering these allegations, together with the claim of subsequent abduction into Uruguayan territory, inasmuch as these acts were perpetrated by Uruguayan agents acting on foreign soil.
The reference in article 1 of the Optional Protocol to ‘individuals subject to its jurisdiction’ does not affect the above conclusion because the reference in that article is not to the place where the violation occurred, but rather to the relationship between the individual and the State in relation to a violation of any of the rights set forth in the Covenant, wherever they occurred.
Article 2(1) of the Covenant places an obligation upon a State party to respect and to ensure rights ‘to all individuals within its territory and subject to its jurisdiction’, but does not imply that the State party concerned cannot be held accountable for violations of rights under the Covenant which its agents commit upon the territory of another State, whether with the acquiescence of the Government of that State or in opposition to it. According to article 5 (1) of the Covenant:
‘1. Nothing in the present Covenant may be interpreted as implying for any State, group or person any right to engage in any activity or perform any act aimed at the (p.176) destruction of any of the rights and freedoms recognized herein or at their limitation to a greater extent than is provided for in the present Covenant.’
In line with this, it would be unconscionable to so interpret the responsibility under article 2 of the Covenant as to permit a State party to perpetrate violations of the Covenant on the territory of another State, which violations it could not perpetrate on its own territory.168
The Committee adopted identical views in Celiberti.169 Note, first, how the Committee conceives of state jurisdiction in purely personal terms, as a reference ‘not to the place where the violation occurred, but rather to the relationship between the individual and the State in relation to a violation of any of the rights set forth in the Covenant, wherever they occurred’. Of course, the Committee somehow had to do away with the explicit mention of territory in Article 2(1) ICCPR, which it did by making an unpersuasive reference to Article 5(1), as well pointed out by Christian Tomuschat in his separate opinion:
I concur in the views expressed by the majority. None the less, the arguments set out in paragraph 12 for affirming the applicability of the Covenant also with regard to those events which have taken place outside Uruguay need to be clarified and expanded. Indeed, the first sentence in paragraph 12.3, according to which article 2(1) of the Covenant does not imply that a State party ‘cannot be held accountable for violations of rights under the Covenant which its agents commit upon the territory of another State’, is too broadly framed and might therefore give rise to misleading conclusions. In principle, the scope of application of the Covenant is not susceptible to being extended by reference to article 5, a provision designed to cover instances where formally rules under the Covenant seem to legitimize actions which substantially run counter to its purposes and general spirit. Thus, Governments may never use the limitation clauses supplementing the protected rights and freedoms to such an extent that the very substance of those rights and freedoms would be annihilated; individuals are legally barred from availing themselves of the same rights and freedoms with a view to overthrowing the regime of the rule of law which constitutes the basic philosophy of the Covenant. In the present case, however, the Covenant does not even provide the pretext for a ‘right’ to perpetrate the criminal acts which, according to the Committee's conviction, have been perpetrated by the Uruguayan authorities.
To construe the words ‘within its territory’ pursuant to their strict literal meaning as excluding any responsibility for conduct occurring beyond the national boundaries would, however, lead to utterly absurd results. The formula was intended to take care of objective difficulties which might impede the implementation of the Covenant in specific situations. Thus, a State party is normally unable to ensure the effective enjoyment of the rights under the Covenant to its citizens abroad, having at its disposal only the tools of diplomatic protection with their limited potential. Instances of occupation of foreign territory offer another example of situations which the drafters of the Covenant had in mind when they confined the obligation of States parties to their own territory. All these factual patterns have in common, however, that they provide plausible grounds for denying the protection of the Covenant. It may be concluded, therefore, that it was the intention of the drafters, whose sovereign decision cannot be challenged, to restrict the territorial scope of the Covenant in view of such situations where enforcing the Covenant would be likely to encounter exceptional obstacles. Never was it envisaged, however, to grant States parties unfettered (p.177) discretionary power to carry out wilful and deliberate attacks against the freedom and personal integrity against their citizens living abroad. Consequently, despite the wording of article 2(1), the events which took place outside Uruguay come within the purview of the Covenant.
The somewhat peculiar reference to Article 5(1) aside,170 note how both the majority and Professor Tomuschat rely on universality in making their argument in favour of extraterritorial application. First, the Committee states that the reason why the ICCPR should apply to Uruguayan actions in Argentina is because ‘it would be unconscionable to so interpret the responsibility under article 2 of the Covenant as to permit a State party to perpetrate violations of the Covenant on the territory of another State, which violations it could not perpetrate on its own territory’. The word ‘unconscionable’ is key—it represents the value judgment that the Committee is making from the standpoint of universality. And so is Tomuschat, when he says that construing ‘the words “within its territory” pursuant to their strict literal meaning as excluding any responsibility for conduct occurring beyond the national boundaries would, however, lead to utterly absurd results’. Again, it is in ‘utterly absurd results’ that we have the universality-speak. This is not an assessment made from pure logic, reason, or law, but rests on the moral and philosophical foundation of human rights—that all human beings should enjoy them by virtue of being human, and that a good reason must be given for denying them.171
Secondly, note how effectiveness emerges in Tomuschat’s arguably more far-sighted opinion. He thus notes that the ICCPR jurisdiction clause has a specific purpose of taking ‘care of objective difficulties which might impede the implementation of the Covenant in specific situations’, difficulties which make a state ‘normally unable to ensure the effective enjoyment of the rights under the Covenant to its citizens abroad’. And as no such difficulties existed here, since Uruguay was in perfect control of the conduct of its own agents in Argentina, the ICCPR should apply extraterritorially.
Note also how Tomuschat’s opinion appears questionable in two elements, at least with the benefit of almost thirty years of hindsight. Firstly, he says that one situation in which there could be ‘plausible grounds for denying the protection of the Covenant’ is military occupation—precisely the one situation which today most of us would consider as undeniably requiring extraterritorial application, per Loizidou and the ICJ’s Wall and Congo v. Uganda decisions. Secondly, Tomuschat relies somewhat problematically on the nationality of the victim when he says that it was never envisaged by the states parties that they would have a ‘discretionary power to carry out wilful and deliberate attacks against the freedom and personal integrity against their citizens living abroad’. This, I hope, was nothing more than a rhetorical point—the result would surely have been no (p.178) different had Mr Lopez Burgos been an Argentinian, rather than a Uruguayan national.172
The Human Rights Committee had few other individual cases dealing with extraterritorial application after Lopez Burgos and Celiberti.173 As we have seen, on its consideration of the reports of states parties to the ICCPR it has also affirmed the validity of the spatial model of jurisdiction.174 Its position, however, remains the same, has solidified due to events surrounding the ‘war on terror’, and is now stated in its General Comment No. 31:
States Parties are required by article 2, paragraph 1, to respect and to ensure the Covenant rights to all persons who may be within their territory and to all persons subject to their jurisdiction. This means that a State Party must respect and ensure the rights laid down in the Covenant to anyone within the power or effective control of that State Party, even if not situated within the territory of the State Party. As indicated in General Comment 15 adopted at the twenty-seventh session (1986), the enjoyment of Covenant rights is not limited to citizens of States Parties but must also be available to all individuals, regardless of nationality or statelessness, such as asylum seekers, refugees, migrant workers and other persons, who may find themselves in the territory or subject to the jurisdiction of the State Party. This principle also applies to those within the power or effective control of the forces of a State Party acting outside its territory, regardless of the circumstances in which such power or effective control was obtained, such as forces constituting a national contingent of a State Party assigned to an international peace-keeping or peace-enforcement operation.175
This paragraph is notable for several reasons. First, the Committee now openly espouses the so-called disjunctive–conjunctive interpretation of the jurisdiction clause of Article 2(1) ICCPR—first argued for by Thomas Buergenthal176—by saying that the words ‘within its territory and subject to its jurisdiction’ mean that ICCPR rights must be respected and ensured to all persons who may be within the state’s territory and to all persons subject to its jurisdiction. Secondly, the Committee again endorses a personal notion of jurisdiction as ‘power or effective control’ over an individual, regardless of the place where such control is exercised.177 Thirdly, the Committee’s final remark that even state forces in a ‘national contingent (p.179) of a State Party assigned to an international peace-keeping or peace-enforcement operation’ have ICCPR obligations may be overbroad, in that it neglects potentially very complex issues of attribution. In cases in which the conduct of such a contingent is actually not attributable to the sending state, there could also be no violation of the ICCPR by that state arising from that particular conduct.178 This issue is, of course, beyond the scope of the present study.
Other UN treaty bodies have largely followed the Human Rights Committee’s lead with regard to the personal model. We have already seen that the Committee Against Torture interprets the CAT jurisdiction clauses referring to a state’s obligations in ‘any territory under its jurisdiction’ as not only extending to ‘all areas where the State party exercises, directly or indirectly, in whole or in part, de jure or de facto effective control’, as well as places or facilities under such control, i.e. within the spatial model, but also that it has adopted a variant of the personal model:
The Committee considers that the scope of ‘territory’ under article 2 must also include situations where a State party exercises, directly or indirectly, de facto or de jure control over persons in detention.179
As a purely textual matter, this interpretation makes very little sense. In fact, it is directly contrary to the text of the treaty; to say that the concept of ‘territory’ includes control over persons is simply mind-boggling. The Committee, of course, had sound reasons of policy to adopt the personal model. While the spatial model could be adapted to cover places rather than large areas of territory, so that, for instance, a Guantanamo detainee would be protected by the CAT since he was, in fact, in an area or place under US control, even the spatial model can only be stretched so far. At the same time, it is quite possible for states to conduct activities that would normally be contrary to the CAT but to do so in places or areas outside their control—this particularly goes for various complicity scenarios. And it is here, to fill this gap, that the personal model becomes especially useful.180 Otherwise, the activities of, say, a UK agent complicit in the inhumane treatment of a person actually detained by the US or Pakistan, or even interrogating the prisoner himself, would simply fall outside the CAT’s scope. Perhaps the Committee’s purpose would have been better served by interpreting the CAT as containing an array of implicit negative obligations unlimited by the jurisdiction clauses, an alternative to which I will turn in more detail below. Be that as it may, I cannot see how the Committee’s interpretation of the word ‘territory’ to accommodate the personal model can withstand serious scrutiny.
(p.180) 2. Inter-American institutions
The Inter-American Commission of Human Rights has also championed the personal model, though the Court itself has not yet had the opportunity to rule on matters of extraterritorial application. The Commission’s mandate is not only to interpret the American Convention on Human Rights, a binding international treaty, but also the world’s first international human rights instrument, the American Declaration on the Rights and Duties of Man. The Declaration is itself not a treaty, but it imposes obligations upon American states indirectly, through their membership in the Organization of American States (OAS). Crucially, while Article 1(1) ACHR contains a jurisdiction clause almost identical to that in the ECHR, the Declaration contains no such clause, nor any other explicit limitation of its territorial scope.181
The Commission’s first case explicitly dealing with the extraterritorial application of the ACHR was Saldaño v. Argentina,182 filed against Argentina by an Argentinian national incarcerated on death row in the United States, who claimed that Argentina had failed to fulfil its positive obligation to protect him. Relying, inter alia, on the European Commission’s decision in Cyprus v. Turkey, which first espoused the personal model of jurisdiction, the Inter-American Commission considered that ‘a state party to the American Convention may be responsible under certain circumstances for the acts and omissions of its agents which produce effects or are undertaken outside that state’s own territory’.183 However, because the conduct complained of was essentially that of the United States, not Argentina, and because ‘the petitioner has not adduced any proof whatsoever that tends to establish that the Argentine State has in any way exercised its authority or control either over the person of Mr Saldaño, prior or subsequent to his arrest in the United States, or over the local officials in the United States involved in the criminal proceeding taken against him’, the petitioner was not within Argentinian jurisdiction.184 The mere bond of Argentinian nationality did not suffice.185
In Saldaño the Commission thus clearly adopted the personal model with regard to the ACHR. It did the same several months later with regard to the American Declaration, in Coard v. United States, a case dealing with the acts of US military forces in Grenada:
While the extraterritorial application of the American Declaration has not been placed at issue by the parties, the Commission finds it pertinent to note that, under certain circumstances, the exercise of its jurisdiction over acts with an extraterritorial locus will not only be consistent with but required by the norms which pertain. The fundamental rights of the individual are proclaimed in the Americas on the basis of the principles of equality and non-discrimination—‘without distinction as to race, nationality, creed or sex.’ Given that (p.181) individual rights inhere simply by virtue of a person's humanity, each American State is obliged to uphold the protected rights of any person subject to its jurisdiction. While this most commonly refers to persons within a state's territory, it may, under given circumstances, refer to conduct with an extraterritorial locus where the person concerned is present in the territory of one state, but subject to the control of another state—usually through the acts of the latter’s agents abroad. In principle, the inquiry turns not on the presumed victim's nationality or presence within a particular geographic area, but on whether, under the specific circumstances, the State observed the rights of a person subject to its authority and control.186
Note how the Commission read into the American Declaration a state jurisdiction requirement that it does not actually contain, only to then adopt the personal model of jurisdiction as authority and control over individuals. The reason it gave for doing so was that ‘individual rights inhere simply by virtue of a person's humanity’, i.e. universality. The Commission has maintained this approach ever since,187 for example in a number of (practically rather fruitless) cases against the United States dealing with Guantanamo.188
3. Early Strasbourg case law
Though in recent years it has been the UN treaty bodies and the Inter-American Commission that have most ardently championed the personal model of jurisdiction, it was in fact first devised by the European Commission of Human Rights. As we have seen, in its first decision on northern Cyprus,189 the Commission opined that
[i]n Art. 1 of the Convention, the High Contracting Parties undertake to secure the rights and freedoms defined in Section 1 to everyone ‘within their jurisdiction’ (in the French text: ‘relevant de leur juridiction’). The Commission finds that this term is not, as submitted by the respondent Government, equivalent to or limited to the national territory of the High Contracting Party concerned. It is clear from the language, in particular of the French text, and the object of this Article, and from the purpose of the Convention as a whole, that the High Contracting Parties are bound to secure the said rights and freedoms to all persons under their actual authority and responsibility, whether that authority is exercised within their own territory or abroad. The Commission refers in this respect to its decision on the admissibility of Application No. 1611/62—X. v Federal Republic of Germany—Yearbook of the European Convention on Human Rights, Vol. 8, pp. 158–169 (at pp. 168–169).
(p.182) The Commission further observes that nationals of a State, including registered ships and aircrafts, are partly within its jurisdiction wherever they may be, and that authorised agents of a State, including diplomatic and consular agents and armed forces, not only remain under its jurisdiction when abroad but bring any other persons or property ‘within the jurisdiction’ of that State, to the extent that they exercise authority over such persons or property. Insofar as, by their acts or omissions, they affect such persons or property, the responsibility of the State is engaged.190
The Commission has affirmed the personal model time and again, and did not limit its scope either to specific places or objects, such as embassies, ships, or aircraft, or to specific agents, such as diplomats, but applied it to all state agents.191 Again, all of the cases that the European Court in Bankovic tried to rationalize (or even ignore) as supposedly being exceptional, tied to the ‘special’ nature of either the agents or the premises on which they acted, were decided on the basis of a general conception of jurisdiction as state authority and control over individuals. The Commission’s approach to Article 1 jurisdiction was, in fact, limitless—whenever there was in substance a potential violation of the ECHR, the Commission found that state jurisdiction existed.192
All of this changed, of course, with Bankovic, where the Court explicitly rejected ‘a “cause-and-effect” notion of jurisdiction’,193 and gave short shrift to the contrary practice of other human rights institutions, on which the applicants relied:
Fourthly, the Court does not find it necessary to pronounce on the specific meaning to be attributed in various contexts to the allegedly similar jurisdiction provisions in the international instruments to which the applicants refer because it is not convinced by the applicants’ specific submissions in these respects (see § 48 above). It notes that Article 2 of the American Declaration on the Rights and Duties of Man 1948 referred to in the above-cited Coard Report of the Inter-American Commission of Human Rights (§ 23 above), contains no explicit limitation of jurisdiction. In addition, and as to Article 2 § 1 the CCPR 1966 (§ 26 above), as early as 1950 the drafters had definitively and specifically confined its territorial scope and it is difficult to suggest that exceptional recognition by the (p.183) Human Rights Committee of certain instances of extra-territorial jurisdiction (and the applicants give one example only) displaces in any way the territorial jurisdiction expressly conferred by that Article of the CCPR 1966 or explains the precise meaning of ‘jurisdiction’ in Article 1 of its Optional Protocol 1966 (§ 27 above). While the text of Article 1 of the American Convention on Human Rights 1978 (§ 24 above) contains a jurisdiction condition similar to Article 1 of the European Convention, no relevant case-law on the former provision was cited before this Court by the applicants.194
As we have seen, in recent years the adoption of the personal model has grown in the jurisprudence of other human rights bodies, and the European Court could not today dismiss it as easily. Be that as it may, having adopted a strict spatial, territorial model of jurisdiction in Bankovic, the Court soon found itself faced with cases where the application of this model would have led to unacceptable results.
4. Post-Bankovic Strasbourg case law
The first post-Bankovic case to directly contradict it was Issa, where the Chamber endorsed the personal model of jurisdiction in addition to the spatial one:
Moreover, a State may also be held accountable for violation of the Convention rights and freedoms of persons who are in the territory of another State but who are found to be under the former State's authority and control through its agents operating—whether lawfully or unlawfully—in the latter State (see, mutatis mutandis, M. v. Denmark, application no. 17392/90, Commission decision of 14 October 1992, DR 73, p. 193; Illich Sanchez Ramirez v. France, application no. 28780/95, Commission decision of 24 June 1996, DR 86, p. 155; Coard et al. v. the United States, the Inter-American Commission of Human Rights decision of 29 September 1999, Report No. 109/99, case No. 10.951, §§ 37, 39, 41 and 43; and the views adopted by the Human Rights Committee on 29 July 1981 in the cases of Lopez Burgos v. Uruguay and Celiberti de Casariego v. Uruguay, nos. 52/1979 and 56/1979, at §§ 12.3 and 10.3 respectively). Accountability in such situations stems from the fact that Article 1 of the Convention cannot be interpreted so as to allow a State party to perpetrate violations of the Convention on the territory of another State, which it could not perpetrate on its own territory (ibid.).195
Of interest is not only that the Chamber considered ‘jurisdiction’ to mean state authority and control over individuals, but how and why the Chamber reached this conclusion. As for the how, note that the Chamber relies precisely on all those cases so brusquely rejected by the Grand Chamber in Bankovic: the old European Commission’s case law; the Coard decision of the Inter-American Commission; and the Lopez Burgos and Celiberti cases of the Human Rights Committee. As for the why, the Chamber made a direct appeal to universality, and did so in the exact language of Lopez Burgos and Celiberti—‘the Convention cannot be interpreted so as to allow a State party to perpetrate violations of the Convention on the territory of another State, which it could not perpetrate on its own territory’.
(p.184) It is also worth mentioning that three of the judges sitting in the Chamber that decided Issa in 2004—Judges Costa, Baka, and Thomassen—also sat on the Grand Chamber that decided Bankovic in 2001. While the Grand Chamber was unanimous in Bankovic, the Court’s (quite unfortunate) practice is to disallow separate opinions of judges in admissibility decisions, which only state whether they were decided unanimously or by a majority.196 Thus, though all of the judges in Bankovic were in agreement that the case should be dismissed, there could have been serious disagreement on the bench on the exact reasoning for doing so—and that disagreement might have re-emerged in Issa three years later. This, of course, is purely speculative. What is not speculative, however, is that no amount of (honest) distinguishing can reconcile Issa with Bankovic, as the former relied on a personal model of jurisdiction which the latter discarded.
Now let us take a look at how the Issa approach fared in subsequent Strasbourg jurisprudence. Its first mention was in that ill-fated case, filed (without any hint of self-irony) by that great friend of human rights Saddam Hussein against several European states, claiming that his continued detention in Iraq was unlawful. A Chamber (Fourth Section) declared his complaint inadmissible since he was being held by US troops, and was therefore outside the respondent states’ jurisdiction.197 Although the Chamber cited Issa, it did not definitively embrace its approach that control over an individual may establish state jurisdiction, though it did allow for the possibility that state involvement in detention could be a basis for jurisdiction.
In several subsequent cases the Issa approach was cited with explicit approval. In Pad and Others v. Turkey,198 the applicants were Iranian nationals, living close to the Turkish border. They were killed by a Turkish helicopter, as Turkey claimed when they attempted to cross the border illegally. Though it was undisputed that they were killed by Turkish forces, it was disputed whether the killing took place on the Turkish or the Iranian side of the border. Relying on Issa, the Chamber (Third Section) held that
…a State may be held accountable for violations of the Convention rights and freedoms of persons who are in the territory of another State which does not necessarily fall within the legal space of the Contracting States, but who are found to be under the former State's authority and control through its agents operating—whether lawfully or unlawfully—in the latter State.
However, in the instant case, it was not disputed by the parties that the victims of the alleged events came within the jurisdiction of Turkey. While the applicants attached great importance to the prior establishment of the exercise by Turkey of extraterritorial jurisdiction with a view to proving their allegations on the merits, the Court considers that it is not required to determine the exact location of the impugned events, given that the Government (p.185) had already admitted that the fire discharged from the helicopters had caused the killing of the applicants’ relatives, who had been suspected of being terrorists.199
In other words, the Chamber in Pad seems to have accepted that because the applicants were killed by Turkish agents, they were brought within Turkish jurisdiction, regardless of whether the killing took place in Turkey or in Iran. This is, of course, again in direct contradiction to Bankovic—even the killing itself took place by missile fire from an aircraft! In Isaak and others v. Turkey 200 the same Chamber declared admissible the complaint by a family of an individual beaten to death by TRNC police in northern Cyprus in a UN buffer zone, and therefore not in an area over which Turkey had effective overall control under Loizidou. The Chamber stated, inter alia, that
[i]n view of the above, even if the acts complained of took place in the neutral UN buffer zone, the Court considers that the deceased was under the authority and/or effective control of the respondent State through its agents (see Issa and Others, cited above). It concludes, accordingly, that the matters complained of in the present application fall within the ‘jurisdiction’ of Turkey within the meaning of Article 1 of the Convention and therefore entail the respondent State’s responsibility under the Convention.
Two similar cases, Solomou and Others v. Turkey 201 and Andreou v. Turkey,202 were decided by a Chamber (Fourth Section)—the same Chamber which later decided Al-Saadoon, though with a somewhat different composition. The former case simply repeated Isaak, while in the latter case, the Chamber declared admissible a complaint by an individual who was shot by Turkish or TRNC soldiers while in the UN buffer zone:
The Government argued that there was no indication that at the time when she was struck by the bullet the applicant had been in the territory of Turkey or of the ‘TRNC’. It was contended that Turkey had no ‘jurisdiction’ and/or control over the UN-controlled buffer zone or the Greek-Cypriot National Guard ceasefire line where the act complained of by the applicant had occurred.
The Court reiterates that, in exceptional circumstances, the acts of Contracting States which produce effects outside their territory and over which they exercise no control or authority may amount to the exercise by them of jurisdiction within the meaning of Article 1 of the Convention. The Court notes that, according to UNFICYP’s press release and the UN Secretary-General’s report on the events of 14 August 1996, the applicant’s injuries were caused by Turkish and/or Turkish Cypriot uniformed personnel, who fired some 25 to 30 rounds into the crowd. These agents of the State were at the time of opening fire in the territory of the ‘TRNC’. The Court further notes that, when she was hit by the bullet, the applicant was standing outside the neutral UN buffer zone and in close vicinity to the Greek-Cypriot National Guard checkpoint. Unlike the applicants in the Bankovic and Others case (cited above) she was accordingly within territory covered by the Convention.
(p.186) In these circumstances, even though the applicant sustained her injuries in territory over which Turkey exercised no control, the opening of fire on the crowd from close range, which was the direct and immediate cause of those injuries, was such that the applicant must be regarded as ‘within [the] jurisdiction’ of Turkey within the meaning of Article 1 and that the responsibility of the respondent State under the Convention is in consequence engaged.203
It is notable that in this particular case, though the Chamber approvingly cited Issa, it seemed to have found that Turkey had jurisdiction either because the bullet was fired at the victim from an area controlled by Turkey, or because when the applicant was hit by the bullet she was located in an area controlled by Cyprus—a situation in the Chamber’s view distinguishable from that in Bankovic possibly because it was in the ECHR’s espace juridique. It is unclear, at least to me, why the Chamber in Andreou did not apply the Issa control over individuals’ basis of jurisdiction, as in Isaak and Solomou.
In Ben El Mahi and Others v. Denmark,204 a Chamber (Fifth Section) declared inadmissible an application lodged against Denmark by Moroccan residents who alleged that Denmark had discriminated against them and violated their freedom of religion by permitting the publication of cartoons disrespectful towards the prophet Mohammed in a private Danish newspaper, which provoked an uproar in the Muslim world. The Chamber found that the applicants were not within Danish jurisdiction. It did, however, cite Issa and did accept in principle that jurisdiction may arise from control over individuals.
In sum, three of the Court’s five sections cited Issa with approval for the proposition that Article 1 ECHR jurisdiction may arise from state control over an individual, rather than territory. This is both contrary to Bankovic and directly undermines it. In none of these cases, of course, did the Chamber in question openly acknowledge the conflict with the Bankovic Grand Chamber, choosing rather to pretend that the Court’s case law is perfectly consistent and coherent.
In that regard, Issa and its progeny notwithstanding, any rumours about Bankovic’s imminent demise would be greatly exaggerated. Yes, it has been chipped away at bit by bit, but this happened mostly in politically relatively unimportant, low-profile cases. It is when the judges in these cases realized that the spatial model failed them—that, for example, they could not establish the effective overall control of Turkey over the UN buffer zone in Cyprus—and that this would lead to manifestly unjust or arbitrary results, that they decided to instead rely on the personal model of jurisdiction. The Grand Chamber has, however, affirmed Bankovic as recently as in its 2010 Medvedyev judgment discussed above:
In keeping with the essentially territorial notion of jurisdiction, the Court has accepted only in exceptional cases that acts of the Contracting States performed, or producing effects, outside their territories can constitute an exercise of jurisdiction by them for the purposes of Article 1 of the Convention (see Banković, cited above, § 67, and Ilaşcu and Others v. Moldova and Russia [GC], no. 48787/99, § 314, ECHR 2004-VII). In its first Loizidou (p.187) judgment (preliminary objections), for example, the Court found that bearing in mind the object and purpose of the Convention, the responsibility of a Contracting Party might also arise when as a consequence of military action—whether lawful or unlawful—it exercised effective control of an area outside its national territory (see Loizidou v. Turkey (preliminary objections) [GC], 23 March 1995, § 62, Series A no. 310). This excluded situations, however, where—as in the Banković case—what was at issue was an instantaneous extraterritorial act, as the provisions of Article 1 did not admit of a ‘cause-and-effect’ notion of ‘jurisdiction’ (Banković, § 75).205
The Court not only reaffirmed the Bankovic rejection of a ‘cause-and-effect’ notion of jurisdiction, but added a new gloss to Bankovic—that it dealt with ‘an instantaneous extraterritorial act’, i.e. a killing. The Court may thus be opening the doors to the application of the personal model to, well, non-instantaneous acts, however defined, most notably detention. This, of course, brings us to the single most crucial problem of the personal model—what does, in fact, amount to state ‘authority and control’ over an individual? In other words, can the personal model be limited so as to serve a useful purpose as a threshold for the application of human rights treaties?
In my view, the answer to these questions is a resounding ‘no’—but this does not mean that attempts have not been made. Let us look at some of them.
C. What amounts to authority and control? Can the personal model be limited?
1. Physical custody
One possible way of attempting to limit the personal conception of jurisdiction is to say that it applies only when a state has physical custody of an individual, and not to any other form of control. Deprivation of liberty is by definition a ‘non-instantaneous act’, to paraphrase Medvedyev, and many European cases applying the personal model, such as Öcalan and Sanchez Ramirez, dealt precisely with detention. By so bounding the personal model, ‘instantaneous acts’ such as an extraterritorial deprivation of life would fall outside the state’s jurisdiction, unless they were committed in the context of detention. Bankovic is, of course, undoubtedly an authority at least for the former proposition.
Though, of course, the European Court could always limit by fiat the personal conception of jurisdiction to physical custody, the real question is whether this would be justified by any consideration of principle. I submit that it would not, and there is no better case to illustrate this, as well as all the perils of the personal model, than Al-Skeini.
Recall that all of the English courts dealing with the case found that the five applicants killed by UK troops on patrol were not within the UK’s jurisdiction, while Mr Baha Mousa, the sixth applicant, who was killed in British custody, was within the UK’s jurisdiction. That outcome can, in my view, only be explained by (p.188) the fact that the jurisdiction issue in that case was reverse-engineered, as it were.206 From the standpoint of the English courts, while the killing of five Iraqis by UK troops on patrol may have been regrettable, it was perhaps perfectly justified in some cases, and could be rationalized as an accident of war in others. However, this could not have been done with the killing of Baha Mousa, the sixth applicant. Nothing could possibly have justified the beating to death of a defenceless prisoner in UK custody, other than an unpalatable, overt adoption of double standards in the name of either relativism or realpolitik, or both. Baha Mousa therefore had to be protected by the ECHR—but what of the other five applicants?
One option, and the only option that would in my view truly have been principled, would have been to engage in a substantive Article 2 ECHR analysis regarding the lawfulness of the deprivation of life of all six applicants. While such an analysis would have assuredly resulted in a violation with regard to Baha Mousa, the result need not have been the same in respect of the applicants killed on patrol, depending mainly on each particular set of facts, and on the degree to which the courts would have been willing to attenuate their Article 2 analysis to better accommodate the extraordinary circumstances of an occupied southern Iraq in the throes of an insurgency.
But this, of course, was not the option favoured by the English courts. It was simply too fraught with peril, political as well as legal. As it seemed to them impossible to reconcile universality with effectiveness, all of the English courts opted to dismiss the cases of the five applicants killed on patrol on the preliminary grounds of lack of UK jurisdiction, and consequent lack of extraterritorial application of the ECHR, and hence had to come up with a way of distinguishing their cases from that of Baha Mousa. The High Court and the House of Lords chose to do so by tweaking the spatial model so that it would cover UK military prisons abroad, by making the extremely dubious analogy to embassies. The Court of Appeal per Lord Justice Brooke thought, however, that such an analogy made little sense, and quite rightly so. It therefore opted for a personal model of jurisdiction as control over individuals,207 but it first had to consider Bankovic, which seemed to have squarely rejected it:
I would therefore be more cautious than the Divisional Court in my approach to the Bankovic judgment. It seems to me that it left open both the ECA [effective control of an area, i.e. the spatial model] and SAA [state agent authority, i.e. the personal model] approaches to extra-territorial jurisdiction, while at the same time emphasizing (in para 60) that because a SAA approach might constitute a violation of another state's sovereignty (for example, when someone is kidnapped by the agents of a state on the territory of another state without that state's invitation or consent), this route to any recognition that extra-territorial jurisdiction has been exercised within the meaning of an international treaty should be approached with caution.
(p.189) Viewed in this light, the Court's conclusion in Bankovic was fairly inevitable. It was not a case in which it could possibly be said that the NATO forces were in effective control of the relevant area when they bombed the television station in Belgrade. What was in issue on the applicants’ arguments was whether, because the bombing was NATO's direct act, any of the states involved in the bombing could be said to have exercised Article 1 jurisdiction over any of the citizens of the FRY who were affected by it, and the court ruled out this possibility in para 71 of its judgment. I do not read the judgment as excluding the possibility that if a person is in the custody or control of the agents of a member state (whether they be military personnel or police officers) that state may not be fixed with having exercised jurisdiction over that person ratione personae, particularly if, like the British army in Iraq, those agents form part of an occupying force in the eyes of international law. This question did not arise for decision in Bankovic. On the other hand, I consider that the attempt by the claimants to rely on some of the conclusions reached by the Commission on SAA grounds in its 1976 Cyprus v Turkey report (see para 58 above) as authority for the proposition that the citizens of Basrah City came within the jurisdiction of the UK simply because they were affected by the activities of the street patrols, without more, as being authoritatively refuted by Bankovic.208
The Court thus considered Bankovic not to exclude the personal model in principle, but to do so with regard to a deprivation of life without physical custody. It then analysed the various cases applying the personal model that I have examined above, including Öcalan, Freda, and Sanchez Ramirez,209 and quite correctly concluded that
[t]hese cases have nothing to do with the principle of public international law relating to activities within aircraft registered with a state when the aircraft is airborne. They reflect examples of the SAA doctrine applying when someone is within the control and authority of agents of a contracting state, notwithstanding that he comes within that control and authority outside the espace juridique of the Council of Europe, and apparently whether or not the host state has consented to this exercise of control and authority on its soil.210
For Lord Justice Brooke, the validity of the personal model was therefore clear, and in his view it covered Baha Mousa:
Throughout the case law to which I have referred in paras 98–106 above there has been the constant refrain that a state may be fixed with having exercised extra-territorial jurisdiction if it has exercised control and authority over a complainant. The court's analysis will then be centred on the particular complaint that is made. If the complaint concerns a breach of Article 2 or Article 3 rights, the court will not only consider whether those rights have been violated but also whether the state was in breach of the positive obligations imposed on it in connection with the duty to secure those rights. The Secretary of State now concedes that the UK had jurisdiction in relation to Baha Mousa throughout the period that led up to his death, because he was being held in a British military prison that was operating in Iraq with the consent of the Iraqi sovereign authorities and contained arrested suspects (see DC 287). In my judgment, Mr Mousa came within the control and authority of the UK from the time he was arrested at the hotel and thereby lost his freedom at the hands of British troops.211
(p.190) Thus, from the moment that Baha Mousa was in British custody, he was within the UK’s jurisdiction under the personal model. As for the other five applicants:
None of them were under the control and authority of British troops at the time when they were killed. This is one of the main points that was decided in Bankovic (see para 81 above). The only case which might give rise to an argument to the contrary is that of Muhammed Salim (on the basis that British troops assumed control of the house when they burst in), but it would in my judgment be thoroughly undesirable for questions about the applicability of the ECHR to turn for their resolution on sophisticated arguments of this kind. The soldiers, for instance, might have found they were by no means in control of the house if they had all been shot dead by hostile gunfire after they had broken in. It is essential, in my judgment, to set rules which are readily intelligible. If troops deliberately and effectively restrict someone's liberty he is under their control. This did not happen in any of these five cases.212
This paragraph is absolutely fascinating, for two reasons. First, the Court of Appeal found that Baha Mousa was under the control and authority of British troops, and thus under UK jurisdiction, but that the five other applicants were not. In other words, it thought that a ‘mere’ killing did not amount to control over an individual for Article 1 ECHR purposes, while the arrest of an individual did. Secondly, the Court briefly considered applying the spatial model as control by British troops over the house where one of the applicants was killed, but decided that this would be stretching it too far—it would be ‘thoroughly undesirable for questions about the applicability of the ECHR to turn for their resolution on sophisticated arguments of this kind’. This is exactly right, and this collapse of the spatial model is exactly why not just the Court of Appeal in Al-Skeini, but other courts as well, have thought it necessary to apply the personal model of jurisdiction.
Now, as for the distinction between physical custody, which does amount to the control over an individual, and killing, which does not, note that the only reason the Court gives for this distinction is that Bankovic said so. While Bankovic did not even openly consider the personal model, the Court was right in interpreting it as precluding a result whereby the power to kill alone could constitute jurisdiction for Article 1 ECHR purposes.
Bankovic aside, of course, the question is whether this distinction makes any sense—and it simply does not. Why, after all, should the state’s power to kill an individual not be considered as an exercise of the state’s control over that individual?213 Is not killing, in fact, the ultimate form of control?214 It is certainly true that if a state has an individual in custody, it can do more to him than just kill him, by for example subjecting him to torture or inhuman treatment. But should this matter? Aside from killing, one can easily imagine a situation where, for instance, (p.191) one state has custody, but another does the torture, or is at least complicit in it—as with the alleged involvement of the US and UK intelligence services in interrogations conducted in Syria, Egypt, or Pakistan. If, say, a CIA agent was to prod a detainee held in Pakistani custody with a hot poker, why should Pakistan be the only state to blame for the torture?
Indeed, if we take a look at some of the case law, we can see that the personal model has been applied even in cases which did not involve physical custody, as, for example, in Pad, Isaak, and Solomou.215 All of them involved a killing that took place outside detention. After all, the whole universality rationale of Issa was that ‘the Convention cannot be interpreted so as to allow a State party to perpetrate violations of the Convention on the territory of another State, which it could not perpetrate on its own territory’.216 From the standpoint of universality, it makes no sense to say that a state acting extraterritorially may not detain arbitrarily, but that it may, in fact, kill arbitrarily. If anything, such a position would precisely create an incentive for states to kill rather than capture,217 as can indeed be seen from the escalation of drone attacks against suspected terrorists in Pakistan and Afghanistan, and a massive decrease in the number of those who are actually detained.218
In short, saying that physical custody is the only form of state authority and control over an individual that is acceptable for the personal model of jurisdiction is precisely the kind of ‘sophisticated legal argument’ that the Court of Appeal rightly decided to disregard in respect of the spatial model. It is, indeed, sophistry rather than sophistication, merely a device to exclude from the ambit of judicial review those cases which present extreme political or policy difficulties, and yet are morally much more blurry than is the killing of a defenceless prisoner—as with the unfortunate, but perhaps justifiable killing of civilians on patrol, or as collateral damage in airstrikes, or what have you. It is to avoid dealing with such cases that the personal model is artificially limited, but such unprincipled limitations by their very nature cannot persist indefinitely.
As of the time of writing, Al-Skeini is pending before the Grand Chamber of the European Court. Compared to the House of Lords’ judgment, the Court could do better or worse, and much depends on just how many issues the Court will reach.
Thus, from the standpoint of the spatial model, the Court could do the following:
(1) Affirm that the effective overall control of an area conception of jurisdiction applies outside the ECHR’s espace juridique, as the Chamber had done in Issa, or, conversely, agree with the House of Lords that the spatial model (p.192) extends only to the territories of the ECHR states parties—certainly the worst possible result;219
(2) Assuming that the spatial model does apply, the question then would be whether Basra was under the UK’s effective overall control. The Court could either say that the UK’s status as an occupying power necessarily meant that it was in control of Basra, despite all the difficulties that it had encountered, and that hence all six applicants were within the UK’s jurisdiction, or it could agree with the English courts that Basra was not under the UK’s effective control;220
(3) A finding that Basra, was, in fact, an area under the UK’s jurisdiction would dispose entirely of the preliminary question of extraterritorial application. If, however, the Court were to find that Basra was not under the UK’s control, it then may decide to apply the attenuated version of the spatial model as control over places in order to bring Baha Mousa under the UK’s jurisdiction upon his detention, as in Al-Saadoon.221 Doing so, however, would leave outside the ECHR’s scope any events that took place between Baha Mousa’s arrest and his ultimate transfer to the UK detention facility.222
The possibilities of the spatial model would thus be exhausted, and the Court would then have to examine the case from the standpoint of the personal model:
(1) It could reject the validity of the personal model altogether, or narrow its application to some arbitrarily selected exceptional circumstances, and thus exclude all of the applicants from the UK’s jurisdiction, unless it opted to fiddle with the spatial model to protect Baha Mousa, as above;
(2) In a similar vein, it could adopt the approach of the Court of Appeal, saying that the personal model requires physical custody, and that accordingly only Baha Mousa was within the UK’s jurisdiction;
(3) Finally, it could radically depart from Bankovic, and say that the five applicants killed by UK troops on patrol were also under the UK’s control, and accordingly within its jurisdiction. This would be a welcome, but not very likely development.
(p.193) The Court may of course very well surprise us. Whatever it does will inevitably be a consequence of how it perceives the tension between universality and effectiveness on the facts of the case, and on their broader policy implications, especially with regard to the personal model.
2. Control over an individual in a specific place, or by specific agents
Another possible way of limiting the personal model would be to mix it with the spatial one so that it would apply only in a certain kind of place, or to limit it only to a certain kind of state agent. For example, one could read cases like Öcalan as implying that custody over the individual was insufficient of itself to create jurisdiction, but that it had to take place in a context where general international law somehow recognizes that states may exceptionally exercise their jurisdiction extraterritorially, e.g. aboard a registered aircraft or in an embassy. This is, in essence, a variation of the special approach to diplomatic agents or premises, or to ships and aircraft, that we have already examined above, and is no more persuasive. None of the relevant cases actually say that the nature of the place in which the control over an individual is exercised matters, nor is it obvious that it should matter. Nor would this, again, be warranted by any consideration of general international law. As I have argued above, it is perfectly irrelevant for the question of the extraterritorial application of human rights treaties that customary law recognizes that the state may regulate conduct aboard registered ships and airplanes, or that diplomatic and consular agents and premises have a distinct status protected by jurisdictional immunities.223
3. Nationality and membership in the armed forces
Yet another potential avenue of limiting the personal model would be to do so on the basis of the victim’s nationality or some other status, most notably membership in the state’s armed forces. So far we have generally considered examples where a state agent affects the rights of a national of another state when acting extraterritorially—for instance, when UK soldiers killed the Iraqi applicants in Al-Skeini. But what of the scenario where one UK soldier kills another, for instance in a friendly fire incident, or in a plain, old-fashioned murder? Would the UK have an Article 2 ECHR obligation to investigate?
Such a scenario is obviously not hypothetical. UK courts first encountered it in Gentle, in which the applicants were the mothers of two UK soldiers who both died while serving in Iraq.224 One was killed by a roadside bomb, the other in a friendly fire incident. Inquests were held regarding both deaths, and the relevant facts were established beyond any doubt. However, the applicants contended that Article 2 ECHR required the UK not only to fully investigate the circumstances of the deaths themselves, but also to establish an inquiry into the 2003 Iraq war as a (p.194) whole, particularly into the process and the legal advice given before the troops were committed to an armed conflict. In effect, they were trying to inject jus ad bellum-type arguments into the framework of Article 2. The House of Lords found unanimously that such a far-reaching inquiry fell outside the scope of state positive obligations imposed by Article 2. It therefore avoided ruling on whether the two soldiers were protected by the ECHR in the first place, i.e. whether it applied to them extraterritorially. That issue was, however, briefly addressed by two of the law lords. Lord Bingham thus noted that
[s]ubject to limited exceptions and specific extensions, the application of the Convention is territorial: the rights and freedoms are ordinarily to be secured to those within the borders of the state and not outside. Here, the deaths of Fusilier Gentle and Trooper Clarke occurred in Iraq and although they were subject to the authority of the respondents they were clearly not within the jurisdiction of the UK as that expression in the Convention has been interpreted…But I think there is a more fundamental objection: that the appellants’ argument, necessary to meet the objection of extra-territoriality, highlights the remoteness of their complaints from the true purview of article 2.225
He thus opted for the spatial model, and considered that the two soldiers were clearly not covered by it. Baroness Hale, on the other hand, stated the following:
Nor have I much difficulty with the proposition that these soldiers were within the jurisdiction of the United Kingdom when they met their deaths. If Mr Baha Mousa, detained in a military detention facility in Basra, was within the jurisdiction, then a soldier serving under the command and control of his superiors must also be within the jurisdiction: see R (Al-Skeini) v Secretary of State for Defence  UKHL 26;  3 WLR 33. The United Kingdom is in a better position to secure to him all his Convention rights, modified as their content is by the exigencies of military service, than it is to secure those rights to its detainees.226
Thus, for Baroness Hale, UK soldiers were within the UK’s jurisdiction simply because they were under the command and control of their superiors—precisely the variation of the personal model that I am interested in here.
The question of extraterritorial application could not have been avoided in Smith, which also concerned a UK soldier who died in Iraq. An inquest was held, which established that he died from heatstroke. The soldier’s mother, however, contended that her son’s death warranted an Article 2 ECHR inquiry, which would not be limited only to the immediate cause of his death, but would also examine any possible systemic failures by the UK government—such as, for instance, not providing the soldiers with appropriate equipment or facilities—that could have led to her son’s death. The question of extraterritorial application thus inevitably arose: was Private Smith protected by the ECHR while in Iraq, or not?
Now, crucially, Private Smith actually died on a UK military base. Per the UK government’s concession in Al-Skeini, the House of Lords’ quite dubious analogy between a military prison or base and an embassy, and the European Court’s (p.195) admissibility decision in Al-Saadoon, that fact alone would have brought Private Smith within the UK’s jurisdiction. In other words, under the spatial model as applicable to places Private Smith would have been within the UK’s jurisdiction, and therefore entitled to protection under Article 2 ECHR. However, issue was raised in the lower courts as to whether he would have been within the UK’s jurisdiction even if he did not die on the base, or in any area under the UK’s effective control, but essentially under the same circumstances.
In the High Court, Mr Justice Collins thought that the answer to this question could only be in the affirmative:
At the material time, the Coalition Provisional Authority in Iraq (following the ouster of the government of Saddam Hussain) had issued an order whereby the Multinational Force (of which British troops formed a part) should be ‘immune from Iraqi legal process’ and that all personnel should be ‘subject to the exclusive jurisdiction of their sending States’: see section 2 1) and 3). Thus the U.K.'s jurisdiction over its own nationals was clearly maintained. In any event, members of the armed forces remain at all times subject to the jurisdiction of the U.K. It would obviously be wholly artificial to regard a soldier sent to fight in the territory of another state as subject to the jurisdiction of that state.227
The judge accordingly applied a personal conception of jurisdiction, as applying to UK nationals and/or soldiers. Now, obviously, this position is of great practical relevance for UK military operations abroad. Many soldiers have lost their lives outside areas under UK effective control, and they would under Mr Justice Collins’ ruling all at least in principle be entitled to the protections of Article 2 ECHR. The government therefore appealed his decision, arguing that Article 1 jurisdiction is exclusively territorial, and would apply only if a soldier died on a UK military base or a UK military hospital (as was in fact the case in Smith)—in other words, in a place under UK jurisdiction. The government relied, inter alia, on Lord Bingham’s opinion in Gentle, which Mr Justice Collins considered to have been dicta.228 The Court of Appeal affirmed the lower court’s ruling in the following terms:
In our judgment, if it is permissible to answer the question posed by Lord Rodger, namely whether there was a sufficient link between Private Smith and the UK when he died, on the assumption for this purpose that he died outside the base or a hospital, in a broad and commonsense way, the answer is in our opinion plainly yes. As the judge put it at  of his addendum, there is a degree of artificiality in saying that a soldier is protected so long as he remains in the base or military hospital but that he is not protected as soon as he steps outside.
As the Commission succinctly puts it in its skeleton argument, there is no question but that members of the British armed forces are subject to UK jurisdiction wherever they are. They remain subject to UK military law without territorial limit and may be tried by court martial whether the offence is committed in England or elsewhere. They are also subject to the general criminal and civil law. Soldiers serve abroad as a result of and pursuant to the exercise of UK jurisdiction over them. Thus the legality of their presence and of their actions depends on their being subject to UK jurisdiction and complying with UK law. As a matter (p.196) of international law, no infringement of the sovereignty of the host state is involved in the UK exercising jurisdiction over its soldiers serving abroad.
It is not in dispute that the army owes soldiers a duty of care while they are in Iraq, as elsewhere. However, it does not follow from this that a soldier in Iraq is within the jurisdiction of the UK under the Convention. We stress that we are not saying that such a soldier is within the jurisdiction merely because the army may owe soldiers a duty of care. We recognise that that is a different question. However, it is accepted that a British soldier is protected by the HRA and the Convention when he is at a military base. In our judgment, it makes no sense to hold that he is not so protected when in an ambulance or in a truck or in the street or in the desert. There is no sensible reason for not holding that there is a sufficient link between the soldier as victim and the UK whether he is at a base or not. So too, if he is court-martialled for an act committed in Iraq, he should be entitled to the protection of article 6 of the Convention wherever the court martial takes place: see in this regard per Lord Brown in Al-Skeini at . In these circumstances we accept the submission made by the Commission that there would have to be compelling reasons of principle for drawing a distinction for the purposes of jurisdiction under article 1 of the Convention between the soldier while at his base and the soldier when he steps outside it, at any rate so long as he is acting as a soldier and not, in the old phrase, on a frolic of his own. In our judgment, no such compelling reasons have been advanced on behalf of the Secretary of State.229
Like the High Court, the Court of Appeal adopted a personal conception of jurisdiction, based on the victim’s status as a member of the armed forces. The Court quite rightly pointed out that it would be artificial and defy common sense to say that a UK soldier was protected by the ECHR while on a UK base, but would lose all protection once he stepped outside it.
The Supreme Court disagreed.230 By a majority of 6 to 3 (Lady Hale and Lords Mance and Kerr dissenting), the justices found that mere membership in the armed forces was insufficient to establish a jurisdictional link for the purposes of Article 1 ECHR. This result is in my view ultimately correct. I am personally less than happy, however, with the reasoning of either the majority or the minority in the process of reaching this result. Because, of necessity, they had to start from Bankovic, with all its methodological flaws, the judgments of both the majority and the minority suffer from a great deal of otherwise needless conceptual confusion. In other words, because they are national judges applying a treaty primarily supervised by an international court, it is natural that the justices followed the European Court’s approach, however flawed, and whatever their misgivings about the mess that Strasbourg has made.
Thus, in his judgment for the majority, Lord Phillips mainly emphasized considerations of effectiveness, by saying that it is for him simply inconceivable that ECHR states parties assumed upon themselves the obligation to conduct an Article 2-compliant investigation into the death of any of their soldiers abroad, nor have they behaved accordingly.231 So did Lord Collins,232 who added that there were no ‘policy grounds for extending the scope of the Convention to armed forces (p.197) abroad. On the contrary, to extend the Convention in this way would ultimately involve the courts in issues relating to the conduct of armed hostilities which are essentially non-justiciable.’ Both justices stressed that there was no ECHR case law directly supporting the proposition that membership in the armed forces could qualify as ‘authority and control’ over an individual, and form a basis for Article 1 jurisdiction.
But, Bankovic and its lamentable offspring aside, is this explanation truly satisfactory? Is the majority really correct that Private Smith was not within the UK’s jurisdiction under the personal model? Is it not true, as Lord Mance pointed out in his dissenting judgment, that it is ‘commonsense’ that UK soldiers are within the UK’s jurisdiction?233 Is it not true, as Lord Kerr thought,234 that ‘the United Kingdom brought its soldiers into Iraq; it not only asserted complete authority over them while they remained there, it explicitly excluded the exercise of authority over those soldiers by any other agency or state; and it has always been clear that soldiers remain subject to the laws of the UK during their service abroad’? And is it not true that, as noted by a commentator, ‘[t]here is a simple, intuitive point—if a soldier is always subject to UK law, wherever he may be, should he not thereby be also entitled to the protection of UK law (including the Human Rights Act)?’235
I would submit that this intuition is wrong, however commonsense it may seem at first glance. It not only assumes that the spatial model can be supplemented by a variant of the personal model based on legal links such as nationality or membership in the armed forces, but it rests on a confusion—confusion that we have seen time and again—between the various meanings that the word ‘jurisdiction’ can have in general international law. We can recall that in its first northern Cyprus decision, the European Commission made the same category error, when it remarked that
[t]he Commission further observes that nationals of a State, including registered ships and aircraft, are partly within its jurisdiction wherever they may be, and that authorised agents of a State, including diplomatic and consular agents and armed forces, not only remain under its jurisdiction when abroad but bring any other persons or property ‘within the jurisdiction’ of that State, to the extent that they exercise authority over such persons or property. Insofar as, by their acts or omissions, they affect such persons or property, the responsibility of the State is engaged.236
As I have explained above,237 it is true that a state may legislate for its nationals and members of its armed forces even when they are abroad, and in fact all states do that to a greater or a lesser extent. It is also true that a state’s soldiers are its de jure organs, and that their conduct performed in their official capacity is attributable to the state. Neither of these premises can, however, lead to the conclusion that a (p.198) state’s soldiers, or diplomats, any other agents or even nationals are, regardless of their location, always within that state’s jurisdiction within the meaning of the jurisdiction clauses of human rights treaties. This ‘jurisdiction’ is simply not the jurisdiction to prescribe in general international law. It is not about the application and extent of UK domestic law, exclusion from Iraqi law, or even some sort of mutuality of obligations.
Likewise, the Court of Appeal’s statement that ‘[a]s a matter of international law, no infringement of the sovereignty of the host state is involved in the UK exercising jurisdiction over its soldiers serving abroad’ is perfectly true. It is also perfectly beside the point. There is equally no infringement on the sovereignty of the host state if a UK soldier is prohibited by UK law or by human rights law from shooting an Iraqi national.238
If the personal model of jurisdiction is valid, its acceptance of nationality or membership in the armed forces as grounds of state jurisdiction would lead to an open embrace of double standards—and there, I may say, is where the intuitive appeal actually lies. Let us, for example, simply transplant Smith to the facts of Al-Skeini—imagine that one of the five applicant Iraqi citizens killed by UK troops on patrol was actually a dual UK/Iraqi national. Imagine further that the sixth applicant was not Baha Mousa, but a UK soldier killed by friendly fire by UK troops on patrol. The reasoning of the lower courts or the minority of the Supreme Court in Smith would lead to the result that the dual national and the UK soldier were within the UK’s jurisdiction and thus entitled to the protection of Article 2 ECHR, while the other applicants were not, even though all of them were killed in exactly the same way, in exactly the same place.
In other words, the appeal of the minority approach lies precisely in the fact that we would feel it unjust if the UK government did not owe human rights obligations to its own soldiers in Iraq,239 while our feelings may be more mixed when it comes to Iraqi civilians shot on patrol. We simply care more about ‘our boys’, than about their enemies or even their victims, and we thus feel them more deserving of protection. Starting from the assumption that Bankovic and Al-Skeini were correctly decided, as they had to, the minority’s approach is ultimately about ‘us’ versus ‘them’, if not openly so, and is antithetical to the whole idea of universality. This is not to say that nationality or membership in the armed forces is totally irrelevant for the substantive application of human rights norms. This, however, only depends on the content of a very limited set of rights, but the application of the ECHR as such cannot depend on nationality or any other personal status.
(p.199) In short, while it may be artificial to say that a UK soldier would be protected by the ECHR while on a UK base, but not off it, this is no more artificial than saying that an Iraqi national is protected by the ECHR while on a UK base or while in UK custody, but not on the streets of Basra,240 as was the result in Al-Skeini. Thus, though Lord Phillips’ judgment for the majority in Smith mainly rests on considerations of practicality or effectiveness to deny the extraterritorial applicability of the ECHR to UK soldiers abroad on account of their status, I would submit that, more fundamentally (and perhaps with a degree of irony), this is the only result consistent with universality as the normative foundation of the ECHR and all other human rights treaties—unless, of course, it is Bankovic and Al-Skeini themselves which were wrongly decided.
4. Exercise of a legal power
A more promising method of limiting the personal model of jurisdiction as authority and control over individuals may be to rely on the ‘authority’ part by requiring the exercise of some sort of purported legal power over the victim of the alleged human rights violation. One could possibly deduce this limitation from a number of cases dealing with what I will broadly term extraterritorial law enforcement, in which no issue of extraterritorial application of human rights treaties was raised, or the applicability of the relevant treaty was assumed. In my view, this limitation would also not rest on any grounds of principle, but let us first examine it in more detail.
Note, first, that as with control over territory, which the European Court in Loizidou thought could be either lawful or unlawful, so too would the personal model have to apply to both lawful and unlawful exercises of state power over an individual. Not only would any other solution open the door to abuse, but it would actually deny human rights protection precisely in those situations where they are being affected by an exorbitant exercise of state enforcement jurisdiction or power pure and simple as a matter of international law. Thus, for example, whether the state of Israel had a right in international law to abduct Adolf Eichmann from Argentina (it clearly did not) should have no bearing on whether Eichmann had any rights vis-à-vis Israel.
Indeed, if we take a look at the European Court cases that applied the personal model, we will see that they explicitly say that jurisdiction arises even when the state concerned acts unlawfully: ‘a State may also be held accountable for violation of the Convention rights and freedoms of persons who are in the territory of another State but who are found to be under the former State’s authority and control through its agents operating—whether lawfully or unlawfully—in the latter State’.241 Likewise, in Lopez Burgos the Human Rights Committee noted that
(p.200) Article 2(1) of the Covenant places an obligation upon a State party to respect and to ensure rights ‘to all individuals within its territory and subject to its jurisdiction’, but does not imply that the State party concerned cannot be held accountable for violations of rights under the Covenant which its agents commit upon the territory of another State, whether with the acquiescence of the Government of that State or in opposition to it.242
This of course makes perfect sense. As I have explained above,243 the sovereignty of the territorial state may or may not have already been violated by the extraterritorial act of another state, depending on whether valid consent was given or whether the latter state had some other authority for its actions, such as self-defence or a Security Council resolution. But the territorial state’s sovereignty is in any event immaterial for the question of whether the individuals in that state are entitled to human rights protection against a third state.
Therefore, the personal model could not be limited only to lawful exercises of state power, as between two states and as a matter of general international law. But could it, perhaps, be limited to those extraterritorial acts which could be unlawful, or not, as a matter of international law, but would still be an exercise of a legal power? In other words, could the personal model of jurisdiction apply only when the state acts extraterritorially under colour of law—say if it arrests or abducts an individual in another state on the basis of a domestic judicial warrant?244
If we examined cases dealing with such instances of extraterritorial law enforcement, or law enforcement which is intraterritorial even though the victim of the alleged violation is outside the state’s territory, we would see that human rights treaties were generally considered to be applicable. To my mind, the best such examples are judicial proceedings, whether as in absentia trials of a person located in another country, or indeed as extraterritorial trials by courts of one state sitting in another. Surely we would say in such cases that the state concerned would be bound by fair trial guarantees? When, for instance, the special Scottish criminal court convened at the neutral venue of Camp Zeist in the Netherlands to try those accused of the Lockerbie bombing, surely the defendants were entitled to Article 6 ECHR rights, even though they were not in the UK? After all, is there anything more jurisdiction-y, as it were, then putting someone on trial?
And, sure enough, if we looked, for example, at Sejdovic v. Italy,245 we would see that neither the Bankovic-aware Grand Chamber of the European Court, nor any of the parties, thought that there was an Article 1 ECHR issue with respect to an applicant, a Montenegrin national, who was tried in Italy in absentia for murder while he had absconded to Germany. Of course he was entitled to Article 6 ECHR fair (p.201) trial rights, even though he was at the time certainly not within Italy’s ‘essentially territorial’ jurisdiction. Note that in this example the possible violation of the applicant’s rights would not have been extraterritorial, as the trial itself took place in Italy, but that it is the victim who needs to be within the state’s jurisdiction at the time of the violation for the ECHR to apply—at least under either the spatial or the personal model.246
Similarly, in Martin v. UK 247 the applicant was the son a UK soldier serving in Germany who was accused of murder, and who was under UK law subject to military law as a family member residing with a member of the armed forces, and was tried by a British court-martial sitting in Germany. Again, neither the parties nor the Court raised any questions about whether he would be protected by the ECHR vis-à-vis the UK, even though the trial itself took place in Germany, and indeed the Court did find a violation of Article 6, on the grounds of a lack of independence and impartiality of the tribunal.248
To give some more relatively random examples, in Mullai and Others v. Albania,249 the applicants were seven Albanian nationals, as well as an Albanian company, who were embroiled in a dispute regarding a building permit. In particular, the applicants alleged that Albanian authorities failed to enforce a final court judgment which recognized the validity of their building permit, and this was in violation of Article 6 and Article 1 of Protocol 1 to the ECHR. Crucially for our purposes, some of the individual applicants were actually not living in Albania, but in Italy and the United States. Again, however, nobody thought that these particular applicants were not within Albania’s jurisdiction. Likewise, in Vrbica v. Croatia,250 the applicant was a Montenegrin national living in Montenegro who complained of the refusal of Croatian courts to enforce a judgment given by a Montenegrin court and subsequently recognized in Croatia, which awarded him damages against two Croatian companies for injuries he sustained in a traffic accident. Again, even though the applicant was outside Croatia at the relevant time, neither the Court nor the parties thought that there was an Article 1 issue to be considered.
(p.202) Nor are such cases limited to Strasbourg. Consider, for example, the well-known Gueye et al. v. France case before the Human Rights Committee,251 in which the applicants were more than 700 Senegalese nationals, all of them residing in Senegal, who had served in the French army prior to Senegal’s independence. They alleged that France discriminated against them on the basis of race or nationality because it had enacted a law reducing their pension benefits when compared to retired French soldiers of French nationality. The Committee duly found a violation of Article 26 ICCPR, but again neither it nor France thought that the applicants were not protected by the ICCPR, even though they resided in Senegal when the relevant law was enacted.252
Common to all of these cases is that even though the applicants were all located outside the relevant state’s territory, they were subjected by that state to some form of legal process. None of these cases could be regarded as being within the relevant state’s jurisdiction pursuant to any variant of the spatial model, however attenuated, as the victims of the human rights violations were not located in any area or place under that state’s control. Likewise, in all these cases we would feel that it would be manifestly unjust to deny the applicability of human rights treaties. It would simply be inconceivable for the ECHR not to apply to in absentia trials in which the accused has fled to another country, or, say, to the seizure of assets by the UK of a tax-evading UK national residing in Monaco. It is obvious that the Convention should apply, and this is why the issue of extraterritorial applicability of human rights treaties is practically never raised in such situations.253
But, other than our intuition, on what grounds could we justify the extraterritorial applicability of human rights treaties in such situations, where the victim is not (p.203) within the state’s jurisdiction conceived of territorially? One answer, which I prefer, would be to simply say that the state jurisdiction threshold does not apply to potential violations of states parties’ negative obligation to respect human rights. Otherwise, we would have to accept a personal conception of jurisdiction as authority and control over individuals, in these particular cases exercised as a legal authority or power by the state over the individual concerned.
Consider, in that regard, Stephens v. Malta, an absolutely fascinating case.254 The applicant was a UK national living in Spain, who was suspected by Maltese authorities of having conspired with other persons in Spain to transport drugs to Malta. A warrant for his arrest was issued by a Maltese court, and he was detained in Spain following a request for his extradition. While still awaiting extradition in Spain, the applicant retained counsel in Malta who challenged the lawfulness of the arrest warrant before Maltese courts, inter alia on the grounds that the court which issued the warrant lacked jurisdiction over a non-Maltese national accused of committing a crime outside Malta. After several judicial instances, the original arrest warrant was found to have been procedurally defective, while in Spain the applicant was released on bail. However, after a new request for extradition was filed, the applicant was rearrested, extradited to Malta, and ultimately convicted on the criminal charges against him. Before the European Court, he challenged the lawfulness of his detention in Spain pending extradition—but he did so by claiming that Malta had violated Article 5(1) ECHR.
Of course, the merits of his case aside, the issue is precisely that until his extradition the applicant did not set foot on Malta. In purely territorial terms, as a UK national residing and then detained in Spain by Spanish authorities, he could not possibly have been within Malta’s jurisdiction. While, as in other cases we have seen above, the parties did not raise the question of the ECHR’s extraterritorial application, the Court apparently realized the problem and did so proprio motu. The Court first set out some general principles:
The exercise of jurisdiction is a necessary condition for a Contracting State to be able to be held responsible for acts or omissions imputable to it which give rise to an allegation of the infringement of rights and freedoms set forth in the Convention (see Ilaşcu and Others v. Moldova and Russia, [GC], no. 48787/99, § 311, ECHR 2004 -…).
According to established case-law Article 1 of the Convention must be considered to reflect this ordinary and essentially territorial notion of jurisdiction, other bases of jurisdiction being exceptional and requiring special justification in the particular circumstances of each case. The Court refers to its case-law on the notion of “jurisdiction” and how that notion has been interpreted and applied in different contexts (see Ilaşcu and Others [GC], no. 48787/99, cited above; Banković and Others v. Belgium and 16 Other Contracting States (dec.) [GC], no. 52207/99, ECHR 2001-XII; Assanidzé v. Georgia, [GC], no. 71503/01, ECHR 2004 -…; Soering v. the United Kingdom, 7 July 1989, Series A no. 161; Cruz Varas and Others v. Sweden, 20 March 1991, Series A no. 201; Vilvarajah and Others v. the United Kingdom, 30 October 1991, Series A no. 215; Loizidou v. Turkey (preliminary objections), 23 March 1995, Series A no. 310; Loizidou v. Turkey, 18 December 1996, Reports of (p.204) Judgments and Decisions 1996-VI; Issa and Others v. Turkey, no. 31821/96, 16 November 2004; Behrami and Behrami v. France and Saramati v. France, Germany and Norway (dec.) [GC], nos. 71412/01 and 78166/01, 2 May 2007; Drozd and Janousek v. France and Spain, 26 June 1992, Series A no. 240, and Hess v. the United Kingdom, no. 6231/73, 28 May 1975, Decisions and Reports (DR) no. 2, p. 72).255
It then applied these general principles—to the extent that it actually identified any of them—to the facts of the case:
The Court considers that for the purposes of the examination of this application and in view of its conclusions in respect of the applicant’s various complaints, it suffices to consider the matter of Malta’s jurisdiction solely in respect of the Article 5 complaints.
The Court notes that the applicant was under the control and authority of the Spanish authorities in the period between his arrest and detention in Spain on 5 August 2004 and his release on bail on 22 November 2004. In so far as the alleged unlawfulness of his arrest and detention is concerned, it cannot be overlooked that the applicant’s deprivation of liberty had its sole origin in the measures taken exclusively by the Maltese authorities pursuant to the arrangements agreed on by both Malta and Spain under the European Convention on Extradition.
By setting in motion a request for the applicant’s detention pending extradition, the responsibility lay with Malta to ensure that the arrest warrant and extradition request were valid as a matter of Maltese law, both substantive and procedural. In the context of an extradition procedure, a requested State should be able to presume the validity of the legal documents issued by the requesting State and on the basis of which a deprivation of liberty is requested. It is to be noted that in the instant case the arrest warrant had been issued by a court which did not have the authority to do so, a technical irregularity which the Spanish court could not have been expected to notice when examining the request for the applicant’s arrest and detention. Accordingly, the act complained of by Mr Stephens, having been instigated by Malta on the basis of its own domestic law and followed-up by Spain in response to its treaty obligations, must be attributed to Malta notwithstanding that the act was executed in Spain.256
First, note how the Court’s reasoning in these three paragraphs has absolutely nothing to do with the general principles that it had considered before. Indeed, the Court seemed to have felt that there was jurisdiction on the facts of the case since the Convention simply had to apply257—but then it realized that whatever it did, this case simply does not fit any of its prior case law on extraterritoriality. Therefore, the Court decided to cite and throw everything in, the kitchen sink and all, from Bankovic and Behrami to Soering and Issa. Secondly, note also how the Court says that ‘it suffices to consider the matter of Malta’s jurisdiction solely in respect of the Article 5 complaints’, thus seemingly being at odds with the Grand Chamber’s regime integrity-inspired holding in Bankovic that ‘the wording of Article 1 does not provide any support for the applicants’ suggestion that the positive obligation in Article 1 to secure “the rights and freedoms defined in Section I of this (p.205) Convention” can be divided and tailored in accordance with the particular circumstances of the extra-territorial act in question’.258 Indeed, if anything, the Court applied precisely the kind of ‘cause-and-effect’ notion of jurisdiction that Bankovic so pointedly rejected—Malta’s warrant was the cause of the applicant’s arrest in Spain, and therefore the applicant was within Malta’s jurisdiction.
Thirdly, the Court pins its analysis on quite sensible grounds of policy—that Spanish courts could reasonably have relied on the validity of the Maltese arrest warrant, and could not have been expected to know of its procedural defects under Maltese law. That said, the Court arrives at the remarkable conclusion that the detention of the applicant, ‘having been instigated by Malta on the basis of its own domestic law and followed-up by Spain in response to its treaty obligations, must be attributed to Malta notwithstanding that the act was executed in Spain’.
By so deciding, the Court was presumably trying to explain why it was Malta, and not Spain, that was the proper respondent in the case, even though the applicant was at all times during his detention under Spain’s ‘control and authority’. And though one cannot say that the Court’s approach is unreasonable as a matter of policy, it is entirely at odds with the general rules on state responsibility to say that the applicant’s detention was not attributable to Spain, but to Malta. The detention was carried out by the de jure organs of Spain, who were not placed at Malta’s disposal nor were acting under Malta’s control.259 What was attributable to Malta was the defective arrest warrant issued by its court, and the request for extradition that it transmitted to Spain. Spain then may have had the legal obligation to comply with that request, but it could just as easily have refused to do so; the applicant’s subsequent detention was still most certainly attributable to it.
Again, the ultimate result that the Court reached was perfectly sensible. Had it been otherwise, Malta would have had human rights obligations towards the applicant only once his extradition was completed, thus putting the period of unlawful detention outside the scope of the ECHR. This still does not mean, however, that the Court produced a satisfactory explanation as to why the ECHR protected the applicant extraterritorially.
As I have said above, I see only two possible explanations. First, the Court could have said, contra Bankovic, that the ECHR could apply even without Malta having jurisdiction over the applicant—but that was not going to happen, for obvious reasons. Secondly, the Court could have openly adopted the personal model, and said that the applicant was within the jurisdiction of both Malta and Spain, with regard to the former because Malta exercised a legal authority over him.260
(p.206) In other words, if we operate within the personal model, that is, if we accept that ‘jurisdiction’ in human rights treaties means not only effective control over territory, but also authority and control over individuals, then we must also accept that such authority over individuals can be exercised by a state simply by putting in motion a legal process, without having physical control of the person in question.261 The bigger question, however, is whether the personal model could be limited to such legal authority or power.
In my view, it clearly cannot, for the same reason that it could not be limited to the lawful exercise of state power under general international law—doing so would open the door to abuse. Any of the possible violations that we have just seen that the state has committed under colour of law it could have done through brute, unadulterated force. Instead of putting someone on trial in absentia, or initiating extradition proceedings, it could simply have sent its agents to abduct him, as the Eichmanns, Öcalans, and Carloses of this world can attest. Instead of taking somebody’s house through a legal procedure, it could have done so simply by burning it down. And instead of France reducing the pensions of Senegalese soldiers by passing a law, it could have simply stopped the payments by fiat.
Or, consider the following example: an ECHR state party organizes parliamentary elections and allows its citizens living abroad to vote by post or by going to designated polling stations in their host state—something that states are not obliged to do in principle,262 but many states in fact do. It then decides, however, to (p.207) blatantly discriminate among its citizens living abroad by saying, for example, that only those of a certain race, ethnicity, or gender may vote, and passes a law to that effect. Would these people residing extraterritorially be protected by the ECHR and its equality guarantee? Well of course they would, either because we consider the jurisdiction threshold not to apply to a state’s violation of its negative obligations, or because we adopt the personal model of jurisdiction, and say that the state has exercised its authority over the individuals concerned, even though it had not laid a hand on them physically.
But what if the state in question discriminated in exactly the same way, yet was smart enough not to pass a law to that effect and in fact prevented the undesirables from voting while having the most pristine electoral legislation imaginable. Would the result be any different? Would the ECHR not apply just because the state has creatively managed to avoid acting under the colour of law, but accomplished the same invidious ends through other means?
Asking this question, I think, is to answer it. To the extent that the personal model of jurisdiction is both valid and necessary, an exercise of a legal power or authority by a state over an individual outside its territory would suffice to satisfy the jurisdictional threshold. Limiting the personal model to such purported exercises of legal power, however, would be entirely arbitrary, and would only serve to undermine the rule of law by creating a perverse incentive for states to act outside their own legal system if they wish to violate human rights.
D. The personal model collapses
What, then, have we learned about the personal model? It cannot be limited to physical custody. It cannot be limited on the basis of nationality or some special status of the victim, or indeed of the perpetrator. It cannot be limited only to lawful exercises of state power over individuals, nor to extraterritorial acts to which the host state consents, nor indeed to acts committed under the colour of law. It cannot, in short, be limited on the basis of any non-arbitrary criterion. ‘Authority and control over individuals’ as a basis for state jurisdiction simply boils down to the proposition that a state has obligations under human rights treaties towards all individuals whose human rights it is able to violate.263
In other words, the main feature of the personal model of jurisdiction—its ability to cover individuals who would be unprotected by the spatial model—is also its main fault. It quite literally collapses, and serves no useful purpose as a threshold for the application of human rights treaties. Unless the personal model is limited artificially on the basis of some essentially arbitrary criteria, there is no threshold.
(p.208) This, I think, was well realized by the Grand Chamber in Bankovic when it rejected a ‘cause-and-effect’ notion of jurisdiction, which ultimately is what the personal model boils down to. The same sentiment was shared by Lord Brown in Al-Skeini, when he remarked that
[a]lternatively it would stretch to breaking point the concept of jurisdiction extending extra-territorially to those subject to a state’s ‘authority and control’. It is one thing to recognise as exceptional the specific narrow categories of cases I have sought to summarise above; it would be quite another to accept that whenever a contracting state acts (militarily or otherwise) through its agents abroad, those affected by such activities fall within its article 1 jurisdiction. Such a contention would prove altogether too much. It would make a nonsense of much that was said in Bankovic, not least as to the Convention being ‘a constitutional instrument of European public order’, operating ‘in an essentially regional context’, ‘not designed to be applied throughout the world, even in respect of the conduct of contracting states’ (para 80). It would, indeed, make redundant the principle of effective control of an area: what need for that if jurisdiction arises in any event under a general principle of ‘authority and control’ irrespective of whether the area is (a) effectively controlled or (b) within the Council of Europe?264
Put aside, for one moment, the understandable fear of judges that the application of the personal model up to its logical conclusion would render them unable to avoid dealing with the merits of legally and factually complex and politically extremely controversial cases. Even if that fear, based as it is on considerations of effectiveness, could be addressed in a satisfactory way—as I think it can be—under the personal model jurisdiction clauses in human rights treaties would essentially be rendered meaningless. Any act capable of violating a person’s rights as a substantive matter would also be capable of bringing that person within the state’s jurisdiction, and no jurisdiction issue would therefore actually arise. And, as Lord Brown well notes, if the personal model is applied consequentially, what would be the use for the spatial model?
A further difficulty with the personal model is its lack of an adequate textual basis in at least some of the treaties. While the words individuals ‘within its jurisdiction’ and ‘subject to its jurisdiction’ can reasonably be interpreted as denoting a power-based relationship between the state and the individual, as we have seen such an interpretation is textually simply impossible for some jurisdiction clauses, such as the CAT’s ‘any territory under its jurisdiction’. These words unambiguously denote a relationship between the state and the territory and only indirectly between the state and the individuals located in that territory. No amount of creative interpretation can plausibly change this result.
But again, more important than the textual difficulties is the fact that the personal model collapses. Because it does, it is unlikely to be endorsed wholeheartedly by courts, especially those in which the stakes are the highest, as in Strasbourg. Rather, they tend to use it in a wholly unpredictable and unprincipled fashion, to carve limited exceptions out of Bankovic and its territorial model of jurisdiction one (p.209) sliver at a time, in cases in which it would be morally difficult not to do so, and yet not too inconvenient politically. Do they have some other choice, other than running headlong into utopia? Is there a better way?
4. A Third Model: Territorial Jurisdiction and the Distinction Between Positive and Negative Obligations
A. Universality unbound
The driving force behind the personal model is easy to spot—how could it be justified against the normative baseline of universality that a state which is in full control of its own agents is dispensed from respecting the human rights of persons whose lives its agents affect, merely on account of their location? Yet it is precisely this same driving force which ultimately leads to the collapse of the personal model of jurisdiction. Why then not simply say that states have to comply with their negative obligation to respect human rights in all circumstances, regardless of whether they exercise jurisdiction over a particular territory or area?
There is truly a fundamental distinction between a state’s obligation to respect human rights, which requires it only to refrain from infringing upon an individual’s rights without adequate justification, and its duty to secure or ensure human rights to the inhabitants of a certain territory, which, in certain circumstances, compels the state to prevent the violations of human rights even by private parties.265 According to the Human Rights Committee,
The article 2, paragraph 1, obligations are binding on States [Parties] and do not, as such, have direct horizontal effect as a matter of international law. The Covenant cannot be viewed as a substitute for domestic criminal or civil law. However the positive obligations on States Parties to ensure Covenant rights will only be fully discharged if individuals are protected by the State, not just against violations of Covenant rights by its agents, but also against acts committed by private persons or entities that would impair the enjoyment of Covenant rights in so far as they are amenable to application between private persons or entities. There may be circumstances in which a failure to ensure Covenant rights as required by article 2 would give rise to violations by States Parties of those rights, as a result of States Parties' permitting or failing to take appropriate measures or to exercise due diligence to prevent, punish, investigate or redress the harm caused by such acts by private persons or entities.266
(p.210) The positive obligation of a state to ensure the human rights of persons within its jurisdiction from violations by private parties is not absolute, as states are neither omniscient nor omnipotent. What they must do is to exercise due diligence, i.e. to take all measures reasonably within their power in order to prevent violations of human rights. As held by the Inter-American Court,
…in principle, any violation of rights recognized by the Convention carried out by an act of public authority or by persons who use their position of authority is imputable to the State. However, this does not define all the circumstances in which a State is obligated to prevent, investigate and punish human rights violations, nor all the cases in which the State might be found responsible for an infringement of those rights. An illegal act which violates human rights and which is initially not directly imputable to a State (for example, because it is the act of a private person or because the person responsible has not been identified) can lead to international responsibility of the State, not because of the act itself, but because of the lack of due diligence to prevent the violation or to respond to it as required by the Convention.267
In order to be realistically complied with, the obligation to respect human rights requires the state to have nothing more than control over the conduct of its own agents. It is the positive obligation to secure or ensure human rights which requires a far greater degree of control over the area in question, control which allows the state to create institutions and mechanisms of government, to impose its laws, and punish violations thereof accordingly.
This is then what my proposed third model would amount to: the notion of jurisdiction in human rights treaties would be conceived of only territorially, as de facto effective overall control of areas and places. Having now looked at the text of the relevant treaties and the treaty practice of states generally, as well as at the case law, this is indeed the most natural way of interpreting the term ‘jurisdiction’. This threshold would, however, apply only to the state’s obligation to secure or ensure human rights, but not to its obligation to respect human rights, which would be territorially unbound.268
I should not be taken as arguing that the state is in exactly the same position with respect to its negative obligations when it acts inside its territory in conditions of normalcy, as when it acts outside it, particularly in wartime or other extraordinary situations. What I am arguing is that the best way to address these difficulties, and to fully take into account all considerations of effectiveness, is not in artificially imposing a threshold for the state’s negative obligations, but in applying the substance of these obligations to the facts at hand with a greater degree of flexibility.269 As Lord Justice Sedley aptly put it in Al-Skeini:
(p.211) [I]t is not an answer to say that the UK, because it is unable to guarantee everything, is required to guarantee nothing. The question is whether our armed forces' effectiveness on the streets [of Basra] in 2003–4 was so exiguous that despite their assumption of power as an occupying force they lacked any real control of what happened from hour to hour in the Basra region. My own answer would be that the one thing British troops did have control over, even in the labile situation described in the evidence, was their own use of lethal force. Whether they were justified in using it in the situations they encountered, of which at least four of the cases before us are examples, is precisely the subject of the inquiry which the appellants seek. It is in such an inquiry that the low ratio of troops to civilians, the widespread availability of weapons and the prevalence of insurgency would fall to be evaluated.270
He did not, however, think that this approach was open to him under existing Strasbourg jurisprudence. And he was right—it is not. Adopting the third model would require a radical rethink of Strasbourg’s approach, and to a lesser extent also that of other human rights bodies. I am well aware that this makes the third model less attractive. But having now extensively examined the European Court’s convoluted case law on extraterritoriality, would it really be such a bad thing to put it on some sensible, principled foundation? I think not. It bears emphasizing, however, that even in regular peacetime conditions the European Court has been prepared to approach positive obligations flexibly.271 Thus, in the context of the right to life it remarked that
[f]or the Court, and bearing in mind the difficulties involved in policing modern societies, the unpredictability of human conduct and the operational choices which must be made in terms of priorities and resources, such an obligation must be interpreted in a way which does not impose an impossible or disproportionate burden on the authorities. Accordingly, not every claimed risk to life can entail for the authorities a Convention requirement to take operational measures to prevent that risk from materialising.…In the opinion of the Court where there is an allegation that the authorities have violated their positive obligation to protect the right to life in the context of their above-mentioned duty to prevent and suppress offences against the person (see paragraph 115 above), it must be established to its satisfaction that the authorities knew or ought to have known at the time of the existence of a real and immediate risk to the life of an identified individual or individuals from the criminal acts of a third party and that they failed to take measures within the scope of their powers which, judged reasonably, might have been expected to avoid that risk.272
Let me now make my argument in more detail. I will first try to establish whether the text of the relevant treaties could accommodate the third model. While I am not saying that my argument is free of all difficulties, I believe that such difficulties can be overcome, particularly if the negative obligation to respect human rights is read into some of the treaties implicitly. Further, as I will explain, I am not advocating a crude distinction between negative and all positive obligations. (p.212) Rather, some positive obligations are procedural or prophylactic in nature, tied solely to the state’s compliance with its negative obligation to respect human rights. In my view, it is only the wide-ranging obligation to secure or ensure human rights, particularly from acts of third parties, that requires a jurisdictional threshold. Finally, I will try to prove that, when compared to its rivals, this third model actually provides the best balance between universality and effectiveness, if at a cost.
B. Textual interpretation and implicit negative obligations
Is the third model even possible under the text of the relevant human rights treaties? Let us again look at what they say. Article 1 ECHR thus provides that ‘[t]he High Contracting Parties shall secure to everyone within their jurisdiction the rights and freedoms defined in Section I of this Convention’ (emphasis added). All this article says, as I read it, is that the positive obligation to secure human rights is contingent on state jurisdiction—something that I certainly do not dispute. Except in the descriptive heading of the article, it says nothing about the negative obligation to respect human rights, which does not mean that this obligation does not exist. Indeed, many human rights treaties with jurisdiction clauses explicitly refer only to the positive obligations of the states parties.
For example, all of the nine jurisdiction clauses in the CAT denote the scope of a positive obligation. Article 2(1) provides that ‘[e]ach State Party shall take effective legislative, administrative, judicial or other measures to prevent acts of torture in any territory under its jurisdiction’ (emphasis added); Articles 5(1)(a), 5(2), and 7(1) require states to criminalize and prosecute torture on the basis of the territoriality and universality principles; Articles 11, 12, and 13 require states to review interrogation rules for the purpose of preventing torture, ensure an effective investigation, and provide remedies to individuals; while Article 16 requires states to prevent cruel, inhuman, and degrading treatment, all in ‘any territory under its jurisdiction’. Notably, these provisions do not expressly say that the state as such has the negative obligation not to torture individuals or treat them inhumanely, probably because it was obvious that such an obligation existed, yet it would have been somewhat impolitic to spell it out. Similarly, Article 3 CERD provides that ‘States Parties particularly condemn racial segregation and apartheid and undertake to prevent, prohibit and eradicate all practices of this nature in territories under their jurisdiction’ (emphasis added).
The treaties which do explicitly mention the obligation to respect are those whose jurisdiction clauses are to a greater or lesser extent based on Article 2(1) ICCPR, which provides that ‘[e]ach State Party to the present Covenant undertakes to respect and to ensure to all individuals within its territory and subject to its jurisdiction the rights recognized in the present Covenant’. The grammatically more natural interpretation of this wording would be to say that the jurisdiction threshold applies to both the obligation to respect and to the obligation to ensure. This interpretation is however not the only plausible one—Article 2(1) could also be read as limiting the jurisdiction threshold only to the obligation to ensure, which (p.213) would be consistent with its object and purpose.273 Article 2(1) CRC and Article 7 of the Migrant Workers Convention are similar in this respect. Notably, however, Article 1(1) ACHR stipulates that the ‘States Parties to this Convention undertake to respect the rights and freedoms recognized herein and to ensure to all persons subject to their jurisdiction the free and full exercise of those rights and freedoms’ (emphasis added). This formulation could quite comfortably be seen as imposing a jurisdictional threshold for the obligation to ensure alone.
Coming back to the treaties which do not mention the obligation to respect explicitly, most importantly the ECHR, that particular obligation is either spelled out in provisions which guarantee specific rights, or can if necessary be implied into the content of the treaty in question. Crucially, if the negative obligation to respect human rights generally, or for example the obligation not to torture specifically, is implicitly read into the treaties, there is no reason why it should depend on the same jurisdictional threshold of application as the positive obligation to secure or ensure.
The best authority for these points is the ICJ’s merits judgment in the Bosnian Genocide case.274 Just like the CAT, which does not say in so many words that states will be responsible if their organs or agents commit torture, so the Genocide Convention does not explicitly provide for state responsibility for the commission of genocide. Thus, Article I of the Convention obliges states to prevent and punish genocide, while other provisions elaborate on these requirements. Both at the preliminary objections and at the merits stage of the case the FRY/Serbia disputed the existence of a separate state obligation under the Convention not to commit genocide, asserting that the Convention was a classical international criminal law treaty, dealing with crimes committed by individuals, not states. All the Convention did, in Serbia’s view, was to require states parties to criminalize in their domestic law the crimes that it defines, and then prosecute the perpetrators of these crimes. Article IX of the Convention, which does mention the responsibility of states for genocide, was, in Serbia’s argument, merely a compromissory clause which did not impose any additional substantive obligations upon states parties.
The Court first reasoned that Article I of the Convention is not merely hortatory or purposive, and that ‘in particular its undertaking to prevent, creates obligations distinct from those which appear in the subsequent Articles [of the Convention]’ so that the ‘the Contracting Parties have a direct obligation to prevent genocide’.275 It then held as follows:
The Court next considers whether the Parties are also under an obligation, by virtue of the Convention, not to commit genocide themselves. It must be observed at the outset that such an obligation is not expressly imposed by the actual terms of the Convention. The Applicant (p.214) has however advanced as its main argument that such an obligation is imposed by Article IX, which confers on the Court jurisdiction over disputes ‘including those relating to the responsibility of a State for genocide or any of the other acts enumerated in Article III’. Since Article IX is essentially a jurisdictional provision, the Court considers that it should first ascertain whether the substantive obligation on States not to commit genocide may flow from the other provisions of the Convention. Under Article I the States parties are bound to prevent such an act, which it describes as ‘a crime under international law’, being committed.
The Article does not expressis verbis require States to refrain from themselves committing genocide. However, in the view of the Court, taking into account the established purpose of the Convention, the effect of Article I is to prohibit States from themselves committing genocide. Such a prohibition follows, first, from the fact that the Article categorizes genocide as ‘a crime under international law’: by agreeing to such a categorization, the States parties must logically be undertaking not to commit the act so described. Secondly, it follows from the expressly stated obligation to prevent the commission of acts of genocide. That obligation requires the States parties, inter alia, to employ the means at their disposal, in circumstances to be described more specifically later in this Judgment, to prevent persons or groups not directly under their authority from committing an act of genocide or any of the other acts mentioned in Article III.
It would be paradoxical if States were thus under an obligation to prevent, so far as within their power, commission of genocide by persons over whom they have a certain influence, but were not forbidden to commit such acts through their own organs, or persons over whom they have such firm control that their conduct is attributable to the State concerned under international law. In short, the obligation to prevent genocide necessarily implies the prohibition of the commission of genocide.276
The Court thus concluded that the ‘Contracting Parties to the Convention are bound not to commit genocide, through the actions of their organs or persons or groups whose acts are attributable to them’.277 It then extended the same reasoning to crimes ancillary to genocide defined by Article III of the Convention, that is conspiracy to commit genocide, direct and public incitement to genocide, attempt to commit genocide, and complicity in genocide.278
Turning back to human rights treaties, to my mind the Article 2(1) CAT obligation of the state ‘to prevent acts of torture in any territory under its jurisdiction’ necessarily implies the state’s obligation to itself refrain from torture, just as with the Genocide Convention. More generally, the Article 1 ECHR obligation to secure human rights implies the obligation to respect them. That implication does not, however, necessarily require the same threshold for its existence, that of state jurisdiction over territory.
Thus, for example, in the Bosnian Genocide case the ICJ thought that the obligation not to commit genocide had no territorial limitation. States were implicitly prohibited from committing genocide anywhere in the world. The Court had the same view with regard to the positive obligation to prevent, an (p.215) obligation of states to exercise due diligence and ‘employ all means reasonably available to them, so as to prevent genocide so far as possible’,279 saying that ‘[t]he substantive obligations arising from Articles I and III are not on their face limited by territory. They apply to a State wherever it may be acting or may be able to act in ways appropriate to meeting the obligations in question.’280
However, in his separate opinion Judge Tomka forcefully argued that the positive obligation to prevent should be limited territorially. According to him,
[U]nder Article I of the Genocide Convention the State does have an obligation to prevent genocide outside its territory to the extent that it exercises jurisdiction outside its territory, or exercises control over certain persons in their activities abroad. This obligation exists in addition to the unequivocal duty to prevent the commission of genocide within its territory.281
In essence, Judge Tomka argued for the imposition of a threshold criterion on the obligation to prevent—state jurisdiction over a territory—the same threshold as in human rights treaties. Under his approach, a state would have to exercise effective overall control of an area in which there is a serious risk of genocide being committed by some other actor in order for its obligation to prevent genocide to arise. His argument was in essence one of policy—that a state should be expected to prevent genocide only when it had territorial jurisdiction, because that obligation was much more onerous then the simple obligation to refrain from committing genocide.282
My point is simply this—there is no inherent contradiction in implying, where necessary, the negative obligation to respect human rights into the relevant treaties and that obligation having a broader, territorially unlimited scope of application than the positive duty to secure or ensure human rights, or prevent violations thereof. Thus, under the CAT for example, we could say that the state has the duty to prevent torture only in territories under its jurisdiction, but that it has the obligation to itself refrain from torturing in all circumstances. Likewise, under the ECHR, the state’s obligation to secure human rights would be limited to areas under the state’s effective overall control, but its duty to respect human rights would apply everywhere, without any territorial limitation. For example, the ECHR would apply to the taking of property by the UK within the UK of a person residing in Monaco, whether in law or merely in fact.
C. Prophylactic and procedural positive obligations
Let me again emphasize that I am not advocating a strict separation between negative and positive obligations. Rather, I am arguing for a separation between those positive obligations which require control over territory in order to be effective, such as the obligation to prevent inhuman treatment or secure human rights generally even from (p.216) third parties, and those obligations whose effectiveness depends only on the state’s control over its own agents. The practice of the European Court and other human rights bodies has long recognized that some positive obligations are of a procedural or prophylactic nature.283 Thus, for example, in the context of the right to life, the state has the negative obligation not to take life unjustifiably, but also the positive obligation to conduct an independent and effective investigation into a possible taking of life by its own agents, e.g. the police or the armed forces. As explained by the House of Lords,
The European Court has also interpreted article 2 as imposing on member states a procedural obligation to initiate an effective public investigation by an independent official body into any death occurring in circumstances in which it appears that one or other of the foregoing substantive obligations has been, or may have been, violated and it appears that agents of the state are, or may be, in some way implicated.284
In Lord Bingham’s view, ‘[t]his procedural duty does not derive from the express terms of article 2, but was no doubt implied in order to make sure that the substantive right was effective in practice’.285 This procedural obligation is, however, not the same as the other positive aspect of Article 2, which flows from the state’s obligation to secure human rights, namely ‘to establish a framework of laws, precautions, procedures and means of enforcement which will, to the greatest extent reasonably practicable, protect life’.286 That obligation is much more far-reaching, as it requires the state to prevent and investigate even acts of purely private violence, in which the agents of the state are not implicated at all.
Therefore, in my view, the first category of positive obligations, which exist solely to make the state’s negative obligations truly effective, should apply co-extensively with the negative obligations themselves. On the other hand, those positive obligations which flow from the state’s duty to secure or ensure human rights or prevent violations thereof—say prevent private violence or discrimination—require a threshold that sets out the limits of realistic compliance. And that threshold is precisely state jurisdiction, i.e. control over territory.
On the other hand, the intensity of these obligations also differs; a state has an absolute duty to investigate violations of human rights committed by its own agents, or committed against persons in its custody.287 It also has to provide persons whom it deprives of liberty with food, clothing, or health care, even though it might not have such obligations towards the general population.288 However, (p.217) when it comes to its positive obligations with regard to purely private conduct, they are indeed more far-reaching in scope but are also more flexible in content. In the words of the European Court in Osman, they must not ‘impose an impossible or disproportionate burden on the authorities’.289
We can thus distinguish between several kinds of positive obligations under human rights treaties, and this distinction should have a bearing on their extraterritorial application. To make this discussion somewhat less abstract, let us go back to Al-Skeini and the five applicants killed by British troops on patrol. Assume, for the sake of the argument, that even though the killings took place in British-occupied Basra, because of the strength of the insurgency Basra could not be qualified as an area under the UK’s effective overall control,290 and was hence outside its jurisdiction. Even so, in my view, the UK would still have not only the negative obligation to refrain from depriving the five applicants of life unjustifiably, but would also have the positive procedural obligation to conduct an effective investigation into their killing. Its existence depends solely on the UK’s own involvement in the killing, and in order to comply with it the UK need not do anything more than investigate the conduct of its own troops, which it is in principle perfectly able to do. However, were the killings actually done by third parties, be they insurgents or indeed the soldiers of an allied country, the UK would have had no obligation to investigate or prevent the deaths, since they took place in an area outside its jurisdiction, and it would in fact be exceedingly difficult, if not impossible, for the UK to conduct an effective investigation without actually having control over the territory.291
Or, take again the example of the assassination of Alexander Litvinenko in London, ostensibly with Russian involvement. The killing clearly took place outside Russia’s jurisdiction, if jurisdiction is conceived of territorially. Nonetheless, Russia would still have the obligation to investigate it, to the extent that a credible case can be made that its own agents were involved in Litvinenko’s death, for example by providing the killers with polonium, the radioactive substance with which Litvinenko was poisoned. If, however, such a credible case could not be made, and on the evidence available there was no reasonable suspicion that Litvinenko’s death was anything other than a purely private act, then it would (p.218) only be the UK which would have the obligation to investigate the murder since it took place within its jurisdiction.292
Similarly, on the facts of Bankovic, the respondent states should have been asked by the Court to justify on the merits their killing of individuals who were not within their jurisdiction territorially conceived, as the killing implicates the states’ negative obligation to which some positive obligations may attach. In doing so, the Court should have taken into account the relevant rules of international humanitarian law and the extraordinary circumstances of armed conflict, and adopted a more flexible approach to Article 2 than in a situation of normalcy. As I have argued above, however, there would be limits to that flexibility, motivated by the need to preserve the integrity of the ECHR regime.293 The Court would also have encountered serious evidentiary difficulties on the merits, but it has managed to deal with such difficulties in the past, as with Chechnya, for example by equitably distributing the burden of proof between the parties. Thus, though in my view the correct result in Bankovic would probably have been that the killings were unlawful, it is far from obvious that this should have been the case. The respondent states would have a case to answer, but they would also have something to answer the case with.
Or, if we take the example of the pending Aerial Herbicide Spraying case before the ICJ while taking the facts alleged by Ecuador as proven,294 Colombia would be responsible for violating the rights of Ecuadorian residents adversely affected by its spraying operation, as the operation was conduct by its own agents or on its own behalf. Colombia would not, however, have the same obligation under human rights treaties with regard to pollution or herbicides used by purely private actors operating within its territory and having effects in Ecuador. Similarly, if we consider (p.219) the scenarios of extraterritorial complicity that we examined above, with say a UK intelligence officer feeding questions and data to a Pakistani torturer interrogating a terrorist suspect in Pakistan,295 the UK would in my view have a territorially unlimited negative obligation under Article 3 ECHR not to be complicit in the torture of any person anywhere in the world. Accordingly, it would also have the positive obligation to investigate complicity in torture by its own agents wherever it might occur. It would not, however, have such an obligation with regard to an extraterritorial human rights violation in which its own agents did not participate at all, unless the act was committed in a territory under its effective overall control.
D. Reconciling universality and effectiveness
To my mind, this third model provides us with the best balance between universality and effectiveness with regard to the extraterritorial application of human rights treaties. Instead of being artificially limited, universality is brought to its logical (and moral) conclusion. States would have the same obligation to respect human rights both within and outside their territories. Whether they use drones for the targeted killings of suspected terrorists, use force in more conventional military operations, abduct or detain combatants or civilians or put them on trial, poison the crops of innocent farmers, or enforce their laws, states would still have to abide by the restrictions that human rights law places on the arbitrary exercise of their power, and do so regardless of territorial boundaries. When, however, states are expected to do more than just refrain from adversely affecting the lives of others, when they need to take positive steps, from preventing domestic violence296 and safeguarding private property to protecting lawful public assemblies297 and the free exercise of religion,298 they cannot fulfil such obligations effectively without having the tools to do so. Such obligations should, therefore, be territorially limited to areas and places under the state’s jurisdiction.
This model would, in principle, be able to accommodate all of the effectiveness concerns that generally militate against the extraterritorial application of human rights instruments, which I have examined above in some detail: flexibility, impact, regime integrity, clarity, and predictability.299 Flexibility is beyond doubt the most important, and can only be achieved on the merits and substance of each concrete instance of extraterritorial application to a specific set of facts. In the transition from threshold to substantive issues lies both the appeal of this model, and its greatest risk. It would be unreasonable to contend that most human rights obligations can be applied in exactly the same way in peacetime in the state’s own territory, and during a military occupation or in an active battlefield abroad. Full account must be (p.220) taken of the extraordinary circumstances in which the treaty is being applied extraterritorially, and substantive adjustments need to be made in order for extraterritorial application to truly become realistic and effective—for example, through the greater use of international humanitarian law in interpreting the norms of human rights law.
But, as I have said, though a price must be paid if universality is not to descend into utopia, that price must not be too steep. There still must be a point to extraterritorial application—it must have an actual impact—and the integrity of the human rights regime as a whole must not be compromised by its being watered down too much. Again, both of these concerns can only be addressed by examining the numerous substantive issues arising with extraterritorial application, such as targeted killings or preventive detention, which are as such beyond the scope of this study. My only point here is that such concerns can be addressed effectively, and that the extraterritorial application of human rights treaties can be both realistic and worthwhile. Obviously, the extension of the negative obligation to respect human rights that I propose can by itself have an enormous impact, as it would at the very least require some sort of justification by states for their acts outside their territory which violate the rights of countless people.
The third model also sufficiently addresses the Bankovic concern that human rights treaties would be compromised if state obligations were divided and tailored to match each specific extraterritorial action. The most onerous obligation by far, that to secure or ensure human rights, is strictly confined to areas over which states have effective control. The treaties are not chopped up and divided. Rather, they require states to respect human rights when they are in the position to do so.
Likewise, by allowing us to have those substantive disputes which are truly important, on matters like targeted killings or security detention, and by detaching these disputes from the preliminary threshold question of extraterritorial application, the third model would lead to a great simplification of the case law on extraterritorial application, its grounding in principle, and a rejection of casuistry. In other words, we would be able to tell with some clarity when human rights treaties actually apply extraterritorially—something that we are in all honesty incapable of doing with the present state of the jurisprudence, at least when it comes to the ECHR. That the current lack of certainty makes the lives of judges and academics more difficult is the least of its problems; it directly undermines the authority of the Court, and makes other participants in the system—above all states—unable to adequately factor the applicability of human rights treaties extraterritorially into their own policy-making.300
(p.221) I should also say that I am aware that it is impossible to conclusively test whether my proposed model strikes the best balance between universality and effectiveness without actually trying it out. From the perspective of the European human rights system, whose strength derives precisely from the fact that it is the European human rights system, the risk of a flood of litigation doing it irreparable damage may be seen by some—particularly judges—as being too great. That risk, it needs to be said, is a realistic one.301 I maintain, however, that the European Convention is sufficiently flexible, and that the tools given to the European Court are sufficiently powerful, to minimize that risk to a great extent. The floodgates will not open, and the European system will not be overwhelmed—at least not any more than it already is—just because the Court would no longer be able to use the preliminary question of extraterritorial application to avoid a number of vexing and controversial substantive issues. Nothing would prevent the Court, for example, from declaring inadmissible those cases which it considers manifestly ill-founded, after conducting even a deferential substantive analysis. And though the risk may be real, it is one worth taking, if universality of human rights is to have any meaning. Whether this will actually happen will depend on those who sit on the European Court of Human Rights, and on whether it is, in their view, the ‘European’ or the ‘Human Rights’ bit of their Court and Convention that matters most.
Finally, to the extent that my textual argument in favour of the third model is not considered to be persuasive, or that its adoption would require too radical a departure from existing jurisprudence, I would (purely as a pragmatic matter) not be opposed to the adoption as a substitute of the personal model of jurisdiction as state authority and control over individuals jointly with the spatial model. If the personal model is considered to be textually necessary, then it should apply to the negative obligation to respect, while the spatial model would apply to the positive obligation to secure or ensure human rights.302 Indeed, if we took another look at the various cases applying the personal model, we would see that they generally deal only with negative obligations of states, or with procedural or prophylactic positive obligations. So long as the personal model is not in any way limited, its application would lead to the same result as my third model, just with the added pretence of applying a jurisdictional threshold to negative obligations when no such threshold in fact exists. This would be an adequate solution for interpreting the jurisdiction clauses in some treaties, such as in the First Optional Protocol to the ICCPR, which limit the right to individual petition only to those persons subject to the state’s jurisdiction. This would also be an adequate answer to Lord Brown’s challenge in Al-Skeini that the spatial model would become redundant if the personal one were to be adopted, since each would apply to different types of state obligations.
(p.222) Having thus outlined the model of extraterritorial application that I would prefer, I again do not wish to be taken as arguing that this model is free of all weaknesses. In some cases at least, the distinction between positive and negative obligations is not as clear-cut as I have made it seem. The same goes for the distinction between positive obligations arising from the duty to secure or ensure human rights, which I would subject to a threshold of territorial jurisdiction, and those which are prophylactic or procedural in nature and thus appurtenant to negative obligations, which I would not. As with most things, there would be grey areas in which my model would not provide a clear answer. And there probably would be areas in which my model would give answers that would run counter to our legal or moral intuitions regarding the territorial scope of human rights treaties. Even so, in my view it provides a principled foundation that allows us to move from the question of when human rights treaties apply extraterritorially to the far more important and difficult question of how they should do so.
5. The Special Problem of the ICCPR
Having discussed the available models of extraterritorial application, I will briefly turn back to an issue that I have mentioned but until now have more or less studiously avoided—the proper interpretation of Article 2(1) ICCPR, providing that ‘[e]ach State Party to the present Covenant undertakes to respect and to ensure to all individuals within its territory and subject to its jurisdiction the rights recognized in the present Covenant’. As we have seen, this clause presents two interpretative difficulties: first, whether the obligation to respect human rights is modified and limited by the remainder of the clause, or whether it is only the obligation to ensure which is subject to limitation; and secondly, whether that limitation is conjunctive or disjunctive, i.e. whether individuals have to be both within a state’s territory and subject to its jurisdiction to have ICCPR rights, or is it rather that states have to guarantee these rights to all individuals within their territories and to those subject to their jurisdiction? As for the former issue, I have just explained why there is in my view no bar to reading Article 2(1) as imposing no limitation on the obligation to respect, similarly to Article 1 ACHR. As for the latter, I have indicated several times how the Human Rights Committee has now espoused the more expansive disjunctive interpretation of Article 2(1),303 and it will come as no surprise to the reader that this is an interpretation I agree with in principle—but let me now offer my reasons for doing so, other than that this is simply the result that I prefer as a matter of policy.
First, I think it fair to say that the conjunctive reading of Article 2(1) is textually or grammatically more natural, and interpretation under Article 31 VCLT does of course start from the text. This does not mean, however, that this is the only (p.223) ‘ordinary’ or ‘plain’ meaning of the text, as e.g. argued by the United States in its appearances before the Human Rights Committee,304 or by Michael Dennis in an academic setting.305 The use of grammatical particles such as ‘and’ or ‘or’ frequently leads to ambiguity, and accordingly to more than one available ‘plain’ or ‘ordinary’ meaning of the text. That ambiguity may be resolved by the (grammatical) context in which the words appear,306 but there is nothing in Article 2(1) which assists us in that regard. Recourse must thus be had to the object and purpose of the treaty, which as we have seen is founded upon the inherent dignity of all human beings and the universal nature of human rights. Although the conjunctive reading may be grammatically the more natural, the disjunctive reading is certainly the one more consonant with the treaty’s object and purpose.
Again, there is nothing truly novel in regarding words such as ‘and’ or ‘or’ as ambiguous and interchangeable in certain respects, if not generally. Not only do such ambiguities frequently arise in everyday linguistic usage,307 but courts have a long history of addressing them. Almost 150 years ago the US Supreme Court declared that
[i]n the construction of statutes, it is the duty of the court to ascertain the clear intention of the legislature. In order to do this, courts are often compelled to construe ‘or’ as meaning ‘and,’ and again ‘and’ as meaning ‘or.’308
It is precisely in the search for the intent of the legislator (as problematic as the whole idea of legislative intent is in the context of interpretation) and the purpose that a legal enactment is supposed to achieve that the ‘[d]isjunctive “or” and the conjunctive “and” may be interpreted as substitutes’.309 And in the international context, for example, it is the very same US administration which rejected a disjunctive reading of the ‘and’ in Article 2(1) ICCPR which employed a similar disjunctive reading of an ‘and’ in a rather important Security Council resolution.310
(p.224) This transformation of an ‘and’ into an ‘or’ might be regarded as a transparent lawyer’s trick, enabling him or her to get to the policy result that he or she wants. Depending on the circumstances, that might even be true. But while interpretation is not a process that should in principle be result-driven, it is equally true that the object and purpose of a treaty are considerations that must be taken into account when interpreting it, and that they may tend to prefer one result over another. In other words, an interpretation of the ICCPR which favours universality and human dignity is within the limits of textual vagueness or ambiguity by definition preferable to an interpretation which runs against the grain of the treaty.
This now brings me to the ICCPR’s travaux préparatoires, on which the United States has relied in particular in support of its argument for a strictly territorial application of the treaty.311 In the US view, the travaux clearly demonstrate that the ICCPR was to have no extraterritorial application. It was in fact on a US proposal that the reference to territory was introduced into Article 2(1), which had previously contained only a jurisdiction clause, and subsequent attempts by some parties to have this language deleted were rejected. The US representative during the drafting of the then single Covenant, Eleanor Roosevelt, explained this wording in the following terms:
The purpose of the proposed addition [is] to make it clear that the draft Covenant would apply only to persons within the territory and subject to the jurisdiction of the contracting states. The United States [is] afraid that without such an addition the draft Covenant might be construed as obliging the contracting states to enact legislation concerning persons, who although outside its territory were technically within its jurisdiction for certain purposes. An illustration would be the occupied territories of Germany, Austria and Japan: persons within those countries were subject to the jurisdiction of the occupying states in certain respects, but were outside the scope of legislation of those states. Another illustration would be leased territories; some countries leased certain territories from others for limited purposes, and there might be questions of conflicting authority between the lessor nation and the lessee nation.312
In the US argument, this explanation not only conclusively excludes the extraterritorial application of the ICCPR to occupied territories, but also with regard to leased territories, foremost among them Guantanamo.
A comprehensive reading of the travaux, however, paints a different picture.313 Rather than giving us clarity, it provides us with an ample dose of confusion and doubt, as is so often the case when having recourse to preparatory work. Even the passage quoted above shows us that the US negotiators at the time were not concerned so much with the application as such of the ICCPR to occupied (p.225) territories, but with a purported obligation to legislate for such territories. This is the result that they wanted to avoid. In truth, bearing in mind the case law we examined, no such general obligation would exist even under the most expansive reading of the obligation to ensure human rights—indeed, as we have seen, Article 43 of the Hague Regulations is designed precisely to prevent occupants from modifying the laws of the occupied territories, except when absolutely necessary to do so.314 In some cases, however, the obligation to ensure might require the occupant to alter the legal system in order to effectively maintain peace and security in the occupied territory, or to fulfil other human rights commitments, and in exceptional circumstances this obligation might lead to an unresolvable norm conflict. Similarly, the US concern with leased territories was that the extension of human rights obligations might lead to questions of ‘conflicting authority between the lessor nation and the lessee nation’, i.e. to norm conflict. The US desires aside, however, whether the language inserted into the treaty actually successfully manages to avoid all such norm conflicts is a different matter—I think it does not.
Other sections of the travaux tell us that the drafters wanted to avoid states parties assuming the obligation to ensure the rights of their own nationals in a foreign country.315 They did not want the states parties to have to do the impossible, for example to ensure the right to a fair trial of their national before a foreign court.316 As the ICJ itself said in the Wall case,
The travaux préparatoires of the Covenant confirm the [Human Rights] Committee’s interpretation of Article 2 of that instrument. These show that, in adopting the wording chosen, the drafters of the Covenant did not intend to allow States to escape from their obligations when they exercise jurisdiction outside their national territory. They only intended to prevent persons residing abroad from asserting, vis-à-vis their State of origin, rights that do not fall within the competence of that State, but of that of the State of residence (see the discussion of the preliminary draft in the Commission on Human Rights, E/CN.4/SR.194, para. 46; and United Nations, Official Records of the General Assembly, Tenth Session, Annexes, A/2929, Part II, Chap. V, para. 4 (1955)).317
The travaux emphatically do not offer us what we most need—a set of first principles as to the proper interpretation of Article 2(1). They simply do not provide conclusive evidence that the drafters wanted a strictly conjunctive reading that would have limited the application of the treaty solely to those territories over which the states parties had title, and that they did so contrary to the basic purpose of the treaty that they themselves professed, that of universality. Nor do the travaux tell us, for example, whether the word ‘jurisdiction’ refers to effective overall control (p.226) over a territory, or to authority and control over an individual, or to something else entirely, such as the exercise of legal authority or power or the right to exercise such power. They only give us glimpses of the particular results or applications of the treaty that (some) of the states parties wanted to avoid.
But it is not such original expected applications of a treaty which form its legally binding meaning. Even if, for example, it was clear from the travaux (which it is not) that states parties did not want the ICCPR to apply during occupation, this does not exclude the possibility that the language that they employed, as well as the broader object and purpose of the treaty, do precisely lead to such a result.
Philip Allott famously remarked that a treaty is a ‘disagreement reduced to writing’.318 I would not go so far as to label the ICCPR itself as such, but its travaux certainly qualify. I am prepared to concede, however, that if we could by some necromantic art revive Eleanor Roosevelt and the other drafters of the two Covenants, and present them one by one with the various scenarios of extraterritorial application that we are concerned with today, they might be disinclined to opt for an expansive approach. But I would also imagine that if we asked them whether under the treaties they were drafting Auschwitz would technically not have been a violation thereof because it was located in occupied Poland, rather than in a territory over which the German Reich had legal title, it would be very doubtful that Roosevelt et al. would have found such an interpretation acceptable. And yet this is precisely what the US position on strict territoriality would entail.
If the disjunctive interpretation of Article 2(1) were to be adopted, what would then be the purpose of the words ‘within its territory’ that were added to the jurisdiction clause? Would they not thus be rendered ineffective? Not necessarily. These words would then make ICCPR obligations more onerous than those under other human rights treaties, as Article 2(1) would require states to ensure the human rights of persons within its territory even when that territory is not under its jurisdiction, e.g. if a foreign state or an insurrectional movement was in control thereof. In other words, à la Ilascu, states parties would have a positive obligation to ensure the human rights of the population of their sovereign territories even in the absence of territorial control, though this obligation would have to be significantly attenuated.
Alternatively, even if the conjunctive reading of Article 2(1) were to be seen as correct, the application of the ICCPR would not have to be confined only to those territories over which the state has both title and effective control. As we have seen, the negative obligation to respect could comfortably be read as being subject to no territorial limitation. Moreover, though the possessive ‘its’ in ‘within its territory’ implies that a state would indeed need to have title or sovereignty over territory, this would not necessarily be the only plausible interpretation. In light of the universality-driven object and purpose of the treaty, the ‘within its territory’ clause could be read as not requiring title, but control.319 In other words, the combination of the reference to territory and the reference to state jurisdiction could be taken as setting (p.227) out a spatial test of jurisdiction over territory that would need to be met in order for the positive obligation to ensure human rights to arise. This interpretation would hence be consonant with my third model of extraterritorial application.
6. Treaties Without Jurisdiction Clauses
This brings me to the problem posed by treaties which contain no provisions on their territorial application such as jurisdiction clauses, or have jurisdiction clauses only for some but not all obligations under the treaty. As we have seen above,320 other than a brief mention of territorial jurisdiction in its Article 14, the ICESCR contains no jurisdiction clauses. Neither does the CEDAW, while most provisions of the CERD, particularly Articles 2 and 5 which protect a wide range of substantive rights, do not have any kind of territorial limitation. What then is the appropriate model of extraterritorial application of such treaties?
There is in principle no need for a one-size-fits-all approach to these treaties. Each should be evaluated and interpreted on its own merits, with regard to its specificities in text and object and purpose. However, as I see it, three basic options present themselves:
(1) The treaty, or particular obligations within it, could be interpreted as being completely territorially limited. It is hard to see how such a restrictive approach could ever be justified from the standpoint of universality in the absence of reasonably clear text to that effect, particularly because state ability to ensure or secure even the most demanding socio-economic rights does not depend on title, but on control over territory. Such a restrictive approach would hence tend to slide into pure apology for legally unrestrained state power.
(2) The treaty, or particular obligations within it, could be interpreted as being completely territorially unlimited. This would be an approach akin to the ICJ’s holding in the Bosnian Genocide case that the obligation to prevent genocide was territorially unlimited, and required states to exercise due diligence and ‘employ all means reasonably available to them, so as to prevent genocide so far as possible’.321 However, even in the context of the prevention of genocide such a far-reaching obligation is undermined by its own vagueness. In other contexts, as for instance with regard to socio-economic rights, this obligation would be even vaguer. What would, for example, the US or EU exactly need to do to safeguard the right to food of say the starving population of North Korea? With one or two possible exceptions, as for example with the Article 2(1) ICESCR obligation to provide ‘international assistance and co-operation’ for the realization of socio-economic rights, this approach would tend to slide into utopia.
(p.228) (3) Finally, my third model of extraterritorial application could be read into the relevant treaties so that the negative obligation to respect the rights in question would be territorially unlimited, while the positive obligations arising from such treaties would generally require the exercise of territorial jurisdiction, i.e. effective overall control of an area, as such control is necessary for their effective realization. Thus, if we take as an example the Aerial Herbicide Spraying case between Ecuador and Colombia that is currently pending before the ICJ,322 we could say that Colombia has the obligation towards the people of Ecuador to respect their right to health and food, which the herbicide spraying would in principle be capable of violating. However, Colombia would not have the obligation (other than possibly as reparation for its prior wrongful act) to actually provide food or health care services to the population of Ecuador, since it has no jurisdiction over the relevant parts of Ecuador. Similarly, if we take as another example the facts of the current Georgia v. Russia dispute, Russian soldiers engaging in military action on Georgian soil would have had the negative obligation under Article 2(1)(a) CERD ‘to engage in no act or practice of racial discrimination against persons, groups of persons or institutions’. However, in the absence of jurisdiction, i.e. territorial control, Russia would not have had the positive obligation under Article 2(1)(d) CERD to ‘prohibit and bring to an end, by all appropriate means, including legislation as required by circumstances, racial discrimination by any persons, group or organization’, in other words to prevent private discrimination.
This discussion is admittedly somewhat simplistic. Each of the treaties in question is deserving of a fuller account, something that I do not wish to engage in at this point.323 It is particularly the lack of concrete cases that necessarily makes this discussion somewhat abstract, but as we have seen, some relevant cases are already being litigated. As a general matter, however, I believe that as with treaties which contain jurisdiction clauses, the third model provides the best balance between universality and effectiveness and could provide a default position even for treaties without jurisdiction clauses, which could then be modified if the particular circumstances of the treaty in question so require.
(1) See below Sections 2.B and 2.C.
(2) See, however, K. Mujezinovic Larsen, ‘“Territorial Non-Application” of the European Convention on Human Rights,’ (2009) 78 Nord J Int’l L 73, who argues in favour of a notion of ‘residual jurisdiction’ flowing from title to territory, out of which only a limited set of positive obligations would arise.
(8) For example, in situations of armed conflict the lawfulness of the killing could be assessed by interpreting human rights law by taking into account the applicable rules of IHL—see Chapter V below.
(9) For general background, see, e.g., ‘Drones Are Weapons of Choice in Fighting Qaeda’, The New York Times, 16 March 2009, available at 〈http://www.nytimes.com/2009/03/17/business/17uav.html?_r=1〉. For an insightful defence of the use of drones for targeted killings in US counter-terrorism policy, see K. Anderson, ‘Predators Over Pakistan,’ The Weekly Standard, 8 March 2010, available at 〈http://ssrn.com/abstract=1561229〉.
(10) See, e.g., ‘Dubai Points Finger at Mossad over Hamas Assassination,’ The Sunday Times, 18 February 2010, available at 〈http://www.timesonline.co.uk/tol/news/world/middle_east/article7031749.ece〉; ‘Dubai Hamas Assassination: How It Was Planned’, The Telegraph, 17 February 2010, available at 〈http://www.telegraph.co.uk/news/worldnews/middleeast/dubai/7251960/Dubai-Hamas-assassination-how-it-was-planned.html〉.
(11) See, e.g., ‘China Defends Arms Sales to Sudan,’ BBC News, 22 February 2008, available at 〈http://news.bbc.co.uk/2/hi/asia-pacific/7258059.stm〉.
(12) See, e.g., ‘Lithuania Hosted Secret CIA Prisons,’ BBC News, 22 December 2009, available at 〈http://news.bbc.co.uk/1/hi/8426028.stm〉; ‘Secret Prisons: Obama’s Order to Close “Black Sites”’, The Guardian, 23 January 2009, available at 〈http://www.guardian.co.uk/world/2009/jan/23/secret-prisons-closure-obama-cia〉; ‘A Window Into C.I.A.’s Embrace of Secret Jails’, The New York Times, 12 August 2009, available at 〈http://www.nytimes.com/2009/08/13/world/13foggo.html?pagewanted=1&_r=2&hp〉. See also M. Danner, ‘US Torture: Voices from the Black Sites’, The New York Review of Books, Vol. 56(6), 9 April 2009, available at 〈http://www.nybooks.com/articles/22530〉 (giving details on a confidential ICRC report on prisoner abuse at such sites).
(13) See further N. Bascomb, Hunting Eichmann: How a Band of Survivors and a Young Spy Agency Chased Down the World’s Most Notorious Nazi (HMH, 2009).
(14) See UNSCR 138 (1960).
(15) See, e.g., ‘US Accused of Holding Terror Suspects on Prison Ships’, The Guardian, 2 June 2008, available at 〈http://www.guardian.co.uk/world/2008/jun/02/usa.humanrights〉.
(18) See, e.g., Joint Committee on Human Rights, ‘Allegations of UK Complicity in Torture’, 21 July 2009, available at 〈http://www.publications.parliament.uk/pa/jt200809/jtselect/jtrights/152/152.pdf〉; ‘Torture—new claim of secret UK complicity’, The Guardian, 26 July 2009, available at 〈http://www.guardian.co.uk/world/2009/jul/26/alam-ghafoor-torture-uk-intelligence〉.
(19) ILC ASR Commentary, at 157, para. 6.
(20) One could argue, however, that they were bound by the same obligation in substance, even though it did not come from the same source.
(21) For the relevance of this distinction between primary and secondary complicity in the Bosnian Genocide case before the ICJ, see M. Milanovic, ‘State Responsibility for Genocide: A Follow-Up’, (2007) 18 EJIL 669, at 680–4.
(23) Similarly, see United States v. Alvarez-Machain, 504 U.S. 655 (1992).
(24) See, e.g., Trail Smelter (United States v. Canada) (1941) 3 RIAA 1905.
(25) Aerial Herbicide Spraying (Ecuador v. Colombia), Application Instituting Proceedings, 31 March 2008, para. 4.
(27) Loizidou v. Turkey, App. No. 15318/89, Judgment (preliminary objections) of 23 February 1995, para. 62 (citations omitted, emphasis added).
(28) For example, the Human Rights Committee has found that the ICCPR was applicable to Occupied Palestinian Territories—see Concluding Observations of the Human Rights Committee: Israel, UN Doc. CCPR/C/79/Add.93, 18 August 1998, para. 10: ‘the Covenant must be held applicable to the occupied territories and those areas…where Israel exercises effective control’, as did the Committee on Economic, Social, and Cultural Rights with regard to the ICESCR—see Concluding Observations of the Committee on Economic, Social and Cultural Rights: Israel, UN Doc. E/C.12/1/Add.27, 4 December 1998, para. 8: ‘The Committee is of the view that the State’s obligations under the Covenant, apply to all territories and populations under its effective control.’ The Committee Against Torture did the same with regard to the CAT in Iraq and Afghanistan—see Conclusions and Recommendations: United Kingdom of Great Britain and Northern Ireland, UN Doc. CAT/C/CR/33/3, 10 December 2004, para 4(b): ‘The Committee expresses its concern at: (b) the State Party’s limited acceptance of the applicability of the Convention to the actions of its forces abroad, in particular its explanation “that those parts of the Convention which are applicable only in respect of territory under the jurisdiction of a State party cannot be applicable in relation to actions of the United Kingdom in Afghanistan and Iraq”; the Committee observes that the Convention protections extend to all territories under the jurisdiction of a State party and considers that this principle includes all areas under the de facto effective control of the State party’s authorities.’ Similarly, the Committee on the Rights of the Child considered that the CRC applied to the Occupied Palestinian Territories, and also seemed to have though that it applied to Israeli army activities with regard to de-mining in Southern Lebanon—see Concluding Observations of the Committee on the Rights of the Child: Israel, UN Doc. CRC/C/15/Add.195, 4 October 2002, paras 2, 5, 57–8.
(29) Legal Consequences of the Construction of a Wall in the Occupied Palestinian Territory, Advisory Opinion, 9 July 2004, ICJ Reports 2004, at 136, paras 109–13; Armed Activities on the Territory of the Congo (Congo v. Uganda), Judgment, 19 December 2005, ICJ Reports 2005, at 168, paras 179, 216–17. The primary basis for extraterritorial application in both cases seems to have been the occupied status of the territories in question—see especially Wall, para. 112: ‘the Court cannot accept Israel’s view. It would also observe that the territories occupied by Israel have for over 37 years been subject to its territorial jurisdiction as the occupying Power.’ However, the Court did not in any way limit the possibility of extraterritorial application to the spatial model alone.
(30) For the best general overview of the spatial model, see R. Wilde, ‘Triggering State Obligations Extraterritorially: The Spatial Test in Certain Human Rights Treaties’, (2007) 40 Israel L Rev 503.
(31) See, e.g., Memorandum for William J. Haynes, II, from Jay S. Bybee, re The President’s power as Commander in Chief to transfer captured terrorists to the control and custody of foreign nationals, US Department of Justice, Office of Legal Counsel, 13 March 2002, available at 〈http://www.justice.gov/opa/documents/memorandumpresidentpower03132002.pdf〉, at 22–5.
(32) Memorandum for John A. Rizo from Steven G. Bradbury, re Application of United States Obligations Under Article 16 of the Convention Against Torture to Certain Techniques that May Be Used in the Interrogation of High Value al Qaeda Detainees, US Department of Justice, Office of Legal Counsel, 30 May 2005, available at 〈http://www.fas.org/irp/agency/doj/olc/article16.pdf〉.
(34) Alternatively, if this argument fails, the OLC turns to the US reservation to Art. 16 CAT, providing that cruel, inhuman, and degrading treatment within the meaning of the CAT will be understood as cruel and unusual punishment as defined by the applicable provisions of the Bill of Rights of the US Constitution. Since, in the OLC’s view, the US Constitution does not apply extraterritorially to aliens, neither does the CAT.
(35) R. (Al-Saadoon and Mufdhi) v. Secretary of State for Defence  EWCA Civ 7 (hereinafter Al-Saadoon CA); R. (Al-Saadoon and Mufdhi) v. Secretary of State for Defence  EWHC 3098 (hereinafter Al-Saadoon HC).
(37) R. (Al-Jedda) v. Secretary of State for Defence  UKHL 58,  1 AC 332; see further M. Milanovic, ‘Norm Conflict in International Law: Whither Human Rights?’, (2009) 20 Duke J. Comp. & Int’l L. 69.
(38) Al-Saadoon HC, paras 24–31.
(39) R. (Al-Skeini and Others) v. Secretary of State for Defence  UKHL 26,  3 WLR 33,  3 All ER 685 (hereinafter Al-Skeini HL); R. (Al-Skeini and Others) v. Secretary of State for Defence  EWCA Civ 1609,  QB 140 (hereinafter Al-Skeini CA).
(40) For commentary on the UK courts’ reasoning regarding the extraterritorial applicability issues in Al-Skeini, see R. Wilde’s note in (2008) 102 AJIL 628, as well as T. Thienel, ‘The ECHR in Iraq: The Judgment of the House of Lords in R. (Al-Skeini) v Secretary of State for Defence’ (2008) 6 JICJ 115.
(41) Al-Saadoon HC, para. 82.
(43) Al-Saadoon HC, paras 89–96.
(44) Al-Saadoon CA, paras 32–40.
(45) See T. Thienel, ‘Cooperation in Iraq and the ECHR: An Awful Epilogue’, Invisible College, 21 January 2009, available at 〈http://invisiblecollege.weblog.leidenuniv.nl/2009/01/21/cooperation-in-iraq-and-the-echr-an-awfu〉.
(46) Al-Saadoon and Mufdhi v. United Kingdom (dec.), App. No. 61498/08, 3 July 2009, paras 84–5.
(48) Note also that the Court in Al-Saadoon relies on the former European Commission’s decision in Hess v. the United Kingdom (dec.), App. No. 6231/73, 2 DR 72, 28 May 1975, which at first glance concerned the lawfulness of the continued detention of the former Nazi leader Rudolf Hess in the Spandau prison in UK-occupied Berlin, who was held after the Nuremberg trials under the joint authority of the four occupying powers, of which only the Soviets continued to oppose Hess’s release. However, it was Hess’s wife who actually lodged the application, complaining that her rights were violated. The threshold question was therefore whether Mrs Hess was within UK jurisdiction, but the Commission examined instead whether Hess himself fell within the scope of Article 1:
The Commission first observes that in the present case the exercise of authority by the respondent Government takes place not in the territory of the United Kingdom but outside its territory, in Berlin.…The Commission is of the opinion that there is in principle, from a legal point of view, no reason why acts of the British authorities in Berlin should not entail the liability of the United Kingdom under the Convention. The question therefore arises whether the Commission, in the particular circumstances of this case, is competent to receive an application against the United Kingdom in respect of the continued imprisonment of Rudolf Hess…
The Commission concludes that the responsibility for the prison at Spandau, and for the continued imprisonment of Rudolf Hess, is exercised on a Four Power basis and that the United Kingdom acts only as a partner in the joint responsibility which it shares with the three other Powers. The Commission is of the opinion that the joint authority cannot be divided into four separate jurisdictions and that therefore the United Kingdom’s participation in the exercise of the joint authority and consequently in the administration and supervision of Spandau Prison is not a matter ‘within the jurisdiction’ of the United Kingdom, within the meaning of Art. 1 of the Convention.
Ibid., at 73–4.
This was undoubtedly a complex case. Even if we accept that the threshold issue was whether Mr Hess, rather than Mrs Hess, was within UK jurisdiction, I can only note that several different concepts are mixed up in the Commission’s approach, namely attribution and jurisdiction as the threshold question; the content of the substantive obligation if the threshold is reached; and the responsibility vel non of the UK in connection with the acts of other states. Then there is also the question of possible norm conflict between the UK’s obligations to the other three occupying powers, and its obligations under the ECHR. At any rate, in my view the Commission’s decision is as such not particularly instructive.
(49) For an excellent recent defence of the wall of separation between the jus in bello and the jus ad bellum, see R. Sloane, ‘The Cost of Conflation: Preserving the Dualism of Jus Ad Bellum and Jus in Bello in the Contemporary Law of War’, (2009) 34 Yale J. Int’l L. 47.
(50) Similarly, see N. Lubell, Extraterritorial Use of Force Against Non-State Actors (Oxford University Press, 2010), at 214–15.
(51) Al-Saadoon (dec.), para. 88.
(52) See also A. Zimmermann, ‘Extraterritorial Application of Human Rights Treaties—The Case of Israel and the Palestinian Territories Revisited’, in I. Buffard, J. Crawford, A. Pellet, and S. Wittich (eds), International Law Between Universalism and Fragmentation: Festschrift in Honour of Gerhard Hafner (Brill, 2008), 747, at 758.
(53) Loizidou (merits), para. 56, internal citations omitted.
(54) Cyprus v. Turkey, para. 77.
(56) Cyprus v. Turkey, para. 77.
(57) Bankovic and Others v. Belgium and Others [GC] (dec.), App. No. 52207/99, 12 December 2001, para. 71.
(58) See Wilde, above note 30, at 516.
(59) Issa v. Turkey, App. No. 31821/96, Judgment, 16 November 2004, para. 71.
(61) Ilascu and others v. Moldova and Russia [GC], App. No. 48787/99, Judgment, 8 July 2004, para. 382.
(65) Similarly, on the importance of potential for control, specifically in the Israeli/Palestinian context, see O. Ben-Naftali and Y. Shany, ‘Living in Denial: The Application of Human Rights Treaties in the Occupied Territories’, (2003) 37 Israel L Rev 17, at 63–4.
(66) Similarly, see Ilascu, Dissenting Opinion of Judge Bratza.
(69) See also Wilde, above note 30.
(70) See, e.g., Y. Dinstein, The International Law of Belligerent Occupation (Cambridge University Press, 2009), at 40, 42 et seq.
(74) See further M. Milanovic, ‘Is Gaza Still Occupied by Israel?,’ EJIL: Talk!, 1 March 2009, available at 〈http://www.ejiltalk.org/is-gaza-still-occupied-by-israel/〉.
(75) United States of America v. Wilhelm List et al., 8 Law Reports of Trials of Major War Criminals 38, 55–6.
(76) Prosecutor v. Naletilic, IT-98-34-T, Trial Chamber Judgment, 31 March 2003, para. 217.
(77) See, e.g., Report of the United Nations Fact Finding Mission on the Gaza Conflict (Goldstone Report), A/HRC/12/48, 15 September 2009, paras 274–80, esp. para. 276: ‘Israel has without doubt at all times relevant to the mandate of the Mission exercised effective control over the Gaza Strip. The Mission is of the view that the circumstances of this control establish that the Gaza Strip remains occupied by Israel’ (emphasis added).
(78) Cf. Arts 69 and 70 of Additional Protocol I.
(79) Gaber Al-Bassiouni v. Prime Minister, H.C.J. 9132/07, Judgment, 30 January 2008.
(80) See Y. Shany, ‘The Law Applicable to Non-Occupied Gaza: A Comment on Bassiouni v. Prime Minister of Israel’, Hebrew University International Law Research Paper No. 13-09, 27 February 2009, available at 〈http://ssrn.com/abstract=1350307〉, at 10 et seq.
(81) See also Wilde, above note 30, at 518 et seq.
(82) Al-Skeini CA, paras 124 and 125, citations omitted.
(84) See Al-Skeini HL, para. 83 (per Lord Rodger):
I would not consider that the United Kingdom was in effective control of Basra and the surrounding area for purposes of jurisdiction under article 1 of the Convention at the relevant time. Leaving the other rights and freedoms on one side, with all its troops doing their best, the United Kingdom did not even have the kind of control of Basra and the surrounding area which would have allowed it to discharge the obligations, including the positive obligations, of a contracting state under article 2, as described, for instance in Osman v. United Kingdom (1998) 29 EHRR 245, 305, paras 115–116.
See also Al-Skeini HL, para. 90 (per Baroness Hale); para. 97 (per Lord Carswell); para. 129 (per Lord Brown).
(85) See Al-Skeini CA, paras 129–40 (per Brooke LJ); Al-Skeini HL, para. 128 (per Lord Brown): ‘Bear in mind too the rigour with which the Court applies the Convention, well exemplified by the series of cases from the conflict zone of south eastern Turkey in which, the state’s difficulties notwithstanding, no dilution has been permitted of the investigative obligations arising under articles 2 and 3.’
(86) Note, however, that Lord Justice Sedley’s preferred solution was simply that the UK had control over the actions of its own troops, and that these troops brought individuals within their authority and control into the UK’s jurisdiction or that no jurisdiction threshold applied, but that he thought this option was precluded under Bankovic (which it was).
(87) See also Wilde, above note 30, at 518–19.
(88) See Dinstein, above note 70, at 42–5.
(89) Loizidou, para. 62.
(90) See UNSC Res. 1483 (2003), with the Council in a preambular paragraph ‘recognizing the specific authorities, responsibilities, and obligations under applicable international law of these states as occupying powers under unified command (the “Authority”)’.
(91) See UNSC Res. 1546 (2004), op. para. 2, in which the Council welcomed that ‘by 30 June 2004, the occupation will end and the Coalition Provisional Authority will cease to exist, and that Iraq will reassert its full Sovereignty’. It is unclear whether the Council was merely recognizing the fact of the occupation ending, even though whether this factual appraisal may have been inaccurate, or whether the Council actually terminated the occupation on its own authority. See further R. Kolb, ‘Occupation in Iraq since 2003 and the Powers of the UN Security Council’, (2008) 90 IRRC 29; M. Sassoli, ‘Legislation and Maintenance of Public Order and Civil Life by Occupying Powers’, (2005) 16 EJIL 661.
(92) The occupation of Germany by the Four Powers presented a similar factual scenario. See also the discussion of the Hess case in note 48 above.
(93) Behrami and Behrami v. France, Saramati v. France, Germany and Norway [GC] (dec.), App. Nos 71412/01 and 78166/01, 2 May 2007.
(94) ILC, Draft Articles on the Responsibility of International Organizations, as adopted on first readings, UN Doc. A/64/10 (2009) (hereinafter ILC DARIO).
(95) See, e.g. A. Sari, ‘Jurisdiction and International Responsibility in Peace Support Operations: The Behrami and Saramati Cases’, (2008) 8 HRLR 151; K. Mujezinovic Larsen, ‘Attribution of Conduct in Peace Operations: The “Ultimate Authority and Control” Test’, (2008) 19 EJIL 509.
(96) See M. Milanovic and T. Papic, ‘As Bad As It Gets: The European Court of Human Rights’ Behrami and Saramati Decision and General International Law’, (2009) 58 ICLQ 267.
(97) See G. Gaja, Seventh Report on Responsibility of International Organizations, UN Doc. A/CN.4/610, 27 March 2009, at 10; ILC DARIO, at 62 et seq.
(99) See generally E. Benvenisti, ‘Occupation, Pacific’, in Max Planck Encyclopedia of Public International Law, available at 〈http://www.mpepil.com〉. See also G. Fox, Humanitarian Occupation (Cambridge University Press, 2008).
(100) See, e.g., Dinstein, above note 70, at 35–7.
(101) See, e.g., ‘Iraq’s Death Squads,’ Washington Post, 4 December 2005, available at 〈http://www.washingtonpost.com/wp-dyn/content/article/2005/12/03/AR2005120300881.html〉; ‘New Order, Same Abuses: Unlawful Detentions and Torture in Iraq’, Amnesty International, 13 September 2010, available at 〈http://www.amnestyusa.org/pdf/Iraqrpt_newordersameabuses.pdf〉; ‘Iraq: Detainees Describe Torture in Secret Jail’, Human Rights Watch, 27 April 2010, at 〈http://www.hrw.org/en/news/2010/04/27/iraq-detainees-describe-torture-secret-jail〉.
(103) ‘Each Party shall permit visits, in accordance with this Convention, to any place within its jurisdiction where persons are deprived of their liberty by a public authority’ (emphasis added).
(104) ‘Each State Party shall allow visits, in accordance with the present Protocol, by the mechanisms referred to in articles 2 and 3 to any place under its jurisdiction and control where persons are or may be deprived of their liberty’ (emphasis added).
(105) Committee Against Torture, General Comment No. 2, UN Doc. CAT/C/GC/2, 24 January 2008, para. 7 (emphasis added).
(107) Al-Saadoon (dec.), para. 88, citations omitted.
(108) See Al-Skeini HL, para. 97 (per Lord Carswell), para. 132 (per Lord Brown). See also Al-Skeini DC, para. 287: ‘a British military prison, operating in Iraq with the consent of the Iraqi sovereign authorities, and containing arrested suspects, falls within even a narrowly limited exception exemplified by embassies, consulates, vessels and aircraft’.
(109) See also T. Thienel, ‘The ECHR in Iraq’, (2008) 6 JICJ 115, at 127.
(110) See also above note 50 and accompanying text.
(112) Bankovic, para. 73.
(116) Cyprus v. Turkey (dec.), App. Nos 6780/74 and 6950/75, 26 May 1975.
(118) See X v. United Kingdom (dec.), App. No. 7547/76, 15 December 1977, concerning the acts of British consular officials in Jordan. Though the application was held to be manifestly ill-founded, the Commission thought that the applicant was within UK jurisdiction even though he was outside UK territory:
It is clear, in this respect, from the constant jurisprudence of the Commission that authorised agents of a State, including diplomatic or consular agents, bring other persons or property within the jurisdiction of that State to the extent that they exercise authority over such persons or property. Insofar as they affect such persons or property by their acts or omissions, the responsibility of the State is engaged.
See also WM v. Denmark (dec.), App. No. 17392/90, 14 October 1993; Gill and Malone v. The Netherlands and the United Kingdom (dec.), App. No. 24001/94, 11 April 1996. See further below, Section 3.
(119) See generally C. Bell, ‘Capitulations’, as well as P. Czubik and P. Szwedo, ‘Consular Jurisdiction’, both in Max Planck Encyclopedia of Public International Law, available at 〈http://www.mpepil.com〉, and the sources cited therein.
(120) See above, Section 2.C.1.
(121) See generally H. Fox, The Law of State Immunity (Oxford University Press, 2nd edn, 2008), at 665 et seq.
(124) See generally A. Mowbray, The Development of Positive Obligations under the European Convention on Human Rights by the European Court of Human Rights (Hart Publishing, 2004), at 7 et seq.
(126) In fact, this is what the ICJ did in the Tehran Hostages case, where it remarked that
[n]aturally, the observance of this principle [of inviolability] does not mean—and this the Applicant Government expressly acknowledges—that a diplomatic agent caught in the act of committing an assault or other offence may not, on occasion, be briefly arrested by the police of the receiving State in order to prevent the commission of the particular crime.
United States Diplomatic and Consular Staff in Tehran (United States v. Iran), Judgment, ICJ Reports 1980, 3, at 40, para. 86.
(127) See also M. Shaw, International Law (Cambridge University Press, 6th edn, 2009), at 754; E. Denza, Diplomatic Law (Oxford University Press, 3rd edn, 2008), at 150: ‘In the last resort, however, it cannot be excluded that entry without the consent of the sending State may be justified in international law by the need to protect human life.’
(128) Similarly, see Zimmermann, above note 52, at 759–60.
(130) Rigopoulos v. Spain (dec.), App. No. 37388/97, 12 January 1999.
(131) Xhavara c. l’Italie et l’Albanie (dec.), App. No. 39473/98, 11 January 2001.
(132) Medvedyev and Others v. France, App. No. 3394/03, Judgment, 10 July 2008.
(134) See also T. Thienel, ‘Oral Argument in Medvedyev v. France’, Invisible College Blog, 13 May 2008, available at 〈http://invisiblecollege.weblog.leidenuniv.nl/2008/05/13/oral-argument-in-medvedyev-v-france〉.
(135) Medvedyev and Others v. France [GC], App. No. 3394/03, Judgment, 29 March 2010.
(138) Similarly, see D. Guilfoyle, ‘ECHR Rights at Sea: Medvedyev and others v. France’, EJIL: Talk!, 19 April 2010, available at 〈http://www.ejiltalk.org/echr-rights-at-sea-medvedyev-and-others-v-france/〉.
(139) Freda v. Italy (dec.), App. No. 8916/80, 7 October 1980.
(140) Illich Sanchez Ramirez v. France (dec.), App. No. 28780/95, 24 June 1996. Note that the later decisions concerning Carlos are correctly labelled as Ramirez Sanchez, rather than Sanchez Ramirez, but I will use the Court’s own nomenclature.
(141) The other being Ramirez Sanchez v. France [GC], App. No. 59450/00, Judgment, 4 July 2006.
(142) Öcalan v. Turkey (dec.), App. No. 46221/99, 14 December 2000.
(143) Öcalan v. Turkey, App. No. 46221/99, Judgment, 12 March 2003, para. 93.
(144) Öcalan v. Turkey [GC], App. No. 46221/99, Judgment, 12 May 2005, para. 91.
(145) The best recent example is Women on Waves and Others v. Portugal, App. No. 31276/05, Judgment, 13 January 2009. The applicants were reproductive rights NGOs which hired a ship to take them to Portugal, for the purpose of staging activities that would promote the decriminalization of abortion in Portugal. The ship was intercepted by a Portuguese warship, which prohibited it from entering Portuguese territorial waters. The Court unanimously found that this amounted to a violation of the Article 10 guarantee of the freedom of expression. Thus, even though the event took place on the high seas and the applicants never entered Portugal, the Court did not think that an Article 1 ECHR issue arose.
(146) CAT General Comment No. 2, above note 105, para. 16.
(147) P.K. et al. v. Spain, Communication No. 323/2007, UN Doc. CAT/C/41/D/323/2007, 21 November 2008 (hereinafter the Marine I case).
(149) See also K. Wouters and M. Den Heijer, ‘The Marine I Case: A Comment’, (2009) 22 International Journal of Refugee Law 1, 8–11.
(150) For an excellent overview of the various issues, as well as for factual background, see D. Guilfoyle, ‘Counter-Piracy Law Enforcement and Human Rights’, (2010) 59 ICLQ 141.
(151) See, e.g., Wouters and Den Heijer, above note 149, at 9.
(152) On extraterritorial refoulement in the context of refugee law, see J. Hathaway, The Rights of Refugees under International Law (Cambridge University Press, 2005), 335–42, as well as UNHCR, Advisory Opinion on the Extraterritorial Application of Non-Refoulement Obligations under the 1951 Convention relating to the Status of Refugees and its 1967 Protocol, 26 January 2007, available at 〈http://www.unhcr.org/refworld/docid/45f17a1a4.html〉.
(153) App. No. 27765/09.
(154) Italy has also been criticized by the CPT for its ‘push-back’ of migrants as constituting a violation of its non-refoulement obligations, even if it took place on ships outside Italy’s territorial sea—see Report to the Italian Government, CPT/Inf (2010) 14, 28 April 2010, esp. para. 29. In that regard, the CPT seems to have adopted a personal conception of state jurisdiction: ‘Italy is bound by the principle of non-refoulement wherever it exercises its jurisdiction, which includes via its personnel and vessels engaged in border protection or rescue at sea, even when operating outside its territory. Moreover, all persons coming within Italy’s jurisdiction should be afforded an appropriate opportunity and facilities to seek international protection.’ Ibid., para. 49.
(155) See, e.g., Convention for the Suppression of Unlawful Seizure of Aircraft [Hijacking Convention], 860 UNTS 105, entered into force 14 October 1971, Art. 4(1).
(156) See also Guilfoyle, above note 150, at 155. Similarly, see the Women on Waves case cited in note 145 above.
(157) See also Guilfoyle, above note 150, at 154.
(159) See, in that regard, R. (Hassan) v. Secretary of State for Defence  EWHC 309 (Admin), where the claimants were the family of an individual detained in the initial stages of the invasion by UK forces in Iraq, who transferred him to the custody of US armed forces pursuant to a memorandum of understanding, under which the UK remained the detaining power under the Third Geneva Convention for the individual in question. Applying the spatial model of jurisdiction, the High Court found that, as the detainee was held on a US military base, he was not within the UK’s jurisdiction. As the detaining power, however, and on the basis of its memorandum of understanding with the US, the UK did retain a legal authority over the detainee, which could have sufficed to bring him within its jurisdiction under the personal model. For a critique of Hassan in that vein, see H. King, ‘Unravelling the Extraterritorial Riddle: An Analysis of R. (Hassan) v. Secretary of State for Defence’, (2009) 7 JICJ 633.
(160) See Section 1.E.
(161) See Section 1.F.
(163) Most ably and prominently by R. Lawson, ‘Life after Bankovic: On the Extraterritorial Application of the European Convention on Human Rights’, in Coomans and Kamminga 83.
(164) Note that this would not mean, as argued by O’Boyle, that the concept of jurisdiction would be conflated with that of attribution, in other words that the inquiries under Art. 2(a) and Art. 2(b) ILC ASR would collapse into each other. Rather, the obligation whose violation is being attributed would not be territorially limited, as is the case for example with the Geneva Conventions. See M. O’Boyle, ‘The European Convention on Human Rights and Extraterritorial Jurisdiction: A Comment on “Life After Bankovic”’, in Coomans and Kamminga 125, and Chapter II, Section 3 above.
(165) Bankovic, para. 75.
(166) Lopez Burgos v. Uruguay, Communication No. R.12/52, UN Doc. Supp. No. 40 (A/36/40) at 176 (1981).
(167) Celiberti de Casariego v. Uruguay, Communication No. 56/1979, UN Doc. CCPR/C/OP/1 at 92 (1984).
(168) Lopez Burgos, paras 12.1–12.3.
(169) Celiberti, paras 10.1–10.3.
(170) See also M. Nowak, CCPR Commentary (Engel, 2nd revised edn, 2005), at 860.
(172) Similarly, see M. Scheinin, ‘Extraterritorial Effect of the International Covenant on Civil and Political Rights’, in Coomans and Kamminga 73, at 75, n. 7; Lawson, above note 163, at 94.
(173) One exception is the so-called passport cases, dealing with the refusal of a state’s diplomatic or consular mission abroad to issue a passport to the state’s citizen—see, e.g., Vidal Martins v. Uruguay, Communication No. R.13/57, U.N. Doc. Supp. No. 40 (A/37/40) at 157 (1982), as well as notes 260 and 261 below and accompanying text. For commentary, see Nowak, above note 170, at 861.
(174) See above note 28.
(175) Human Rights Committee, General Comment No. 31, Nature of the General Legal Obligation on States Parties to the Covenant, UN Doc. CCPR/C/21/Rev.1/Add.13 (2004), para. 10 (emphasis added).
(176) T. Buergenthal, ‘To Respect and to Ensure: State Obligations and Permissible Derogations’, in L. Henkin (ed.), The International Bill of Rights: The Covenant on Civil and Political Rights (Columbia University Press, 1981), at 72–91. See also T. Meron, ‘Extraterritoriality of Human Rights Treaties,’ (1995) 89 AJIL 78, as well as the works cited in Chapter III, above note 11.
(177) The Committee has continued doing so in its examination of state party reports—see, e.g., Concluding Observations of the Human Rights Committee—United Kingdom, UN Doc. CCPR/C/GBR/CO/6, 18 July 2008, para. 14 (requesting of the UK to ‘state clearly that the Covenant applies to all individuals who are subject to its jurisdiction and control’).
(179) Committee Against Torture, General Comment No. 2, UN Doc. CAT/C/GC/2, 24 January 2008, para. 16.
(180) Indeed, the Committee seems to have adopted the personal model precisely because of Guantanamo—see Committee Against Torture, Conclusions and Recommendations: United States, UN Doc. CAT/C/USA/CO/2, 25 July 2006, para. 15: ‘The State party should recognize and ensure that the provisions of the Convention expressed as applicable to “territory under the State party’s jurisdiction” apply to, and are fully enjoyed, by all persons under the effective control of its authorities, of whichever type, wherever located in the world’ (emphasis added).
(181) See generally C. Cerna, ‘Extraterritorial Application of the Human Rights Instruments of the Inter-American System’, in Coomans and Kamminga 141.
(182) Saldaño v. Argentina, Report No. 38/99, Annual Report of the IACHR 1998, esp. paras 15–20.
(186) Coard et al. v. United States, Case No. 10.951, Report No. 109/99, Annual Report of the IACHR 1999, para. 37 (emphasis added).
(187) See Armando Alejandre Jr and Others v. Cuba (‘Brothers to the Rescue’), Case No. 11.589, Report No. 86/99, 29 September 1999, para. 23. See further J. Cerone, ‘The Application of Regional Human Rights Law Beyond Regional Frontiers: The Inter-American Commission on Human Rights and US Activities in Iraq’, ASIL Insights, 25 October 2005, available at 〈http://www.asil.org/insights051025.cfm〉.
(188) See, e.g., Decision on Request for Precautionary Measures (Detainees at Guantánamo Bay), 12 March 2002, (2002) 41 ILM 532. For more extensive discussion, see Cerna, above note 181, at 159 et seq; Gondek, at 217–19.
(189) Cyprus v. Turkey (dec.), App. Nos 6780/74 and 6950/75, 26 May 1975.
(190) Cyprus v. Turkey (dec.), App. Nos 6780/74 and 6950/75, 26 May 1975, at 136, para. 8.
(191) See, e.g., Stocké v. Germany, App. No. 11755/85, Commission Report, 12 October 1989, para. 166. See also Gondek, at 126.
(192) Sarah Miller suggests that the first Cyprus case can somehow be interpreted as being about control over territory, rather than individuals—see S. Miller, ‘Revisiting Extraterritorial Jurisdiction: A Territorial Justification for Extraterritorial Jurisdiction under the European Convention,’ (2010) 20 EJIL 1223, at 1237, arguing that ‘the essential predicate was that Turkey had established so great a presence in Cyprus that it was in a position effectively to control administration in the region. Likewise, Banković emphatically suggests that one cannot conflate questions of effective control of a region with subsequent attribution of acts to state officials’. With respect, Miller is simply mistaken. There is nothing in the Commission’s first decision on Cyprus, nor in its progeny, that suggests that the Commission saw ‘jurisdiction’ as being about control over territory, rather than individuals. That strand of Strasbourg jurisprudence begins in the Court’s Loizidou judgment many years later. It is likewise simply anachronistic to read the Commission’s case law in light of Bankovic. While it is true that, under the Commission’s approach, ECHR states parties would be responsible for violating the ECHR whenever state agents exercise authority and control over individuals, this does not mean that the issues of attribution and the existence of breach of obligation are conflated, just that the obligation is not limited territorially.
(193) Bankovic, para. 75.
(194) Bankovic, para. 78.
(195) Issa, para. 71.
(196) This was also most notably the situation in Behrami—see further Milanovic and Papic, above note 96, at 293–4.
(197) Saddam Hussein v. Albania and others (dec.), App. No. 23276/04, 14 March 2006. One could question whether the United Kingdom was in a different position than the other respondent states, due to its status as a joint occupying power in Iraq. See above Section 2.C.4.
(198) Pad and Others v. Turkey (dec.), App. No. 60167/00, 28 June 2007.
(200) Isaak and Others v. Turkey (dec.), App. No. 44587/98, 28 September 2006.
(201) Solomou and Others v. Turkey, Judgment, App. No. 36832/97, 24 June 2008, paras 44–5, 51.
(202) Andreou v. Turkey (dec.), App. No. 45653/99, 3 June 2008.
(204) Ben El Mahi and Others v. Denmark (dec.), App. No. 5853/06, 11 December 2006.
(205) Medvedyev [GC], para. 64.
(207) See Al-Skeini CA, para. 49 (per Brooke LJ), distinguishing between the spatial and personal models, with the latter being referred to as ‘state agent authority’, or SAA.
(213) See also Scheinin, above note 172, at 77–8; Lubell, above note 50, at 221–4.
(214) In that regard, one could perhaps draw a helpful analogy to the interpretation of the Fourth Amendment to the US Constitution, which prohibits ‘unreasonable searches and seizures’. The US Supreme Court has thus held not only that ‘[w]henever an officer restrains the freedom of a person to walk away, he has seized that person, but also that ‘there can be no question that apprehension by the use of deadly force is a seizure subject to the reasonableness requirement of the Fourth Amendment’. Tennessee v. Garner, 471 U.S. 1, 7 (1985). See also ibid., at 25 (O’Connor J., dissenting).
(215) See above, Section 3.B.4.
(216) Issa, para. 71.
(217) See also Lubell, above note 50, at 224.
(218) With this particular example, the incentive to kill rather than capture was created in great part by the cautious inroads of US courts into the Executive’s detention policies abroad, as with Guantanamo, and pending litigation regarding Bagram—see, e.g., ‘Special Report: How the White House learned to love the drone’, Reuters, 18 May 2010, available at 〈http://www.reuters.com/article/idUSTRE64H5SL20100518〉.
(220) See above, Section 2.C.3.
(221) See above, Section 2.D.
(222) This issue was manifest in the parties’ arguments in Al-Skeini—see Al-Skeini CA, para. 183 (per Sedley LJ):
[H]ow, Mr Rabinder Singh QC asks, can one draw a rational line at the prison wall, so that if Mr Mousa had been beaten to death by troops in the hotel where he was arrested rather than in the prison to which he was taken, there would have been no violation of his Convention rights? The logic of Mr Greenwood’s submission is that, bar the special status of the military prison in international law, there can be no extra-territorial reach to the ECHR. The logic of Mr Singh’s submission is that, given that the Convention reaches the prison, by parity of reasoning it reaches everywhere in Iraq where British forces are the only functioning form of government. That at least is his narrower submission, based on what he calls effective control authority (ECA). His broader and preferred submission is that the Convention reaches everywhere that agents of a state signatory operate (SAA).
(223) See above, Section 2.D.
(224) R. (Gentle) v. Prime Minister  1 AC 1356.
(227) R. (Smith) v. Secretary of State for Defence  EWHC 694 (Admin), para. 12.
(228) R. (Smith) v. Secretary of State for Defence  EWCA Civ 441, paras 13–15.
(230) R. (Smith) v. Secretary of State for Defence  UKSC 29.
(235) See A. Bailin, ‘Case preview—R (Smith) v Secretary of State for Defence—on appeal from  EWCA Civ 441’, UKSCblog, 13 March 2010, available at 〈http://ukscblog.com/case-preview-%E2%80%93-r-smith-v-secretary-of-state-for-defence-on-appeal-from-2009-ewca-civ-441〉.
(236) Cyprus v. Turkey (dec.), App. Nos 6780/74 and 6950/75, 26 May 1975, at 136, para. 8 (emphasis added). Similarly, see Saldaño v. Argentina, paras 20–2.
(239) The irony is that before the progress of internalization of human rights has progressed to the extent that it has today, soldiers were considered to have waived the rights that they would have had as civilians, and had no recourse to judicial review merely because of their membership in the armed forces. Similarly, it has been a point of some dispute within UK academic literature whether ‘public authorities’ within the meaning of s. 6 of the Human Rights Act 1998 could themselves enjoy rights under the Act against other public authorities. See further D. Oliver, ‘The Frontiers of the State: Public Authorities and Public Functions under the Human Rights Act’, (2000) PL 476, at 491–2; H. Davis, ‘Public Authorities as ‘Victims’ under the Human Rights Act’, (2005) 64 CLJ 315.
(240) See also B. Silverstone, ‘R. (on the application of Smith) v Oxfordshire Assistant Deputy Coroner: Human Rights and the Armed Forces’, (2009) 4 EHRLR 566, esp. at 576.
(241) Issa, para. 71 (emphasis added). See also Pad and others v. Turkey (dec.), para. 53; Isaak (dec.), at A.2; Solomou, para. 45; Andreou (dec.), at A.3; Ben El Mahi (dec.), section ‘The Law’.
(242) Lopez Burgos, para. 12.3 (emphasis added).
(244) This is, for example, what was argued by the respondent governments in Bankovic, para. 36:
As to the precise meaning of ‘jurisdiction’, they suggest that it should be interpreted in accordance with the ordinary and well-established meaning of that term in public international law. The exercise of ‘jurisdiction’ therefore involves the assertion or exercise of legal authority, actual or purported, over persons owing some form of allegiance to that State or who have been brought within that State’s control (emphasis added).
(245) See, e.g., Sejdovic v. Italy [GC], App. No. 56581/00, Judgment, 1 March 2006.
(247) Martin v. United Kingdom, App. No. 40426/98, Judgment, 24 October 2006.
(248) Of interest may be a similar US case, where a US court sitting in West Berlin was trying an East German national for hostage-taking and the hijacking of an airplane which he used to defect to West Germany. In this case, however, the extraterritoriality issue was explicitly raised, with the US prosecutors arguing that the US Constitution and its guarantees of due process and jury trial simply did not apply to a US court sitting in West Berlin, which was even obliged to follow instructions issued by the State Department. The presiding judge was quite understandably appalled, and promptly went on to empanel a jury composed of West Berliners. The judge later wrote a book about the trial, which was then made into a decent, if slightly preachy low-budget movie (Judgment in Berlin, 1988), remarkably starring Martin Sheen and Sean Penn. See also K. Raustiala, Does the Constitution Follow the Flag? The Evolution of Territoriality in American Law (Oxford University Press, 2009), at 151–4.
(249) Mullai and Others v. Albania, App. No. 9074/07, Judgment, 23 March 2010.
(250) Vrbica v. Croatia, App. No. 32540/05, Judgment, 1 April 2010.
(251) Ibrahima Gueye et al. v. France, Communication No. 196/1983, U.N. Doc. Supp. No. 40 (A/44/40) at 189 (1989). While examining the merits, the Committee only cryptically observed that ‘the authors are not generally subject to French jurisdiction, except that they rely on French legislation in relation to the amount of their pension rights’. Ibid., para. 9.4.
(252) Similarly, see El Orabi v. France (dec.), App. No. 20672/05, 20 April 2010, where the European Court declared inadmissible the application of an Algerian national living in Algeria who was married to an Algerian former member of the French armed forces, whose request for a survivor’s pension was denied. The Court declared the application manifestly ill-founded, but it did not say that there was any Article 1 issue in the case, even though both the applicant and her husband were residing in Algeria at the time of her husband’s death. See also Carson and Others v. the United Kingdom [GC], App. No. 42184/05, Judgment, 16 March 2010, similarly dealing with UK pensions of persons residing outside the UK, where likewise no jurisdiction issue was raised.
(253) We could also consider the situation where it is not the state which initiates legal proceedings against an individual outside its territory, but when it is rather that individual who does so. Thus, in Markovic and Others v. Italy [GC], App. No. 1398/03, Judgment, 14 December 2006, dealing with domestic proceedings initiated in Italy by the victims of the NATO bombing of the Belgrade TV station at issue in Bankovic, the European Court held that (para. 54):
[e]ven though the extraterritorial nature of the events alleged to have been at the origin of an action may have an effect on the applicability of Article 6 and the final outcome of the proceedings, it cannot under any circumstances affect the jurisdiction ratione loci and ratione personae of the State concerned. If civil proceedings are brought in the domestic courts, the State is required by Article 1 of the Convention to secure in those proceedings respect for the rights protected by Article 6. The Court considers that once a person brings a civil action in the courts or tribunals of a State, there indisputably exists, without prejudice to the outcome of the proceedings, a ‘jurisdictional link’ for the purposes of Article 1.
(254) Stephens v. Malta (No. 1), App. No. 11956/07, Judgment, 31 March 2009.
(255) Stephens v. Malta (No. 1), App. No. 11956/07, Judgment, paras 48 and 49.
(257) This of course rests on the assumption that the applicability of all of the Convention rests on a single threshold, that of state jurisdiction, which need not be the case.
(258) Bankovic, para. 75.
(259) See Arts 4, 6, and 8 ILC ASR.
(260) See, in that regard, Lichtensztejn v. Uruguay, Communication No. 77/1980, UN Doc. CCPR/C/OP/2 at 102 (1990), which concerned the refusal by Uruguayan authorities to issue a passport to an Uruguayan national residing in Mexico. The question, therefore, is whether the individual was within Uruguay’s jurisdiction, and the Human Rights Committee found that he was (para. 6.1):
When considering the admissibility of the communication, the Human Rights Committee did not accept the State party’s contention that it was not competent to deal with the communication because the author did not fulfil the requirements of article 1 of the Optional Protocol. In that connection, the Committee made the following observations: article 1 applies to individuals subject to the jurisdiction of the State concerned who claim to be victims of a violation by that State of any of the Covenant rights. The issue of a passport to a Uruguayan citizen is clearly a matter within the jurisdiction of the Uruguayan authorities and he is ‘subject to the jurisdiction’ of Uruguay for that purpose. Moreover, a passport is a means of enabling him ‘to leave any country, including his own’, as required by article 12 (2) of the Covenant. Consequently, the Committee found that it followed from the very nature of that right that, in the case of a citizen resident abroad, article 12 (2)imposed obligations both on the State of residence and on the State of nationality and that, therefore, article 2 (1)of the Covenant could not be interpreted as limiting the obligations. of Uruguay under article 12 (2)to citizens within its own territory.
(261) A commentator has suggested that factual conceptions of the word ‘jurisdiction’ in human rights treaties, such as the one that I advocate in this study, cannot explain cases in which jurisdiction was found to exist because of the exercise of a legal power by a state, such as the issuance of a passport as in the Lichtensztejn case cited in n. 260 above. In his view, such cases can only be explained if the word ‘jurisdiction’ referred both to a state’s factual control over territory and individuals and to its lawful competence in international law—see H. King, ‘The Extraterritorial Human Rights Obligations of States’, (2009) 9 HRLR 521, at 537 et seq. To my mind, such an approach is unsatisfactory for two simple reasons. First, it fails to explain how semantically, as a matter of treaty interpretation, a single word—‘jurisdiction’—can mean two radically different things, both a state’s act affecting an individual and its right to so act. Secondly, if the personal model of jurisdiction is accepted as valid, it is entirely unnecessary for its application to require that a state exercise its lawful competence in international law to prescribe legal rules and enforce them. It is enough, for example, that a state has subjected an individual to a trial in absentia. Whether it had a right to do so in international law—in other words, whether it violated, or not, the sovereignty of another state when it did so—is to my mind perfectly irrelevant. After all, why should the applicability of fair trial guarantees depend on whether the in absentia trial was an exorbitant exercise of the state’s jurisdiction to prescribe or enforce? And if it does not, then the enquiry into the lawfulness, vel non, of the state’s act vis-à-vis a third state becomes entirely beside the point.
(262) See, e.g., X. and Association Y. v. Italy (dec.), App. No. 8987/80, 6 May 1981.
(263) See, in that regard, Scheinin, above note 172, at 76–8. This does not mean that a state’s authority and control would exist whenever a state could affect an individual through its actions in a more general way (e.g. by cutting exports or development aid). Rather, the act in question would need to be capable of directly violating the treaty—see Lawson, above note 172, at 95, 104.
(264) Al-Skeini HL, para. 127 (per Lord Brown).
(265) See, e.g., T. Buergenthal, ‘To Respect and to Ensure: State Obligations and Permissible Derogations’, in L. Henkin (ed.), The International Bill of Rights: The Covenant on Civil and Political Rights (Columbia University Press, 1981), at 72–91. P. van Dijk et al., Theory and Practice of the European Convention on Human Rights (Intersentia, 4th edn, 2006), at 13; Nowak, above note 170, at 37–41. In the context of socio-economic rights, Eide has in particular developed a more detailed typology of rights, separating the broader obligation to ensure into the obligations to protect and fulfil—see generally ‘The Right to Adequate Food as a Human Right’, Final Report submitted by Asbjørn Eide, Special Raporteur, UN Doc. E/CN.4/Sub.2/1987/23 (1987). I do not believe that such a disaggregated typology would be of direct relevance for the issue of extraterritorial application, so I will not address it any further.
(266) Human Rights Committee, General Comment No. 31, para. 8.
(267) Velasquez Rodriguez Case, Judgment of 29 July 1988, Inter-Am.Ct.H.R. (Ser. C) No. 4 (1988), para. 172.
(268) Similarly, see E. Roxstrom, M. Gibney, and T. Einarsen, ‘The NATO Bombing Case (Bankovic et al. v. Belgium et al.) and the Limits of Western Human Rights Protection’, (2005) 23 B.U. Int’l L.J. 55, at 72 et seq; J. Cerone, ‘Jurisdiction and Power: The Intersection of Human Rights Law & the Law of Non-International Armed Conflict in an Extraterritorial Context’, (2007) 40 Israel L Rev 72, at 122–3.
(270) Al-Skeini CA, para. 197 (per Sedley LJ).
(271) See Lawson, above note 172, at 106.
(272) Osman v. United Kingdom [GC], App. No. 23452/94, Judgment, 28 October 1998, para. 116 (emphasis added).
(273) See, e.g., Cerone, above note 268, at 124; R. Künnemann, ‘Extraterritorial Application of the International Covenant on Economic, Social and Cultural Rights’, in Coomans and Kamminga 201, at 227–99.
(274) Application of the Convention on the Prevention and Punishment of the Crime of Genocide, Judgment, 27 February 2007 (hereinafter Bosnian Genocide merits judgment).
(276) Application of the Convention on the Prevention and Punishment of the Crime of Genocide, Judgment, 27 February 2007 (hereinafter Bosnian Genocide merits judgment), para. 166 (paragraph breaks added).
(280) Bosnian Genocide merits judgment, para. 183.
(281) Bosnian Genocide merits judgment, Separate Opinion of Judge Tomka, para. 67.
(282) For more on the territorial scope of application of the Genocide Convention, see M. Milanovic, ‘Territorial Application of the Convention and State Succession’, in P. Gaeta, ed., The UN Genocide Convention: A Commentary (Oxford University Press, 2009), 473.
(283) See generally Mowbray, above note 124.
(284) R. (Middleton) v. West Somerset Coroner  2 AC 182, para. 3 (citing Strasbourg case law).
(285) Gentle, para. 5 (per Lord Bingham).
(286) Middleton, para. 2 (citing Strasbourg case law).
(287) See, e.g., Selmouni v. France [GC], App. No. 25803/94, Judgment, 28 July 1999.
(288) Even a system which has an extremely limited acceptance of positive obligations generally, such as the one in the United States—see e.g. DeSheaney v. Winnebago, 489 U.S. 189 (1989)—recognizes that the state may accrue positive obligations towards an individual if it acts beforehand so as to restrict or affect the rights or liberty of the individual concerned, for example by incarceration. In the words of the US Supreme Court, ‘elementary principles establish the government’s obligation to provide medical care for those whom it is punishing by incarceration’ as ‘it is but just that the public be required to care for the prisoner, who cannot by reason of the deprivation of his liberty, care for himself’. Estelle v. Gamble, 429 U.S. 97, 102–4 (1976). See also Farmer v. Brennan, 511 U.S. 825 (1994).
(289) Osman, para. 116 (emphasis added).
(290) See above, Section 2.C.3.
(291) I am not denying that the UK might be faced with serious evidentiary and forensic difficulties even when investigating the conduct of its own troops in a territory outside its control, e.g. because it does not have access to the crime scene. In such a situation the prophylactic positive obligation to investigate should again be interpreted flexibly, so that the UK is obliged to do only what it can in fact do. Above all, however, the UK would be expected to put in place reasonable safeguards and procedures before it mounts a military operation that would allow it to investigate allegations of misconduct by its own troops.
(292) See, in that regard, W v. United Kingdom (dec.), App. No. 9348/81, (1983) 32 DR 190, a Commission admissibility decision. The applicant’s husband was killed in the Republic of Ireland, while her brother was killed in Northern Ireland. She complained that the United Kingdom had failed to secure her husband’s and brother’s right to life. With respect to the husband, who was killed in the Republic of Ireland, the Commission declared the application to be incompatible ratione loci (at 199):
The Commission further considers that, in determining its competence ratione loci in relation to the jurisdiction of the United Kingdom, regard must be had to the position, at the relevant time, of the direct victim (i.e. the applicant’s husband) and not of the indirect victim (the applicant herself) of the alleged violation of the Convention. It finds that, at the time of his death at G. Sales Yard, in the Republic of Ireland, the applicant’s husband was not ‘within the jurisdiction’ of the United Kingdom in the sense of Article 1 of the Convention. The Commission has also considered whether any active measures by United Kingdom authorities could have contributed to the murder of the applicant’s husband in the Republic of Ireland. However, it notes that even the applicant has not alleged any such action by these authorities.
Note how the Commission applied a spatial model of jurisdiction—the victim was not within UK’s jurisdiction because the killing took place in Ireland, and therefore the UK had no obligation to investigate it. However, the Commission did entertain the possibility that the result could have been different if the applicant had shown that UK authorities were involved in the murder.
(293) A further problem is that even though Article 15(2) ECHR permits derogations from Article 2 ‘in respect of deaths resulting from lawful acts of war’, which could serve to add more flexibility to an Article 2 analysis in times of armed conflict, no such derogation was made by ECHR states parties engaging in the bombing of Serbia. See also Chapter V below.
(294) See Section 1.F above.
(295) See Section 1.D above.
(296) See, e.g., Opuz v. Turkey, App. No. 33401/02, Judgment, 9 June 2009.
(297) See, e.g., Plattform ‘Ärzte für das Leben’ v. Austria, Judgment, 21 June 1988, Ser. A, No. 139, (1991) 13 EHRR 204.
(298) See, e.g., 97 members of the Gldani Congregation of Jehovah’s Witnesses and 4 others v. Georgia, App. No. 71156/01, Judgment, 3 May 2007.
(300) See also Miller, above note 192 at 1230, arguing that
[t]his uncertainty creates the twin risks that states will either under-estimate the jurisdictional scope of the Convention and violate human rights which might otherwise be protected, or that they will over-estimate the Convention’s reach and refrain from actions which are strategically essential. Either way, the Court’s doctrinal ambivalence prevents signatory states from accurately weighing the legal liabilities associated with particular extraterritorial actions, to the detriment of both human rights protection and security.
(301) See, in that regard, Miller, above note 192, at 1235.
(302) For a similar contextual approach, see Lawson, above note 172, at 120, as well as R. Lawson, ‘Really out of Sight? Issues of Jurisdiction and Control in Situations of Armed Conflict under the ECHR’, in A. Buyse (ed.), Margins of Conflict: The ECHR and Transitions to and from Armed Conflict (Intersentia, 2010), 57; Lubell, above note 50, at 227 et seq; S. Skogly, Beyond National Borders: States’ Human Rights Obligations in International Cooperation (Intersentia, 2006), at 66 et seq, 206.
(303) See Buergenthal, above note 265; Human Rights Committee, General Comment No. 31, para. 10: ‘States Parties are required by article 2, paragraph 1, to respect and to ensure the Covenant rights to all persons who may be within their territory and to all persons subject to their jurisdiction.’
(304) This argument was developed to the fullest in the US second and third periodic reports to the Human Rights Committee, submitted in one document, as UN Doc. CCPR/C/USA/3, 28 November 2005, Annex I.
(305) See M. Dennis, ‘Application of Human Rights Treaties Extraterritorially in Times of Armed Conflict and Military Occupation’, (2005) 99 AJIL 119, at 122 et seq.
(306) See R. Gardiner, Treaty Interpretation (Oxford University Press, 2008), at 178 et seq.
(307) If, for example, I were to ask a guest of mine: ‘would you like some milk and sugar in your coffee?’ we would not normally interpret that question as presenting my guest only with the options of taking both the milk and the sugar or neither, even though ‘and’ is conjunctive in nature. Similarly, if we replaced the ‘and’ in the question with an ‘or’, we would not interpret it as precluding my guest from taking both milk and sugar with his coffee, though the disjunction through an ‘or’ is normally regarded as being exclusive. (I imagine that not even lawyers and logicians ask their guests whether they would like milk ‘and/or’ sugar with their tea or coffee.) The ambiguity of this sentence is not resolved by the grammatical context in which the words were uttered, but by their social context—i.e. by the fact that plenty of people drink their coffee only with milk, or only with sugar, and that it would be quite impolite not to offer these options to one’s guest. Likewise, in treaty interpretation it is the wider context, not merely the grammatical one, taken together with other considerations such as object and purpose, which can help us resolve such ambiguities.
(308) United States v. Fisk, 70 U.S. (3 Wall.) 445, 447 (1865).
(309) N.J. Singer, 1A Statutes and Statutory Construction (West, 6th edn, 2002), § 21:14.
(310) See OLC, Memorandum Opinion for the Counsel to the Vice President: Whether False Statements or Omissions in Iraq’s Weapons of Mass Destruction Declaration Would Constitute a ‘Further Material Breach’ under U.N. Security Council Resolution 1441, 7 December 2002, available at 〈http://www.justice.gov/olc/2002/materialbreach.pdf〉.
(311) See above note 304.
(312) See UN Doc. E/CN.4/SR.138, at 10.
(313) See generally N. Rodley, ‘The Extraterritorial Reach and Applicability in Armed Conflict of the International Covenant on Civil and Political Rights: a Rejoinder to Dennis and Surena’, (2009) EHRLR 628; N. Lubell, above note 50, at 195 et seq; M. Satterthwaite, ‘Rendered Meaningless: Extraordinary Rendition and the Rule of Law,’ (2007) 75 Geo. Wash. L. Rev. 1333, at 1361 et seq.
(315) See M. Bossyut, Guide to the ‘travaux préparatoires’ of the International Covenant on Civil and Political Rights (Nijhoff, 1987), at 53–5, and the sources cited therein.
(316) See Nowak, above note 170, at 43: ‘the adoption of these words was intended to avoid obligating States parties to protect persons under their jurisdictional authority but outside their sovereign territory. This would be possible only through diplomatic channels’; Lubell, above note 50, at 201: ‘the drafters…wished to avoid the risk that the Covenant would create positive duties outside the scope of a state’s authority and ability to execute such obligations’.
(317) Wall, para. 109.
(318) P. Allott, ‘The Concept of International Law, (1999) 10 EJIL 31, 43.
(319) I have argued for a similar reading of Article VI of the Genocide Convention—see Milanovic, above note 282, at 481–2.
(321) Bosnian Genocide merits judgment, para. 430.
(322) See above, Section 1.F.