The European Convention on Human Rights in Non-International Armed Conflict – Revisiting Serdar Mohammed

Written by

The UK Government’s re-commitment in May to replacing the Human Rights Act (HRA) immediately followed the Supreme Court’s further hearings on one of the more controversial cases under the Act – the Serdar Mohammed claim against the Ministry of Defence (on which additional hearings are expected later this year). The claimant, who on the assumed facts was a Taliban commander detained by the UK military in Afghanistan for 110 days in 2010, alleges a breach of his right to liberty under the European Convention on Human Rights (ECHR).

As readers will recall, the lower courts upheld this claim, prompting controversy in the press and in academia. Preventative detention (or “internment”) of the enemy is widely regarded as an essential incident of armed conflict. The suggestion that the ECHR prohibited the UK from detaining a Taliban commander to prevent his engagement in hostilities against British forces raised obvious concerns about the application of the ECHR in armed conflict, also fuelling further criticism of the HRA.

Since international humanitarian law (IHL) norms designed for the context of hostilities do not prohibit internment in non-international armed conflicts (NIACs) like the Afghan conflict in 2010, much of the legal debate focused on the content of these norms and their relationship with the ECHR. The High Court decision, declining to use IHL to override the ECHR, was criticised as “an outright rejection of the applicability of IHL to the question of who may be detained for what reasons and following which procedure” in NIACs.

Rather than rehearsing the extensive debates (see a small sample here and here) over whether IHL norms authorise detention in NIACs, this post challenges an assumption about the interpretation of the ECHR which underlies the arguments raised by both parties to the claim. Its focus is on a specific provision of the ECHR and its application to situations like that in which the claimant was detained – state participation in NIACs outside their own territory (extra-territorial NIACs).

The result is an alternative approach, based on a context-sensitive interpretation of the ECHR complemented by IHL, which helps address the concern that the ECHR and HRA are inherently unsuited to conditions of armed conflict.


Article 5 of the ECHR, on which the Serdar Mohammed claim largely rests, enshrines the right to liberty. Readers interested in that provision’s application to internment in armed conflict will be familiar with the Strasbourg ruling in Hassan v UK, arising from the early stages of the Iraq war – when the UK was engaged in an international armed conflict (IAC). The court effected an “accommodation” by which the IHL norms applicable in IACs expanded the limited grounds for detention permitted under Article 5(1) to allow internment. See Milanovic and Hill-Cawthorne’s responses to Hassan for detailed background and discussion on this point, including the grounds on which it is open to criticism.

The case for the UK Government in Serdar Mohammed principally depends on extending that criticised “accommodation” concept from the IAC situation considered in Hassan to that of a NIAC. The Court of Appeal held that this founders on the different IHL norms which apply to NIACs. It considered that these do not include a legal right to intern, in contrast with the position in IACs but consistent with the views of Akande and Hill-Cawthorne (among others) in the NIAC IHL academic debate mentioned above.

That position however leaves open the question of whether Article 5 can be reconciled with the realities of participating in an extra-territorial NIAC. Both sides to the case implicitly accept that the provision on its face prohibits internment by a state engaged in an extra-territorial NIAC, and so to the extent internment is required must be disapplied, either by derogation (on which see below) or by using IHL to override it. This post contends instead that the ECHR would allow such a state to detain enemy fighters with a continuous combat function. Such detention could be considered necessary to prevent the detainee committing an offence, and so be permitted by Article 5(1)(c) of the ECHR, even if the threshold for derogation is not met.

Preventative detention and Article 5(1)(c)

Article 5(1)(c) of the ECHR sets out one of the permitted grounds for detention:

“the lawful arrest or detention of a person effected for the purpose of bringing him before the competent legal authority on reasonable suspicion of having committed an offence or when it is reasonably considered necessary to prevent his committing an offence or fleeing after having done so” (emphasis added).

The purpose – of bringing the detainee before the competent legal authority – governs each of the three options which follow (reasonable suspicion, preventing commission of an offence or preventing flight). That interpretation, adopted in the early case of Lawless v Ireland (No 3), would not appear to preclude preventative detention, provided it is accompanied by the necessary procedures for an appropriate officer to review its lawfulness.

After Lawless however, the court moved further by holding that the combination of Articles 5(1)(c) and 5(3) with the purpose of Article 5 (protection from arbitrary detention) precluded preventative detention entirely, absent an intention to bring the detainee to a full criminal trial. The detainee must therefore be suspected of having already committed an offence – see for example the majority opinion in Ostendorf v Germany.

Macken has previously noted that this interpretation seemingly deprives Article 5(1)(c)’s wording on prevention of the independent effect which the most natural reading of the text (noting the word “or”) suggests. She further notes the explicit support in the ECHR’s travaux preparatoires for interpreting Articles 5(1)(c) and 5(3) as allowing preventative detention even where there is no suspicion of a prior offence. A minority in Ostendorf also suggested that the court had historically erred in absolutely prohibiting preventative detention under Article 5(1)(c).

Nonetheless the more recent position has force as applied to the domestic peacetime cases in which it developed, given the purpose of Article 5 (and the ECHR generally) in protecting individuals from arbitrary restrictions on their liberty. Two propositions in particular support a restrictive approach:

1. It is difficult to justify allowing preventative detention where less repressive administrative, police and surveillance measures are available to prevent criminality.

2. The principle of legality requires the parameters of lawful detention to be set out in law, in clear and foreseeable terms, whereas the inherent uncertainty associated with determining which individuals are sufficiently likely to commit a future offence creates significant risk of abuse.

The question is whether these points are best addressed with blanket assertions that Article 5(1)(c) could not be used purely preventatively in any context. They could have been addressed in the previous cases by holding that detention could never be “reasonably considered necessary to prevent [a detainee] committing an offence” in a state’s own territory, or in peacetime conditions. In such cases the availability of alternative measures renders such detention unnecessary in principle, and the absence of an objective standard for assessing propensity to future criminality fails to satisfy the requirement of legality. Accordingly the text of Article 5(1)(c) could preclude such detention in the cases where the court has explicitly rejected it, without the need to resort to a general absolute prohibition.

Article 5(1)(c) and extra-territorial NIACs

Such an analysis permits a different approach to extra-territorial NIAC cases, like Serdar Mohammed, where the two propositions noted above would not necessitate the same result:

1. Regarding the availability of less repressive measures than detention, the UK was acting outside its territory, without recourse to domestic law enforcement capabilities. Although co-operation with the Afghan government would normally be the appropriate mechanism to address that deficit, the ongoing NIAC severely limited its capacity as well. Neither state could respond effectively to such levels of violence and organised armed opposition (characteristic of NIACs) using peacetime law enforcement measures.

These conditions would presumably have met the threshold for derogation by the Afghan government from its own international human rights law obligations. It is less clear that they could meet the threshold for the UK to derogate under Article 15 of the ECHR, where it had “chosen to conduct an overseas peacekeeping operation, however dangerous the conditions, from which it could withdraw” – to adopt Lord Bingham’s description of the Iraq extra-territorial NIAC in R (Al-Jedda) v Secretary of State for Defence.

2. As to whether it is possible to define which specific individuals can be detained with sufficient certainty, at this point IHL can assist, informing the question of whether an individual’s detention can reasonably be considered necessary under Article 5(1)(c).

According to the International Committee of the Red Cross (ICRC), in NIACs there exists a category of individuals who by virtue of their expected future conduct are liable under IHL targeting rules to the use of lethal force. These are members of an organised armed group who have assumed a “continuous function for the group involving” their direct participation in hostilities (DPIH). This continuous combat function (CCF) concept was elaborated by the ICRC in its Interpretive Guidance on the Notion of Direct Participation in Hostilities.

This is not to suggest that the IHL targeting rules described in that Guidance establish an implied legal basis in IHL for detention in NIACs. Rather, they provide a framework for addressing the concern that preventative detention is inherently too uncertain and open to abuse to satisfy the ECHR’s requirement of legality. The relevance lies in the distinction between organised armed group members with a CCF and civilians who engage in DPIH on a “spontaneous, sporadic, or unorganized basis.”

For civilians, engagement in DPIH (even if repeated) is not “a sufficiently reliable basis for the assumption that [they] would necessarily do so again” and so only justifies their targeting during such engagement. Those with a CCF, on the other hand, are liable to attack even when they are not engaged in DPIH, since their “membership of the group and assumption of a fighting capacity within the group is itself an indication of the dangers posed by the individual.” Hence a person “recruited, trained and equipped by such a group to continuously and directly participate in hostilities on its behalf can be considered to assume a [CCF] even before he or she first carries out a hostile act.” Expected future conduct, assessed on the basis of objective identifying factors, underlies the distinction.

If the evidence described in the Guidance can confirm the probability of an individual’s future engagement in DPIH with sufficient certainty to justify lethal force to prevent it, that evidence should also suffice to justify detention for the same preventative purpose. The fighter’s role within an organised armed group also links them directly to the circumstances which give rise to the necessity of internment, since the existence of such a group is a constitutive element of a NIAC (and explains the associated inadequacy of less restrictive measures to address the threat). The question remains however as to whether such an approach is legally sustainable.

Interpreting the ECHR

To fall within Article 5(1)(c), internment of those with a CCF requires as its purpose the prevention of a criminal offence. That focus on criminality explains the ECHR’s statement in Hassan that the detention of “prisoners of war [with] combatant privilege, allowing them to participate in hostilities without incurring criminal sanctions” would not fall within Article 5(1)(c).

While prisoners of war in IACs benefit from such privilege, in NIACs organised armed group members do not. Furthermore in practice, as was agreed during the expert discussions which preceded the ICRC Interpretative Guidance, engaging in DPIH or assuming a CCF “would normally constitute a criminal offence” leading to “arrest and prosecution” under the applicable domestic law of the territorial state. Accordingly in a NIAC, in contrast to the Hassan situation, internment can be consistent with the ordinary meaning of Article 5(1)(c).

The Serdar Mohammed situation provides an example. British forces, supporting the Afghan authorities, had a legitimate interest in preventing (criminal) violence, in which the claimant was alleged to be involved, against themselves and civilians. While Afghan criminal proceedings might have been preferable, in an armed conflict legal and practical obstacles often arise to transferring detainees to the territorial state authorities (see the Serdar Mohammed Court of Appeal ruling, paragraph 213).

Altering the assessment of compliance with human rights norms, during armed conflict, is not novel. The International Court of Justice’s Advisory Opinion on The Legality of the Threat or Use of Nuclear Weapons considered that the determination of whether an action is “arbitrary” under Article 6 of the International Covenant on Civil and Political Rights could then be determined using IHL rather than normal peacetime standards. That approach has been relied on by various human rights bodies since, and was cited by the Strasbourg court in Hassan (where IHL supported a far more radical departure from Article 5’s text than is here proposed).

Whether detention can be “reasonably considered necessary” for preventing criminality could equally be determined differently, or with different results, during armed conflict. The use of IHL to inform this human rights assessment could be regarded as an example of “systemic integration”, comparable to that recently discussed on this blog in relation to the African Charter on Human and Peoples’ Rights.

The acknowledgement in Al-Skeini v UK that it is possible to “divide and tailor” ECHR rights when they are applied extra-territorially could also support permitting flexibility in an extra-territorial NIAC. A state’s means to suppress criminality within its territory are greater, as discussed, and where a domestic NIAC seriously compromises that capacity the threshold for derogation is likely to be met. Where states participate in extra-territorial NIACs, on the other hand, the flexibility which Al-Skeini acknowledged (in imprecise terms) is necessary for the extra-territorial application of the ECHR could provide a basis for departing from the established jurisprudence on Article 5 without the need to derogate.


The above considerations suggest sufficient space for the proposed departure, and justification for limiting it to the specific and problematic circumstances of an extra-territorial NIAC, both on policy and legal grounds.

The proposed adjustment does not resolve all of the issues raised by the Serdar Mohammed litigation. Although it suggests the claimant’s detention could have fallen within the ECHR’s permitted grounds, it does not address the need for a legal basis and adequate procedural safeguards, both of which the Court of Appeal found lacking in that case.

Nonetheless, by providing additional flexibility within limits which could be overseen by the courts, it demonstrates the possibilities for a constructive relationship between the ECHR regime and the norms of IHL. The approach, it is hoped, can support the effective protection of liberty in extra-territorial NIACs through an interpretation which respects the text and purpose of the ECHR, without imposing limitations states will not realistically accept on military operations.

Print Friendly, PDF & Email


No tags available

Leave a Comment

Comments for this post are closed


Aurel Sari says

August 22, 2016

Good arguments, but isn't there an easier route altogether? If membership in an OAG opens up an individual to lethal targeting, as this argument accepts and is based on, then it follows that detention in lieu of death is permissible. No need to mix in criminal law. To this effect in more detail, see

Ed Robinson says

August 22, 2016

Thanks Aurel, appreciate the comment. Although the outcome is similar, for me the difficulty is that I find the approach Akande and Hill-Cawthorne outlined here (that IHL rules in NIACs are prohibitory only, unlike in IACs) compelling . Accordingly when I refer to the evidence of CCF being sufficient to justify targeting I don't mean to suggest it provides an IHL authorisation to do so, but that it justifies displacing an IHL prohibition that would otherwise exist (if they were a civilian).

That view doesn't preclude those targeting rules from being relevant to the question of preventative detention, given the underlying policy considerations the CCF concept rests on and the evidence it uses, but if correct it removes the basis for finding an authority to detain existing in NIAC IHL. Further, even if such authority does exist I would also have a concern over using lex specialis to interpret Article 5(1) of the ECHR in a manner I don't consider its wording can sustain.

Jordan says

August 22, 2016

Aurel raises an important point regarding interpretation of international agreements, in view of the ordinary meaning of textual provisions in light of the object and purpose (e.g., to lessen death, injury and suffering while allowing otherwise unprohibited reasonably necessary and proportionate measures of warfare -- thus also considering CIL and subsequent practice as background re object and purpose and textual (even implied) meaning). Result: otherwise lawful detention is not prohibited and is permissible in view of CIL and subsequent practice.
Yet, re article 15 of the European Convention, does the law of war provision only apply re killings? Or should this also be interpreted in view of the object and purpose of the treaty and in view of CIL and subsequent practice?

Jordan says

August 22, 2016

p.s. might this approach to treaty interpretation under the Vienna Convention and CIL re treaty interpretation sometimes compel detention where feasible as opposed to killing, except that CIL laws of war allowing killing until surrender. At least detention would generally be preferable to killing, from a policy-oriented point of view (which the Vienna Convention favors at least partly because of the need to consider object and purpose). At least rigid, if not self-deceiving, textual alone approaches are unacceptable.

Jordan says

August 22, 2016

Another point: the need to interpret text also in view of other textual provisions. If one focuses merely on the claim of a single rights claimant, one can miss the need to accommodate the rights of others. Clearly, a human rights treaty is not meant to protect merely a right of a single rights claimant at the expense of the human rights of others.
Therefore, a right re detention should be interpreted in view of rights of others to personal security, liberty, dignity, and vice versa, esp in view of the object and purpose of a human rights treaty. This, of course, is far more easy to recognize re the ICCPR articles that use a limiting word "arbitrary."

Clive Walker says

August 28, 2016

I wonder how this argument about art.5(1)(c) fits with the views of the ECtHR about the length of detention in Magee v United Kingdom, App no 26289/12, 29062/12, 29891/12, 12 May 2015, para.105.

Ed Robinson says

August 29, 2016

Thanks Clive, it's an interesting point to bring out. I'd note that Magee's starting point is different from mine, in that they're dealing with a case about detention where there's a reasonable suspicion of an existing offence rather than purely preventative detention. It does however appear the Magee approach could also be relevant to some cases like Serdar Mohammed's – on the assumed facts a reasonable suspicion that he had already committed an offence could presumably have been justified, and indeed in due course he was convicted in the Afghan courts. The 12 day period Magee held lawful would also give more flexibility than the 96 hours permitted by the UK courts in the Serdar Mohammed case so far.

That said, the High Court ruling on Serdar Mohammed from paragraph 330 to 332 – issued before Magee of course – discusses and dismisses similar arguments by the UK given the 110 day period of his detention. That, and the need for suspicion that an offence has already been committed, suggest to me that Magee doesn't provide a complete answer to the problem and there's still a need for genuinely preventative detention. On that aspect Magee paragraph 105 still appears to start from the assumption in the existing jurisprudence that preventative detention is unlawful, which as I've noted seems reasonable in the contexts in which it has developed (and the Magee case itself) but less so in extra-territorial NIACs.