Reconceptualising the right of self-defence against ‘imminent’ armed attacks

ABSTRACT A state’s right to act in self-defence against ‘imminent’ armed attacks remains an unsettled question of international law. Yet, states persist in justifying military actions on this basis. Absent a common definition of imminence, assessing the legality of these operations is practically impossible. Although imminence is traditionally understood as referring solely to the temporal proximity of an armed attack, for some this approach is insufficient. This article examines scholarship and examples of state practice that indicate that imminence may be viewed as comprising several contextual indicators that determine whether states may have recourse to self-defence. This conception of imminence raises fears of an expansive right of self-defence. Yet, this author concludes that such ‘contextual imminence’ stands as a proxy for jus ad bellum necessity. This conflation is perhaps unfortunate, but an orthodoxy regarding all forms of self-defence is thereby maintained, subject to the enduring legacy of the Caroline formula.


Introduction
Article 51 of the Charter of the United Nations ('UN Charter') recognises a state's inherent right of self-defence 'if an armed attack occurs'. 1 This right is one of only two explicit exceptions to the Article 2(4) prohibition on the threat or use of force between states, the other being force authorised by the UN Security Council pursuant to Chapter VII of the UN Charter. 2 likewise focused on the question of imminence, 9 and it was a feature of the report of the Special Rapporteur on Extrajudicial, Summary or Arbitrary Executions that considered the legality of the strike. 10 Imminence was also prominent in the ensuing media coverage and scrutiny by American lawmakers. 11 Despite imminence being so pivotal to the legal analysis of such claimed acts of self-defence, its precise meaning has been the focus of relatively little attention, academic or otherwise. 12 Different states, scholars, courts, and international organisations employ the term 'imminence' and use it to justify and/or review a use of putatively defensive force in very different ways. Without a common understanding of the term, however, assessing legality is extremely difficult, if not impossible. This article examines, therefore, one of the most contemporary and contentious issues in the jus ad bellum.
Because the notion of anticipatory self-defence is so contentious, debates on this topic are often polemic and zero sum. We may be presented with either an outright rejection of any right of anticipatory self-defence or a wholesale acceptance of an expansive version of it. For this author, however, there is space for more nuanced consideration. This conclusion is based on a study that this author conducted elsewhere, where imminence is considered in the context of a broader review of jus ad bellum necessity and proportionality. In that study, this author concludes that imminence, as understood by certain scholars and states, stands as a proxy for the customary requirement of necessity. 13 The purpose of this article is to develop this argument by exploring in much greater detail the place and importance of imminence in the jus ad bellum, the fears that surround the prospect of a flexible appreciation of imminence and, crucially, to insert some balance into the debate by considering how such fears might be mitigated and why greater flexibility might be justifiably required by states. The hope, therefore, is that this article provides a comprehensive and evenhanded examination of an under-theorised subject.
Section 2 of this article sets out the context from which the ensuing analysis proceeds. It provides an overview of the different types of anticipatory self-defence and touches on the issue of whether, as a matter of lex lata, there exists a right of self-defence in international law against armed attacks that are imminent. The law is unsettled, but the prospect of countering imminent armed attacks, especially by terrorist NSAs, clearly persists as a feature of state justifications for using force. Important for present purposes, section 2 also highlights some of the concerns associated with any such right, and these concerns run as a theme throughout the article as we examine the possible meanings of imminence.
Section 3 considers how certain states and scholars understand imminence and its role in the jus ad bellum. Section 3 centres on recent case studies comprising public attempts by states to articulate the meaning of imminence. This article shows that several scholars and, most importantly, certain states do not simply regard imminence as a matter of the temporal proximity of an anticipated armed attack. Imminence is understood more flexibly, being comprised of various contextual indicators, both temporal and non-temporal, that reflect the nature of the threat that states might face. This article dubs this conceptualisation 'contextual imminence'. Although it is doubtful that contextual imminence currently represents the accepted customary international law standard, contextual imminence explains how certain militarily powerful states might use force in their international relations and, moreover, it is potentially indicative of how international law might develop if other states adopt it.
This author recognises that the prospect of states relying on this standard raises fears of an overly broad conceptualisation of the right of self-defence, one that potentially gives states greater freedom to project force beyond their own borders. However, increased flexibility of action does not equate to complete freedom to use force. The key consequence of the adoption of contextual imminence is explored in section 4, which sets out the aforementioned conclusion that the ability of states to act in self-defence against imminent armed attacks understood on a contextual basis essentially translates into a question of jus ad bellum necessity. Section 4 explains this argument and develops it further, including by introducing a case study that illustrates the conflation of imminence and necessity. Furthermore, section 5 provides a novel examination of the normative implications of adopting contextual imminence and suggests how concerns associated with this more flexible concept might be assuaged.
This author ultimately argues that even if contextual imminence is or becomes the accepted standard for states to be able to act in self-defence anticipatorily, an orthodoxy in the jus ad bellum is nevertheless preserved. Contextual imminence permits a state to exercise its right of self-defence where military action in advance is required to ensure an effective defensive response. Yet, a use of force must still be the only reasonable choice of means at the relevant point in time to counter a positively identified future armed attack. Military responses to latent threats are precluded, regardless of the nature or gravity of such threats.

Anticipatory self-defence: concept and types
Before engaging with the meaning of imminence, we must first consider the concept of anticipatory self-defence in general terms. It is well known that Article 51 of the UN Charter recognises a state's inherent right to selfdefence 'if an armed attack occurs'. Given the wording of Article 51, the issue under consideration is whether, and under what circumstances, defending states 14 possess a right of self-defence against future armed attacks that, by definition, have not yet 'occurred'. To examine this issue, a terminological inconsistency must first be clarified. States and scholars are rarely consistent in how they describe the right to take military action in respect of future threats. Labels such as 'anticipatory', 'preventive' and/or 'pre-emptive' self-defence may be used variously to describe very different rights of action. For the purposes of this article, the term 'anticipatory self-defence' describes any use of defensive force employed to counter the threat of a future armed attack. This umbrella term is subsequently divided into two forms of anticipatory self-defence: 'preventive selfdefence' and 'pre-emptive self-defence'. 15 Preventive self-defence represents the most expansive type of anticipatory self-defence. It refers to the possibility of countering potential future threats 14 A 'defending state' is a state that is, or claims to be, the victim of an armed attack by another state or NSAs. The controversies regarding whether NSAs have the capacity to carry out armed attacks are also well known. For an overview, see Ruys (n 3) 368-510. Although the matter is still debated, it is assumed for the purposes of this article that states possess a right of self-defence against armed attacks by NSAs regardless of any attribution of such attacks to a state. See 15 See Ruys (n 3) 250-4 for a general discussion of the terminology that this article adopts. that have not materialised and might not do so. A classic example is the potential future acquisition of nuclear weapons by states such as North Korea and Iran. Importantly for our undertaking, preventive self-defence refers to a use of force to counter possible future armed attacks that are not 'imminent'. This is regardless of how imminence is understood (see section 3). The doctrine of preventive self-defence is thus based on conjecture. Its purview is uncertain and unspecified threats existing at an unknown point in a potentially distant future. As such, it relies on an overly expansive interpretation of Article 51, one that gives states enormous flexibility to act unilaterally outside of the collective security framework of the UN Charter. The obvious consequence of this interpretation is the ability of militarily powerful states to abuse their right of self-defence. It gives them great freedom to project military force beyond their own borders in the broadest possible range of scenarios, whilst claiming to operate within the confines of legality.
Preventive self-defence has been most notably advocated by the USA under the auspices of the notorious 'Bush Doctrine', 16 yet such right has been almost universally denounced by scholars and is broadly rejected by states. 17 In 2016, even the USA appeared to reaffirm the requirement of an imminent armed attack for the purposes of engaging a right anticipatory self-defence, thereby implicitly ruling out the previously asserted right of preventive self-defence. 18 Beyond the USA's position, with the exception of Israel, state practice in support is virtually non-existent. 19 The vast majority of states prefer to avoid such a 'fuzzy and dangerous notion'. 20 The Non-Aligned Movement, representing over half of the world's states, 16 This doctrine, contained in the USA's 2002 National Security Strategy, set out the USA's need for 'anticipatory action to defend ourselves, even if uncertainty remains as to the time and place of the enemy's attack. To forestall or prevent such hostile acts by our adversaries, the United States will, if necessary, act preemptively': The White House, 'The National Security Strategy of the United States of America' (September 2002) https://2009-2017.state.gov/documents/organization/63562.pdf, 15 (accessed 9 June 2021). The reference to pre-emptive self-defence highlights the aforesaid terminological issue. For present purposes, the defensive action described in the National Security Strategy is preventive. The policy was principally adopted to counter threats posed by 'rogue states', NSA terrorists and WMDs. The rationale is that such threats are potentially so devastating that states should not be required to wait until they materialise before responding defensively to forestall them. 17  has explicitly rejected the doctrine of preventive self-defence, 21 as has the UK, typically a staunch ally of the USA. 22 In considering the meaning of imminence and whether it should be construed in narrow or broad terms, it is important to note that even if a right of preventive self-defence is no longer referenced explicitly as it once was, its core idea (being the ability to counter unspecified potential threats) subsists as a point of academic contemplation and, most importantly, in the words and deeds of states. For example, while the American National Security Strategy of 2017 did not explicitly promote the idea of preventive self-defence along the lines of the Bush Doctrine, it maintained the possibility that the USA 'will act against [terrorist] sanctuaries and prevent their reemergence, before they can threaten the U.S. homeland.' 23 Although this statement is ambiguous, concern was rightly raised that this so-called 'Trump Doctrine' might not be so different from the Bush Doctrine. 24 As to current American policy, the incumbent Biden administration is yet to publish its full National Security Strategy. Its Interim National Security Strategic Guidance does reference the need to 'prevent an ISIS resurgence', yet it explicitly notes that military force is not the answer to the challenges in the Middle East and emphasises that force should generally be employed as a last resort. 25 It is yet to be seen whether this less belligerent language will translate into a more cautious approach to the issue of anticipatory self-defence. In the meantime, that states might still seek to respond to threats to their security that are unspecified and potentially temporally remote is troubling. As the following sections explore, even where states affirm a right to respond to armed attacks that are 'imminent', potentially expansive interpretations of that term may have dangerous consequences for the legal regulation of armed force.
Pre-emptive self-defence, in this article, refers to defensive force employed to counter armed attacks that are 'imminent'. This type of anticipatory self-defence is the focus of the current inquiry. for states to act in this way, even in the post-UN Charter era, is typically said to derive from the celebrated Caroline incident of 1837 26 and American Secretary of State Daniel Webster's assertion that the necessity of selfdefence must be 'instant, overwhelming, and leaving no choice of means and no moment for deliberation.' 27 This part of the so-called 'Webster formula' or 'Caroline formula' is the traditional starting point for any discussion of a right of pre-emptive self-defence, even if the facts of the incident do not necessarily comprise a response to an anticipated future armed attack. 28 In its most orthodox form, a right of pre-emptive self-defence is typically understood as a right to respond militarily to an armed attack that is about to be launched in the reasonably foreseeable future. 29 Pre-emptive self-defence is, therefore, a response to the 'sitting duck dilemma', 30 meaning that Article 51 should not be interpreted in a way that 'requires a state to passively accept its fate before it can defend itself.' 31 In contrast to the almost uniform rejection of preventive self-defence, whether states currently possess the right to respond to imminent armed attacks occupies a much less clear position in the lex lata. There is no agreement amongst scholars as to whether the UN Charter as interpreted (or reinterpreted) admits of pre-emptive action, and/or whether states possess such a right as a matter of pre-Charter customary international law (that has survived the signing of the Charter), and/or by virtue of customary development that has occurred since 1945. 32 Credible arguments can be made either way. Although some commentators reject any notion of anticipatory 26  self-defence, 33 this is arguably a minority view. There is a growing academic consensus that supports a limited right of pre-emptive self-defence as a general premise, even if disagreement persists over the meaning of 'imminence'. 34 The UN has also appeared generally supportive of such a right. 35 The International Court of Justice ('ICJ') meanwhile has not offered its opinion on a right of pre-emptive self-defence, having twice avoided expressing a view on it. 36  were it ever to be called upon to consider it. 37 Most significant is that the ICJ has consistently interpreted the right of self-defence in a conservative manner. In Armed Activities, it held that Article 51 of the UN Charter may only justify a defensive use of force within the strict confines of that article and that it 'does not allow the use of force by a state to protect perceived security interests beyond these parameters'. 38 Such a pronouncement might appear self-evident, given that the trigger for the right of self-defence under Article 51 is an 'armed attack' and not any broader notion of 'security interests'. Nevertheless, this conclusion stands as an implicit rejection of a right of preventive self-defence against non-imminent threats. 39 Therefore, the ICJ's jurisprudence potentially indicates scepticism regarding any right of anticipatory self-defence, but it offers no clear answers. The existence or otherwise of this right in the lex lata remains subject to state practice. Such practice must be considered to interpret Article 51 40 and, coupled with opinio juris, determines the position under customary international law. 41 3. State practice and the meaning of imminence

State practice
It is not the purpose of this article to provide an exhaustive review of post-1945 state practice that indicates whether international law currently accommodates a right of self-defence against armed attacks that are imminent. Eminent jus ad bellum scholars have already penned excellent commentary on this practice and there is no need to repeat this analysis here. 42 Rather, regardless of whether international law clearly recognises a right of preemptive self-defence (which is arguable), the focus of the ensuing analysis is to show that states have had, and continue to have, recourse to a right of pre-emptive action to justify using force. In such cases, the presence or absence of imminence is often the determining feature of other state responses as to the legality or legitimacy of such putatively defensive acts. It is only possible to assess states' adherence to international law as it stands currently, or might develop in the future, if we know how and under what circumstances states rely on imminence to justify their actions and review the actions of other states. The following references to state practice are therefore employed to highlight a potential shift in how certain states interpret imminence and might subsequently resort to self-defence against future armed attacks. For present purposes, it is sufficient to note that a number of states since 1945 have expressed support for a right of pre-emptive self-defence, either as a general premise and/or to respond to a particular incident involving a use of force. Likewise, and conversely, several states have rejected the doctrine of pre-emption entirely. 43 This divergent state practice led Ruys in 2010 to conclude rightly that it was impossible to identify a right of pre-emptive selfdefence in the lex lata. Ruys noted that although opinio juris in support of a right to counter imminent armed attacks had increasingly become more common and explicit, consistent opposition by a large group of states meant that there was no widespread acceptance of such right. 44 More recent state practice on this issue has occurred principally in the context of combatting international terrorism. In the post 9/11 era, states have increasing had recourse to the right of self-defence to justify combatting threats posed by hostile NSAs, including terrorist groups. In so doing, they have claimed that such groups, as well as individuals, pose an imminent threat to national security. 45 This most recent practice is particularly important. It serves to bring up to date Ruys' assessment of whether a right of preemptive self-defence exists as a matter of lex lata. Moreover, for our purposes, it speaks to our understanding of imminence. This is because the context of combatting a persistent terrorist 'threat' has seemingly informed how certain states now interpret imminence. These incidents of state practice are, therefore, key to the present endeavour and will be referred to as case studies in the sections to follow.
A well-known and controversial case study occurred in August 2015, when the UK killed British citizen, Reyaad Khan, in Syria using a RAF drone. 46 David Cameron, the then British Prime Minister, publicly rationalised the strike by reference to the UK's inherent right of self-defence. 47  UN Charter, 48 the UK invoked the right of individual self-defence based on the claim that Khan was 'actively engaged in planning and directing imminent armed attacks against the United Kingdom'. 49 Following the drone strike, the UK Parliament's Joint Committee on Human Rights issued a report entitled The Government's Policy on the Use of Drones for Targeted Killing. 50 The report examined, inter alia, the meaning of imminence, 51 agreeing with the UK government's long held view that pre-emptive self-defence is available to a state in response to imminent armed attacks. 52 This acceptance of a right of pre-emptive selfdefence has also been articulated by former British Attorneys General. 53 Following this incident, British allies have likewise explicitly asserted their right of pre-emptive self-defence. Most important for this article is that, in so doing, such states have spelled out their understanding of imminence. From the USA's perspective, the U.S. Department of State Legal Adviser confirmed in 2016 the long-standing American position that the jus ad bellum allows a state to exercise its right of self-defence 'not only in response to armed attacks that have occurred, but also in response to imminent ones before they occur.' 54 The USA's stance on this issue is incorporated into their current Department of Defense Law of War Manual, as well as being reflected in the initial justifications for targeting General Qassem Soleimani in 2020. 55 Australia likewise explicitly asserts a right of pre-emptive selfdefence, 56 52 Ibid, para 3.30, although the report refers to 'preventive' self-defence in making this assertion, highlighting the aforementioned terminological issue. The report went on to note the imprecision surrounding the meaning of imminence, whilst noting the well-known Caroline test for imminence (see section 3): Ibid, para 3.31. 53 Former UK Attorney General Lord Goldsmith has asserted that the UK's position regarding the right to respond defensively to imminent armed attacks has been consistent 'over many years': New Zealand, might also adopt this common position, but they have not publicly confirmed this. 57 Other states that have recently asserted and/or recognised a right of self-defence against imminent armed attacks include Estonia, 58 India, 59 Iran, Israel, 60 Japan, 61 The Netherlands, 62 and Turkey. 63 Subsequent to the Soleimani killing, Lithuania and South Africa both recognised the right of states to respond to 'imminent threats', and Liechtenstein offered implicit support for a right of pre-emptive self-defence. 64 In summary, at this stage we can say no more than state practice remains inconclusive in terms of establishing as lex lata a clear right to respond in selfdefence against imminent armed attacks. 65 Post-UN Charter claims to a right of pre-emptive self-defence are relatively uncommon and several states remain hostile to any idea of anticipatory self-defence. Yet, recourse to such a right is undoubtedly a feature of contemporary state practice. A limited number of states remain very pro a right of pre-emptive selfdefence and continue to explicitly espouse it, most notably in the modern context of combatting the persistent 'threat' of international terrorism. Other states, in actively supporting anti-terrorist military action, and/or not condemning it in circumstances that call for such condemnation, 66  66 We should be careful in drawing conclusions from states remaining silent regarding the acts and legal claims of other states, in particular whether silence should be construed as acquiescence or support of a legal right or course of action. This is a controversial subject, but the ICJ has recognised in principle (albeit in a different context) that '[t]he absence of reaction may well amount to acquiescence … That is to say, silence may also speak, but only if the conduct of the other State calls for a response': might be viewed as implicitly supporting a right of pre-emptive action. 67 Where states do resort to pre-emptive self-defence to justify their actions, the meaning of 'imminence', whether as lex lata or lex ferenda, potentially has huge significance, given the ramifications of states using force extraterritorially. However, the most recent references by states to the right of pre-emptive self-defence generally contain no explanation regarding how those states understand imminence. Only the UK, the USA and Australia have provided any detail regarding their positions on this term, which is why these examples of state practice are examined in detail in the following section. However, even these explanations do not provide the consistency or clarity that we might hope for.

The rise of contextual imminence?
Having established that the law remains in flux on the question of whether states currently possess a right of pre-emptive self-defence against imminent armed attacks, the remainder of this article focuses on the question of how states might now interpret imminence. In approaching this analysis, we should first note that there exists no authoritative legal definition of imminence insofar as such term relates to an armed attack. 68 Scholarship provides no consensus on this issue and the ICJ's jurisprudence contains no answers. In terms of state practice, the USA, the UK and Australia have each explained their respective positions on imminence in support of their aforementioned assertions of a right of preemptive self-defence. 69 As this section will demonstrate, their views remain open to varying interpretations and have raised as many questions as answers. 70 Nevertheless, they represent important case studies for this article to consider. Most importantly, these public pronouncements are made by militarily powerful states and indicate how such states understand imminence and might act upon that understanding in exercising their right of selfdefence. States rarely promulgate such general statements on the jus ad bellum. They therefore stand as rare and explicit examples of state practice that offer valuable (if not entirely coherent) detail that speaks to the current inquiry. 71 The positions adopted by these states illustrate the relationship that imminence has with the 'armed attack' trigger of the right of selfdefence and, of particular significance, the relationship that imminence has with 'necessity', which conditions the exercise of that right. As will be seen in the ensuing analysis, the distinction between these two latter concepts is increasing blurred by how these states appear to conflate imminence and necessity, which has meaningful repercussions for the development of the jus ad bellum.
Whilst this author highlights the importance of these case studies, it is accepted that the practice of three states (even militarily powerful ones) is unlikely on its own to be considered widespread enough to generate custom. It remains to be seen if other states will publicly endorse or adopt their views. Yet, it would be facile to dismiss such practice without recognising its potential normative significance. The practice of a relatively small number of states might be sufficient to generate customary international law 'if such practice, as well as other States' inaction in response, is generally accepted as law (opinio juris).' 72 Even though the silence of other states in response to the UK, USA and Australia's respective articulations of imminence does not necessarily equate to acquiescence, 73 there is potential for '"Imminence" in the Legal Adviser's Speech', Lawfare (6 April 2016) www.lawfareblog.com/imminencelegal-advisers-speech (accessed 9 June 2021); Marty Lederman, 'The Egan Speech and the Bush Doctrine: Imminence, necessity, and "first use" in the jus ad bellum', Just Security (11 April 2016) www. justsecurity.org/30522/egan-speech-bush-doctrine-imminence-necessity-first-use-jus-ad-bellum/ (accessed 9 June 2021); Monica Hakimi, 'The UK's most recent volley on defensive force', EJIL: Talk! (12 January 2017) www.ejiltalk.org/the-uks-most-recent-volley-on-defensive-force/ (accessed 9 June 2021); Adil Ahmad Haque, 'The United Kingdom's "Modern Law of Self-Defence" -Part I', Just Security (12 January 2017) www.justsecurity.org/36235/united-kingdoms-modern-law-self-defence-part/ (accessed 9 June 2021); Marko Milanovic, 'What is an imminent armed attack? A hopefully helpful hypo', EJIL: Talk! (12 January 2017) www.ejiltalk.org/what-is-an-imminent-armed-attack-a-hopefullyhelpful-hypo/ (accessed 9 June 2021); James A Green, 'Initial Thoughts on the UK Attorney General's Self-Defence Speech', EJIL: Talk! (13 January 2017) www.ejiltalk.org/initial-thoughts-on-the-ukattorney-generals-self-defence-speech/ (accessed 9 June 2021); Adil Ahmad Haque, 'Imminence and self-defense against non-state actors: Australia weighs in', Just Security (30 May 2017) www. justsecurity.org/41500/imminence-self-defense-non-state-actors-australia-weighs/ (accessed 9 June 2021); Henderson (n 17) 297-307. such standard to be sufficiently widespread that it becomes custom. If the UK, USA, and Australia have acted, or proceed to act, together on the same understanding with their allies and coalition partners, such possibility is more likely. Consequently, these case studies do not necessarily reflect the contemporary lex lata, but they might have sown the seeds for the evolution of international law.
Turning to the positions of these three states, we see that although imminence is often said to reflect the Caroline formula of 'instant, overwhelming, and leaving no choice of means and no moment for deliberation ', 74 what this means in practice is open to varying interpretations. The UK, the USA and Australia have each adopted Bethlehem's 'Principle 8' amongst other factors that they consider when approaching imminence: Whether an armed attack may be regarded as 'imminent' will fall to be assessed by reference to all relevant circumstances, including (a) the nature and immediacy of the threat, (b) the probability of an attack, (c) whether the anticipated attack is part of a concerted pattern of continuing armed activity, (d) the likely scale of the attack and the injury, loss, or damage likely to result therefrom in the absence of mitigating action, and (e) the likelihood that there will be other opportunities to undertake effective action in self-defense that may be expected to cause less serious collateral injury, loss, or damage. 75 Scholars have noted that it is unclear how these factors relate to each other, or whether they carry equal or differing weights. 76 For example, in addition to considering the timing of an attack, the reference to the scale and effects of an armed attack might indicate that a different threshold of violence should apply to threatened armed attacks. Equally, this reference might simply pertain to the Nicaragua de minimis gravity threshold. 77 Furthermore, Bethlehem's Principle 8 and the three states' adoption of it are stated to apply in the context of combatting armed attacks by NSAs, but whether they are intended to be equally applicable in an interstate context is unclear.
In any event, as well as not necessarily reflecting the lex lata in whole or in part, the Bethlehem Principles are highly controversial and have been 74 The Chatham House Principles (n 34) 967. 75  subjected to understandable criticism. 78 Yet, even if not representative of contemporary international law, and despite the criticisms and controversies, Bethlehem's formulation and its adoption by the three states concerned are highly instructive for our examination of imminence. As case studies, they serve as a catalyst for intellectual investigation, requiring us to consider both the advantages and disadvantages of appreciating imminence more flexibly and what adoption of this approach might mean for the right of self-defence. This inquiry includes scrutiny of the possible dangers that this meaning might pose and how such dangers might be mitigated. The result is a better realisation of the law as it potentially stands or might develop in the future. meaning. 83 Conceptualising imminence based on the temporal proximity of the armed attack reflects a common understanding of the word that something is impending. This approach was adopted, for example, by the UK's All Party Parliamentary Group on Drones Inquiry Report of 2018 that followed the 2015 drone strike on Reyaad Khan. The Report postulated that the ordinary meaning of imminence 'requires an assessment of temporal factors only and translates to an attempt to answer the question: is the attack about to happen?' 84 It also contended that imminence understood along the lines of the Caroline formula 'emphasises the need for a specific and identifiable threat which is being prepared at the time and about to be delivered in a very short amount of time.' 85 Focusing on the temporal element of the Caroline formula and the requirement that there be 'no moment for deliberation' might indeed suggest that, to be imminent, an armed attack must be impending. Certain scholars have also focused on the temporal proximity of the armed attack. For example, in one of the rare academic studies that seeks to examine the meaning of imminence in any detail, Lubell argues that an imminent armed attack 'must be an impending attack over which there is a reasonable level of certainty that it will occur in the foreseeable future' and the threat must be 'specific and identifiable'. 86 It is uncontroversial that imminence, as it relates to pre-emptive self-defence, must refer to an 'objectively verifiable, concretely imminent attack', 87 rather than to unmaterialised and speculative threats, which are the domain of unlawful preventive self-defence. However, regarding the temporal proximity of the armed attack, Lubell maintains that an armed attack might be imminent, but selfdefence will not be necessary where non-forcible alternatives are available, or where the action by the UN Security Council precludes the need for defensive action. 88 This logic, like that of the UK's All Party Parliamentary Group on Drones Inquiry Report, reduces imminence solely to a question of the timing of a specifically identified future armed attack.
As noted in section 2, this position perhaps represents the orthodox view of imminence, but it is not the only approach that one might take. Academics have long argued that imminence comprises additional non-temporal components. Notably, in their commentary on whether an armed attack is 83 See, e.g. Onder Bakircioglu, Self-Defence in International and Criminal Law: The Doctrine of Imminence (Routledge, 2011) 196, emphasising that imminence signifies the 'temporal facet of self-defence'. 84 UK's All Party Parliamentary Group on Drones Inquiry Report (n 81) 36 (emphasis added). 85 Ibid (emphasis added). 86 Lubell (n 7) 702-5, 718. Lubell describes imminence as a separate, third, customary requirement for measuring defensive action, in addition to necessity and proportionality. This minority position is not generally shared by scholars, nor in the state practice referred to herein from the USA, UK and Australia. As set out in this article, imminence is best understood as pertaining to necessity. Recommendations each include a set of contextual indicators that speak more generally to the threat that a state is facing. Over and above the timing of the armed attack, such factors include the nature and gravity of the attack. The Chatham House Principles also include the capability of the relevant attacker, the geographical situation of the defending state and the past record of attacks. 89 Akande and Liefländer also conclude that imminence involves an assessment of the type of attack threatened, its likelihood of occurring, its gravity and timing, 90 whilst Milanovic emphasises that imminence is a certainty or likelihood criterion, rather than a temporal one. 91 Henderson appropriately labels this explanation of imminence, which includes both temporal and non-temporal factors, as 'contextual imminence' (a term that this article adopts). 92 As will be made clear, it is this relationship between a number of factors that blurs the line between imminence forming part of the armed attack trigger, and imminence inhering in the contextual determination of necessity.
Bethlehem is not, therefore, the first commentator to argue that imminence is broader than simply the timing of the anticipated armed attack. However, the USA, the UK and Australia are the first states to publicly set out such a broad and detailed policy position on this point. The stated positions of these three states, which include Bethlehem's Principle 8, indicate that determining imminence depends on non-temporal factors that relate to the wider circumstances of the threat. Justifications regarding the ongoing military action against Daesh also point in this direction. 93 The explicit nature of this state practice and the consistency of these states' reliance on Bethlehem's scholarship provide highly instructive material for exploring what imminence might look like when we venture beyond purely temporal considerations.
If the foregoing description of contextual imminence is, or becomes, the accepted standard, the temporal proximity of the anticipated attack is clearly important. However, other non-temporal factors, including the nature and likelihood of the threat and the prospect of peaceful alternatives to counter it, also pertain to the question of imminence and whether, in the circumstances, a state is facing a situation of genuine emergency that requires recourse to self-defence. Contrary to Lubell's position, therefore, where peaceful alternatives alone might be effective to prevent the attack from happening, ipso facto, such attack is not truly imminent. The necessity of selfdefence will also be absent, as force is not the only reasonable option in the circumstances. As such, contextual imminence does not establish an independent temporal requirement, meaning that the temporal remoteness of a threat does not, on its own, constitute a bar to an exercise of selfdefence. 94 This conclusion supports the logic that states should not be denied a right of self-defence in the face of a highly probable and severe threat, the realisation of which might be temporally remote, but where there will be no future opportunity to eliminate the threat. 95 This issue is discussed further in the next section.

A last window of opportunity to act
Contextual imminence encapsulates the idea of a 'last window of opportunity' for a state to respond effectively to an anticipated armed attack. This is the standard of imminence suggested by several scholars, 96 and which is reflected in Bethlehem's Principle 8 and the stated positions of the USA, the UK and Australia. By this standard, a state may only act in preemptive self-during the last window of opportunity that it has to defend itself against an armed attack that is forthcoming. The critical question, as the Tallinn Manual contends: is not the temporal proximity of the anticipatory defensive action to the prospective armed attack, but whether a failure to act at that moment would reasonably be expected to result in the State being unable to defend itself effectively when that attack actually starts. 97 Crucially, as noted in the previous section, temporal considerations are clearly significant in evaluating the last window of opportunity to act, but they do not act as an independent injunction against a defensive response. The temporal question is instead: when will the last possible window of opportunity to act close, such that the ability to avert an attack will be 94 Akande and Liefländer (n 12) 565. See also the Tallinn Manual 1.0 (n 34) 64; Marty Lederman, 'The Egan speech and the Bush Doctrine' (n 70). The ICJ has employed this reasoning in the context of necessity as a circumstance precluding wrongfulness, noting that a 'peril' appearing in the long term might be held to be 'imminent' as soon as it is established, at the relevant point in time, that the realization of that peril, however far off it might be, is not thereby any less certain and inevitable': Gabčíkovo-Nagymaros Project (Hungary v Slovakia) (judgment) [1997] ICJ Rep 7, para 54. Although potentially prima facie supportive of the foregoing analysis, as Lubell notes, we should be cautious in transposing this precedent from the laws of state responsibility into the jus ad bellum: Lubell (n 7) 703. Yet, the present article suggests that the contemporary jus ad bellum position might not be so different. 95 See Akande and Liefländer (n 12) 565. 96  lost? 98 The timing of the armed attack interacts with other factors (including the likelihood that the future armed attack will be launched, its nature and gravity and the availability of reasonable alternatives to force) to determine how long a state has to respond in order to defend itself effectively, before it is too late. The natural fear that such formulation elicits is whether the window of opportunity to act defensively might allow states to respond to latent threats long before they occur. Such fear and how it might be mitigated is discussed in section 5.1.

Imminence as necessity
There is an urgent need for each of the UK, USA and Australia to provide further clarity regarding their views on contextual imminence. Potentially dangerous uncertainties persist, and each state has ostensibly acknowledged the perils of an overly broad interpretation of self-defence. Yet, verbally at least, these states have responded to that danger by articulating and accepting constraints on anticipatory responses. They have set out, albeit imperfectly, limitations based on the particular context. Although concern rightly persists regarding the explanations put forward by this small groups of states regarding when they are prepared to resort to using force in self-defence, the governing question regarding whether a response to the threat of a future armed attack is lawful will always be whether force is the only reasonable option to counter that threat. This is the essence of necessity.
It is well known that necessity derives from the aforementioned assertion by Webster that a state's ability to act in self-defence is limited to cases in which the 'necessity of that self-defense is instant, overwhelming, and leaving no choice of means and no moment for deliberation.' 99 The ICJ has consistently affirmed necessity as a requirement of customary international law. 100 Following an armed attack, necessity conditions the exercise of the right of self-defence in response. It is a notoriously indeterminate concept, and although the Webster formula is not synonymous with the contemporary lex lata, the essential elements of necessity are derived from it. 101 For present purposes, necessity encapsulates the idea that it should be exceptional for states to use defensive force. Self-defence is a measure of last resort, where the particular situation compelled a certain course of conduct. 102 Necessity requires that there must not be any non-military alternative to using force that is practical and likely to be effective in averting a threat or bringing an attack to an end, or have a reasonable chance of so doing. 103 If a state can counter an actual or imminent armed attack by peaceful means, it has no justification for using force. 104 For it to be necessary for a state to respond to an actual or imminent armed attack using force, therefore, there must be no reasonable choice of means available to it in the particular circumstances. 105 Necessity requires that i) a state has either resorted to peaceful measures before using defensive force (and they have failed), or ii) peaceful measures are unfeasible and/or, on their own, they will be ineffective to halt, repel or prevent an armed attack. 106 Ultimately, the use of defensive force is only necessary (either on its own, or in combination with non-forceful measures) if using peaceful means exclusively is unfeasible and/or will be ineffective. This conception of necessity emphasises the availability of genuine alternatives to force at the relevant time and in the particular circumstances.
For this author, the analysis in section 3 of contextual imminence exposes a conflation of imminence and necessity. 107 A summation of the UK's stated position on imminence demonstrates this conflation: 108 is action against an identifiable threat necessary now, 109 before the last clear opportunity to act disappears, 110 or are effective alternatives to force available? 111 On this analysis, imminence stands as a proxy for necessity. Therefore, those states and scholars that adopt contextual imminence (or may do so in the future) blur the conceptual demarcation between the trigger for the exercise of the right of self-defence (being armed attacks, in this case that are imminent) and one of the requirements that conditions the exercise of that right (being necessity). 112 This is because a contextual analysis of the anticipated armed attack to determine whether it is imminent and a contextual analysis of necessity is duplicative in answering the single governing question: can a state exercise its right of self-defence now to prevent an armed attack that it anticipates in the future?
This convergence of these two previously distinct concepts might be regarded as unhelpful or even misguided. At the very least, it might cause confusion regarding how states justify their military acts and how third parties review such acts based on the justifications provided. Ultimately, however, it is necessity that provides the legal litmus test for determining whether self-defence is lawfully exercisable in response to any form of armed attack. Imminence (as associated with the armed attack) does no additional or independent legal work in answering that question. Nevertheless, this conflation between the two concepts is clearly a feature of the scholarship and the state practice referred to in this article. It must, therefore, be examined further.

Necessity: the governing factor for any right of pre-emptive selfdefence
The aforementioned conflation illustrates the ambiguities of Bethlehem's Principle 8 and the UK's, USA's and Australia's respective accounts of necessity, imminence and self-defence more generally. The jus ad bellum, by its nature, will always contain a degree of indeterminacy, yet the possibility that different states might use force based on divergent understandings of the applicable legal rules raises real concerns, not least regarding the potential for abuse of the right of self-defence and post facto accountability of the exercise of that right. For allies and coalition partners, practical operational difficulties might also arise regarding coordinating defensive military action when states interpret imminence differently. 113 Further public discussion and elucidation by these and other states on the question of imminence can only be beneficial. Such public explanations, or 'legal diplomacy', enable better international cooperation and joint action between states, as well as a common understanding of international law and a way to manage differences in interpreting legal rights and obligations. 114 Further clarification around the meaning of necessity and imminence would, for example, assist in a review of ongoing military action against Daesh and other NSAs around the world. That states continue to speak in terms of confronting a terrorist 'threat', which includes the fear of future armed attacks, demonstrates the need for greater understanding regarding how states delineate the 'threat' (see section 4.2). Absent clarification, it is difficult to assess state assertions that they want to, and do, comply with international law when resorting to armed force beyond their borders. 115 However, whereas the conflation of imminence and necessity essentially renders the former term nugatory (so far as it applies to an armed attack), the foregoing account of contextual imminence does capture the substance of how these states conceive of the necessity of responding anticipatorily to future armed attacks. The very fact of conflation might help to assuage concerns regarding an overly flexible right of self-defence. This is because the ability to exercise any right of pre-emptive self-defence (whether as a matter of lex lata or lex ferenda) will always depend on a case-by-case contextual assessment of necessity. 116 It is not, therefore, that the armed attack must be 'imminent' in any legally significant and independent sense. Rather, to quote Webster once more, it is the necessity of using force in self-defence, that must be 'instant, overwhelming, and leaving no choice of means and no moment for deliberation' (not the armed attack).
For those states and scholars that have adopted contextual imminence, therefore, 'imminence' simply describes the type of positively identified future armed attack that triggers a lawful defensive response today. 117 It is this understanding of imminence (and/or necessity if we are using the terms interchangeably) that provides defending states with a degree of flexibility and the means to protect themselves, rather than having to sit idly by and suffer an armed attack. Yet, at the same time, necessity requires that the defending state is suffering a situation of genuine emergency, whereby the recourse to force at a particular time is the only reasonable option in the circumstances. If a defending state can demonstrate that it used force when it did because failure to do so would have deprived it of the ability to defend itself effectively, on the preceding analysis, the necessity of self-defence will be established. 118 115 Regarding such claims, see ibid, 236-7; UK Attorney General Speech 2017 (n 22) 20; Australian Attorney-General Speech 2017 (n 56). 116 If a positively identified armed attack and necessity are established, any defensive action that a state takes must also conform to the customary requirement of proportionality. 117 See Milanovic (n 70), also concluding that the approach to imminence described here looks very much like necessity. See also Lederman's comments in the same blog post, regarding imminence determining necessity. Following the Soleimani strike (see n 8n 11 and accompanying text), Milanovic further elaborated on his view of how the USA and its allies construe imminence and necessity: 'an armed attack will be regarded as imminent if responding to the attack is necessary now, regardless of when and how exactly the attack will take place … An imminent attack is thus one where the attacker has committed to particular aggressive course of action which they will not desist from absent some kind of intervention in the causal chain, such as a use of force in self-defence': Milanovic (n 9). The Chatham House Principles (n 34) 967-8 also recognise this close, if not fully conflated, relationship. The authors maintain that '[t]he criterion of imminence is closely related to the requirement of necessity', and necessity may 'determine imminence'. 118 As to the evidence required to establish such necessity, see section 5.2.
Alternatively, if a use of force is not the only reasonable option in the circumstances, then imminence will not, ipso facto, be established and an exercise of self-defence will be unnecessary. As with the 1981 Osiraq incident, there will be no instant or overwhelming necessity of self-defence. 119 This incident, involving an Israeli airstrike on an Iraqi nuclear facility, is perhaps the most familiar and widely cited example in the modern Charter era pertaining generally to anticipatory self-defence. Moreover, the associated state practice reflects how states may potentially conflate imminence and necessity or use the former as a proxy for the latter. It is well known that Israel justified its actions as an act of self-defence in response to a threat of 'nuclear obliteration'. It claimed that the facility was designed to produce atomic bombs that Iraq would use to target Israel 120 and that that it was required to strike the nuclear reactor before it went 'hot'. 121 Israel has been a long-standing proponent of the right of pre-emptive self-defence, 122 yet this rationale might also be read as an Israeli claim of preventive selfdefence in respect of a non-imminent threat. 123 Regardless, all states intervening in the UN Security Council debates explicitly denounced Israel's attack, with many labelling it as an act of aggression. 124 The Council itself unanimously characterised the strike as a clear violation of the UN Charter, 125 whilst the UN General Assembly also condemned Israel's actions as aggression. 126 The Osiraq incident illustrates how states might respond to the legality of using force by reference to a lack of imminence. State rationales for denouncing Israel's actions were mixed, 127 yet a review of the UN Security Council debates and the reasons given by states for condemning the airstrikes show that, for certain states, there is little to no distinction between imminence and necessity. The bottom line for such states appeared to be that Israel was not facing a situation of emergency that meant that force was the only reasonable option at the time that Israel resorted to using it. Similarly, and conversely, the UK's justification for killing Reyaad Khan in 2015 appeared to rest largely on the lack of any reasonable alternatives to defensive force at the time he was killed by the British drone. 128 4.2. A case study of conflation 4.2.1. Self-defence against a persisting terrorist 'threat' When considering the potential conflation of necessity and imminence, we should note that it has been rare for states to rely purely on a right of anticipatory self-defence to justify using force. Absent a history of violence between the defending state and the attacker, defending states tend not to claim a right of response to the mere threat of an armed attack. This was particularly so prior to 9/11. 129 Instead, states have tended to invoke the need to respond to an actual armed attack or, when acting anticipatorily, they justify their defensive response to future armed attacks when they have already been the victim of a previous one. 130 In the latter case, the prior armed attack has evidential significance. It is treated as being indicative of further attacks, meaning that a defending state has the potential to identify a credible ongoing threat that necessitates a defensive response. 131 This phenomenon is apparent in justifications of pre-emptive action where states have been subjected to a series of armed attacks by terrorist NSAs or, as in the case of the UK's killing of Reyaad Khan, a series of actual and foiled attempts to attack. 132 Bethlehem, in his Principle 8, calls this 'a concerted pattern of continuing armed activity' 133 and, in setting out his Principles generally, emphasises the need to avert further imminent attacks by terrorist groups. 134 Likewise, the UK Attorney General, in setting out his understanding of imminence, spoke of a 'proven track record' when referring to the need to respond to ongoing terrorist threats. 135 The notion of a persisting terrorist threat poses an interesting conceptual issue for considering imminence and necessity. Ruys suggests that an analytical distinction should be made between states that are responding to prior 129 Green (n 123) 97; Gray (n 14) 170. 130 See Ruys (n 3) 342-3; Lederman, 'ASIL Speech by State Legal Adviser Egan' (n 70); Lederman, 'The Egan speech and the Bush Doctrine (n 70); Milanovic (n 70). 131 Considering a series of attacks as a whole and combining prior attacks with imminent attacks so as to collectively amount to an ongoing threat is sometimes referred to as the 'accumulation of events' or 'pin-prick' theory of self-defence and might be equated to an ongoing armed attack: see Claus Kress, 'The International Court of Justice and the "Principle of Non-Use of Force"' in Ruys and Corten (eds) (n 26) 561, 588. Although a controversial notion, the ICJ appears to have accepted in principle that a number of small-scale uses of force, individually falling below the level of an armed attack, may be accrued such that, collectively, they amount to an armed attack. This is most clearly seen in Oil Platforms (judgment) (n 100) para 64. See also the implicit acceptance of this principle in Nicaragua armed attacks with the stated objective of preventing the occurrence of additional attacks and situations of pre-emptive (or preventive) selfdefence, where there has not been a prior armed attack. In the former case, where a state is responding to a series of armed attacks, Ruys argues that the ability to react defensively reflects the prospective element of necessity, which allows defensive action to prevent further anticipated attacks from the same source. 136 The ILA likewise notes this distinction, being based on 'whether the risk of further attacks can be seen as a continuation of the initial armed attack and prevention of these being a part of the same self-defence action.' 137 A conceptual separation is thereby suggested between the ability to respond to imminent armed attacks i) when the defending state has already been the victim of a prior armed attack from the same source and defensive action is thereby preventing a reoccurrence (understood as an issue of necessity), and ii) the defending state has not yet been a victim of an armed attack and defensive action is thereby purely future-orientated (understood as an issue of pre-emption). For this author, such a distinction is not so readily apparent. Whether we conceive of a persisting terrorist threat as a series of armed attacks from the same source and/or as an issue of pre-emption, the legal analysis is essentially the same. Both types of defensive responses ultimately depend on an assessment of imminence. This is because the necessity of self-defence falls away when an armed attack is fully complete, meaning that there are no ongoing hostilities and/or there is no occupation or annexation of the defending state's territory. 138 With a series of armed attacks, and absent such occupation or annexation, it can only be the prospect of a further imminent armed attack that establishes an ongoing threat and which, in turn, maintains the necessity of self-defence. 139 Without the prospect of a further imminent armed attack, completed armed attacks are just that. They are over, and it is unlikely that force will be the only reasonable response in the circumstances. 140 Indeed, Ruys accepts that the right to 136 See Ruys (n 3) 290-1, 342-3. 137 ILA, Final Report on Aggression and the Use of Force (n 4) 11. 138 In such circumstances, non-forcible options are likely available to the defending state to resolve the issue. Any force used to respond to armed attacks that are fully complete risks being characterised as an unlawful armed reprisal. For an overview of reprisals, see Shane Darcy, 'Retaliation and Reprisal', in Weller (ed) (n 7) 879. 139 This is despite the USA's assertion that 'once a State has lawfully resorted to force in self-defense against a particular armed group following an actual or imminent armed attack by that group, it is not necessary as a matter of international law to reassess whether an armed attack is imminent prior to every subsequent action taken against that group, provided that hostilities have not ended': USA State Department Legal Adviser Speech 2016 (n 54) 239. This assertion by the USA might be a conflation of the jus in bello and jus ad bellum, however: see Henderson (n 17) 304-5. 140 If an armed attack is fully complete, as opposed to being part of a series of armed attacks, there is no pressing need to halt, repel or prevent it. Peaceful alternatives to force (e.g. diplomatic negotiations leading to reparations) may well be sufficient to resolve the issue. In such cases, the necessity of self-defence will be absent.
respond defensively to a series of armed attacks depends on compelling evidence of further imminent armed attacks. 141 This academic debate and the examples of state practice referred to in this section reflect and reveal the potential for conflation between imminence and necessity. The jus ad bellum necessity analysis regarding the right of states to respond to a persisting terrorist threat comprised of a series of armed attacks logically depends on imminence. Without the prospect of further imminent armed attacks, states may not exercise their right of selfdefence solely on the basis that they have been the victim of a prior completed armed attack. As such, we may query the value of making a distinction between the prospective element of necessity and a right of pre-emptive selfdefence against armed attacks that are imminent. They amount to one and the same thing. The bottom line is whether there is a necessity of responding militarily to a positively identified future armed attack at a particular point in time.
As noted, the context of a persisting terrorist frames how the UK, USA and Australia approach the question of imminence, as well as how these states conceive of the necessity of defensive action. For example, the prospect of a persisting terrorist threat appears to explain the UK Attorney General's adoption of the troubling part of Bethlehem's Principle 8 that relates to the uncertainties associated with anticipated armed attacks: … we will not always know where and when an attack will take place, or the precise nature of the attack. But where the evidence supports an assessment that an attack is imminent it cannot be right that a state is prevented from meeting its first duty of protecting its citizens without nailing down the specific target and timing of an attack. 142 There are a number of observations that might be made regarding this statement. For present purposes, even if one accepts a prima facie right for a state to act in self-defence against an unclear but persisting threat, an issue arises from this statement regarding when and for how long a state may act. As with the U.S. Department of State Legal Adviser's speech discussed in detail in section 5.1, the UK Attorney General's quoted assertion gives a state leeway to act in self-defence when there is a 'proven track record' of armed attacks, but the place and nature of further imminent armed attacks are uncertain. However, there is a troubling addition to this statement. By emphasising that states will not always know when an attack will take place, the UK Attorney General also appears to reserve the right to act in self-defence when the timing of the anticipated armed attack is unknown.
The UK Attorney General makes this comment in the context of a series of armed attacks by terrorist NSAs, where there have been previous armed attacks and the evidence shows that further armed attacks are imminent. 143 It might be argued, as the UK seems to, that where there exists an identified and persisting threat with no other means to counter it, a state may retain flexibility in responding militarily to that threat. 144 This flexibility extends to the timing of its defensive response. Despite the Attorney General's insistence that the UK's approach in no way dispenses with the requirement of imminence, 145 such a view has serious ramifications for our understanding of imminence. The concern is that increased flexibility regarding when a state may respond to an anticipated armed attack raises the issue of how a state is to determine when the last window of opportunity to act will close and, therefore, when the necessity of mounting a defensive response is established. The inference from the UK's stated approach might be that the 'permanent imminence' 146 of the terrorist threat absolves a state from making such a determination. If this is indeed the understanding, then a persisting terrorist threat gives rise to a continuing necessity of self-defence. This claimed flexibility raises significant concerns regarding whether necessity can act as a meaningful constraint on states exercising their right of selfdefence in such circumstances. This conclusion is explored further in the next section.

Global Coalition intervention in Syria
The most recent state practice since 2014 of the Global Coalition's military action against Daesh in Syria helps us to examine further the issue of a persistent terrorist threat. 147 The intervention provides an important case study that supports and elaborates on the foregoing analysis, revealing the conflation between imminence and necessity. The Global Coalition action also serves as a note of caution regarding how states employ contextual imminence to respond militarily to a persisting threat that includes anticipated future armed attacks. In this regard, we must first consider the UN Security Council's implicit approval of defensive action against Daesh in UN Security Council Resolution 2249, 148 as well as its repeated calls to the international 143 Ibid 144 That this is the UK's position was noted as a matter of concern by the UK's All Party Parliamentary Group on Drones Inquiry Report (n 81) 36-7. 145 UK Attorney General Speech 2017 (n 22) 17. 146  community thereafter to combat Daesh and other terrorist groups. 149 In particular, the Council has affirmed that Daesh 'has the capability and intention to carry out further attacks'. 150 For this author, this affirmation prima facie recognises an ongoing threat comprised of past and anticipated future armed attacks. Such understanding refers us back to the aforementioned point that states tend to rely on imminence when faced with a series of armed attacks in order to establish the necessity of defensive action. 151 The recognition by the UN Security Council that Daesh constitutes a permanent and active threat raises the worrying prospect of what one scholar has described as the 'permanent imminence' of anticipated armed attacks. 152 The natural concern is that, in acting to combat such threat, states might be adopting a more flexible right of defensive action. The UK Parliament has likewise expressed concerns regarding the notion of permanent imminence in relation to how the UK responds to terrorist threats. 153 In terms of state practice, Global Coalition states intervening in Syria against Daesh have tended not to rely on imminence alone to justify claims of self-defence. 154 Rather they have had recourse, in whole or in part, to the more easily established justification of the collective self-defence of Iraq. 155 Nevertheless, in support of their right of action, such states have also tended to refer to the need to combat the ongoing 'threat' posed by Daesh. Such threat is identified as being either to that state specifically, and/or to other states, and/or to international peace and security more generally. 156 In so doing, a handful of Global Coalition states have explicitly employed UN Security Council Resolution 2249 to support their self-defence claims. 157 There may be situations, therefore, where specific future armed attacks against a particular state are not identified, yet NSA terrorists are deemed by states and the UN Security Council as posing an ongoing threat justifying an enduring exercise of self-defence. Of course, the Global Coalition intervention in Syria is controversial and by no means universally supported. Yet, in considering the relationship between imminence and necessity, the Syrian example is indicative of how certain states conceive of the necessity of responding to enduring terrorist threats where further imminent armed attacks are anticipated. Moreover, the intervention in Syria is not the only case study where a significant number of states have acted in concert to combat what is perceived to be an active and persistent terrorist threat. Following 9/11, the USA justified its invocation of self-defence by reference to the need to respond to the 'ongoing threat' posed by Al-Qaeda and the need That not every Global Coalition partner operating in Syria has explicitly made a self-defence claim, or made a report to the UN Security Council, is indicative of the inconsistent state practice pertaining to this incident. This makes evaluating such practice difficult, especially as the absence of a report to the UN Security Council may be indicative of whether a state genuinely believes itself to be acting in self-defence: Nicaragua to prevent and deter further attacks. 158 The UK likewise asserted the need 'to avert the continuing threat of attacks from the same source', 159 acting in selfdefence 'in circumstances where there is evidence of further imminent attacks by terrorist groups'. 160 The UN Security Council recognised and reaffirmed the right to self-defence in those circumstances. 161 Significantly, the ensuing Operation Enduring Freedom received almost universal support from the international community. 162 In cases like these, where states refer to the need to counter a persisting terrorist threat, imminence and necessity are clearly conflated, either in whole or in part. Moreover, the idea of 'permanent imminence' strains the conceptual boundaries 163 of each concept to breaking point. Most unsettling, however, is the prospect that imminence and necessity are being disregarded as requirements for lawful self-defence against the 'threat' posed by terrorist NSAs. Alternatively, and more optimistically, we could conclude that the aforementioned examples of state practice point to the continuing necessity of self-defence against an enduring terrorist threat that is comprised of past and imminent armed attacks. The unique nature of that threat is regarded by states and the UN Security Council as constituting an enduring threat to international peace and security and peaceful alternatives to satisfactorily remove that threat have been absent. Furthermore, the UN Security Council has been unable or unwilling to take effective action using its Chapter VII enforcement powers. As such, this alternative conclusion recognises that states simply have no choice of means to counter that threat using defensive force.
In such instances, one might view the case for acting in self-defence as strong. A history of armed violence potentially serves, together with other relevant factors, as cogent evidence of future armed attacks and of the necessity of defensive action to confront them. The evidentiary burden for the defending state to show the necessity of self-defence in such circumstances is logically lighter than where there has been no prior armed attack. 164 This is because the prospect of a future armed attack is not merely speculative. The intention and capability to attack is already established, so the threat is regarded as 'genuine' and a further attack may occur at any time and without warning. 165 Indeed, intent and capability to attack seem to constitute the hallmarks of an actual, as opposed to potential, threat that requires an exercise of self-defence. 166 Yet, this argument in favour of a military response does not fit comfortably with the principles and purposes of the jus ad bellum, including the requirement of necessity. A right to respond to an enduring threat constitutes an extremely broad conception of the right of self-defence, stretching it far beyond a temporary right of states to respond to situations of emergency. If this expansive view of self-defence is accepted by states generally, whether confined to anti-terrorist defensive actions or more broadly, necessity as a customary requirement intended to condition the exercise of the right of selfdefence would seem to have very little meaning at all. Yet, despite the difficulties in applying necessity to long-standing terrorist threats posed by the likes of Al-Qaeda and Daesh, the necessity requirement can still operate as a tool to govern and assess the legality of ongoing military action against such groups. At some point in time, defending states taking anti-terrorist military action will have degraded and dismantled the operational capacity and supporting networks of the terrorist organisations to such an extent that they will have been effectively destroyed and will no longer be able to attempt or launch strategic attacks. 167 As a practical matter, being able to pinpoint a moment in time when this has occurred will be challenging. It is essentially a question of fact, but might be subject to differing opinions. Yet, there is likely be a tipping point when the military action against a particular group of NSAs has had such an effect that a counterterrorist law enforcement operation will be capable of replacing it. 168 Where hostilities are reduced in this way, force (as a measure of last resort) will not be the only reasonable option in the circumstances. Consequently, the exercise of the right of self-defence will be unnecessary and must cease.

Normative implications and mitigation
The previous sections have set out a broader conception of imminence which, like necessity, is based on context. The purpose of this section is to explore the potential ramifications of such a conclusion. If contextual imminence is, or might become, the accepted legal standard for anticipatory selfdefence, this has clear normative implications for the development of the jus ad bellum and the right of states to act militarily to counter anticipated threats. Although concerns regarding this more flexible standard have already been noted, it would be overly simplistic to dismiss contextual imminence as being too dangerous without considering why certain states might have adopted contextual imminence and what steps might be taken to mitigate genuine concerns.

The fear of abuse and mitigation of the fear
A contextual approach to imminence has led to understandable unease amongst certain scholars who regard it as too expansive. Gray queries whether such a wide conception of imminence provides any significant constraint on the use of force. 169 Indeed, an overly broad interpretation of imminence, one that includes a right of response well in advance of an anticipated armed attack, edges the jus ad bellum dangerously close to adopting a right of preventive self-defence. It potentially allows militarily powerful states to deploy force more easily beyond their borders in response to armed attacks that have not yet materialised. Such concern is reflected, for example, in the UK's All Party Parliamentary Group on Drones Inquiry Report. 170 Kattan also fears that the Bethlehem Principles reintroduce preventive self-defence 'through the back door'. 171 Green, in his commentary on the UK Attorney General's speech, notes the apparent move away from the requirement of a specific, identifiable and concrete imminent armed attack. His concern is that the UK Attorney General's understanding of imminence allows for too much 'eye-of-the-beholder discretion', which is open to abuse. 172 O'Connell likewise argues against the adoption of imminence in any form, which she views as an unwarranted and dangerous development in international law. 173 These concerns are serious and valid. They demand closer examination. The dividing line between preventive and pre-emptive self-defence is not as bright as one might wish. If we consider, for example, the last possible window of opportunity to act discussed in section 3.2(b), it will be recalled that the timing of the armed attack does not act as an independent injunction against a defensive response. As such, although the last window of opportunity to act might occur immediately before the anticipated armed attack, this is not necessarily so. Indeed, the window might present itself long before the attack occurs. 174 This latter possibility is problematic as it clearly raises the spectre of a right preventive self-defence against potential latent threats. The risk of abuse of the right of self-defence is evident in such instances. However, it is also possible to mitigate the fear if we consider in detail the theory of pre-emptive self-defence and the recent state practice referred to in this article.
First, is the timing element. Temporal considerations have a heavy impact on the possibility of making accurate predictions pertaining to future threats. 175 States may not have recourse to self-defence based on mere speculation. The more distant in time the armed attack is, the harder it will be for a state to establish that it is concretely identifiable and/or probable. This is so regardless of the nature and gravity of the threat. Although it is illogical and impossible to require absolute certainty of the impending armed attack, as we can never be certain about something that has not yet happened, the degree of uncertainty can only increase the more temporally distant a threat to a state is deemed to be. 176 This temporal hurdle tempers the risk of abuse, as assessing the probability of an armed attack becomes harder with time.
A long period between an anticipated attack and a proposed defensive response also introduces more variables into a state's decision-making process. Much could happen in that intervening period, such as the attacker reversing its course of action, peaceful measures being effective to head off the perceived threat, or the UN Security Council taking effective action that renders self-defence unnecessary. The longer the period, the more pressure there will be on the potential victim state to resolve the matter peacefully, rather than resorting to military force. 177 Therefore, it is unlikely in these circumstances that there will exist a genuine state of 'irreversible emergency' that necessitates the resort to defensive military force at a particular point in time. 178 As such, the last window of opportunity to act in the face of an anticipated armed attack is not thrown wide open to allow states to counter non-specific anticipated threats. This is so regardless of how deadly such threats are deemed to be. States and scholars have roundly rejected the idea of preventive self-defence 179  Charter 'does not allow the use of force by a State to protect perceived security interests'. 180 Belligerent rhetoric or the possible future acquisition of WMDs will not, therefore, be sufficient to trigger a right of self-defence. From the British perspective, this position is clear. The UK does not countenance a right to respond to remote threats that have not yet materialised, instead requiring a concrete anticipated armed attack and no choice of means to combat it. 181 Australia has explicitly concurred with this position. 182 From the American perspective, the rhetoric of preventive self-defence is no longer as explicit as it once was, but its core premise nevertheless potentially persists. The 2016 U.S. Department of State Legal Adviser's Speech exemplifies this fear. In it, the Legal Adviser stated: The absence of specific evidence of where an attack will take place or of the precise nature of an attack does not preclude a conclusion that an armed attack is imminent for purposes of the exercise of the right of self-defense, provided that there is a reasonable and objective basis for concluding that an armed attack is imminent. 183 Although the incumbent Biden administration's position is yet to be clarified, 184 this statement by the then Legal Adviser echoes the Bush Doctrine's assertion of a right of preventive defensive action 'even if uncertainty remains as to the time and place of the enemy's attack'. 185 Indeed, criticism of the Legal Adviser's comments at the time he gave them centred on whether the then Obama administration had implicitly adopted the Bush Doctrine through an expanded conception of imminence. 186 This unease, expressed in 2016, was not new. It echoed, for example, the furore that arose around an American Department of Justice white paper from 2011 ('DOJ White Paper'). This document set out the legal framework for evaluating when the American government could use lethal force in a foreign country against one of its own citizens believed to be a senior operational leader of Al-Qaeda or an associated force actively engaged in planning operations to kill Americans. 187 The DOJ White Paper was part of a wider attempt by the USA to justify its much-criticised programme of extraterritorial targeted killings of NSA terrorists operating in the territory of other states that were unwilling or unable to suppress that threat. 188 A particularly troubling section of the DOJ White Paper read as follows: [T]he condition that an operational leader present an 'imminent' threat of violent attack against the United States does not require the United States to have clear evidence that a specific attack on U.S. persons and interests will take place in the immediate future. 189 The issue, therefore, is whether the USA and its allies are claiming a right of self-defence against potential latent threats. On close inspection, however, the USA's position as outlined in the 2016 Legal Adviser's Speech appears to be approximately in line with both the UK and Australia in maintaining a distinction between pre-emptive and preventive self-defence.
Although the Legal Adviser does not explicitly rule out preventive selfdefence, as do his British and Australian colleagues, 190 in explaining when a state may use force in self-defence, he does confine himself to discussing actual and imminent armed attacks. 191 In respect of the meaning of imminence, he conceives of it in relation to 'an attack'. This prima facie means an identified anticipated attack, rather than a less certain threat that has not yet materialised and which is characteristic of preventive selfdefence. 192 Bethlehem, for example, understood this distinction as representing the USA's position in 2016. In response to Goldsmith, who drew parallels between the Bush Doctrine and the Obama administration's policy on imminence (as expressed by the Legal Adviser), Bethlehem emphasised that the latter policy 'is some distance, and materially different, from the broad, unilateralist brush of the [Bush Doctrine]'. 193 As such, and accepting that there remains some ambiguity, the Legal Adviser appears implicitly to rule out a right to respond defensively to potential unmaterialised threats, regardless of their nature and gravity.
Likewise, the earlier DOJ White Paper frames its analysis and conclusions within the context of a continuing (if sporadic and unpredictable) terrorist threat comprised of Al-Qaeda leaders who are 'continually planning attacks' 194 and 'a terrorist organization engaged in constant plotting against the United States'. 195 The potentially targetable Al-Qaeda leader is also posited as one who is 'actively engaged in planning operations to kill Americans'. 196 How the notion of a persisting terrorist threat informs the imminence analysis is covered in detail in section 4.2. For present purposes, there is a distinction to be drawn between the Bush Doctrine that envisaged the USA being able use force in respect of unmaterialised potential future threats and the subsequent American policy. The latter policy is certainly broad in nature, but it appears to require an identifiable concrete threat (comprised of anticipated future armed attacks) in order for the USA to be able to deploy defensive force anticipatorily.
If this distinction holds true, and a close inspection of the Legal Adviser's speech suggests that it does, then the prior policy of preventive selfdefence and the current policy of pre-emptive self-defence are indeed materially different. This is so, even accepting a contextual understanding of imminence as set out by the Legal Adviser. We might well debate the degree of distinction. There is space for states to manoeuvre within a contested area of regulation after all. However, it is one thing to say that a state may lawfully use force defensively in response to a potential threat that might or might not materialise in the future (preventive self-defence) and quite another to accept that states may take advantage of a last window of opportunity to respond to an identified attack that is forthcoming (pre-emptive self-defence). Although all the details of the anticipated armed attack might not be apparent, the USA does require 'a reasonable and objective basis for concluding that an armed attack is imminent' before military action in self-defence may be taken. 197 This conclusion assumes that there has not been a radical shift in the USA's position on this issue since the Legal Adviser's speech in 2016. 198 In this regard, it is noteworthy that The White House Legal and Policy Frameworks Report 2016 adopts the same language as the Legal Adviser (viz 'a reasonable and objective basis for concluding that an armed attack is imminent') and the current Department of Defense Law of War Manual likewise refers to imminence in respect of an attack. 199 These official documents appear to confirm a continuation of policy on this matter, therefore, rather than a return to the Bush Doctrine. Regardless of the current American policy, the Australian and UK Attorneys General have adopted the same 'reasonable and objective' standard for their respective states. 200 Following the Qassem Soleimani targeted killing in 2020, South Africa similarly confirmed that when responding in self-defence to imminent threats, 'such threats must be credible, real and objectively verifiable'. 201 The concern regarding an expansive interpretation of imminence rightly persists, however. The spectre of preventive self-defence hovers everpresent in the background of academic consideration of this issue. Most important, however, is that these case studies exemplify how elements of preventive self-defence potentially linger in the policy and decisionmaking of states regarding when they might have recourse to force in their international relations. This general unease around anticipatory self-defence is well-founded. Any exception to Article 2(4)'s prohibition on the use of force should be construed narrowly, and primacy must be given to force being exercised collectively under UN authorisation and not individually by states. Where states do depart from this framework in exceptional circumstances, they must provide convincing justifications and evidence for doing so.

The evidentiary standard
The main issue in determining how far each of the UK, USA and Australia interpret the law as they understand it to be (or are pushing it towards how they want it to be), is that it is unclear how each state interprets Bethlehem's Principle 8 and assesses the various contextual factors that go into their respective conceptualizations of imminence. A further pressing issue for our consideration of imminence is establishing an acceptable evidentiary standard for identifying a threatened armed attack. Bethlehem suggests that the 'reasonable and objective basis' formulation for determining that an armed attack is imminent 'requires that the conclusion is capable of being reliably supported with a high degree of confidence on the basis of credible and all reasonably available information.' 202 What this means in practice, however, is debatable. As with all claims of self-defence, lawfulness relies on a good faith appraisal of all the circumstances based on credible information and capable of objective assessment. 203 States may form an initial subjective view, involving Green's 'eye-of-the-beholder discretion', 204 but this view will be subject to post facto review by third parties. As the ICJ has made clear, the test of whether there is a necessity of self-defence is 'strict and objective, leaving no room for any "measure of discretion".' 205 In respect of imminent armed attacks, the need for defending states to articulate clearly their justifications for taking military action is particularly important given the uncertainties and concomitant speculation inherent in responding to possible future events.
Ideally, detailed evidence regarding a specific anticipated attack should be publicly demonstrable. 206 States, international organisations, courts and scholars are likely to hold high expectations that states that use force preemptively will share the intelligence that led them to take such action. 207 These third-party reviewers should be satisfied on the basis of such evidence that, but for pre-emptive action, the armed attack would have occurred. Yet, an ex post facto review of the lawfulness of pre-emptive action (assuming that the relevant third-party reviewer accepts the potential for lawfulness) will be significantly hindered by the fact that it is unlikely that defending states will be willing and able to release all of the intelligence relating to the threat, due to much of it being sensitive or classified. That the anticipated armed attack may have been thwarted also means that the full facts surrounding the claimed threat will likely be unavailable.
The Reyaad Khan incident illustrates the potential for an evidentiary black hole in these circumstances. Despite the UK Prime Minister having insisted publicly that there was 'clear evidence of these individuals planning and directing armed attacks against the UK', 208 such evidence was not forthcoming. The same can be said of the more recent strike by the USA against Iranian General Qassem Soleimani. The fact that President Trump and his administration provided little evidence by way of intelligence reports to support the existence of an imminent armed attack was widely commented upon and criticised by American lawmakers, journalists and academic commentators. 209 According to the USA, imminence appeared to be satisfied on the basis that 'General Soleimani was actively developing plans to attack American diplomats and service members in Iraq and throughout the region.' 210 President Trump later stated that the targets were four American embassies, although it was disputed, including by the American Defence Secretary, whether intelligence existed to justify this assertion. 211 This lack of detail in support of using force extraterritorially raises concerns regarding the transparency of decision making and holding to account states that take (potentially unlawful) military action.
When considering this issue of evidence, we should also recall the context in which Bethlehem, the USA, the UK and Australia have set out their respective conceptions of imminence. That context is the ongoing threat posed by terrorist NSAs discussed in section 4.2. Anticipated armed attacks by terrorist NSAs are often much more difficult to identify than traditional interstate threats. 212 In such cases, specific details concerning the location and nature of the attack are likely to be less clear than anticipated attacks from other states, and the precise timing of the armed attack might be impossible to ascertain with any precision. Demanding absolute certainty regarding all of the details of an anticipated terrorist attack, therefore, is too high an evidentiary burden if, at the same time, states are to have the ability to act legitimately to protect their territory and the lives of their citizens. There is truth, therefore, in the assertion that what constitutes imminence must develop to meet new circumstances. 213 It is right that states publicly justify their actions and provide sufficient supporting evidence. Yet, they should not be hobbled unduly so they cannot act before it is too late, i.e. before the last window of opportunity to respond defensively closes. An appropriate balance needs to be struck. As such, we might question the degree of specificity that is required regarding the anticipated armed attack provided that, on a good faith assessment of all the evidence, there is indeed a reasonable and objective basis for concluding that an armed attack is truly imminent. To satisfy this test, the analysis in section 5.1 suggests an apparent consensus amongst the UK, USA and Australia that a defensive response must relate to a concrete threat of an armed attack. What is sufficient in the circumstances to objectively establish that threat and the necessity of responding to it with force will need to be determined on a case-by-case basis.
To mitigate the inevitable lack of transparency regarding claims of preemptive defensive action, it will be for international courts and tribunals, international organisations, other states, and scholars to review claims of pre-emptive self-defence and, where necessary, demand satisfactory evidence to support those claims. As few jus ad bellum cases involving issues of selfdefence ever make it before the ICJ, the most natural forum for explanation, review and accountability is the UN Security Council. States should abide by their obligation under Article 51 of the UN Charter to report exercises of self-defence so that such reviews may properly occur. Furthermore, improvements to the UN Security Council's working methods regarding such communications would be an important step in enhancing transparency and accountability. 214 Although deficiencies are likely to persist regarding monitoring claims of pre-emptive self-defence, states and international organisations may nevertheless be swift to condemn anticipatory action where there is clearly no reasonable and objective basis to conclude that an armed attack is imminent. The Osiraq incident is an obvious example of such a negative response. As noted, although the reasons for denouncing Israel's airstrikes were mixed, 215 in respect of the right of pre-emptive self-defence, states were clearly concerned over the lack of any evidence pointing to an immediate or imminent threat to Israel, with certain states explicitly citing the Caroline formula. 216 Although there will always be uncertainties surrounding any claim of self-defence, therefore, evidence rather than abstract principle is likely to determine how other states respond to claims of pre-emptive selfdefence. 217

Conclusion
The right of states to act in self-defence in response to a threat of future armed attacks is a longstanding topic of debate. The notion of preventive self-defence has largely disappeared from the legal lexicon, but a right of pre-emptive self-defence against armed attacks that are imminent persists as a feature of legal scholarship and state practice. However, it occupies an unclear position in the lex lata. A right to respond to armed attacks that are imminent might be lawful based on a review of state practice. Regardless, certain states persist in justifying their actions on the basis of pre-emptive self-defence. In such cases, imminence might determine the lawfulness of a resort to self-defence and is often the key factor upon which the legitimacy of anticipatory defensive action will turn. Yet, absent a common understanding of imminence, assessing the legality of putatively defensive action is extremely difficult, if not impossible. There is also the potential for misunderstanding and conflict between allies and coalition partners based on how they interpret imminence and subsequently undertake military action. We require a better understanding of legal justifications involving imminence if we are able to comprehend the scope and content of the law and assess claims of conformity with it. Greater dialogue between states and scholars on this topic is needed. A multilateral consensus would avoid future confusion and possible conflict and the onus is on those states that have already set out their views to take the lead. 218 Although imminence has traditionally been understood as referring to the temporal proximity of an anticipated armed attack, reducing imminence purely to an issue of timing does not reflect the views of a majority of scholars. Academics have long pointed to other non-temporal contextual indicators to refer to imminence. State practice on this topic is limited, but recent examples provided by the UK, USA and Australia also indicate that imminence is not regarded by such states as being confined to the timing of the armed attack. Their explanations are not perfectly conceived, or necessarily very clear, statements of the meaning of imminence, but they are certainly indicative of how these states understand and justify their ability to respond to the threat of future armed attacks based on a range of contextual factors.
The practice of three states, on its own, is not sufficient to create custom. Yet, there is potential for such standard to be sufficiently widespread so as to become normatively constitutive. They are also important case studies to be employed to tease out how we might and/or should understand imminence. Furthermore, this most recent practice might explain past state responses to such incidents as the Osiraq incident of 1981, where the timing of a future armed attack was only one of the concerns that states articulated in highlighting the absence of imminence. 219 More recently, the notion of contextual imminence speaks to how states might justify the necessity of selfdefence against persisting terrorist threats. Such claims logically rest on whether a further armed attack is imminent. Without imminence, armed attacks that have already occurred are fully complete and the necessity of 218 The Joint Committee's Drones Report (n 50) paras 6.17-6.19, for example, called upon the UK government to take the lead internationally to provide clarity regarding the meaning of imminence. 219 See n 119n 127 and accompanying text. exercising self-defence falls away. Combatting a threat comprised of a series of past and future armed attacks relies on the notion of imminence.
If a range of contextual factors rightly explains how certain states and scholars conceive of imminence, then a circularity or conflation between imminence and necessity is revealed. The former is used as a proxy for the latter: 'imminence' is shorthand for Caroline necessity. Whether this is a whole or partial conflation is uncertain from state practice. Further clarification is needed. Yet, on this construction of imminence, we may conclude that an imminent armed attack means a positively identified future armed attack where the necessity of exercising the right of self-defence is established today. This is to say that a defending state may act in self-defence before it is too late to do so, namely before the window closes on the opportunity to mount an effective defence. Absent such pre-emptive action, the defending state will be the victim of an armed attack.
Conceived of along these lines, an imminent armed attack is a flexible concept. This flexibility raises concerns regarding the ability of international law to curtail the resort to military force. Too much latitude risks states exploiting their right of defensive action and points to a reengineered version of the much-maligned Bush Doctrine. That the threat may not be temporally proximate potentially risks the idea of preventive self-defence making a reappearance by the back door. However, in a world where states face unpredictable threats, like international terrorism, there is an understandable logic to states conceptualising imminence contextually. States should be able to respond effectively to contemporary threats, provided that force is the only reasonable choice in the circumstances and that such response is proportionate. 220 The fact that a danger is more remote does not make it any the less real or immediate if the opportunity to counter that danger will be lost without effective military action at a particular time. States must not be left defenceless in the face of real danger.
Moreover, the concern regarding an expansive interpretation of imminence is tempered precisely by how states are interpreting it. Albeit not a picture of clarity, the states referred to in this article appear to conceive of pre-emptive self-defence within the general confines of the Caroline formula. That the necessity of self-defence must be 'instant, overwhelming, and leaving no choice of means and no moment of deliberation' requires the existence of a situation of irreversible emergency. A fear of a possible unmaterialised future attack is well beyond these parameters. The circularity between, or conflation of, imminence and necessity therefore tempers the risk of abuse of the right of self-defence. Contextual imminence, whether it stands as necessity or is closely associated with it, precludes a defensive response to an unspecified latent threat, regardless of its nature or gravity.
Whereas the threatened armed attack need not be immediately anticipated, the further into the future such armed attack is envisaged by a defending state, the harder it will be for it to establish necessity. In such circumstances, there will always be a 'choice of means'. It will be difficult indeed for states to argue that peaceful alternatives to force are not available to counter such distant future possibilities. Each incident of pre-emptive self-defence will have to be reviewed on a case-by-case basis. Accountability rests on transparency. It is not utopian to demand that states be open and unequivocal, providing clear justifications regarding their decision-making process and sufficient grounds to support their claims. There must be clear evidence of a reasonable and objective basis to establish the necessity of resorting to defensive force at a particular point in time. Any response must also be proportionate.
Although necessity has greater significance when military responses to potential future events are being contemplated, this customary requirement applies to all claims of self-defence, regardless of the timing of the anticipated armed attack. As such, this author questions the legal significance of describing an armed attack as imminent. Perhaps it is time to retire this word from the legal lexicon, so far as it relates to an armed attack. While reference to imminence persists as a rhetorical tool in the words of states, the underlying legal analysis for all claims of self-defence remains the same. This is true regardless of whether such claims appear in the context of anti-terrorist operations against NSAs, whether the attacker is a state, or the perceived threat comprises conventional weapons or WMDs. Notwithstanding any conflation between imminence and necessity, the bottom line will always be whether it is necessary for a state to respond militarily to a future armed attack before it occurs. An orthodoxy regarding the right of all forms of self-defence is thereby maintained, understood along the lines of the enduring legacy of the Caroline formula.