The Controversy over Territorial State Referrals and Reflections on ICL Discourse

This article examines some of the prominent critiques concerning territorial state referrals to the ICC (also known as ‘self-referrals’), in order to test and refine the arguments. Despite wide acceptance of the drafting history claim that such referrals were not contemplated in the negotiation of the Statute, the records expressly show the opposite. Critiques about the potential for political manipulation are significant, but reflect a tension inherent to all international criminal justice efforts, regardless of trigger mechanism. The concern about ‘selective externalization’ of prosecutions is compelling; however, the legal and normative implications are more subtle and multi-faceted than is commonly assumed. The article also ven-tures some preliminary observations about international criminal law (ICL) discourse in general. One observation is that although discourse focuses on points of disagreement, the interpretive community also implicitly absorbs assumptions that limit and shape legal debate. For example, the widespread but incorrect assumption that territorial state referrals were ‘not contemplated’ by the drafters has eclipsed the actual drafting history, and has framed the present legal debate concerning the supposed ‘innovation’. A related observation concerns the prospect of assuming a single vision or model of the Court and allowing that model to dictate interpretation. Multiple plausible models are compatible with the Statute, and open-minded assessment of the merits and implications of each is needed. Such models of the Court may include, for example, an ‘antagonistic’ model, a ‘catalyst’ model, a ‘reverse cooperation’ model and a ‘facility’ model.


Introduction
The literature surrounding the International Criminal Court (ICC) has developed in various ways over the last 15 years. When the ICC was being established, criticisms of the ICC were almost entirely 'external', in that they came from those opposed to the ICC itself, most notably from among right-leaning thinkers in the United States. 1 Conversely, apart from arguing that the Statute should be stronger still, ICC supporters tended to close ranks and gloss over problems with the Court in order to help maintain support for a vital institution in its nascence. As the Court's roots became established, some scholars quite rightly ushered in a new phase, in which supporters of ICL offered criticisms of decisions and approaches of the Court. The position of supportive critic (or critical supporter) was initially an outlier position, but is now a popular one.
Such a cultural shift is welcome. However, robust scholarship also requires that the critiques themselves be subject to scrutiny. 2 As will be demonstrated below, several critiques that have gained great traction in the interpretive community appear to be predicated on perplexing foundations. Given that the ICC already faces formidable challenges from its actual missteps, difficulties and 'teething problems', 3 it must be appropriate to examine, clarify and reduce misplaced concerns.
It is for these reasons and in this spirit that this article aims to examine the most prominent critiques advanced against territorial state referrals (commonly but imprecisely described as 'self-referrals'). This article will focus on (i) the textual claim that such referrals are based on a 'creative' interpretation; (ii) the drafting history claim that such referrals were not contemplated by the drafters; (iii) concerns about political manipulation by states; (iv) concerns that such referrals cannot displace the admissibility test; and (v) concerns about the impact on the duty to prosecute. 4 The article will demonstrate that the first two claims are verifiably empirically incorrect. The third argument, while intriguing, is a special case of a more general challenge inherent in international criminal law and not a product of any trigger mechanism. The fourth argument is correct, but attempts to refute a position that the Court has never 1 See e.g. J. Bolton, 'The Global Prosecutors: Hunting War Criminals in the Name of Utopia', 78 Foreign Affairs (1999) 157; A.P. Rubin, 'Challenging the Conventional Wisdom: Another View of the International Criminal Court' , 52 Journal of International Affairs (1999) 783. 2 If discourse is critical without also being auto-reflective, it raises risks of new bandwagon effects, i.e. a risk that the community will shift from one that is uncritically accepting to one that is uncritically critical. 3 A. Cassese,'Is the ICC Still Having Teething Problems?' 4 Journal of International Criminal Justice (JICJ) (2006) 434. 4 For reasons of space, this article focuses on the most prominent arguments and hence will not address other arguments, such as that self-referrals lead to 'queue-jumping' despite gravity requirements; that referrals suggested or invited by the Prosecutor are not 'real' referrals; or that referrals inappropriately circumvent the Pre-Trial Chamber.
advanced. The fifth concern is a significant one, worthy of attention, but the interpretive and normative implications are more nuanced than is commonly assumed. This article will also venture some reflections on ICL discourse itself. One observation concerns the surprising extent to which ICL discourse transmits and continually reconstructs a shared understanding of ICL's history, goals and problems and the significance of texts and cases. While this process is essential to the building of a coherent discipline, it also entails that opinion within the interpretive community can coalesce around incorrect propositions, in which case our received wisdom is distorted. For example, it is today an almost universally-shared belief that the drafters of the Rome Statute did not contemplate a state referral of a situation on its own territory and thus that such referrals are an 'innovation'. Surprisingly, as will be shown below, the belief is directly opposite to the actual drafting history, which show repeated express anticipation of such referrals. Yet, the contrary belief has become widely shared and has framed legal debate concerning the alleged innovation. Another example is the common belief that the 'unwilling or unable' exception in Article 17 is the entirety of the Article 17 test, a belief so firmly entrenched that the Court's efforts to point out that Article 17 contains other words and prior conditions have been met with disbelief and confusion. The power of such beliefs within an interpretive community, and their impact on discourse, warrants serious reflection.
A further observation concerns the danger of assuming a single model of how the Court will function and elevating that expectation into a perceived Statute requirement. Such assumptions are often asserted through the malleable medium of 'drafters' intent': the proposition 'I did not expect this' becomes articulated as 'the drafters did not want this'. It is argued below that we need to cultivate an awareness of our pre-conceptions and an openness to other viable visions. The conclusion will suggest that multiple models of how the Court should function are equally consistent with the Statute.

Critiques of Territorial State Referrals
Article 14 of the ICC Statute authorizes states parties to refer situations to the Prosecutor, who then must assess the situation under Article 53 to determine whether to initiate an investigation. The first three state party referrals ç from Uganda, the Democratic Republic of the Congo (DRC) and the Central African Republic (CAR) ç all concerned situations on the territory of the referring state. As mentioned above, 'self-referral' is the term now popularly used when a state party refers a situation on its own territory (although it will be argued below that the term is ambiguous and unhelpful). 5 5 The ambiguities and limited utility of the term are discussed below in Sections 4.B. and 8.A.

Territorial State Referrals and Reflections on ICL Discourse
This development has generated a significant literature. It is not possible here to examine all of the literature, so this article will consider the most influential and oft-cited works on the topic. One of the most prolific writers on this theme is William Schabas, who argues that territorial state referrals are an 'invention of the Office of the Prosecutor'. 6 This analysis will focus on the arguments advanced in two of his seminal pieces on the topic, one on prosecutorial discretion and judicial activism 7 and one offering 'uncomplimentary thoughts' on complementarity 8 while drawing on other works that buttress his critique. 9 Professor Schabas objects that such referrals '[flow] from a creative interpretation of Article 14 of the Rome Statute that was not seriously contemplated by the 1998 Diplomatic Conference and during the prior negotiations'. 10 Elsewhere, he writes that self-referrals are based on a 'novel interpretation of Article 14 ::: of which there is not a trace in the travaux pre¤ paratoires'. 11 These and other passages advance two claims: (i) a textual claim that self-referrals are based on a 'creative' interpretation and (ii) a claim that such referrals were not contemplated during the drafting history. Schabas asserts that self-referrals are an 'interpretative deviation' and an 'opportunistic construction' 12 of the ICC Statute driven by the desire to generate activity. 13 He also raises a teleological argument, that these 'flawed sophomoric experiments' 14 present a 'trap' for the Court, 15 which 'distort[s] the proper role of the Court', 16 and he accordingly expresses the hope that the Court will grow out of this phase. 17 In a similar vein, an influential article by Mahnoush Arsanjani and W. Michael Reisman expresses strong concerns about the 'law-in-action' of the ICC, 18 a term connoting the 'divergences between nominal and actual law '. 19 They warn that ICC interpretations are developing in unexpected ways and in departure from the text of the Statute. 20 They assert that 'before and during the Rome negotiations, no one ::: assumed that governments would want to invite the future court to investigate and prosecute crimes that occurred in their territory', and that there is 'no indication that the drafters ever contemplated' such referrals. 21 They also argue that 'such a referral does not seem to meet the requirement of admissibility'; 22 that a system of such referral 'could open the way to ::: selective externalization of difficult cases'; 23 and that 'the innovative allowance of voluntary referral in future cases may take the ICC into areas where the drafters of the Rome Statute had not wished to tread.' 24 Such concerns have been echoed in subsequent literature, up to and including the most recent article on the topic in this journal, in which Andreas Mu « ller and Ignaz Stegmiller reiterate that such referrals were 'not foreseen by the delegates at Rome'; 25 that such referrals are 'often' accompanied by 'understandings' between the referring state and the Court; 26 and that they are not a 'panacea' but a 'patient' and should be avoided. 27

The Textual Claim: A 'Creative' Interpretation?
We will start with the claim that self-referrals reflect a 'creative' or 'novel' interpretation of the text of the Statute. 28 An interpretation is typically considered 'creative' if it strains against the ordinary meaning of the words in their context, for example by assigning an unusual meaning or reading in words.
The text of Article 14(1) ICC Statute, with the key passage emphasized, is as follows: A State Party may refer to the Prosecutor a situation in which one or more crimes within the jurisdiction of the Court appear to have been committed requesting the Prosecutor to investigate the situation for the purpose of determining whether one or more specific persons should be charged with the commission of such crimes. The essence of Article 14(1) is contained in the opening clause: a 'State Party' may refer a 'situation'. No exception is stated. The non-italicized passages provide clarifications but do not limit that proposition. 29 Thus, as a matter of ordinary meaning and classical logic, where a given entity falls within the term 'state party' and a given phenomenon falls within the term 'situation', that entity may refer that phenomenon.
Suppose, for example, that the DRC wishes to refer to the Prosecutor 'the situation in the territory of the Democratic Republic of the Congo since 1 July 2002'. 30 An application of the ordinary meaning of the words and basic syntax is as follows: (a) a 'State Party' may refer a 'situation' (Article 14); (b) the DRC is a 'State Party'; 31 (c) the situation in the territory of the DRC since July 2002 is a 'situation'; 32 (d) therefore, the DRC may refer the situation in DRC since 2002.
This syllogism is a basic application of the text. No reading in or strained interpretations were involved.
Importantly, Professor Schabas acknowledges that '''self-referral'' is not explicitly prohibited in the text' and concedes that 'legal instruments take on a life of their own, and creative interpretation of their provisions is usually desirable'. 33 However, these concessions still slightly misconceive the textual situation. It is not merely that the Statute fails to prohibit territorial state referrals: the Statute provides an express general authorization for state parties to make referrals without any qualifiers. Thus, it is excluding a class of state party referrals that would require 'creative' or extra-textual arguments to restrict the general authorization in the text. Accordingly, on a textual level, the charge of 'creative' interpretation reverses the situation. The ICC jurisprudence permitting territorial state referrals is based on the ordinary meaning of the words. To preclude self-referrals would require that we read in an unexpressed exception, such as 'except a situation 29 The second segment is an adjectival clause modifying 'situation' , clarifying that it is a situation in which one or more crimes appear to have been committed. The third segment provides clarification on the significance of a referral: it is a request to the Prosecutor to investigate and possibly bring charges. Jurisprudence has clarified that situations may be defined by territorial and temporal parameters and in some cases personal parameters; see e. concerning its own territory'. Of course, there might be sound extra-textual arguments to read in such a limitation; such arguments are considered in the following sections. But in so far as text is the starting point of international legal interpretation, the debate should not be framed not as whether there is a legal basis to allow territorial state referrals, but whether there is a legal basis to exclude them. Similarly, Professor Schabas asserts that 'there had never been even the slightest suggestion, in the drafting history of the Statute, that a State might refer a case ''against itself'''. 36 He later repeats, in equally categorical terms, that '[t]here is not a trace in the travaux pre¤ paratoires or in the various commentaries by participants in the drafting process to suggest that a State referring a case against itself was ever contemplated'. 37 He asserts that the paradigm of 'a consensual relationship between the State of territorial jurisdiction and the international body ::: was not what was contemplated when the Rome Statute was drafted'. 38 Accordingly, self-referrals 'distort the vision of those who struggled so hard to create the Court'. 39 These arguments, if their premises are true, could present a plausible argument to depart from the text. However, as a preliminary point, let us recall that a failure by the drafters to foresee an eventuality would not by itself constitute an interpretive argument sufficient to overcome the text; a stronger form of the claim is required. For example, it is often said, probably correctly, 34 Arsanjani and Reisman, supra note 18, at 386^387. 35 Ibid., at 397. 36 Schabas, supra note 8, at 7 (emphasis added). 37 Schabas, supra note 8, at 13 (emphasis added); Schabas, supra note 7, at 751 (emphasis added). 38 Schabas, supra note 8, at 16. 39 Schabas, supra note 7, at 760. As will be discussed below (Section 8), this belief as to drafting history is often repeated in the literature, including recently by Mu « ller and Stegmiller, supra note 25, at 1269: 'the mechanism has not operated as intended', self-referral 'was not foreseen by the delegates at Rome'. that the drafters of the Statute did not expect that states with crimes on their territory would ratify the Statute, as it was widely believed that states would not want to expose themselves to international intervention. 40 Nonetheless, such ratifications have occurred. They have not been rejected on the grounds that the drafters did not foresee ratifications in such circumstances. Instead, we apply the terms of the Statute to the development. Similarly, even if it were true that the drafters did not imagine that a state party would further invite international scrutiny referring a situation on its own territory, that lack of imagination would not by itself suffice to prevent an application of the text of Article 14. 41 Accordingly, to overcome the text of the Statute and to read in a limit, a stronger form of the 'drafters did not foresee' argument is required. One would have to show, for example, that the outcome is manifestly absurd or unreasonable or that the drafters ascribed a 'special meaning' to the term. 42 Setting that aside, the drafting history objections encounter even graver difficulties when we test them against the drafting history records. The critiques assert, rather categorically, that there is 'no indication', 'not a trace' and 'not the slightest hint' in the drafting history that territorial state referrals were ever contemplated. Surprisingly, the cited records establish the opposite proposition.

The
For example, the early work of the International Law Commission (ILC) introducing the concept of a 'complaint' expressly contemplated a complaint by a territorial state, and indeed proposed a territorial link as one of three required grounds for standing to make a complaint. 43 Similarly, the official report of the debates of the Ad Hoc Committee reads: [S]ome delegations expressed the view that any State party to the statute should be entitled to lodge a complaint with respect to the [core crimes] :::. However, the view was also expressed that only the States concerned that had a direct interest in the case, such as the territorial State, the custodial State or the State of nationality of the victim or suspect ::: should be entitled to lodge complaints. 44 40  The record expressly shows that territorial state referrals were contemplated.
More remarkably still, it shows that the controversy was not whether to allow territorial state referrals; it was whether states parties without a direct interest such as territory should also be allowed to make referrals.
The same configuration appears throughout the drafting history records, from the ad hoc Committee to the Preparatory Committee to the Rome Diplomatic Conference. The official reports of the Preparatory Committee attest to the same debate and the same two camps: Some delegations felt that only those States parties to the statute with an interest in the case should be able to lodge a complaint. Interested States were identified as the custodial State, the State where the crime was committed, the State of nationality of the suspect, the State whose nationals were victims and the State which was the target of the crime. Some other delegations opined that the crimes under the statute were, by their nature, of concern to the international community as a whole. 45 The draft texts on state party referrals prepared by the Preparatory Committee and forwarded to the Rome Conference also reflect the same two major options. One option would allow referrals only by states with a 'direct interest', defined as '(a) a State on the territory of which the act [or omission] in question occurred; (b) a State of custody; (c) a State of the nationality of a suspect; (d) a State of the nationality of victims.' 46 The other option would allow referrals by any state party, without restricting the right to interested states.
The Rome Conference records again report the same two camps. In debates, delegations explicitly contemplated referrals by 'the state concerned', with some delegations going further and arguing that only the state concerned should be able to make referrals. 47 The ultimate decision was to reject such limits, so that any state party may refer any situation. 48 Academic commentary by conference participants Philippe Kirsch and the present author, cited by Professor Schabas as support for the view that territorial state referrals were not contemplated, 49 also expressly notes that territorial state referrals were discussed and that the controversy related to whether disinterested states should be allowed to make referrals: Many delegations supported a proposal to restrict the ability to refer a situation to those States Parties with an 'interest' in the matter. 'Interested' States would include those whose territory was the locus of the crimes in question, or whose nationals were victims or perpetrators, or those States with custody over particular perpetrators. Other delegations were of the view that these crimes were, by their very nature, of fundamental concern to the international community as a whole ::: . Article 14 accordingly permits any State Party to refer a situation, without being restricted to 'interested' States. 50 The implications for the common narrative of the drafting history are remarkable. It is commonly asserted that there is 'no trace' of such referrals in the records, whereas we find instead that such referrals were a recurring and explicit topic of deliberation throughout the negotiations, from the first to the final discussions. Equally startlingly, territorial state referrals were not only expressly discussed, they were the uncontroversial common ground between the two camps. One camp wanted to allow referrals only by states with an 'interest' in the situation (e.g. a linkage such as territory or nationality); the other camp would not limit referrals to interested states and would allow any state party to refer any situation. The overlap is that both camps supported referrals by interested states, which expressly included territorial states. Ironically, it was referrals by states parties without a direct interest that attracted controversy.

B. Alternative Critique: The Alleged Requirement to Target Another State
One could attempt an alternative critique, in the hopes of reconciling the 'not contemplated by drafters' claim with the drafting history. There is a latent ambiguity in the term 'self-referral'. Commentators routinely present a dichotomy between (i) a referral concerning a state's own territory and (ii) a referral targeting another state. However, those categories are neither mutually exclusive nor jointly exhaustive. They are not mutually exclusive, because a referral could concern a state's own territory yet implicate another state. They are not jointly exhaustive, because a referral could concern a state's own nationals acting on another territory. 51 This article has conformed to the popular usage of 'self-referral', which focuses on the 'same territory' aspect (i.e. territorial state referrals). However, an alternative critique could emphasize instead the alleged obligation to target another state. Thus, the alternative critique would (i) acknowledge, as one must, that territorial state referrals were indeed expressly contemplated, but (ii) insist that referrals must target 'another state' and hence assert that (iii) a state may refer a situation on its territory but only where another state is implicated in the commission of the crimes. The alternative critique could also draw on a potentially plausible 'special meaning' argument advanced by Professor Schabas. Schabas points out that earlier drafts of the Statute used the term 'complaint' which, he argues, 'clearly' connotes a complaint about another state. 52 He does not argue that the drafting history at any point actually states that a complaint had to be against another state; the connotation is simply asserted to be 'clear'. 53 From this connotation, we are invited to draw an inference that the drafters did not contemplate a referral concerning a state's own territory or not implicating another state. The ascribed connotation is perfectly plausible if one thinks of an analogy of a human rights treaty, in which states issue complaints against other states.
However, the inference is not as self-evident as the critique suggests. First, it is rather state-centric to assume that 'complaints' can only be about acts carried out on behalf of other states and not, for example, non-state actors, given the growing power of non-state actors. 54 The state-centric assumption, while traditional in human rights law, should not be transposed lightly to international criminal law, concerned as it is with acts of all individuals, whether or not directed by a state. Second, as the Statute is an international criminal law instrument, it is equally plausible (or perhaps more plausible) that the term 'complaint' carries its connotation not from human rights law but criminal law, i.e., the notitia criminis that triggers enquiry, in which case it would lack the ascribed antagonistic connotation. Indeed, the ILC commentary accompanying the draft provision on 'complaint' reflects the more neutral concept of a triggering mechanism: 'The court is envisaged as a facility available to states parties to its statute, and in certain cases to the UN Security Council. The complaint is the mechanism that invokes this facility and initiates the preliminary phase of the criminal procedure.' 55 Even larger problems arise when we consider the entirety of the records. First, as demonstrated above, the records show extensive discussion of the prospects of territorial state referral. It is implausible that delegates, in discussions of territorial state referrals, had in mind yet repeatedly failed to mention an exception that would almost entirely eclipse the rule; i.e. territorial state referrals are allowed but only in the minority of situations when another state is active on the territory. It would be particularly odd that delegates submitted formal proposals for binding legal texts expressly authorizing territorial state referrals and yet forgot to mention an exception that precludes most of the rule. 52 Schabas, supra note 8, at 13^14. 53 Ibid., at 13. 54 See also, P. Akhavan, 'Self-Referrals Before the International Criminal Court: Are States the Villains or the Victims of Atrocities', 21 Criminal Law Forum (2010) 103^120. 55 ILC, supra note 43, at 45 (emphasis added). The ILC expresses its view that the Court will rely on the cooperation of the triggering state to secure 'evidence, witnesses and the like' (ibid., at 46), indicating an assumption that the triggering state is sufficiently linked to the situation to have such evidence. The ILC again reiterates the neutral concept of 'complaint' as 'intended to bring to the attention of the court the apparent commission of a crime' and refers to the Court as a 'mechanism that should be available' when necessary (ibid., at 46).
Second, the posited assumption is incompatible with the content of the arguments articulated in the debates. Whereas the popular narrative today is that the drafters assumed or preferred referrals by disinterested or victim states (described as 'referrals in the classic sense' 56 ), the records show a debate focused on an entirely different question. The debates revolved around which states had a sufficient 'interest' to give them a legitimate basis to call for ICC intervention. Suggested grounds were the familiar links of jurisdiction (e.g. territory, active nationality, passive nationality, custody). One argument given for requiring such links was that third states were considered more likely to bring frivolous, politically motivated or unsubstantiated complaints. 57 A consideration alluded to by the ILC was the need for concrete cooperation by the referring state. 58 Indeed, one of the oft-mentioned grounds to legitimize a referral,'nationality of the suspects', is even more irreconcilable with the claim that drafters assumed that referrals must be directed against another state. It does not seem plausible that delegates discussing this ground implicitly meant only the exceedingly rare circumstances where another state was directing the referring state's nationals, and failed to mention this caveat.
In conclusion, the narrative about drafters preferring or assuming referrals by third states is not supported by the records and is contradicted by the proposals made and arguments advanced during the negotiations. The records show that delegations were discussing whether objective links of proximity to the situation should be required, with an apparent preoccupation being to reduce frivolous referrals. Ultimately, the decision was that, given the nature of the crimes, all states parties have sufficient 'interest' to refer any situation. 59

C. Implications
The foregoing findings have implications at two levels. At a doctrinal level, the arguments to read an unexpressed limit into Article 14 based on drafters' intent are unsustainable. Their premises are unsupported by, and indeed contrary to, the drafting history. At a broader level, the findings raise intriguing implications about our discourse and legal reasoning. As will be discussed in the conclusion (Section 8), the belief that drafters did not foresee such referrals is now widespread in the interpretive community and has framed legal debate. As a result, a straightforward application of the Statute text, in a context specifically foreseen by drafters, has been framed as an innovation and has generated an enormous literature. Given that the referrals were made less 56  than ten years after the Rome Conference, the transformation of our collective recollection is a humbling reminder that we must always be vigilant about the assumptions that frame debate.

Teleological Argument: Manipulation of the Court
One of the most prominent and recurring teleological critiques is that states may provide referrals to advance their own agenda, which may diverge from the aims of the ICC. 60 As Schabas notes,'when a State is actively engaged in initiation of the process, there is potential for manipulation. In effect, the state quite predictably uses the international institution to pursue its enemies.' 61 Professor Schabas notes that, in the context of Uganda, President Museveni was likely resorting to the ICC as one more tool to bring pressure on the Lord's Resistance Army. 62 He warns that 'self-referral, far from being an expedient to provide a fledgling institution with some cases, is actually a trap. If a State refers a situation against itself, that is, against its rebels, it is doing so with a result in mind.' 63 Similarly, Arsanjani and Reisman warn that states may seek to co-opt or use the ICC to pursue adversaries or to resolve internal political problems. 64 However, the familiar observation that states may have their own agendas and interests is nothing terribly new, and does not alter the legal applicability of Article 14. As the International Court of Justice (ICJ) has noted, the motives of a state or body of states in triggering a legal process are legally irrelevant. 65 60 Another critique outlined by Arsanjani and Reisman is that territorial state referrals may draw the Court into 'political problems' (supra note 18, at 393) and 'political struggles' in which 'despite all the attendant violence ::: negotiation and settlement' might be the only practicable mode of resolution (at 392). States will refer difficult situations, such as Ituri, where 'the state of chaos and insecurity and the prevalence of land mines' make it 'difficult, if not impossible, for the Office of the Prosecutor to conduct any meaningful investigation on the ground or to have any access to any accused.' (at 398) Thus, 'the failure of governments will simply become the failure of the ICC' because 'the Court will find that it is unable to conduct any serious investigation and prosecution.' (at 395) This argument invites significant questions such as: (i) whether labeling a mass violence situation as 'political' should per se preclude international criminal justice; (ii) whether such concerns should be considered under the 'interests of justice' rather than by precluding a priori a class of referrals; (iii) whether there exist any situations of mass atrocity that are not 'political'; and (iv) whether self-referrals are any more likely to lead to politicized situations than other triggers (consider the situation in Darfur, which has divided a continent and which was referred to the ICC by the Security Council). Moreover, the prediction of 'de facto deferrals' due to the impossibility of investigating situations such as Ituri appears in retrospect to be unduly pessimistic. None of this affects the legal validity of the act; the role of the Prosecutor is to carry out a preliminary examination of the situation, initiate an investigation if warranted and carry out the investigation in accordance with his or her Statute duties. Nonetheless, the insight is still valuable, not as a legal objection to territorial state referrals, but rather as a sound note of caution in planning and managing an investigation. 67 It is certainly true that states might vacillate from supportive to obstructionist if the work of the ICC diverges from their interests. However, this hazard can arise in any relationship and any investigation, irrespective of the trigger mechanism. 68 Indeed, the root of the problem is the intractable paradox of independence and dependence, which is inherent to all international justice efforts. That paradox engages yet another paradox, that of detachment and engagement: the Prosecutor must be analytically detached, in order to select situations and cases in a principled manner based on Statute criteria, yet the Prosecutor must also engage with the world, in order to build and maintain cooperation from various actors.
It is understandable that observers of the ICC would have reservations about relationships with territorial states, because these tensions, independence^dependence and detachment^engagement, are particularly visible in relation to territorial states. Territorial states ç from the eyes of the international prosecutor ç have a duality of nature. As a matter of international law, a territorial state is the lawful authority in the territory, whose cooperation is required to carry out meaningful operations on its soil. As a matter of criminal law, those authorities are also potential targets of investigation. Combining these strands, in international criminal law, territorial states are both lawful authorities whose cooperation is valuable, and also objects of analysis and investigation. Thus, for example, a prosecutor might need to request military security support in a dangerous region, and at the same time to reiterate that senior army members remain subjects of analysis and, where warranted, investigation and prosecution. 69 The tensions of dependence^independence and detach-ment^engagement create a profoundly difficult line, but it is one that every Court's legal assessment of request to the ICJ, which 'invites it to discharge an essentially judicial task'). 66 See also, Akahavan, supra note 54. 67 See e.g. C. Kress, '''Self-Referrals'' and ''Waivers of Complementarity''', 2 JICJ (2004) 944^948; P.
Gaeta, 'Is the Practice of ''Self-Referrals'' a Sound Start for the ICC?' 2 JICJ (2004) 949^952. 68 As one hopefully hypothetical example, the Security Council might agree at one moment to refer a situation, and then its members' enthusiasm may subsequently wane for any number of reasons. 69 Such a principle is expressed in the OTP draft Policy Paper on Preliminary Examinations, October 2010, available at http://www.icc-cpi.int/Menus/ICC/StructureþofþtheþCourt/ OfficeþofþtheþProsecutor/PoliciesþandþStrategies/ (visited 29 November 2010). Selection international prosecutor is obliged to walk. The tensions arise regardless of trigger mechanism. Schabas warns that 'the moment [the Prosecutor] seeks charges against pro-government forces, cooperation from the government is likely to become less enthusiastic.' 70 This prediction certainly may prove accurate. However, a negative reaction would be an issue of cooperation, not of trigger mechanisms. The challenge of securing arrests of high officials is a perennial problem of international criminal justice, not one peculiar to self-referrals. Moreover, it is premature to speculate that referring states will cease to cooperate if the acts of their agents are found to meet the threshold for case selection. The history of international criminal law is replete with 'realist' warnings about what states will never do, which have frequently proven incorrect. This is not to advocate a rosy triumphalism, but rather to inject a dose of healthy scepticism about Realpolitik predictions. As experience with President Milos› evic¤ and others has shown, such predictions underestimate the unpredictable and dynamic ways in which the world unfolds.
Professor Schabas argues that self-referrals entail an 'implied compact with governments' 71 or a 'degree of complicity' between the state and the Office of the Prosecutor (OTP). 72 Interestingly, the prospect of improper understandings attached to territorial state referrals has graduated in the literature from speculation to presumed fact. Thus, in the most recent article, Mu« ller and Stegmiller assert as a simple matter of fact that self-referrals 'will often be accompanied and burdened by ''understandings'' between The Hague and the referring state that can taint the subsequent proceedings.' 73 Such assertions presume the impropriety that they purport to prove. Moreover, if such understandings existed, or if inappropriate selection decisions were made, then it would be the understandings and the decisions that violate independence and impartiality, not the referral. 74 There remains of course a problem of perceptions. Territorial state referrals may entail a particular practical disadvantage of inviting worries and speculation about secret or implicit understandings. Thus, it would be prudent for a prosecutor to be particularly clear about independence and impartiality and to explain case selection decisions. It would seem that the OTP has not done enough in this respect to address such questions.
decisions 'shall not be altered by the presumed or known wishes of any party or by the cooperation seeking process (at x 34). ' 70 Schabas, supra note 8, at 19. 71 Schabas, supra note 7, at 753. 72 Ibid., at 751. 73 Mu« ller and Stegmiller, supra note 25, at 1285 (emphasis added). 74 Unless one were to respond that all territorial state referrals must entail inappropriate understandings. However, to assume without evidence such automatic subservience of the ICC to power is to reject a priori the basic possibility of impartial international justice.
Professor Schabas raises important questions about case selection decisions which are important and which merit careful analysis. 75 Without engaging in the case selection debate (which would be beyond the scope of this article), one may query the presumed linkage to self-referrals and the apparent assumption that the OTP would trade (and has traded) its independence and impartiality for a referral, as the motivation to do so is unexplained. Even without a territorial state referral, investigations may still be triggered in other ways, such as proprio motu initiation. In the DRC situation, the state agreed to provide a referral; in Kenya, the state did not. In both situations, an investigation was launched. A referral has some modest value as a signal to those inside and outside the country who might raise objections about perceived interventionism, and as a signal to officials that they should provide cooperation going beyond the very limited regime of Part IX of the Statute. However, it has not been explained how those modest benefits are thought to have motivated the OTP to enter into the implicit compacts that appear to be assumed.
Observers are entitled to their concerns, fears or suspicions about the potential for territorial states to withhold cooperation or try to influence case selection, or that a prosecutor's selection decisions might be influenced by cooperation needs. However, it must be recalled these risks and challenges are perennial: all situations feature a territorial state, regardless of trigger mechanism. Indeed, a territorial state referral may provide some modest additional leverage against the state, because of the domestic and international reputational costs of a reversal of position.

The Argument that Territorial State Referrals Do Not 'Trump'Admissibility
A recurring critique aims to refute the perceived argument that a territorial state referral somehow satisfies or bypasses the admissibility test. In their seminal work in this strand of thought, Arsanjani and Reisman frame 'the relevant question' as 'whether voluntary referral will henceforth trump what had heretofore been deemed to be inherent requirements for ICC admissibility'. 76 Note their framing of the issue: 'we refer to situations in which the sole basis for satisfying the Court's admissibility test is the referral.' 77 They conclude that 'such 75 Up to the present, the OTP has yet to bring charges against agents of a referring state. Accordingly, Schabas argues that the selection decisions are driven by political considerations: Schabas, supra note 7, at 753; Schabas, supra note 8, at 33. The OTP account is that the case selection decisions were based on gravity. Space here does not permit an adequate dissection of the intricate issues of gravity and case selection, on which there are very different views as to the appropriate standards and the facts on the ground, and hence the correctness of the case selection decisions to date. 76 Arsanjani and Reisman, supra note 18, at 387. 77 Ibid., at 383. a referral does not seem to meet the requirements of admissibility under Article 17 of the Statute', 78 'in strict legal terms, a voluntary referral such as the one by Uganda appears to fail to satisfy the threshold for admissibility set out in Article 17 of the Statute.' 79 The Arsanjani and Reisman analysis is hampered by a few technical errors, such as conflating trigger mechanisms with jurisdiction, thereby giving rise to fears that a referral might be 'withdrawn', 80 and conflating trigger mechanisms with admissibility. 81 Putting those technical details aside, however, the core of the critique has remained a focus in the literature. Following this framing of the question, a great volume of literature has been devoted to analysing the perceived argument that self-referrals 'trump', alter or set aside Article 17. 82 What is puzzling about this critique and others in the same vein is that the argument which they purport to refute is one that has never been advanced or relied upon by the Court. The critiques are commonly conceived as responding to the experts group on complementarity, the OTP or the Chambers, but so far none of those sources have ever argued that the trigger mechanism trumps or even affects the admissibility test. 83 To date, admissibility in the cases before the ICC has been based on the absence of national proceedings and thus the failure to satisfy the explicit requirement of Article 17(1)(a)^(c) that, for a case to be rendered inadmissible, it must be in the process of being 78 Ibid., at 392. 79 Ibid., at 395. 80 For example, Arsanjani and Reisman assert that the DRC referral 'vastly expanded the ICC's geographic scope of jurisdiction': supra note 18, at 398. However, the referral did not expand the ICC's territorial jurisdictional by a single centimetre: the Court already has jurisdiction over DRC territory by virtue of its ratification. The conception of the referral as the basis for jurisdiction makes possible the concerns that referrals might be 'withdrawn' (ibid., at 395 and 397). However, a referral is a procedural mechanism that can trigger the initiation of an investigation; the Court's jurisdiction is not predicated on the referral. 81 This section discusses the separation between trigger mechanisms and the admissibility test.
As a further minor point, in the cited passages Arsanjani and Reisman repeatedly speak of applying the admissibility test to the referral itself, whereas the admissibility test is of course applied to the existence and nature of national proceedings in relation to cases of interest. 82 See e.g. G. Gaja, 'Issues of Admissibility in Case of Self-referrals', in M. Politi  The confusion about 'self-referrals' 'trumping' admissibility is connected to confusion surrounding 'admissibility-due-to-inaction'. The perceived innovation of territorial state referral was followed by the perceived innovation of admissibility-due-to-inaction, which appears to have led some to assume post hoc ergo propter hoc: the latter must be caused by the former. Legally, however, the concepts are entirely separate.
Confusion over admissibility-due-to-inaction flows from a widespread and surprisingly entrenched belief that the Article 17 complementarity test is a one-step test focused exclusively on the all too famous prongs of 'unwilling' and 'unable'. For example, Schabas, Arsanjani and Reisman have firmly insisted that Article 17 consists only of the two requirements of 'unwilling' or 'unable'. 85 When confronted with assertions that Article 17 contains express prior requirements that the state must be investigating or prosecuting, Schabas has dismissed such suggestions as 'gloss', 86 'activist', 87 'invented, 88 an 'unwritten third criterion', 89 an 'interpretive deviation', 90 an 'opportunistic construction', 91 'read in' despite 'the rather clear wording' 92 and 'not in the ICC Statute'. 93 Similarly, Arsanjani and Reisman dismiss the suggestion as 'gloss' 94 and 'innovative', 95 the 'glossators'advancing it only 'purport' to find the requirement in the text, 96  Schabas and Williams, supra note 9, at 615^616. 86 Schabas, supra note 7, at 757. 87 Ibid., at 757. 88 Ibid. 89 Schabas, supra note 6, at 349; Schabas, 'Rise and Fall', supra note 9. 90 Schabas, supra note 7, at 760. 91 Ibid. 92 Ibid. 93 Ibid., at 757. There is reason to hope that the interpretive gap is narrowing. For example Professor Schabas in a recent and important work on the Rome Statute acknowledges and addresses the proceedings requirement: supra note 6, at 341 and 345. However, other passages leave this unclear, by asserting that the unwilling and unable tests must be applied even where there are no proceedings to which to apply them. Schabas acknowledges that such an approach 'defies common sense' but asserts that it is required by the text (ibid., at 341).With respect, it is not explained why an exception would be considered when the condition precedent from which it derogates is not satisfied. As the Appeals Chamber has noted in Katanga, supra note 83, at xx 74^79, to elevate the exception into a general requirement 'put[s] the cart before the horse' and is 'irreconcilable with the wording of the provision' . 94 Supra note 18, at 397. 95 Ibid., at 396. 96 Ibid., at 390 and 391. 97 Ibid., at 395. The phenomenon is a curious and remarkably persistent interpretive disconnect; the author demonstrates and explores this 'mysterious mysteriousness of complementarity' elsewhere. 98 This trend of regarding 'unwilling' and 'unable' as the entire Article 17 test and not detecting a prior requirement of national proceedings is remarkably common in the literature. 99 Due to the overlaps with the issues in this section, it is worthwhile to summarize that the author has sought to demonstrate the following propositions. First, the requirement of national proceedings is expressly stated in 55 words of Article 17, which require either that 'the case is being investigated or prosecuted by a State' (17(1)(a)), 'the case has been investigated by a State' (17(1)(b)) or that it has been tried (17(1)(c)) if the case is to be rendered inadmissible under complementarity. Second, the 'unwilling'/'unable' tests are expressly stated as exceptions that arise if and only if there are national proceedings to which they can be applied. Third, the foregoing is not only stated in the text but also confirmed by the drafting history, contextual interpretation and teleological analysis. 100 Curiously, the 'textual' critiques of admissibility-due-to-inaction have not offered an alternative interpretation but have simply failed to engage with the logically confined role of the unwilling/unable exception or the 55 prior words expressly requiring proceedings. The author has sought to demonstrate that this disconnect has led to good faith but misplaced accusations of Statute violations and has obscured many intriguing questions about viable interpretations. 101 In conclusion, while the critiques are correct that the choice of trigger mechanism does not displace the admissibility test, they are purporting to refute a position that the Court has not to date advanced. The admissibility of cases to date has been based on the absence of national proceedings and thus the non-satisfaction of Article 17's requirements for inadmissibility.

Territorial State Referrals and the Externalization of Cases
The remaining critique is a significant one. As the previous section clarified, territorial state referrals and admissibility-due-to-inaction are legally separate concepts and can arise independently of each other. 102 However, it is also quite correct that the two may in practice be 'intimately linked'. 103 A state wishing to see cases brought to the ICC can provide a territorial state referral and also decline to initiate national proceedings, thereby inviting ICC investigation and also clearing away admissibility barriers. As Arsanjani and Reisman note, this 'could open the way to using the Court as a backup to national judicial systems that are otherwise competent and to the selective externalization of difficult cases' 104 and 'encourage governments to externalize to the Court ::: domestic political problems'. 105 This concern is compelling. The legal implications and the normative implications are however more multifaceted than the critiques appear to contemplate. The cited critiques oversimplify the legal implications because they incorrectly perceive territorial state referrals and admissibility-due-to-inaction as departures from the Statute; hence, the perceived solution is simple: to return to the Statute. However, given that both concepts flow from the Statute text, the actual legal question is whether to adopt extra-textual interpretive arguments to limit or alter the text of Articles 14 or 17. Alternatively, the question arises whether there are policy steps to be taken, working within the Statute as it is, to mitigate such concerns.
The cited critiques also oversimplify the normative question, by assuming that deliberate externalization of cases to the ICC is ipso facto undesirable. However, that normative question can only be judged against some model of how the Court should work, and there is more than one plausible model. If one assumes a model in which the Court's role is exclusively catalytic, or in which the Court can only operate in antagonism with territorial states, then self-referral plus inaction is always undesirable. If one assumes that the aim of the relevant provisions is to protect state sovereignty, then self-referral plus inaction is always unobjectionable, because the state has invited the justice intervention. On the myriad of plausible models in between those extremes, the more subtle answer to the normative question of desirability is: it depends.
Thus, the critique, while not necessarily incorrect, fails to consider the full richness and complexity of different viable models or visions of the Court's role. Those different models can generate different answers as to when self-referrals may be 'good'or 'bad', which in turn guides interpretive arguments or operational policies on how to respond to such referrals. A space must be opened to allow debate on the plausible models.
An essential part of the puzzle is the preamble to the Statute, including the sixth paragraph, which 'recall[s] that it is the duty of every State to exercise 102 Admissibility-due-to-inaction flows from the absence of proceedings and hence nonsatisfaction of the proceedings requirements in Art. 17(1)(a)^(c), regardless of trigger mechanisms. Conversely, a territorial state referral may or may not be accompanied by national judicial inactivity; the territorial state or another state may be investigating or prosecuting relevant cases. 103 Schabas, supra note 6, at 343. 104 Arsanjani and Reisman, supra note 18, at 390. 105 Ibid., at 392. its criminal jurisdiction over those responsible for international crimes'. 106 The passage is important for contextual and teleological interpretation; any plausible model must in some way reconcile Articles 14 and 17 and the preamble. Indeed, the Statute poses a complex picture, reflecting the diversity of visions that underlie it. On the one hand, the text of Article 14 allows state party referrals without any qualifier, and the text of Article 17 expressly leaves a case admissible where there are and have been no genuine national proceedings in relation to the case. 107 On the other hand, there is also a preambular reference to a duty to exercise criminal jurisdiction. How are the permissive Statute provisions to be reconciled with such a duty?
Any plausible account must consider and address many probing questions. What exactly does the preamble reference to a duty mean? 108 How can it be fostered? The experts group on complementarity took first steps to open this debate, in a passage that advances six propositions (numbers are added here to delineate the separate ideas): [1] Article 17 specifies the consequences for admissibility where a state is investigating or prosecuting, [2] but does not expressly oblige states to act.
[3] However, paragraph 6 of the preamble refers to the 'duty' of States to exercise criminal jurisdiction.
[4] While the preamble does not as such create legal obligations, the provisions of the Statute may be interpreted in the light of the preamble.
[5] The duty to 'exercise criminal jurisdiction' should be read in a manner consistent with the customary obligation aut dedere aut judicaire, and is therefore satisfied by extradition and surrender, since those are criminal proceedings that result in prosecution.
[6] However, as noted above, the reference to a duty also reflects the spirit of the Statute that States are intended to carry the main burden of investigating and prosecuting. This is necessary for the effective operation of the ICC. 109 The issue is whether that refers to the traditional duty aut dedere aut judicaire or a more novel duty to 'prosecute-and-not-to-extradite'. There is a valuable debate to be had here, but we are not yet engaged in it.
There are many viable ways to interpret the meaning of the duty, as well as its scope, limitations and impact. Since the passage 'recalls' a pre-existing duty, rather than imposing a new conventional duty, the experts group suggested a legal interpretation according with established maxim aut dedere aut judicaire (proposition 5). On this view, the provision would serve, in the words of one noted expert on the preamble, as 'a sort of Martens clause' recalling that just because crimes are not dealt with by the Statute, 'this does not mean that there is now impunity for them'. 111 However, the experts group suggested that it could in addition reflect a 'spirit' or systemic goal 112 that national systems should in general carry the main burden of prosecuting (proposition 6). Accordingly, the experts group identified methods to try to encourage national proceedings and to prevent overburdening of the Court, including applying public pressure and declining situations based on 'the interests of justice '. 113 Nonetheless, notwithstanding the experts group approach, plausible arguments could be advanced to interpret the preamble as referring to a narrower 'duty-to-prosecute-and-not-to-extradite'. However, to pursue this route, questions must be thoughtfully addressed. To which states does this more stringent duty apply (territorial states, states of nationality, others)? Is the duty absolute or are there limits (such as force majeure)? 114 How can it be reconciled with the Statute text? What is the significance of 'recalling' a duty in the preamble as opposed to imposing a duty in the operative text of the treaty? Failure to reflect on these questions will result in interpretations with scant legal support and absurd consequences.
An obvious interpretive strategy to give effect to the duty would be to creatively read limits into Article 14 or 17 to preclude some referrals or to declare cases inadmissible before the ICC even where there are no national proceedings. The problem with such solutions is that they do not advance the goal of ensuring state action. Legally, paralyzing the ICC even where no state is prosecuting only fosters impunity. 115 Thus, it appears that the way to encourage state action is not through rejecting a class of referrals or rewriting the admissibility test, but rather through a prosecutorial policy of trying to encourage states, where feasible, to carry out effective proceedings themselves.
Thus, the most urgent questions on the appropriate role of the Court may lie not in the realm of legal interpretation but of policy formulation. For example, there is no legal question that cases are admissible where states are inactive, given the express terms of Article 17. But when the OTP first encounters a scenario of state inaction, what can and should it do as a policy matter? In what circumstances can it and should it encourage and press for national action? In what circumstances can it and should it accept state inaction to facilitate admissibility, as part of a burden-sharing arrangement to ensure that effective justice is done? The Statute does not answer these questions. Crucially, the appropriate demarcation points between a policy of encouraging national action and a policy of burden-sharing remain to be clarified. The experts group on complementarity suggested some scenarios in which burden sharing might be the desirable option, 116 but a valuable debate awaits, because different viable visions are entirely plausible, 117 depending inter alia on whether one focuses exclusively on the aim of promoting national proceedings or whether one also considers other aims of the preamble, such as effectiveness and cooperation in ending impunity. 118

A. Implications
This article has examined some of the most prominent critiques of territorial state referrals. The textual critique is demonstrably unsound; given that Article 14 provides an express general authorization of state party referrals, it is reading in a restriction on a class of referrals that would require 'creative' interpretation. The drafting history critique, that the drafters never contemplated territorial state referrals, is also demonstrably incorrect; such referrals were repeatedly and expressly contemplated by drafters. The political critique, that states may have political agendas and may wish to cooperate selectively, is significant. However, the challenge of maintaining independence despite significant practical dependence on cooperation partners, most particularly territorial states, is a tension inherent in all cooperation relationships, and not a product of any trigger mechanism. The objection that territorial state referrals do not trump the admissibility test is correct, but it is misconceived in that it aims to refute a position that the Court has never advanced or relied upon. Finally, the concern that such referrals may lead to externalization of cases is compelling; however, the legal and normative implications are not as straightforward as is often assumed. A space must be created for debate on different plausible models. The term 'self-referral' is of very limited value, because of its ambiguity, legal irrelevance and perceived novelty. The term is ambiguous, because of the flawed dichotomy by which it is commonly defined. 119 The term is legally useless, because it does not denote a distinct legal category; a territorial state referral is simply a subset of state party referral. 120 The perceived novelty of the concept has led to unfruitful efforts to imbue it with distinctive legal characteristics that it does not possess. While a policy debate will surely continue on the advantages and disadvantages of different trigger mechanisms, the legal debate is misconceived, because all of the trigger mechanisms are legal.
On a broader level, this article also opened with observations about discourse in international criminal law. An initial observation was that discourse within the epistemic community appears to have shifted toward more enthusiastic criticism of the Court. This is valuable, as the assumptions and decisions of the Court must be open to thoughtful scrutiny. However, the critiques themselves must be open to thoughtful scrutiny, as otherwise we risk falling into new and untested collective opinions ('groupthink'). A rigorous scholarship to test and refine arguments and advance knowledge must surely be our collective endeavour.
This article also invites reflection on ICL discourse itself, and the manner in which assumptions may influence what we perceive as evident and shape our legal debate. This is important for two reasons. First, ICL discourse and the potential influence of discursive assumptions is in itself a phenomenon worthy of enquiry. Second, awareness of any such trends and tendencies may help us to improve our debate and understanding.

B. The Power of Collective Assumptions to Frame Debate
While ICL takes the form of robust debate, it also has an implicit function of consensus formation. As we select issues for debate, we implicitly absorb and acquiesce in countless assumptions and propositions. The process is essential for discipline formation. However, it is valuable to be aware of the extent to which we absorb beliefs that bound our discourse and shape our legal debate and the extent to which the interpretive community continually reconstructs the history, aims and reality of ICL. This awareness counsels us to be vigilant and sceptical of even our most widely shared assumptions, as inaccurate shared assumptions may have a remarkable influence on discourse.
For example, as was shown above, the drafting history as generally presented today appears to be not merely incorrect, but almost the opposite of 119 See Section 4.B. For example, some might argue that a referral of a state's nationals is a self-referral (on the targeting approach), others may argue that it is not (on the territory approach). Any such debate would be unproductive since no legal consequence arises from the term. 120 A similar point is made by Mu « ller and Stegmiller, supra note 25, at 1275. the recorded drafting history. It is today a nearly universally shared commonplace that the Rome Statute drafters did not contemplate a state referring a situation on its own territory. 121 In the literature, this or similar propositions are routinely asserted, sometimes with a reference to Schabas or to Arsanjani and Reisman, 122 and sometimes as an uncontroverted proposition. 123 Furthermore, what is today narrated as the consensus understanding among drafters ç i.e. that referrals should be only by third states but not by territorial states ç does not appear to have been advanced by any faction in the debate.
Ironically, it appears the positions described as 'consensus' and those of which there was 'not the slightest trace' have become perfectly reversed. 124 This article does not suggest that the collective belief is causally attributable to any particular commentators, 125 nor does it attempt to explain how collective belief came to diverge so dramatically from history. However, one brief speculation will be ventured in passing. Consider the difference between the proposition that the drafters thought an eventuality unlikely and that they did not contemplate it at all. It is often reported that most or many drafters thought that a state party referral would be rare or unlikely. 126 This author, relying on his own subjective and fallible perceptions and recollections, would agree that this was a common assessment during the negotiations. 127 Thus, while the drafters clearly contemplated territorial state referrals as a legal option, it is also plausible they considered the prospect unlikely in practice. Legally, of course, such probability assessments of delegates would not affect interpretation: whatever their private predictions, the drafters made legal provision for referrals, and those provisions apply. However, the distinction may help explain the belief shift: 'the drafters considered it unlikely' may have become elided with 'the drafters never contemplated it'.
In any event, a reconstructed history appears to have eclipsed the actual drafting history. The shared belief about history has framed and shaped the present-day legal debate. Today, both opponents and supporters of territorial state referrals generally start from the shared premise that they are an unforeseen innovation. 128 Thus, the issue has been framed as one of whether to indulge the perceived innovation instead of as a routine application of the Statute. Accordingly, discussion has often been cast as whether other creative reinterpretations of the Statute are required to absorb the supposed innovation. For example, a great deal has been written on whether Article 17 must be adjusted, set aside or applied to self-referrals, 129 and on whether other provisions, such as Article 59 (arrest warrants), must also be adapted to accommodate self-referrals. 130 It is intriguing that a straightforward textual application of Article 14, in a context that was expressly contemplated and uncontroversial in the negotiation of the text, has generated such a substantial literature and debate, framed in terms of innovation. 131 Similarly, it appears that collective belief can influence our understanding not only of history but also of text. For example, as touched on above, the one-step unwilling/unable conception of complementarity has become so firmly entrenched in the popular imagination that when the Court faithfully applies the other express terms of Article 17, many react with outright disbelief and insist that the Court is 'inventing' terms contrary to the 'rather clear' text, resulting in various recriminations against the Court. 132 Such examples 127 Indeed, the assessment that state party and Security Council referrals would be unlikely was a major argument advanced for the proprio motu procedure: see e.g. Ad Hoc Committee, supra note 44, at x 113; PrepCom, supra note 45, at x 165. The same set of assumptions that led delegates to doubt the likelihood of state party referrals and Security Council referrals would presumably have led them to assume that referrals by a territorial state would be particularly unlikely. 128 See examples, supra note 121. Some analyses are silent as to whether territorial state referrals were foreseen by drafters; the author has found none that start from the assumption that they were expressly contemplated. 129 See examples supra note 82. 130 El Zeidy, supra note 82, at 463^465. 131 This is not to suggest that the analyses have been unimportant; indeed several intriguing arguments have been advanced: supra note 82. Moreover, the mere fact that such referrals were indeed foreseen does not preclude the need for further theorizing and reflection. The concern here is the manner in which debate has been framed. 132 See examples in Robinson, supra notes 98, 100. counsel humility, care, scepticism and open-mindedness about even our most deeply held and shared beliefs about the Statute, its interpretation, its drafting history and its objects and purpose.

C. Open-mindedness to Multiple Plausible Models
The previous theme considered transmission of assumptions. The remaining theme concerns the prospect that we may have expectations about the model by which the Court will work, and then elevate those expectations into requirements. We may do this by assuming that only one vision is consistent with the Statute, or that only one vision was shared monolithically by the drafters.
For example, one might expect an antagonist model, in which states will resist intervention, resist admissibility and resist cooperation, and the Prosecutor must stare down the obstructionist state, using measures such as media, public shaming and the pressure of other states to bring the recalcitrant state to heel. Several passages of the cited critiques appear to assume a model of the ICC locked in antagonism with states, and to regard departures from this model as contrary to the 'Statute' or to 'drafters' intent', even when the actual sources might not be quite so conclusive. This is illustrated and reinforced in repeated assertions about the intent, vision, design and philosophy behind the Statute: the paradigm of 'a consensual relationship between the State of territorial jurisdiction and the international body' 'was not what was contemplated when the Rome Statute was drafted'; 133 an antagonistic relationship such as that with Sudan 'better suits the underlying philosophy of the Rome Statute'; 134 territorial state referrals 'distort the vision of those who struggled so hard to create the Court'; 135 the Court's approach draws energies away from 'situations and cases for which it was truly designed'; 136 the Court's approach 'distort[s] the proper role of the ICC' and may 'demoralize the strongest supporters of the Court'. 137 This article does not aim to reject the antagonist model. First, it is perfectly likely that the antagonist model was in the minds of many at the Rome 133 Schabas, supra note 8, at 16. But see Part 9 of the Statute, suggesting that territorial state cooperation could not have been completely unexpected to delegates. 134 Ibid., at 33. 135 Schabas, supra note 7, at 760. But see the explicit references to territorial state referrals in the drafting history, above. 136 Schabas, supra note 8, at 33. 137 Schabas, supra note 7, at 761. Arsanjani and Reisman, supra note 18, similarly assert an antagonistic model that was assumed by 'the drafters'; see e.g., 'the drafters of the Rome Statute assumed that governments would be reluctant ::: to surrender the national criminal jurisdiction to the Court' (at 386), 'no one assumed that governments would want to invite' the court to investigation crimes on their own territory (at 386); allowing such referrals risks taking the ICC 'into areas where the drafters of the Rome Statute had not wished to tread' (at 397). As will be suggested below, some or many drafters may have had such views, but one should not assume monolithic views without greater substantiation; the record shows many conflicting views on important issues.