Protect and punish: norm linkage and international responses to mass atrocities

Since being founded in 2002, the International Criminal Court has frequently intervened in ongoing conflicts and alongside other forms of coercive intervention, specifically sanctions and military measures. In this article, I argue that this pattern has been enabled by governments engaging in strategic norm linkage. To justify their positions on both judicial and non-judicial interventions, governments have discursively linked international prosecutions to the protection of civilians – in specific ways that have favoured joint judicial and non-judicial crisis responses. My argument, which I test through qualitative and quantitative content analyses of UN Security Council debates, contributes not only to debates on the politics of international criminal justice, but also to general theory-building on international norm dynamics. Adding to recent research on norm complexity and norm interactions, my study underlines and disaggregates the potential for discursive agency at the intersection of multiple international norms.


Introduction
The International Criminal Court (ICC), created in 2002 to prosecute genocide, crimes against humanity, war crimes and aggression, has since become active in different world regions. Its activities have followed a pattern that is puzzling in several respects. First, most ICC investigations took place amid ongoing conflicts or atrocities. This pattern departs from the pre-ICC practice of initiating prosecutions after the cessation of violence. 1 It has persisted although investigations in ongoing conflicts have been found to face practical impediments (Unger and Wierda, 2009) and, in some cases, complicate peace processes (Ku and Nzelibe, 2006;Rodman, 2011;Snyder and Vinjamuri, 2003) and peacekeeping (Buitelaar and Hirschmann, 2021). Second, ICC interventions have often -but not always -occurred alongside other forms of intervention, specifically sanctions and military operations. This raises the question of whether and how judicial and non-judicial interventions are related. 2 In this article, I seek to understand the observable pattern of ICC interventions by mapping the discursive strategies of justification that have enabled it. I argue that the court's in-conflict interventions alongside other forms of intervention have been made possible by strategic norm linkage, that is, by governments constructing discursive connections between prosecution and the protection of civilians to justify their positions on judicial and non-judicial interventions.
International Relations (IR) scholars have studied both the prosecution of atrocity crimes and the protection of civilians as (contested) international norms stipulating how states should collectively respond to mass violence (Bower, 2019;Gholiagha, 2022;Mills and Bloomfield, 2018;Welsh, 2013). 3 The prosecution norm holds that atrocity crimes should be prosecuted internationally if the responsible state is 'unwilling or unable' to do so (ICC Statute Art. 17.1); 4 the protection norm demands that outside states collectively protect civilians from severe violence -with coercive tools if necessary -if the responsible state fails to do so. 5 A simple explanation of the co-occurrence of judicial and non-judicial interventions is that protection and prosecution norms are applied in parallel. Both emerged in response to the same events, most notably the Rwandan genocide and the Balkan wars of the 1990s. Hence, cases of mass violence should trigger simultaneous, but independent deliberations about judicial and non-judicial responses. The notion of two independent norms appears to be confirmed by the ICC's own claims that 'the "interests of justice" must not be confused with the interests of peace' and that '[t]he prospect of peace negotiations is therefore not a factor' in the court's decision-making (Bensouda, 2012: 440-441).
And yet, I argue in this article, the story of two parallel, independent norms does not fully capture observable intervention patterns. Protection and prosecution norms have become closely linked by strategic political actors -in specific ways that have favoured the co-use of judicial and non-judicial measures in ongoing crises.
My argument builds on and contributes to existing scholarship on international criminal justice. Scholars have long discussed whether international criminal tribunals deter both future and ongoing violence -and can thus serve as tools of protection -or endanger peace and protection at least in the short run (Mendeloff, 2018;Snyder and Vinjamuri, 2003;Weerdesteijn and Holá, 2020;Wegner, 2015). My study does not add to this discussion by studying the effects of prosecutions on mass violence. Instead, it shows how the debate itself has been politicized and used by political actors. In analysing their discursive strategies, I draw on anecdotal arguments in the international criminal justice literature that highlight potential political interaction effects between prosecution and protection.
First, some observers note critically that justice advocates tend to 'sell' international tribunals (Vinjamuri, 2010: 199) by emphasizing their ability to contain violence against civilians -a rhetorical move that encourages the ICC's use as a crisis-response tool. According to the critics, this framing is problematic because it not only papers over tensions between justice and protection, but also raises unrealistic expectations that could ultimately undermine support for justice (Arbour, 2014;Bosco, 2012;Vinjamuri, 2010: 205).
A second argument holds that international criminal justice is used to legitimize the use of force against perpetrators (Mégret, 2011;Nouwen and Werner, 2010;Stahn, 2015). This strategy could account for the frequent co-use of judicial and non-judicial interventions. While it could help overcome barriers to effective protection (Carpenter, 2022), it has also been argued to facilitate ineffective or problematic uses of force (Nouwen and Werner, 2010).
A third, contrary argument holds that international tribunals and the claim that they protect civilians could be used as a 'fig leaf' by governments seeking to avoid more robust -and more costly -protective measures (Mégret, 2001: 209; see also Rudolph, 2001: 681-682;Smith, 2002). This 'moral hazard' dynamic ) could explain not only ICC interventions in cases in which no other collective responses can be agreed, but also parallel judicial and non-judicial interventions: governments not contributing to, for instance, a collective military intervention could support a parallel ICC intervention to show that they are 'doing something' about atrocities. Either way, moral hazard could lead to an under-provision of protection.
All three arguments are largely based on anecdotal evidence, 6 but continue to resonate in discussions about international criminal justice. For instance, the 'fig leaf' argument is echoed in a recent commentary on the ICC and Ukraine: Tribunals can get in the way of deterring war crimes by substituting them for genuine action. They can create the impression that the international community is holding an aggressor accountable while enabling the international community to avoid the hard choices that could actually save civilian lives and enforce the U.N. Charter [. . .]. And in Putin's case, this stopping power can come only from military force, not solely from judges. (Carpenter, 2022) In my study, I subject anecdotal propositions about political interaction effects between prosecution and protection to a systematic, comparative test. Theoretically, I conceptualize them as highlighting different justification strategies based on different forms of norm linkage. Methodologically, I use a qualitative and quantitative content analysis of UN Security Council debates about the Protection of Civilians in Armed Conflict and about conflicts in Darfur/Sudan, Libya and Syria to investigate (1) to what extent actors have relied on norm linkage in debating judicial interventions, and (2) which linkage-based justification strategies they have employed. While I do not aim to provide a complete explanation of individual interventions decisions, I seek to capture prevalent discursive patterns that favour specific intervention outcomes and thus contribute to the pattern observed at the outset.
My analysis shows that political actors have relied heavily on two justification strategies favouring clustered judicial and non-judicial interventions in ongoing conflicts: they have used linkage arguments not only to legitimize judicial interventions as a tool of protection, but also to boost support for concurrent sanctions and military interventions. In contrast, my study casts doubt on the notion that international justice is used as a 'fig leaf' to avoid more costly protective interventions. In light of these findings, I revisit the argument that justice and protection should be more 'segregated' (Arbour, 2014). I conclude that while a general de-linking of both norms is neither feasible nor desirable, political actors can and should avoid using linkage to legitimize sanctions and military interventions.
My study contributes to this strand of research by demonstrating and disaggregating the potential for and consequences of discursive agency at the intersection of multiple norms. It has become widely accepted in IR norms research that political actors re-interpret and contest individual norms even after their adoption. 7 This should logically also extend to relations between different norms, yet most researchers treat norm compatibilities, conflicts or clusters as exogenous conditions that can facilitate or obstruct the spread, evolution or decay of individual norms (e.g. Deitelhoff and Zimmermann, 2019: 13;Lantis and Wunderlich, 2018;Saltnes, 2017;Sandholtz, 2008: 101, 109;Wiener and Puetter, 2009: 14). In contrast, my study highlights how the same set of norms can be connected through different, even contradictory linkage arguments which actors vary strategically across situations.
Early norms research already showed that non-governmental 'norm entrepreneurs' often connect emerging norms to pre-existing norms to boost support for them (e.g. Finnemore and Sikkink, 1998: 98). My study expands on this observation by highlighting governmental actors' (re)construction of norm relationships long past the initial phase of norm emergence, including in discourses of norm application. 8 Drawing on scholarship on legitimacy and justification in international politics (e.g. Kornprobst, 2014;Niemann, 2018), I argue that norm linkage allows governments to enlarge the repertoire of justifications they can use to defend their policy preferences, but also to reverse them and come to agreements. Over time, certain linkage arguments may gain discursive dominance, thus again constraining actors' repertoires.
In the following, I develop an analytical framework that distinguishes different types of linkage arguments and strategic uses of linkage. This framework allows me to systematize and test anecdotal arguments on the politics of judicial interventions, but can also be applied to other issues to capture and compare norm linkage in different areas of international politics.

Conceptualizing norm linkage
In line with other scholars (particularly ; see also Jurkovich, 2020), I conceive of norms as consisting of different elements: statements of guiding values (e.g. justice); descriptions of problems to which these values should be applied (e.g. mass atrocities); descriptions of situations in which norms are applicable (e.g. conflicts involving ICC member states); descriptions of behaviours deemed to resolve or cause the problem (e.g. ICC investigations); and prescriptive or proscriptive rules stating whether norm addressees ought to engage in or refrain from a certain behaviour (e.g. initiate an ICC investigation). 9 Political actors engage with these components in applying and (re)interpreting norms. They may leave all elements in place and debate merely whether a given problem matches a norm's abstract descriptions (e.g. are these acts atrocity crimes?), but they may also change the norm by replacing or adding components .
Norm linkage is a special case of such discursive re-arrangement. It can be understood as a discursive act through which political actors attach an individual norm component to two (or more) distinct international norms. For this study, I distinguish three types of linkage arguments that discursively construct different norm relationships.
First, political actors can construct a norm conflict ( Figure 1) by claiming that in a given situation (e.g. the Syrian civil war), a behaviour (prosecution) concerning a problem (mass atrocities) which is subject to a rule A (prescription to prosecute) in light of the guiding value of norm A (justice) is subject to a different rule B (prohibition to prosecute) in light of the guiding value of norm B (protecting civilians) concerning the same problem.
Second, actors can construct a norm synergy ( Figure 2) by claiming that a behavioural rule (applying to a given problem in a given situation) does not only serve the guiding value of the norm A it is primarily associated with, but also the guiding value of an additional norm B. The argument that ICC investigations both serve justice and help protect civilians is a case in point.  Third, norms can be framed as a norm cluster ( Figure 3) by claiming that different behavioural rules stipulated by norms A and B are based on the same guiding value and/ or problem perception . Norm entrepreneurs often use this form of linkage in claiming that a new norm they advocate for is consistent with the values of preexisting norms or addresses the same type of problem (Carpenter, 2007;Finnemore and Sikkink, 1998: 98). Yet, cluster arguments can also emerge past the stage of norm acceptance, as exemplified by the notion of a 'responsibility to protect and prosecute' (Ralph and Gallagher, 2015). This phrase, coined years after the establishment of the ICC and the UN's adoption of the Responsibility to Protect (R2P) principle, expresses the idea that prosecution and protection are 'twin tasks' (Thakur and Weiss, 2009: 42) that apply distinct values (protection/justice) to the same problem (mass atrocities).

Strategic norm linkage and public justifications
Why do political actors construct different norm relationships? I understand norm linkage as a resource which actors can use strategically to broaden the discursive repertoire they dispose of to publicly justify their policy choices.
This argument is based on three assumptions. First, in line with constructivist and pragmatist approaches to legitimacy and justification in international politics, I assume that political actors on the global stage are under a constant 'imperative' to justify their decisions by giving public reasons for them (Kornprobst, 2014: 197; see also Gadinger, 2016;Hurd, 2007b: 196;Reus-Smit, 2007: 44). In doing so, they seek to convince their audiences -governments, domestic publics, international media and expert communities -that their actions are legitimate in the sense of being based on generally accepted standards. To achieve this, actors must draw on discursive repertoires shared by their audiences (Kornprobst, 2014;Niemann, 2018: 6). International norms are important parts of such collective repertoires.
Second, I assume that actors interpret and use norms strategically to support their political aims, that is, as a discursive resource they can tap into (Bower, 2015;Hurd, 2005Hurd, , 2007b. I understand 'strategic' in a wide sense, as referring to the selective and creative enactment of justificatory repertoires adapted to the aims that actors pursue in specific situations (Gadinger, 2016: 196; see also Kornprobst, 2014: 202). This can, but does not need to imply that public justifications are insincere. Whether or not an actor privately believes in a public justification she advances, we can assume that she uses the argument because she believes it to bolster the legitimacy of her position (Hurd, 2007a: 32; Niemann, 2018: 100). Neither does strategic justification necessarily seek to defend fixed (national) interests. One of the main functions of the public exchange of justifications is to enable agreement and collective action (Gadinger, 2016;Kornprobst, 2014), so an actor may use a justificatory argument strategically to explain why she is reversing her position to agree with others.
Third, I assume that public justifications shape political outcomes. As Holger Niemann (2018: 201) argues, political actors engage in the time-consuming work of crafting public justifications because they believe that these normative arguments matter. They do not necessarily matter in the sense of convincing others to endorse the policy that a speaker is justifying; this cannot be ruled out, but is an unlikely result of highly formalized public exchanges. Yet, even without persuading others to change course, public justifications matter: they enable an actor to claim legitimacy for her position, by reassuring audiences -and herself -that she is acting based on generally accepted standards. This, in turn, makes it easier for the actor to take the stance she is justifying, and ultimately for the respective policy to be realized, in ways consistent with the justificatory argument.
This enabling effect must be understood in terms of a soft, probabilistic causality: justificatory discourses do not determine policy choices but make certain choices more likely, all other things being equal. Accordingly, studying public justifications does not suffice to give a full explanation of individual political outcomes. For instance, a Security Council resolution referring a situation to the ICC can be influenced by numerous factors, including horse-trading across Council resolutions. Neither does a study of justificatory discourse measure the impact of discursive strategies on audiences. To do this, we could trace the internal deliberations through which individual governments arrived at specific intervention decisions or test through survey experiments which of several alternative justifications generates higher public approval for interventions. In my study, in contrast, I map and compare uses of discursive strategies that make certain intervention choices more likely, thus shedding light on patterns of intervention outcomes.
Based on these assumptions, I propose to understand norm linkage as a discursive resource which actors can use to enlarge their justificatory repertoires. For example, arguing that an ICC intervention helps protect civilians adds a justification to the repertoire that justice advocates can draw on.
As with individual norms, I argue that actors interpret norm relationships strategically, that is, that they craft linkage arguments adapted to situation-specific political aims. This does not imply that linkage is 'mere rhetoric' and inconsequential for political outcomes. Like individual norms, linkage is a resource that enables actors to publicly support or oppose certain policies and that shapes the ways in which these policies are implemented in practice.
Like individual norms, norm relations are malleable within limits. Actors may link any norm to any other norm, but to be persuasive they must build on evaluative claims and 'cause-effect stories' (Kim and Sharman, 2014: 430) that are familiar to audiences (such as the popular idea that punishment 'deters' wrongdoing). Where multiple accounts of a norm relationship exist, as in the case of protection and prosecution norms, this offers actors a range of linkage options and the possibility to justify different policy choices. Even then, however, a specific linkage argument may over time become dominant through a process of discursive 'funnelling' (Holzscheiter, 2010), making it difficult to de-link norms or connect them in alternative ways. It is thus an empirical question whether a set of norms becomes consolidated into what Phil Orchard (2020) describes as a 'regime': a mechanism that 'brings increased regularity to state practices' by bundling together distinct standards of behaviour.

Linkage-based justification strategies
How can different linkage arguments be used to justify policy choices? I distinguish four justificatory strategies: legitimation, delegitimation, cross-legitimation and cross-delegitimation. With respect to prosecution and protection norms, the four strategies give rise to four hypotheses about why actors discursively link both norms. While I do not measure the impact of discursive strategies on audiences, each strategy can be expected to favour certain intervention outcomes. Consequently, its use indicates the (relative) plausibility of a particular explanation of ICC intervention patterns.
A legitimation strategy can rely on synergy or cluster arguments. In using a synergy argument, actors seek to boost support for a policy prescribed by a norm A by mobilizing an additional evaluative standard (norm B). In the case at hand, they can seek to strengthen the legitimacy of an ICC intervention by claiming that it is prescribed not only by the prosecution norm but also by the protection norm. Alternatively, actors can seek to legitimize a policy by arguing that the value guiding it forms part of a cluster of values which ought to be realized jointly. Presenting prosecution and protection as 'twin' responses to atrocities can help speakers to legitimize judicial interventions as part of a standard package of crisis responses and deflect calls for delaying justice in the search for peace. In summary, Hypothesis 1 holds that actors supporting international judicial interventions will use synergy or cluster arguments to justify this position. Reliance on this strategy can be expected to favour the use of judicial interventions as a crisis-response tool in an early phase of conflict.
Linkage can also be used to delegitimize policies associated with a norm and enable actors to justify deviations from it. By constructing a norm conflict, actors can challenge a behaviour prescribed by a norm A (e.g. an ICC investigation prescribed by the prosecution norm) by connecting it to an additional norm B (the protection norm), which is said to prohibit it in the same situation. This justification strategy can also be used as a tactic of 'norm spoiling' to undermine support for the norm in general (Sanders, 2018: 273). Hypothesis 2 thus holds that political actors opposing judicial interventions will use conflict arguments to justify this stance. If we find actors using such a delegitimation strategy, this would suggest that linkage has contributed to obstructing, rather than facilitating, judicial interventions.
Legitimation and delegitimation strategies attempt to increase or decrease support for a policy by linking it to an additional norm. Yet, political actors may also link a policy to an additional norm to support or oppose not the linked policy itself, but other policies associated with this second norm. I term this a cross-(de)legitimation strategy.
In a cross-legitimation strategy, actors link a policy prescribed by norm A to an additional norm B to add legitimacy to other policy options associated with B. In the case at hand, a cross-legitimation strategy would be at work if judicial interventions were used to justify other forms of protective intervention.
A cross-legitimation strategy can rely on synergy arguments. Judicial measures are often presented as a tool of protection preceding or complementing other instruments in a protective 'toolkit' (Evans, 2008;Mégret, 2011). The corollary of this argument is that judicial and non-judicial tools should be used in conjunction, or that judicial measures should be followed by more forceful measures if they fail to achieve the intended protective effect. Thus, a 'toolkit' narrative can contribute to what Thomas Weiss (2011: 2) refers to as 'sandwiching' military force between other measures to make humanitarian intervention 'more palatable'. This cross-legitimation strategy is most evident if prosecution is explicitly framed as preceding other measures, but it may already suffice to mention the protective effect of justice to reassure audiences that sanctions or military measures are used as part of a balanced and graduated protection strategy.
Alternatively, a cross-legitimation strategy can rely on cluster arguments stressing shared values and/or problem descriptions guiding different behavioural rules. If international prosecutions are presented as a measure that ought to go hand in hand with the protection of civilians but cannot itself ensure protection, the explicit or implicit conclusion is that investigations of atrocities (behaviour prescribed by norm A, prosecution) should be paralleled by measures to protect civilians from the same atrocities (sanctions or military intervention prescribed by norm B, protection). Thus, justice can become a 'trigger' for other forms of intervention (similarly Roach, 2006: 171).
Summarizing both options, Hypothesis 3 holds that political actors supporting protective sanctions or military interventions will construct synergy or cluster arguments linking justice to protection to legitimize this stance. The use of this cross-legitimation strategy can be expected to favour simultaneous uses of judicial and non-judicial interventions.
Finally, the notion that judicial interventions can be used as a 'fig leaf' to avoid more robust measures exemplifies what I call a cross-delegitimation strategy. This strategy relies on a synergy argument: by linking a behaviour prescribed by norm A (an ICC intervention prescribed by the prosecution norm) to an additional norm B (protection), actors can not only boost support for justice but also enlarge the menu of policy options available for implementing norm B. This, in turn, makes it easier for them to oppose other options connected to B (e.g. military intervention). The cross-delegitimation strategy is most evident if prosecution is explicitly framed as effective 'alternative' to other protection measures. Yet, it can already be at work if a synergy argument merely presents prosecution as an effective protection tool, letting audiences draw their own inferences about the necessity of additional protective interventions. Hypothesis 4, then, holds that political actors opposing protective sanctions or military measures will use synergy arguments to justify this stance. This cross-delegitimation strategy can be expected to contribute to early ICC interventions in cases in which no other enforcement measures can be agreed. In cases in which sanctions or military measures are agreed, it will incentivize individual governments opting out of these collective measures to push for the parallel use of justice as a 'fig leaf'.
In summary, political actors can legitimize or delegitimize prosecution or protection measures by linking both underlying norms with different types of arguments (Table 1).

Methodology and case selection
In the following, I use my theoretical framework to investigate whether and how political actors have relied on norm linkage to support or oppose judicial interventions, as well as sanctions and military interventions aimed at protecting civilians. By probing the four hypotheses, I seek to understand both why certain linkage arguments have prevailed and how observable intervention patterns have been discursively enabled as a result.
My analysis focuses on official statements made by government representatives in UN Security Council (UNSC) debates, specifically thematic debates about the Protection of Civilians in Armed Conflict (PoC) in the years 2000-2016 and debates about the conflicts in Darfur/Sudan (2004), Libya and Syria (both 2011. UNSC debates constitute a particularly suitable body of text for the task at hand. IR scholars interested in legitimacy and justification emphasize 'the fundamental role played by processes of legitimation and delegitimation in all that the Council does' (Hurd, 2007a: 2). Since the Council cannot implement its decisions without the consent and cooperation of UN members, its members 'attempt to legitimise their decisions by claiming them as rightful' (Niemann, 2018: 2), thus at the same time engaging in a practice of 'normative ordering' (Niemann, 2018: 4-10). Furthermore, the UNSC is a key site of decision-making on judicial and non-judicial interventions. It can authorize sanctions or the use of force as part of its competence to restore international peace and stability (Ch. VII, UN Charter), and it can refer situations to the ICC (ICC Statute Art. 13 (b)). 10 Accordingly, governments should use the forum to justify the use (or non-use) of all these forms of intervention. In past studies, scholars have analysed UNSC debates to understand normative controversies over the use of force, in particular (e.g. Booth Walling, 2015;Hurd, 2007b;Niemann, 2018).
From a methodological viewpoint, UNSC debates have features that facilitate comparative and longitudinal analyses of justificatory discourses. They constitute a clearly defined universe of texts permitting systematic sampling. A comparison of different UNSC debates holds constant the audience addressed by government representatives, and for thematic PoC debates also the topic of debate. A potential disadvantage is the changing composition of elected Council members. This problem is softened by including thematic PoC debates, which are open to all UN members. Due to the higher number of speakers this entails, PoC debates are also particularly amenable to quantitative analysis. Furthermore, my selection of thematic PoC and country-specific debates allows me to compare processes of norm (re)construction at a more abstract level (PoC) and situations of norm application (country debates), and to test for differences between both contexts that the notion of strategic linkage would lead us to expect. For country-specific debates, I focus on the three cases in which the UNSC either referred a situation to the ICC (Darfur/Sudan in 2004 and Libya in 2011) or failed to reach agreement on a proposed referral resolution (Syria in 2014). The three cases differ on both intervention outcomes and voting patterns: the UNSC authorized an ICC intervention (unanimously), strong sanctions and a coercive military intervention in Libya; an ICC intervention (with four abstentions), more limited sanctions and non-coercive peacekeeping in Darfur; and no collective intervention measures in Syria. This allows me to test for differences across strategic contexts predicted by my hypotheses. Despite these advantages, my case selection also introduces two sources of potential bias: the high level of public attention for the three conflicts, and the focus on a specific pathway of initiating ICC investigations. While all four justification strategies can be used in other cases as well, incentives to do so are arguably strongest in high salience crises in which governments have full control (collectively) over the initiation of prosecutions. A positive finding on either of my hypotheses would therefore require confirmation in follow-up studies of other cases, whereas a negative finding would suffice to reject it.
In total, I analysed 954 statements, 377 made in PoC debates and 577 made in country-specific debates. 11 I conducted a qualitative content analysis of these statements, measuring frequencies of arguments with the software MaxQDA. Quantification is important because it allows me to measure the salience of linkage and the relative weight of different justification strategies. While I argue that linking protection and prosecution makes a difference, I do not expect that political actors use only linkage arguments to justify or oppose ICC interventions; nor do I claim that different linkage-based justification strategies are mutually exclusive.
In coding the text material, I used a codebook developed in two steps: 12 Following the deductive development of a preliminary scheme, codes were inductively refined by precoding a text sample. The codebook was designed to capture three types of arguments that allow me to operationalize my hypotheses: linkage arguments (including synergy, cluster and conflict arguments), principled arguments for or against judicial interventions, and expressions of support for/opposition to judicial measures, sanctions and military interventions.

Operationalization of hypotheses
I argue that discursive linkage between protection and prosecution norms has contributed to the observable pattern of judicial crisis interventions, through one or several of the justification strategies identified by the four hypotheses. To test this argument, I break it down into several sets of observable implications.
The first set captures the overall salience of linkage. The general proposition that linkage shapes intervention patterns would lead us to expect that linkage arguments form a prominent part of the discursive repertoire that UN members have developed in discussing prosecution and protection norms, and that governments leverage this repertoire in specific situations when prosecutions are being considered. Accordingly, in PoC debates discussing protection at an abstract level, majorities of speakers should both mention judicial measures and advance linkage arguments connecting them to the task of protecting civilians (OE1). 13 Furthermore, majorities of speakers should advance linkage arguments in discussing (potential) judicial interventions in Darfur, Libya and Syria (OE2). In both PoC and country-specific debates, we would also expect linkage arguments to be used more frequently than principled arguments. Principled arguments justify or criticize judicial interventions without linking them to protection, thus indicating an independent application of the prosecution norm. Principled justifications of prosecutions include retribution (the claim that crimes 'deserve' to be punished), redress for victims of atrocities and expressivism (the notion that criminal trials express fundamental community norms). 14 Principled objections include concerns about sovereignty, 'politicized' justice and double standards.
Zooming in on specific justification strategies, the legitimation hypothesis (H1) suggests that political actors supporting international judicial interventions advance synergy or cluster arguments to justify this stance. Accordingly, we should find both types of arguments being used in a qualitative content analysis of PoC and country-specific debates (OE1). Quantitatively, we would expect positive linkage to be more prevalent (relative to conflict and principled arguments) in strategic contexts in which actors have little interest in criticizing prosecutions. This should be the case in the PoC debates focusing on abstract norm construction and in the applicatory case of Libya, in which an ICC referral was unanimously agreed (OE2). If positive linkage is used as a 'sales pitch' for judicial interventions, we would also expect it to be used more by speakers supporting international prosecutions than by those merely professing support for 'justice' in general or for national accountability efforts (OE3). Finally, the utility of a linkage-based legitimation strategy would be corroborated if we could observe individual governments in particularly difficult political positions relying on it. Governments publicly supporting a judicial intervention despite being strongly predisposed to oppose it -by their own previous positions or external pressures -should use linkage to justify their choice (OE4). Conversely, the legitimation hypothesis would be weakened if governments in such positions did not resort to fitting linkage arguments.
The delegitimation hypothesis (H2) generates analogous expectations. In a qualitative content analysis of UNSC debates, we should be able to identify uses of conflict arguments to justify opposition to international criminal prosecutions (OE1). A delegitimation strategy should be most attractive in contexts in which actors have a strategic interest in resisting prosecutions. Accordingly, conflict arguments should be articulated most frequently in the Darfur and Syria debates (OE2), and by actors opposing international judicial measures, in particular (OE3). Finally, the delegitimation hypothesis would be supported if we could see individual countries relying on conflict arguments to publicly oppose international prosecutions despite having countervailing reasons to support them (OE4). According to the cross-legitimation hypothesis (H3), actors supporting sanctions and military interventions should be more likely than others to advance synergy or cluster arguments to legitimize these practices (OE1). Furthermore, we would expect to find in a qualitative content analysis that linkage arguments are not only correlated with support for protective interventions but are explicitly connected to them through 'trigger' or 'toolkit' narratives (OE2). The use of this cross-legitimation strategy should be particularly attractive in the case of Libya, in which the strongest intervention measures were adopted and needed to be justified. Finally, the hypothesis would be supported by cases in which a cross-legitimation strategy enabled individual countries to support sanctions or military measures against countervailing incentives or pressures (OE3).
The cross-delegitimation hypothesis (H4) would lead us to expect that actors opposing sanctions or military interventions are more inclined than others to use synergy arguments (OE1). Again, we would expect a qualitative content analysis of UNSC debates to confirm that synergy arguments not only correlate with opposition to sanctions or military measures but are explicitly used to frame justice as a viable alternative to them (OE2). This 'fig leaf' strategy should be particularly attractive in the case of Darfur, when the UNSC could agree on an ICC referral but only limited sanctions and no military measures. Finally, the hypothesis would be strengthened if individual countries under particularly intense pressure to justify their non-intervention in a conflict used a cross-delegitimation strategy to fend off criticism (OE3).

Salience of norm linkage
My analysis shows shat discursive linkage between protection and prosecution norms has been highly salient across UNSC debates. In PoC debates (Figure 4)  In UNSC debates about Darfur, Libya and Syria ( Figure 5), majorities of speakers drew on this discursive repertoire in discussing proposed international judicial interventions, with only the Syria debate remaining just under the 50 percent threshold (OE2). 16 In both types of debates, linkage arguments outweighed principled arguments for or against judicial interventions (OE3). The only exception was the Syria debate, which suggests that principled arguments can prevail in specific contexts. In the Syrian case, supporters of an ICC referral knew that Russia would block the referral resolution. This could explain why they focussed less on protective effects of a hypothetical ICC intervention and more on 'shaming' Russia with principled arguments. A similar blockade also shaped later discussions of an investigative mechanism to ensure accountability for chemical weapons use.

Legitimation: using linkage to support judicial interventions
Did actors use linkage strategically to legitimize international judicial interventions (H1)? A qualitative analysis of UNSC debates provides many examples of speakers justifying judicial measures with synergy and cluster arguments (OE1). In the 2014 PoC debate, for instance, the Estonian statement stressed the ICC's 'power to deter further atrocities or prevent their escalation', while Chad's representative argued that 'international tribunals' efforts [. . .] could very well dissuade those who, absent such consequences, would commit the most serious crimes against civilians'. 17 Deterrence was also discussed in country-specific debates. In the 2014 UNSC debate about the proposed ICC referral of the Syrian conflict, for instance, the representative of Luxembourg argued: To fight against impunity is to deter perpetrators from committing more crimes. We deeply believe that justice is an essential ingredient -an indispensable ingredient -to restoring peace in Syria. 18 Speakers also stressed the contribution of prosecutions to post-conflict peacebuilding, as illustrated by the Algerian statement in the 2005 debate on the Darfur ICC referral: Algeria firmly believes that fighting impunity is a crucial element for the entrenchment of peace and stability. That need is all the more essential in the case of Darfur because the conflict, which has raged for several years, has damaged relations between the communities. 19 In other statements, protection and prosecution were presented as clustered obligations without attributing a direct protective effect to justice. In the 2008 PoC debate, for instance, Australia's representative argued: Just as the international community has a responsibility to protect communities from such mass atrocity crimes, so too do we have a responsibility to bring those who abuse fundamental principles of international law to justice. 20 As predicted by the legitimation hypothesis, a quantitative comparison of debates shows that the use of positive linkage varied across strategic contexts (OE2). Positive linkage arguments (particularly synergy arguments) were most prevalent in the PoC and Libya debates, that is, in contexts in which governments had no strategic interest in criticizing prosecutions ( Figure 6). The notion that positive linkage is used as a 'sales pitch' for international judicial interventions is also supported by a comparison of speakers with different positions on justice. In PoC debates, speakers supporting international justice relied significantly more on positive linkage than speakers merely professing support for justice in general or the use of national courts (OE3). This is reflected both in the percentage of statements using linkage (Figure 7) and in the frequency of linkage arguments compared with principled justifications (Figure 8). 21  Finally, can we identify instances in which positive linkage helped individual countries to support judicial interventions despite being strongly predisposed not to do so, thus enabling collective agreement (OE4)? India's discursive strategy in the Libya debate is a case in point. India, an ICC non-member, was among the countries most sceptical about referring the situation in Libya to the ICC. In informal discussions, it warned against escalation risks and negative repercussions for negotiations with Libyan leader Muammar Al-Ghaddafi (Fung, 2022: 9). When India nevertheless joined the consensus on a referral resolution, it explained this reversal by pointing to positive linkage arguments: [W]e would have preferred a calibrated and gradual approach. However, we note that several members of the Council, including our colleagues from Africa and the Middle East, believe that referral to the Court would have the effect of an immediate cessation of violence and the restoration of calm and stability. The letter from the Permanent Representative of Libya of 26 February addressed to you, Madame President, has called for such a referral and strengthened this view. We have therefore gone along with the consensus in the Council. 22 Drawing together the findings on all observable implications, I conclude that norm linkage has been prominently used to support judicial interventions.

Delegitimation: using linkage to oppose judicial interventions
Was linkage also used to oppose judicial interventions in UNSC debates (H2)? A qualitative content analysis highlights some instances in which speakers relied on this strategy (OE1). In most of these instances, they combined 'conflict' arguments against international prosecutions with reassurances of being generally supportive of justice, as in China's statement explaining the country's abstention on the Darfur referral: Undoubtedly, the perpetrators must be brought to justice. The question is: What is the most effective and feasible approach in this connection? In addressing the issue of impunity, we believe that, when trying to ensure justice, it is also necessary to make every effort to avoid any negative impact on the political negotiations on Darfur. 23 A quantitative comparison of debates (OE2) shows that conflict arguments were articulated exclusively in the Darfur and Syria debates, and more so in the earlier debate on Darfur (22 out of 119 coded segments) than in the debate on Syria (only 7 out of 442 coded segments). This finding underlines, on one hand, the strategic nature of linkage: although potential tensions between prosecution and protection were discussed informally in all three conflicts including Libya (Fung, 2022), they were publicly articulated only by governments openly opposing ICC interventions. On the other hand, the rare use of conflict arguments in the Syria debate despite strategic incentives to use them (particularly compared to much more frequent principled objections to justice, see Figure 6) appears to indicate a discursive 'funnelling' towards positive linkage that has made conflict arguments less usable over time.
The scarcity of conflict arguments does not permit a quantitative analysis of their correlation with actor positions on national and international justice (OE3). Yet, the strategic use of such arguments is underlined by the fact that most articulations (19) came from governments directly targeted by international judicial measures (Sudan and Syria) and their principal UNSC allies (Russia and China), with other governments accounting for only 10 statements.
Still, conflict arguments were not always used where they could have been helpful in justifying politically difficult positions (OE4). This is illustrated by the US position in the debate on Darfur. The US government decided to abstain on the Darfur referral resolution, although it faced a powerful domestic campaign to intervene in the conflict and respond to allegations of 'genocide' (Seymour, 2014). In justifying its stance, US representatives did not mention potential risks to peace negotiations (although these were discussed informally, see Prunier, 2005: 143) but emphasized principled objections to the ICC, specifically the court's perceived encroachment on US sovereignty. 24 In summary, linkage was sometimes used strategically to oppose judicial interventions, but within limits that appear to reflect an increasing discursive dominance of positive linkage. Indeed, speakers increasingly criticized the very articulation of conflict arguments, as exemplified by the Argentinian statement in the debate about the proposed Syria referral: It is untenable and even offensive today to seek to present a false dichotomy between peace and justice. We have all learned that there is no justice without peace and that peace is endangered without justice. 25

Cross-legitimation: using linkage to support sanctions and military measures
What are the indications that by linking prosecution and protection norms, speakers in UNSC debates sought to justify not only judicial but also non-judicial interventions (H3)?
As predicted by the cross-legitimation hypothesis, speakers supporting sanctions or protective military interventions (PMI) used positive linkage more frequently than speakers not supporting these measures in PoC debates (OE1; see Figure 9). 26 The joint articulation of linkage arguments and expressions of support for non-judicial interventions is one indication, but does not yet suffice to demonstrate that speakers used the former to legitimize the latter. An additional indicator is the use of explicit 'trigger' or 'toolkit' arguments in UNSC debates (OE2). 27 Regarding sanctions, most relevant statements were made in debates about the Syrian civil war, with a focus on the use of chemical weapons. In this context, many speakers pointed to the findings of a specific accountability mechanisms, the Joint Investigative Mechanism (JIM) of the UN and the Organisation for the Prohibition of Chemical Weapons, to underline the necessity of sanctions. In 2017, for instance, the representative of France demanded taking vital action on our part to sanction those who have flouted the prohibition: an action drawing the necessary conclusions from established determinations by a Mechanism that we unanimously renewed, and an action, finally, to send a warning and a message of firmness against those who would consider breaking the taboo in the future.
In the same debate, the Italian representative stated that Italy supported sanctions 'to show our credibility and the deterrence role of the JIM', stressing the need 'to ensure meaningful follow-up to the work of the Joint Investigative Mechanism'. 28 Both statements exemplify a cross-legitimation narrative in which the findings of a judicial mechanism trigger non-judicial measures to protect civilians.
Regarding military interventions, trigger arguments were used more rarely. It was particularly US representatives who pointed to the JIM's findings to justify the US-led intervention against the Islamic State in Iraq and the Levant, as in this 2016 statement: The JIM has also found the Islamic State in Iraq and the Levant (ISIL) at fault for the chemical weapons attack on Marea. We have seen the allegations of additional use and agree that use by terrorists poses a profound threat. Use by ISIL is yet one more affirmation of the importance of the United States-led coalition's fight against that monstrous group. 29 More frequently, speakers legitimized military measures -specifically the 2011 military intervention in Libya to enforce a no-fly zone over the country -by arguing that non-military means including the ICC referral of the situation had been tried but failed to stop the violence. An example of this 'toolkit' narrative can be seen in the 2011 Lebanese statement, which explains Lebanon's support for the no-fly zone: Faced with that great danger [of innocent civilians being killed, the author], the Security Council has not stood by idly. We adopted resolution 1970 (2011), which demanded the immediate cessation of all acts of violence. Given the grave danger posed by these crimes, the situation in Libya was referred to the International Criminal Court. [. . .] However, Colonel Al-Qadhafi's regime disregarded the demands and yearnings of his people as well as international resolutions.
Speaking on the same occasion, Colombia's representative argued: [T]he new resolution that we have just adopted [on the no-fly zone, the author] represents the continuation of a process involving gradual measures that is in keeping with the Charter and that we began with resolution 1970 (2011) [. . .]. 30 In the case of South Africa, a similar justification strategy helped the country to adopt a position in support of military intervention in Libya which it found particularly challenging to justify vis-à-vis countervailing internal and external pressures (OE3). The South African government was in a 'very uncomfortable position' in the run-up to the adoption of UNSC resolution 1973, 'being pressurized to help stop the killing of civilians' (Adler-Nissen and Pouliot, 2014: 903-904) while also realizing that a military intervention would be domestically controversial and was 'in direct contradiction to the position of the AU [African Union] Ad hoc Committee [. . .] that had been tasked with finding a negotiated settlement to the crisis' (Smith, 2016: 396). To explain why his government nevertheless voted for the military enforcement of a no-fly zone, the South African representative argued: In adopting resolution 1970In adopting resolution (2011 [on the ICC referral and sanctions, the author], the Security Council had hoped that the Libyan authorities would act responsibly and stop committing more acts of violence against their own people. The authorities have defied that resolution and have continued to kill and displace numerous civilians while continuing to violate their human rights. We believe that the United Nations and the Security Council could not be silent, nor be seen to be doing nothing in the face of such grave acts of violence committed against innocent civilians. [. . .] 31 This statement uses a 'toolkit' argument to justify the use of force: after earlier protective measures including the ICC referral have been tried but failed, a no-fly zonedescribed as 'additional measure' in another passage of the South African statement -has become necessary.
While no arguments of this type were advanced to justify the use of force in the Syrian conflict, Syria itself and its Russian ally repeatedly claimed that Western powers supported judicial measures to legitimize a military intervention in the conflict. Thus, crosslegitimation strategies in the case of Libya were echoed critically and turned into a discursive strategy for delegitimizing judicial measures in the case of Syria. In 2014, for instance, the Russian representative denounced the proposed ICC referral of the situation in Syria as 'an attempt to use the ICC to further inflame political passions and lay the ultimate groundwork for eventual outside military intervention'. 32 In 2017, he alleged that Western advocates of the JIM 'needed the Joint Investigative Mechanism for the sole purpose of laying responsibility for the use of chemical weapons at the door of Al-Assad's Government, thereby creating additional reasons for regime change in Damascus'. 33 Pointing to Libya as an alleged precedent, the Syrian representative contended in 2017: The three Western States on the Council have become used to incorporating similar language in previous draft resolutions, leading to their misuses later on by those States to justify intervention in the internal affairs of other States and military aggression against them, as was the case in Libya and other countries. 34 Equally in UNSC debates about Syria, we can observe yet another variant of linking judicial and non-judicial measures. In addition to arguing that judicial investigations legitimize sanctions or the use of force, speakers framed both these forms of intervention as directly serving the values of justice and accountability. For instance, the representative of Uruguay argued in 2017 that a proposed sanctions resolution on Syria 'did constitute a first step towards ensuring accountability for the use of chemical weapons', while Italy stated that it supported sanctions because 'those who planned, ordered and executed the attacks must face justice'. 35 Similarly, some speakers sought to legitimize the use of force, specifically air strikes against Syrian chemical weapons facilities, by tying them directly to the accountability norm. The representative of France, for instance, justified the 2017 air strikes as 'a legitimate response to a mass crime that could not go unpunished'. 36 In summary, my analysis suggests that linkage between prosecution and protection norms was used strategically to legitimize sanctions and military interventions, but also that the use of this justification strategy was critically seized upon by others to delegitimize prosecutions.

Cross-delegitimation: using linkage to oppose sanctions and military measures
The cross-delegitimation hypothesis (H4) holds that actors opposing sanctions or protective military interventions are more likely to frame justice as an effective (alternative) tool of protection to justify this stance. Yet, the relevance of this 'fig leaf' strategy is supported neither by a quantitative analysis of the arguments preferred by intervention opponents (OE1, Figure 10) 37 nor by a qualitative analysis of the narratives used to justify non-intervention (OE2). In contrast to 'trigger' and 'toolkit' arguments, explicit arguments advocating against the use of sanctions or force by presenting judicial measures as effective alternative could not be identified in the UNSC debates analysed. Even in the Darfur debate, in which speakers should have had strong incentives to point to the ICC referral to justify their limited use of other intervention measures, the cross-delegitimation strategy was not used. Furthermore, Germany in the Libya debate is a case in which justice was not presented as an alternative tool of protection although this argument could have helped the country to defend its controversial decision not to support a military intervention (OE3). In what Rebecca Adler-Nissen and Vincent Pouliot (2014: 903) describe as 'reversed logic' from the South African position, Germany abstained on resolution 1973 authorizing the use of force in Libya, although this decision was criticized both domestically and by NATO partners. In justifying the abstention, the German representative emphasized that his government was supporting alternative measures to protect civilians. However, the measure he emphasized was not deterrence through the ICC, but the adoption of economic sanctions, characterized as a response to the 'plight of the Libyan people' and 'an effective way to end the rule of Muammar Al-Qadhafi'. 38 Taken together, my findings lead me to reject the 'moral hazard' hypothesis that is still regularly advanced in discussions about justice in conflict: there are no indications that governments used criminal prosecutions as a ' fig leaf' to avoid other forms of intervention.

Conclusion
My analysis of UN Security Council debates demonstrates that prosecution and protection norms have become closely linked in political discourse. In both thematic and situation-specific debates, governments have used norm linkage strategically to broaden their justificatory repertoires. Specifically, they have relied on synergy and cluster arguments to legitimize international judicial interventions, but also -somewhat less prominently -sanctions and military interventions aimed at protecting civilians. In almost all debates analysed (except for the Syria debate), governments placed greater emphasis on links between justice and protection than on principled arguments evaluating justice independently from protection. Strategic norm linkage, I conclude, has contributedmore than the parallel application of two independent norms -to the ICC's frequent interventions in ongoing conflicts and alongside other forms of intervention.
My analysis also shows that governments have at times constructed conflicts between justice and protection to delegitimize judicial interventions. Yet, this strategy was used within limits that indicate a (growing) discursive dominance of positive linkage.
In contrast, the proposition that governments link justice and protection to avoid other forms of protective intervention can be rejected based on my analysis. A 'moral hazard' dynamic was not found to be at work in any of the debates analysed.
Taken together, my findings suggest that judicial interventions linked to civilian protection have come with some negative political repercussions, but also that not all warnings have been equally warranted. Under-provision of protection (expected by the 'moral hazard' argument) does not appear to be a problematic consequence of ICC interventions. However, the prevalence of instrumental, protection-related justifications of prosecution risks raising exaggerated expectations and thus eventually undermining support for justice, while also favouring judicial interventions at an early stage of conflicts where they can be less effective or even counterproductive. Finally, the cross-legitimation of military measures through linkage enabled military protection measures in Libya but also backfired in the Syrian conflict, allowing opponents of judicial measures to portray the latter as a slippery slope towards the use of force.
Considering these problems, would it be beneficial to 'segregate as much as possible the juridical from the political' in the pursuit of international criminal justice and thus the 'agendas of peace and justice', as proposed by Louise Arbour (2014), the former Prosecutor of the International Criminal Tribunal for the Former Yugoslavia? My study suggests that a general de-linking of protection and prosecution norms would be neither feasible nor inherently more desirable.
First, the high salience of positive linkage indicates that de-linking both norms is practically impossible, as Arbour (2014) concedes. Even actors who warn against factoring prospects for peace into decisions about prosecutions -like former ICC Prosecutor Bensouda (2012: 439) also portray the ICC as 'a form of protection' and a framework 'to manage conflicts'. Rather than segregating justice and protection, such assertions of judicial independence amount to one-sided linkage.
Second, would de-linking protection and prosecution norms be desirable? Understanding norm linkage as the strategic use of normative repertoires for political purposes implies that de-linking norms is also a strategic political choice. If justice advocates limited themselves to principled arguments (as suggested by Bosco, 2012), this would cut off deliberations about judicial interventions from one evaluative repertoire. From an empirical perspective, however, this repertoire should remain relevant, since judicial interventions have been found to have both positive and negative consequences for protection.
The better alternative is to openly acknowledge and discuss both positive and negative interactions between justice and protection while managing expectations about what justice can and cannot achieve. Addressing tensions may result in 'yielding more to the political imperatives of peace' (Arbour, 2014), but above all would require actors to spell out what potential risks to peace and protection are (not) worth taking to fulfil the principled goals associated with international criminal justice. Academics who have devoted much energy to exploring the nuances of the relationship between justice and protection can contribute to ensuring that political discussions about criminal justice are more reflective of those nuances and less of a caricature of scholarly debates. Such discursive change is as important as institutional reform (such as the development of guidelines for investigators, mediators, or peacekeepers) to manage tensions between justice and protection.
And yet, one form of de-linking can and should be attempted. A cross-legitimation strategy that uses linkage to justify non-judicial coercive interventions ties the pursuit of justice not only to an abstract protection norm but also to other, highly controversial intervention practices -in ways that risk undermining support for justice. The short-term benefits of this strategy (some sceptics are pushed or enabled to support coercive protection measures) do not outweigh the long-term risks to the prosecution norm. Russian arguments denouncing the ICC as a pretext for military intervention may be cynical, yet they are easy to connect to existing narratives of a 'neo-colonial' court instrumentalized by Western powers.
Contributing to broader discussions on norm dynamics in international politics, my study demonstrates that relationships between norms can be as malleable and contested as individual norms, far beyond the stage of norm emergence. It illustrates how political actors can use norm linkage strategically to broaden justificatory repertoires. In the case of protection and prosecution norms, linkage discourses have been versatile enough to enable actors to take different, even contradictory positions. This highlights the empowering and permissive potential of linkage: in navigating an ever-denser web of norms, political actors can use linkage to loosen normative constraints on their actions. And yet, the study also indicates that over time, some linkage arguments may achieve a discursive dominance that makes it hard to de-link norms or construct alternative norm relationships. Whether and under what conditions norm relations thus become stabilized into authoritative 'regimes' is a question worth exploring in many areas of global politics.
trigger or toolkit arguments are not always needed to legitimize sanctions or military interventions. The mere mention of justice as a 'tool of protection' can already suffice to frame other coercive instruments as being part of a graduated approach.