Multiperspectival Normative Assessment: The Case of Mediated Reactions to Terrorism

This article provides a model for how communication phenomena can be normatively assessed using multiple normative perspectives simultaneously. We exemplify the procedure of multiperspectival normative assessment (MNA) using mediated reactions to terrorism as our case in point. We ﬁrst identify the normative challenges related to the speed and substance of terrorism communication and the ways in which relations of solidarity are communicatively constructed in reacting to terrorism. We link these challenges to four distinct normative theories that prioritize competing values for public discourse (freedom, community values, empowerment of the marginalized or constructive debate). The resulting set of competing normative expectations, which help assess the performance of terrorism communication, are eventually translated into recommendations for professional and non-professional communicators. In conclusion, we show how MNA can help ground empirical scholarship in ﬁrmer theoretical foundations while simultaneously demonstrating the usefulness of normative theory in analyzing a wide range of issues.

Communication researchers have devoted considerable effort to studying the mechanisms driving such debates, i.e., the media logics in which they are embedded, the diversity of voices, the media framing, and possible shifts in opinion, among other things (see, for example, Baden & Springer, 2017).
However, mediated contestation is not only characterized by internal value conflicts. Normative disagreement also exists on a meta-level that addresses the forms and functions of mediated contestation itself. Different normative models of democracy inform different normative conceptions of the public sphere, which in turn play out in diverging quality standards for mediated public debate. Such quality standards can then be used to conduct systematic normative assessment of empirical realities in mediated contestation (Althaus, 2012). Given the diversity of normative models, such normative assessment will have to be multiperspectival. Thus, in this article we aim at explicating multiperspectival normative assessment (MNA) as a standard research procedure that can help illuminate what a particular empirical finding means for the democratic public sphere as understood within these diverging models.
This pluralism of normative horizons poses two fundamental problems for communication researchers that we seek to tackle in this article. First, how can communication researchers, as theorists, make sense of the cacophony of values and normative horizons at hand? And second, as public-facing scholars, how can communication researchers develop clear and targeted recommendations despite the plurality of normative horizons?
On the first question, we argue that the role of communication theory lies in systematizing the diverging normative horizons that implicitly underlie public debates. For example, Ferree, Gamson, Gerhards and Rucht (2002) distinguish a representative-liberal, a participatory-liberal, a discursive and a constructionist model of the public sphere, which prioritize different core values. Other authors such as Baker (2002), Benson (2009), Dahlberg (2011), Freelon (2015, Holbert (2013) and Wessler (2018) propose slightly different sets of normative models. Among communication theorists there is no agreement on the most appropriate set of models, nor is there sustained debate about these meta-theoretical questions, partly due to the fact that different distinctions follow different knowledge interests. We do not aim at resolving such abstract disagreements across issues. Instead we choose to anchor our MNA procedure in journalists' own implied beliefs about their normative obligations in dealing with one particularly problematic media issue, namely terrorism. Such beliefs, which are sometimes expressed in editorials and published self-reflection, have been usefully reconstructed by Horsbøl (2016). They help us identify four contrasting normative horizons, which suggest distinctive responses to the normative challenges of communicatively reacting to terrorism, roughly corresponding to more general normative models of democratic governance and public contestation. When normative horizons are clearly delineated in this way, discussions about which normative standards should be applied become much more clear-cut, well-informed, and accessible to non-specialists. This theoretical contribution becomes socially consequential when we consider the second question faced by communication researchers: If normative horizons diverge, how can communication researchers even hope to help guide communicative practice? Some would answer that researchers should articulate a single "correct" understanding that can be translated into a clear set of practical rules. Our answer is different. We argue that the inherent plurality of normative orientations calls instead for specific decision rules about how to arrive at a clear set of practical prescriptions on the basis of plural values. The solution we offer in this article is two-pronged: In cases where different normative models advocate for more or less of a particular quality in public communication that addresses a particular issue, we develop our prescriptions from the most demanding model in order to maximize potential quality gains and societal relevance. In cases where the models favor contradictory qualities, we choose an inclusive approach to cover as much of the different priorities as possible in the recommendation. To our knowledge, our effort is the first to explicate such decision rules on the basis of diverging normative horizons.
Our argument will proceed in three steps. We start with an elaboration of MNA as a step-by-step research procedure. We then detail the three main theoretical MNA steps (Identifying normative challenges; Mapping competing normative horizons; Specifying decision rules to derive recommendations) using mediated reactions to terrorism as our case in point. In conclusion, we reflect on the transferability of the MNA procedure and the normative horizons mapped here to other controversial issues.
We are not aware of any previous attempts to address the full range of normative challenges in terrorism communication through systematic theoretical elaboration of normative standards that can inform recommendations for action. Many articles and books on terrorism communication (e.g., Matusitz, 2012;Norris, Kern, & Just, 2003, Ruggiero & Vos, 2013 do contain implicit or explicit normative judgments and recommendations, but: (a) the normative standards underlying these judgments and recommendations are rarely explicated, (b) recommendations often cover only isolated aspects of the problem (e.g., only graphic video portrayals of violence, or only unverified speculation, etc.), and (c) the bulk of the extant literature targets journalists and news media; recommendations are rarely addressed to active social media users. For example, UNESCO's (2017) Terrorism and the Media. A Handbook for Journalists is appropriately comprehensive and detailed in its coverage, but its purpose is not to provide a theoretical elaboration of normative standards, nor to address social media users. Thus, ours is the first attempt at applying a comprehensive MNA to mediated terrorism communication.

Multiperspectival normative assessment
MNA of communicative performance was first developed in the field of political communication (Althaus, 2012;Rinke, Wessler, Löb, & Weinmann, 2013), but its remit is quite general. The purpose of MNA is to move normative analysis in communication research away from ad-hoc judgments in the concluding sections of otherwise purely empirical research publications. The procedure results in a systematic assessment of empirical communication phenomena using multiple, explicated normative standards. On the highest level of normative assessment as defined by Althaus (2012, p.100), assessment judgments take the form of the following statement: "Finding A is troubling from the standpoint of Theory Z, because Theory Z requires Y amount of Quality X." An assessment statement like this "[p]ositions the author's evaluative stance within larger theoretical debates" (Althaus, 2012, p.100). Such a positioning promises to provide a clearer and deeper justification for the assessment. It also makes the normative standards more transparent and more easily criticizable in academic debates on the matter. Both of these gains, in turn, promise to make the contentious discussions on proper communicative conduct more rational.
The procedure of MNA comprises four essential steps: Step 1: Identifying the normative challenges. MNA starts with an investigation of what seems normatively problematic in the domain under study. This can be based either on accounts circulating in public discourse or explicit problematizations in the scholarly literature. It can also be based on extant empirical, especially comparative, results that hint at differences in performance and thus suboptimal outcomes in some instances, which can then be framed as normative challenges in MNA.
Step 2: Mapping competing normative horizons. The central element in any MNA is the specification of competing normative horizons against which actual communicative performance will later be compared. "Normative horizon" serves here as an umbrella term for explicit normative theories, for more concrete normative models or ideal-types, and for professional normative meta-reflection, all of which can, separately or conjointly, inform the normative expectations that lie at the heart of the MNA procedure.
Step3: Measuring performance levels empirically. In order to serve as empirical yardsticks, normative horizons must be broken down into sets of operational criteria applicable to the domain and material under study. As normative horizons vary, each horizon will require its own unique set of operational criteria, resulting in a multidimensional measurement tool (for an example, see Freelon, 2015). The result of Step 3 is an assessment matrix or narrative that specifies how well each normative horizon is matched by empirical reality based on the measurement criteria used.
Step 4: Devising decision rules to derive recommendations. While the assessment matrix technically concludes the academic part of the MNA exercise, in many cases MNA will be performed with the goal of improving communicative practice. In this case, researchers must decide how practical recommendations can be derived from a multidimensional matrix or narrative. There are in principle two ways to do this. Researchers can either retain multiple normative horizons and formulate recommendations from each perspective separately, or they can first devise decision rules that help resolve diverging recommendations into theoretically harmonized guidelines.
The MNA procedure we are advocating here might raise two questions or objections. First, some might argue that it is intellectually unsatisfying to simply acknowledge competing normative horizons as different but equally worthy. As scholars, should we not be advocates for our preferred normative perspective and assess reality from that unique vantage point? It is of course still possible even within the MNA approach to advocate one particular normative horizon and base normative assessment on that particular one only. In this case the presence of other perspectives will hopefully lead at least to a stronger transparency of the options and awareness for one's own choices and justifications. But the full enlightening potential of normative assessment is realized when competing standards are taken seriously as legitimate perspectives.
Second, critics might ask whether it should not be possible to combine the normative aspirations expressed in two or three of the normative horizons to construct a new, composite horizon? In specifying normative expectations vis-à-vis the mediated public sphere, why can we not say: Freedom of expression is important (liberal), but only in conjunction with constructive, self-reflexive debate (deliberative) and with an open eye for the dangers of vilifying societal minorities (agonistic model)? We agree that such a synthetic approach is of course possible. In fact, the prior identification of competing ideal-types prepares one well for such a merging of normative standards. But such complex composite norms must be theoretically justified much like single ideal-types. And using composite norms also comes with an increase in complexity and a decrease in specificity. In purely academic contexts this might be perceived as an advantage. But as soon as the discussion of problematic tendencies leaves the corridors of academia and enters the marketplace of public debate, simplicity and clarity become paramount. This is why we will work with separate rather than merged ideal-types in the following description of MNA.
Step 1: identifying the normative challenges As mentioned earlier, the identification of normative concerns to be addressed by MNA can be based either on societal problem discourses or on explicit academic problematizations.
Step 1 thus amounts to picking up the challenges most relevant for the domain under study and to systematizing those challenges.
In the domain of mediated reactions to terrorism, which we take as our example here, there is no shortage of critical statements concerning the media's alleged symbiosis with terrorism (for an overview, see Spencer, 2017) both in public and scholarly debate. Even though we advocate for a more contingent understanding of the media's role, which avoids positing sweeping necessities, the challenge lies in ordering and prioritizing the plethora of normative claims that might be made for terrorism-related communication. We do this here with the help of two heuristic devices. One is to use a generalized account of what terrorists want to achieve communicatively and match these goals with aspects of mediated terrorism communication. The other heuristic device is to arrange the normative challenges on a temporal, a substantive and a social dimension. Using this approach, we can cluster normative problems that arise from the increased speed of terrorism communication (Backes et al., 2016;Liebes, 1998); from concerns pertaining to who gets to speak and what is being said about the motivations and consequences of terrorism (substance) (Kampf & Liebes, 2013); and from the selective relations of solidarity that terrorism communication establishes with different categories of victims and with so-called "suspect communities" (Hillyard, 1993) unduly held responsible for supporting terrorism (see Table 1).

Terrorists' communicative goals
Terrorist activity can be defined as "the deliberate creation and exploitation of fear through violence or the threat of violence in the pursuit of political change" (Hoffman, 2006, p. 40). Even though there is no consensus in the literature on how to define terrorism, the widely-respected Hoffman definition provides a useful starting point by highlighting the communicative nature of terrorist behavior: Terrorism exploits fear not just in relation to those directly attacked, but in the larger community or society that terrorists target to achieve political change. In order to achieve this wider emotional effect, terrorist groups create what has been called the "theater of terror" (Weimann & Winn, 1993) to make sure local attacks and the associated terrorist messages are widely publicized by professional news media and circulated by social media users. Incidentally, the definition also shows that terrorism is an activity that actors engage in or refrain from; it is a tactic rather than an essential trait of such actors. We acknowledge that calling somebody a terrorist or labeling an act of violence as a terror attack constitutes an act of communicative construction and is thus dependent on the perspective applied. As the old saying goes, one person's terrorist is another person's freedom fighter.
Beyond terrorists' attempts to maximize visibility, their interests are also furthered by particular questionable qualities of public communication that might confer legitimacy and prestige in addition to visibility (Althaus et al., 2018) (see the third column in Table 1). For one, terrorists' interests are tremendously helped by a climate of immediate and widespread confusion, panic or outrage following an attack, rather than by measured, well-considered and delayed responses (Althaus, Bajjalieh, Jungblut, Shalmon, van Atteveldt, & Wessler, 2020). Beyond the upheaval, however, terrorists also want to get their messages across. They would like their professed grievances, their ideology and political aims represented in public debate. And they would appreciate accounts that both signal understanding for their grievances and portray themselves as effective actors that pose credible threats and are taken seriously by the authorities. While terrorism is generally considered a "onesided" issue (Norris et al., 2003, p. 12) pitting terrorists against humanity, and widespread condemnation of attacks is typical, public representations of terrorists' motivations and effectiveness can vary in ways that either amplify or undermine their strategic communication goals. When victims come into focus, terrorists' interests can be furthered by the depiction of extreme helplessness and horror because these bolster their claims to effectiveness. Finally, many terrorist groups want to be seen as the legitimate spokespeople, or avant-garde, of those aggrieved communities for which they purport to speak. Media representations that equate perpetrators with their proclaimed larger constituencies support this communicative goal.

Speed
The problems associated with the instantness of terrorism communication can be highlighted when we consider the following scenario: A terrorist attack happens. The identity of the perpetrator(s) and their motivation is initially unclear. During the ongoing attack the first livestreams, eyewitness videos, and photos appear on social media. Journalists repost that material, then start probing its authenticity and asking over social media for the permission to use them in their outlets' reporting (Rauchfleisch, Artho, Metag, Post, & Schäfer, 2017). Media outlets start live blogs, broadcasters abandon their usual schedule to run live special feature programs for hours to come. People in the vicinity of the attack post more impressions, rumors, and speculations. More and more people repost, some journalists amplify. Eventually, a climate of panic takes hold leading people to misinterpret things they hear and see as indications of multiple parallel attacks and continuing danger. Law enforcement officials have a hard time confirming what is known and disclaiming what isn't, and also being listened to on social media while attempting to make that distinction. This is roughly the scenario that materialized in Munich, Germany, on 22 July 2016, as documented by the quality daily Süddeutsche Zeitung in an article entitled "Timeline of panic" (Backes et al., 2016). On that day, an 18-year-old attacker killed nine people and eventually himself near a shopping center. The shooting was motivated by the attacker's right-wing extremist, racist ideology and happened on the fifth anniversary of the mass shooting in Norway. All of the victims had a migrant background or were Sinti. In an unusual flurry of speculation and panic, social media users not present at the shopping center falsely identified no less than 66 additional attack sites (Backes et al., 2016). Some hurt themselves and damaged property in a well-known traditional Munich brewhouse several kilometers away from the shopping center while fleeing in panic. For several hours large numbers of rescue workers, doctors, and law enforcement officers (some in plain clothes) were mobilized to react to the expected massive onset of wounded and dead that, fortunately, never happened. 1 In sum, the accelerated and unregulated flow of social media (mis)information, differentially selected and amplified by journalists, created problems for both the veracity and the emotional tone of the messages (see Table 1). These problems have always existed to some degree in relation to terror attacks, but are exacerbated by the real-time nature and distributed character of public communication today. A normative problematization of terrorism communication must therefore address the question: (a) how information can be verified under conditions of extreme uncertainty and strong competition for attention. It must also find an answer to: (b) how the tone of terrorism communication can do justice to the seriousness of the situation without contributing to potentially harmful panic or uninhibited outrage.

Substance
The second set of normative challenges relates to who is accorded legitimate voice in terrorism communication and how terrorist acts are substantively contextualized by talking about motives and consequences (see Table 1). Concerning voice, it seems less likely today than it used to be that journalists would treat terrorists as normal interviewees, but it does happen and can confer celebrity status to some perpetrators (Kampf & Liebes, 2013, pp. 42-64;Rosenberg & Maoz, 2012). Perpetrators can also bypass journalistic gatekeepers and exercise voice directly on the Internet. But this does not automatically confer legitimacy to their causes. To achieve legitimacy terrorists still need coverage that validates their professed grievances, ideologies, and aims. A normative conception of terrorism communication today will therefore have to develop models for how to publicly talk about what terrorists think and want without giving them direct, legitimate voice (see Table 1). This could be done through interpretations by experts or testimony by self-critical ex-terrorists.
In addition, guidelines are required for how to analyze the phenomenon of terrorist violence beyond immediate event coverage. How can the causes of terrorism be addressed in public communication without simply taking terrorists' professed grievances as those causes? One possibility would be to talk with experts about why nonviolent options do not seem to exist for certain groups in particular situations. Sophisticated analysis is even more important in relation to the consequences of terrorism: If perpetrators are stylized as highly effective actors exerting credible threats, such portrayals could increase the potential for aggressive overreactions accompanied by sustained curtailment of fundamental civil rights within targeted populations. Research shows that strong fear appeals tend to boost support for security measures (Saleem, Prot, Anderson, & Lemieux, 2015;von Sikorski, Schmuck, Matthes, & Binder, 2017), which are hard to take back later on.

Solidarity
Apart from speed and substance, terrorism communication is also about the communicative construction of solidarity between groups. When a community or state is violently attacked by someone construing that entity as their enemy, strong ingroup/outgroup dynamics invariably set in. Mediated reactions strongly depend on whether the victims are ingroup members, that is, whether they belong to the community that collectively self-identifies as the entity under attack, or whether the victims are members of an outgroup with which the majority ingroup does not identify (see Wolfsfeld, Frosh, & Awabdy, 2008).
In the case of ingroup victims, the typical reaction would be a communicative closing of ranks in the ingroup expressed, for example, by highlighting support for victims and survivors, postulating shared ingroup values, and reinforcing the demarcation toward outgroups. However, dramatic expressions of compassion can sometimes expose ingroup victims in ways that endanger decency and respect for those deceased and mourning relatives, especially in the tabloid press (see Table 1). In the case of outgroup victims reacting to terrorist attacks poses a peculiar normative challenge for both professional and nonprofessional communicators, namely, how to deemphasize the marginal character of the outgroup by symbolically including outgroup members in the ingroup category. This can be achieved, for example, by highlighting the shared humanity and possibly also giving voice to outgroup members.
Finally, terrorism communication also unavoidably frames, and constructs relations to, perpetrators and the larger communities for which they claim to act. As with victims, relationships to perpetrators follow a clear ingroup/outgroup dynamic. In the case of outgroup attackers, who construe the dominant societal ingroup as their enemy, it is not uncommon that the community which these terrorists claim to represent is constructed as a "suspect community." The term was coined by Hillyard (1993) in relation to the Irish living in Britain under the constant suspicion of supporting the terror committed by the Irish Republican Army (IRA). "Suspect community" denotes a group who is suspected of supporting terrorism because its members share one or several features with the perpetrator of a terror attack. Importantly, the term "suspect community" refers to a process of communicative construction: Groups are not naturally more or less suspect, they are constructed as potential supporters of terrorism through an overgeneralization of some externally perceptible similarity between them and particular perpetrators. The undifferentiated construction of communities suspected of supporting or perpetrating terrorism will generally aggravate the situation of the communities' innocent members by subjecting them to intensified prejudice as well as selective repressive measures. In addition, categorizing outgroups as suspect "may be serving to undermine national security rather than enhance it" (Pantazis & Pemberton, 2009, p. 646) because it might contribute to the radicalization of outgroup members and generally boost hostility and violence in intergroup relations.
Thus, if the perpetrator of a terror attack shares externally perceptible features with members of a societal outgroup (e.g., an Islamist terrorist in the Western context), the "suspect community" constructed in public discourse will be that larger outgroup (e.g., Muslims in Western societies). The normative challenge for terrorism communication lies in reporting about outgroup perpetrators without constructing the community with which they share some salient characteristic as suspect of terrorism. If on the other hand the perpetrator shares characteristics with the dominant societal ingroup, e.g., a light-skinned man with domestic citizenship in a Western country such as the Norwegian (2011) or the Christchurch (2019) attacker, the construction of the entire majority ingroup as a "suspect community" is culturally incongruent. The normative challenge here lies in avoiding the temptation to trivialize, or even implicitly excuse, the actions of ingroup perpetrators.
Step 2: mapping competing normative horizons Journalists and citizens follow different, often conflicting and sometimes mutually exclusive values in publicly addressing contentious issues. MNA is not designed to resolve these disputes, but to disentangle them and make the respective claims more mutually understandable and criticizable. As mentioned earlier, the mapping of normative horizons can be based on explicit normative theorizing directly or on communicators' own normative meta-reflections. Our argument below concerning the competing normative horizons is in part a reconstruction of actually held beliefs, in part a delineation of underlying or associated philosophical traditions, and in part also a constructive attempt at creating ideal-types. Ideal-types specify the constellation of defining characteristics of a phenomenon or worldview while abstracting from some empirical variation. The ideal-types we will define are designed to bring into sharp relief the normative choices that researchers face when they engage in systematic normative assessment of communication as well as the choices professional and nonprofessional communicators must make when they make themselves heard.
The contentious nature of public communication is particularly evident in the domain of terrorism communication. Different actors and groups fundamentally disagree on many aspects including which acts of political violence to call terrorism in the first place and which to label acts of insurgency or resistance (Moeller, 2009). In his discourse analysis of debates surrounding the so-called Copenhagen killings 2 of 14 February 2015, Horsbøl (2016) demonstrates that public articulations on the issue of terrorism often carry implicit normative conceptions of publicness and the public sphere in them. 3 We aim at uncovering and systematizing these implicit normative horizons to generate distinctive alternative responses to the normative challenges identified above (see the four models in Table 2). Horsbøl (2016) bases his analysis on a case of terrorism in which an outgroup attacker (a radical Islamist) victimized ingroup members. As we strive to define normative horizons that can cover both outgroup and ingroup terrorism, we will amend and extend his findings in the following. Horsbøl (2016) identifies four distinct conceptions of publicness with alternative normative reference points that connect well with more general, established normative theories of the public sphere. 4 (1) Horsbøl's discourse of "Enlightenment and freedom of speech" aligns well with the broad liberal tradition in normative public sphere theory, which revolves around the central value of individual freedom (Ferree et al., 2002, pp. 206-209).
For a corresponding Freedom model of terrorism communication, therefore, the preferred option for action lies in using freedom of speech to openly deal with conflicts, courageously defying threats from terrorists and protecting those targeted by them. Normatively, public communication in the face of terrorist threats is characterized as a fight for freedom, justice, and equal rights against the forces of intimidation and obscurantism (Horsbøl, 2016, pp. 12-13). According to the Freedom model, the public sphere is thought of as a "free marketplace of ideas," which must be defended against the interference of a potentially authoritarian state and oppressive actors as well as against undermining tendencies such as self-censorship and cowardice in the face of threats.
(2) Horsbøl's "Western culture and freedom of speech" discourse with its Huntington-style juxtaposition of conflicting community values can be read as a combative, culturally essentialist variant of a communitarian conception of the public sphere. While this second normative conception is also opposed to self-censorship and wants to protect those threatened or harmed by terrorists, this posture is defended not by recourse to universal values (freedom, justice, equality) but to proclaimed Western values and a Western model of society. Public communication is equally characterized as a fight or even war, but this time it is not the struggle between Enlightenment and obscurantism, but between Western and non-Western values, culture, and religion that serves as the main metaphor. According to this Community Values model, public communication is supposed to serve as an arena in which citizens who share the same basic values can exchange views to determine what is best for them. Values are thought to be "owned" by clearly demarcated groups (Ezzati, 2020) and must therefore be defended and reaffirmed through public communication. 5 (3) In direct opposition to this ingroup-oriented perspective stands the discourse of "tolerant dialogue in a multicultural society" that Horsbøl (2016) identifies as a third normative conception of publicness. In this view public communication is likened not to a fight or war, but to the ideal of peaceful interaction expected in families or the schoolyard. Respect and tolerance, particularly toward societal outgroups, is the main public virtue favored in this perspective with the aim to stop bullying, mocking and insulting minorities. The existing freedom of speech should instead be used against the powerful to foster an "open and tolerant form of interaction" between majority and minority groups (p. 11). This conception of publicness aligns well with some intentions of the agonistic tradition of public sphere theory (Mouffe, 2013). In light of fundamentally irreconcilable conflicts between different political groups, including those between a dominant majority and underprivileged minority groups, agonists privilege the empowerment of the marginalized. However, in the spirit of "agonistic respect" political opponents are supposed to see each other as adversaries, not as enemies.
(4) Finally, Horsbøl's discourse of "Democracy and moderation" is very close to the deliberative tradition of public sphere theory, which prizes well-considered, constructive debate across lines of disagreement (Habermas, 1996). In relation to terrorism, the central metaphor for public communication is an insistence on the power of everyday life. "Moderation" here means to avoid overreaction and to reflect and navigate the "media dilemma" that lies at the heart of all terrorism communication, namely that terrorists might become "media stars" in their own action drama if communicators don't restrain themselves (Horsbøl, 2016, p. 12). According to the Constructive Debate model everybody potentially affected by an issue or political decision should participate, or at least all relevant perspectives should be represented, in public debate. Positions and opinions should be justified, and communicators should address each other's concerns explicitly in a respectful demeanor. Mutual perspective-taking (Muradova, 2021) is the key to finding common solutions or at least arriving at a more "reasoned dissent" (Wessler, 2008).
To further specify the substance of the normative horizons the MNA procedure entails a mapping of their demands onto the set of normative challenges identified in Step 1 (see Table 2).
(1) The Freedom model suggests that citizens need verifiable information to both protect themselves and form opinions about terrorist threats. This will also be supported by a composed, unemotional tone in terrorism communication. The Freedom model also places considerable confidence in public debate by assuming that, in the long run, terrorist messages will self-defeat in the marketplace of ideas and that fact-based analysis on causes and consequences will help combat the terrorist threat. Public communication should also, according to this model, signal support for those in danger of being intimidated by terrorist threats (Horsbøl, 2016, p. 11). In relation to outgroup communities who could be constructed as suspect of terrorism, the liberal model aims at upholding "zones of indifference," highlighting the principle of cultural and religious neutrality of the liberal state rather than an obligation to actively seek a closing of ranks within the majority ingroup or opening up to outgroup minorities.
(2) The concerns of the Community Values model are best served by a resolute, even combative tone against terrorists in order to defend the community values against their enemies. In comparison, the verification of information is not a core concern in this normative perspective. Terrorists should be denied public voice, and analysis should aim at finding the best solution for guaranteeing the safety of the majority community. While the Community Values model is clearly concerned with supporting ingroup victims, it is unclear how it would play out in cases if ingroup terrorism against vulnerable minorities. Finally, outgroup members should only be symbolically included in the ingroup community if they publicly dissociate from terrorists and pledge allegiance to the essentialized ingroup values.
(3) If we apply the Empowerment model to the normative challenges of reacting to terrorism it becomes clear that the focus of concern is on outgroup minorities. The model aims at giving these groups public voice and including them as equals in a multicultural conversation. The demand that outgroup leaders should publicly distance themselves from terrorists is seen as discriminatory in itself and as perpetuating a general-and undue-suspicion that outgroups support terrorism against the majority community. In the case of ingroup terrorists public communication should signal support for outgroup victims; conversely, the relation to ingroup victims remains unclear in this model. In addition, the marginalization of outgroups is seen as a contributing cause of the kind of radicalization that might lead some alienated and unstable outgroup individuals to support or commit terrorist acts (see Horsbøl, 2016, p. 10). Compared to these solidarity-related aspects the Empowerment model has no special priority concerning verifiable information. Finally, in terms of the tone of coverage and social media exchanges this model privileges a composed atmosphere in which marginalized actors can assertively criticize tendencies of exclusion or discrimination and respectful exchanges across ethnic and cultural divides can develop.
(4) According to the Constructive Debate model, finally, verifiable information is supposed to provide the information base for well-informed debate that includes all parties affected except perpetrators. A special feature of the Constructive Debate model lies in the demand for meta-reflection on the rules and conduct of public debate itself in order to temper reactions and bind them to normative standards. The tone is supposed to be composed on the one hand, but compassionate with the victims and survivors of terrorism whoever they might be. In order to facilitate a truly open discussion no community should be excluded or relegated to a lower rank so that, much like in the Empowerment model, outgroups should not be asked to distance themselves from outgroup terrorists because that tacitly raises unfounded suspicions.
Overall, the Freedom (1) and the Constructive Debate model (4) are complemented by a Community Values model (2) primarily concerned with the societal ingroup and an Empowerment model (3) essentially aiming at protecting marginalized outgroups. Some limited overlap becomes visible between models 1 and 2 and models 3 and 4, respectively. But on the whole the models provide distinctive normative responses to the same set of normative challenges by prioritizing particular challenges and target values over others and by specifying unique normative expectations. In addition, Horsbøl's (2016, pp. 14-16) analysis has shown that there is explicit counter-positioning between the four models in public discourse so that the delineation is not simply imposed by an outside observer but inherent in the public articulation of the models itself.
Step 3: measuring performance levels empirically In this theoretical contribution we have no space to actually conduct the empirical measurement of communicative performance that constitutes the third step of any MNA. Suffice it to say that this step, in which the normative horizons are operationalized in performance indicators, eventually results in an assessment matrix or narrative that specifies the degree to which empirical performance levels fulfill the expectations of each normative horizon.
Step 4: devising decision rules to derive recommendations The final, public-facing, step of MNA reconnects the procedure to communicative practice in actual media settings. But what is the best way to proceed from a multiperspectival assessment result to actionable recommendations? Three alternative routes present themselves: 1. Choose the lowest common denominator on which all four models could agree 2. Privilege the most demanding variant to make sure the exigencies of all models are substantively included or 3. Look for a middle ground that levels out particularly low and high expectations.
For the recommendations presented below we have generally opted for the "most demanding" approach. This increases the normative stakes at play and thus potentially also the extent of improvement brought about by following the respective recommendations. By contrast, the "lowest common denominator" approach would have generally lowered the standards to a point where the normative leverage and distinctiveness of the four models would have been lost. The "middle ground" approach was followed in cases where the standards proposed by the models directly contradict each other. The result of these considerations is summarized in Table 3.
The recommendations are not designed as utopian ideals but rather as things that journalists and actively contributing social media users can actually do or avoid. 6 In this way we hope to contribute not to an unattainable ideal, but to a

16
Communication Theory 00 () 1-24 "social science of the possible" (Wright, 2013; see also Wessler, 2020). Having said that following these prescriptions will require that communicators develop a considerable level of self-reflection as well as resistance to cognitive shortcuts and entrenched routines. But that is what normative discussions are all about: that social practices are not just reflected, but improved in relation to the normative goals specified. The recommendations listed in the following paragraphs constitute guidelines, not absolute demands. We acknowledge that in concrete situations different exigencies might compete and circumstantial conditions need to be taken into account to arrive at ethical decisions. But this does not and should not preclude the prior formulation of guidelines for action. On the question of verification both the Freedom and the Constructive Debate models call for a high-quality information base. Thus we recommend that journalists and social media users refrain from circulating unconfirmed information such as rumors, speculation, and unedited live streams from attack sites. In a viral sharing environment not doing something (i.e., sharing unconfirmed information) can be a positive contribution to responsible terrorism communication in its own right.
Similarly, a composed, even sober tone seems to be the most demanding standard in situations of horror and uncertainty. Journalists would therefore be well advised to use neutral, factual language and refrain from narrative or musical dramatization of the attacks. Both journalists and social media users would honor the seriousness of the situation by reflecting for a short while before posting emotional statements.
On the issues of voice and analysis we recommend that journalists and social media users alike refrain from giving perpetrators direct voice through videos, online manifestoes or material from sympathizers. Journalists should be careful to not give perpetrators too much attention and prominence while social media users should not circulate such perpetrator profiles. 7 Journalists are also called upon to talk about the causes and consequences of terrorism in an analytical manner and with the help of expert advice, while media users should seek out such analysis from the available menu of sources. Finally, the Constructive Debate model goes one step further and asks that journalists publicly reflect on the performance of public debate and on the appropriateness of their coverage as a mechanism of collective self-improvement.
Concerning victimization, we have seen that the Community Values model privileges the ingroup while the Empowerment model prefers the outgroup. In our recommendations on this point we therefore opt for an inclusive "middle ground" approach that accounts for both the ingroup and the outgroup and asks communicators to treat all victims with equal standards. In addition, special decency standards apply in relation to human suffering. To preserve the dignity of victims and not victimize them further we thus recommend that communicators do not expose victims or make them directly identifiable, at least not without obtaining prior consent by victims or their relatives.
Finally, communicators are called upon to counter the construction of suspect communities by clearly differentiating between individual perpetrators and the larger communities for which they purport to act (Chan, Wessler, Rinke, Welbers, van Atteveldt, & Althaus, 2020;von Sikorski et al., 2017). This includes, as posited by the Empowerment and the Constructive Debate models, not pushing for public denouncements of terrorism by outgroup representatives because that would subtly perpetuate precisely those terror suspicions. In addition, ingroup and outgroup perpetrators should be treated with the same standards to resist trivializing ingroup terror. In fact, communicators should judge and label acts of violence according to transparent and stable attributes rather than according to the origin of the perpetrator.

Conclusion
In this article we hope to have shown that both time-tested and more recent normative theories of the public sphere can be specified and applied to an issue domain-in this case terrorism communication-in fruitful and insightful ways. If the normative challenges and dilemmas to be addressed are defined in sufficiently concrete terms (step 1), the responses derived from different normative traditions can be both internally consistent and distinctive (step 2). Those divergent responses can then be used for systematic normative assessments of the empirical realities of communication in the issue domain (step 3). In addition, such a specification of normative expectations does not have to remain an ivory-tower pursuit simply for the sake of theoretical clarity. When unambiguous decision rules are developed for the reconciliation of diverging demands, then the competing normative considerations can indeed be translated into concrete recommendations for action (step 4). These, in turn, might inform broader professional and public debate on how to communicatively deal with contentious issues.
Beyond the issue-specific instantiation of relevant normative demands and practical recommendations we present in this article, we hope to have provided a general model for how systematic, multiperspectival normative analysis can be conducted across other research domains. Apart from defining "responsible terrorism communication," therefore, we hope to encourage with this article more explicit-and nuanced-normative reflection and assessment in future communication scholarship. The MNA procedure is sufficiently generic to accommodate public communication about a very broad range of issues spanning from existential threats such as terrorism or the coronavirus pandemic through moral issues like abortion or same-sex marriage to resource-based conflicts over job creation or taxation. In all these cases the sequence of steps does not need to be altered.
What is at stake, however, in future applications of the MNA procedure is the generalizability of the normative horizons that we have specified. It remains to be seen how much of public contestation can be fruitfully covered by the Freedom, the Community Values, the Empowerment, and the Constructive Debate models. As we have seen in Step 2 above, these models are linked to general theories of democracy and the public sphere (i.e., the liberal, communitarian, agonistic and deliberative traditions). At the same time, applying these normative traditions to the case of terrorism communication required some alterations or shifts of emphasis in specifying the issue-specific normative horizons. For example, the Community Values model took on a particularly combative and culturally essentialist flavor when it was connected to the ingroup/outgroup dynamics present in terrorism coverage. There may also be issues that are publicly discussed in ways that only speak to two or three of the models and traditions, for example when the issue does not involve marginalized minorities and the agonistic tradition consequently seems less insightful. We would like to encourage researchers across the discipline to probe the set of normative horizons we have mapped in this contribution. With more applications of MNA in communication research in the future the discipline's explicit knowledge about the nature of its normative foundations should be able to grow. But even if the set of normative horizons would require major alterations in relation to a particular class of issues, we contend that the MNA procedure itself is flexible enough to guide normative inquiry in illuminating ways.
Funding 1 To explain the extent of the panic, it should be noted that the Munich attack happened only eight days after the attack in Nice, France, in which an Islamist terrorist killed 86 people by driving a truck through the crowds on the seafront promenade. Naturally, many people in Munich were more sensitized to the dangers of violent attacks than they would otherwise have been. 2 The Copenhagen killings consisted of two attacks on the same day committed by the same radical Islamist with Danish citizenship killing two people, one in a café and one in front of a synagogue. On the following day, the perpetrator was found and killed in a gunfire exchange with the police. 3 The following delineation of the four alternative normative visions of publicness is based on three analytical dimensions employed in Horsbøl's (2016) discourse analysis: (a) agency: antithetical options for public action (such as self-censorship versus defense of values, etc.), (b) main public virtue (such as courage or composure or respect), and (c) main metaphors for public communication (such as fight or war or friendly interaction). Each discourse or vision of publicness is characterized by a typical combination of tropes on these dimensions. 4 It is striking that these alternative models of publicness correspond well with some of the prima facie duties as defined by ethicist W.D. Ross and further differentiated by Meyers (2011), pointing to promising additional sources of theoretical justification for these models that have been all but ignored in normative public sphere theory (Ferree et al., 2002;Wessler, 2018). The strongest connections seem to exist between the Freedom model and the duties of honesty, respect, and formal justice; the Empowerment model and the duties of nonmaleficence and beneficence or even reparation (vis-à-vis marginalized minorities); and between the Constructive Debate model and the duties of honesty, respect, nonmaleficence, and distributive justice. 5 To be sure, focusing on the public defense of essentialized community values does not do justice to communitarian theorizing of democracy and the public sphere as a whole (see, for example, Barber 2003). We concur here with Freelon's (2015, p. 775) reading that in media debates communitarianism "manifests itself in part as heavy levels of conversation and participation with like-minded others with comparatively little direct interest in outsiders except as adversaries." Communitarian conceptions of the public sphere also bear resemblance with inspiring, newer approaches in the field of virtue ethics, which conceptualize virtues not merely as a form of individual self-improvement but also as a project of communal human flourishing (for a programmatic outline, see Plaisance, 2016). 6 We restrict our recommendations to users actively posting about terrorist activity on various social media platforms as we are concerned with how such active posting can help spread terrorists' own messaging and create panic that potentially endangers people not directly involved in an attack. We also acknowledge that not only users, but also social media platforms carry responsibility for the quality of terrorism communication.
But an in-depth discussion on this issue is beyond the scope of this article, particularly because the proper locus and forms of "content moderation" are controversial in both academic and popular discussions. The big technology companies including Facebook, Microsoft, Twitter and YouTube have founded the Global Internet Forum to Counter Terrorism (gifct.org), which supports content filtering and deletion with "a shared, but secretive database of known terrorist images, video, audio, and text" (Katzenbach & Ulbricht, 2019, p. 10). 7 These recommendations are in line with the media protocol promoted by the No Notoriety movement in the USA (https://nonotoriety.com/).