Emotions in the European Union's decision-making: the reform of the Dublin System in the context of the refugee crisis

Abstract

The European Union's decision-making process has mostly been pictured as a technocratic and consensual process with little or no place for the logics of passion. The reform of the Dublin System in the wake of the refugee crisis (2015–16) has been analyzed from different angles but not from a political psychology perspective in which emotions are fully taken into account. Crises are often seen as a window of opportunity because key players engage in a strategic process of sense-making. This article explores the role of emotions in this process: to what extent do emotions – seen via crisis-framing processes – contribute to the understanding of the reform of the Dublin System? To this purpose, I show first how emotions are related to crisis-framing efforts. Second, I discuss how the study of emotions contributes to the understanding of decision-making processes. I carry out a critical-frame analysis illustrated by a systematic qualitative content analysis of primary documentation. Data include European Parliament debates on the topic of the refugee crisis (covering also the points of view of the Commission and the Council), and policy documents and speeches by the main actors in the decision-making process, including relevant member states and non-governmental organizations.

Keywords:

• refugee crisis
• emotions
• European Union
• framing
• dehumanization

1. Introduction

The so-called refugee crisis is a critical case for improving our understanding of the emotional component of crisis-framing and its potential effects on the decision-making process. This particular crisis is interesting for analytical purposes because it has at least two main formal decision-making moments with two different policy outcomes: the decision to adopt the quota relocation system in 2015; and the Reform of the Asylum system, a legislative process which started in 2016 and which – after the removal of the main legislative proposal in 2018 – is still ongoing.

The decision-making process regarding the relocation of asylum seekers during the so-called refugee crisis has been analyzed from different angles (Ripoll Servent 2019; Zaun 2018; Niemann and Zaun 2018; Trauner 2016), but not from a political psychology perspective in which emotions are fully taken into account. The political psychology perspective challenges the rationality models according to which there would be a rational policy process (Dyson and 't Hart 2013). Rather than assuming a dualistic rational ideal in which emotions and cognition are separated and in which the blame is usually placed on human psychology for its alleged violations of rationality, this article takes into account the latest developments in cognition/emotion research: there is increasing consensus around the idea that emotions serve a productive function and that emotions have both preconscious and conscious aspects (McDermott 2004; Brader and Markus 2013).

A state of crisis assumes the perception of a serious threat to the fundamental values and basic structures of a given society and is characterized by a high degree of uncertainty and a sense of urgency (Dyson and 't Hart 2013). From the political psychology perspective, crises constitute a window of opportunity, not only because of the urgent need to find policy solutions, but also because policy-makers and key players engage in a process of sense-making (Boin, 't Hart, and McConnell 2009; Dyson and 't Hart 2013). Actors inside and outside government need to diagnose the nature of the situation and produce an authoritative account of what has happened. While political psychology usually focuses on individual-level analysis (Demertzis 2013), my perspective aims at establishing links at the macro level; thus, it is also aligned with the research agenda of the political sociology of emotions.

In crisis situations, policy leaders engage in a process of crisis-framing: they need to explain the significance of the situation and the causes of it, and devise policy solutions. In this process of crisis-framing, policy-makers often engage in crisis exploitation, defined as the purposeful utilization of crisis rhetoric to alter the level of support for public office-holders and public policies (Boin, 't Hart, and McConnell 2009). Emotions lie at the heart of responses to crises, particularly those emotions that are considered to be negative, such as anger, fear and anxiety. Given the prevalence of these so-called negative emotions, crises offer an effective stage for exploitation by populist groups (Moffitt 2016). Crisis-framing and emotions are obviously not the only explanations that matter for the understanding of any decision-making process. The study of emotions and the world-views that they bring need to be understood in combination with other relevant factors already highlighted in mainstream research on European studies.

The main question that this article seeks to answer is: To what extent have emotions – via crisis-framing processes – contributed to the understanding of the decision-making process and the choice of policy solutions? To answer this question attention is drawn to the following sub-questions: How was the so-called refugee crisis framed by different key players? To what extent did emotions support existing frames? How have emotions contributed to the shaping of the decision-making process in 2015 and in 2016/18?

The empirical part is based on an in-depth critical-frame analysis and a systematic content analysis of policy documents from key European Union (EU) players, including European Parliament (EP) debates, and speeches from Council representatives, Commission officials and Civil Society Organization (CSO) representatives. First, the analysis shows that key players hold positions along two axes: a human dignity policy, based on compassion; and a security policy, based on fear. This article also discusses the emotion-based strategies used by key actors, including naming and shaming, fearmongering and de/humanization. The last section explores the evolution of the emotional environment during the main two decision-making moments and shows how the prevalence of different emotions can help us understand the policy choices made.

2. Understanding the role of emotions in decision-making processes

This first section introduces existing research on the decision-making processes during the refugee crisis; it proposes an analytical framework and methodology to explore the role of emotions in decisions taken during a crisis situation.

2.1. Current understandings of the decision-making process in the refugee crisis

The decision to retain the Dublin System has often been perceived by academic researchers as a failure (Zaun 2018; Ripoll Servent 2019; Trauner 2016). The Dublin System is aimed at rapidly determining which EU member state is responsible for the examination of an application for asylum. In practice, border countries are responsible for any asylum seeker entering the Schengen area through their territory.

The perception of failure is mainly related to the malfunctioning of the system, which was especially visible during the so-called refugee crisis. Front-line member states (mainly Greece and Italy) did not implement the first-country-of-entry principle and allowed refugees to move on to other countries. This so-called wave-through approach resulted in secondary movements toward northern Europe. Hungary even stopped taking candidates altogether as early as June 2015 (EP 2016). In August 2015, Germany decided to admit asylum seekers into their national asylum system irrespective of their first country of entry. However, after a couple of weeks of Willkommenspolitik (welcome policy), Germany introduced border controls to avoid secondary movements, and so did many EU member states.

How did EU key players respond to the malfunctioning of the Dublin System? The EU response consisted of two formal decision-making processes, each of which had a different outcome. The first led to the adoption of an emergency relocation scheme for a total of 160,000 refugees. This system was adopted in combination with a hotspot approach aimed at ensuring that front-line member states would fulfill their obligations in terms of registering migrants’ claims. The Council took these decisions in the midst of the crisis, after the Commission triggered the emergency response system envisaged in Art. 78(3) of the Treaty on the Functioning of the European Union. The introduction of a fairness mechanism could be interpreted as a complete overhaul of the Dublin System and as a first step toward a truly supranational asylum system. The creation of the relocation system and the hotspots approach could also be interpreted as simply adding new layers of policy instruments to existing regulations to maintain a sub-optimal policy core (Trauner 2016).

The second decision-making process started in 2016 with a package of legislative proposals aimed at reforming the Dublin System. The main innovation was a corrective allocation mechanism to help member states under pressure. Despite substantive divisions in the EP, this EU institution reached a common position (a mandate for negotiations) in November 2017; this included the solidarity mechanism proposed by the Commission. In sharp contrast, the Council focused on outsourcing the problem via the EU–Turkey deal¹ and rejected the idea of a common relocation system. In the absence of the adoption of a common position by the Council, the Commission's proposal was withdrawn and the Dublin System was not substantially reformed.

In September 2020, the Commission adopted a proposal for a new migration pact, including the idea of a mandatory solidarity system (Euractiv, September 21, 2020). While mandatory, this pact includes a flexible type of solidarity that allows reluctant countries to focus on the return of asylum seekers; for this, it has been criticized by non-governmental organizations (NGOs), which consider that the Commission's proposal is like ‘asking the school bully to walk the kid home’.²

It is not surprising that the EU's incapacity or unwillingness to reform a malfunctioning system for so many years has attracted scholarly attention (Ripoll Servent 2019; Zaun 2018; Niemann and Zaun 2018; Trauner 2016). Researchers tend to agree that the inability to reform the Dublin System can mainly be explained via liberal inter-governmentalist arguments. Liberal inter-governmentalism understands European integration as a series of rational choices made by national leaders based on a process of national preference formation. In the reform of the Dublin System, governments were pressured into adopting restrictive policies in response to the electoral gains made by right-wing populist parties (Zaun 2018). The dynamics of the politization and mediatization of the refugee crisis at the national level have already been analyzed from the perspective of critical discourse analysis (Krzyzanowski et al. 2018). However, there is as yet little research about such dynamics at the EU level.

Liberal inter-governmentalism has also been considered to be particularly applicable to crisis situations. During the so-called refugee crisis, heads of government and state would have been the key decision-makers, casting a shadow over agenda-setting and day-to-day decision-making (Ripoll Servent 2019). While most articles on this topic point to insufficient European cooperation, giving some credit to liberal inter-governmentalism, some authors have highlighted that neo-functionalism provides crucial insights for the understanding of certain aspects of the EU response to the refugee crisis, as for example, the regulations on the European Border and Coast Guard. According to this view, the existing dysfunctionality between a supranational Schengen system and a weak external border regime was exposed during the refugee crisis. By creating new functional pressures, the crisis would have contributed to a change of mentality: external borders would increasingly be perceived as common borders, giving supranational powers to the European Border and Coast Guard and implementing an integrated border management system (Niemann and Speyer 2018).

My analysis builds on existing research and adds to existing findings in two ways. First, it compares the decision-making moments of 2015 and 2016–18. Existing articles explain the non-adoption of the refugee quota system in 2018, but there has been little discussion about why and how this same system was temporarily adopted in 2015. Second, while traditional theories of European integration have greatly contributed to the understanding of the refugee crisis, they tend to simplify or overlook the role of emotions by assuming an instrumental rationality based exclusively on cognition. According to Dyson and 't Hart (2013), to fully understand the outcome of decision-making processes, attention needs to be turned to sense-making and meaning-making. Existing studies on European migration policies have often emphasized the securitization of migration (Huysmans 2000; Bonansinga 2019). Securitization refers to the definition of migration as a security issue, by for example establishing links with terrorism, drug trafficking and organized crime. While the securitization of asylum policies seems to be a confirmed ongoing process, little is known about what makes this process possible and acceptable by public opinion and policy-makers. The analytical framework presented in the next section also contributes to a better understanding the securitization of asylum policies.

2.2. Analyzing emotions in crisis situations: framing and emotion-based strategies

A crisis can lead to a process of crisis exploitation in which actors engage in a frame contest to exploit the window of opportunity opened by the crisis (Boin, 't Hart, and McConnell 2009). A frame is ‘an interpretive schema that simplifies and condenses the world out there by selectively punctuating and encoding objects, situations, events, experiences, and sequences of actions within one's present or past environments’ (Snow and Benford 1992, 137). The outcome of the crisis in terms of decision-making depends on which crisis narrative has taken hold. The theoretical framework proposed by Boin et al. to explain policy outcomes focuses on the framing process of the status quo players and of the advocates of change. While this model is a promising starting point, I consider that to better understand framing processes and which narratives prevail in the crisis contest, it is of utmost importance to take into account the role of emotions. While there is not a generally accepted definition or typology of emotions, the following elements are often highlighted: the activation of key bodily systems; the appraisal of situational stimuli; overt or inhibited expression; and socially constructed labels and rules (Demertzis 2013). In social sciences, emphasis is usually placed on cognitive interpretations (appraisals) as the main triggers of emotion (Brader and Markus 2013). This article focuses on fear, anger and empathy, which are generally considered to be both biological processes and intersubjectively mediated reactions to perceptions that depend on the context and on pre-existing social understandings (Crawford 2014). The biological basis of emotions such as compassion³ and shame is not so clearly established.

Research on framing processes has lately given increasing focus to the role of emotions (Lecheler, Schuck, and de Vreese 2013; Lecheler, Bos, and Vliegenthart 2015; Gross and D’Ambrosio 2004). The general idea is that frames lead individuals to experience different emotions and that these emotional reactions mediate framing effects on opinions and political behavior (Lecheler, Schuck, and de Vreese 2013; Lecheler, Bos, and Vliegenthart 2015). Emotions can thus help us understand how framing effects take place since they play a decisive role when decisions are made after exposure to a frame.

In taking the role of emotions into account, the literature on framing has started to fully integrate research on the role of emotions in decision-making processes. The role of emotions in judgment and decision-making has gained increased attention across disciplines ranging from philosophy to neuroscience (Lerner et al. 2015; McDermott 2004). While current studies include competing theories and few definitive conclusions, they tend to agree that emotions powerfully and pervasively influence decision-making. Existing research has shown that emotions influence individual and interpersonal decision-making processes, for example, via depth of thought or goal activation. They have also shown that the possible unwanted effects of emotions can be reduced under certain circumstances.

How can emotions influence decision-making processes and how can they affect policy solutions? According to appraisal-tendency theory, an emotion, once activated, can trigger a cognitive predisposition to assess future events in line with the central appraisal dimensions that triggered the emotion (Lerner et al. 2015). In other words, emotions such as fear change what we look for, what we see and the way we think (Crawford 2014). More often than not, individuals are not aware on the effects that emotions have on their assessment of reality or on their choices. This theory is in line with the argument that specific emotions carry action tendencies or implicit goals. For example, anger has been associated with the desire to change a situation and to move against another person/obstacle by fighting, harming or conquering (Frijda 2007; Lerner et al. 2015). Emotions can also become institutionalized in a process in which emotion-based perception leads to specific arguments and policy responses that, with time, become institutionalized (Crawford 2014). Institutionalization of emotions occurs when the framing of problems and solutions are normalized and become taken for granted. Not only the dominant beliefs, but also the feelings of the groups are translated into practices and procedures designed to meet emotional needs and organizational goals.

If emotions can shape cognitive predispositions and can be associated with specific goals or policy outcomes, they can play a key role in crisis-framing processes in the dynamics of crisis exploitation. Since the two decision-making moments previously highlighted led to different types of outcome, the role of emotions in the framing contest may have evolved. This change is captured by the three sub-questions: What are the different competing crisis frames? What is the role of emotions in each one of these? How did emotions contribute to shaping the decision-making process in 2015 and in 2016/18?

The so-called refugee crisis gave rise to competing policy frames and the different decisions adopted can be partially explained by different configurations of these policy frames games in each one of the decision-making moments. Some examples of competing frames identified in previous studies are the pull-factor frame; the migration as a tragedy frame; and the push factor frame (Abdou 2020). The so-called pull-factor migration frame focuses on the increasing flows of asylum seekers, while the migration as a tragedy and the push factor frames focus on the effects of migration interpreted as a human tragedy in which the migrants are seen as victims.

The following analysis gives specific attention to the role of emotions. First, attention is drawn to emotions, such as fear, compassion and anger, which seem to have a relevant role in debates on migration (Sanchez Salgado 2018 and 2020). The attention is then drawn to the connection between the emotions prevalent in the frames, the emotion strategies and the policy solutions (see Table 1). Many studies have established connections between emotions and policy solutions. For example, compassion and empathy imply a movement of moving forward (Frijda 2007) or a motivation to better other's situation (Halperin and Pliskin 2015). Compassion and empathy have thus been related to humanitarian impulses, justice and peace (Käpylä and Kennedy 2014; Crawford 2014). The development of out-group empathy is also considered essential for the promotion of human rights and democracy. Policy solutions related to empathy and compassion have also been associated with helping behavior or, more specifically, with pro-social activism.

Table 1. Connection of basic emotions to action tendencies and emotion-based strategies.

Master emotions	Action tendencies	Emotion-based strategies (examples)
Fear	Neutralization of danger Closure Avoidance of risk	Fearmongering Dehumanization
Compassion Empathy	Improve others’ situations Helping behavior Pro-social activism	Humanization Promoting communication and cultural interaction
Anger (combined with fear or compassion)	Punishment Regulation Risky behavior Change	Naming and shaming
No emotion	No action/decision/urgency	Dehumanization (mechanistic) De-dramatization Toning-down

Fear involves low certainty and low sense of control, and produces a perception of negative events as unpredictable and situationally determined (Lerner et al. 2015). Regarding action tendencies, fear leads to the neutralization of danger (Frijda 2007), closure and avoidance of risk (Halperin and Pliskin 2015). The study of fear and empathy reveals that these two emotions cannot easily be combined. Along with other conditions such as stress and competition, fear tends to diminish empathy (Crawford 2014). Even incidental fear (random fearful imaginary) negatively affects empathy toward out-group suffering (Richins et al. 2019). Explicit competition among groups (about, for example, access to jobs or healthcare) can even create counter-empathic reactions.

Anger in this research has a different status, since it has often been pictured as very close to both fear and compassion. Anger often emerges in situations when people are threatened or frustrated and, in this sense, situations that elicit fear can also produce anger (Brader and Markus 2013). However, people also get angry without suffering the negative consequences themselves. In cases of third-party anger, anger is connected to empathy and compassion since people get angry when they perceive that an injustice or a moral violation has taken place (Landmann and Hess 2017). The combination of fear or empathy with anger leads to action tendencies and policy solutions that are different from the ones related to fear and empathy previously mentioned. Angry people tend to view negative events as predictable and caused, and as under the control of individuals (Lerner et al. 2015). Anger has also been connected to punitive, risky and confrontational attitudes (Brader and Markus 2013). Anger can motivate people to take control of a situation and ameliorate the problem at hand (Turner 2007).

Attention will then be turned to emotion-based strategies, which will help us understand how emotions can contribute to the shaping of decision-making processes. The decision-making process can be understood as a process of institutionalization of different emotions (Crawford 2014). This process of institutionalization can take place through emotion-based strategies such as fearmongering, naming and shaming, and de/humanization. The institutionalization of a specific emotion implies that the action tendencies previously related to this emotion – and the connected policy solutions – are also institutionalized.

For the institutionalization of fear, two typical strategies that can be employed are fearmongering and dehumanization. Fearmongering is usually attributed to populist groups and consists of deploying fear through narrative techniques to normalize errors in reasoning through repetition, misdirection and the presentation of isolated occurrences as trends (Glassner 2004). Dehumanization involves the denial of human characteristics to human beings by presenting them as animal-like (animalistic dehumanization) or as objects or automata (mechanistic dehumanization) (Haslam 2006). Animalistic dehumanization includes depicting human beings as animals (e.g. describing them as rats, vermin, etc.) or denying human characteristics such as rationality, self-control, intelligence and rationality. In debates regarding mobility, the use of water metaphors in the process of dehumanization is also frequent (Kainz 2016).

Strategies for the long-term institutionalization of empathy and compassion include promoting trade, communication and cultural interaction (Crawford 2014). Following these strategies, empathy is considered to be institutionalized in the EU and in other pluralistic and amalgamated communities. Another emotion-based strategy that can be employed in the short-term dynamics of framing and argumentation is humanization, which, in discursive dynamics, can be related to an emphasis on personalization and emotional content.

When fear and empathy/compassion are combined with anger, the process of institutionalization can take place through naming/blaming and shaming. This is widely used by NGOs and consists of appealing to guilt or shame to generate a desired action (Käpylä and Kennedy 2014). Existing research on the effects of naming and shaming is not conclusive. Some studies consider that it is a successful practice to reduce, for example, human rights violations, but it has also been often considered as an unpredictable policy tool (Krain 2012; Sanchez Salgado 2020). The use of naming and shaming can indeed alienate CSOs from potential allies.

Lastly, key players may adopt impartial, objective and rational language so that they appear more convincing (Palm 2018). This strategy can sometimes contribute to mechanistic dehumanization, in which human beings are denied warmth, emotion and individuality. Most early literature on dehumanization perceived it as an extreme phenomenon circumscribed to racial and ethnical inter-group conflict. More recently, dehumanization has also been understood as an everyday social phenomenon (Kalina 2014; Haslam 2006). Mild dehumanizing attitudes and behaviors are often considered beneficial and necessary. For example, people in power are often expected to make ‘tough’ decisions that may cause suffering to others in a cold, distant and rational manner. It is also widely accepted that policy-makers engage in strategies aiming at toning-down emotions. However, these supposed benefits of mild denumanizing behavior have not been supported by scientific evidence (Kalina 2014).

2.3. Analyzing emotions in political speeches

The analysis combines an in-depth critical-frame analysis with a qualitative content analysis. Following the critical-frame analysis approach, policy frames will be analyzed along two key dimensions: the diagnosis (what is the problem?) and the prognosis (what is the solution?) (Verloo and Lombardo 2007). Other questions address specifically the role of emotions (what are the underlying emotions that are being institutionalized within each frame?) and situate these in a dynamic perspective (which emotion-based strategies contribute to this process of institutionalization?). The qualitative content analysis illustrates the in-depth critical-frame analysis and interpretation.

This analysis also includes a comparative temporal dimension to determine if there was a difference between the first decision-making moment (September 2015) and the second (2016–2018). It is important to consider that the two official decisions were not taken following the same timing and procedures. The first followed an emergency procedure in the context of the climax moment of the refugee crisis. The second was a long-term process that should have culminated with the reform of the Dublin System. It is also important to take into account that the relocation system adopted in 2015 was temporary, while the proposal of the Commission would have made this temporary mechanism permanent.

To establish a more comprehensive understanding of the role of emotions in the decision-making processes, I analyzed documents from the most relevant decision-makers and interest representatives active in this decision-making process. The key players selected include the Council of Ministers and the European Council, the EP, the Commission and NGOs. For better insight into the position of the European Council and Council of Ministers, I analyzed policy documents from Germany and the Visegrad Group (VG), including Hungary, the Czech Republic, Poland and Slovakia. Germany is the larger member state and in this specific case, it was one of the countries that supported the relocation of refugees. In sharp contrast, the VG created a unified front against obligatory relocation and was for this dubbed the coalition of the unwilling (Duszczyk, Podgórska, and Pszczółkowska 2020).

Regarding the EP, the debates analyzed include the perspectives of the different political groups, including populist groups. At the time of the refugee crisis, in the EP there were mainly two right-wing populist groups: the Europe of Nations and Freedom, composed of political parties such as the Dutch Party for Freedom and the French National Rally, a party led by Marine Le Pen; and the Europe of Freedom and Direct Democracy, composed mainly of UKIP MEPs, including Nigel Farage. As is well known, EP political groups are broad umbrellas including a diversity of factions. Right-wing groups, such as the European People's Party and the European Conservatives and Reformists, also have members from national political parties that are considered to be populist. On the other side of the political spectrum, the European United Left/Nordic Green Left political group includes political parties that are considered to be left-wing populists, such as the Spanish Podemos or the French La France Insoumise. For the present analysis of the so-called refugee crisis, the emphasis will be placed on right-wing populism, which emphasizes ethnic nationalism and tends to be exclusive. The documents analyzed from each one of these key players are presented in detail in Table 2; they were systematically analyzed using the Atlas.ti automatic coding function. The keywords used to illustrate how different actors employed different frames are available in the annexes.

Table 2. Documents analyzed.

	Content	Total pages /words
European Council and Council	Speeches and press releases by the president of the European Council and the presidency of the Council from 2015 to 2018 on the topic of migration/refugees	86 pp. 42,294 words
EP	20 EP debates from 2015 to 2019 (selected using keyword migration/refugee in title)	772 pp. 362,840 words
Germany	51 articles, speeches and statements from 2015 to 2019, from the website of the Federal Chancellor (selected using keyword Auswartiges)	108 pp. 43,066 words
VG	Press releases from 2015 to 2019 on the topic of migration	43 pp. 19,427 words
Amnesty International	Press releases from Amnesty International on the topic of migration/refugees from 2015 and 2016	130 pp. 58,922 words

3. Emotions and decision-making during the refugee crisis

The following results sections focus first on the crisis-framing contests, placing specific emphasis on their emotional basis and on the differences between policy diagnosis, policy solutions and emotion-based strategies. Attention is then specifically drawn to the differences between the two decision-making moments.

3.1. The framing contest: crisis-framing and emotion in the refugee crisis

The content analysis showed that there are two basic frames: one based on security and a second based on human dignity. The security frame can be related to the pull-factor frame, while what I define here as the human dignity frame would include both migration as a tragedy and the push factor frames (Abdou 2020). While key players adhere to different interpretations of these frames, they tend to share a common core (see Table 3). Each frame contributes to the institutionalization of certain emotion-based policy solutions.

Table 3. Crisis frames based on fear and compassion.

	Security frame (fear)	Human dignity frame (compassion)
Definition of problem	Security problem Focus on irregular arrivals and illegal migrants	Refugees welcomed Focus on asylum seekers fleeing war; focus on human rights and human dignity, on moral obligations and legal responsibilities
Policy solutions	Border control Focus on registration and finger-printing Focus on pull factors Focus on preventing arrivals through Outsourcing/externalization of the problem	Legal paths to asylum seekers Rescue missions in the Mediterranean Focus on the well-being of migrants Focus on relocation, resettlement and integration
Emotion-based strategies	Fearmongering Dehumanization Blaming/shaming	Humanization Blaming and shaming
Examples of key players	VG, right-wing populist groups	Germany, NGOs, Liberal, Left and Green EP groups
Divided key players	European Council and Council, EP, European People's Party

In the security frame, the main problem is the arrival of irregular migrants, which are considered as a threat. Policy-makers often appeal to fear within this policy frame and thus one could argue that this frame contributes to the institutionalization of fear (see box). The main fear for the Council representatives was fear of a Europe without external borders. It was also considered that the implementation of alternative policies would lead to fears among European citizens (Bert Koenders, president in office of the Council, February 2, 2016).

Fear as the main emotion in policy framing

When fear takes over, nations increasingly turn away from each other. What we need to do here and now is to regain a sense of security, which is a fundamental need for every human being, as important as the need for freedom. Donald Tusk, president of the European Council September 21, 2016.

And that brings me to my second key value: security. Not being able to control and manage migration is a great cause for concern. This has to do with fear of criminality and terrorism. Jeroen Dijsselbloem, president of the Eurogroup, Decembre 9, 2016.

I think the so-called refugees on our borders need to be repatriated to Muslim countries, as their values are clearly incompatible with our liberal western democracies. This will avoid the current clash of cultures that denigrates the achievements of Western civilization and flouts the protection of women, the gay community and vulnerable children, who are being attacked by Muslim gangs and migrants who deplore our way of life. Janice Atkinson, Europe of Nations and Freedom, April 12, 2016.

In the security frame, the main policy solutions proposed were the protection of external borders, and cooperation with third countries to stop the arrivals. These solutions are congruent with the action tendencies related to fear, such as neutralization of danger, closure and minimizing risks (Frijda 2007; Halperin and Pliskin 2015). According to this frame, citizens think that the role of public authorities is to protect the territory and borders, and to enforce the law. The relocation system and quotas are seen as pull factors resulting in more migrants wanting to reach European borders. The frame also sees relocation as voluntary on the grounds that no immigrants should be imposed on any member state. This corresponds to a specific understanding of solidarity, which well conveyed by a former president of the European Council:

The principle of ‘one for all, and all for one’ sounds nice, but we cannot force it upon others. After all, we are talking about solidarity here, which is something that cannot be enforced – either it's there, or it isn't. I remember when during a debate on relocation at a European Council summit, it was suggested that, in our conclusions, we include the term ‘the obligation of solidarity’. I protested, saying this was an obvious oxymoron. And I know what I am saying, I still feel like an expert in solidarity. (Donald Tusk, 14 November 2019)

On the human dignity view, the EU should help those seeking protection in Europe, fully respecting international obligations. The human dignity frame argued explicitly that public policy on this topic needed to be based on compassion (see box) and, in this sense, it contributed to the institutionalization of empathy.

Compassion as the main emotion in policy framing

Migration is a global phenomenon but one that requires measured, compassionate and effective European leadership. Roberta Metsola, European People's Party (EPP), June 7, 2016.

Beyond the issues of relocation and identification procedure, which must be ethical and compassionate, we must ensure access to education and put in place mechanisms to support and heal children who are very often experiencing immense suffering and trauma, often resulting in serious mental health issues. Julie Ward, Progressive Alliance of Socialists and Democrats (S&D), October 26, 2016.

We are witnessing the worst refugee crisis of our era, with millions of women, men and children struggling to survive amidst brutal wars, networks of people traffickers and governments who pursue selfish political interests instead of showing basic human compassion, Amnesty International (AI), June 15, 2015.

The policy solutions proposed within this human dignity frame are directed toward helping refugees and thus correspond to the action tendencies associated with empathy and compassion. The first proposed solution is ensuring safe and legal routes for refugees into Europe and the protection of refugees and migrants. Within this vision, a relocation system is seen as a step in the right direction. According to NGOs and some MEPs, the reform of the Dublin System should also include elements such as mutual recognition of asylum decisions and greater freedom of movement for recognized refugees. Here, solidarity means helping each other and doing things together, and that all member states adhere to the same standards. Solidarity is also understood as a responsibility to the community as a whole and, in this way, is opposed to national egotism (Angela Merkel, November 13, 2018).

Tables 4 and 5 illustrate how both frames were employed by showing how different key players used different keywords in their speeches. For example, while discussing the topic of migration, the VG placed much more emphasis on the word-markers related to the security frame, while NGOs focused on word-markers related to the human dignity frame.

Table 4. Word-markers used by different key-players in the decision-making process^a.

Keywords	VG	%	CoM	%	Germany	%	EP	%	NGOs	%	total
Border control	2 (0,01)	1,5	4 (0,009)	2,06	1 (0,002)	0,63	48 (0,01)	4,69	10 (0,02)	2,61	65 (0,01)
Control	23 (0,12)	17,42	29 (0,07)	14,95	11 (0,026)	6,33	280 (0,07)	27,37	28 (0,05)	7,31	371 (0,08)
Threat	10 (0,05)	7,58	6 (0,01)	3,09	11 (0,026)	6,33	29	2,83	25 (0,04)	6,53	187 (0,04)
External Borders	29 (0,15)	21,97	35 (0,08)	18,04	40 (0,09)	25,32	48 (0,01)	4,69	7 (0,01)	1,82	159 (0,03)
Islam	1 (0,005)	0,76	0	0	0	0,63	90 (0,024)	6,74	8 (0,01)	2,09	99 (0,02)
Pull factor	6 (0,03)	4,55	1 (0,002)	0,51	1 (0,002)	0,63	6 (0,002)	0,59	0	0	14 (0,003)
Security	31 (0,16)	23,48	18 (0,04)	9,28	26 (0,06)	16,46	72 (0,02)	7,04	41 (0,07)	10,71	188 (0,04)
Terrorism	12 (0,06)	9,09	6 (0,01)	3,09	8 (0,02)	5,06	29 (0,008)	2,83	8 (0,01)	2,09	63 (0,01)
Fear	0	0	9 (0,02)	4,64	3 (0,007)	1,90	45 (0,01)	4,40	54 (0,07)	14,10	111 (0,02)
Sub-total	114 (0,59)	86,36	108 (0,26)	55,67	101 (0,23)	63,92	647 (0,18)	63,25	181 (0,31)	47,26	1151 (0,24)
Human rights	0	0	2 (0,004)	1,03	1 (0,002)	0,63	66 (0,02)	6,45	101 (0,17)	26,37	170 (0,035)
Humanity	0	0	2 (0,004)	1,03	1 (0,002)	0,63	11 (0,003)	1,98	6 (0,01)	1,57	20 (0,004)
Legal response	0	0	0	0	0	0	2 (0,001)	0,20	0	0	2 (0,001)
Quotas	1 (0,005)	0,76	10 (0,02)	5,15	4 (0,009)	2,53	32 (0,009)	3,13	6 (0,01)	1,57	53 (0,01)
Relocation	6 (0,03)	4,55	21 (0,05)	10,82	2 (0,005)	1,27	69 (0,02)	6,74	29 (0,05)	7,57	127 (0,03)
Solidarity	11 (0,06)	8,33	45 (0,11)		46 (0,11)	29,11	131 (0,04)	12,81	11 (0,02)	2,87	244 (0,05)
Save lives	0	0	2 (0,004)	1,03	3 (0,007)	1,90	25 (0,007)	2,44	15 (0.03)	3,91	45 (0,01)
Shame	0	0	3 (0,007)	23,20	0	0	26 (0,007)	2,54	26 (0,04)	6,79	55 (0,01)
Compassion	0	0	1 (0,002)	0,51	0	0	14 (0,003)	1,39	8 (0,01)	2,09	23 (0,005)
Sub-total	18 (0.09)	13,64	86 (0,20)	44,33	57 (0,13)	36,08	376 (0,10)	36,75	202 (0,34)	52,74	739 (0,15)
Total	132		194		158		1023		383		1890
Total words	19427 (100)		42294 (100)		43066 (100)		362840 (100)		58922 (100)		483483 (100)

Note: Elaborated by the author. Speeches and press releases by VG, and by the Councils very often discussed several topics (and not only refugees/migration). Thus, I deleted all codes that did not correspond to the topic of the refugee crisis/migration.

Keywords: The keywords were selected because they represent key aspects of each one of the identified frames. The list is by no means exhaustive. It only aims at illustrating the use of different keywords that can be related to frames.

^aThis figure includes two types of percentages:

• the number between parenthesis indicates the % in relation to the total amount of words

• the % in a separate column indicates the percentage in relation to the total amount of keywords selected for this figure.

Table 5. Examples of keywords used by key-Players at the European Parliament.

	CoM**	EC	HR	EPP	S&D	ALDE	Green	GUE/NGL	ECR	EFDD	ENF	NI
Border	13 30,2%	45 33,3%	1 3,0%	22 22,4%	9 13,0	10 33,3%	6 13,6	1 11,1%	10 25,6%	5 35,7%	7 31,8%	0
Control	13 30,2%	10 7,4%	1 3,0%	30 30,6%	23 33,3%	8 26,7%	8 18,2	4 44,4%	12 30,8%	12 42,9%	7 31,8%	0
Pull factor	0	1 0,7%	0	0	1 1,4%	0	0	0	3 7,7%	0	0	0
Security	4 9,3%	11 8,1%	17 51,5%	12 12,2%	3 4,3%	2 6,7%	1 2,3%	0	1 2,6%	5 17,8%	0	0
Terrorism	0	0	2 6,1%	8 8,2%	7 10,1%	0	1 2,3%	0	1 2,6%	2 7,1%	2 9,0%	2 66,7%
Human Rights	0	5 3,7%	9 27,3%	3 3,1%	3 4,3%	4 13,3%	15 34,1%	2 22,2%	1 2,6%	0	6 27,2%	1 33,3%
Legal paths	0	3 2,2%	0	1 1,0%	0	0	1 2,3%	0	0	0	0	0
Relocation	9 20,9%	21 15,6%	0	8 8,2%	7 10,1%	2 6,7%	2 4,5%	2	8 20,5%	1 3,6%	0	0
Solidarity	4 9,3%	39 28,9%	3 9,1%	14 14,3%	16 23,2%	4 13,3%	10 22,7%	0 22,2%	3 7,7%	3 10,7%	0	0
Total	43 100%	135 100%	33 100%	98 100%	69 100%	30 100%	44 100%	9 100%	39 100%	28 100%	22 100%	3 100%

Note: CoM: Council of Ministers; EC: European Commission; HR: High Representative of the Union for Foreign Affairs and Security Policy; EPP: European People's Party; S&D: Progressive Alliance of Socialist and Democrats; ALDE: Alliance of Liberals and Democrats for Europe; Green: European Green; GUE/NGL: Confederal group of the European United Left/Nordic Green Left; ECR: European Conservatives and Reformists; EFDD: Europe of Freedom and Direct Democracy; ENF: Europe of Nations and Freedom; NI: Not attached to any political group.

*Table 5 covers 16 EP debates instead of 20 (the coding per political groups was not available for 4 out of the 20 debates). It is important to consider that after 2012 EP Debates are not translated into English. Most MEPs speak in their own language while representatives of EU institutions usually express themselves in English. **Table 5 only includes analysis of speeches given by representatives of the CoM in the European Parliament.

Table 5 shows more specifically how different political groups at the EP used keywords and illustrates how key players engaged in different types of frame. The security frame was shared by conservatives and right-wing populist groups. These groups used keywords such as ‘borders’ and ‘control’, while liberals or left-wing parties such as the Greens frequently used keywords such as ‘quotas’ and ‘human rights’.

3.2. Emotion-based strategies

In line with the different frames, the key players used very different emotion-framing strategies. As expected, right-wing populist groups engaged in a fearmongering strategy to promote the security frame, while NGOs often used a humanization strategy to promote the human dignity frame. Naming and shaming were used to promote both policy frames. Commission officers and mainstream policy-makers tended to engage only occasionally in emotion-based strategies, and their lack of emphasis on emotions may have contributed to dehumanization.

Only right-wing populist groups engaged in the dynamics of fearmongering. They pictured refugees as terrorists or as profiteers from social welfare regimes and, in this way, they contributed to reducing levels of empathy. Blaming the undeserving other, which is widely used by populist groups (Abts and Baute 2021), also contributed to dehumanization. This fearmongering strategy supported more extreme fear-based policy solutions, such as reducing social benefits for refugees and even shoot-to-kill orders on European borders (Frank Engel European People's Party, February 2, 2016).

Examples of fearmongering

But there is a real and genuine threat. When ISIS say they want to flood our continent with half a million Islamic extremists, they mean it, and there is nothing in this document that will stop those people from coming. Nigel Farage, Europe of Freedom and Direct Democracy (EFDD), April 29, 2015.

You cannot solve the mess created by the open borders with more European Union. That's why you should close the borders and des-islamize. Vicky Maeijer, Non attached (NI) April 12, 2016, translated from Dutch.

I think the so-called refugees on our borders need to be repatriated to Muslim countries, as their values are clearly incompatible with our liberal western democracies. This will avoid the current clash of cultures that denigrates the achievements of Western civilization and flouts the protection of women, the gay community and vulnerable children, who are being attacked by Muslim gangs and migrants who deplore our way of life. Janice Atkinson, Europe of Nations and Freedom (ENF), April 12, 2016.

The security frame was also promoted through naming and shaming. One of the main targets of the blame game was Germany and its ‘refuges welcome’ approach (Willkommenspolitik). Angela Merkel was seen as inflicting chaos on her country and as forcing the crisis onto Germany's neighbors. NGOs, MEPs and European officials in favor of protective measures were also targeted (see box).

For the human dignity camp, the main target of blame was the Council of Ministers and the European Council, whenever they failed to adopt decisions. Within the Council, the attitude of the VG countries, especially Hungary, and its treatment given to asylum seekers was also the object of criticism. As summarized by the president of the Council, Donald Tusk: ‘The advocates of tight external borders were accused of xenophobia and a lack of solidarity, while the proponents of Willkommenspolitik of letting in terrorists and unacceptable submission’ (November 14, 2019).

Examples of naming and shaming

Naming and shaming following the fear-based frame

Not content with inflicting the migrant chaos on its own – in some cases, beleaguered – citizens, Germany is using bully-boy tactics to force this crisis on to its neighbors.

German domination might well mean that when Ms. Merkel says jump, we all ask ‘How high?’, but I – on behalf of the United Kingdom and Britain – do not agree. Diane James, EFDD, September 14, 2015.

We must not allow our compassion to imperil our security. Nigel Farage, EFDD, May 20, 2015.

The EU, and many people in this chamber, need to realize that their naïve, soft-hearted approach to migration has not only failed, but has actually cost lives and funded terrorism and crime, Jane Collins, EFDD, February 17, 2017.

Naming and shaming following the compassion-based frame

That is why the Council conclusions, I would say to Mr. Tusk, are nothing less than shameful because they do not focus on saving lives and opening this continent to people seeking refuge. Malin Björk, European United Left/Nordic Green Left (GUE/NLG), April 29, 2015.

I think what was decided at the Council was a disgrace. It was unfortunately a waste of time. Even though it was good that you sat together, nothing came out of it, Ska Keller, Greens, April 29, 2015.

Therefore, let us turn our fire to where it should be: on the Council. Claude Moraes, S&D, April 20, 2015.

European Union (EU) governments should hang their heads in shame at the ongoing reluctance of many to ensure a collective and concerted search and rescue operation in the Mediterranean, said Amnesty International, Amnesty International, March, 12, 2015.

Key players also engaged in processes of humanization/dehumanization. Supporters of the securitization camp often blamed the victims, using vocabulary such as ‘irregular migrants’. They also used euphemistic labeling and a distortion of consequences, as for example, not directly talking about deaths in the Mediterranean. Water metaphors such as wave or flow were also used in debates regarding migration and the so-called refugee crisis (see Table 6). These metaphors convey a sense of danger and urgency of action since it is very difficult to contain, control or channel a liquid (Kainz 2016). The metaphor of a flow suggests that policy-makers were controlling uncontrolled flows rather than stooping human beings from moving to a safe place.

Table 6. Examples of keywords indicating dehumanization and humanization^a.

	Germ.	%	VG	%	CoM	%	EP	%	NGO	%	Total
Death	2 (0,005)	2,53	1 (0,005)	1,81	1 (0,002)	1,54	25 (0,007)	6,23	43 (0,07)	16,35	72
Children	6 (0,01)	7,59	1 (0,005)	1,81	3 (0,007)	4,05	152 (0,04)	37,90	91 (0,15)	34,60	253
Family	4 (0,009)	5,06	1 (0,005)	1,81	2 (0,005)	2,70	20 (0,005)	4,99	45 (0,08)	17,11	72
Victim	4 (0,009)	5,06	1 (0,005)	1,81	1 (0,002)	1,54	21 (0,006)	5,24	10 (0,002)	3,80	37
Tragic	3 (0,007)	3,80	1 (0,005)	1,81	6 (0,01)	8,11	20 (0,005)	4,99	10 (0,002)	3,80	40
Life	7 (0,02)	8,86	2 (0,01)	3,64	4 (0.009)	5,4	60 (0.02)	14,96	34 (0,06)	12,93	107
Subtotal	26 (0,06)	32,91	7 (0,04)	12,72	17 (0,04)	22,97	298 (0.08)	74,31	233 (0,40)	88,59	581 (0,12)
Irregular migrant/migration	9 (0,02)	11,39	9 (0,05)	16,36	12 (0,03)	16,22	23 (0,006)	5,74	5 (0,008)	1,90	58
Mass migration	0	0	2 (0,01)	3,64	0	0	4 (0.001)	1,00	0	0	6
Migration/Migrant pressure	0	0	2 (0,01)	3,64	0	0	4 (0,001)	1,00	0	0	6
Illegal migrant/migration	35 (0,08)	44,30	14 (0,07)	25,45	16 (0,04)	21,62	5 (0,001)	1,25	0	0	70
Flow	9 (0,02)	11,39	21 (0,1)	38,18	19 (0,04)	25,68	60 (0,02)	14,96	20 (0,03)	7,60	129
Wave	0	0	0	0	10 (0,02)	13,51	7 (0,002)	1,75	5 (0,008)	1,90	22
Subtotal	53 (0,12)	67,09	48 (0,25)	87,27	57 (0,13)	77,03	103 (0,03)	25,69	30 (0,05)	11,40	291 (0,060)
Total	79	100	55	100	74	100	401	100	263	100	872 (0,18)
Total words		43066 (100)	19427 (100)		42294 (100)		362840 (100)		58922 (100)		483483 (100)

Keywords: Selecting keywords to grasp a de/humanization process is challenging. The keywords selected are by no means exhaustive. They only aim at giving a preliminary illustration. Further research would be needed to show a dehumanization process more systematically.

^aThis Table includes two types of percentages:

• the number between parenthesis indicates the % in relation to the total amount of words. It is important to take into account that the EP debates are written in several languages while all other texts are in English. Thus, the % for the EP is lower than it would be if only the number of English words had been considered.

• the % in a separate column indicates the percentage in relation to the total amount of keywords selected for this figure.

In sharp contrast, key players emphasizing human dignity used terms aimed at humanizing the debate on asylum seekers. There was thus a contest between a strategy of humanization and a strategy of dehumanization. For example, as is shown in Table 6, NGOs and some MEPs often referred to children and families to create a sense of empathy and compassion. Table 6 also illustrates the clear difference between NGOs and some members of the EP, which tended to employ humanizing language, and official authorities, especially the VG and the Council of Ministers.

Interestingly, the analysis reveals that NGOs were also using dehumanizing metaphors such as ‘flows’. In sharp contrast, supporters of security politics never use humanizing language when they refer to migrants or refugees. Also, while the focus on women and children can help to trigger emotions and a sense of urgency, it may also have contributed to developing a restrictive – even dehumanizing – conception of compassion. Compassion understood as pity pictures victims as pure and undeserving of their suffering, while a higher conception of compassion also sees the suffering whether or not this is seen as deserved (Hoggett 2006). A restrictive conception of compassion can lead to rapid ‘compassion fatigue’ and offers little defense against populist dehumanizing arguments that place the blame on victims.

The infrequent use of emotions by policy officers and decision-makers may also have contributed indirectly to the process of dehumanization. Dehumanization was also promoted more directly through a strategy of toning-down emotions. Representatives of the councils affirmed that they used a strategy of toning-down emotions in the debates (see box). To calm down emotions (in a context where emotions are perceived to be opposed to rationality) may seem a sensible strategy. However, toning-down emotions could contribute to policy solutions that would not have been accepted had emotions such as empathy and third-party anger prevailed.

Toning emotions down

But we have succeeded in persuading most actors to calm their emotions and focus on more pragmatic initiatives that protect our borders, on the fight with people smugglers, as well as cooperation with Libya and other African countries, and of course the deal with Turkey. Donal Tusk, president of the European Council, January 15, 2019.

Above all, we must put an end to the destructive emotions surrounding the issue of relocation, as they continue to fuel populism and divide Europe. Donald Tusk, President of the European Council, February 13, 2018.

3.3. Framing and emotion-based strategies at work: emotions in the two decision-making processes

While the triggering of the emergency system led to the adoption of a relocation system in 2015, this relocation principle was not adopted in the reform of the Dublin System, initiated by the Commission in 2016, due to the absence of a common position in the Council. During this period, there was not a significant change in the composition of the councils, but changes in the political context, internal politics, the perception of failure of the quota system and the role of the presidencies of the EU can be considered potentially relevant explanatory factors.

The first relevant finding is that most key players tended not to change their crisis frames or emotion-based strategies. The only actor that significantly changed its framing was the European Council and Council. The evolution of the position of EU leaders is visible in the European Council statements (Duszczyk, Podgórska, and Pszczółkowska 2020). While in 2015 European leaders seemed to be concerned with saving lives, the main focus after 2016 was about exporting the problem. Table 7 illustrates the change of position of the councils, as reflected in the usage of keywords in documents by the presidencies of the councils. In 2015, there was much more emphasis on solidarity than in the subsequent years. The position displayed by the councils evolved from a position that was open to a relocation mechanism to a position that was contrary to one. Even the president of the European Council, Donald Tusk sent a message to the leaders before the summit in December 2017 suggesting that mandatory quotas should not be used (Duszczyk, Podgórska, and Pszczółkowska 2020). The position of the councils thus was increasingly aligned with the position of right-wing populist groups. Tusk even occasionally engaged in the blaming-the-victim emotion-based strategy (dehumanization). An illustration of this trend is his proposal (in the wake of the Paris terrorist attacks) to detain asylum seekers for 18 months for screening to see if they were potential terrorists; he also described Merkel's policies as ‘dangerous’ (Traynor 2015).

Table 7. Use of keywords at the European Council and the Council of Ministers between 2015 and 2018**.

	2015	%	2016	%	2017	%	2018	%	Total	%
Border control	2	2,78	1	3,57	1	5,56	0	0	4	2,96
Security	4	5,56	7	25	1	5,56	6	35,29	18	13,33
External borders	16	22,22	7	25	6	33,33	4	23,53	33	24,44
Fear	1	1,39	7	25	1	5,56	0	0	9	6,67
Human rights	2	2,78	0	0	0	0	0	0	2	1,48
Quotas	4	5,56	0	0	5	27,78	1	5,88	10	7,4
Relocation	10	13,89	1	3,57	3	16,67	4	23,53	18	13,33
Solidarity	32	44,44	5	17,86	1	5,56	2	11,76	40	29,63
Compassion	1	1,39	0	0	0	0	0	0	1	0,74
Total	72	100	28	100	18	100	17	100	135	100

*The great majority of speeches correspond to Donald Tusk, president of the European Council during this period.

In contrast to the councils, many member states did not change their position, as was the case with the VG, which remained very consistent in its opposition to the relocation system. The evolution of Germany's position is, however, interesting. During 2015, the ‘we can do it’ policy was promoted and defended with enthusiasm, and Germany clearly took a leadership role. As the German Minister of Economic Affairs put it, referring to the solidarity that people were demonstrating toward people in need, ‘Germany is showing a side of itself of which it can rightly be proud’ (website of the German Chancellor, 7 September 2015). While Germany always supported a solution based on solidarity, its position became more nuanced after 2016, and included the possibility of the so-called flexible solidarity. The enthusiastic ‘we can do it’ of 2015 changed into the more nuanced ‘I said we could do it but I did not say that it was going to be easy’ of 2016 (website of the German Chancellor, June 28, 2016).

Emotions can help – in interaction with other factors – to explain why the European Council and Council of Ministers changed their position and why Germany nuanced its position. The emotional environment in different moments helps us to understand how the compassion-based frame prevailed in 2015, while the fear-based frame was more prevalent in subsequent years.

A compassion-based naming and shaming and humanization strategy seems to have contributed to policy change within a divided Council at the climax of the crisis. The adoption of the relocation system could indeed be interpreted as a first preliminary step toward the institutionalization of empathy in the EU asylum system. During the crisis in 2015, key players in the human dignity camp had more opportunities to create a sense of urgency and empathy, and thus place pressure on the Council. At that moment, there were more debates on the EP (see Table 8) on this topic, and arguments and decisions often referred to the events that helped to bring a wave of empathy and compassion into public opinion and a sense of urgency into the political discussions. The symbols of this time were the deadly shipwrecks off the coast of Italian island of Lampedusa, the discovery of 71 bodies of migrants in a lorry in Austria (August) and the tragic death of Aidan Kurdi (September). The coverage by the media of these events has been described as ‘ecstatic humanitarianism’ since articles included frequently claims to European benevolence, and references to empathy and solidarity were much more often attributed to citizens than fear (Chouliaraki and Zaborowski 2017). The above-mentioned events were present in the arguments of key-players:

It is our European decency that lies lifeless on the Turkish beach together with Aylan Kurdi, the three-year-old toddler in his red T-shirt and small sneakers. He lost his life but he brought passion – and compassion – back to the European debate on migration and therefore we will remember him for a very long time. (Celilia Wilstrom, Alliance of Liberals and Democrats for Europe, September 9, 2015)

In sharp contrast, the context that surrounded the decision-making process starting in 2016 was very different. Media attention turned to events such as the 2015 Paris terrorist attacks (November) and the sexual aggression in Cologne on New Year's Eve. In November 2015 (in the refugee news narratives), refugees swiftly had turned into malevolent actors and fear was more frequently attributed to citizens than empathy (Chouliaraki and Zaborowski 2017). While the impact of previous humanitarian events was considered to be immediate but transitory, these latest events were considered to have a more lasting impact on the public debate. The media coverage of these events contributed to a climate of fear. As previously announced, in a context of fear, empathy toward out-groups is diminished. In the new emotional environment, the conditions were thus much more favorable for the cultivation of fear and for the success of fearmongering strategies and dehumanization. In the context of Brexit, a number of EU leaders such as Angela Merkel, but also Jean-Claude Juncker, did not want to exacerbate east–west divisions (Duszczyk, Podgórska, and Pszczółkowska 2020).

Table 8. Evolution of usage of keywords Council and urgency at the European Parliament.

2016							2017			2019	Total
	Debate 12–04	Debate 11–05	Debate 07–06	Debate 04–07	Debate 14–09	Debate 26–10	Debate 18–01	Debate 17–02	Debate 12–09	Debate 15–01	EP
Conseil	1	1	1	0	10	0	1	0	0	0	97
Council	1	1	8	1	5	1	2	1	1	14	196
Total	2 (0,01)	2 (0,01)	9 (0,03)	1 (0,01)	15 (0,11)	1 (0,01)	3 (0,02)	1 (0,01)	1 (0,01)	14 (0,10)	293
Urgent	6 (0,03)	3 (0,01)	3 (0,01)	0	3 (0,02)	4 (0,04)	28** (0,20)	3 (0,03)	3 (0,01)	1 (0,01)	170
Total	20166	17233	22506	11116	13386	9139	13846	9610	16927	13852	362840
	2014	2015						2016
	Debate 25–11	Debate 29–04	Debate 20–05	Debate 09–09	Debate 14–09	Debate 06–10	Debate 16–12	Debate 02–02	Debate 08–03	Debate 08–03
Conseil	1	20	5	7	22	9	10	8	0	1
Council	13	23	39	3	16	6	5	42	12	2
Totals	14(0,08)	43(0,15)	44 (0,16)	10(0,06)	38(0,16)	15(0,20)	15(0,11)	50(0,11)	12(0,06)	3 (0,02)
Urgent*	8 (0,05)	14(0,05)	19 (0,07)	12(0,08)	21(0,10)	8 (0,11)	3 (0,02)	15(0,03)	8 (0,04)	8 (0,05)
Total	16692	28187	27962	15752	23568	7604	13431	47090	18587	16186

Note: Elaborated by the author. The keyword Council/Conseil was chosen because this institution was the main target of blaming and shaming by the We can do it supporters. The keyword giving a sense of urgency (urgent, urgence and emergency) was selected because the use of emotions and a blaming and shaming strategy is often used by NGOs to create a sense of urgency for the adoption of policy changes (Sanchez Salgado 2018). The gray boxes indicate the debates where the keywords Council/Conseil where the most used as well as the keywords related to urgency.

• the number between parenthesis indicates the % in relation to the total amount of words.

**The great number of occurrences is explained by the title of the Debate: Emergency aid for refugees and migrants facing severe weather conditions in European Camps.

EU leaders therefore decided to ‘solve’ the crisis outside of the formal EU procedures by signing the so-called EU–Turkey deal; this represented a significant step toward the institutionalization of fear in this policy area. With this decision, the policy solution adopted in 2015 to relocate refugees (based on empathy) was replaced by the decision to close borders and send refugees back to Turkey (based on fear). The illegal character of this agreement, considered by NGOs as an historic blow to human rights, can be understood as reflecting a lack of empathy/compassion. To give an example of alternative scenario, had the combination of empathy with anger prevailed, one would expect that it would have been easier to impose sanctions on the non-compliant member states, giving the quota system more chance of success.

The decisions regarding the reform of the Dublin System were taken in 2017 and 2018 at a moment when decision-makers did not perceive a situation of urgency. The strategy of toning-down emotions can be easily applied within this new context. Only two EP debates discussed the reform of the Dublin System in 2016; one took place a posteriori when the proposal had already been withdrawn (Table 8). The lack of emotional intensity led to different policy solutions in different EU institutions. While the EP adopted a policy solution related to empathy in its common position adopted in November 2017, the European Council and Council of Ministers focused on decisions such as closing borders and the externalization of refugees (fear-related). This could be explained by the fact that these different institutions respond to different models of emotion-based institutionalization (Crawford 2014). The EP seems to fit the model of an amalgamated policy community that is more prone to the institutionalization of empathy.

4. Conclusion

This article showed that emotions can help us understand why certain policy frames and policy solutions can prevail in a certain given context. In the crisis situation analyzed, different key players were competing for the institutionalization of different emotions. These emotions served as affective foundations of the prevailing policy frames. The article first showed that the proponents of security policies were using a frame based on fear, while the proponents of the relocation system placed the emphasis on compassion. The article also showed how key players engaged in emotion-based strategies, such as fearmongering, naming and shaming, and de/humanization.

The article also showed how emotions have played a role in the choice of policy solutions. Policy solutions can even be interpreted as the institutionalization of specific emotions. This article showed that the extent to which emotion-based frames and strategies prevailed depended on the content and intensity of the emotional context. While in 2015 a wave of empathy contributed to the success of the humanizing and naming and shaming strategy, and to the institutionalization of compassion with the adoption of the relocation system, at the beginning of 2016 a climate of terror seems to have contributed to the success of fearmongering and to the institutionalization of fear with the EU–Turkey agreement. The decision-making process regarding the reform of the Dublin System took place in a climate of less emotional intensity and, in this context, it seems understandable that strategies such as toning-down emotions and a non-decision as policy outcome prevailed.

This article's main contribution is thus to bring attention to the relevance of the emotional environment in which specific decisions are taken, and the interaction of this with emotion-based frames and strategies. The contribution of emotions can only be understood in combination with other relevant contextual factors, such as external events, and with institutional factors such as the role of the presidencies and the decision-making procedures. Further research could focus on the extent to which the emotion environment affects policy-making, and to what extent incidental fear (such as Covid-19 fear) contributes to the adoption of policy solutions with little empathy toward out-groups.

Acknowledgements

All my thanks to the guest editors of this special issue and the anonymous reviewers for their constructive and useful feedback. I would also like to thank Abhijit Mahale for his assistance with the data collection.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Rosa M. Sanchez Salgado

Rosa M. Sanchez Salgado is Assistant Professor of European Politics in the Department of Political Science, University of Amsterdam. She has for many years researched how the European Union has contributed to the shaping of civil society organizations, including the effects of EU funding. More recently, she has investigated the role of emotions in European politics. Her work has appeared, among others in West European Politics, Journal of Common Market Studies, Journal of European Integration, Global Society and Politics & Governance.

Notes

1 The EU-Turkey Deal is a statement of cooperation between the Turkish government and European member states aiming at controlling the number of refugees and migrants crossing from Turkey to the Greek Islands.

2 This comparison was written by Judith Sunderland, deputy Director of Human Rights Watch in a tweet (Chadwick and Montalto Monella 2020).

3 In this article, compassion is considered as a secondary emotion related to empathy. The biological basis of empathy is much more established than the biological basis of compassion. While these concepts are not identical, they are sufficiently related to use them as synonymous for the purposes of this article.