Polarization has the potential to weaken the governability, social cohesion, and democratic processes within societies worldwide (McCoy et al., 2018). Fortunately, not all polarization does. Political polarization, the polarization on political issues and the ideological distance between political parties, is largely considered to be beneficial for democracies (Campbell, 2016). To identify when polarization endangers or weakens democratic systems, we need to look beyond political polarization and conceptualize polarization as both a political and a relational process rooted in affect and identity (Iyengar et al., 2019). When polarization extends beyond the political realm and the animosity enters other aspects of social relations, causing citizens to increasingly dislike and distrust those with other political preferences, affective polarization is evident (Iyengar et al., 2019; McCoy et al., 2018). Affective polarization intertwines identity, interests, and attitudes with political viewpoints (McCoy & Somer, 2019).

Nevertheless, not all affective polarization endangers democratic societies; like Mouffe (2016), we differentiate between “agonistic” (struggle among adversaries) and “antagonistic” (struggle among enemies) affective polarization. The difference lies in how the “other” is referenced. Adversaries are defined in political terms as opponents within the democratic system; enemies are defined in moral terms as foes threating our existence (Mouffe, 2008, 2016). The latter may enable antagonistic polarization, a series of processes by which the normal multiplicity of variances in societies are boiled down to a single dimension and make “people increasingly perceive and describe politics and society in terms of ‘Us’ versus ‘Them’” (McCoy et al., 2018, p. 18).

In our view, leadership plays a crucial role in the development of these processes. Therefore, we argue, based on identity leadership (Haslam et al., 2022) and the social identity approach to leadership (Van Knippenberg, 2023), that leaders use social identity as a political tool to convince followers they are part of the group they seek to represent and to create opposition between us and them. Reicher and colleagues' Five Step Model (2008) explains that social identity is crucial in understanding why, how, and when groups display hostility towards each other, as it plays a key role in forming and spreading outgroup hatred. The model shows how for the benefit of their in-group, people are prepared to commit inhumane acts (Reicher et al., 2008).

According to the Five Step Model, the development of collective hate starts with (1) the identification with an in-group. As SIT shows, this entails the positive self-definition of people in terms of group-membership (Tajfel & Turner, 1979). A shared social identification, an individual’s sense of internalized group-membership, and an associated feeling of “we-ness” makes people feel they belong to an in-group (Tajfel, 1972). Next, (2) is the exclusion of targets outside the in-group. Creating an in-group also means placing others outside the group. SCT explains this as defining category boundaries, group-distinctiveness, and formulating who does and does not belong to the in-group (Hornsey, 2008). Collective hatred comes into being when (3) the out-group is framed as a threat endangering the existence of the in-group. Importantly, out-group hatred is actively constructed in each situation and can be seen as a way to simplify and organize the differences between us and them (Reicher et al., 2008). Furthermore, to enhance the moral self-esteem of groups (4) the differences between the in- and out-group are portrayed in terms of us as (inherently) good and them as bad (Van Stekelenburg, 2014). Lastly, the development of collective hate is complete when (5) groups celebrate the defeat of the out-group. This means that immoral collective acts can be celebrated, because the good in-group needs to be defended against the dangerous out-group (Reicher et al., 2008).

Using insights from the Five Step Model and the theory of identity leadership in the context of affective polarization, we theorize that political leaders use identity-claims to convince people they belong to the same in-group and that they are the best politicians to represent that group, and to portray others as outsiders undermining our identities. They do so by using pronouns such as we and us and phrases as “us Dutch people”, “our country”, and “we Social Democrats” (Mols et al., 2023). In order to be accepted as the identity leaders of specific social groups, political leaders need to be (1) “identity prototypes”, who need to be seen as one of us, (2) “identity champions” who are doing it for us, (3) “identity entrepreneurs”, who rhetorically refine and shape who we are, and (4) “identity impresarios”, who make us matter by performatively and creatively organizing local events to exemplify who they are to the world (Haslam et al., 2020).

The present study contributes to the research on political rhetoric and the discursive construction of us and others by proposing a stepped approach to analyze how Dutch PLs from across the political spectrum deploy discursive strategies which might enable antagonistic polarization. The research combines a critical discursive approach rooted in the sociocognitive approach to context (Van Dijk, 2009) with identity leadership derived from of the SIT and SCT traditions. Within the sociocognitive approach to context social identities are approached as something one “gradually adopts, grows into, etc., but also something one must ‘learn’, that is, social knowledge” (Van Dijk, 2009, p. 72). Within this view, social identities are understood as gradually constructed by the interaction and discourse surrounding social actors (Van Dijk, 2009). By combining the approaches, we clarify how diverse PLs construe not only their identity leadership (identity claims) but also shed light on the ways (discursive strategies) they define group distinction and enable antagonistic polarization.

We subject the speeches of ten PLs at eight different moments to three analytical steps based on their acts of identity entrepreneurship deriving out of identity leadership. We do so to examine the political rhetoric of politicians; the way PLs refine and shape the shared social identity of their followers (Haslam et al., 2022). Acts of identity entrepreneurship comprise the ways in which PLs describe who we are and by whom we are threatened, depersonalize others, and present themselves as collaborative partners in changing social reality (Choi et al., 2022; Mols et al., 2023; Reicher et al., 2005). We examine the following steps in the process: (1) the use of identity-claims by PLs making themselves indispensable category prototypes for the group; (2) how PLs draw category boundaries between us and them; and (3) the discursive ways in which PLs communicate the identity-claims that their political points appear to be those of “the people” (Haslam et al., 2020; Mols et al., 2023; Van Knippenberg, 2023).

Once the identity-claims of the PLs are identified in Step 1, Step 2 tackles the question how to differentiate the rhetoric strategies upon which they draw category boundaries between us and them. Which words are used to indicate the other as adversary or as enemy? Here we enter uncharted territory. Literature defining antagonistic rhetoric without reference to populism or radical-right is elusive. As mentioned, we examine how political leaders across the political spectrum produce antagonistic rhetoric as part of the political game. In general, antagonistic rhetoric is used to emphasize irreconcilable differences among groups of people (Braun, 2002). Antagonistic rhetoric can be found in the form of personal and political insults, sarcasm, and rhetorical figures such as hyperbole, labelling, and oxymorons (Gonawela et al., 2018). In this study, we define identity-claims as antagonistic when references to irreconcilable differences among groups of people are intended to maintain or replace the existing social configuration in society by ideas on mutually exclusive identity camps (Braun, 2002). In other words, while agonistic identity-claims leave room for discussion and acceptance of the different views of the other, antagonistic identity-claims undermine this democratic principle by portraying others as those who put the in-group’s identity into question, thereby threatening the existence of the in-group (Mouffe, 2016).

In step 3, to analyze how PLs strategically communicate their identity-claims, we turn to a critical discursive approach. This approach helps in understanding the world through language and the daily use and reproduction of ideologies (Van Dijk, 2006; Wetherell, 1998). Ideologies can be seen as the organizing principle to divide people and society in antagonistic terms (Van Dijk, 1998). As previous research on populist and radical-right discourse of politicians shows, there are multiple discursive strategies politicians deploy to construct inside (elites, political opponents) and outside (ethnic or religious minorities) enemies (Burns & Stevenson, 2013; Sakki & Pettersson, 2016; Verkuyten, 2013). In doing so, they often portray their enemies as being different, deviant, and posing a threat for the in-group (Van Dijk, 1993). Moreover, discursive strategies such as “reversal of victim-perpetrator positions”, scapegoating”, “conspiracy theories”, and “collective memory and nostalgia” enable these leaders to claim they represent the “ordinary people” while constructing a positive image of themselves acting on behalf of the people (Sakki & Pettersson, 2016). In this study, we analyze how PLs across the political spectrum deploy diverse discursive strategies to mark the difference between us and them.

We collected the political texts of ten PLs of right-, center, and left-wing parties of both coalition and opposition from eight political events prior to and following the March 2021 Parliamentary election. These ten parties all played a significant role in the electoral campaign by participating in (national) debates and hold more than one seat in the Second Chamber, both pre and post-election. See Table 1.

The data analyzed come from a variety of sources, including written statements on official party websites, speeches during debates in the Second Chamber, and spoken messages of PLs on social media platforms (Table 2). For each political event we analyzed similar texts per PL. All spoken texts were transcribed in Dutch and the analysis was conducted by the first author on the original Dutch-language speeches and statements. Afterwards, we translated the extracts highlighted in this study from Dutch to English. The length of the extracts varied for each political moment and per PL. The campaign video (Event 1) is the shortest entry, consisting of a couple of sentences, and the longest entry examined was a speech for the own followers (Event 7).

These political events were chosen because they are illustrative of Dutch politics in 2021 and 2022. Our aim was to include moments before and after the March 2021 election. The first three events were obvious: our starting point was (1) analyze the texts of the campaign videos of the parties for the election. Shortly after the start of the campaign, there was (2) the Dutch political moment most discussed in domestic and foreign press in 2021: the resignation of the government due to the Dutch childcare benefits scandal. The discriminatory application of algorithms caused thousands of parents to be falsely accused of fraud by the Dutch tax authorities (European Parliament, 2022); this resulted in the fall of the Dutch Parliament. We included the reactions of the PLs as political texts. Another logical event to include (3) was the live reaction of the PLs on the election results.

The political events following the 2021 election were chosen from a selection of 20 key political moments in the Netherlands, as identified by nearly 25.000 people in the EenVandaag Opiniepanel. This is a political barometer in which a selection of 80.000 people across the Netherlands give their opinion on sociopolitical issues (EenVandaag, n.d.). The selection of the five events included in this study was based on the amount of media coverage and determined by the availability of texts of all PLs or their representatives. Based on these two steps we included, (4) the event called “position Omtzigt, job elsewhere” which comprises the unintentional disclosure of a sensitive political negotiation document during the coalition-negotiation. Chief negotiator Ollongren was photographed by the press while she (Ollongren) was carrying a document with the following words clearly visible: position Omtzigt, job elsewhere. Though Prime Minister Rutte denied having stated a desire to replace politician Omtzigt, later it became clear that he had done so (Valk, 2021). As a result and in accordance with the Dutch political process, the opposition brought a no-confidence motion against Rutte and accused him of lying in public (EenVandaag, n.d.).

Next, we included political event (5) in which Democrats 1966 (center-left) leader Kaag lifts the blockade on coalition-formation with a Christian party. Kaag instituted the blockade during the coalition-negotiations because of fundamental disagreements on medical-ethical issues such as abortion and euthanasia and stated that she wanted to govern “over left” with left-wing parties such as GreenLeft and the Labor Party (Meijer, 2021). Because of this decision, a majority could not be reached. This blockade was in place for several months; eventually she lifted the blockade in September 2021 and the coalition could be formed. A logical next political event was (6) the presentation of the coalition-agreement by the coalition parties and reaction of the opposition parties.

Because this study focusses on social identity, we also added (7) a text of a political event such as a congress or a video made in 2021 for the own party followers. Finally, we included (8) the most impactful political event of 2022, according to a poll of the EenVandaag Opiniepanel, “Cabinet leaves Parliament in protest”. Nearly 30.000 people participated in the 2022 poll. This event entails the first time in Dutch history the entire Cabinet walked out of the debate in the House of Representatives. The walk-out was triggered by TB’s (Forum for Democracy, radical-right) accusations that SK (Democrats 1966, center-left) had studied at a spy (Oxford) college (Den Dool & de Witt Wijnen, 2022).

Atlas.ti was used to structure the analysis of the 80 texts. All texts were labeled with the initials of the PLs and numbered according to the event (see Table 2). We labeled the quotes accordingly, using the abbreviation of the PLs, the political party, and political position. For example, “LM6 (Socialist Party, left)” refers to the text of Lilian Marijnissen during event six (the coalition agreement). There are two important exceptions: during event six there was no statement from TB (Forum for Democracy, radical-right) and during the scope of data collection, the PLs of three political parties (Labor Party/left, Democrats 1966/center-left, Christian Democratic Appeal/center-right) changed between event seven and eight. Therefore, the texts of the new PL or a representative of the party were analyzed using the same abbreviation.

As mentioned, all texts were subjected to three analytical steps inspired by the concept of identity entrepreneurship. In the first step, we specified the identity-claims, that is how PLs make themselves category prototypes by using pronouns such as we, referring to collective self-understanding of people (Mols et al., 2023). The texts were searched for personal pronouns used by PLs to address social identities. This step was executed by thoroughly reading and searching the texts for pronouns and references to social identities using terms such as I in combination with we, you, our, us, they, their, and them and exploring whether these words were followed by rhetoric to address social identities. In other words, we collected personal pronouns followed by phrases and text passages referring to the collective self-understanding. In addition, the texts were searched for words and phrases which indicate that PLs claim to represent a specific group as in the following examples: (1) “.. and – I say as a young parent..” (SK8, Democrats 1966, center-left), (2) “… our country is only being destroyed further..”(GW6, Party for Freedom, radical-right), and (3) “…I understand what is going on in the Netherlands, in the villages and in the cities and I will stand up for you..” (LP1, Labor Party, left).

In the second step, we zoomed in on how PLs draw category boundaries by redefining followers’ shared idea of who does or does not belong to us (Haslam et al., 2020). We did so by coding the identity-claims following Mouffe’s distinction between agonistic rhetoric in which others are considered opponents, and antagonistic rhetoric in which others are considered enemies. The latter happens when the others’ right to defend their ideas is questioned and when the sentiment that those others threaten our collective self-identity arises (Mouffe, 2008). Additionally, identity-claims are often made in implied, hidden, denied, or assumed ways. To understand this layering within agonistic and antagonistic identity-claims, we differentiated between explicit and implicit claims. For an identity-claim to be called explicit, a PL clearly mentions the opponent or enemy in the text. If this is not the case, it is an implicit identity-claim. This results in four categories: (1) explicit antagonistic, (2) explicit agonistic, (3) implicit antagonistic, and (4) implicit agonistic.

In the third step, we examined how PLs strategically use social identities by identifying which discursive strategies they deployed while constructing a positive image of us and a negative image of them (Mols et al., 2023). We did so by labelling the discursive strategies used to express identity-claims as political and rhetoric tools, as “discursive patterns or ‘ways of talking’ about the ‘other’” (Augoustinos & Every, 2007, p. 124). These strategies can undermine the position of the political opponent or outline the own moral position. Our aim here was to identify how the discourse of the other was constructed by the politicians and whether this construction might enable antagonistic polarization.

We found 249 unique identity-claims in the examined data. That is to say, 249 pronouns such as I in combination with you, we, us, they, them, our people, our nation followed by phrases and text passages in which PLs refer to the collective self-understanding of people and present themselves as prototypical for their followers (Mols et al., 2023). As expected, from an identity leadership perspective all ten PLs voice identity-claims to connect with their own followers via social identity. For example, JK (GreenLeft, left) makes the apparent claim of being “one of us” in a response to the resignation of the government due to the Dutch childcare benefits scandal when he discusses being a father:

In contrast, FA (Political Movement Denk, center-left) aligns himself with all Dutch people struggling to pay high gas bills due to the war in Ukraine:

Of the 249 claims, nine could not be assigned since they are not made in opposition to others. They include phrases such as “(..) we are here not for ourselves, but for the Netherlands. For Dutch people” (MR4, People’s Party for Freedom and Democracy, center-right). The remaining 240 identity-claims are subjected to further analysis.

In the second step, the 240 identity-claims are assigned to the following categories: explicit antagonistic (70), explicit agonistic (79), implicit antagonistic (5), and implicit agonistic (86). These claims illustrate how PLs draw category boundaries between us and them whereby they redefine the shared self-understanding of who we are (Haslam et al., 2022).

Just as analyzing pronouns alone is insufficient to assess whether an identity-claim is agonistic or antagonistic (Step 1), assigning identity-claims to categories of claims (Step 2) does not indicate whether claims might enable antagonistic polarization. For that, we need to consider the discursive strategies of PLs (Step 3). Discursive strategies are understood as acts of identity entrepreneurship which help to shape and redefine the shared self-understanding of followers (Mols et al., 2023). Based on the theory of identity leadership, we understand leadership as a mutual process produced by interactions among followers and leaders (Haslam et al., 2020). We analyze this relationship by using a critical discursive analysis (CDA) to argue that social identities are not only gradually constructed by leaders and followers but are also based on acquired social knowledge (Van Dijk, 2009). This knowledge is obtained by experience and discourse on social identities.

As mentioned, the analysis shown in this study shows four discursive strategies. These four strategies illustrate how social identities are strategically applied and how the strategic use of identities is multidimensional. We acknowledge that the identity-claims found in this study are the result of subjective interpretations as to what qualifies as an identity-claim. We have attempted to transparently and clearly explain our methods and requirements for statements to be defined as identity-claims, but we understand that especially the implicit identity-claims can be called into question by others. Furthermore, this study does not give insight into the ways in which the identity-claims and discursive strategies are received by society. We do not know whether these claims and strategies indeed change the social realities of people, let alone if the claims lead to decreased or increased levels of polarization. We invite future research on this issue.

Future research directions can include an examination of who are being influenced by the discursive strategies of political leaders, and in which ways that occurs. It is crucial to discover which part of the message appeals to which groups and how people act upon the claims they hear. An additional thought-provoking line of research is to identify how the rhetoric of political leaders resonates in the identities of young people in their formative years. We assume that such claims influence the way they develop their identities, however, how and in which ways is unknown. How do political leaders discursively influence the identity formation of young people in such polarized times? Are young people cognizant of the ways in which they are being influenced and what does this mean for in-group and out-group dynamics within societies?