A warrant for violence? An analysis of Donald Trump's speech before the US Capitol attack

On January 6th, 2021, Donald Trump's speech during a ‘Save America’ rally was followed by mass violence, with Trump's supporters storming the U.S. Capitol to prevent the certification of Joe Biden's victory in the presidential election. In its wake, there was a great deal of debate around whether the speech contained direct instructions for the subsequent violence. In this paper, we use a social identity perspective on leadership (and more specifically, on toxic leadership) to analyse the speech and see how its overall ar-gument relates to violence. We show that Trump's argument rests on the populist distinction between the American people and elites. He moralises these groups as good and evil respectively and proposes that the very existence


Abstract
On January 6th, 2021, Donald Trump's speech during a 'Save America' rally was followed by mass violence, with Trump's supporters storming the U.S. Capitol to prevent the certification of Joe Biden's victory in the presidential election. In its wake, there was a great deal of debate around whether the speech contained direct instructions for the subsequent violence. In this paper, we use a social identity perspective on leadership (and more specifically, on toxic leadership) to analyse the speech and see how its overall argument relates to violence. We show that Trump's argument rests on the populist distinction between the American people and elites. He moralises these groups as good and evil respectively and proposes that the very existence of America is under threat if the election result stands. On this basis he proposes that all true Americans are obligated to act in order prevent Biden's certification and to ensure that the good prevails over evil. While Trump does not explicitly say what such action entails, he also removes normative and moral impediments to extreme action. In this way, taken as a whole, Trump's speech enables rather than demands violence and ultimately it provides a warrant for the violence that ensued.

K E Y W O R D S
Capitol attack, identity leadership, mass mobilisation, social identity, toxic leadership, Trump, violence

I N TRODUC T ION
At noon on 6 January 2021, Donald Trump addressed the crowd at a 'Save America' rally in Washington, DC, just prior to a joint meeting of Congress to certify the victory of Joe Biden in the Presidential Election of 3rd November 3 months earlier (Kambhampati et al., 2021). Within an hour, barricades around the Capitol Building, where Congress was convening, were under assault. Within 2 h the Capitol had been invaded, the certification meeting had been suspended and the building had been put on lockdown. Five people died, including one police officer and one protestor who was shot by the police (Evelyn, 2021).
To what extent were Trump's speech and the assault on the Capitol connected? This question became a matter not only of academic debate, media discussion, and popular argumentation, but also the subject of an impeachment trial for Trump in the Senate (the second that he had faced in his Presidency). The debate largely turned on whether Trump had explicitly told crowd members to invade the Capitol and disrupt the certification process. This led both accusers and prosecutors to search for specific passages within the speech that contained explicit guidance about how to act. In particular, the article of impeachment (Naylor, 2021), presented to the Senate on 25 January 2021, alleged that Trump 'wilfully made statements that, in context, encouraged -and foreseeably resulted in -lawless action at the Capitol, such as "if you don't fight like hell, you're not going to have a country anymore"'.
Conversely, those supporting and defending Trump argued that such passages conveyed no such instructions. In a response to the article of impeachment, the former President's lawyers argued that Trump did not incite the crowd to engage in destructive behaviour. They supported that 'It is denied that the phrase "if you don't fight like hell you're not going to have a country anymore" had anything to do with the action at the Capitol as it was clearly about the need to fight for election security in general' (Castor Jr & Schoen, 2021, p. 6). During the subsequent impeachment process, in their trial memorandum, Trump's lawyers contended that any arguments regarding the mobilization of mass violence were based on a false narrative that rested on a 'cherry-picked, non-contextual parsing' of the speech itself (Castor Jr et al., 2021, p. 10), that 'the president didn't mention violence ' (p. 11), and that it is 'important to read the speech in its entirety ' (p. 10).
In this paper, we provide precisely such an analysis of Trump's speech in order to explore the extent to which the overall argument can be seen as advocacy for violence. Our reading of the speech is grounded in the social identity model of leadership (Haslam et al., 2011) and more specifically in the social identity analysis of toxic leadership (Haslam et al., 2020;Reicher et al., 2008), as well as in previous social identity analyses of Trump's leadership (Reicher & Haslam, 2017a, 2017b. We outline each of these in turn before turning to the specific contribution of this paper.

The social identity model of leadership
The core premise of the social identity approach in social psychology (Turner et al., 1987) is that people have multiple identities relating both to their individual characteristics (personal identity) and their membership of social groups (social identity). Different aspects of the self-system will be salient in different contexts, and when any specific social identity is salient enough to serve as a basis for selfdefinition, people will act in terms of the norms, values, beliefs and interests associated with that particular group.
A key feature of the approach is that it contains an explicit model of social influence (Turner, 1991). That is, it argues that when-and to the extent that-group members seek to behave according to a salient group identity, they will be influenced by those who are in a position to define that group identity (i.e. ingroup members) and by messages that are seen as consonant with that group identity. For example, in a context where Bob defines himself as a Republican, he will be influenced more by someone who he understands to be 'a good Republican' than by someone else (e.g. 'a Republican in name only' or a Democrat).

IDENTITY LEADERSHIP AND COLLECTIVE VIOLENCE
Moreover, in addition to containing a model of social influence, the social identity approach also contains an implicit model of leadership (Turner, 1991). This link has been made explicit by a number of authors. In particular, Hogg (2001) noted that effective leaders need to be seen not just as ingroup members but as prototypical of the groups they seek to represent (see also Turner & Haslam, 2001). More recently, Haslam et al., (2011) and Haslam et al., (2020) have proposed a more comprehensive model of identity leadership. This characterizes leadership as a process of social identity management in which leaders exert influence by defining 'who we are' and hence what we should do. Effective identity management is contingent not only on being seen as a prototypical ingroup member ('being one of us') but also on being seen to act for the group's interest (as opposed to personal interest or that of some other group) and on achieving group goals ('doing it for us').
However, critically, the definition of social identity is not a given. It is not something predetermined that the leader simply draws upon to guide group action. Instead, leaders are actively involved in the definition of social identity ('creating a sense of us') both in word and deed. This operates at various levels, including defining people as belonging to a common group, defining the boundaries of inclusion, defining the content (norms, values, etc.) and defining the prototypes of the category. In this way, leaders are able to construct themselves, their policies and their accomplishments as instantiations of the group identity. That is, effective leaders are accomplished entrepreneurs of identity (Reicher & Hopkins, 2001).

Social identity and toxic leadership
From a social identity perspective, toxic leadership has been analysed at two levels. The first is intragroup (Haslam & Reicher, 2016). If influence derives from group identity, then authority in a group depends upon who is allowed and able to define that identity. Accordingly, one can identify a continuum from fully democratic leadership (in which the leader claims no precedence in defining 'who we are' but acts to facilitate a conversation amongst group members around group identity and its implications for action), to hierarchical leadership (in which leaders present their definitions of identity and its implications as authoritative, but do not disallow rival versions), to authoritarian leadership (where leaders elide their own personal selves with the group identity such that anyone who disagrees with them personally is against the group and liable to exclusion). Once the leader is seen as the embodiment of the 'general will' and opponents of the leader become 'enemies of the people', democratic debate and dissent become impossible (Pettersson, 2019;Rapley, 1998).
The second level of toxic leadership is intergroup (Reicher et al., 2008). This has to do with setting up certain populations as an outgroup and as a threat to the ingroup. For example, migrants can be portrayed as outsiders to a national ingroup and be accused of threatening 'us' economically ('they take our jobs'), culturally ('they change the character of our high streets') or existentially (in narratives such as the 'great replacement theory', which posits that an indigenous ethnic community will be replaced by ethnic outgroups; Obaidi et al., 2022). This then creates the conditions for violence against those groups, especially where ingroup norms endorse violence (Littman & Paluck, 2015) and under conditions of perceived existential threat (Obaidi et al., 2022).
Moreover, when group relations are moralized such that the ingroup is represented as 'good' and outgroups as 'evil', then toxic behaviours against the outgroup can be celebrated as acts of virtue that preserve the 'greater good' (cf. Haslam et al., 2019; see also Gere, 2017;Koonz, 2003). Supporting this analysis, studies by Verkuyten (2013) in the Netherlands, as well as Pettersson (2019), and Sakki and Martikainen (2021) in Finland, provide empirical demonstrations of how the use of such 'ingroup virtue threatened by outgroup' rhetoric justifies and promotes hate and discrimination.
Yet while it is possible analytically to distinguish the intragroup and intergroup dimensions of toxic leadership, in practice, they are often entwined. This point was recognized in the Nazi jurist (and inspiration to many contemporary far-right figures) Carl Schmitt's analysis of 'enemyship' (Schmitt, 1927(Schmitt, /1996; see also Engels, 2010). For by identifying an enemy who threatens the ingroup and proposing action against that ingroup, a leader can claim to be acting for the ingroup's interest at the same time that they frame their rivals as failing to defend the group (either through collusion, ignorance or weakness) and hence being unfit to lead. Along these lines, Portice and Reicher (2018) show how the (anti) immigration speeches of British leaders presuppose that migrants form a threat while simultaneously explaining how they are uniquely qualified to deal with this threat. What appears to be an intergroup argument is fundamentally about an intragroup claim.

Donald Trump's identity leadership
Trump's rhetoric fits into a tradition of populism whereby 'the people' are in opposition to 'the establishment' (Bracciale & Martella, 2017;Mudde, 2004). More specifically, 'the American people' are contrasted with both internal and external enemies. The former is made up of a political elite that is presented as being in collusion with the latter, which includes governments, criminals and migrants. This is encapsulated in Trump's 'Argument for America', the broadcast advert with which he signed off his successful 2016 Presidential campaign, which opened with the claim that 'Our movement is about replacing a failed and corrupt political establishment with a new government controlled by you, the American people' (see http://www.livin groom candi date.org/comme rcial s/2016/donal d-trump s-argum ent-for-ameri ca#).
Trump positions himself rhetorically by describing himself as an ordinary guy from the wrong side of the tracks who had achieved the American dream of success through sheer hard work: 'we didn't learn from MBAs, we learnt from people who had doctorates in common sense', he says (quoted in Reicher & Haslam, 2017a, 2017b. But he also does it performatively. Thus, while they might be the undoing of other candidates, his repeated violations of the normal rules of political decorum (crude language, swearing, offensive slurs against his opponents) actually bolstered his campaign. For in violating 'their' rules-those of the political establishment-Trump firmly established himself as 'one of us' (Reicher & Haslam, 2017a, 2017bTaibi, 2018).
Trump likewise positions himself as acting and achieving for the group, arguing that (unlike his opponents), his wealth means that he cannot be bought by corporate interests and that his business skills in 'the art of the deal' (Trump & Schwartz, 2009) will be turned to the interests of the group. According to him, after years of being fleeced by rivals like China, a Trump-led America will prevail once more.
But while Trump's identity management is based on a fairly conventional populist set of categories, he also brings something distinctive to the table (albeit subsequently copied by others) in the form of his MAGA ('Make America Great Again') construction. This neatly summarizes the idea that America was great but has fallen from greatness and will only be returned to its rightful greatness by Trump. At one level, the argument is familiar in the USA, drawing on a form of rhetoric that Bercovitch (1978) dubs 'the American Jeremiad'. This is the idea that America is morally exceptional, a City on a hill that has a unique destiny to do good in the world. But, by that very token it must be continually interrogated to see if it is living up to its mission. Where it is found to have fallen short by failing to apply our standards to our treatment of others, the nation must be renewed in order to recover its true self. As others have shown, this Jeremiad discourse is at the core of many of the seminar moments of American leadership performance, including Lincoln's Gettysburg address (Wills, 2006) and Martin Luther King's 'I have a dream' speech (Howard-Pitney, 1986).
In its original formulation, the Jeremiad argument is progressive in so far as it is aimed at overcoming division and conflict. However, Trump's version differs from this in two critical respects. First, whereas the original formulation ascribes the decline of the group to the moral failures of its members, Trump ascribes it to the acts of its enemies (both internal and external). It is the fault of the Washington 'swamp' and foreign cronies. As Trump puts it in his 'Argument for America', it is the 'political establishment' in alliance with 'a global power structure' that 'has brought about the destruction of our factories and our jobs as they flee to Mexico, China and other countries all around the world' (Charteris-Black, 2018, p.23). Correspondingly, where the original sees renewal as coming from an internal act of rededication and renewal, Trump sees it as coming from defeating one's enemies and laying them low: 'draining the swamp', building a wall to keep out aliens.
Second, whereas the original Jeremiad places agency in the action of ordinary Americans themselves, the Trump version portrays the people as incapable of defending themselves and needing a champion (himself) to do so. Renewal will come from uniting under Trump, whose agency will win the day. Thus he concludes his 'Argument for America' by saying 'I'm doing this for the people and for the movement and we will take back this country for you and we will make America great again'.
Once again, Trump advances these arguments performatively as well as rhetorically. In particular, analysis of Trump rallies, which played a key part in mobilizing support for his election campaign (Stone, 2017), shows how they functioned as 'morality plays' in which internal enemies (the press) and external enemies (protestors) who normal oppress 'the people' were silenced, cowed and ejected as a result of Trump's agency, thereby showing how-on a larger stage-America could become great again (Reicher & Haslam, 2017a, 2017b.
In this way Trump offers up what can be understood as a twisted Jeremiadin effect a fusion of the original with concepts of enemyship-that shifts the focus from personal renewal to the defeat or destruction of enemies. It is less about concern for others who have been mistreated than about resentment at other who have taken what is 'ours' by right (cf. Wodak, 2015Wodak, , 2017. It is less about intragroup harmony than intergroup conflict. Moreover, in this context, the moral exceptionalism of the ingroup becomes a warrant for extreme acts against those who threaten it.

The present study
In this paper, we apply the foregoing analysis of leadership, toxic leadership and Trump's twisted Jeremiad to his 6th January speech prior to the assault on the Capitol. Importantly, we do not directly examine the impact of the speech on crowd members. We have dealt with this elsewhere both in academic form  and in an unpublished paper to the US House Committee investigating the Capitol assaults . Rather, our focus here is on the nature of the speech itself-on whether and in what ways it constituted advocacy for, and legitimation of, violent action.
It is also important to recognize that the context in which Trump spoke was very different from that of his previous speeches. Before, Trump was speaking in a bid to become President, as President or to stay as President, and he was talking in general terms about enemies, the need to defeat them, and the need for people to support him in defeating them. By contrast, the 6th January speech was oriented to preventing a very specific event-the certification of the 2020 election result by Congress. Moreover, it required direct action by followers, not just support for Trump himself. So how did this context, which we spell out in more detail below, impact on Trump's rhetoric and how it relates to his 'twisted Jeremiad' rhetoric?
The speech was part of a movement of protest against the results of the 2020 US Presidential elections. As early as May 2020, Donald Trump had started to question the authenticity of postal voting due to the COVID-19 pandemic and to claim that it would damage the Republican Party and the nation. As the election came closer, Trump increasingly attacked COVID-19 lockdowns as a cloak for election fraud that would risk the election being stolen from him. Even before the election, he suggested that he might not accept its results, especially in the swing states. In this he was not alone, and a 'stop the steal' movement grew on social media from early September 2020 (Holt, 2021). This movement became more active during the election with armed protests at both Michigan and Arizona election counts on election day (Pilkington, 2021). It became especially prominent after 7 November 2020, the day Joe Biden (with 80 million votes compared to Trump's 74 million) was projected to be the next President. At this point there were 'Stop the Steal' protests across the country attended by armed groups such as the Proud Boys. At a so-called 'Million Maga' march on 14th November there was violence against counter-protestors. Trump himself justified the violence claiming it was a response to antifa attacks and calling on the police to 'do your job and don't hold back' (Holt, 2021). Soon, though, the various groups contesting the election result settled on 6 January 2021 as the key date for protest. This was the date on which Congress, presided over by the US Vice-President was to formally certify the election result. The certification is normally a ceremonial occasion (Dacey, 2020) since, by then, the election result has been accepted. But formally, it is the deciding moment and in 2021 Trump and others hoped it could be used to reject the result and declare him President for a second term. Accordingly, a 'Rally to Save America' was planned for the day. Publicly, this was organized by groups independent of Trump such as 'Women for America First' and 'Stop the Steal', but Trump's staff were centrally involved in organizing and promoting the event (Steakin et al., 2021). Trump himself tweeted on 19th December: 'Big protest in D.C. on January 6th. Be there, will be wild!' (Barry & Frenkel, 2021). Over the following days and weeks, he continued to post similar messages on social media, encouraging people to attend the event and reposting multiple other advertisements for it (Sherman, 2021).
It was in this context that the 'Rally to Save America' began at 9 AM on January 6th January. There were 12 speakers before Trump, the first of whom-Mo Brooks, the Republican representative from Alabama-addressed the crowd as 'American patriots'. He asked them 'Are you willing to do what it takes to fight for America?' and then urged them to 'Carry the fight to Capitol Hill'. This was echoed by several other speakers. Katrina Pieson, for instance declared 'we are going to fight for our country', Amy Kramer urged the crowd to 'keep fighting for America' and Donald Trump Jnr did likewise: 'stay in this fight! Stay loud! Don't be suppressed!… Stand up and hold your representatives accountable'. More concretely, Eric Trump told the crowd: 'we need to March on the Capitol today'.
President Trump himself started speaking just before noon and continued for some 70 min. Even before he began, though, the very title of the rally ('Save America') and those before him had provided a clear categorical frame for his own remarks. The context was defined as an existential threat to the nation, in which American patriots needed to fight for their country to stop it being stolen from them. But how, and to what extent, did Trump reinforce this frame? And how did he elaborate on it? These are the questions that our analysis aims to address. More specifically we ask: 1. How did Trump characterize the nature of the categories (sides) involved in the dispute over the election result? 2. Did Trump moralize those categories as good or evil-and, if so, how? 3. How did Trump characterize the relations between the 'sides' to the dispute? 4. What forms of action did Trump advocate for his followers? 5. Are there any other aspects of Trump's speech which are critical to his advocacy?

M E T HOD Data
Our analysis focuses on President Donald Trump's speech at the 'Rally to Save America' which took place in Washington on 6 January 2021, just before Congress was about to certify the 2020 US presidential election results. The transcript of the talk is presented in Appendix S1 and also became immediately available online on various websites (e.g. see https://www.rev.com/blog/trans cript s/donal d-trump -speec h-save-ameri ca-rally -trans cript -janua ry-6).

Analytic approach
Considering the textual nature of our data, and our focus on the rhetorical construction of group identities, our analysis is primarily informed by the rhetorical approach commonly taken by researchers in the social identity tradition (e.g. Drury & Reicher, 2000;Mols et al., 2022;Reicher et al., 2006;Reicher & Hopkins, 1996. This approach can be seen as a form of hybrid thematic analysis, combining deductive (theory driven) and inductive (data driven) approaches (Braun & Clarke, 2022;Fereday & Muir-Cochrane, 2006). That is, we deductively code the speech according to our theoretical analysis of toxic leadership and Trump's 'twisted Jeremiad' rhetoric in particular. This is encapsulated in the first four of the above questions. Accordingly, we code for (a) the definition of social categories; (b) the moralization of social categories; (c) category relations; and (d) calls to action. Additionally, as encapsulated in our fifth question, we code inductively for any other elements in the speech that impel action by followers.
In practical terms, our procedure is broadly based on the standard six-step model of conducting thematic analysis: familiarizing oneself with the data, initial coding, creating themes, reviewing themes, determining the significance of themes, writing up (Braun & Clarke, 2022). We started with all five authors reading the speech and discussing its overall categorical structure as well as how this relates to Trump's previous rhetoric. From this, one of the authors did a first coding, abstraction, review and evaluation of themes; and wrote a draft analysis. A second author then did an independent coding, analysis of themes and revised the previous draft. This was then discussed between the two analysts and an agreed text was circulated and finalized with the other authors.
To substantiate our arguments, we use numbered extracts in italics which are verbatim passages from Trump's speech. The use of […] denotes missing material. We use this, principally for reasons of space, in order to exclude material that is repetitive or irrelevant to the development of the point at hand. Where we use italicized material within the body of the analysis this also is verbatim material from the speech, but generally not of sufficient significance or length to warrant a separate extract.

A NA LY SIS
Our analysis is structured as follows: First, we present the core categories of 'us' and 'them' that appear throughout the speech. Second, we show how these categories are moralized as 'good' and 'evil'. Third, we consider how Trump construes the relations between groups and the significance of the outgroup ('them') for the ingroup ('us'). Fourth, we examine how these various constructions are used to legitimize and impel action-particularly potentially violent courses of action on the part of his audience. Our analysis is exemplified in Figure 1.

F I G U R E 1
The process whereby leaders warrant the use of mass violence by their followers.

Constructing categories
Near the beginning of his speech, Trump says: [Extract 1] It's just a great honor to have this kind of crowd and to be before you. Hundreds of thousands of American patriots are committed to the honesty of our elections and the integrity of our glorious Republic. All of us here today do not want to see our election victory stolen by emboldened radical left Democrats, which is what they're doing and stolen by the fake news media. That's what they've done and what they're doing. We will never give up. We will never concede, it doesn't happen. You don't concede when there's theft involved. Our country has had enough. We will not take it anymore and that's what this is all about.
The categories could not be clearer. On the one hand we, the ingroup, are American patriots. These are the country. They have had enough. Importantly, though, they are not just of the country, they are for the country. Being American is defined as a politicized activist identity in which people stand up for the nation and defend it when it comes under attack. More specifically, they are committed to challenging electoral malpractice. In Trump's own words: [Extract 2] American patriots are committed to the honesty of our elections and the integrity of our glorious Republic. All of us here today do not want to see our election victory stolen by emboldened radical left Democrats, which is what they're doing and stolen by the fake news media.
On the other hand, they, the outgroup, are described not just as Democrats, but as 'emboldened', as 'radical' and as 'left'. It is important to stress that this characterization is meant to distinguish a subset of extremist Democrats from the rest of the party. But elsewhere in the speech Trump insists that all Democrats are complicit in the in election fraud:

[Extract 3] Democrats attempted the most brazen and outrageous election theft. There's never been anything like this. It's a pure theft in American history, everybody knows it.
Here, then, Trump implies that to be a Democrat is, in and of itself, to be radical, left and now emboldened. That is, Democrats are ideologically alien to the American tradition-an alien status already hinted at in Extract 1 when Trump contrasts them to 'our country' and which he argues explicitly further on in the speech that 'Democrats enacted policies that chipped away our jobs, weakened our military, threw open our borders and put America last'. Sometimes, though, Trump insinuates, it is much simpler than that. Democrat voters are just plain foreigners. In Georgia, he claims, they stopped citizenship confirmation 'Are you a citizen? You're not allowed to ask that question. Because they want to steal the election'. In every way, Trump argues, Democrats are at odds with 'us' and what is 'ours' as Americans.
Yet while (emboldened, radical, left) Democrats are the major component of the outgroup it is clear that he does not see them as acting alone. From the very start Trump enjoins the media to 'turn your cameras please and show what's really happening out here' while simultaneously declaring that 'the media will not show the magnitude of this crowd', and as we saw in Extract 1 the 'fake news' media are portrayed as complicit in 'the steal'. Like the Democrats, the media are not of the people and they are most certainly not for the people. On the contrary, Trump claims 'Our media is not free. It's not fair. It suppresses thought. It suppresses speech, and it's become the enemy of the people'. But the media, however bad, are not alone. According to Trump:

| 9
IDENTITY LEADERSHIP AND COLLECTIVE VIOLENCE So the list lengthens-an anti-American outgroup made up of Democrats, of the media and of big tech. There is just one more element in this coalition-one that is particularly important in terms of Trump's overall argument. He sometimes refers to them as 'weak Republicans' and sometimes by the acronym 'RINOs' (Republicans in name only). If, as we saw above, Democrats are accused of trashing American interests, this is only because some Republicans collude in allowing this to happen: [Extract 5] The weak Republicans, and that's it. I really believe it. I think I'm going to use the term, the weak Republicans. You got a lot of them, and you got a lot of great ones, but you got a lot of weak ones. They've turned a blind eye even as Democrats enacted policies that chipped away our jobs, weakened our military, threw open our borders and put America last. This takes us back to a point we made at the start of this analysis. That is, the ingroup is defined in activist terms-by acting to 'save America', by challenging the election result. Accordingly, those who fail to support this activism are, by definition, outgroup. In many ways, the whole point of Trump's speech-and the rally more generally-is to pose the question of their identity. Will the Republican Vice-President, Mike Pence and other Republicans in Congress block the certification and reclaim their patriotic, Republican credentials or not?
[Extract 6]: The Democrats are hopeless. They're never voting for anything, not even one vote. But we're going to try and give our Republicans, the weak ones, because the strong ones don't need any of our help, we're going to try and give them the kind of pride and boldness that they need to take back our country.

Moralizing categories
The moralization of the outgroup is evident all through Trump's speech. After all, the rally is defined as a response to their act of theft. This, and cognate definitions of the Democrats and their fellow travellers are used repeatedly by Trump. They constitute a 'criminal enterprise', they are 'brazen and outrageous' in their criminality, and their crime is uniquely extreme:

[Extract 7] This is the most corrupt election in the history, maybe of the world. You know, you could go to third world countries, but I don't think they had hundreds of thousands of votes and they don't have voters for them. I mean, no matter where you go, nobody would think this. In fact, it's so egregious, it's so bad, that a lot of people don't even believe it.
That's just the start of it. They defraud the people. They cheat. Democrats along with the cities, counties and states they run are corrupt. Equally the media are corrupt, suppress the truth and lie. Big tech along with the media are people 'who want to deceive you and demoralize you and control you' and who also 'want to indoctrinate your children in school by teaching them things that aren't so'. In the simplest and clearest of moral terms, they are bad. In sum, then: But while Trump's characterization of the outgroup is replete with negative moral labelling, he is just as elaborate and explicit when it comes to moralizing the ingroup. For instance, he observes: [Extract 9] As this enormous crowd shows, we have truth and justice on our side. We have a deep and enduring love for America in our hearts. We love our country. We have overwhelming pride in this great country, and we have it deep in our souls. Together we are determined to defend and preserve government of the people, by the people and for the people.
There are multiple good moral qualities here: the patriot crowd are devoted to democracy. They are characterized by truth and justice. But what predominates over all of these, and what Trump insists upon at multiple points in the speech is love: love of country, love of each other, love of our movement, love for Trump, and Trump's love for them. This loving quality characterizes two telling interchanges in which the crowd responds spontaneously to Trump's words: [Extract 10] Trump: There's never been a movement like this ever, ever for the extraordinary love for this amazing country and this amazing movement. Thank you. Crowd: We love Trump! We love Trump! We love Trump!" [Extract 11] Trump: I could go on and on about this fraud that took place in every state and all of these legislatures want this back. I don't want to do it to you because I love you and it's freezing out here, but I could just go on forever. I can tell you this… Crowd: We love you. We love you. We love you. We love you. We love you. We love you. We love you. We love you.
The contrast between outgroup and ingroup, then, is a stark contrast between evil and good. But, more than that, it is a struggle between those who want to create and those who want to destroy: 'You're the real people. You're the people that built this nation. You're not the people that tore down our nation', he says. This takes us to our next set of themes which concern Trump's depiction of the relations between groups.

Defining category relations
Just as the very concept of a rally to 'Stop the Steal' implicitly moralizes social categories, so too the notion of a rally to 'Save America' also implies an antagonistic relationship between categories. If America has to be saved, then America is under threat. But saved from whom and saved from what? We have already addressed the first question in the previous sections: saved from the combined anti-American forces of Democrats, media, big tech all aided and abetted by 'weak Republicans'. As for the second, again Trump provides a range of answers.
First of all, as we saw in Extract 6, there is a threat to material interests such as jobs and standards of living. Indeed one of Trump's arguments about the election being stolen is that he so obviously improved America's economic outlook, and those gains would so obviously be at risk under the Democrats, that he would self-evidently have won any fair election: [Extract 12] So we've taken care of things. We've done things like nobody's ever thought possible. And that's part of the reason that many people don't like us, because we've done too much, but we've done it quickly. And we were going to sit home and watch a big victory. And everybody had us down for a victory. It was going to be great. And now we're out here fighting.
Second, a Democratic victory is a cultural threat. It endangers America's most prominent and cultural symbols-which Trump would protect: [Extract 13] We will not take the name off the Washington monument. We will not. Cancel culture. They wanted to get rid of the Jefferson Memorial, either take it down or just put somebody else in there. I don't think that's going to happen. It damn well better not. Although with this administration, if this happens, it could happen. You'll see some really bad things happen. They'll knock out Lincoln too, by the way. They've been taking his statue down, but then we signed a little law. You hurt our monuments, you hurt our heroes, you go to jail for 10 years and everything stopped.
Third, the Democrats threaten core American institutions. The crowd, says Trump, is there 'for the sake of our democracy, for the sake of our Constitution and for the sake of our children'. The conjunction is telling, for these institutions are not incidental aspects of American life, they are what makes | 11

IDENTITY LEADERSHIP AND COLLECTIVE VIOLENCE
America what it is. Take them away and 'we won't have a country'. The threat is to the very future of the group, as represented by 'our children'.
Overall, then, the threat is existential. It is about enemies of America who threaten 'illegally to take over our country'. It is about the very real prospect that 'you're not going to have a country anymore'.

Enjoining action
Let us start this section by providing a fuller context for the quote with which we concluded the previous section: [Extract 14] We fight like hell and if you don't fight like hell, you're not going to have a country anymore.
There is a degree of pragmatism in this argument. That is, if you face an existential threat to the group, you cannot decide to err on the side of caution and leave the fight for another day. For, if you wait, there will no longer be a group to fight for. In this sense, the construction of an enemy who poses an existential threat to the ingroup impels action now.
At the same time, there are also moral dimensions to Trump's argument, ones that operate at both a universal and a particularistic level. That is, on the one hand, there is a universal responsibility to challenge what is immoral. On the other hand, our group in particular (American patriots) is characterized by standing up for what is morally right (America).
We saw both arguments in operation in Extract 1. There, the universal was expressed in the general claim that 'you don't concede where there's theft involved'. The particularistic was expressed both in the characterization of 'American patriots' as people who 'never give up' and also in the characterization of American institutions in terms of 'honesty' and 'integrity', and of the American Republic as 'glorious'.
We also saw previously the corollary of this activist definition of American patriotism. Those who fail to act against 'the steal' have their American (and Republican) identities called into question. Whether they oppose certification or not will determine whether they are part of the ingroup: [Extract 15] Today we will see [not only] whether Republicans stand strong for integrity of our elections, but whether or not they stand strong for our country, our country.
If you are to be considered a true Republican and American patriot, you are obliged to take action. If you do not, you will be excluded (or, as Trump says of those Republicans who fail to oppose certification, 'we have to primary the hell out of those who don't fight'). More than this, though, you are obliged (and justified) in abandoning all niceties and fighting tooth and nail to preserve America (the good) against its enemies (the bad): [Extract 16] Republicans are constantly fighting like a boxer with his hands tied behind his back. It's like a boxer, and we want to be so nice. We want to be so respectful of everybody, including bad people. We're going to have to fight much harder.
In these various ways, Trump employs the construction and moralization of categories, along with the sense of an existential threat to America, to oblige strong action. He does so on multiple pragmatic and moral grounds. But action to what end, and what sort of action? Trump is forthcoming about the former. The action is to 'Stop the Steal', which, in context, means stopping the certification of the election results in Congress later that day. More specifically, as we saw in Extract 6, it means pressuring Republicans in Congress to oppose certification since Democrats are a lost cause.
Trump is less forthcoming about the precise forms of action necessary to achieve this end. He talks on several occasions about walking or marching to the Capitol. He tells the crowd to 'peacefully and patriotically make your voices heard'. But this does not get us very far in understanding what people should do along the way or once they get to Congress since the definition of violence is notoriously open (De Haan, 2009;Gorringe, 2006): one person's violent action is another's legitimate self-defence; another's 'assault' may be our non-violent direct action. So, ultimately, while Trump is clear that the crowd, as American patriots, is obliged to act, to abandon their traditional restraint and to stop Congress (especially Republicans in Congress) from certifying the election, he is much less clear about exactly what they should and should not do to reach these ends. But he does promise them that if they do what is necessary to defeat America's enemies, then a glorious future awaits. America will be great again: [Extract 17] Our exciting adventures and boldest endeavors have not yet begun. My fellow Americans for our and movement, for our children and for our beloved country and I say this, despite all that's happened, the best is yet to come.

DISCUS SION
The aim of this paper was twofold. First, examine how Trump's 6th January speech relates to toxic leadership rhetoric in general and to Trump's specific 'twisted Jeremiad' version of it. Second, to address how the overall structure of his argument constitutes advocacy for, and legitimation of, violence.
As regards the first, Trump's speech contains all the elements we have previously identified as associated with intergroup toxicity (Reicher et al., 2008). It creates a narrow ingroup of 'American patriots, from which certain Americans (including Democrats and Republicans who do not challenge the election result) are excluded and defined as an outgroup. It moralizes the groups such that the patriots are virtuous individuals sustaining a virtuous cause; they are brave crusaders for a cause motivated by selflessness and love. By contrast, those who endorse the election results are immoral on multiple counts: thieves, corrupt, and just plain bad. It then characterizes the patriot ingroup as under existential threat from the outgroup. The conclusion is that patriots are obligated to go all out to defeat their enemy. By stopping the certification of the elections in Congress, they thereby prevent the triumph of evil over good.
Trump's speech also reflects his particular rhetorical form of toxic leadership, the 'twisted Jeremiad'. America is a 'glorious' Republic, diminished by its enemies, who need to be defeated if America is to reclaim its rightful place and achieve its full potential. If every patriot does as they should, the 'steal' will be stopped, Trump will regain his rightful place as President, and then 'the best' will be 'yet to come'.
At the same time, the context of the 6th January rally is distinctive, and this is reflected in the nature of Trump's rhetoric. As we noted in the Introduction, his previous speeches and those we have previously analysed (Haslam et al., 2020;Reicher & Haslam, 2017a, 2017b were given in a context of soliciting general support for him to become or remain President. They were not about specific actions or the audience acting for themselves against his enemies. They were about sanctioning Trump's actions against their enemies. This speech was much more clearly about support for a specific outcome (stopping the upcoming certification of the election by Congress) to be achieved by the actions of followers. The emphasis is therefore much less on what Trump would do than on what his followers should do.
Additionally, Trump's position was radicalized compared to previous MAGA rhetoric. Whereas, previously, he asserted that America's enemies diminished the nation (which only he could restore), he did not put the existence of the nation in question. On January 6th he argued that the certification of the election would destroy the nation so that there would be nothing left to restore. Consequently, on 6th January, action was required. Anyone who considered themselves an American and who cared about America's virtues could not abstain or delay. This helps explain the repeated emphasis on a moral obligation to act. This is something that has recently achieved considerable attention in the social movement literature as a key determinant of participation (Agostini & van Zomeren, 2021;Sabucedo et al., 2018). Arguably, the greater the cost of inaction, the more that moral obligation comes into play.

IDENTITY LEADERSHIP AND COLLECTIVE VIOLENCE
In Trump's speech, such costs and obligations are greatly overdetermined. 'Glorious' America, he argues, is on the brink of annihilation. American patriots are obligated to respond since moral action is the norm for the group. But more than that, it is a defining feature of American patriotism and so one cannot be an American patriot without taking such action. And if that is not enough, there is a general moral injunction to defend the good against evil. By 'stopping the steal' crowd members cannot be accused of pursuing sectarian group advantage. They are simply doing what is right.
In general terms, then, Trump's speech fits the pattern of toxic leadership. It is also broadly consistent with his previous 'twisted Jeremiad' albeit with some significant elaborations. The most important of these is that Trump locates agency primarily in his audience rather than himself. This then radicalizes his appeal in the sense that, on January 6th, he calls on his followers to defeat their enemies and indeed obligates them to do so.
Let us now turn to the issue of Trump's rhetoric and violence. Our analysis here suggests four things. First, to repeat, Trump explicitly obligates his followers to do something in order to stop the certification of the election results. Second, it asks them to discard their customary constraints. Third, it proposes that followers march on Congress. Fourth, however, Trump is never explicit about what actions they should take once they are there.
Having analysed at the speech as a whole (rather than focusing on single phrases) and having placed it in its wider context (particularly the fact it was happening just before the final act of certifying the election), we are now in a stronger position to argue that, on 6th January Trump's rhetoric does constitute advocacy for violence. At the same time, a contextualized analysis leads us to a more nuanced understanding of the nature of this advocacy. Trump does not give instructions to be violent. He is largely mute on what people should do, and the few snippets one can extract confuse rather than clarify the matter: alongside telling people to 'fight like hell' he tells them to act 'peacefully'. Instead, by combining notions of identity, morality and existential threat-thereby representing the issue as preserving good in the face of evil-Trump removes restraints on action. For, as we showed in the Introduction, the notion that we are acting to advance 'a greater good' (in this case the survival of the 'glorious' American republic) has frequently been shown-both in historical studies (Koonz, 2003;Reicher et al., 2005) and studies in the laboratory (Birney et al., 2023;Haslam & Reicher, 2017)-to provide a warrant for toxic behaviour. In a phrase, Trump does not demand violence, but he does enable it. This is not surprising, especially when considering that leaders can recognize and manage accountability concerns when enabling particular courses of action (see . Moreover, while enabling may seem less extreme than instructing, that is not necessarily the case. In his influential work on the operation of the brutal Nazi state, Kershaw (1993) explodes the myth that Nazi Germany was run by a system of strict orders. Rather, Hitler would set broad objectives (such as 'eliminate the Jewish threat') without specifying what exactly should be done. In response, his subordinates would seek to outdo each-other in realizing their leader's wishes, becoming ever more radical in the process (Sofsky, 1993; see also Haslam & Reicher, 2007). This process, Kershaw calls 'working towards the Fuhrer'. Through a similar process (but, of course, with very different ends and means), one can see Trump as setting the general goal of stopping the certification and his followers then seeking to outdo each other as they 'work towards Trump' .
As for the wider significance of our analysis, we have already made it clear that there are certain elements of Trump's rhetoric that draw on specific cultural forms-most obviously the 'American Jeremiad'. In particular, the general structure of his MAGA rhetoric -that 'we' are exceptional, that 'we' have been brought low (or even face annihilation) by venal enemies, and that these enemies must be defeated (or even destroyed) so we can rise to our true glory once more and good can prevail-is far from unique. The rhetoric of many contemporary leaders-Orban in Hungary (Lendvai, 2019), Modi in India (Jaffrelot & Schoch, 2023), Duterte in the Phillipines (Miller, 2018)-is structured in very much the same way. But perhaps the clearest example is that of Putin in Russia. As Sharafutdinova (2020) argues, Putin mobilized a militant nationalism by activating a sense of Russian exceptionalism ('Holy Russia') combined with 'a sense of foreign threat to the state and its people ' (p.18). This is central to his justification for the war in Ukraine. In a speech on 21 February 2022, 3 days before the invasion, Putin characterized the situation as part and parcel of a longstanding attempt by a corrupt and immoral West to bring greater Russia (of which Ukraine is part-an 'inalienable part of our own history, culture and spiritual space') because the West, and America in particular, 'just do not need a big and independent country like Russia around' (http://en.kreml in.ru/event s/presi dent/news/67828). This characterization then allowed the Russian leader, at a speech in the Luzhniki stadium shortly after the invasion, to represent the war as a heroic moral act to save the Russian people from Western oppression: 'I recall the words from the Holy Scripture: 'Greater love hath no man than this, that a man lay down his life for his friends. And we are seeing how heroically our military are fighting during this operation' (http://en.kreml in.ru/event s/presi dent/news/68016).
What we are suggesting, then, is that while the present analysis must be situated in a specific context, it can contribute to the elaboration of a more general model of the mobilization of collective hate and violence. This model involves a number of elements: the historical exceptionalism (both moral and material) of the ingroup; the threat (in particular, the existential threat) of an outgroup; the necessity to eliminate the outgroup threat in order to restore the true nature and status of the ingroup; and the moral obligation of group members to act against the threat.
It is worth noting that such a model has implications for our understanding of group processes in general and social identity processes in particular, which go beyond the issue of hate and violence. Most importantly, it brings a temporal dimension to our understanding of group prototypicality and social influence. That is, these are not perceptions of what makes the ingroup distinctive from the outgroup in the present (Rosch, 1978;Turner, 1991;Turner et al., 1987). Rather, they are constructions of moments in history when the group was 'truly itself' (with the implication that the group in the present may not represent the group's identity at all). What is more, it suggests that a particularly powerful form of influence may be to argue that we need to change from what we currently are in order to return to the prototype (i.e. the 'true self').
Clearly, a range of research is needed before talk of a 'model' is warranted-both rhetorical to examine the use of these forms of argument by different leaders in different places and experimental in order to examine the impact of the various elements of the rhetoric (both separately and in combination). For now, however, our analysis helps us to better understand the toxic character of Trump's 6th January speech and to see how it functioned not as an order but as a warrant for the subsequent violence that unfolded on Capitol Hill.

AU T HOR C ON T R I BU T IONS Evangelos Ntontis:
Conceptualization; writing -original draft; methodology; writing -review and editing; formal analysis; project administration; data curation. Klara Jurstakova: Conceptualization; writing -original draft; methodology; writing -review and editing; formal analysis; data curation. Fergus Neville: Conceptualization; methodology; writing -review and editing; formal analysis. S. Alexander Haslam: Conceptualization; methodology; writing -review and editing; formal analysis; writing -original draft. Stephen Reicher: Writing -original draft; conceptualization; methodology; writing -review and editing; formal analysis.

CON F L IC T OF I N T ER E S T S TAT E M E N T
EN sits on the editorial board of BJSP. However, he had no involvement in the processing of this paper. The other authors have no conflict of interests to declare.

DATA AVA I L A BI L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analyzed in this study.