The power of Trump-speak: populist crisis narratives and ontological security

Abstract For most observers, the election of Donald Trump as the 45th president of the United States (US) came as a shock. This has been widely recast as the culmination of the American public’s long-standing dissatisfaction with the political elite and deep-seated frustrations with broader socio-economic conditions. We argue that the Trump campaign’s success also stemmed from its effective use of an emotionally charged, anti-establishment crisis narrative. With insights from political psychology, we examine the socio-linguistic mechanisms that underlie the effectiveness of ‘Trump-speak’ through both quantitative and qualitative content analysis of Trump’s communications toolkit during the 2016 US presidential election campaign. We show that his leadership legitimation claims rest significantly upon ‘crisis talk’ that puts his audience in a loss frame with nothing to lose and explain why ‘crisis talk’ impacts on political behaviour. As we demonstrate, the crisis stories that political agents tell simultaneously instil ontological insecurity among the American public and serve to transform their anxiety into confidence that the narrator’s policy agendas are the route back to ‘normality’. Through these rhetorical mechanisms, the Trump campaign manipulated individuals’ ontological (in)security as a tool in the politics of reassurance at the broader, societal level.


Introduction 1
In November 2016, the American people elected Donald Trump, an eccentric businessman and TV-show host, as the 45th president of the United States (US). To many commentators, the steep rise of an outsider with no experience in politics came as a shock. This has since been recast as the culmination of white voters' long-standing dissatisfaction with the political elite and their This article has been republished with minor changes. These changes do not impact the academic content of the article. deep-seated frustrations with broader social and economic conditions: recent research links Trump's electoral success more explicitly to nativism (Young 2017), racism and xenophobia (Lopez 2016;Pew Research Centre 2016) and economic populism (Oliver and Rahn 2016).
In this paper, we provide an alternative account of the President's popular appeal which revolves around the way he speaks. Expanding upon recent studies on Trump's communication practices-such as those that frame his communication style as comedic and entertaining (Hall et al 2016), polarizing (Sclafani 2018), anti-intellectual (Degani 2016) and conversational (Golshan 2016;Liberman 2016a; 2016b)-we look at how the lure of 'Trump-speak' is rooted in the way it triggers voters' cognitive biases. Specifically, we connect Laing's (1960) and Giddens's (1991) notion of ontological security with Tversky and Kahneman's (1990) work on 'loss frames' and prospect theory to establish a conceptual connection between Trump's crisis talk, individuals' need for ontological security, and voting behaviour that favours strongmen outsider candidates.
Members of the policymaking community have long acknowledged that rhetoric is not simply bluster but is useful in the pursuit, exercise and retention of political power. To paraphrase Alasdair MacIntyre (1981,201), political agents recognize that in both their actions and their practices they are storytelling animals. Often, the better story, not the more rational argument, determines electoral success (Polletta 2006;Marcus 2002;Lakoff 2016). We may not intuitively equate Trump's 'idiosyncratic and impromptu style of speaking' (Thompson 2017) with finely calibrated, politically impactful rhetoric. Yet if we look at the 'grip' of Trump's communication style more closely through a socio-cognitive lens, a picture emerges not only of clear linguistic patterns, but also of a decidedly populist crisis rhetoric that relies on the activation of cognitive biases to generate electoral support (see Chilton 2017).
Scholars within cognate fields, including those studying populist movements (for example, Laclau 2005; Mudde and Kaltwasser 2017; Jansen 2011), have widely discussed the importance of language in popular mobilization. Those studying international relations (IR), however, have thus far shown little interest in unravelling the psycho-linguistic mechanisms that foster both the rise of nationalist strongmen like Donald Trump and the political parties that support them in their proclaimed endeavour to return to a better past (no matter whether one existed or not). This is surprising because, over the past two decades, the scope of IR has expanded to include approaches that emphasize meaning as unstable and contested, which matters for understanding and explaining international politics. These approaches have stimulated a conceptual interest in everyday actions that do not necessarily involve conscious, representational knowledge. These include habits, rituals, emotions, symbolic structures and popular culture-all of which used to be considered trivial, mundane and irrelevant to IR. Work on 'the unnoticed' (Hviid Jacobsen 2009, 2) thus potently critiques the tendency of mainstream IR scholars to prioritize powerful political agents and emphasizes not only that agency is varied and complex, but also that the choices we make in our everyday lives play a part in shaping world politics (Hobson and Seabrooke 2007;also Acuto 2014;Kessler and Guillaume 2012).
Scholars have recently imported the concept of ontological security-the 'security-of-Being', defined by Mitzen (2006, 342) as 'the need to experience oneself as a whole, continuous person in time'-into IR. Prominently associated with the works of Robert D Laing and Anthony Giddens, ontological security has a long history in psychology and sociology research and is increasingly used in IR to further underscore the importance of subconscious and routine-centric explanations within the discipline. Such works include the reproduction of security dilemmas between states (Mitzen 2006) and states' autobiographical continuity in interstate conflict (Suboti c 2016); the role of emotions in securitization processes (van Rythoven 2016); the impact of narrative reconstructions of traumas and collective memory (Kinnvall 2004;; and the radicalization of individuals in the era of global terrorism (Croft 2012). In this article, we argue that the concept of the security-of-Being can also help us understand how 'Trump-speak' enabled a political novice like Trump to attract nearly 63 million votes-and how he continues to maintain his appeal. Importantly, the concept of ontological security also provides a useful entry point into detangling the complex relationship between crisis-centric political rhetoric and the politics of reassurance that favours both simplistic policy solutions and the populist strongmen that articulate them.
The paper proceeds as follows. The first section analyses the speeches given by Donald Trump during the 2016 US presidential election campaign between 16 June 2015, when he announced his run for presidency, and election day on 9 November 2016. 2 It draws findings from a computer-assisted, quantitative corpus-linguistic analysis supported by a qualitative analysis to increase the level of detail in uncovering 'Trump-speak' features. As this empirical investigation of Trump's political rhetoric reveals, the language used by candidate Trump during his campaign should be understood as a rhetorical strategy-intentional or not-that helped him control meaning and mobilize the public via 'crisis talk'. The paper's second part will discuss our findings through the conceptual lens of ontological security and apply insights from political psychology and prospect theory to show how crisis talk enables someone like Trump to target individuals' feelings of insecurity and loss in order to, paradoxically, create public support for a risky outsider candidate with no expertise in the workings of the American government. We argue that the way in which 'Trump-speak' discursively constructed a broader affective crisis setting (Solomon 2014) during the 2016 presidential election amplified voters' negative images of themselves, the country they live in and the challenges they faceall of which contributed to Trump's electoral success.
Research on ontological security in IR has thus far ignored the possibility that political agents may actively target individuals' drive towards the security-of-Being for political purposes-and their efforts may include discursively constructing crisis for electoral gain. Rather, as Rossdale (2015) pointed out, the IR ontological security literature has focused predominantly on biographical continuity and stable narratives of selfhood, largely obscuring power relations and politics inherent in the security-of-Being. By putting a psycho-linguistic spotlight on the communications toolkit that Donald Trump used during his 2016 presidential campaign, this article brings politics into IR ontological security research. It also extends the field of IR beyond its descriptive path towards an understanding of how language impacts on political behaviour at both the individual and societal levels. As we show, the crisis stories political agents tell simultaneously instil ontological insecurity within the public and transform this anxiety into confidence in the narrator's policy agenda as the effective route back to normalcy. Manipulating individuals' security-of-Being becomes a populist tool in the politics of reassurance at the broader societal level.
The building blocks of ' Trump-speak' As Street (1997, 60) pointed out, political agents are actively 'engaged in creating works of popular fiction which portray credible worlds that resonate with people's experiences'. In the following section, we focus on the world Donald Trump discursively constructed as Republican nominee during the 2016 US presidential election. Combining both quantitative and qualitative elements, we show that his main rhetorical building blocks involved creating and fuelling an imaginary 'Crisis America' and identifying the culprits responsible for America's descent. Our analysis also reveals a discursive legitimation strategy that centrally features black-and-white rhetoric that divides Trump's world into contradistinctive notions of other, we and I.

Constructing 'Crisis America'
Our corpus of the Republican presidential nominee's campaign rhetoric-a computerized collection of texts-comprises 74 speeches given by Donald Trump between 21 March 2016 and 9 November 2016, as well as the announcement of his candidacy on 16 June 2015. With a word total of 217,325, this corpus is comparatively small but representative of Trump's campaign speeches. We explore its 'aboutness' through quantitative linguistic analysis, which includes generating word frequency counts that allow us to explore salient occurrences and collocations in more detail. Uncovering statistical regularities and relations in Trump's campaign speech language (that is, what is quantitatively distinctive) is a first but important step in revealing how 'Trump-speak' fosters audiences' 'psychological perception of subject matter' (Phillips 1989, 7). It allows initial speculation about his discursive leadership style.
Quantitatively determining the level of thematic concentration (TC) contained in a corpus is one way to gain insights into Trump's discursive legitimation strategy. This method mathematically captures the manifestation of the 'effort to communicate some topic(s) more intensively than other topics, or-importantly-more intensively than would be expected from "neutral" language' ( Cech et al 2015, 216). A higher TC generally indicates a communication practice that reduces complexity and problems to simple, ideologically charged sound-bites-and is associated with undemocratic discursive leadership styles ( Cech 2014). Wang and Liu's (2017) study of Donald Trump's political discourse during the 2016 election showed that the Republican candidate had a significant TC level in his campaign speeches, which was higher than that of both his presidential predecessor and the democratic nominee, placing his rhetoric in the strongman corner of TC.
Ranking content nouns and proper names based on their frequency is another important step in the study of rhetoric features; it provides a more detailed insight into what themes (lexis) centrally featured in Trump's campaign speech repertoire.
Our analysis shows that Trump addressed the 'people' most prominently in his speeches, followed by 'country'. This is illustrated in Table 1, which lists the 15 most frequent content words and proper names of the Trump Campaign Corpus. The high frequencies of 'country', 'American', 'America', 'United' and 'States' are quantitative testament to the extraordinary discursive position that the American nation and populace occupied in his speeches.
Certainly, rhetorical identification with America, often as an exceptional territory of the imagination constructed retrospectively from the past (Levinger and Lytle 2001;Taggart 2006), is not uncommon for US presidential candidates (for example, Tyrrell 1991). 'Trump-speak' stands out because of its 'minimalist triggering' of a populist narrative by glorifying the nation's past (Chilton 2017) while characterizing the country as in a state of decline. Indeed, although Trump used 'America' over 2000 times in his campaign speeches, the context in which he did so shows that he presents the US not as currently extraordinary-as was common among his predecessors-but rather as a project that must be rebuilt according to the blueprint of the (imagined) past. In Trump's words, 'America is going to be strong again. America is going to be a reliable friend and ally again … we must make America great again' (Trump,27 April 2016, Washington, DC, emphasis added). Paraphrasing Taggart (2006), Trump projected a romanticized and ahistorical (but past-derived) vision of the country onto the present to show that something essential to American existence had been lost-but could be regained.
The textual context in which the mentions of 'country', 'American', 'America', 'United' and 'States' occur underscores that Trump's communicative strategy painted a rather grim picture of the US. Figure 1, which groups all word forms that occur at least 50 times in the corpus together with the words with which they predominantly co-occur, graphically illustrates that the word clusters 3 surrounding 'American' and 'country' centrally featured the interrelated themes of crime and violence, killing jobs, and poverty, as well as illegal immigration and drugs, Islamic terrorism, trade and infrastructure, respectively. Figure 1 directly references crises across six thematic fields: security, government, drugs, economy, education and environment.
Trump's characterization of America as in a state of crisis both echoed and amplified a broader public sentiment. Since 2009, the Public Religion Research Key sources of personal anxiety cited in the survey (PRRI 2015) were health care (63 per cent), terrorism (62 per cent) and jobs and unemployment (60 per cent). There was also less tolerance towards immigrants who do not speak English, a particularly major concern for white working-class Americans (63 per cent). These worries comingle with a bleak outlook on both the economy and America's future: 72 per cent of respondents believed that the country was still in a recession while 49 per cent supposed that America's best days lay in its past, white mainline Protestants being markedly more pessimistic than other groups (60 per cent). Fiftythree per cent of respondents also believed that the American culture and way of life had changed for the worse since the 1950s-a view more pronounced among white Americans (57 per cent). In turn, respondents showed little faith in the idea that a candidate clearly aligned with America's governing elite would do much to alleviate these concerns. Indeed, 54 per cent agreed that electing another president from the Bush or Clinton dynasties would be bad for the country; this view was particularly resonant with those who identified as Republican (61 per cent) and Tea Party members (69 per cent). Just after Trump entered the 2016 presidential race, Americans were anxious and depressed, white working-class Americans being the most anxious and depressed group surveyed.
Although it is outside the scope of this study to investigate whether there was a direct causal line between Trump's entrance into the presidential race and a negative shift in the opinions and attitudes surveyed by the PRRI, there is little room to assume that his rhetoric alleviated this sense of gloom and doom. A comparison of the PRRI survey results before and after Trump's candidacy announcement suggests that Americans felt more troubled after the start of his campaign-although this has limited explanatory power (detailed in PRRI 2015; PRRI 2016). As Figure 2 illustrates, worries about personal economic security rose after Donald Trump announced his candidacy and did not decline again during the 2016 election cycle.
This contrasts with much of the economic reality at the time: during the 2016 presidential election campaign, the US economy showed significant signs of recovery from the Great Recession. Stocktaking in August 2016 suggested that the Obama presidency's long streak of job creation was uninterruptedseeing the 77th consecutive month of private-sector job creation and the 71st consecutive month of overall job growth-white men generally fairing best in terms of both employment and income (US Census Bureau 2017; Kasperkevic and Diehm 2016). Using data from the Bureau of Labour Statistics, the New York Times's last economic snapshot before the 2016 presidential election also showed a healthy job growth rate; an unemployment rate as low as it was in February 2008, before the financial crisis; and the strongest labour market since the Great Recession (Cohen 2016). At the same time, US Census Bureau (2017) Figure 2. How important are the following issues for you personally: jobs and unemployment (PRRI 2015versus PRRI 2016 data shows that poverty had gone down significantly by 2016, dipping to only 0.2 per cent above the 2007 level, prior to the Great Recession. The consumer price index adjusted median and mean incomes for those classified as either white or white non-Hispanic rose above pre-2008 levels-although they remained below the 1999 peak-with increases most significant in the south and west of the US (US Census Bureau 2017). Nevertheless, as Turchin (2016, 248) observed, contemporary America is experiencing 'intensifying structural-demographic pressures for instability' almost as high as the stress level that triggered the American Civil War, and this should not be dismissed as insignificant.
The culprits of 'Crisis America' To conjure a picture of American 'carnage', Donald Trump relied upon a language of division. His presidential campaign speeches constructed vivid images of who gets to belong and who gets to be excluded, of who is strong and who is weak and of who is a winner and who is a loser. The 'others' that Trump consistently framed both as responsible for America's state of crisis and as diametrically opposed to Trump, himself and the American people consist of two sets of social actors: (1) the weak, self-serving and detached 'establishment'; and (2) foreigners who either flow into the country or lead countries that want to exploit the US for economic and security reasons. Donald Trump persistently criticized the failures of America's elites-often represented by Hillary Clinton in his speeches-by negatively portraying their members as not smart, stupid or losers. To distance himself from the failures of American's elites, Trump pointed out that, while he once 'used to be establishment … when I decided to run, I became very anti-establishment' (quoted in Washington Post 2016). Trump generally underscored his own political outsider status by foregrounding his business credentials-based on language considered as outside the norms of 'political correctness'-to set himself apart as a leader who will not budge easily. In contrast, Trump firmly pushed his democratic opponent into the establishment corner, rhetorically identifying Hillary Clinton as someone who embodied precisely those institutions and political agents that were responsible for leading America into a state of crisis. According to Trump, Clinton was 'weak, ineffective, pandering' and embodied 'the failed status quo' (Virginia Beach, VA, 11 July 2016; Charlotte, NC, 18 August 2016).
Trump, of course, makes many references to the name of his opponent, Hillary Clinton. Table 1 shows, for example, that both 'Hillary' and 'Clinton' occur over 1000 times in the corpus. Counterintuitively, though, the lexis of Trump's campaign speeches contains little evidence of his otherwise public display of sexism and misogyny apart from gender-laden references, like characterizing Clinton as weak, lying and neither fit nor strong enough to command. Rather, as Figure 1 shows, Trump connects Clinton's name to a number of negatively connoted words-'lied', 'corrupt establishment', 'disaster', 'trillion dollars debt', etc-all of which serve to signal his opponent's moral bankruptcy. In Trump's eyes, Hillary Clinton was the most corrupt person ever to seek the presidency (for example, Bangor, MA, 15 October 2016; Delaware, OH, 20 October 2016; Green Bay, WI, 17 October 2016; Orlando, FL, 2 November 2016; Raleigh, NC, 7 November 2016). While his own indifference to facts is well established, 5 Trump draws a stark division between himself and the democratic nominee on the basis of truth aversion. For example, he declared that, rather than lying, his problem was that 'I can be too honest, [while] Hillary Clinton is the exact opposite: she never tells the truth' (Charlotte, NC, 18 August 2016). And Trump's critique of the democratic candidate does not stop at creating an image of Hillary Clinton as 'crooked'. He also prominently claimed that she would both pose a threat to the US because of her lax attitude towards immigration (for example, Jackson, MI, 24 August 2016; Gettysburg, PA, 22 October 2016), and would only make things worse if she was elected (see, for example, Phoenix, AZ, 31 August 2016).
The absence of directly sexist and misogynist words in the corpus of Trump's campaign speeches does not suggest that such words did not matter in his presidential bid. On the contrary, Trump's vulgar and demeaning attacks on female candidates on the campaign trail-alongside his attempts to stoke male anxieties-showcase his tendency to play upon gender stereotypes for electoral support (see, for example, Katz 2016). Indeed, recent research found that Trump gained an unexpectedly large share of the popular vote both because voters-particularly white, working-class men (Francis 2018)perceived Clinton as more objectionable than her Republican counterpart and because Clinton significantly underperformed Obama among most demographic groups (Weinschenk 2018). Our analysis suggests that 'Trump-speak' actively fostered widespread dislike for Hillary Clinton by describing her simultaneously as weak and dangerous along with characterizing her as lying and corrupt. By chanting 'lock her up' (and worse) and wearing T-shirts with the slogan of 'Trump that Bitch', Trump's rally audience signalled their receptiveness to his framing of the Democratic candidate.
While studies have found that holding an anti-Hillary position was an important predictor of Trump support in the 2016 presidential election-in addition to hostile sexist attitudes (Schaffner et al 2017;Wayne 2016), racism (McElwee andMcDaniel 2017) and Christian nationalist ideology (Whitehead et al 2018), for example-many of Trump's diverse supporters also showed hostility towards immigrants and Muslims (Ekins 2017). Indeed, Trump's campaign speeches heavily implicated this second category of 'others'-foreigners-in the creation of America's current state of misfortune. In much of Trump's rhetoric, foreigners appear either as the 'strong' leaders of countries that have helped to push America into economic decline or as 'bad' individuals who flow into the country, threatening Americans' jobs and physical safety.
The way that Trump characterized (potential) immigrants condensed foreigners into concrete groups to be feared-a key marker of populist rhetoric (Wodak 2015). Alongside his constant repetition of divisive phrases, harsh words and violent imagery more broadly, Trump framed newcomers to the US as 'criminals', 'rapists', 'killers', 'snakes' and 'terrorists' who like 'chopping off heads' (for example, Trump 2016b; also Phoenix, AZ, 31 August 2016; Manchester, NH, 13 June 2016). In Trump's campaign stories of America in crisis, border-crossing foreigners served as culprits behind America's economic deterioration, as well as its loss of individual security and social cohesion, fuelling anxieties about difference among his audience.
Trump also painted a picture of an internationally defeated America caused by foreigners' actions. Across his campaign speeches, Trump consistently condemned other countries for what they do to us-for 'killing us', 'beating us', 'laughing at us', 'taking our money', 'lending our money back to us' and for 'sending people to us'. Without an intervention, Trump claims, the state of the US will continue to deteriorate because America's 'enemies are getting stronger and stronger [while] we, as a country, are getting weaker … We don't have victories anymore' (New York City, 16 June 2015).
In Trump's story about America's decline, countries like Russia and China no longer respect the US and America's allies are not 'paying their fair share' (Henderson, NV, 5 October 2016; Trump 2016a). Even as Trump's crisis narrative implicated the broader US establishment in the loss of US standing in the world, Trump signalled respect for the nationalist strongmen whom he framed as key contributors to America's decline. For example, Trump expressed a 'love' for China, pointing out specifically that 'their leaders are much smarter than our leader' (New York City, 16 June 2015); he also declared that 'Putin has much better leadership qualities than Obama, but who doesn't know that?' (Doral, FL, 27 July 2016). The Republican presidential nominee thus characterized the ability of foreign countries to swindle America as a strong leadership trait that the US political elite lacked. Prominent in Trump's crisis talk, then, was his claim that America had forfeited-and the US government, as he puts it, had actively given away-money, jobs and influence. In the process, the US had lost esteem around the world. 'Our' country, Trump claimed, was now in 'serious trouble' (New York City, 16 June 2015).
The above suggests that Trump organized his messages along a firm rhetorical line between 'us' and 'them'. Specifically, Trump's attack on internal and external others relied on three key rhetorical tactics: (1) name-calling based on moral corruption, weakness and racism ('Low Energy Jeb', 'Little Marco', 'Lyin' Ted', 'Crooked Hillary' and 'Pocahontas)'; (2) using attributes that put the character of political opponents in a bad light (stupid, horrible, bad); and (3) labelling the actions of political opponents as horrific and destructive (chopping off heads, rapists).

Trump as first among equals
Much of Trump's antagonistic 'black-and-white' campaign focused on establishing otherness through a rhetorical frontier between the 'real' American people and their enemies. He drove a discursive wedge between 'our people' and those who do not belong (Canovan 1999, 5). At the same time, Trump's leadership legitimation strategy centrally featured the 'people'-an idealist conception of the community he claimed to both represent and serve-as a fundamental focus of his policy agenda. While this aspect of Trump's campaign rhetoric oriented 'Trump-speak' towards populism (Taggart 2006;Norris and Inglehart 2018), it also served as a discursive springboard that Trump could use to present himself as the only electoral choice who put the 'people' first and could reverse the status quo of 'Crisis America'.
As Figure 1 shows, 'we-ness'-the country, people, I-is generally characterized by positive attributions such as 'great', 'amazing', 'working', 'love' and 'beautiful'. Similar to otherness, 'we-ness' is also split into two interlinked subsets. Although each subset prominently features a notion of 'we' in which 'Trump' and 'I' take centre stage, one presents Trump as an integral part of 'you', the American people, while the other characterizes Trump as a core element of the solution-as part of the winning team. Trump is with the audience, and the audience is with him.
As we have shown above, Trump characterized America as a state in crisis-as dead, in 'serious trouble', a 'third world country' and a 'dumping ground for everyone else's problems' (for example, Erie, PA, 12 August 2016; Prescott Valley, AZ, 4 October 2016; Newtown, PA, 21 October 2016). To locate himself among the 'we' and 'us' who had previously lost out, Trump used several rhetorical devices in his campaign speeches. For example, Trump told his audience that 'we live in fear in our churches and our synagogues we live in fear that we're going to lose our tax-exempt status if we say anything that's even slightly political' (New York City, 16 July 2016, our emphasis). Again, Trump (Cleveland, OH, 21 July 2016, our emphasis) made clear that 'the problems we face now, poverty and violence at home war and destruction abroad, will last only as long as we continue relying on the same politicians who created them in the first place'. By framing the establishment-other as the culprit behind the sorry state of the US, Trump, in emotionalized populist fashion, simultaneously absolves 'us' from any responsibility for 'Crisis America' (Hameleers et al 2016, 872). Bourdieu (1991, 69) pointed both to the significance of political agents' ability to strategically adapt their language to their audiences' social field and to the symbolic power that agents hold when they can adapt their language 'without appearing to be ignorant or incapable of satisfying [their] demands' despite their distance. In addition, rhetoric that mimics 'we, the people'-what  has described as 'public language'-is characterized by the absence of the concrete; 'it is a language of implicit meaning' (Bernstein 1971b, 29). By creating a 'we-ness' that discursively establishes him as primus inter pares among 'the people', Trump forged a rhetorical bond with Americans over worries that the country was disintegrating and radically changing, the economy was deteriorating and ferocious enemies and minorities were growing emboldened. In forging this bond, Trump identified with his audience, who in turn recognized themselves in him (see Laclau 2005; Levinger 2017). Trump's way of speaking renders his actual, contrasting background of advantage, privilege and wealth invisible. The American people are his tribe, and he is reassuring them that he is destined to protect them from internal and external outsider threats while guiding a crisis-stricken US towards a 'New American Future' (Asheville, NC, 12 September 2016).
In addition to situating himself among the people, Trump also rhetorically positioned himself as the only one within the 'we' that can lead the people out of their current misery. He emphasized that changing the course of the country was a joint effort, proclaiming that 'together we will lead our party back to the white house and we will lead our country back to safety prosperity and peace; we will be a country of generosity and warmth' (Cleveland, OH, 21 July 2016, our emphasis). Trump also created differences within the 'we' to situate himself clearly both outside the existing governing elite and above the people by following a two-fold rhetorical strategy. On one hand, Trump signalled his departure from the establishment and the status quo, noting that 'we need new leadership; we need new thinking; we need strength; we need in our country law and order; and if I'm elected president that will happen' (New York City, 16 July 2016). On the other hand, Trump made what many commentators waved off as delusions of grandeur (see, for example, Lancer 2016) an integral part of his leadership legitimation strategy towards audiences that held pre-existing preferences for candidates who signal strength and dominance (see Lakoff 2016); he rhetorically paints himself as someone with fantastic knowledge, high intelligence and the ability to think critically. Trump believes his superiority is self-evident: 'I understand things. I comprehend very well, better than I think almost anybody' (quoted in The Week 2017), 'I understand the tax laws better than almost anyone, which is why I'm the one who can truly fix them' (Pueblo, CO, 3 October 2016) and 'I know more about renewables than any human being on Earth' (quoted in Fox News 2016). When it comes to building a border fence to keep out unwanted immigrants, Trump declared that 'nobody in the history of this country has ever known so much about infrastructure as Donald Trump. I build infrastructure. Do I know how to build a wall? Do I know how to build infrastructure' (quoted in CNN 2016). By representing himself as someone who knows best and who is a sensor of cognitive processes, Trump puts another wedge between himself and the self-serving, elitist out-group: Trump knows, thinks, and assumes; they want and need, and they don't know, and don't have the answer. Shifting between 'nobody knows' references and a 'trust-me-and-onlyme' rhetoric not only emphasizes Trump's cognitive abilities, but also sets him apart-as the leader of his flock-from his discursively constructed in-group. 6 Overall, the empirical analysis of Trump's campaign rhetoric gives his crisis talk a decidedly populist orientation, even if it is far from what Laclau (2005) understands as 'populist mobilization'. 'Trump-speak' mobilized elements of a 'people's ideology' by (1) creating an imaginary community within the American territory based on radical exclusion and antagonistic frontiers; (2) fostering a belief in the possibility that this imagined community can return to sovereignty; and (3) expressing strong scepticism vis-a-vis the political elites who are suspected of betraying the people's will and fuelling the strength of 'others' (Anderson 1983;M eny and Surel 2000, 177-222;297-312).
For anyone following Trump's political spectacle, it may seem intuitive to associate Trump's communicative strategy with populism. As our above analysis confirmed, well-known populist contradistinctions both between the people and the elite and between real Americans and others-as well as Trump's representation of himself as the only candidate capable of returning America to its past glory in the name of the 'people'-are integral to his crisis-centric campaign speeches. If one views 'Trump-speak' through the lens of ontological security, however, its effectiveness stems less from the populist orientation of his campaign speeches. Rather, as we argue below, it stems primarily from the Trump who pushed the image of American carnage onto his audiences. Through his imaginary 'Crisis America', Trump created a critical situation in the socio-cognitive sense-that is, 'a set of circumstances which-for whatever reason-radically disrupts accustomed routines of daily life (Giddens 1979, 124).

Ontological security in the context of populist crisis narratives
The terminology of ontological security was coined by the psychiatrist RD Laing (1960) in his book The divided self, which drew attention to the social dimension of mental health. Laing underscored the importance of someone having a sense of presence in the world, describing such a person as a 'real, alive, whole, and, in a temporal sense, a continuous person' (Laing 1960, 39). Being ontologically secure allows us to 'encounter all the hazards of life, social, ethical, spiritual, biological' with a firm sense of both our own and others' reality and identity. However, ontological security only prevails in the absence of anxiety and danger (Laing 1960, 39, 41).
Three decades later, the sociologist Anthony Giddens (1991) picked up Laing's concept of the security-of-Being in his enquiry into Modernity and selfidentity, which also established the conceptual groundwork for much contemporary scholarship on ontological security. Giddens stressed that our identity is defined primarily by our capacity to sustain a particular biographical narrative that must be able to 'continually integrate events which occur in the external world, and sort them into the ongoing "story" about the self' (Giddens 1991, 54). In the absence of our own historical contingency, we require 'a sense of continuity and order in events' to be ontologically secure (Giddens 1991, 243). To be able to 'keep on with everyday activity', we need to be able to trust that we-as well as our environment-will remain constant, stable and predictable (Giddens 1991, 37, 92).
Crisis narratives can remove this sense of stability by disrupting our ability to insulate our environment and ourselves. Indeed, widespread perceptions of ruptures and instability can temporarily strip individuals of their presumed biographical continuity, sparking episodes of anxiety and ontological insecurity (see, for example, Chernobrov 2016). Although such feelings of insecurity negatively impact upon individuals' ability to accept fluid conditions and an unknown future, they also motivate individuals to foster cognitive and behavioural certainty through routinized behaviour and interpretations (Mitzen, 2006, 342;Marshall 2014, 155-170). People thus make sense of what happens, why it happens and to whom it happens by reverting to familiar behaviours. While rhetorical choices can help individuals navigate their changing and continuous autobiography-thereby ensuring their self-identity is stable (Suboti c 2016)-individuals can also create a substantial status quo bias towards deep-rooted ideational casings if they emphasize crises. Discursive reverberations of the past should thus be understood as historical imprints on sense-making. The narrated past acts as a reference point for ontological (in)security that influences political behaviour in the present.
Crises also increase the need for cognitive closure: our psychological discomfort with ambiguity and desire for predictability, order and definite conclusions-all of which help us feel more in control (for example, Webster and Kruglanski 1994). Because crisis settings, which are inherently uncertain, challenge individuals' ability to continue with their everyday lives, they increase our need to restore our sense of ontological security. Rather than depriving individuals of agency (see Mitzen 2006, 342, 345), crisis situations push us to seek action and resolution to escape a state of insecurity-of-Being, often without us carefully considering the root of the problem. Crisis settings act as broader affective environments in which individuals urgently seek to (re-) establish a secure sense of self; since individuals are not divorced from their wider cultural-affective milieu, crisis settings impact on individuals' electoral decisions (Solomon 2018).
The influence of crisis settings on our electoral decisions suggests that our drive to restore ontological security in a crisis situation is susceptible to promises of restoring what once was-to a retrospective recasting of the past as stable, predictable and comforting. But this is only part of the story. As our empirical analysis of Donald Trump's campaign speeches shows, political rhetoric often encourages individuals to interpret a situation as a crisis by using terminology that evokes images of defeat, costs and unfairness. While this interpretation heightens the sense of ontological insecurity within individuals, it also creates a 'loss frame' (Tversky and Kahneman 1981), which paradoxically fosters preferences for both more unconventional, riskier policy options and outsider political candidates. Political rhetoric influences attitudes towards risk-taking.
Prospect theory (for example, Kahneman 1981, 1990;Quattrone and Tversky 1988) suggests that this cognitive bias is particularly prevalent when people face two choices: one with little risk and one with more risk. While people tend to be risk averse in their political behaviour, they become more risk-seeking when they perceive that they have already been left behind or that their country has taken a turn for the worse. If the choices in this context include an option that continues this negative trend rather than a radical change of course, people are more willing to choose a riskier option in order to prevent further loss. Putting people in a loss frame-for example, by nurturing perceptions of crisis-thus increases their risk-taking inclination and fosters preferences for outsider leadership contenders-even if the candidates are a political lottery with unspecified policy platforms. In Trump's own words: 'what do you have to lose? It can't get any worse' (Washington, DC, 9 September 2016;see also Asheville, NC, 12 September 2016).
Perhaps unsurprisingly, recent works have found that a feeling of deprivation-the belief that something was not merely lost but taken away-is associated with sympathy for populism (Pettigrew 2017;Marchlewska et al 2017). But creating loss frames by constructing political crises, disenchantment and unease is a well-known element of broader discursive leadership legitimation practices (Schnurr et al 2015). Relevant linguistics research suggests, for example, not only that legitimation strategies become particularly prevalent in crisis settings (van Dijk 1998, 257-262), but that political agents construct crisis situations in order to legitimize and authenticate their claims to leadership (Kerr 2008, 202). Much of the populist appeal of nationalist strongmen like Trump is their ability to intertwine discursive legitimation strategies built around crises and loss with affective stories of a past filled with national greatness and biographical stability; it is the source of the strongmen's narrative attraction and 'soft power' (Solomon 2014;Watanabe and McConnell 2008;Ahmed 2004). At the socio-cognitive level, a crisis-centric, populist-leaning communicative strategy is built around what is sometimes referred to as a 'nostalgia effect' (Steenvoorden and Harteveld 2017). A sentimental longing for a positive past is a near-universal phenomenon-even if one that cannot plausibly recur-that is often characterized by feelings of both sadness and joy (for example, Sedikides et al 2004). For those in perceived crisis situations, however, this nostalgia effect leads them to seek out what they imagine to be a better pastthat is, to alleviate their generalized anxiety and restore their ontological security. Generally, they recall a comforting past (complete with hindsight bias and rosy retrospection) to make the present reassuring and restore notions of belonging, inclusion and continuity. 7 Such restorative (rather than reflective) nostalgia is characterized by a pathological drive to reclaim the imagined past by reinstating 'a particular vision of a neglected, forgotten or defeated set of cultural or social arrangements' (Kenny 2017, 262). 'Usable pasts' (Anderson 1983) and 'chosen' episodes of triumph and loss (Volkan 1988) are interlinked with the present to cater to an individual's need for stability and to allow each individual to manage their anxiety and ontological insecurity in crisis settings. Because individuals are predisposed to rhetorical appeals using the nostalgia effect, they privilege those political agents and policy agendas that can discursively (re)establish notions of familiarity and control. These political agents provide their audiences with 'ontological security in the present' (Kinnvall 2014, 322) by promising a return to the past instead of change and instabilityand by reassuring their audiences of their place in the world.
Perceptions of crisis also play a major role in activating authoritarian inclinations that would otherwise remain insignificant as motivations for individuals' political behaviour. Here, authoritarianism 8 is a predisposition or psychological profile characterized by leanings towards obedience and conformity (oneness and sameness) combined with traits of 'moral absolutism, intolerance and punitiveness towards dissidents and deviants, [and] racial and ethnic prejudice' (Stenner 2009, 142;also Feldman and Stenner 1997). While scholars continue to debate the underlying causes of authoritarianism, research over the past two decades has suggested that authoritarians generally convey deeper fears and display less tolerance for diversity and ambiguity than the rest of the electorate; they tend to see the world in black-and-white terms (Altemeyer 1996;Napier and Jost 2009;Hetherington and Suhay 2011). An authoritarian predisposition chronically inclines individuals to hold and express negative attitudes towards 'outgroups that seem to threaten collective security' (Duckitt and Sibley 2009, 103) and to support, in turn, attitudes and behaviours that 'enhance sameness and minimize diversity of people, beliefs, and behaviors' (Stenner 2005, 19). In constructing his 'Crisis America', Trump tapped into rhetoric linked to both antagonism towards 7 On the significant role of emotion in the generation of ontological (in)security, see contributions by Subotic and M€ alksoo in this special issue. 8 Authoritarianism is often assessed through four child-rearing questions that ask respondents to choose between two options each: is it more important for a child to be respectful or independent, obedient or self-reliant, well-behaved or considerate, and well-mannered or curious? Survey respondents who pick the first option in each of these pairs are strong authoritarians. These questions first appeared on the 1992 ANES survey and are widely used to estimate authoritarianism (Feldman and Stenner 1997;Hetherington and Suhay 2011). It is important to note that authoritarian inclinations, while more strongly associated with the political right, does not stop at party lines. out-groups and amity towards the in-group, directly targeting authoritarian voters' psychological predispositions. And Trump succeeded: in the 2016 election, authoritarian inclinations were a key predictor of Trump support (for example, MacWilliams 2016).
Heightened perceptions of fear and crisis most significantly affect the political behaviour of non-authoritarians, since such perceptions shift non-authoritarians away from their usual electoral preferences and towards nationalist strongmen. During normal times, individuals who feature comparatively low on the authoritarianism scale express different political preferences from those situated on the higher end of the scale. However, as Hetherington and Suhay (2011) show, these non-authoritarians are swayed towards 'authoritarian thinking' when they perceive themselves to be threatened (see also Hetherington and Weiler 2009;MacWilliams 2016). A sense of destabilizing social change-as well as threats to ontological and physical security-mobilize non-authoritarians to behave more like authoritarians by activating latent authoritarian predispositions (Stenner 2009); these perceptions foster widespread political behaviour aimed at reinstating normalcy and supporting the nationalist strongmen who promise to decisively deal with the problems of instability, crisis and social disintegration. Importantly, individuals do not need to hold significant authoritarian inclinations to express an electoral preference for an authoritarian leader-a self-styled strongman political agent like Trump. Rather, political agents can actively evoke such a preference by creating the perception that the world is chaotic, uncertain and alien and that major public institutions are failing to inspire confidence-all of which were integral to Trump's campaign rhetoric.
While we never make judgments in a discursive vacuum, the likelihood that we respond to cognitive cues contained in the message text increases when we are only indirectly or vicariously familiar with an issue. The less people have first-hand experiences of what political agents frame as threats-like foreigners, terrorism or the breakdown of the socio-political order-the more they tend to rely upon images of reality that are built from what they see, hear and read through their networks, through their communities, in media and through communication outlets (Bandura 2001, 271;Browning 2018). Ontological security is not immune to these framing effects.
As the Cambridge Analytica scandal has shown (see Persily 2017), attempts to manipulate political behaviour by targeting people based on their psychological predispositions has become a critical tool for political campaigning in the digital age. Our analysis suggests that Donald Trump's rhetorical strategy, built around populist-leaning crisis talk, may have tapped into many voters' everyday cognitive processes and predispositions to shift political behaviour in his favour. Our analysis implies that political agents choose how they discursively legitimate their leadership claims in order to manipulate individuals' psychological traits and fuel pre-existing socio-economic grievances for short-term political gain.

Conclusion
As Roland Barthes (1970, 16) once noted, 'Language is never innocent.' The rhetorical devices that political agents use on the campaign trail and the types of stories they tell can resonate with voters' emotions and subjective experiences in a range of consequential ways. 'Trump-speak' is no exception. While there are other significant aspects of the 45th president's campaign rhetoric that should be explored in future research-like the effects of false-speak, fearmongering, banter and repetition-we have focused on his exaggeration of America's problems, his anti-establishment critique and his division of the domestic and international sphere into warring camps-into zones of 'us' and 'them'-to examine how rhetoric can influence political behaviour at the sociocognitive level.
At the heart of 'Trump-speak' is a politics of reassurance, which relies upon a three-fold rhetorical strategy: it tells audiences what is wrong with the current state of affairs; it identifies the political agents that are responsible for putting individuals and the country in a state of loss and crisis; and it offers an abstract pathway through which people can restore past greatness by opting for a high-risk outsider candidate. For many Trump voters, rational arguments or detailed policy proposals pale in comparison with the emotive pull and self-affirmation of an us-versus-them crisis narrative, which creates a cognitive feedback loop between individuals' ontological insecurity, their preferences for restorative policy, and strongmen candidate options. In short, 'Trumpspeak' relies on creating the very ontological insecurity that it promises to eradicate for political gain.
Conceptually linking insights from prospect theory and ontological security, our analysis has also shown that Trump's communication style evokes a range of cognitive biases because of its crisis talk element. As Laclau (2005, 85) stated before the rise of Trump, 'without this initial break-down of something in the social order-however minimal that something could initially be-there is no possibility of antagonism, frontier, or, ultimately, "people"'. By building his rhetorical toolkit around augmenting existing grievances and emphasizing the prospect of further rupture and defeat, Trump generated ontological insecurity manifested simultaneously as a sense of loss and a desire for belonging. This widened the frame of political possibility in American elections, opening a pathway to success for an outsider candidate like Trump. Although the context behind which a significant segment of the American electorate perceived alienation, fragmentation and economic hardship limited how much audiences could be mobilized through linguistic means, Trump's populist crisis talk should still be considered an integral part of his electoral victory.