# The Dark Side of Politics: Why Struggles for Power Turn Disagreement into Hatred Politics organizes genuine conflicts about income, property, taxes, religion, customs, security, authority, and belonging. Without politics these disagreements would not disappear: private force, inherited hierarchy, or unilateral imposition would settle them instead. The paradox is that the activity through which societies negotiate competing interests can also convert differences into moral hostility. At a certain point citizens no longer say that opponents are mistaken; they come to regard their very presence as a threat to the country, the family, liberty, or human dignity itself. A distinction is essential between **ideological or issue polarization**, **affective polarization**, and **democratic breakdown**. The first concerns the distance between policy preferences and party programmes. The second concerns dislike, distrust, and exclusion directed towards people associated with another political group. The third appears when competition ceases to recognize elections, rights, institutional limits, or the opponent's legitimate status. A country can disagree intensely about public policy without turning that disagreement into persecution. Another can have parties whose programmes overlap while their supporters dislike each other intensely. A 2025 systematic review by Esteves and Junqueira of Brazilian research covering 2012–2022 found evidence of affective polarization, especially among the politically engaged, but not comparable evidence of widespread ideological polarization in that period. The darkest political behaviour does not require most people to become violent. Public incentives may reward humiliation; political leaders may cast differences as existential danger; trusted intermediaries may lose authority to arbitrate conflicts. The most severe expressions include aggression and attacks against constitutional government. Everyday manifestations include broken family ties, exclusion in workplaces, friendships ended over affiliations, and public delight in strangers' misfortune. These outcomes differ in severity and legal character, but identity, dehumanization, and symbolic punishment can link them. ## When an opponent becomes a type of person rather than a position Henri Tajfel and John Turner offered a durable explanatory framework through social identity theory: some of our self-esteem and orientation in the world comes from groups to which we belong. Categorizing people into “us” and “them” does not require hatred, but it can enable in-group favouritism. When political membership becomes a certificate of intelligence, virtue, or patriotism, criticism of a party may be experienced as an attack on the individual. Debating a tax proposal is then no longer simply discussing tax policy; it is defending personal dignity. Shanto Iyengar and Sean Westwood documented how partisan signals in the United States shaped judgments and behaviour outside formal political disputes. Subsequent research described affective polarization as a divide in sympathy and trust between identities, rather than simply distance between beliefs. Identities can also be built **negatively**. A person may know precisely whom they oppose but disagree with a significant share of positions held by those they support. Rejection of an adversary can mobilize more readily than a detailed debate about a public budget, institutional design, or necessary compromise. Motivated reasoning contributes to this process. Someone may scrutinize an unfavourable report for every possible flaw but accept a flattering claim about their political allies without comparable examination. The relevant psychological literature does not mean that people are universally incapable of reason. It shows that under particular conditions a desired conclusion affects the search for, weighting of, and memory of evidence. Education can strengthen critical examination; it can also provide sophisticated arguments for conclusions already embraced. Moralization further intensifies the divide. Jonathan Haidt draws attention to intuitive moral judgments involving care, fairness, loyalty, authority, and purity, among other values. His framework is influential but contested regarding measurement and cross-cultural generalization. Its more modest contribution is useful: arguments about income distribution or social customs become difficult to negotiate when each group presumes the opposing side possesses a morally repulsive intention. Concession becomes betrayal and listening becomes confused with endorsement. ## Power, fear, conflict, and the political tradition Machiavelli observed that politics combines conflict, reputation, fear, and power preservation. Reading *The Prince* alone as a celebration of ruthlessness would misrepresent his wider work: the *Discourses on Livy* also treated conflict among social interests as a component of republican life. The lesson is ambivalent. Leaders can exploit public passions, while well-designed institutions can channel rivalry into accountable public rules. Thomas Hobbes developed a theory of political authority from vulnerability, fear, and insecurity. His state of war is a conceptual construction rather than an empirical description of every human society. Yet a contemporary mechanism resembles his dilemma: when groups believe an electoral loss would render them defenseless before an enemy, they become more receptive to extraordinary measures promised as protection. Alexis de Tocqueville argued that democracy does not abolish pressures to conform. Majority opinion, the desire for acceptance, and social associations help determine what individuals feel able to say. His concern about majority tyranny helps explain how political communities can impose social costs on dissent even without direct coercion from the state. Karl Marx linked political contests to class, property, and the material organization of society. Max Weber emphasized struggles over legitimate domination, bureaucracy, and the tension between an ethic of conviction and an ethic of responsibility. Antonio Gramsci examined how authority is sustained partly through cultural consent and institutions of civil society. These thinkers do not offer interchangeable psychological theories. Their value is to explain how individual emotions gain political force when connected to employment, resources, rights, organizations, and elite strategy. Carl Schmitt made the friend–enemy distinction central to his concept of the political. This framework can help identify the point at which opponents cease to be treated as interlocutors. It is not a democratic prescription; Schmitt's association with Nazism makes it particularly important to distinguish analytical description from normative approval. Hannah Arendt studied totalitarianism, ideology, and the dissolution of plural public life. She examined the danger of reducing human beings to categories and subordinating individual judgment to an all-encompassing movement. Her analysis of Eichmann and the “banality of evil” does not establish that all political extremists are thoughtless or that every political crime results from automatic obedience. ## Crowds, obedience, and permission to behave worse Gustave Le Bon pictured crowds as particularly vulnerable to suggestion and imitation. His 1895 work remains part of the intellectual history of mass psychology, but its broad generalizations and elitist assumptions should not be substituted for contemporary empirical research. Freud's *Group Psychology and the Analysis of the Ego* (1921) considered identification with leaders and fellow members. Erich Fromm's *Escape from Freedom* (1941) related insecurity to the appeal of submission and authoritarian power. Theodor Adorno and colleagues, in *The Authoritarian Personality* (1950), explored authoritarian dispositions; the study's scales and inferences also face significant methodological limitations. None of these approaches justifies diagnosing individual voters from their electoral choices. They illuminate social conditions, not psychiatric labels. Stanley Milgram's obedience research and subsequent studies of conformity can inform questions about authority and norms, but they do not prove that every contemporary electorate would behave as experimental subjects once did. The transition from hostile rhetoric to actual violence depends on organizations, opportunities, sanctions, leadership, institutional protections, and specific circumstances. René Girard approached social rivalry through mimetic desire and scapegoating: disparate frustrations may be simplified by assigning them to a collective enemy. His account is broad and not always easily falsifiable. In contemporary political communication an observable analogue appears when one category of people is cast as the sole source of corruption, unemployment, moral decay, or insecurity, all of which have multiple causes. | Tradition or thinker | Contribution to understanding the mechanism | Limitation | | --- | --- | --- | | Hobbes, Tocqueville, Weber | Fear, conformity and legitimacy | Political theory does not measure current polarization | | Marx, Gramsci, Bourdieu | Class, hegemony and status distinction | Economics alone cannot explain personal hatred | | Freud, Fromm, Adorno | Identification, insecurity and authority | Historical models are not diagnoses of voters | | Tajfel, Turner, Iyengar | Group identity and affective discrimination | Findings depend on design and population | | Arendt, Girard, Haidt | Dehumanization, scapegoating and moralization | Their claims have different empirical status | ## The political economy of resentment: income, status, and abandonment Job loss, persistent inflation, and housing insecurity do not automatically produce radicalization. What may make them politically explosive is the gap between expectations and lived experience: the sense of having worked, followed the rules, and nonetheless lost an expected future. Material threats coexist with symbolic threats, including declining status, ridicule, or exclusion from a valued community. Pierre Bourdieu showed how social class involves not only income but also cultural capital, prestige, and practices of distinction. Arlie Russell Hochschild investigated feelings of being disregarded and narratives of deserved recognition in American politics. Pippa Norris and Ronald Inglehart interpreted part of recent conflict through cultural backlash, while Dani Rodrik examined distributional disruption and the political supply of nationalist and populist responses. No single pathway explains every country. Amory Gethin, Clara Martínez-Toledano, and Thomas Piketty assembled evidence from more than 300 elections in 21 Western democracies between 1948 and 2020. They documented a gradual separation of the electoral relationships with income and education. Coalitions conventionally associated with the left became relatively more attractive to highly educated voters, while high-income voters remained relatively closer to the right. This is consistent with an expansion of political conflict across redistribution, cultural identity, immigration, merit, and territory, rather than a simple disappearance of economic class. It describes a long-run pattern in Western democracies, not a law transferable without testing to Brazil. Institutional design shapes how economic disappointment becomes political distrust. The OECD's 2024 Trust Survey found that around 30% of respondents believed the political system gave people meaningful input on relevant measures of voice and consultation. A citizen facing high prices may still regard public institutions as legitimate if representation and correction are credible. When decisions appear inaccessible, policy losses can look like punishment imposed by a group unwilling to listen. Distrust may stimulate constructive participation, but can also be exploited in narratives seeking democratic rupture. ## Brazil: long roots and recent acceleration Brazilian polarization did not begin with digital platforms or a single election. Slavery, unequal access to rights, concentrated wealth and land, authoritarian practices, patronage, and recurring disputes about the state have shaped conflicts often settled unequally. Yet projecting today's party labels directly onto the Empire, the Old Republic, or the military dictatorship would be anachronistic. Political categories and coalitions changed across those eras. The democratic transition in 1985 and the 1988 Constitution institutionalized competition and expanded guarantees, without eliminating inequality, corruption, violence, or distrust. Monetary stabilization during the 1990s and phases of income growth and greater consumption subsequently changed expectations. The recession and institutional disputes of the 2010s, corruption allegations, disagreement over distribution, and contested authorities affected public confidence. No one event, party, or social group provides a sufficient explanation of the entire sequence. The 2013 demonstrations involved heterogeneous causes and actors, not a coherent political camp. In the years after the 2014 election, struggles increasingly spilled into demonstrations challenging the legitimacy of political competition itself. In a 2025 article in *Lua Nova*, Priscila Delgado de Carvalho and Luciana Tatagiba traced how protests surrounding presidential elections from 2010 to 2022 helped construct anti-Workers' Party identity, new right-wing mobilizations, and affective polarization. In 2018, disputes over public security, corruption, morality, religion, and opposition to the Workers' Party converged around Jair Bolsonaro, while opponents rallied around competing projects and against some of his positions and public statements. In 2022, a substantial share of these competing political identities centered on the Lula–Bolsonaro contest. The Brazilian Superior Electoral Court (TSE) recorded 60,345,999 valid second-round votes for Luiz Inácio Lula da Silva, or 50.90%, and 58,206,354 for Jair Bolsonaro, or 49.10%, in 2022. A narrow vote is an electoral fact, **not a direct measure of hatred**: close contests also occur in democracies where political losers accept the institutions. ```chart type: bar title: Brazilian presidential runoff in 2022 — share of valid votes unit: % Lula | 50.90 Bolsonaro | 49.10 ``` The January 8, 2023 attack on Congress, the Supreme Federal Court, and the presidential palace demonstrated that some radicalized participants crossed from protest into invasion and destruction of democratic institutions. The Federal Public Prosecutor's Office documents the attacks and material losses. Criminal responsibility must be assessed individually and with due process. Participants in violent acts cannot simply be equated with everyone who voted for a particular candidate. In December 2025, Datafolha found 40% of respondents more aligned with Workers' Party politics and 34% with Bolsonarism on a political-identification scale. Yet some respondents in the first category positioned themselves on the ideological right, while some in the second identified with the ideological left. Personal allegiance, negative identity, and ideological preference do not perfectly overlap. On October 13–14, 2026, a face-to-face Datafolha survey of 2,520 people in 153 municipalities found 78% perceiving Brazil as divided between opposing political camps, 17% perceiving harmony and respectful coexistence, and 5% not answering. This is a **measure of perceived polarization**, not a direct estimate of the percentage who personally hate their opponents. ```chart type: bar title: Perceived political climate in Brazil — Datafolha, October 2026 unit: % of respondents Polarized | 78 Harmonious | 17 Don't know | 5 ``` A longitudinal study published in *Nature Communications* in 2026 followed responses around Brazil's 2022 election across several survey waves, with samples of approximately 2,000–3,000 people per wave. It found marked affective polarization between Lula and Bolsonaro supporters before the campaign, stability during the contest, and partial decline afterward. Experimental treatments correcting mistaken beliefs about the other group's stance on contested policies reduced measured antagonism by up to 6.66 percentage points on the study's thermometer. This does not imply a guaranteed national reconciliation. It establishes that mistaken perceptions can contribute to hostility under experimentally tested conditions. ## The world has no single model of polarization The United States offers particularly extensive evidence about Democrat–Republican affective polarization. Pew Research Center surveyed 5,103 American adults on April 20–26, 2026: 58% held an unfavorable view of the Republican Party and 59% held an unfavorable view of the Democratic Party. Negative ratings of institutions are not the same as direct hatred of their supporters. Indeed, 26% viewed both parties negatively, compared with 21% in 2020. A polarized two-party system can coexist with considerable frustration toward the entire party system. European disputes over immigration, European integration, austerity, security, secularism, and sovereignty differ among countries. In some places the relevant divide is territorial, generational, or educational. Multiparty systems often produce coalitions crossing ideological categories. In Latin America, inflation, insecurity, inequality, and crises of representation can feed antagonism, but Argentina, Mexico, Chile, and Brazil do not share one political trajectory. Parts of Asia and other regions have distinct nationalist, religious, and institutional disputes. The concept of polarization should preserve differences between regimes and societies rather than flatten them. ```map title: Comparative settings — mechanisms, not comparable hostility scores Brazil | Personalized antagonism and negative identity | Surveys from 2022 to 2026 show distinctions between ideology and emotional alignment United States | Party identity and segmented media | Studies measure cross-party hostility and test informational interventions Europe | Diverse socioeconomic and cultural conflicts | Multiparty systems and electoral coalitions differ among countries Latin America | Inequality and representational crises | There is no uniform political trajectory across its countries ``` Research published in *Nature Human Behaviour* in April 2026 drew on quota-sampled online surveys of 15,202 respondents across 30 countries. Reported online political hostility was higher in less democratic and more economically unequal societies, and people who were hostile online also tended to be hostile offline across the sampled countries. These observational relationships do not establish that inequality, institutional conditions, or platforms independently cause hostility. Quota-sampled surveys should not be read as censuses. The findings nonetheless place online antagonism within broader distributions of resources, status, and political institutions. The V-Dem Institute's *Democracy Report 2026* puts current political disputes in the wider setting of pressures on representative institutions. Three questions must remain separate: how distant are competing policy positions, how strongly do social groups dislike each other, and do institutional rules for alternation in power remain legitimate? Democracy indices, opinion surveys, and election results do not measure the same variables. They should not be combined into an imaginary single global “hatred index.” ## Algorithms and the economics of outrage The business model of many digital services rewards attention, retention, and interaction. Provocative content can attract replies and sharing, securing additional exposure; political entrepreneurs can convert antagonism into audience, revenue, donations, and mobilization. This does not establish that every algorithm creates extremism or that users have no agency. Platforms, political parties, media organizations, and audiences continually affect one another. Experimental evidence requires caution. Research published in 2023 found that Facebook users encountered substantial same-minded content, but reducing exposure to such content within the experiment did not detectably reduce political polarization. A 2025 feed-reranking experiment on X found that increasing or decreasing exposure to partisan animosity changed out-group dislike. A 2026 study of X's feed algorithm found that activating algorithmic ranking shifted some policy attitudes in a more conservative direction without a statistically significant effect on measured polarization or partisanship. Another 2026 experiment tested redesigned feeds that could improve users' perceptions of norms in political discussion. These studies tested different platforms, populations, periods, treatments, and outcomes. Exposure, hostility, ideological preference, voting, and real-world violence cannot be treated as interchangeable variables. Political communication has also become less dependent on institutional intermediaries. Politicians can address segmented audiences directly; short videos and isolated statements turn opponents into caricatures. A person may encounter the other political camp chiefly through its most offensive moments, selected precisely because they spread well. The informational environment does not merely transmit disagreements; it can distort beliefs about the ordinary person on the other side. ```flow Material or symbolic insecurity → search for belonging and explanations → moralized political identity → selective interpretation of information → exaggerated perception of opposing hostility → social rewards for humiliation → distrust and retaliation Responsive institutions and pluralism → representation of interests → negotiated conflict → legitimate alternation in power → fewer incentives to pursue institutional rupture ``` ## Why humiliation can feel like winning Political hostility can offer immediate psychological rewards: group membership, peer approval, displays of loyalty, and a simple assignment of blame. Circulating an insult may feel like public service; humiliating an opponent may be framed as fighting injustice. Physical distance and limited personal contact can reduce emotional barriers to attacking strangers through screens. No ideology is the sole possible source of dehumanizing rhetoric. Such rhetoric and attempts to delegitimize opponents occur in different political traditions, yet their frequency, organizational structure, severity, and consequences must be assessed separately. **Analytical neutrality does not require factual equivalence between actions.** An unpopular proposal, an offensive remark, a coordinated deception campaign, and physical assault are not equivalent conduct. Democratic freedoms protect opposition, not an unlimited license to violate other people's rights. Hatred can also conceal distributional contests. Political organizations must raise funds, recruit volunteers, and gain visibility. It can be electorally less costly to describe an opponent as a threat to civilization than to defend a technically complex reform. The incentives are unequal: concentrated actors may benefit from mobilization, while the costs of weakened trust spread through households, schools, employers, and public bodies. | Mechanism | Evidence worth observing | Invalid shortcut | | --- | --- | --- | | Affective identity | Feeling thermometers, willingness to associate, discrimination | Close vote equals hatred | | Material insecurity | Employment, real earnings, inflation, regional differences | Poverty determines radicalization | | Elite strategy | Speeches, campaigns, incentives, coordination | All anger is manipulated | | Media systems | Feed experiments, reach, exposure, engagement | Algorithms automatically control choices | | Institutional fragility | Trust, acceptance of elections, tolerance | Criticizing a court is inherently anti-democratic | ## How politics enters families, faith communities, education, and work When partisan identification blends with family, faith, occupation, and community, the conflict leaves the ballot box. Within families, disagreement may be judged as ingratitude or moral failure; at work, as a defect of trustworthiness; in religious environments, as a test of doctrinal loyalty. Bonds that protect people from isolation can also produce conformity and exclusion when alignment becomes a condition of belonging. Religion is not synonymous with extremism, and secularism is no guarantee of tolerance. Religious and nonreligious traditions can articulate human dignity and ethical limits, but actors may also exploit either tradition to justify hierarchy, exclusion, or absolute certainty. Critical education should not impose artificial political uniformity; it should teach the differences among verifiable facts, interpretation, propaganda, interests, and legitimate disagreement. People can hold firm political positions while preserving reciprocal rights. Exposure to disagreement alone is no guaranteed cure. A 2018 field experiment by Christopher Bail and colleagues found that exposure to opposing political messages on social media, under certain conditions, could increase polarization rather than reduce it. Sustained personal contact, shared goals, and fair discussion rules are different from encountering provocative messages from strangers. A poorly designed intervention may make mistrust worse. ## Breaking the cycle — and the evidence still missing Conditions that could reduce antagonism include competition that recognizes losers' legitimacy, distributive policies that make costs explicit, institutions responsive to public grievances, effective representation, responsible journalism, proportionate sanctions for threats and violence, and public speech that does not manufacture unanimous hostility among opponents. None requires citizens to give up political struggle. The goal is to prevent competition from depending on denying the rights or humanity of rivals. One testable hypothesis is that correcting exaggerated beliefs about the opposition may reduce part of the resentment. The 2026 Brazilian study provides experimental evidence, but only for particular interventions and measures. Another is that changing the incentives for exposure and engagement may affect discussion norms; Facebook and X experiments, however, have yielded different results. Short-run effects cannot establish long-run transformation, and a more civil tone is not by itself proof of greater participation or social justice. Several findings could weaken the explanation advanced here. If hostility remained unchanged after correcting perceptions of opponents, misinformation would be less important in that setting. If communities facing similar economic conditions displayed sharply different levels of antagonism after accounting for institutions and media, insecurity alone would be insufficient. If changes in ranking affected engagement but not measured attitudes, algorithmic effects should be confined to the distribution of content. A rigorous account of polarization must remain open to these tests. Political conflict is not the failure of democracy; competition over power, values, and resources is central to it. The danger lies in the interaction of insecurity, hardened identities, incentives for indignation, and efforts to discredit shared rules. In that environment cruelty may be celebrated as courage, intolerance as loyalty, and falsehood as protection of the group. The boundary between democratic contest and persecution is maintained when citizens recognize that opponents remain fellow citizens whether they win, lose, or disagree. ## Sources and references - [Esteves and Junqueira (2025), systematic review of Brazilian political polarization — TSE Digital Library](https://bibliotecadigital.tse.jus.br/items/61ea4883-4079-405e-93ee-bb220bc8b412). - [Carvalho and Tatagiba (2025), protest and elections in Brazil, 2010–2022 — *Lua Nova* / SciELO](https://www.scielo.br/j/ln/a/mBmbX8KwYSmdktmL58bfXJn/?format=html). - [*Nature Communications* (2026), polarization dynamics and corrective information around Brazil's 2022 election](https://www.nature.com/articles/s41467-026-72990-9). - [Brazilian Superior Electoral Court, official 2022 runoff result](https://www.tse.jus.br/comunicacao/noticias/2022/Outubro/100-das-secoes-totalizadas-confira-como-ficou-o-quadro-eleitoral-apos-o-2o-turno/). - [Federal Public Prosecutor's Office (2026), documentation of January 8, 2023 attacks](https://www.mpf.mp.br/atuacao/como-e-a-atuacao-do-mpf/casos-de-destaque/atos-antidemocraticos/entenda-o-caso-1). - [Datafolha / *Folha de S.Paulo* (October 16, 2026), perceived polarization in Brazil](https://www1.folha.uol.com.br/poder/2026/10/datafolha-78-dizem-que-brasil-vive-polarizacao-politica.shtml). - [Datafolha / *Folha de S.Paulo* (December 24, 2025), political alignment in Brazil](https://www1.folha.uol.com.br/poder/2025/12/datafolha-polarizacao-atinge-74-dos-brasileiros-com-petistas-superando-bolsonaristas.shtml). - [Iyengar and Westwood (2015), affective polarization — *American Journal of Political Science*](https://doi.org/10.1111/ajps.12152). - [Iyengar et al. (2019), origins and consequences — *Annual Review of Political Science*](https://doi.org/10.1146/annurev-polisci-051117-073034). - [Gethin, Martínez-Toledano and Piketty (2022), political cleavages in 21 democracies — *Quarterly Journal of Economics*](https://doi.org/10.1093/QJE/QJAB036). - [OECD (2024), Trust Survey](https://www.oecd.org/en/publications/oecd-survey-on-drivers-of-trust-in-public-institutions-2024-results_9a20554b-en/full-report/executive-summary_7c5c9309.html). - [Pew Research Center (May 1, 2026), perceptions of US parties](https://www.pewresearch.org/politics/2026/05/01/americans-continue-to-view-both-the-republican-and-democratic-parties-negatively/). - [Bail et al. (2018), exposure to opposing views — *PNAS*](https://doi.org/10.1073/pnas.1804840115). - [Nyhan et al. (2023), like-minded Facebook sources — *Nature*](https://www.nature.com/articles/s41586-023-06297-w). - [Piccardi et al. (2025), partisan animosity feed intervention — *Science*](https://pubmed.ncbi.nlm.nih.gov/41308156/). - [Gauthier et al. (2026), X feed algorithm — *Nature*](https://www.nature.com/articles/s41586-026-10098-2). - [Brady et al. (2026), feed design and social-norm misperceptions — *Nature*](https://www.nature.com/articles/s41586-026-10536-1). - [Bor et al. (2026), political hostility in 30 countries — *Nature Human Behaviour*](https://www.nature.com/articles/s41562-026-02432-5). - [V-Dem Institute, *Democracy Report 2026* — publications](https://v-dem.net/publications/democracy-reports/). **Conceptual bibliography:** Machiavelli, *The Prince* and *Discourses*; Hobbes, *Leviathan*; Tocqueville, *Democracy in America*; Marx, *The Eighteenth Brumaire of Louis Bonaparte*; Weber, *Politics as a Vocation*; Le Bon, *The Crowd*; Freud, *Group Psychology and the Analysis of the Ego*; Fromm, *Escape from Freedom*; Adorno et al., *The Authoritarian Personality*; Arendt, *The Origins of Totalitarianism* and *Eichmann in Jerusalem*; Girard, *Violence and the Sacred*; Bourdieu, *Distinction*; Gramsci, *Prison Notebooks*; Haidt, *The Righteous Mind*; Hochschild, *Strangers in Their Own Land*. These texts offer different forms of explanation and should not be mistaken for direct measurements of polarization in 2026.