The Psychology of Speech Repression

Why do people across ideologies keep trying to silence dissent — and what does that reveal about the way our minds process threat, loyalty, and truth?

Author’s Note: In a recent review paper—“Speech Repression and Threat Narratives in Politics: Social Goals and Cognitive Foundations,” published in the Annals of the New York Academy of Sciences — I attempt to develop a theory explaining why activists across different historical and cultural contexts seek to regulate and suppress political speech. In this essay, I summarize the main ideas of that paper and focus in particular on the strategic and psychological motivations that lead people to restrict expression on moral and political topics.

I am grateful to Jacob Mchangama for inviting me to present this work here. I’d be more than grateful if you, reader, could take the time to read the paper and tell me what you think of my theorization — especially if you have critical thoughts. I’m going to keep working on this topic, theoretically and experimentally. I will end the post with a few thoughts on my own experience of self-censorship on “taboo” topics of human behavior approached from a naturalistic lens in French academic social science. All the references to the claims I’m making in this post can be found in the paper.

---

If you’ve spent time online in the past decade, you’ve witnessed it: someone says something judged as “offensive,” and the backlash is swift. Calls to “cancel” them, demands they be deplatformed, accusations of spreading “harmful” or “offensive” ideas. This happened on both the left and the right, particularly in the U.S., despite both sides claiming to champion free speech. The left worries about speech that might harm marginalized groups, while the right tries to ban critical race theory or discussions of the social influences on gender from schools.

Instances of speech repression taking place online are modern instances of an eternal tendency to restrict speech judged as “dangerous.” In the 1950s, the U.S. was seized with an anti-communist panic where anyone suspected of having left-wing sympathies was considered a facilitator of Soviet influence. The Catholic Church persecuted heretics. Etc.

What is striking is how little this dynamic respects ideological boundaries. Progressive activists typically seek to restrict speech they perceive as harmful to marginalized groups. Conservative activists, meanwhile, have pushed to prohibit the teaching of critical theories of race or the family entirely from school curricula. The censored content differs, but the underlying impulse is the same. Speech repression, in other words, is politically and historically ubiquitous.

An Interdisciplinary Perspective

So why, across such different ideological contexts, do people engage in speech repression? My paper reviews the research I believe is most relevant for understanding why motivations to repress speech tend to recur during periods of perceived social threat — highly polarized political environments, moral panics, or moments when narratives about hostile outgroups become prominent.

My analysis draws primarily on experimental political science, cultural evolution, evolutionary psychology, and social cognition research, supplemented by historical studies (but those are somewhat outside of my specialization as a political psychologist). Academic career tracks and our own cognitive limitations push us to read mostly and contribute to work within a given discipline. I believe this is quite suboptimal because interdisciplinary efforts are absolutely required to make theoretical progress in social science and psychology.

Speech repression, I argue, is precisely the kind of cross-cultural phenomenon that benefits from such an interdisciplinary approach.

That Political Narratives Are Simplistic and Often Held Rigidly Is A Feature, Not A Bug

Attempts to restrict speech often aim to protect mobilizing political narratives.

These narratives typically point to a collective threat and encourage coordinated collective action against it: racism, sexism, communism, capitalist exploitation, enemies of the faith, and so on. The ideological narratives rarely capture the full complexity of social reality. Instead, they tend to be simplified, dualistic, and moralized accounts of conflict between groups, often attributing greater intentional malice to adversaries than the empirical evidence would justify.

This simplification may appear puzzling from the perspective of rational cognition. One might expect individuals to benefit from forming nuanced and accurate beliefs about the social world.

In many domains, this is indeed the case. If I enter a dark room and collide with a table I did not know was there, I immediately revise my mental model of the room. Updating beliefs based on new evidence is clearly adaptive when dealing with directly observable physical phenomena, i.e., when dealing with what psychologists call “intuitive beliefs.”

But political narratives are rather different psychological objects. They are “reflective beliefs”: explicit beliefs about highly abstract social phenomena — macroscopic patterns of inequality, group conflict, or hidden conspiracies — which are next to impossible to test and falsify through everyday experience. As a result, these narratives can persist even when they oversimplify reality and are contradicted by evidence. Moreover, political narratives often incorporate features that make them cognitively attractive: strong threat cues, which capture attention, and conceptual simplicity, which makes them easy to understand, remember, and communicate.

From this perspective, political narratives function less as precise descriptions of the world than as tools for mobilizing and coordinating cooperation and signaling moral commitments. Political narratives are the best illustration that there can be social benefits from mischaracterizing social reality.

Hyper-Sensitive Threat Detection

Human cognition appears to be calibrated to detect threats very easily — much like a smoke detector designed to trigger false alarms rather than miss a real fire. This principle helps explain a wide range of phenomena, from the prevalence of conspiracy theories to disproportionate fears about unlikely dangers. In evolutionary fitness terms, the cost of ignoring a genuine threat could be much higher than the cost of reacting to a false alarm.

In particular, it is the detection or avoidance of social threats to which human minds appear deeply wired. Coordinated social threats from other human groups can be highly lethal, so there have been deep selection pressures on abilities to detect them and coordinate group cooperation in response, often using fear-mongering rhetoric. Our minds evolved a high responsiveness to information suggesting that a group—whether ethnic, ideological, or religious—poses a danger to our community or threatens our way of life.

This is where engaging in speech repression can become useful: to preserve prosocial commitments and mobilizations, in particular against perceived threats and political outgroups.

Controlling Beliefs and Controlling Common Knowledge

“Let’s hope it’s not true; but if it is true, let’s hope that it does not become widely known.” (possibly apocryphal quote attributed to Lady Ashley, speaking of Darwin’s theory that humans descended from apes)

Given their mobilizing power, ensuring that political narratives are widely disseminated in society and protected from criticism becomes at stake in fierce political battles of influence among political leaders and activists. This is because statements about threat operate as cognitive “vectors” that push, with varying degrees of force, toward more investment in cooperative efforts.

From the perspective of an ideological entrepreneur concerned with maintaining widespread support for their cherished cause, free speech poses a risk. It creates the danger that views nuancing the salience of a threat or the benefits of a mobilization — against say structural racism or communism — can travel through social networks and reach an increasing number of people.

Often, I suspect that activists repress speech because they seek to prevent the social dissemination of communicated information perceived as having the potential to “dilute” accounts of social threat and undermine prosocial commitments. Activists often worry that dissenting views—even if true—might weaken their cherished mobilization against their most feared foe.

In this context, trying to regulate and repress speech can make sense through at least two mechanisms.

The first is straightforward persuasion. Activists fear that exposing people to arguments defending racism, authoritarianism, communism, or other morally condemned positions might lead some individuals to adopt those views privately and adopt less prosocial, less regulated behaviors.

But a second mechanism is arguably equally important: the creation of common knowledge. When ideas are expressed publicly — through textbooks, speeches, or widely shared media — not only do individuals encounter them; everyone knows that everyone else has encountered them. This shared awareness can alter expectations about social norms.

If discriminatory or authoritarian arguments become widely visible, people may infer that such views are becoming more socially acceptable. This perceived shift in norms could embolden individuals who might otherwise hesitate to act on prejudiced attitudes.

The fear, then, is not merely that individuals will change their private beliefs. It is that public discourse might reshape shared perceptions of what others believe and what behaviors are socially tolerated. Hence, efforts to restrict the “dangerous” speech.

Punishment and Deterrence

To the extent that “dunking on” or punishing dissenting discourse is done publicly — in front of a large audience — it can also be a way of saying: “don’t mess with us,” “don’t you dare question the cause / qualify this sacred idea.”

I see this as a variant of the mobilization-maintenance motivations I just described. Here, you try to prevent the dissemination of demobilizing discourse by intimidating the source and publicizing the punishment to others. This motivation for cracking down on contrarian discourse is essential and falls in line with Steven Pinker’s analyses in “The cancelling instinct” from his last book, When Everyone Knows That Everyone Knows (which I highly recommend: It develops ideas that are fairly similar to mine on the motivations for censoring political and scientific discourse, and came out the same month as my paper).

Speech Repression as Signaling

Attempts to restrict political speech are not only outward-facing and meant to exert ideological influence. Punishing “dangerous” speech can also function as a signal of loyalty within the group, for personal benefit.

To care about a cause or group means being willing to sacrifice other things for it. By publicly condemning someone who challenges a particular ideological narrative, activists can try to demonstrate that they prioritize the cause above competing values such as intellectual openness, truth-seeking (if the censored discourse has evidence to back it up), or maintaining good relationships with the groups whose speech is restricted. In effect, the censors decide to intentionally burn bridges, precisely because this operates as a credible signal of commitment.

This can bring reputational benefits within one’s political community. Demonstrations of moral righteousness may be interpreted as evidence of integrity and dedication.

In that sense, speech restriction can be part of the wider phenomenon by which political communities reward intolerance of disagreement with an ingroup’s orthodoxy and punish efforts to understand other moral perspectives.

In sum, speech repression can simultaneously serve strategic goals both inside and outside one’s group (protecting a mobilizing narrative) and status goals within one’s political tribe.

Self-censorship as a form of signaling

The logic of reputational signaling also helps explain why people sometimes censor themselves — or even misrepresent their beliefs — when operating within environments dominated by strong political orthodoxies.

Political communities often expect their members to endorse certain core moral commitments. Under these conditions, refraining from expressing opinions that might be interpreted as expressing insufficient loyalty to a narrative or as politically demobilizing becomes a rational reputational strategy.

Human beings are, in fact, quite skilled at this kind of self-censorship. Elisabeth Noelle-Neumann famously described spirals of silence, situations in which individuals collectively refrain from expressing opinions they believe are socially disapproved.

Timur Kuran identified even more active forms of this phenomenon. In what he called “preference falsification” and “knowledge falsification,” individuals publicly endorse beliefs they do not actually hold because expressing their real views would carry reputational costs.

Those individuals who don’t spontaneously self-censor to sound politically correct and fit in often strike most people in society as being “weird,” neuro-atypical.

When people underestimate how widespread this self-censorship is, the result can be pluralistic ignorance: a situation in which public discourse becomes grossly misaligned with private beliefs. Everyone hears the same dominant opinions expressed in public, and each person assumes others genuinely believe them. In reality, many individuals privately harbor doubts that no one dares to articulate. Some examples in U.S. contexts: criticism of Trump’s authoritarianism in MAGA bubbles; scepticism towards open border immigration policies or biological sex denialism among strong liberals.

Does Speech Repression Reach Its Goals? Not Always.

Critically, speech restriction measures may or may not produce the effects activists intend. They may successfully suppress the diffusion of certain ideas, but the censors may very well provoke backlash, discredit themselves, and intensify conflict.

For the purposes of explaining why people try to cancel speech, however, it is sufficient that activists believe such measures will help achieve their goals. This is not an isolated phenomenon. Research in cultural evolution suggests that people often promote beliefs or practices when they believe doing so advances their objectives, even if those beliefs ultimately have little effect. The same logic likely applies to attempts to regulate speech.

This has important implications for democratic societies. Activists rarely anticipate the broader consequences of widespread self-censorship and echo chambers. When dissenting arguments disappear from public discussion, societies lose opportunities to identify errors in prevailing narratives or policies.

Similarly, if audiences are on average very attached to tolerance of disagreement or the pursuit of truth, active measures to repress putatively dangerous ideas by a political side can contribute to discrediting that political side in ways that its stronger proponents typically don’t anticipate. For instance, part of the anti-racist U.S. far left’s canceling attempts over the last ten years has had the effect of discrediting it among classical liberals attached to free expression, pushing those people more towards the right (e.g., James Lindsay, Peter Boghossian, Bret Weinstein, etc).

Cycles of mutual speech repression between political factions can thus escalate hostility and erode social trust in the long run. Speech repression is not often framed as such, but it is an essential manifestation and amplifier of affective polarization and distrust in elites.

Speech Repression Is Often Motivated By Sincere Convictions

It would be a mistake to assume that speech repressors are simply cynical or power-hungry. As far as the behavior of ordinary citizens is concerned, I would suspect that much speech repression is motivated by genuine prosocial commitments rather than by cynical power-seeking.

If one sincerely believes that certain ideas will lead to serious harms — racism, tyranny, moral decay, the collapse of the traditional family — restricting those ideas will feel like an act of protection rather than suppression. People often hold both moral and factual beliefs not because they have been proven true but because they seem necessary to uphold a moral order, a certain conception of a just society.

Moreover, there are strategic and reputational advantages to sincere belief. Individuals who genuinely endorse the values they advocate are typically more persuasive. They produce more convincing arguments, are more creative at finding rebuttals and post hoc rationalizations to objections, and are less likely to appear inconsistent.

Robert Trivers’ hypothesis of self-deception captures this dynamic: individuals’ cognitive systems may be evolutionarily calibrated to endorse socially important moral beliefs and narratives sincerely so that they can persuade others of their importance more effectively, and signal devotion to them.

If there’s one key finding in social psychology, it’s that human beings are remarkably adept not only at presenting their motives as virtuous but also at convincing themselves that those motives are genuinely noble.

Hypocrisy and the Defense of Free Speech

Psychological research consistently finds that people who appear motivated by universal moral principles are more trusted than those openly pursuing self-interest. It is therefore common for individuals to frame their actions in terms of broadly accepted principles such as justice, equality, or freedom. Free speech is one such principle.

Yet if the analysis developed above is correct, we should expect every political group to support restrictions on expression when doing so serves its perceived interests (e.g., mobilizations-maintenance and signaling). This means that commitments to free speech will often coexist with selective applications of that principle.

In practice, many people claim to endorse free expression in the abstract — often sincerely — while supporting censorship when speech threatens causes they care about. This does not necessarily reflect conscious duplicity. If Trivers’ ideas about self-deception are correct (they seem broadly correct to me, but they’re controversial), our cognitive architecture should make it easy to believe we’re guided by universal principles without seeing the myriad ways in which we apply the principles selectively.

This means that the institutional protection of free speech requires constant vigilance against deeply rooted features of human psychology. Namely, against our tendency to grant our own stated principles only to the ideas and groups we happen to like.

Self-Censorship in Academia: A Case Study

The dynamics of speech repression are not limited to professional politics or to online citizen mobilizations. They also appear in academic environments (the one I know best).

In the social sciences and humanities, political attitudes are heavily skewed toward the left. In France, this imbalance is particularly pronounced, shaped by the historical influence of Marxism, Bourdieusian sociology, anti-Americanism, and strands of anti-positivist and “post-modern” anti-science thought.

I personally identify with many left-leaning positions. I support stronger taxation of extreme wealth, share concerns about police violence (a big problem in la patrie des droits de l’homme), and stand with feminist movements against sexual violence. At the same time, I often find myself diverging from dominant leftist narratives on certain topics. I believe markets are extraordinarily powerful mechanisms for generating prosperity (by leveraging our natural selfish impulses), even if they require careful regulation. I also think that there can be genuine tensions between the tenet of sex equality and the cultural mindsets of certain immigrant populations from Muslim countries — the recognition of which has become incomprehensibly taboo on the left in my country.

My concern is that the strong (far) left-wing ideological homogeneity in academia in France can generate collective forms of motivated reasoning and motivated ignorance that undermine the truth-seeking mission of scientific research.

One example concerns the study of psychological sex differences and naturalistic approaches to human behavior more generally.

There is a basic principle in philosophy, articulated by David Hume and taught in every introductory ethics class: factual claims about the world do not logically imply normative conclusions. What it does not tell us is what ought to be. Full stop. The existence of average biological differences in motivation or ability between men and women — even if well-established — would not justify discriminating against any individual, any more than the existence of height differences between populations justifies treating tall people as superior.

Yet many professional researchers (paid by my parents’ tax euros to conduct impartial research) treat this as a live danger. Publish findings that suggest biological influences on socially desirable psychological sex differences, and someone, somewhere — the far right, masculinists — might use them to justify sexism. I’ll concede that the concern is not entirely without sociological basis. People do sometimes misuse scientific results to advance theses they already endorse. That’s impossible to control.

But the response — avoiding the empirical questions and making evolutionary hypotheses altogether taboo — is a non sequitur. You don’t protect against motivated misreading by refusing to publish anything. There will always be a risk that some people somewhere commit the fallacy of using empirical findings to justify questionable ends. And a bit of exposure to “biological” takes on sex differences won’t easily make Western generations committed to female emancipation fall back into XIXth century patriarchy.

The consequence is visible in the literature and seminar rooms. Strong social-constructionist accounts of gender have become effectively mandatory in parts of the social sciences — even when they conflict with cross-cultural data, with developmental psychology’s findings on the consistency of interest divergences between boys and girls across very different societies, and with the most elementary principles of evolutionary theory, whose predictions are richly confirmed in animal behavior. (Of course, the field is still moving, and some psychological sex differences on some traits have been found to be smaller than we thought. But the probability they don’t exist and never matter socially is close to zero).

Young scholars pick up the signal fast. Raising certain hypotheses attracts reputational costs. Worse, they are rarely exposed to naturalistic frameworks in the first place, both because few scholars feel safe working in that space and because the intellectual competencies needed to engage with it are never transmitted. When career advancement, passing peer review, and satisfying committees depend on peer approval, self-censorship becomes a rational strategy. It is, in the language of behavioral ecology, locally adaptive — even as it distorts the collective enterprise.

The result is a subtle distortion of academic inquiry and of the public discourse that looks at what academics say. Researchers avoid certain questions not because they lack scientific merit but because pursuing them could threaten prevailing moral narratives. This is already very bad for the goal of accurately informing the public.

But there is something worse: The costs of this dynamic are not confined to seminar rooms. They are concrete, and in some cases, they may fall on the very groups the dominant consensus claims to protect.

Medical research and women. For decades, biomedical research used male subjects as the default, on the tacit assumption that male and female bodies were interchangeable for research purposes. The results have been damaging sometimes: drug dosages calibrated on male physiology, cardiovascular disease systematically underdiagnosed in women, autoimmune conditions — which affect women at twice the rate of men — have been long under-studied. Recognizing biological and potentially psychological sex differences in medicine is a precondition for treating women properly.

Research on sexual violence and women. Evolutionary psychology has generated some of the most productive frameworks for understanding the conditions under which men commit sexual harassment and assault — including research on the situational triggers and male coalitional dynamics that modulate behavior that consists of surveilling and coercing women. Prevention programs that ignore the psychological mechanisms driving these behaviors — that treat sexual violence as purely a product of “patriarchal culture” or a willingness to exert “power” with no roots in sexual motives — are working with an impoverished model. You cannot effectively intervene in a phenomenon you do not seriously try to explain. Here too, the reluctance to engage with naturalistic accounts does not protect women. Ironically, it leaves them with weaker tools.

A glaring democratic contradiction. Our Western constitutional traditions rest on a commitment to pluralism, open inquiry, and the free contest of ideas — a hard-won recognition that no group, however well-intentioned, can be trusted with a monopoly on legitimate thought. Yet certain academic fields have become among the most ideologically homogeneous and censoring professional environments in modern society. The irony is sharp: institutions formally dedicated to truth-seeking, largely funded by the public, operating on a de facto consensus often on a par with that of political parties. A democracy should expect better from its universities.

A generation of scholars quietly squeezed out. There is today a cohort of young researchers — methodologically sophisticated, well-read in evolutionary biology and psychology, familiar with current cross-cultural datasets and experimental methods — who have learned to keep their heads down, or drifted out of academia altogether. Hiring committees are still dominated by older scholars who formed their identities in the 1980s when “evolutionary explanation” meant “genetic determinism” and “justification” of traditional sex roles. That caricature has long been overtaken by the data, with studies finding equally or almost equally left-wing mindsets in evolutionary social sciences and standard social sciences. But the profession is still largely, in effect, selecting for ideological and intellectual conformity under the guise of scientific judgment.

Conclusion

If the arguments outlined in my paper are correct, the struggle over free speech is not merely a conflict between enlightened defenders of liberal principles and illiberal censors. It requires a continuous fight against deep features of human psychology.

Political movements require motivating narratives. Human minds are highly sensitive to perceived threats. Public condemnation of dissent can signal loyalty and deter deviation. And individuals often sincerely believe that restricting speech serves the greater good.

These forces make pressures toward censorship persistent and recurrent. Defending free expression, therefore, requires institutions and cultural norms capable of resisting psychological tendencies that, in many other contexts, have historically served important social functions.

---

Antoine Marie is a postdoctoral researcher at CEVIPOF, Sciences Po in Paris. He is a political psychologist with a background in social psychology, political science, and cognitive science.