When Informal Guardrails Fail: The Erosion of Democratic Taboos and the Risks of Normalising Extremism

Democracies are judged by their visible institutions. Elections, constitutions, courts, legislatures and a free press are treated as the cornerstones upon which democratic systems stand. When coming to assess democratic health, they tend to focus on voter turnout, constitutional protections, judicial independence, or the conduct of political leaders.

Many of the rules that sustain democratic societies are neither written down nor formally enforced. They exist in the realm of social expectation: shared beliefs about what constitutes legitimate political behaviour, how opponents should be treated, and where the boundaries of acceptable public conduct lie. These informal norms rarely attract attention when they are functioning well. Their significance comes apparent only when they begin to fray.

This was the central theme of a recent The International Risk Podcast interview with Erez Levin. Levin’s argument is straightforward but epochal: liberal democracies depend not only on legal protections, but also on a set of informal social guardrails that discourage overt forms of hateful bigotry and dehumanisation. When those guardrails are impaired, the effects extend far beyond immediate public discourse.

The question is not whether norms change. They always do. The question is what happens when the norms that govern restraint change faster than the institutions designed to contain their effects.

The Slow Drift of Democratic Boundaries

Political breakdown is often depicted as sudden. In reality, it is usually preceded by a long period of adjustment in which previously unthinkable positions become discussable, then defensible, then routine.

Weimer Germany is often invoked in this context not as an analogy, but as a caution about tempo. The decisive shift was not the arrival of dictatorship, but the gradual diminution of shared constraints in public discourse. Political opponents were fabricated not as rivals within a common system, but as existential threats to the political community itself. Once that shift occurred, compromise no longer appeared as democratic practice but as moral compromise. Institutions remained in place, but their foundational legitimacy was hollowed out.

A similar logic, though unfolding under very different historical conditions, can be seen in Rwanda in the years preceding 1994. The genocide did not emerge from rhetoric alone, nor from a single ideological shift. It was the result of organised political violence layered onto long-standing grievances and administrative structures capable of mobilisation. It was also preceded by a sustained effort to recode social identity in progressively dehumanising terms, particularly through the categorisation of Tutsi and Hutu populations in political and media discourse.

Leading up to the genocide, segments of the Tutsi population were portrayed through propaganda and political messaging as an internal threat, often as infiltrators or enemies of the state. The prominence of this process lies not in the claim that language causes violence mechanically, but in how it alters the perceived boundaries of moral obligation. As categories of belonging were narrowed, the psychological distance between speech and permissible action diminished too. Violence, when it came, did not appear as a rupture with language, but as something that had already been rehearsed within it.

Across these cases, institutions tend to erode more slowly than the norms that sustain them. By the time institutional stress becomes visible, the social constraints that once made certain forms of political action unthinkable may already have shifted.

Taboos as Democratic Infrastructure

Levin’s argument becomes salient here. His focus is not on censorship or legal restriction, but on the informal fabrics that make pluralistic politics workable. Democracies, he suggests, are sustained not only by rights but by taboos, social limits that make certain forms of expression costly even when they remain lawful. These are a stabilising function enforced through reputational consequence, institutional boundary-setting, and collective refusal rather than formal sanction.

The aftermath of the 2020-21 wave of political violence and unrest in the United States exemplifies this, when several widely publicised incidents triggered immediate and uneven social and professional consequences for those who appeared to endorse or celebrate acts of political violence. What is notable is not the consistency of those responses, but their visibility: universities, media organisations, and private employers were forced to articulate, in real time, where the line between protected speech and disqualifying conduct should sit. In some cases, individuals were suspended or dismissed; in others, similar expressions produced little response institutionally. The resulting ambiguity becomes part of the political signal itself, shaping perceptions of what kinds of rhetoric could be expressed without cost.

What Levin is describing is not necessarily the elimination of objectionable ideas, but their containment within a predictable perimeter, one in which the social costs of endorsement remain legible, even in moments of heighted political emotion.

Visibility Without Proportion

The difficulty in the present moment is that visibility is no longer a reliable proxy for prevalence. Digital communication has disrupted one of the basic assumptions of social judgement: that repeated exposure to a view reflects its consequential distribution in society.

Online platforms don’t just transmit information; they reorder it. Content that is emotionally salient, conflict-driven, or identity-reinforcing is systematically amplified, while more routine or moderate position tend to circulate with less intensity. This results in a distortion in which the loudest signals are often mistaken for the most representative.

Such became visible during the 2016-2024 electoral cycle across multiple Western democracies, where fringe political narratives, often concentrated within highly active online sub-communities, achieved high levels of visibility that far exceeded their organisational or electoral strength. In a number of cases, small but highly networked online communities were able to set the tempo of public debate by repeatedly surfacing framings of immigration, identity, or institutional legitimacy. While these were not necessarily new beliefs, nor widely held ones, they were optimised for circulation within algorithmic environments that reward repetition, affect, and conflict.

What is important is not simply that such narratives exist, but that they become structurally overrepresented in the informational environment in which broader publics form their impressions. This is the gradual sense that certain positions are more socially present, and therefore more socially permissible, than they actually are.

Risk Implications for Businesses

Levin’s argument can easily be directed to political terms, but some immediate consequences are arguably corporate. If social taboos function as informal guardrails for public discourse, then businesses are increasingly finding themselves operating as de facto enforcement nodes within that system, whether they intend to or not.

The first and most visible risk is reputational volatility. In an environment where contested speech can travel instantly across platforms, firms are frequently drawn into disputes that originate far outside their operational remit. A single employee post, customer interaction, or public association can rapidly become a test case for perceived organisational values with the absence of stable reference points for a proportionate response.

A second, more structural risk is workforce polarisation. Large organisations now contain employees whose political and moral expectations of the workplace diverge significantly. For some, neutrality is itself interpreted as complicity; for others, visible enforcement of behavioural norms is seen as overreach. This produces a managerial tension in which companies are expected to act as arbiters of acceptable discourse while simultaneously maintaining internal cohesion and psychological safety.

The third risk is what might be described as boundary diffusion, the gradual decay of clarity around where professional conduct ends and personal expression begins. High-profile cases in recent years, including disputes over employee speech on social media following politically charged events like Charlie Kirk’s death, have illustrated how quickly off-duty expression can be reinterpreted as organisational liability. The result is not simply disciplinary pressure, but uncertainty: employees and employers alike become less confident about where institutional tolerance begins and ends.

In Levin’s framing, this is precisely where the erosion of shared taboos becomes economically material. When social constraints weaken or fragment, firms are left to navigate an unpredictable space between legal permissibility and social legitimacy, a gap that no compliance framework fully resolves.

When these expectations are widely shared, they function as a stabilising layer between disagreement and disorder. When they fragment, organisations are left exposed to rapid shifts in sentiment, inconsistent enforcement, and reputational shocks that are difficult to anticipate or contain. There is a large danger of habituation.