Desmet on the crowd before the ringleader

Matias Desmet has a precise and contested claim about the twentieth century, and he makes it again, slowly, for Glenn Diesen. The Soviet Union and Nazi Germany were not classical dictatorships. They were the first totalitarian states. The difference is not a matter of degree — a bit more brutality, a bit more reach — but of kind, and the kind is psychological. In Desmet’s account, what made the totalitarian state new was that the mass came first and the ringleader came after. A fanatical fraction of the population — he puts it at roughly 20 to 30 per cent — buys into a narrative before any seizure of power. Stalin and Hitler did not impose belief on a quiet society; they mounted a society that had already started to believe.

The crowd, not the dictator

Desmet’s first move is a definitional one, and almost everything that follows depends on it: the totalitarian state and the classical dictatorship are not the same animal.

A classical dictatorship, in his telling, is psychologically simple. A small group with a credible threat of violence imposes a social contract on a frightened population. If the regime collapses, the contract collapses with it. The mechanism is downward — from a regime to the people it intimidates — and it is reversible the moment the regime is no longer feared.

A totalitarian state is structurally different. The first thing that happens in its emergence is not coercion but conversion. A fraction of the population — Desmet’s persistent figure is 20 to 30 per cent — begins to believe fanatically in an ideology: historical materialism in the Soviet case, racial theory in the Nazi case. Only after this conversion is underway does a gifted rhetorical leader put themselves at the head of the converted and use them to capture the state. The state, once captured, is not propped up by police alone; it is propped up by the converted fraction, who report their neighbours and their families to it.

This is why, Desmet argues, classical strategies of resistance fail against totalitarian regimes. Removing the ringleader does not remove the crowd. The crowd, once formed, will find another.

The distinction matters for what the rest of the conversation tries to argue. If the new technocratic order Desmet sees forming is a totalitarian state in his sense, then the resistance script that works against a dictatorship — depose the strongman, replace the regime — does not apply. There is no strongman to depose. There is a converted fraction and a technocratic apparatus, and the converted fraction will simply find a new apparatus.

The four conditions, and the loneliness underneath them

If the mass is the engine, what makes one form? Desmet names four preconditions. They are not exotic; they describe the present.

The four are: pervasive social isolation, a profound lack of meaning, free-floating anxiety, and free-floating frustration or aggression. The adjective doing the work is free-floating: anxiety and frustration that the person cannot attach to any specific cause. This unattachable distress is, in Desmet’s clinical training, the most aversive psychological state a human can occupy. To feel afraid without knowing what one is afraid of is to feel out of control of one’s own interior.

When many people are in this state simultaneously, a narrative that names an object of anxiety — a virus, a class, a race, a foreign leader — and prescribes a collective ritual to act on it produces an immense relief. The relief is twofold. First, the unbound anxiety is given an object, which feels like control. Second, the ritual — masking, marching, salutes, vaccinations, denunciations — produces a sense of being connected to others performing the same ritual. Loneliness recedes. The narrative does not have to be true, or even plausible. It has to be performable.

The empirical question is whether Western societies are in fact in this condition. Desmet, in the conversation, reaches for a Gallup figure he half-remembers (“somewhere between 40 and 17 per cent”) and the David Graeber claim that “up to 60 per cent” of people consider their jobs meaningless. Both numbers are a little loose. The actual Meta-Gallup global figure, from a 2022–2023 survey of 142 countries, is that 24% of people worldwide in the Meta-Gallup study said they felt “very” or “fairly” lonely, with the figure rising sharply among the young — a high of 45% in Comoros to a low of 6% in Vietnam. The Graeber figure most often cited is closer to 37 percent: British workers who, asked by YouGov, said their job did not make any meaningful contribution to the world. The fact that Desmet’s numbers were imprecise does not falsify the underlying claim. It does mean the claim sits on softer empirical ground than the certainty of his delivery suggests.

The four conditions, in Desmet’s argument, are not an accident of modernity but its product. The atomised self that Enlightenment philosophy idealised — the autonomous, rational, self-sufficient individual — turns out to be lonely. The mechanistic worldview that delivered industrial productivity has stripped most jobs of meaning. The press, originally invented to give the citizen access to the invisible parts of an industrialised world, became dependent on the same industrial capital it was supposed to oversee, and so became propaganda. The result, in his telling, is a population pre-prepared for mass formation — anxious, lonely, performance-ready, and missing only the narrative.

The bond is to the ideal, not the person

This is the part of Desmet’s theory that seems most extreme and turns out to be the most argued for: a mass is not a group of connected people, it is a group of separately connected people.

The distinction is fine but it is the centre of the argument. In a healthy group, the bonds run laterally — A is attached to B, B to C, C back to A. In a mass, the bonds run upward. Each individual is attached to the same collective ideal, and only secondarily — through that ideal — to anyone else. The lateral bond between mother and child, or between neighbour and neighbour, is real but it now runs through the ideal. When the ideal demands betrayal of the lateral bond, the lateral bond breaks.

Desmet’s example, lifted from a conversation he had with an Iranian woman who lived through the early Khomeini years, is of a mother who reported her own son to the regime for disloyalty, attended his execution, and accepted a medal from the state. The example is gothic, and Desmet uses it that way; it makes his point in a single image. The general claim is that the bond to the ideal becomes stronger than the bond between persons, and once that asymmetry is established, the betrayals follow.

A mass is a group that emerges because each individual is connected to a collective ideal. After a while, that connection gets stronger than even the strongest connection between individuals.

— Matias Desmet

This is the mechanism Desmet uses to explain a particular paradox of the COVID years: that “solidarity” became the slogan of a period in which the elderly died alone in care homes, neighbours stopped helping strangers in distress, and informal household visits became reportable offences. The reader does not have to accept his entire reading of the COVID response to notice that the paradox itself was real. Whether Desmet’s mechanism is the correct explanation is a separate question.

There is a corollary worth naming. If the bond is to the ideal, then attacking the ideal — and not the people in its grip — is the only thing that has any effect. Persuading individuals one by one fails because the relevant bond is not interpersonal. Aiming at the narrative, calmly and publicly and at scale, is the only intervention Desmet considers viable. He returns to this point at the end of the conversation, and the closing recommendation of his book is its name: sincere speech.

The technocratic turn

Desmet’s most consequential claim — and the one most likely to determine whether the theory ages well — is that the next totalitarianism will look nothing like the first two.

He attributes the prediction to Hannah Arendt in 1953. The attribution is loose. Arendt did write extensively about the affinity between totalitarian rule and bureaucracy — most famously in Eichmann in Jerusalem and in her later journals — and her 1953 essay “Ideology and Terror,” which became the closing chapter of the 1958 edition of The Origins of Totalitarianism, names bureaucratic dehumanisation as a structural feature of the form. “The essence of totalitarian government, and perhaps the nature of every bureaucracy, is to make functionaries and mere cogs in the administrative machinery out of men, and thus to dehumanize them”, she wrote that year. The specific Desmet formulation — that Arendt predicted a coming totalitarianism “led by dull bureaucrats and technocrats” rather than by ringleaders — is a paraphrase he tightens for clarity. The argument is in the Arendt corpus; the sentence is his.

What he means by technocratic is more than just rule-by-experts. He means a public power whose face is faceless. The classical totalitarian state had a single recognisable image — Stalin’s moustache, Hitler’s salute. The technocratic version will have committees, dashboards, model outputs, and the recurrent invocation of “the science” or “the data” as the source of decisions no person can be held responsible for having made. The mass formation underneath remains the same: a converted fraction, a free-floating anxiety, a ritual of compliance. What changes is the surface. The ringleader has been replaced by a methodology.

The intellectual scaffolding for this turn, in Desmet’s view, is the mechanistic worldview — the seventeenth-century inheritance that treats the human being as a biological machine, fully describable in physical and computational terms. He names Yuval Noah Harari’s Homo Deus as a popularised exemplar: “Organisms are algorithms. This simple statement forms one of the key assertions underpinning Harari’s view of the future. This reductionist view claims that all life boils down to ‘a methodical set of steps that can be used to make calculations, resolve problems, and reach decisions.’” A society that has internalised this picture, Desmet argues, will be uniquely susceptible to a regime whose authority rests on having more data than its citizens. The expert, in this picture, replaces the prophet.

This is where the conversation arrives at its most distinct empirical claim. Desmet states that bureaucratic work has grown from “about 10 per cent” of jobs in the early nineteenth century to roughly 70 per cent today. The exact numbers are imprecise — there is no single coding scheme that distinguishes bureaucratic from non-bureaucratic work cleanly — but the directional claim is uncontroversial: a far larger share of paid work today is administrative, procedural, or managerial than was the case in 1900. Whether this transformation produces a technocratic class with the inclinations Desmet attributes to it is the harder question, and the one his theory cannot answer from the inside.

The Crimean parallel, and what Desmet thinks resists

Diesen, who is not a psychologist but a Norwegian political scientist with a long-standing realist reading of NATO, pulls the conversation toward the Crimean War. The detour is not a digression; it is where the theory tries to earn its keep on a specific historical case.

The Crimean War of October 1853 to February 1856 was, in the standard British framing, a war to defend the Ottoman Empire against Russian expansion. Lord Palmerston, then Foreign Secretary and shortly to become Prime Minister, treated the conflict as an opportunity to neutralise Russian naval power in the south. His two great aims were to prevent Russia establishing itself on the Bosphorus and to prevent France doing likewise on the Nile, a strategic framework that has proven remarkably durable. The Treaty of Paris in 1856 prohibited Russia from basing warships in the Black Sea — a limited measure, which did not reduce the totality of Russian power, but was still deeply resented as a national humiliation. The clauses were unilaterally repudiated by Russia in 1871.

Diesen’s reading, which Desmet endorses, is that this template — isolate Russia from the southern seas — is the same strategy NATO has pursued since the 1990s, and that the propagandistic technique attached to it is also continuous. The Turks of the 1840s, who had been described in the British press as barbaric easterners, were reframed in the 1850s as innocent victims of Russian aggression. The Russians, who had been a trade partner, were reframed as a civilisational threat. Marx and Engels, writing for the New-York Tribune, supported the British position; Marx saw the Crimean War as a conflict between the democratic ideals of the west that started with the “great movement of 1789” against “Russia and Absolutism.” The propagandistic compression of complex strategic interests into a binary of good versus evil, in this account, predates radio and predates television; it is a feature of modern mass politics from its first moments.

Desmet uses the Crimean comparison to make a methodological point about his theory. Mass formation is not new — he traces episodes back to the Crusades. What is new is the strength of the formations, which has tracked the increasing atomisation of the modern citizen. The 1850s mass was thinner and less totalising than the 1930s mass, which was thinner and less totalising than the 2020 mass, which will be thinner and less totalising than what is coming. If the loneliness data trend in the direction the WHO and the U.S. Surgeon General have been saying it is trending — and the Gallup figures above are consistent with that — then the precondition is being laid faster than the resistance is being organised.

This is where Desmet places his only constructive recommendation. He does not argue that the mass can be woken. He argues that the mass cannot be reasoned out of the narrative because the bond is not to reason. What sincere speech does, in his account, is prevent the final stage of mass formation — the stage at which destroying the dissident becomes an ethical duty. The persistence of calmly articulated dissent does not break the mass. It keeps the dissident alive, and through them keeps the option of return open for whoever, eventually, will want to return.

What is and is not settled

A coda for the careful reader, who should be holding several things at once.

What is genuinely contested in this conversation: that mass formation, as Desmet describes it, is a real and identifiable psychological mechanism rather than a sophisticated narrative frame. The theory has not been operationalised in a way mainstream psychology recognises. It explains a great deal, perhaps too much; theories that explain everything are often theories that predict nothing. A reader who finds the theory illuminating should be honest that they are responding to its descriptive force, not to its empirical validation.

What is not contested: that twentieth-century totalitarian regimes used propaganda systematically, that they recruited and depended on a fanatical minority, that they relied on neighbour-to-neighbour denunciation, and that the technocratic register has become more prominent in contemporary governance than the strongman register. These are historical and observational facts that Desmet’s theory organises but does not need to invent. A reader can reject mass formation as a unified mechanism and still take the COVID-era paradoxes Desmet names — the solidarity-at-a-distance, the elderly dying alone, the household informant — as worth a serious explanation, just one that may take a different shape than the one offered here.

What this conversation is for: it is not the right venue for deciding whether mass formation is true. It is a useful venue for hearing a worked-out account of why a careful reader might come to suspect that the dominant interpretive frame of the past five years is itself the kind of thing the frame is supposed to identify. That is an uncomfortable thought to hold steadily, and Desmet’s contribution is to articulate it in a register that — for all the imprecision in his numbers and the looseness of his historical attributions — does not raise its voice. Diesen, hosting his third openly heterodox guest in this collection, is doing the same. Whether the reader ends agreeing or disagreeing, that register is the part worth preserving.