Atheism and Atrocities: The Argument Examined

The Accusation and Why It Deserves a Real Answer

The argument arrives with the confidence of a finishing blow. You have spent ten minutes laying out the historical record of religious violence, the Inquisition, the Crusades, the sectarian slaughters, the systematic persecution of heretics, the blessing of colonial conquest, and then it lands: what about Stalin, what about Mao, what about Pol Pot? Those were atheist regimes, and they killed more people than the Church ever dreamed of. The person delivering this line usually considers the matter closed. The atheist, in their telling, has been hoist by the very logic they invoked.

This is not a stupid argument. It is, in fact, a serious one, and it deserves to be treated as such rather than waved away with slogans. The twentieth century’s totalitarian regimes killed on a scale that staggers the moral imagination. Stalin’s purges, collectivisation campaign and Gulag system are estimated to have caused somewhere between six and twenty million deaths, depending on the methodology applied and whether one counts the famine deaths attributable to deliberate policy. Mao Zedong’s Great Leap Forward produced one of the worst man-made famines in recorded history, with credible estimates ranging from fifteen to fifty-five million dead, a range whose very width reflects the difficulty of obtaining honest statistics from a system designed to falsify them. Pol Pot’s Khmer Rouge, governing a country of perhaps seven million people, killed between one and a half and two million of them in under four years. These are not talking points. They are catastrophes, and any honest account of political violence in the modern era must confront them directly.

The question is not whether those catastrophes occurred, or whether the regimes responsible were wicked. The genuine question is whether those catastrophes were caused by atheism, whether the absence of belief in gods was the operative factor driving men to build extermination camps and engineer famines. That is a genuinely different question from whether the regimes were horrible, and conflating the two is precisely where the argument goes wrong.

This essay argues the history rather than the personalities. It examines what actually drove those regimes to kill, compares the logic of totalitarian violence with the logic of religiously motivated violence across centuries, and addresses the three standard comebacks that appear whenever this territory is disputed: that communism was “really a religion” anyway; that atheism left a moral vacuum which ideology filled; and that no society sustains decent governance without some form of religious foundation. Each of these deserves more than a dismissive sentence, so each will receive its own sustained treatment. The conclusion, reached by examining the evidence rather than asserting it, is that the accusation depends on a causal claim that the historical record does not support.

What Atheism Actually Is, and What It Is Not

Before the historical argument can be made properly, a definitional point is necessary, because the entire accusation depends on a category confusion that is almost never examined. Atheism is not a worldview in the sense that Christianity, Islam or Marxism-Leninism is a worldview. It is not a doctrine, a manifesto, or a creed. It contains no theory of history, no account of human nature, no instructions about how to organise agriculture or industry, no guidance about the relationship between state and individual, no demand for ideological conformity, no eschatology driving believers toward a final terrestrial or celestial reckoning, and no ecclesiastical structure authorising its conclusions. Atheism is the absence of belief in gods, and that is, in the most literal sense, the whole of it.

This may seem an obvious point, but it carries enormous consequences for the argument being made against it. When a person kills in the name of Islam, Christianity, Hinduism or any other theistic tradition, there is a direct chain of doctrinal causation available for inspection. The doctrine commands, authorises, or at minimum makes intelligible the killing. The Quran contains specific passages about jihad and the treatment of apostates. The Bible contains explicit instructions for the execution of those who worship other gods. The Inquisition operated under papal authority and drew its justification from specific theological claims about heresy and the danger it posed to immortal souls. In each of these cases, the doctrine and the violence are connected, and the connection can be traced through texts, institutions and specific acts of authority.

With atheism, there is no such line to trace. There is no atheist text commanding the collectivisation of agriculture. There is no atheist theology demanding the elimination of class enemies. There is no atheist eschatology promising salvation through revolutionary violence. When Stalin ordered the Gulag or Mao ordered the denunciations of the Cultural Revolution, they were not acting on deductions from the proposition that gods do not exist. They were acting on deductions from Marxist-Leninist theory of history, Stalinist theory of the vanguard party, and the operational logic of a one-party state that had defined dissent as an existential threat to the revolution. The driving force was a specific political ideology, and political ideology is not the same thing as atheism, even when that ideology includes a rejection of religion among its many other components.

The parallel worth considering here is instructive. Many violent regimes across history have also been notable for other incidental features of their leadership, particular dietary preferences, architectural tastes, or cultural prohibitions. Nobody argues that vegetarianism caused the Holocaust because Hitler was reportedly averse to meat. The co-occurrence of two features within the same regime does not establish that one caused the other. The burden of proof lies with the person claiming causation, and that burden has not been met. To say “these regimes were atheist and they killed people, therefore atheism kills” is to reason no more carefully than to say “these regimes banned jazz and they killed people, therefore banning jazz kills people.” The logical structure is identical, and it fails in both cases for the same reason: correlation is not causation, and an incidental feature is not a mechanism.

What Actually Drove the Killing: Totalitarian Logic

The actual mechanism of totalitarian killing is well understood by historians, and it has very little to do with theology in either direction. The conditions required to produce mass political violence on the scale achieved by Stalin, Mao and Pol Pot are consistent across regimes regardless of their religious or irreligious character. They include: a single-party state that has eliminated institutional opposition and destroyed the independent judiciary; a cult of personality that renders the leader’s judgement practically unchallengeable within the system; an official ideology that defines enemies of the state as threats to a cosmic or historical project; a securitised internal apparatus empowered to identify and eliminate those enemies; and a population rendered politically inert by systematic terror and the destruction of civil society. These conditions appear in fascist regimes, communist regimes, theocratic regimes and nationalist regimes alike. The presence or absence of a god in the official ideology is not the critical variable.

George Orwell, writing in 1946, identified the structural dynamic with characteristic precision: “A totalitarian state is in effect a theocracy, and its ruling caste, in order to keep its position, has to be thought of as infallible. But since, in practice, no one is infallible, it is frequently necessary to rearrange past events in order to show that this or that mistake was not made, or that this or that imaginary triumph actually happened. Then, again, every major change in policy demands a corresponding change of doctrine and a revaluation of prominent historical figures.” Orwell was not claiming that totalitarianism is religious in the sense of requiring a supernatural deity. He was identifying that the structural logic of totalitarianism, the demand for infallibility, the rewriting of history, the doctrinal rigidity, the persecution of those who question the official line, replicates the structural logic of theocracy. The implication is not that Stalin was secretly religious. The implication is that the pathology runs deeper than theology: it is a pathology of unchecked power that can clothe itself in any available ideological garment, divine or secular, sacred or revolutionary.

The Stalinist cult of personality is instructive on this point. Stalin was not venerated as an atheist thinker whose calm reasoning about the non-existence of God had led him to wise conclusions about agricultural policy. He was venerated as the infallible interpreter of historical forces, as the embodiment of the revolutionary will of the proletariat, as a near-sacred figure whose portrait hung in every public building and whose published words constituted a form of law in a manner that would be entirely familiar to any student of divine-right monarchy or papal infallibility. The absence of a supernatural god did not create a disenchanted, rational political culture in which evidence and argument held sway. It created a political culture in which the functions previously performed by God, the authorisation of power, the promise of a perfect future, the identification of enemies deserving annihilation, were transferred wholesale to the party and its leader. The god-shaped structural role was filled with total efficiency; only the specific occupant changed.

Mao’s Cultural Revolution is equally revealing on this point. The Red Guards were not motivated by calm, secular reasoning about the non-existence of deities. They were driven by a quasi-sacred text, the Little Red Book, which was waved at mass rallies and treated with a reverence functionally indistinguishable from religious devotion in any other cultural context. The denunciation sessions in which teachers, intellectuals and party officials were forced to confess ideological crimes before crowds demanding punishment bear a striking structural resemblance to the auto-da-fé of the Spanish Inquisition. The content of the required confession differed between the two systems. The mechanism of enforced public self-abasement before a righteous collective empowered to pass judgement on correct belief was functionally identical. Removing the supernatural element had not produced rationality, restraint or compassion. It had produced a different species of fanaticism, one just as capable of filling a stadium with screaming accusers and just as indifferent to evidence, mercy or individual circumstance.

Pol Pot’s Cambodia takes this logic to its most extreme expression. Year Zero, the foundational concept of the Khmer Rouge project, was an explicitly eschatological idea dressed in secular language. History was to be annulled and remade from nothing. The agrarian utopia to be constructed from the ruins of existing society was not a modest policy proposal amenable to revision in light of results; it was a vision of redemption, of a purified people emerging from the cleansing violence of total revolution. The language of Year Zero, with its imagery of purity, contamination, enemies of the people, and the absolute erasure of the past, was soteriological in structure even while rejecting religious vocabulary. It promised salvation through destruction. The lesson this offers is not that the Khmer Rouge were secretly religious. The lesson is that one need not have a god to construct an apocalypse.

Comparing Like With Like: The Religious Body Count Across Centuries

The standard version of the “atheism kills more” argument compares twentieth-century secular totalitarianism against a kind of sanitised, best-case version of religious governance, implicitly selecting the Church at its most charitable moments and the secular state at its most murderous. This is not an intellectually honest comparison. A fair comparison requires examining what theistic governance actually produced across the full span of human history, and examining it with the same willingness to attribute deaths to their ideological cause that the argument demands when applied to communist regimes.

The Crusades were launched by Pope Urban II in 1095 with the explicit promise of spiritual reward, including the remission of sins, for those who fought and died in the recapture of Jerusalem. The enterprise was organised, commanded and theologically justified by the Church, and its violence was celebrated by participants as a glorious act of Christian devotion. Contemporary accounts of the sack of Jerusalem in 1099 describe Crusader forces massacring the Muslim and Jewish population of the city and giving thanks to God for the victory in the same breath. The First Crusade alone is estimated to have caused several hundred thousand deaths, and four subsequent major Crusades, along with numerous smaller campaigns, extended the project over two centuries. The Albigensian Crusade in southern France, launched in 1209 against the Cathar heresy, killed an estimated two hundred thousand to one million people in a campaign lasting twenty years. The reported instruction attributed to the papal legate at the massacre of Béziers, “Kill them all; God will know his own,” may or may not be verbatim, but the massacre itself was entirely real and entirely sanctioned by the theological framework that organised it.

The Spanish Inquisition, established in 1478 and operating for three and a half centuries, is often cited by its defenders as less lethal than popular imagination suggests, and this is partly accurate: modern historians estimate around three to five thousand executions over the full period of its operation. But the appropriate comparison is one of logic rather than merely of scale. The Inquisition operated within a framework that explicitly authorised the torture and judicial killing of human beings for holding incorrect opinions about the nature of God, for privately practising Judaism or Islam after nominal conversion to Christianity, and for reading prohibited texts. The doctrine was the mechanism. Remove the theological framework and there is no Inquisition, because there is no coherent concept of heresy, no imperilled immortal soul requiring rescue through interrogation, and no authority to define correct belief and enforce it by violence. The same analytical test applied to Stalinist terror yields a different result: remove the atheism and the Gulag persists, because the Gulag was generated by Leninist vanguardism, Stalinist paranoia and the mechanics of a totalitarian police state, none of which requires atheism as a logical premise. The Gulag would have operated under a Christian Bolshevik regime, as indeed it did operate under the brief governance of regimes that were not atheist in their formal ideology but were equally authoritarian in their structure.

The early modern European witch trials, running roughly from the fifteenth to the eighteenth centuries across Catholic and Protestant territories alike, produced somewhere between forty thousand and sixty thousand executions by conservative scholarly estimates, with some researchers placing the total considerably higher. Every one of those deaths was theologically motivated. The belief that certain individuals had entered into compact with Satan, that they could cause harm through supernatural means, and that execution was the appropriate remedy for this crime, derived entirely from Christian demonology. Without the theology, no witch trial is logically possible, because the entire concept of diabolical witchcraft is incoherent outside a theological framework that posits a devil, a divine order threatened by his activity in the world, and a church with the authority and obligation to defend that order by identifying and destroying his human agents.

The Wars of Religion in Europe, fought across the sixteenth and seventeenth centuries primarily between Catholic and Protestant powers, killed millions. The Thirty Years War alone, from 1618 to 1648, is estimated to have reduced the population of some German territories by a third through a combination of battlefield deaths, displacement, famine and disease generated by the military campaigns. This was a war explicitly organised around the question of which version of Christianity should govern Christendom. Theological difference was the casus belli. It was not the only cause, since dynastic and territorial interests were also plainly in play, but the religious dimension was not decorative. It organised the military coalitions, supplied the motivating rhetoric, and prevented the kind of negotiated settlement that disputes over secular interests more readily permit, precisely because no ruler could afford to be seen compromising on the truth of the Gospel without undermining the entire basis of their own sacral authority.

Colonial violence authorised by religious doctrine constitutes another substantial and often undercounted category. The Spanish conquest of the Americas was accompanied by explicit theological justification throughout. The Requerimiento, a legal document read aloud to indigenous populations before military assault, informed them that they were being given the opportunity to submit to the authority of the Church and the Spanish Crown, and that failure to do so would result in war waged against them with fire and sword, and that the deaths, losses and harms resulting would be their own fault. The document was routinely read in Latin to populations who spoke neither Spanish nor Latin, sometimes from ships still offshore, as a theological formality preceding the assault that had already been planned on entirely other grounds. The population of the Americas fell from an estimated fifty to one hundred million at the time of first contact to perhaps ten million by 1600, through a combination of epidemic disease and deliberate violence. The violence was authorised, organised and rhetorically framed by Christian theology. The doctrine of papal donation, which held that the Pope possessed the authority to grant sovereignty over non-Christian lands to Christian monarchs, provided the entire legal framework for the conquest. This is doctrine doing direct causal work, in precisely the way that atheism does not do causal work in the Soviet purges.

The First Comeback: “Communism Was Really a Religion”

The most intellectually interesting response to this line of argument is the claim that communism, Maoism and similar ideologies were themselves forms of religion, that they possessed sacred texts, prophets, holy days, inquisitions, heresies and martyrs, and that calling them “atheist” is therefore misleading at best. The regimes were, on this account, not examples of godless rationalism but of secular religion, and their violence is accordingly evidence not against religion but for its ubiquity and apparent inescapability as a feature of organised human society.

This argument has genuine analytical force, and it would be intellectually dishonest to dismiss it. There is real and documentable structural similarity between the way Marxist-Leninist states functioned and the way theocratic states function. The reverence for sacred texts, the infallible authority of the party’s interpretation of those texts, the public confession of ideological sin before a judging collective, the identification and punishment of heretics who deviate from correct doctrine, the promise of a terrestrial paradise corresponding in structural terms to the eschatological promise of theistic religion: all of these parallels are real and have been documented by serious historians and social theorists, from Hannah Arendt’s analysis of totalitarianism as a wholly new political form to Eric Hoffer’s work on the psychology of mass movements. Orwell identified the same structural correspondence in the passage already cited above.

But notice carefully what this concession actually does to the original argument. If the theist wants to claim that communist regimes were “really” religious in their structure and psychology, then the deaths produced by those regimes cannot honestly be placed in an “atheism” column at all. They belong instead in a “religion-like fanaticism” column, which is precisely where critics of institutional religion have always wanted to put them. The argument that communism was a pseudo-religion does not function as a defence of religion; it functions as an extension of the critique of religion into secular territory. If the absence of a supernatural deity is insufficient to prevent a movement from developing all the pathological features of a religion, including sacred authority, enforced orthodoxy, the persecution of those who question the official line, and utopian violence in the service of a promised future, then the implication is that religion’s most dangerous features are not generated by the supernatural element itself. They are generated by the structural demand for unquestionable authority, and that demand can be satisfied by a party, a leader, a nation or an ideology just as efficiently as it can be satisfied by a god. The problem is not the existence of God as a specific philosophical claim. The problem is the style of thinking that requires an unchallengeable foundation for authority, that treats doubt as betrayal and dissent as heresy. Theism provides one such foundation, and has historically been among its most durable and powerful expressions. Totalitarian secular ideology provides another. Neither is reliably safe when it achieves unchecked institutional power.

Jerry Coyne has noted the observation, made by the journalist Nick Cohen, that “when people say of dozens of political and cultural movements from monetarism to Marxism that their followers treat their cause ‘like a religion,’ they never mean it as a compliment. They mean that dumb obedience to higher authority and an obstinate attachment to dogma mark its adherents.” The comparison to religion is itself deployed as a term of criticism. It signals the presence of dogmatism, of deference to authority over evidence, of the suppression of legitimate doubt. When commentators describe Stalinism as quasi-religious in its structure, they are indicting Stalinism by reference to a failure mode that religion exemplifies in its most familiar institutional forms. They are not exonerating religion; they are using religion as the paradigm case of what rational, evidence-responsive discourse should not become. The critique travels in one direction only.

Grant the structural parallel between communism and religion, and you have not refuted the atheist critique of religious institutions. You have instead generalised it into a broader critique of authoritarian systems of thought that suppress rational scrutiny, of which institutional religion has historically been among the most prominent, most durable and most consequential examples. The ground has shifted, but it has shifted against religion, not in its favour.

The Second Comeback: “Atheism Left a Vacuum”

A softer and more nuanced version of the argument does not claim that atheism directly caused the violence of totalitarian regimes, but suggests instead that it created a moral vacuum. When religion was driven out of public life by communist regimes, something had to fill the space it had previously occupied, and what filled it was nationalism, ideological fervour, and the cult of personality. The implicit claim is that religion, whatever its historical faults, provides a kind of moral ballast that moderates the worst tendencies of purely secular governance, and that deliberately removing it leaves a society exposed to precisely the pathologies that appeared in Soviet Russia and Maoist China.

This argument is more sympathetic than the direct accusation, and it deserves a careful response rather than a dismissive one. The careful response begins by noting that a vacuum is not itself a cause. If the furniture is removed from a room and someone subsequently places a bomb in it, the missing furniture did not cause the explosion. The causal question is what motivated the person who brought the bomb, not what was absent from the room when they arrived. To say that the removal of religion left a vacuum that totalitarian ideology filled is to describe a sequence of events; it is not to identify a mechanism. It does not establish that religion was the only possible content for that institutional space, or that a different kind of secular culture, one built on liberal institutions, democratic accountability, separation of powers, and the rule of law, would have produced the same catastrophic outcome. Sequence is not causation, and a description of what happened is not an explanation of why it had to happen in that specific way.

The empirical record of actually existing secular societies is, on this point, decisive, and it points in exactly the opposite direction from what the vacuum argument predicts. The most secular societies in the world today are not dystopian nightmares of moral collapse and political violence. They are, by virtually every available measure of human welfare, among the best places on earth to be alive. The Scandinavian countries, which consistently rank among the least religious in terms of regular attendance at religious services and self-reported personal belief in a personal god, also consistently rank at or near the top of global indices measuring happiness, life expectancy, social trust, press freedom, political stability, educational attainment and absence of public corruption. Denmark, Sweden, Norway, Finland and Iceland have not descended into the kind of totalitarian horror that the vacuum argument implies must follow from the absence of religious governance. They have instead built some of the most functional, humane and stable societies in recorded history, and they have done so without requiring their citizens to maintain sincere religious belief as a condition of civic virtue or political participation.

The vacuum argument also faces a straightforward and largely ignored historical problem: the most intensely religious societies in history have been perfectly capable of generating catastrophic political violence without any secular vacuum to explain it. The Thirty Years War was fought between deeply religious populations in a deeply religious continent, governed by rulers who drew their authority from theological claims and prosecuted their wars with theological justifications. The Ottoman massacres of Armenian Christians between 1915 and 1923 were conducted by a state with an official Islamic identity against a predominantly Christian minority, in a context of explicit religious differentiation. The Rwandan genocide of 1994 took place in one of the most heavily Christianised countries in Africa, where over ninety percent of the population identified as Christian and where, by documented historical record, church infrastructure was actively complicit in organising and facilitating the killing. The religion was present throughout in full institutional force and with no secular vacuum anywhere in the picture. If religious belief reliably prevents mass political violence by supplying an irreplaceable moral foundation, then Rwanda in 1994 requires an explanation that the vacuum argument simply cannot provide.

For a fuller examination of whether secular societies demonstrate evidence of moral collapse in the absence of religious governance, the essays on the myth of moral collapse without religion and on morality without god develop the positive case at length. The short version is that there is no reliable correlation between a society’s measured religiosity and its prosocial behaviour, and that secular ethical frameworks are more than capable of grounding a robust commitment to human dignity, rights and welfare without requiring supernatural enforcement.

The Third Comeback: “Secular Societies Cannot Sustain Decency”

The third comeback is the most philosophically ambitious of the three. It does not claim that atheism caused specific historical atrocities in any direct sense. It claims, rather, that without a transcendent moral foundation, a society has no principled basis for resisting the slide toward might-makes-right ethics of the kind that totalitarian regimes embody and practise. If there is no God, the argument runs, then there is no objective moral order, and in the absence of objective moral order, the only ultimate currency in political life is power. Stalin and Mao, on this reading, were not aberrations from secular reason but its logical conclusion: if God does not exist, everything is permitted.

This argument has a long philosophical pedigree, and its most famous formulation comes from Dostoevsky’s Ivan Karamazov, whose claim that “if God does not exist, everything is permitted” has been quoted approvingly in theological literature for well over a century. The serious response to it has several distinct components, and each deserves attention.

First, the empirical point already established: if secular societies were systematically and reliably less moral than religious ones, this should be visible in the data across multiple measures of human welfare and social behaviour, and it is not visible. The secular democracies of northern Europe have not concluded that everything is permitted, and their citizens have not descended into nihilistic amorality. They have instead built welfare states that protect the vulnerable, secured civil liberties, abolished the death penalty, established robust legal protections for minorities, and maintained functioning institutions of democratic accountability across multiple generations. The philosophical claim that godlessness entails amorality is flatly and comprehensively contradicted by the lived experience of tens of millions of people in societies that have moved substantially away from religious governance without collapsing into barbarism or moral anarchy.

Second, the philosophical claim itself confuses two distinct questions: the ontological question and the motivational question. The ontological question is whether moral facts require a divine lawgiver to be objectively real. The motivational question is whether human beings require belief in divine punishment and reward to behave decently toward one another. These are entirely separate questions, and both can be answered without God. Moral realism of a non-divine variety, grounded in the verifiable facts of sentient experience, the reality of suffering, the measurable consequences of cruelty and cooperation for individuals and communities, constitutes a defensible and well-developed philosophical position with a literature stretching back to Aristotle and carrying through Hume, Mill and contemporary philosophers such as Derek Parfit and Peter Singer. And even setting the metaethical debate aside entirely, human beings are demonstrably capable of acting on genuine care for others, sustained commitment to fairness, and authentic horror at suffering, without requiring a supernatural enforcer monitoring their behaviour from outside the physical world. The evolutionary and psychological literature on the origins of moral behaviour is extensive and points consistently toward the conclusion that empathy, reciprocity and cooperation are features of our evolved social nature, arising from selection pressures that operated over millions of years before any of the world’s major religions existed.

Third, and perhaps most pointedly of all: theistic governance has historically produced its own ample share of “everything is permitted” reasoning, and has done so on explicitly theological grounds rather than in spite of them. The doctrine of holy war, in various forms across Christianity and Islam, holds that divine command supersedes ordinary moral prohibitions on killing, and that violence carried out in God’s service is not merely permitted but meritorious. The doctrine of heresy holds that the souls of millions are worth the torture and judicial execution of individuals who hold incorrect theological opinions, a calculus that only makes sense within a theological framework. The doctrine of divine election, appearing in various forms across multiple religious traditions, holds that a group specially chosen by God is entitled to land, resources and dominance over those not so chosen, with violence in the service of that entitlement thereby sanctified. Each of these doctrines constitutes a form of “everything is permitted” reasoning, and it is reasoning produced not despite the existence of God but because of the specific claims made about what God requires and rewards. The divine command removes the moral constraint rather than supplying it. If the analytical question is which framework has proven historically more capable of generating rationalisations for atrocity, the honest answer is that theological frameworks have been at least as effective as secular ideological ones, have the advantage of several thousand years of practice, and carry the additional feature of placing their authority beyond rational challenge.

Christopher Hitchens stated the underlying political principle with characteristic precision: “The secular state is the guarantee of religious pluralism. This apparent paradox, again, is the simplest and most elegant of political truths.” The paradox he identified is entirely real and deserves more attention than it typically receives. Only a state that does not privilege any single religious truth claim can guarantee the freedom of conscience that all believers theoretically value and that persecuted minorities have historically needed most urgently. The secular framework is not a threat to moral order or a vacuum waiting to be filled by violence. It is the precondition for the kind of moral and political order that protects everyone, including those whose religious convictions are in the minority. The question of whether atheists can be moral is addressed more directly in the essay on whether atheists are immoral, but the short and empirically supported answer is that they demonstrably can be, and in measurable terms generally are.

The Asymmetry of Doctrinal Causation

It is worth pressing further on the question of doctrinal causation, because this is where the “atheism kills” argument is at its most logically vulnerable and where the comparison with religious violence is at its most analytically damaging. The most useful question to ask of any act of mass political violence is not merely whether the perpetrators held a particular belief among their many beliefs, but whether that specific belief did the causal work, whether removing it would have made the violence impossible or at least substantially less likely to take the specific form it took.

Consider the September 2001 attacks on New York and Washington. Would those attacks have taken place if the nineteen hijackers had been atheists rather than believers in jihad and paradise? The honest answer is almost certainly that they would not have, because the entire operational logic of the attacks depended on specific and identifiable religious claims: that the United States was waging war on Islam, that jihad against that enemy was a religious obligation binding on individual Muslims, and that death in the course of such jihad guaranteed paradise for the martyr, transforming the act from mass murder-suicide into a form of worship. Remove the theology and you remove the motivation, the justification, the recruitment mechanism, and critically the willingness to die in the act. The specific grievance, American foreign policy in the Middle East, existed independently of the theology. But the theology transformed that political grievance from a basis for political opposition into a cosmic confrontation requiring the specific form of violence that occurred. The religion was doing irreplaceable causal work, not serving as incidental decoration on a pre-existing plan.

Now apply the same analytical test to Stalin’s Great Terror. Would the purges have taken place if Stalin had been a devout Russian Orthodox Christian rather than a former seminary student turned Bolshevik atheist? There is very little in the historical or structural analysis to suggest that they would not have. The conditions that generated the Great Terror, the Leninist theory of the vanguard party requiring absolute internal discipline, the logic of Stalinist consolidation of personal power against potential rivals, the institutional momentum of the NKVD’s apparatus, the dynamic by which each forced confession produced a new list of names requiring investigation, none of these depended on atheism as a logical premise or an operational requirement. Remove the atheism from Stalin and leave the Leninist party structure, the cult of personality, the show trial mechanism, and the securitised internal apparatus entirely in place, and the Great Terror almost certainly still occurs. Remove the theology from the Crusades and leave everything else nominally in place, and the Crusades become literally incoherent, because the entire enterprise requires the theological premise of sacred territory, divine command and spiritual reward for violence in the Church’s service. Without those premises, there is no reason to march an army two thousand miles to capture a city, and no mechanism by which popes could have persuaded European nobility to undertake the attempt.

This asymmetry in doctrinal causation is one of the most important points in the entire debate, and it is almost never acknowledged by those deploying the “atheism kills more” argument. It means that attributing the crimes of communist regimes to atheism involves a logical error that attributing the crimes of the Crusades or the Inquisition to theology does not. The connection between theological doctrine and religiously motivated violence is direct, explicit and in many cases textually verifiable in the actual documents that organised and authorised the violence. The connection between atheism and totalitarian violence is, at best, structural and indirect, operating through the mediating mechanism of a specific political ideology that happened to include irreligion among its many other features. A structural resemblance between ideological fanaticism and religious fanaticism is, as already argued, an indictment of fanaticism as a mode of thought, not a defence of religion as an institution.

The Longer Arc: Religion’s Institutional Power and Its Uses

There is a further dimension to this comparison that is rarely introduced into the debate about relative body counts, and it concerns temporal scale. Marxism-Leninism as a governing ideology had roughly seventy years of serious political power, from the Russian Revolution of 1917 to the collapse of the Soviet Union in 1991. Maoism in its most radical and most lethal phase had perhaps a decade of fully unrestrained implementation, from the late 1950s through the early 1970s. Pol Pot’s Khmer Rouge governed Cambodia for less than four years before being driven from power by the Vietnamese invasion of 1979. These are extraordinarily brief periods in the history of human governance, and they were periods in which industrial technology made possible a concentration and intensity of organised violence that previous centuries could not have achieved at the same scale or speed.

Organised religion, by contrast, has been a major force in human governance for several thousand years. The question of what the aggregate death toll of religious governance across that span amounts to cannot be answered with the same numerical precision as the communist figures, because the historical record is incomplete for earlier centuries and because distinguishing religious from secular motivation is genuinely difficult in premodern societies where the two categories were rarely cleanly separated. But the partial catalogue of what is documented and attributable is nevertheless instructive. The Crusades and the Albigensian campaign. The Spanish, Roman and Portuguese Inquisitions. The European Wars of Religion. The early modern witch trials. The colonial violence authorised by doctrines of papal donation and religious hierarchy. The sectarian violence of the Indian partition in 1947, in which between two hundred thousand and two million people died as Hindu, Muslim and Sikh populations were sorted and expelled along religiously defined lines. The ongoing violence in multiple regions of the world with roots in theological claims about sacred territory, divine mandate and religious identity. The institutionalised and systematically concealed abuse of children within religious institutions across multiple countries and multiple denominations, not a body count in the conventional sense but a catalogue of deliberate, serial harm protected by institutional power for decades, enabled by precisely the culture of unquestionable authority and deference to clerical hierarchy that secular critics have identified as a structural danger. Adding the category of deaths attributable to religious opposition to medical knowledge, children dying from faith-healing practices that substituted prayer for medicine, the AIDS fatalities attributable to Catholic opposition to contraception in sub-Saharan Africa, and the mortality consequences of religiously motivated anti-vaccination movements, no final tally is calculable but by any honest reckoning it is enormous.

The point of assembling this partial catalogue is not to win an arithmetical competition in human suffering, which would be morally grotesque and analytically unserious. The point is that the “atheism kills more” argument almost never applies anything remotely like the same standard of causal attribution to religious violence that it demands for secular violence. The moment a consistent standard is applied, the picture becomes far more complicated than the simple accusation allows. Religious institutions have had immeasurably longer to operate, have governed far larger populations over far longer periods, have been far more consistently in a position to deploy state power in the service of their doctrines, and have done so with direct doctrinal justification that is a matter of preserved historical record. That this record exists and is accessible is not erased by any quantity of pointing at Stalin, Mao and Pol Pot, however genuine the horrors those names represent.

The Specific Case of Anti-Religious Persecution in Communist States

One genuine concession that defenders of the “atheism kills” argument can legitimately make is that communist regimes specifically targeted religious believers for persecution, and that this persecution was presented by the regimes themselves in terms of their irreligion. Soviet priests were imprisoned or shot by firing squad. Churches were demolished or repurposed as storage facilities and cinemas. Buddhist monks were sent to labour camps in Maoist China. Clergy and religious figures were prominent among the victims of Khmer Rouge violence in Cambodia. All of this is historically documented and worth acknowledging directly, rather than minimising or explaining away.

But examine the logic driving that persecution carefully, and it becomes clear once again that atheism was not the operative mechanism. Religious institutions were targeted by totalitarian regimes primarily because they constituted alternative centres of authority, loyalty and organised social life that the state did not control. A church, a mosque, a temple or a monastery is a site where people gather regularly under their own organisation, form bonds of communal solidarity, defer to leadership structures that derive their authority from sources independent of the state, and draw on accounts of obligation and meaning that are not generated by the party and cannot be revised by it. For any totalitarian regime, that structural independence is intolerable, not because God specifically threatens the ideology, but because divided loyalty as such threatens a system that requires total loyalty. The Soviet state was equally hostile to non-religious forms of organised autonomous life: trade unions that were not state-controlled, political parties other than the Communist Party, journalists who maintained commitments to factual accuracy independent of party doctrine, scientists who insisted on following evidence regardless of ideological preference. The persecution of religious institutions was one instance of a general and systematic persecution of all autonomous institutions. A totalitarian state would have persecuted equally autonomous secular civic organisations with precisely the same vigour, as the Soviet treatment of independent labour movements amply demonstrates.

The suffering of religious believers under communist regimes was genuine, severe and deserves to be remembered honestly. Insisting on the accurate causal explanation for that suffering is not a way of minimising it. It is a way of understanding it correctly, which is the prerequisite for preventing its recurrence. The cause was totalitarianism’s structural intolerance of any rival source of authority and meaning, not atheism’s philosophical opposition to religion as a set of knowledge claims. Conflating those two different claims does not honour the victims; it simply misidentifies what killed them, and a misidentified cause cannot be effectively prevented from recurring.

What the Argument Is Actually Doing

It is worth pausing to consider what the “atheism kills more” argument is actually doing in the context of public debate, because its rhetorical function is substantially different from its stated logical function. Logically, it purports to be an empirical claim about the comparative death tolls of religious and irreligious governance across history. Rhetorically, it functions as a defensive deflection: a way of stopping the conversation about religious violence before it reaches an uncomfortable conclusion, by immediately redirecting attention to secular violence as though the one somehow neutralises or answers the other. It is, in essence, a tu quoque argument of the kind that logicians classify as a fallacy, and such arguments are characteristically deployed not to advance the truth but to prevent the opposing side from scoring a rhetorical point without engaging with the evidence they have actually presented.

The honest version of this debate would require both parties to engage with the full historical record of violence carried out under their respective ideological banners, to apply consistent and agreed standards of causal attribution to both sets of evidence, to acknowledge the structural conditions that generate mass violence regardless of a regime’s specific ideological content, and to ask which institutional and political arrangements have proven most effective at preventing such violence in the future. That honest debate, conducted with consistent standards rather than selective ones, almost uniformly favours the secular liberal position, not because secular governance is automatically or invariably virtuous, but because the institutional structures that have demonstrably proven most effective at preventing mass political violence, functioning liberal democracy, genuine separation of executive, legislative and judicial powers, an independent press, protection of minority rights, and the rule of law applied equally to governors and governed alike, are not only compatible with secularism but emerged historically in substantial tension with theocratic authority.

The separation of church and state was not conceived as an attack on religion, whatever its enemies have claimed. It was conceived, as Hitchens argued, as its guarantee. It was the arrangement that allowed religious diversity to exist within a single polity without producing the Wars of Religion that had devastated Europe across the previous century. The secular state does not require citizens to abandon their faith; it requires the state to remain neutral among competing faith claims, thereby protecting every faith from the political ambitions of every other. The historical record of theocratic alternatives to this arrangement is not one that religious apologists should be eager to invoke when they are attempting to defend the proposition that belief in God makes governance more humane.

The broader empirical question of whether secular liberal governance produces better outcomes for human welfare than theocratic alternatives is not, in truth, a difficult one to answer from the available evidence. The data from the United Nations Human Development Index, the World Happiness Report, the Freedom House indices of political and civil liberties, comparative international surveys of social trust and institutional integrity, and the academic literature on the relationship between religiosity and various measures of social pathology all point with reasonable consistency in the same direction. Secular liberal governance, with its institutional protections, its tolerance of dissent, its separation of ultimate metaphysical claims from legislative authority, and its structural accountability mechanisms, has produced better outcomes for more people across more dimensions of wellbeing than theocratic governance has managed to achieve anywhere in the historical record. No rhetorical device can change that empirical finding, and the “atheism kills more” argument does not change it. It merely redirects attention away from it, which is precisely its rhetorical purpose.

A Note on Intellectual Honesty

Intellectual honesty requires acknowledging with equal care what this argument does not establish. It does not establish that atheism is automatically or universally morally superior to religious belief as a basis for individual ethical life. It does not establish that all religious institutions are equally harmful or that individual religious practice produces no genuine benefit to those who engage in it. It does not establish that secular governance is immune to the pathological dynamics of authoritarianism, as the history of fascism, which was not an atheist project in the same ideological sense that Bolshevism was, amply and terrifyingly demonstrates. It does not establish that the particular historical forms that secular totalitarianism took under Stalin, Mao and Pol Pot were inevitable or necessary consequences of atheism as a philosophical position, or that a different set of political choices operating within a secular framework would have produced the same outcomes.

What this argument does establish is more modest in its scope but considerably more important in its implications: the proposition that atheism caused the crimes of Stalin, Mao and Pol Pot in the way that theology caused the Crusades and the Inquisition is not supported by the historical evidence, the causal mechanism implied by that proposition does not exist in any traceable form, and atheism as the simple absence of belief in gods carries no political programme that could generate the specific forms of violence attributed to it. The crimes of those regimes were real and their scale was appalling. The attribution of those crimes to atheism as their primary cause is analytically false, and the two facts, the reality of the crimes and the falsity of the attribution, must be distinguished rather than collapsed into a rhetorically convenient but historically empty equation.

The impulse to say “but your side killed people too” is understandable as a human response to criticism. It is considerably less impressive as a logical argument. A critique of religious institutions on the basis of the documented harm their doctrines and practices have caused does not stand or fall on whether other institutions have also caused harm. The question of whether a particular institution’s doctrines and practices have done damage is answered by examining that institution’s doctrines and practices with honesty and consistency, not by pointing across the room at someone else’s failures. A prosecution does not collapse because the defence can identify other criminals. It collapses only if the specific evidence against the defendant fails to establish the specific charge. The evidence for the harm done by religious institutions across centuries of governance is substantial, documented, textually verifiable and direct. The evidence that atheism, as opposed to totalitarian ideology operating in a secular register, caused the atrocities of the twentieth century’s worst regimes is, on honest examination, essentially non-existent.

Conclusion: Argue the History

The “atheism kills more” argument is not, in the end, an argument about history. It is a rhetorical gesture that relies for its force on the audience not examining the history too closely. Once the history is actually examined with consistent analytical standards, the gesture loses all its apparent force. Atheism, understood as the proposition that gods do not exist, contains within it no instructions for operating a gulag, engineering a famine, or declaring a Year Zero in which all previous human culture is to be annihilated. The regimes that committed those crimes drew their operational logic from Leninist theory of the vanguard party, Stalinist consolidation of absolute personal power, and the utopian millenarianism of revolutionary movements that promised earthly salvation through purifying violence. None of those elements requires atheism as a logical premise. When the structural features of those regimes are examined honestly rather than rhetorically, they look less like secular rationalism and considerably more like the authoritarian demand for unquestionable sacred authority that critics of religious institutions have identified as a structural pathology. That resemblance is not a coincidence. It is a lesson about the dangers of unchecked power, not a lesson about the non-existence of God.

The religious violence of the Crusades, the Inquisitions, the European Wars of Religion, the theologically authorised colonial slaughter, and the ongoing sectarian killing of the present century is directly traceable to specific doctrines held by specific institutions wielding specific power. The causal chain is visible, documentable and explicitly recorded in the texts, the papal bulls, the theological treatises, the legal instruments and the institutional structures that organised and authorised the violence. The supposed causal chain from “no god” to “gulag” is, by contrast, not visible in the historical record in any comparable form, because it operates through the entirely separate mechanism of a specific political ideology that happened to include irreligion among its other features. When the same standard of causal attribution is applied consistently to both sides of the historical ledger, the picture that emerges does not vindicate religious governance. It vindicates the secular liberal institutions that arose, in significant part, as a considered response to what religious governance had repeatedly demonstrated it was capable of producing when it held unchecked power over human life.

The honest answer to the accusation with which this essay opened is that the accusation is malformed at its foundation. Atheism did not kill anyone, because atheism is not an agent, a doctrine, a programme, or an institution capable of issuing commands. Totalitarian ideology killed millions, and it did so by replicating the structural logic of religious authority in a secular register, which is an indictment of that structural logic rather than an exoneration of religion. Religious institutions, wielding specific doctrinal authority over human behaviour across centuries and continents, have killed on a scale and with a directness that their modern defenders have not reckoned with honestly. Both of these things are true simultaneously, and holding them both is not moral equivalence between atheism and religion. It is the minimum that honest historical reasoning demands of anyone who wants this conversation to be about what actually happened, rather than about what is rhetorically convenient.

Did Atheism Cause History’s Worst Atrocities?