The Crowd as the Weapon

On the evening of 4 August 1996, roughly two thousand people marched from a mosque in Gatesville, on the Cape Flats of Cape Town, to the Salt River home of Rashaad Staggie, a leader of the Hard Livings gang. They beat him, shot him, and set him alight in the street; he burned to death in front of a row of television cameras. The footage ran on the national news that night, and anyone living in Cape Town then can still describe it. On its face it was the purest image of spontaneous popular justice — a community pushed past endurance by a gangsterism the state had failed to control, taking the law into its own hands.¹

What the cameras did not show was underneath it. The organization that led the march, People Against Gangsterism and Drugs, had been infiltrated by South Africa’s National Intelligence Agency; a former agent later confessed to a state campaign against the movement, including an arms cache used to steer and discredit it, and the group’s own leadership accused intelligence of being behind bombings and killings later pinned on them.² The documented core is narrow and certain: a recorded mass killing, carried out by a crowd, run by a movement the state’s intelligence service had penetrated and was working to direct. Whether intelligence manufactured that particular mob remains contested, and it is held exactly there. But the mechanism the case exposes is not contested at all.

The operation has two moving parts, the same two tracked separately, now bolted together. First, a target is marked: named, framed, made into the thing the crowd already fears, through a trusted channel the audience does not think to distrust. Second, the crowd does the violence, so no order need ever be traced: on paper nobody ordered anything, there was only popular feeling, and it acted. It is the firehose of the previous chapter given a body.

The fullest documented version, with convictions attached to both parts, is Rwanda in 1994, where the marking was done from two channels at once. The radio station RTLM read the names and locations of Tutsi to be found and killed, and called them, night after night, cockroaches and enemies to be exterminated. In the “Media Case,” the International Criminal Tribunal for Rwanda convicted RTLM’s founder Ferdinand Nahimana, its board member Jean-Bosco Barayagwiza, and the newspaper editor Hassan Ngeze of incitement to genocide, conspiracy, and crimes against humanity — establishing in law that marking a population for a crowd to destroy is itself the crime, distinct from swinging any machete.³ The pulpit did its share directly: the Catholic priest Athanase Seromba was convicted by the same tribunal, his sentence raised on appeal to life for committing genocide, after he had the church at Nyange bulldozed onto the roughly fifteen hundred to two thousand Tutsi who had taken refuge inside it.⁴ The narrative marked; the crowd, the Interahamwe and ordinary neighbors, carried out the killing. Neither half could have produced a genocide alone. Together they produced one that its organizers could, and did, call a spontaneous eruption of ethnic hatred.

Where an intelligence service wants a crowd but cannot rely on one forming, it manufactures one directly, and has done so on the record. The clearest case is the CIA’s 1953 operation to overthrow Iran’s elected prime minister, Mohammad Mossadegh. The agency’s own declassified history is explicit: CIA officers hired some of Tehran’s most feared gangsters, and paid and bussed in demonstrators, to stage the “pro-Shah” riots then reported to the world as a popular uprising against Mossadegh; the station chief Kermit Roosevelt organized the mob and let it wear the face of the Iranian people.⁵ The same logic ran domestically through the FBI’s COINTELPRO, which did not always need to supply a crowd because it could set two existing ones against each other. The Bureau forged letters between Black Panther leaders, and between the Panthers and a rival organization, specifically to incite them to violence and, in its own internal language, to murder — turning targets into each other’s mob while the Bureau’s hand stayed in a filing cabinet until the Church Committee pried it out.⁶ The staged pogrom is the oldest form of this: the Nazi state orchestrated the November 1938 violence against Jews across Germany, Kristallnacht, and presented it to the world as the spontaneous anger of ordinary Germans.⁷

The trusted channel need not be a radio or an intelligence front; the most effective one, historically, is the pulpit, because religious authority can frame a target not merely as dangerous but as damned — a license no secular propaganda can match. The Russian Orthodox Church under Patriarch Kirill has sanctified the invasion of Ukraine as a holy war and a metaphysical struggle, lending a state’s aggression the one endorsement its population is least equipped to question.⁸ South Africa’s own Dutch Reformed Church supplied the theological justification for apartheid for decades, until the World Alliance of Reformed Churches declared that justification a heresy in 1982 and the Belhar Confession was written against it — a marking of a whole people as divinely ordained subordinates, issued from the pulpit and believed from the pew.⁹ The reach is not a modern or a foreign problem: in the United States of the 1930s, the radio priest Charles Coughlin broadcast antisemitic and pro-fascist narratives to an audience estimated at thirty million, carrying the authority of the collar into tens of millions of homes.¹⁰

Every case here shares the same deniability — the reason the mechanism is chosen at all. When the enforcement is done by a crowd, there is no order to subpoena, no operative to flip, no chain of command to reconstruct — only a population that, it is said, felt strongly and acted, or a faith that, it is said, spoke its sincere conviction. The hand that marked the target and the hand that lit the match are, by design, never shown holding each other — and that is exactly the point, whether the marked man burns on a Cape Town street or a country burns to a voice carried over the radio. Reconstruction is the only answer — because the crowd’s violence is visible to every camera present, while the hand behind the marking is visible only to whoever is willing to look for what the footage left out.

Reconstruction is also the discipline that separates a documented case like this one from its paranoid double. The claim that a crowd was secretly steered against a person is only ever as strong as what can be shown from outside that person’s own account — a seeded rumor caught in the act, a paid organizer, a record a stranger can verify. Where the directing hand can be documented, as in the cases here, the mechanism is real; where the only evidence is the strength of the feeling that one has been mobbed, it is exactly the belief — clinically, a delusion of persecution — that this book is at pains never to feed.