Stochastic Intimidation
Naming the Invisible Architecture of Modern Political Threats
I’m going to tackle a puzzle that has been bothering me for a long time. Can we do something about this pattern in our political life: A powerful person condemns someone publicly, as a traitor, as a danger to society, as an enemy of all that is right and good. That statement is amplified on television, in other media and virally online. And then someone acts on that narrative frame to deliver death threats to the named individual.
We can scarcely turn around any longer without hearing such stories.
This dynamic contributed to Marjorie Taylor Greene’s break from MAGA. In November, on Truth Social, President Trump labeled the former Georgia congresswoman “Marjorie Traitor Greene.” Then, as The New York Times reported, “Terrified by the ensuing wave of death threats aimed at her and her family from apparent supporters of Mr. Trump, she could no longer see any upside to duking it out in the political arena.”
The dynamic operates further down the political food chain, too. A Houston mother attempted spoke out against the creation of a club inspired by Turning Point USA, the group founded by conservative activist Charlie Kirk, at her children’s high school. According to The Times, the woman believed that “the club could ‘sow division and hate.’” Texas Attorney General Ken Paxton had an answer: “The far-left tried to silence these patriotic young Texans, but we will never surrender.” The mom reported receiving death threats.
I had a similar experience with Turning Point. In 2017, I was added to the group’s Professor Watchlist along with defamatory claims about what I taught. Kirk appeared on Tucker Carlson’s show, presenting my photo and those claims. I got death threats on my phone’s voicemail.
We need to name this problem.
Alexandria Ocasio-Cortez said this of Carlson: “Every time that dude puts my name in his mouth, the next day, I mean, this is like what stochastic terrorism is. ... It’s like when you use a very large platform to turn up the temperature and target an individual until something happens, and then when something happens, because it’s indirect, you say, ‘Oh, I had nothing to do with that.’”
The phrase Ocasio-Cortez used — “stochastic terrorism” — doesn’t get it quite right, but it’s close. It was coined in the early 2000s by a risk analyst, then appropriated and redefined by an anonymous blogger in 2011. It has been used to refer to a pattern whereby degrading or dehumanizing statements by a powerful person create a climate of hate or threat for a group or individual, and then someone acts in violence against that target.
The word stochastic comes from statistics and probability, and it refers to this aspect of the phenomenon: Who acts against the target is unpredictable, but that someone will do so is highly predictable. The killings of Minnesota legislator Melissa Hortman and Charlie Kirk himself and the assassination attempts against Trump are examples of stochastic terrorism. At a certain point, the virality of inflammatory language makes it likely that someone is triggered to act violently, even if predicting who it will be is impossible.
Rand Europe recently did a study of the pattern for a Dutch governmental agency. A goal was to clarify possible strategies for interrupting the resulting political violence, with the hope that naming and characterizing the phenomenon will make that easier. They conclude, though, that the phrase stochastic terrorism falls short. Causal links between an initial speaker, the sources of virality, and the violent actor are too hard to establish.
So here’s another concept: “stochastic intimidation.” We have to distinguish between cases in which the inflammatory speech results in violence and those where what results is intimidation through threats. What clearly exists is a pattern in which a speaker denigrates a target; the denigration goes viral; and someone delivers threats to the target. Who delivers the threat may be random, but that a threat will follow from social media fueled condemnatory discourse is nearly guaranteed. The cases where this pattern results in violence are the tail of the distribution of the phenomenon of stochastic intimidation.
This now very common pattern has driven many people out of public office, introduced a need for more proactive protection of journalists, and frightened ordinary citizens into keeping quiet about issues they care about. Silencing the voices of democratic citizens is the purpose.
We have laws against voter intimidation, against witness intimidation, and against threats to public officials and their families. In other words, where speech or action degrade a function necessary for the healthy operations of democracy, we have protected those functions. These laws are generally consistent with the First Amendment because true threats and certain intimidation and harassment conduct are not protected speech. It might be tempting to think we could also have laws against journalist intimidation, election worker intimidation, and citizen intimidation. There is a chance we could pull off the first two, modeled on voter and witness intimidation. Indeed, since 2020, several states have passed laws against election worker intimidation.
But the law won’t help us protect citizens generally from verbal intimidation and threats. That’s a game of whack-a-mole. One powerful speaker, connected to mechanisms of amplification, can generate hundreds of threats. Prosecuting all of them would be impossible. We have no legal way of braking the behavior of the amplifiers; and the powerful speaker is protected by First Amendment rights and often has the cover of plausible deniability. We will need to look for our solutions elsewhere — in institutional norms and practices and in our civic culture.
In the late 1980s and early 1990s, a rash of shootings in post offices by disgruntled employees terrified the nation. This is where the phrase “going postal” came from.
The U.S. Postal Service was able to interrupt this pattern. It had to acknowledge a toxic, abusive management culture. Then the combination of a Congressional hearings and a USPS Commission process yielded a multi-prong strategy of response that included: “stepping up the current violence prevention program, improving job applicant screening, increasing training to improve the interpersonal skills of managers, and overhauling the dispute resolution process.” USPS also had to address a massive backlog of employee grievances, reducing a backlog from 135,000 cases to 15,000.
Of course, tackling violence inside of a single bureaucracy is a substantially different project from addressing an emergent property of a distributed social system of communication and grievance. But some lessons might apply. The USPS didn’t solve workplace violence primarily through prosecution — it changed the culture, created early warning systems, and restructured grievance processes.
If the only civic culture that people have is one of tribal combat, the pattern of stochastic intimidation will keep recurring. Building a thick civic culture of pluralism, disagreement without dehumanization, and reflective patriotism is a generational project, but can address the root causes of this problem. That’s the culture change we need, and elites can help it along if they break ranks with instigators and name the instigation of stochastic intimidation as an unacceptable practice.
For early warning systems, we should establish social media platform design obligations such as a duty to interrupt amplification cascades. In other words, if we make platforms responsible for the virality that they cause, we can interrupt it when they are accelerating intimidation.
As to restructured grievance processes, we have to acknowledge the limits of criminal law in addressing stochastic intimidation — it can’t touch the instigator or the amplifier. But we might also ask whether civil liability frameworks could interrupt this pattern. Perhaps we could create a civil cause of action, modeled on tortious interference in the business context, to let a target sue either the powerful speaker or the amplifier who engage in a pattern of public communication that they know or should know will foreseeably result in intimidation against an identifiable target. In business contexts, you can be liable for predictable harms your conduct causes to third parties. Platform companies might themselves also be sued for civil liability on the model of tortious interference.
Greene said a threat aimed at her son was particularly decisive for her: “Derek will have his life snuffed out soon. Better watch his back.” According to The New York Times, the email came with a subject heading using Trump’s new nickname for her, “Marjorie Traitor Greene.”
Greene then texted the president, and The Times reports: “According to a source familiar with the exchange, his long reply made no mention of her son. Instead, Trump insulted her in personal terms. When she replied that children should remain off limits from their disagreements, Trump responded that she had only herself to blame.”
A White House spokesperson offered this comment on the situation: “President Trump remains the undisputed leader of the greatest and fastest growing political movement in American history — the MAGA movement. On the other hand, Congresswoman Greene is quitting on her constituents in the middle of her term and abandoning the consequential fight we’re in — we don’t have time for her petty bitterness.”
Stochastic intimidation is not a pretty name for the phenomenon that Trump and his staff here pretend to erase through redirection. But it is a real phenomenon. Its power to terrify and silence citizens is only spreading and growing. We must interrupt it.
Stochastic intimidation is one of the president’s most powerful tools, but he is by no means alone in wielding it. Figures on both left and right, holding both high public office and that highest office of mere citizen, are suffering harm from it. The very staff who offer absurd statements such as the one above from the White House press office probably suffer from it. Odds are that their own effort to find a way to live with a constant barrage of threats is what leads them to pretend that the barrage doesn’t exist. That pretense might help them endure. It doesn’t help us all end the onslaught.
Here’s what will help: We have to acknowledge that the problem exists, that it has a name — stochastic intimidation — and that we need to stop it. I wonder what it would take to get us some Congressional hearings on stochastic intimidation and a commission on how to interrupt it.
Stochastic intimidation. Not a pretty name. Not a pretty thing. But a definite pattern to interrupt.
Make sure to check out these new announcements for The Renovator Community! And join us next week, Feb. 25, at 12:15 P.M. EST for Danielle’s Substack Live with Bill Kristol!



You're certainly right that we should use the existing legal categories to the full extent that we can. However, they leave a gap. The effort to treat a distributed, multi-actor system via simple two-party legal categories routinely leaves the instigator untouchable. This is the core of what I am trying to name. The problem involves three distinct actors: (1) the powerful speaker who denigrates, (2) the platforms and media that amplify, and (3) the anonymous individuals who deliver threats. Libel, incitement address the relationship between a single speaker and a target but we are routinely seeing that existing law doesn't easily reach the instigator, because the powerful speaker can typically maintain plausible deniability. The causal chain runs through amplification and probability instead of direct instruction. I think there are real gaps that we need to address.
Yes, yes, but how to make such a system that's in any way "fair"? Remember, Trump's survived assassination attempts, and he will surely claim that all the people saying he's a dictator are encouraging more assassination attempts.
And there is a very scary leftist version, with all the rhetoric of "punch Nazi" and "kill TERFs" and similar. See, for example:
https://www.thefp.com/p/jesse-singal-bluesky-has-a-death-threat-problem
Should that journalist be able to sue Bluesky? His more powerful haters?
The problem with "break ranks with instigators and name the instigation of stochastic intimidation as an unacceptable practice." is that the hate-mongers don't care.