The quotation marks were doing more work than I expected. When I saw the title, I first assumed that this would be an essay about why we shouldn't engage in stochastic terrorism.
Anyway, I propose my own rule:
If you're accusing someone of conduct that might plausibly lead to someone committing violence against them, such as killing 700,000 people or attempting to destroy the American way of life, you should consider that you have a particular moral duty to check your facts and assumptions.
So lets say you are going to post a social media rant that suggests that cops murder purpletonians with impunity, and that neither purpletonians nor anyone who cares about purpletonians should ever call the cops to report a crime as a result. If this statement is reasonably likely to be true, then it's an important thing to warn people about! But if it's not, it might lead to violence against cops, or it might lead to crime victims not being protected from criminals. So while you might normally say it as part of normal exaggeration or because it's "directionally true" or because it fits with your existing assumptions, the possible downstream effects should raise a moral duty to do some extra fact-checking.
I basically agree with this, except that I think you shouldn't do the "directionally true" thing in any case. Maybe the violence-inciting bit adds some extra urgency.
The higher the stakes, the *more* certain you should be before airing the claim, not less. A claim that Haitians are killing and eating people's pets has far more potential for harm if it's false than a claim that a Haitian man was rude to you at the coffee shop.
I don't see why we should treat those claims identically just because they're in the same "direction" of expressing dislike for Haitian immigrants.
A claim that "some Haitians are stealing and eating pets" doesn't justify violence against Haitians in general. It justifies violence maybe against the theoretical subset of Haitians that did steal and eat other people's pets, but if that subset doesn't exist then, other than smearing the image of Haitians, it's not advocating for violence.
Contrast - a claim that <politician> is destroying democracy, ending elections, locking up <minority> and putting them in concentration camps - if true, this completely justify the assassination of said politician if you believe said politician cannot be stopped through normal means, or even creates the moral imperative to do so.
Do you think we have the moral imperative to murder Hamas politicians? They would seem to qualify (in that they actively support (and do!) murder gays*).
*unless, of course, you think that... genocide is different from putting people in concentration camps?
Not all human communities hold the same view of "Pets" vs "Food." I know where to get dog in Shanahai, and horse was routinely on the breakfast buffet at my favorite hotel over there. The fact that neither appeal to me, so I don't eat them, is not an excuse to express surprise and revulsion that someone has different food preferences.
Denial that other humans Do have different food preferences is every bit as racist as an expression of revulsion at someone else's preferences, absent cases of deliberate cruelty or theft of the critter in question.
Someone's willing to document a tradition of eating cats. Apparently eating dogs is utterly taboo.
I don't expect that this would be a tradition kept in the United States (unlike, say, the Hmong practicing ritual animal sacrifice as part of culturally approved medicinal practice), and even if it was*, cats are practically free from shelters (and stray cats don't require paperwork).
*over a million Haitians (or native born of Haitian origin) in America? Is it possible one of them decided to do a culturally "okay" practice for back home, where it is "not okay" here? Sure, okay, maybe one person (said person is probably on the lower side of intelligence and cultural assimilation).
Indeed. Without taking a position on whether or not Haitians were/are eating people's pets, the idea that it's *obviously absurd* that people from a radically different culture would regard animals like cats and dogs as food animals rather than pet animals strikes me as itself obviously absurd.
Oh, I absolutely agree, I did enjoy dog meat in my childhood and bemoan its scarcity these days. Pigs are just as intelligent as dogs, after all.
To clarify, I'm saying people who steal pets and eat them deserve to be met with violence, not that people who eat <animals which are commonly considered pets> deserve to be met with violence.
I think you overestimate the rationality of those people who commit violence in response to stochastic terrorism.
I think that if I were the sort of hooligan who could easily be spurred to violence against Haitians, then I would not distinguish between the statements "some Haitians steal and eat pets" and "Haitians in general steal and eat pets".
Much, much less than half of political rhetoric is directionally true. The fact that so many reasonably-intelligent people implicitly accept the absurd assumption that there are only ever two possible responses to any issue is the kind of thing that (almost) makes me wish the paperclip maximizers would hurry up and get here already.
What does 'directionally true' mean. Google said 'moving in the right way'. And I'm like, how does that have anything to do with what is really true. (and of course who is defining 'right'?)
"Directionally true" is basically "fake but accurate" or whatever the old defense of playing loose with the facts was. When you get the details wrong, but you're, say, slandering a bad person, so it's you move peoples opinions in the direction you want them to go. Obviously highly subjective and basically begging the question.
For what it's worth, my observation is that a bunch of the people I talk to seem to have the implicit belief that fact-checking accusations against someone they know to be bad (Epstein, for example), is morally suspect.
I don't think they have that *explicit* belief, but they seem to have a strong disgust reaction when presented with evidence that a particular accusation or statement in that direction is incorrect, and I suspect that applies in their interior thought processes too. They know (correctly) that Epstein was a bad guy, so it feels immoral to fact-check statements that support the conclusion that Epstein was bad.
I agree with Scott that this is bad, but what I was trying to say was that even if this is your attitude, I would still argue that you should apply extra caution when your statements about Epstein are likely to cause harm. (For example, accusing a possibly innocent person of conspiring with Epstein or accusing him of being part of a larger pattern of Jewish pedophile conspiracies.)
It means "Well maybe Bad Guy Politicians didn't call for all Purpletonians to be hanged from lampposts, but this guy went out and beat up a Purpletonian, and that is because of the climate of hate and fear against Purpletonians, and that climate is encouraged and created by the likes of Bad Guy Politician who said 'there has been an increase in crime by Purpletonians who deal drugs', so in the end it *is* all Bad Guy's fault the other guy felt free to go out and beat up Purpletonians!"
The term may be recent, but the underlying views go back as far as you'd like to go. See for example the attempted assassination of Gabby Giffords in 2011. In the immediate aftermath, Sarah Palin was blamed for encouraging the shooter by using a crosshairs ad in 2010 as part of a campaign about fighting elections against perceived vulnerable Democrat-held seats:
"Crosshairs is a political phrase that emerged from Palin's political action committee SarahPac that targeted congressional districts for the Tea Party campaign in the last election, including the district of Rep. Gabrielle Giffords.
Although Palin later denied she meant the graphic over the districts to look like a gun sight, it is part of the hunting lexicon that critics say she prefers.
Comedian Frank Conniff tweeted: "Hey, Sarah Palin, hows that hatey, killy, reloady, crosshairsy thing working out for ya?"
Facebook executive Randi Zuckerberg said many people on the social networking site are asking whether Sarah Palin is to blame."
The shooter turned out to be probably mentally ill and also shot six other people during the attempt, but the publicity was all about "Right-wing Republican rhetoric encourages shooters!"
"The US was tonight seized by a fierce debate over whether inflammatory rightwing rhetoric was to blame for a shooting spree in Tucson, Arizona, that targeted congresswoman Gabrielle Giffords and left six dead, including a nine-year-old child.
Giffords, 40, who remains critically ill in hospital after being shot through the head but is expected to live, criticised Sarah Palin last year for putting her and 19 other Democrats on a hitlist of districts, each shown as being in crosshairs.
"When people do that, they have got to realise there are consequences to that action," Giffords said. Palin today distanced herself from the shooting, as did leading figures of the Tea Party movement. Conservative bloggers accused liberals of seeking to exploit the attack."
I mean, I hate Sarah Palin as much as the next guy, but this blame-mongering is ridiculous. We had a Labour MP in the UK who was accused of "betraying" the leadership (Jeremy Corbyn) and responded by saying "I wouldn't stab you in the back, I'd knife you in the front", and a bunch of right-wingers tried to get people to treat this as a death threat. Absurd.
If I say "group X wants to murder me and everyone like me", when the truth is that some member of group X might have been a little less than perfectly polite to me once, what I said is "directionally true"; that is, there is some other, true but entirely unremarkable and useless for whatever rhetorical point I am trying to score, thing that could have been said, but instead I chose to say something literally untrue then rapidly backpedal and point to the true thing when called on it.
It means "false, but will cause the casual listener to believe something else that I think is true".
So, "Hitler was cruel to anmials" is directionally true. It's absolutely false, but it leads people to believe that Hitler was a cruel person and Hitler was a cruel person.
It depends on the number of dimensions. The more there are, the harder it is for two vectors to line up. For an extreme example, consider the semantic space of LLMs. 'Stone', 'cat' and 'river', say, are all practically orthogonal to each other.
Political arguments don't have infinite dimensions, but they could reasonably have more than one or two.
It's possible to have reasoned opposite positions on Haitian immigration that are both opposed to 'they're eating the pets'.
Say in your location it rains 80% of the time. If most people think it rains only 50% of the time, then saying "it always rains" is directionally true, even though it does not rain 100% of the time, because averaging it against the current belief moves everyone closer to the true statement. However, if most people think it rains 90% of the time, then "it always rains" is now directionally false, because it actually rains less than that.
Since this is so dependent on what most people think to start with, which things people think are directionally true largely depends on their social group, which makes it unreliable.
Oh, I agree, I was just trying to say whatever your usual standard for fact-checking is, it should be higher if you're accusing someone of something that would justify violence against them.
Sounds a bit like proposing to expand the definition of what counts as "per se" defamation to things that might get you killed if someone believed them.
I like how purpletonian is supposed to kind of elevate the argument to its most rational and general, shedding any prior notions, but there's only one demographic at least in America known for being killed by the cops so its entirely rhetorically superfluous, I was amused by this and good post
The entire Internet-smart-person thing of pretending to abstract away preconceptions by recasting arguments as "If group A says B about group C" -- which is a) no simpler and b) not fooling anybody, since it usually follows a posting which A, B, and C are all clearly specified -- infuriates me.
I mean the intent is good, the intent being to emotionally depolarize a statement so we can take out our syllogisms and our prepositions and actually get to work. Scott does it in this article where he brings up "stochastic terrorism" accusations from all sides of the aisle
The intent is fine; I can't imagine being a person who it works on, or being a person who thinks it can possibly work on anyone. If such people exist, obviously I'm the one failing at theory of mind.
I feel like I remember it working on me sometimes. It's kind of a reminder to (if you want to think of yourself as fair) think about it more generally. Like it triggers your conscience, at least if you're constitutionally aligned that way anyway.
I’m curious: what do you mean by “being a person who it works on”?
Because I have no trouble imaging a person who has some negative emotional response to, say, some abortion doctors stating fetuses aren’t full humans and have no value, but then a lower emotional response to group A says B about C where the context implies it could easily be abortion doctors not thinking fetuses are full persons. (If I am clearly missing the key point of what you mean, please enlighten me.)
In general if you value abstract logic and the ability to talk about any hypotheticals/principles/etc., but also have negative emotional reactions, biases and preconceptions, then abstracting away can have value. Such values are true of myself as far as I can tell
I was mostly trying to minimize any implication that left wingers are particularly guilty of irresponsible statements that are also false.
I don't think removing the parts of your argument that are extraneous and more likely to be inflammatory or derail the discussion is bad (obviously, I guess, since I did it.) You and gus are right that anyone with basic knowledge can decode it, but I think generalizing it is at least inoffensive, and probably slightly good.
I think I want to see an analysis of where this standard comes down on the Nuremberg Files website, which provided private identifying information about abortion providers, and was widely understood to serve the purpose of enabling harassment of those people, and was known to be circulated on message boards where people were saying that abortion was murder and that killing those who performed abortions would be virtuous (even if none of them was dumb enough to state, even anonymously, that they personally planned to murder anyone).
Like, we have court doctrine about "true threats", and we have some ideas about where people have the "expectation of privacy". I largely agree with the thrust of this post, but is there _some_ kind of line, short of explicitly doxxing somebody and saying "hey folks, you should go murder this guy"? I get that this is a tricky concept, and "I know it when I see it" is not a satisfying standard (because anyone can decide they see it in all criticism by their enemies, and none by their friends). I'm not sure how you build a workable standard that doesn't immediately slippery-slope into stupidity.
Yes, this is exactly the character of the problem.
I fundamentally agree with Scott that the accusation of engaging in stochastic terrorism has expanded drastically beyond the bounds of what is justifiable, and is used to try to chill speech that the person making the accusation doesn't like.
At the same time, it still seems like stochastic terrorism is a real thing, I can point to examples where I believe we all ought to agree, "Yup, these people were engaged in speech that had the clear intent of getting people hurt or killed."
I guess I kind of agree with None of the Above that a decently written law, that could be used as a standard by a jury, might be the right way to deal with this. Proving the mens rea for this crime would be a high bar, but surmountable in extreme cases (like the Nuremberg Files). At least, assuming it's possible to find twelve normie citizens without getting at least one or two who are hyperpartisan cultists for either side of the argument that the stochastic terrorist was trying to influence through violence.
We have this, right? Credible death threats are assault, incitement is a crime, conspiracy to commit other crimes is a crime. But none of those lets you do anything to the guy who says "f--ker had it coming" about Charlie Kirk or George Floyd, or who says "From the river to the sea" or "Blue lives matter."
I try to censor in my comments any comment that reveals too precisely where anybody lives. For example, one commenter had a funny story about riding the elevator with Monica Lewinsky in the XYZ Apartment Building, but I spiked the comment.
With me, this goes back to the 1980 murder of John Lennon outside his Dakota apartment building.
John had doxxed himself. Around 1979 he'd gotten over his drug problems and started doing witty interviews again. One of his favorite jokes was how he'd written the line "Imagine no possessions," but now he owned several apartments in The Dakota.
Yes, I agree it's bad, and I don't think people use it to defend their own ideas, so maybe it's unfair.
However, my impression is a lot of people *do* limit their bs detector or fact check etc. if a particular fact implies something they believe to be true. So for example, if you're confident that President Camacho is a scumbag, then maybe you retweet a story that the Pope gave a sermon scolding President Camacho without checking whether it's a hoax or not.
Not sure I understand your point. Yes, people definitely subject claims they dislike to greater scrutiny than claims they like. And the fact that we do this is bad; it leads to widespread distorted judgment on many important topics, and undermines important norms of truth, honesty, and evidence.
If you're saying "when stakes are high, you should try EXTRA hard to avoid this bad thing that you should actually ALWAYS avoid but that most people aren't ethical enough to consistently avoid" then, sure, I guess.
Technically, I'd argue that people should spend a rational amount of effort fact-checking their beliefs, and that the rational amount of effort should be higher when they are publishing a belief that is more likely to lead to harm.
Also, since not everyone is a rationalist, I'd be happy with a moral value that it's OK to criticize insurance executives and Chinese exchange students and Israelis, but when the criticism might cause harm, there's a particular moral duty to do your homework and not exagerate.
I think of "directionally true" as being like the lies/approximations that we tell children.
"There are two sorts of chemical bonds -- covalent and ionic" is the sort of directionally-true thing you say to introduce an eleven-year-old to basic chemistry. If they continue to study chemistry then eventually they'll learn that it's more complicated than that.
The difference is, we tell those things to children because smoothing over the intricacies of complicated subjects is a necessary part of teaching anyone anything.
But when people use the "directionally true" line in politics, it's pretty much always either because they are too lazy to engage with the complexities or too cowardly to admit that they were wrong or (most commonly) both. When expressing political opinions to adults you *can* just qualify them with things like "well, it's a complex issue, but I expect [Trend X]" to broadly hold.
If instead of doing that, you bloviate about how Those People Are Ruining Everything, then you deserve exactly zero credit for getting the directionality right *even if* it turns out that Those People are Sometimes Ruining a Few Things. And in reality, claims of "directional correctness" basically never actually wait for better data to come in: they just beg the question while retreating to a more defensible motte and waiting for people to get distracted and move on.
Fair, and since I am both lazy and lack influence, it will be a heavy lift!
I'd love to see it get a catchy name and see people scolding the other side with something like: Ok, we get that the Frankist child grooming conspiracy would be essential to reveal if true, but since it's both (a) a dangerous accusation and (b) you verifiably haven't done your homework, you should be shamed.
Sadly, I think there are too many people around who care little or nothing for our shaming. Might even be proud of it. I think some level of restriction of speech is the only consequence that might moderate the effect of such speech on innocent victims.
I literally just wrote an essay on this! I take a slightly different tack. I'll share here when it's published, but in the meantime here's the gist which I think is a critical lens you didn't drill into enough.
Stochastic terrorism, the worst and non-imaginary instances of it, can cause very serious even if usually very indirect harm. Like pretty much all speech, it should be legally free to express. But we can and must treat overt hateful bigotry and the endorsement of violence as a violation of our universal moral TABOOS, with principled ostracism and social consequences to push this hate and the people who espouse and sanitize it out of polite society. This is how we marginalized and defeated the KKK. If we don't, that hate will normalize and turn into violence, as we are seeing.
I love the 1A and I loathe cancel culture, but we have to make certain speech and behavior socially unacceptable, and I believe we can do that in a principled way that aligns to our universal, liberal values. I wrote more here criticizing Greg Lukianoff and FIRE for their over-simplified "free speech culture" absolutism: https://whiterosemagazine.com/the-dangerous-myth-of-more-speech/
I think the first KKK was put down by the military occupation, the second by its own scandals, and later iterations were infiltrated by the FBI.
Your linked post makes some fair points, but also some lousy ones.
> Greg often asks, “Am I safer for knowing less about what people think?” Of course not—I want to know exactly who the bigots are. But we can monitor the fire without letting it burn in the living room
Speech does not burn your living room, just as it doesn't break your leg or pick your pocket (as Thomas Jefferson put it).
> Granting an unrepentant bigot the same professional standing and platform as those who uphold our shared humanity isn’t “openness”—it’s a subsidy
No, neutrality is never a "subsidy". I think that speeding cars should have traffic fines, but it's not a "subsidy" to not have any speed limits.
> This is the Paradox of Tolerance: if you extend unlimited tolerance to the intolerant, the tolerant end up destroyed
The United States has a First Amendment which protects the speech of all political ideologies, even the most tolerant. There were times when that was less protected (such as Schenk v. United States), but since SCOTUS had started protecting it more, the US has not gotten destroyed but instead kept trucking. In contrast, Weimar Germany actually did have some laws we would now categorize under "hate speech" protections, that would be prohibited under our First Amendment, and yet they fell to the Nazis. The American approach of prohibiting actual violence rather than speech appears to have worked out better, regardless of what people might imagine in theory.
> We need to stop pretending that “freedom of association” only protects the speaker
Fair enough. Does FIRE actually deny that?
> A truly free society doesn’t require us to be indifferent to the destruction of our fellow citizens’ dignity
I agree with Steve Pinker's "The Stupidity of Dignity". We can make consesquentialist arguments about what makes people better vs worse off without bothering with such notions as "dignity".
1. "the first KKK was put down by the military occupation, the second by its own scandals, and later iterations were infiltrated by the FBI."
There were numerous reasons the Klan was defeated, but for the Third Klan especially, their hate was seen as TABOO and socially unacceptable and that absolutely limited their ability to recruit and make their ideology normalized. You couldn't be an open Klan member and be openly welcomed in most of polite society. You couldn't even be friends with an open Klan member. Their masks were to hide their faces and identities.
2. To your points re "burning" and "subsidy", those are analogies. I'll grant they're imperfect (like all analogies) and easy to poke holes in. I won't spend time elaborating here. Same with "dignity", which I really mean "letting people live without the fear that they'll face random violence or discrimination that is actively and openly encouraged by many people in society without consequence."
3. "The United States has a First Amendment which protects the speech of all political ideologies"
I'm only referring to social/private consequences, not ones imposed by the state. The latter is what the 1A protects from, and I support it fully for all the reasons you listed.
4. "We need to stop pretending that “freedom of association” only protects the speaker Fair enough. Does FIRE actually deny that?"
Yes, FIRE would and does argue that if for example somebody was fired for celebrating Charlie Kirk's murder after a mob demanded that they get fired, that would be cancel culture, and is thus bad/wrong. Greg used the broadest possible definition and continues to double down on it, not acknowledging that there are and should be material social consequences for certain speech that society broadly (nearly universally) believes is unacceptable and antithetical to our liberal democratic values.
If we do not uphold this speech as TABOO, it will undoubtedly normalize, and that will undoubtedly lead to catastrophe. And we can do it in a principled way. We must, or we are done for.
It is not covered by the first amendment and should not be, but a common norm of "the boss will fire you if he doesn't like your off-the-job political speech" would yield a massive reduction in peoples' ability to speak their minds, since only retired and independently wealthy people would dare take public positions that their employer disapproved.
This is why I created principles that limit the ability for this to succumb to the excesses of "cancel culture". It's not just "opinions I don't like". It's only when people cross a very specific line and refuse to clarify or apologize.
4 Principles for Restoring our Universal Moral Taboos
1. The Red Line: Limit actionable taboos to overt bigotry, dehumanization, and the endorsement of violence.
2. The Consensus Test: Distinguish between subjective offense (which is partisan and open to individual interpretation) and a Shared Moral Violation (which reflects a broad, trans-tribal consensus and warrants consequence).
3. The Private Mechanism: Enforce standards through civil society (employers, associations), never government coercion.
4. The Open Door: The goal of consequence must be correction, not permanent destruction; always offer a path to redemption.
The obvious placethis breaks down is the consensus test (many high profile cancellations during the Great Awokening were enforcing a moral consensus of a small subset of adademic, media, and tech employees.
In 2020, what fraction of people in the US would consider deadnaming a transperson to be a serious infraction that ought to lead to someone being fired? What fraction of Americans find someone saying "serves the bastard right" when they heard about Charlie Kirk being murdered a serious infraction of this kind? I think you are looking at >25% of people who would find neither of these very offensive, and yet people in a position to enforce consequences for speech were very happy to punish both when they had the chance. ISTM that doing this almost always ends up involving the consensus of some powerful people, whether or not that reflects a society wide consensus. And indeed, it looks to me like a lot of cancellation campaigns, including the woke ones and the Charlie Kirk ones, were tranparent attempts to impose a small group's consensus on everyone.
The point of the consensus test is that almost nothing reaches it. Only overt, hateful bigotry and the endorsement of violence, things that directly go against our core universal values of liberty and equality. Nearly everything else is tribal and should be rejected.
"cancelling" the people who celebrated Kirk's murder was absolutely the right thing to do, as ~99% of Americans believe political violence for speech is unacceptable. So somebody endorsing that political violence is violating our universal moral taboos. Conversely, somebody who said "I don't care that he's dead, he was an asshole" was being insensitive and crass, but that is not a universal taboo, and thus they shouldn't get fired.
If we don't fire people who explicitly endorse Hamas, terrorism, political violence, etc. that shit will normalize and spread like wildfire until we have mass anarchy and violence. It doesn't mean 20% of Americans will hold those views. But even a small minority empowered to think their views are socially acceptable, and are being encouraged to burn it all down, will try to do exactly that, in big and small ways.
Klan members wore masks to hide their identities in their first iteration, because they were insurrectionary terrorists fighting a military occupation. The second Klan were cosplaying as terrorists.
> I'm only referring to social/private consequences, not ones imposed by the state. The latter is what the 1A protects from, and I support it fully for all the reasons you listed.
You may be, but Popper was explicitly talking about a government and the two questions of whether it should have the legal right to suppress certain speech, as well as whether it would be prudent to do so if it had such a right (most people bringing up his supposed paradox forget his more nuanced take on the latter question).
Empirically speaking, I just don't believe that "we are done" if we don't stigmatize the celebrations of death/murder via firing. Solving for the equilibrium, I think that firing people in response to a mob of outsiders getting mad about their off-the-clock speech just incentivizes "offense archaeology" trying to dig up things to be mad about and spreading them. Instead, the best response to bad speech is not firing, but good speech counteracting what they say. https://slatestarcodex.com/2013/12/29/the-spirit-of-the-first-amendment/
I just asked Gemini if Popper was talking only about government interference:
"No, Karl Popper was absolutely not speaking only about government interference.
The popular modern internet meme version of the Paradox of Tolerance often formats it as a neat justification for immediate state censorship or legal bans, but Popper’s actual text in The Open Society and Its Enemies (1945) argues the exact opposite. He viewed state suppression as a last resort, placing the primary responsibility on civil society, public opinion, and social norms."
I understand where you're coming from encouraging free/more/good speech, and I'm on board with that too. But that's not always an option, especially in the internet age. Do you think the people celebrating Kirk's murder shouldn't have been fired? What would happen if they weren't? Is it more likely that their friends would talk them out of saying that kind of stuff, or that they and others would feel emboldened to start calling for the deaths of others?
When people celebrated Oct 7th and Hamas and they didn't all immediately get fired, the taboo shattered in a big way. That doesn't mean it can't be restored. But it must be restored. The incentives are too strong to pull young people towards expressing this hate, until the permission structures are created for anarchy and violence at a level we won't be able to contain.
I agree he said that suppression is a last resort. That's what I was referring to when I wrote "most people bringing up his supposed paradox forget his more nuanced take on [whether it would be prudent to do so if it had such a right]". Saying that it would be unwise for the government to do so is explicitly talking about the government, as I said.
> Do you think the people celebrating Kirk's murder shouldn't have been fired?
I am with Scott Alexander that unwanted speech should not result in firing. The only exceptions would be if it's somehow related to their job.
> But it must be restored. The incentives are too strong to pull young people towards expressing this hate, until the permission structures are created for anarchy and violence at a level we won't be able to contain.
I disagree. Rules against actual violence are enough.
> I am with Scott Alexander that unwanted speech should not result in firing. The only exceptions would be if it's somehow related to their job.
Respectfully, and humbly, I think you and Scott have not thought this through enough. I'll use a simple example of an extreme edge case (but seen not infrequently in real life) that I think will show where your principle fails:
I worked at Google for a long time. If a colleague of mine put a post on social media celebrating Charlie Kirk's death, I and many of my colleagues would rightly feel incredibly unsafe in a workplace with somebody who thinks political violence against speech is acceptable, because surely I have some views they'd find "offensive" and "dangerous" too. There's a 100% chance that person would be fired, and that would be true in most other serious businesses that have codes of conduct and employees that would raise those concerns. And of course, the employer would not just have the legal right but plenty of justification to do so (and arguably a big legal risk if they didn't fire him).
The same could apply to a business. If the supermarket I shop at has a checkout kid who posted such a video, I'd feel unsafe going to that supermarket. I'd tell their employer to fire them or I would not be returning. I probably wouldn't be the only one doing so, and the manager would almost certainly fire them.
If you were the manager at a company and your employee celebrated Kirk's murder (I don't just mean a crass comment, I mean open unabashed endorsement of political violence), and your other employees or customers said they felt unsafe and made that ultimatum to you that it's him or them, what would you do? If you decided to fire them, would that be "cancel culture"?
Endorsing the chilling effect as a way to encourage democracy strikes me as a genuinely naïve approach. Well-intentioned, but naïve. You are welcoming this being weaponized against whatever opinions you hold.
So when people openly celebrated Kirk's murder or Hamas, and others reached out to their employees to get them fired, was that wrong?
I'm aware of the risks of this getting weaponized, which is why I created 4 principles (below) to restore these taboos in a way that doesn't violate the 1A or succumb to the excesses of cancel culture.
It's totally fair if you say that these can't work, but would you argue that they aren't probably the right approach for society to follow? And if we don't, what's the alternative? "cancel culture" gets weaponized on a partisan/tribal basis, and every tribe permits the most abhorrent behavior within their in-group? Sounds like a recipe for the downfall of society to me.
1. The Red Line: Limit actionable taboos to overt bigotry, dehumanization, and the endorsement of violence.
2. The Consensus Test: Distinguish between subjective offense (which is partisan and open to individual interpretation) and a Shared Moral Violation (which reflects a broad, trans-tribal consensus and warrants consequence).
3. The Private Mechanism: Enforce standards through civil society (employers, associations), never government coercion.
4. The Open Door: The goal of consequence must be correction, not permanent destruction; always offer a path to redemption.
Yes, this was wrong, and also a number of people with legal or contractual employment protections won fairly big lawsuits against their employers for the wrongful terminations. At will employees are out of luck, however.
"Yes, FIRE would and does argue that if for example somebody was fired for celebrating Charlie Kirk's murder after a mob demanded that they get fired, that would be cancel culture, "
FIRE lobbies two kinds of organizations.
1. public educational institutions, whose hiring and firing decisions are directly implicated by the first amendment, not private associative decisions.
2. private educational institutions that claim to subscribe to equivalent norms of "academic freedom."
if you're a private educational institution that doesn't make claims about "academic freedom" or similar, you go on their "none of our business" list and they never really bother criticizing you. if you're not an educational institution at all, you're totally out of scope.
I know FIRE's scope, but unfortunately Greg and co extend the same principles to apply to all manner of consequences for speech. He wrote a whole book about cancel culture, which he defined incredibly broadly, and afaik has never said anything about when "cancelling" is actually ok. He calls himself an "opinion absolutist" and explicitly says he doesn't want anybody to be fired for their legal speech. That means he wants people to tell us they're Nazis but to not get fired from their jobs for being Nazis.
FWIW, I approached Greg privately asking him to clarify this position which I saw as flawed. He refused to acknowledge it, and continues to. He could really easily squash all of this but he chooses not to.
>Speech does not burn your living room, just as it doesn't break your leg or pick your pocket
I expect that pretty much every criminal conspiracy to do any of those things involved the perpetrators talking to each other to make a plan first.
'Speech is always harmless' is a really misguided defense of free speech, because anyone with half a brain will quickly realize it is false, and most of those will not be impressed by 'well actually speech and speech acts are two different things' or other such apologia to defend the idea.
The better approach is just to admit that some speech is harmful, obviously, but we allow all kinds of harmful things to be legal because we don't want or trust a complete police state which intervenes in every moment of our lives. Free speech is important enough that we will accept great harms to maintain it, that's a good and proper stance to take.
Denying those harms makes it sound like you are not actually committed to the importance and benefits of free speech. Like, *if* it caused any harm ever at all obviously we would ban it, but it just fortunately happens that it is never harmful so luckily we can support it. That's not a stable position.
I'm always a little baffled by people who want to construct their models or morality such that the things they want to be legal/allowed are always completely morally blameless. Much better to accept that we all do harm, we all do good, we should try when possible to do less of the former and more of the latter, but there's no way to live a perfectly blameless life and anyone who pretends to is trying to sell you something.
And yet NYT reporters are openly flouting those laws (specifically, about the Situation Room reporting, which is potentially/likely to compromise the intelligence the POTUS is receiving from his subordinates).
In short, there is a difference between "there outta be a law about that" and "this is actively working to prevent the escape of state secrets."
So, I agree with this statement. I think where we differ is about what norms should be enforced with what kind of social sanction. I think norms of broad tolerance for even pretty offensive speech, ideas, and lifestyles are very valuable for letting all of us coexist in the very diverse society we live in, and for tolerating weird people who often have a lot to offer. And especially, coordinated efforts to punish speech by getting people fired have the potential to go terribly wrong, and have visibly done so in the last decade. A world where nobody will say what they believe lest they get fired for it is a world where nobody actually knows what their neighbors believe, and a world where a lot of unpopular opinions, some valuable and true, are seldom heard.
At the same time, social sanction well short of arresting or firing people is how we maintain norms of all kinds, from "don't be a Nazi" to "don't make sexual jokes/references about children" to "don't speak ill of the recently dead." A lot of online speech is not subject to the kind of informal thing where everyone at the table winces and someone mutters "Jesus, Fred, dial it back a notch."
Yes, 'conspiracy' is a great example of the law distinguishing between 'speech' and 'speech acts', thanks for giving an example of the type of thing that is not at all convincing here.
I'm talking about overt, hateful, eliminationist bigotry e.g. people explicitly endorsing violence and terrorism. If they are earnest* about that, enough to openly and unabashedly express those views, they are dangers to society. I'll still protect their legal right to express those views (short of immediate incitement) but we must punish those taboo violations with social consequences.
*FWIW, one of my 4 principles for restoring these taboos and arguably the most important one that prevents it from becoming a version of cancel culture is "The Open Door" - we can always let people apologize and clarify their views. And I have also developed a framework that does not assess somebody based only on a single statement, but based on what they say/do and then how they respond to pushback. If somebody says something incredibly inflammatory, but then is willing to apologize and not cross that line again, we should extend them grace. But if they double down on that overt hateful bigotry, they must be ostracized.
You wouldn't argue that KKK members were "cancelled" when they lost their jobs, that theirs was an imagined or exaggerated "bigotry", right?
You are conflating two things there though. Expressing "overt, hateful, eliminationist" opinions is NOT the same as "explicitly endorsing violence and terrorism".
Incitement to violence is normally where liberal societies draw the line on free speech, but expressing hatred to certain things strikes me as a perfectly reasonable thing to do if that thing is actually worthy of being hated.
I'm not talking about incitement, which can be illegal. I'm talking about things that are short of direct/immediate incitement as defined by the law, but still are encouraging violence and terrorism. Saying "Hamas is great I'm glad they killed those Jews and I wish they killed more" or "I'm glad Kirk was killed I hope they get Shapiro next" is not incitement but it is and must remain socially unacceptable. People must lose their jobs and standing in polite society, at least until they renounce that hateful bigotry.
Would you keep in your employ an avowed KKK member or Nazi, or somebody who openly celebrated Charlie Kirk's murder?
This would be an easier sell if we did not have ten years of worked examples where someone was widely denounced, cancelled, fired, and unpersoned for being a racist or nazi or some such thing, and when you dug into things a bit, the evidence was that they made a weird smirk while a bunch of weirdos were jeering at them, or that they posted lyrics to a rap song with the n word when they were teenagers, or told a tasteless joke in what they thought was a private forum, or some similar thing.
I agree that social pressure including ostracism is a natural, unavoidable consequence of speech. That is morally neutral (in Imperial Japan in the 1930s, you'd get in trouble for advocating for peace with the neighbors), but it is an inevitable part of human interaction. However, I think US society is extremely diverse, and norms that allow for a great deal of tolerance for weird or even offensive ideas, ways of life, etc. are neccessary to make it work.
I agree and understand. I call this "moral hyperinflation", and I think we need to reverse it, and certainly not accept that we simply can't socially punish the actual instances of that 'overt hateful eliminationist bigotry' because a lot of people abused the terms "bigot" and "fascist" for 10+ years. https://elevin11.substack.com/p/moral-hyperinflation
How do you feel about the cancelling of James Watson? This was expressing offensive ideas (probably in an offensive way, because Watson was famously a huge jerk), but not in a way that anyone could reasonably feel physically threatened by and with a set of factual claims that were at least plausible enough to not be obviously wrong.
I'm not super familiar with the case but I carve out "Observation" e.g. of differences across groups as socially acceptable (or at least not universally unacceptable) as long as it's not hateful and with intent or indifference to harm, which can be determined through dialogue. I would generally use Charles Murray for this example, which is easier bc he wasn't a jerk about it, and generally tried to be careful with what he was saying. If Watson was a jerk and careless with his words, indifferent at all to how people were going to interpret them in a way that might harm certain groups, then I think it'd be fair for many to criticize him for it.
If you're only talking about bigots who also endorse violence, then why bring up the concept of bigotry at all? Your red line already prohibits endorsement of violence, so all that adding "bigotry" to the list does is invite misunderstanding.
It is, in the sense that you are doing something for the other person that they would never do for you if the situation was the other way round.
If in a Green town the Purple minority is allowed the freedom of speech, but in a Purple town the Green minority is not allowed, then the Purple people have more freedom globally than the Green people. In that sense, the Green people subsidize the Purple people.
If in a Green town the Blue minority is allowed the freedom of speech, and also in a Blue town the Blue minority is allowed the freedom of speech, there is no subsidy, it is free speech for everyone.
Therefore, for a tolerant Green living in a Green town it makes sense to tolerate the Blue speech, but oppose the Purple speech. Especially when the Purple people say openly what kind of regime they prefer and are proud about it.
At the bottom is the expectation of reciprocity. I treat you nicely even if I disagree with you, because I imagine that in a similar situation you would do the same for me. If I know for sure (because you told me openly, because your entire philosophy is about that) that you would hurt me if you could, what incentive do I have to be nice towards you? Virtue signaling? Thank you, I am not signing up for suicide pacts. I am nice towards people who would be nice to me... or to people I do not have enough information about (because I prefer to err on the side of niceness).
Years ago Scott argued about whether it was justified to lie on behalf of a good cause https://slatestarcodex.com/2014/02/23/in-favor-of-niceness-community-and-civilization/ and Scott said he was still unwilling to do that, even against people who believed that tactic was justified. That is an asymmetry between Scott and his opponent, but it's not a subsidy.
>But we can and must treat overt hateful bigotry and the endorsement of violence as a violation of our universal moral TABOOS
As we've seen over the last several years, or possibly all of history, the problem here is that people *will* carve out exceptions to those taboos, and find excuse for any bigotry they themselves find palatable. This is bad for a multicultural society, and in the long run is likely to make multiculturalism unstable.
Few liberal-progressives are against *racism*; they're against one form in particular, sometimes a couple others when convenient, and the exclusions are not hard to notice. Few are against *sexism*, likewise. Few are against *violence*, likewise.
>This is how we marginalized and defeated the KKK. If we don't, that hate will normalize and turn into violence, as we are seeing.
What we're seeing is what happens when those exceptions get enforced. *That* normalizes hate.
I agree. I'm trying to say here "no more exceptions". These are universal moral taboos that we need to enforce no matter who it's directed at or coming from. We need to override the tribalism and partisanship and all hold the line, especially against our own, and condemn this hate. By excusing and tolerating it within partisan tribes, it becomes normalized. Everyone needs to stand up and say that hateful bigotry is unacceptable, and pressure others to enforce social consequences against those who espouse and sanitize it.
> If we don't, that hate will normalize and turn into violence, as we are seeing.
I feel like "the problem of our times is we didn't push hard enough against hate" is the wrong lesson to take from the last 15 years. We pushed harder than before, because before it wasn't necessary; and the more we pushed, the more the hate pushed back and came out squeezed between our fingers and said "see, see, I'm being repressed!" and gained sympathy thanks to it.
Ultimately I just don't think the root causes are that not enough people pushed. If a vast, vast majority of people share some principles then those principles are effortlessly upheld. If there's a significant minority that doesn't then the rest shouting about how bad they are or even punishing them won't fix things. I think there is a mix of underlying material conditions and fundamental loss of trust and belief in the liberal project, sadly, at the root of all this. It's not going to be fixed by screaming louder at people that they're being bad.
I think the problems of the last 15 years were we called things bigotry that weren't, that we excused hateful bigotry when it came from our own tribe, and we taught a generation that it was ok to be bigoted towards anybody deemed an "oppressor". We have to undo that and collectively hold the line with universal principles against overt, hateful, eliminationist bigotry, the thing that ~99% of us agree is socially unacceptable.
> I loathe cancel culture, but we have to make certain speech and behavior socially unacceptable
Making certain speech and behaviour socially unacceptable is literally what cancel culture is. We should be careful that any rules we decide on are applied fairly to both sides - yours, and your outgroup's; we don't want to end up creating yet another one of those irregular verbs so prevalent in political rhetoric, with a conjugation table something like - I align with universal, liberal values; you make certain speech unacceptable; he/she/it engages with cancel culture.
If your definition of "cancel culture" includes firing a KKK member or somebody who joyously celebrated Charlie Kirk's murder and wished for others to be next, then your definition is too broad.
I hear you on your concerns, and it's why I've laid out 4 Principles for Restoring our Universal Moral Taboos (below), which helps prevent this from becoming what I think is the better definition of cancel culture which was more often than not social consequences imposed based on offenses taken by a small group, and completely unforgiving.
I'm not naive enough to think that we can follow these perfectly, but I do think this is how society functioned until not that long ago, and it's something we must get back to or we're going to see a very quick backsliding into uncontrollable anarchy and violence.
1. The Red Line: Limit actionable taboos to overt bigotry, dehumanization, and the endorsement of violence.
2. The Consensus Test: Distinguish between subjective offense (which is partisan and open to individual interpretation) and a Shared Moral Violation (which reflects a broad, trans-tribal consensus and warrants consequence).
3. The Private Mechanism: Enforce standards through civil society (employers, associations), never government coercion.
4. The Open Door: The goal of consequence must be correction, not permanent destruction; always offer a path to redemption.
Can you expand on which of your four rules an employer's decision to fire "somebody who joyously celebrated Charlie Kirk's murder and wished for others to be next" fails to meet? It's not at all obvious from here.
What do you mean fails to meet? The people who unambiguously celebrated Kirk's murder crossed the red line (1) that nearly all of us would agree (2) is unacceptable. If they were willing to apologize and recant that endorsement of violence, their employer and others could decide to offer them grace (4) and not fire them.
What was actually quite inspiring to me was that the people who didn't celebrate his murder but were merely "remorseless", even though there was pressure to fire them, many employers did not, because 1 and 2 were not met.
Hm, I took "your definition is too broad" to mean you thought the people in your example shouldn't have been subjected to social consequences, but now it sounds like you actually do want social consequences for them, you just don't want your particular set of social consequences to have the "cancel" label? Am I misunderstanding again?
That's right. I think "cancelling" is too broad of a term. Some people/behaviors MUST get "cancelled", or they will normalize and destroy America and the West. My principles are meant to re-impose those social consequences where they should/must be, without this devolving into the excesses of "cancel culture."
Do you have a principled disagreement with this view?
If your definition of "cancel culture" includes firing a KKK member or somebody who joyously celebrated Charlie Kirk's murder and wished for others to be next, then your definition is too broad.
I hear you on your concerns, and it's why I've laid out 4 Principles for Restoring our Universal Moral Taboos (below), which helps prevent this from becoming what I think is the better definition of cancel culture which was more often than not social consequences imposed based on offenses taken by a small group, and completely unforgiving.
I'm not naive enough to think that we can follow these perfectly, but I do think this is how society functioned until not that long ago, and it's something we must get back to or we're going to see a very quick backsliding into uncontrollable anarchy and violence.
1. The Red Line: Limit actionable taboos to overt bigotry, dehumanization, and the endorsement of violence.
2. The Consensus Test: Distinguish between subjective offense (which is partisan and open to individual interpretation) and a Shared Moral Violation (which reflects a broad, trans-tribal consensus and warrants consequence).
3. The Private Mechanism: Enforce standards through civil society (employers, associations), never government coercion.
4. The Open Door: The goal of consequence must be correction, not permanent destruction; always offer a path to redemption.
> I'm not naive enough to think that we can follow these perfectly, but I do think this is how society functioned until not that long ago,
I assure you, no society ever functioned the way you imagine. Certainly US never had a "trans-tribal consensus" of what is or is not taboo. I am 60 years old; I remember when it was entirely normal for an employee break room to have cutouts of porn magazines on a wall, and if any woman did not like it, too bad.
Now the pendulum swung in the opposite direction, but it does not mean it ever resided in some happy middle.
You're proving my point. There was no consensus on that example. Most women and some men found it offensive. It wasn't universal.
Compare this to the KKK. Pretty much everyone agreed their overt hateful bigotry was not welcome in polite society. That doesn't mean nobody could make or enjoy racist jokes.
Steelman my argument. Find the most extreme case where you think I'm right. I think the KKK (especially the Third Klan) is the right one.
I just published my essay that tries to re-frame this concept of "Stochastic Terrorism", acknowledging that the term has been abused, and that even its most horrid examples should never result in state-imposed consequences, but that we must ostracize real hateful bigotry out of polite society: https://elevin11.substack.com/p/dismantling-stochastic-terrorism
I was expecting this post to be about the extent to which stochastic terrorism is a demonstrable or quantifiable effect, so in lieu of that I spent a little bit of time researching it myself and it appears that there is strong evidence of correlation but little evidence of any actual causative effect. (Pretty much what I expected, but still interesting.)
The censorship question is another matter entirely, to me - even if there were strong, conclusive evidence that certain types of speech led to violent act by other people, I would be against censoring that speech.
Going to court is more expensive in America. And you won't recover attorney costs if you win, so settling is more attractive even if you're only offered a discounted amount
More to the point, all 50 states and the Federal government have laws against incitement (to violence or other criminality) and US courts already have existing doctrine for what type of speech can actually be prohibited as incitement
Yeah, I agree with Scott's political argument that stochastic terrorism ought not to become a justification for arresting people for their political speech, but still think it is an interesting empirical question how much various kinds of hostile rhetoric lead to actual violence.
I think where I land is that "will no one rid me of this meddlesome priest" is only remotely a thing you can hold the king liable for. If you're just a guy, you aren't under the obligation to understand your positions and opinions as potentially being interpreted as commands - actual material power on an individual level gives you that responsibility, because it's something more reasonably foreseeable. That's why I think we can say "important politicians have a responsibility not to incite things" - it's at least not crazy for them to think the things they say can lead to real actions - without having to wonder what it is that Jodie Foster did to stochastically induce certain events in 1981
Does Father Coughlin fall under "important politicians"? How about Stephen Colbert? Perhaps when you can hold a rally in Washington DC and bribe people to be there with funnel cakes, you're enough like an important politician to Watch Your Language?
(But don't we already have laws for libel and slander? How is this not covered by that? Can you really incite violence without distorting the truth?)
Well, fair, I don't think generally accurate guidelines are necessarily going to fall along definitional lines that can be demarcated clearly. If anything, it's the sort of "reasonable person" test that's really difficult to determine legally, and is just sorta the way I think it makes sense to consider it for the "social stigma" angle. Testing intent and "he should've known" type things are more of a thought experiment, since we rarely *actually know*.
More than the "what's the brightline for sufficient amount of individual power" question (which I think varies - for instance, the leader of a small but fervent cult probably has less actual power than a mid-level politician, but probably contends with *more* likelihood that errant words turn into disproportionate action), I find the problem of collective power a bigger thorn. The most natural response to my position, IMO, is "a society with the power to set dangerous norms is made up of a lot of different people without much power in their own right - components of a powerful and culpable force for incitement of violence aren't innocent just because they couldn't do it by themselves". You see this a lot in modern politics, where the idea is "each person in X group made society more amenable to the conditions that allowed bad thing Y to happen, and this was a natural and foreseeable consequence of the world they worked for".
Personally, I think it's just practically unworkable to think this way - no one can say anything, ever, because everyone with that position might have the joint power to incite bad things to happen. But it's a bit inelegant, I suppose.
Dangerous norms include "It's okay to riot when a black person dies at the hand of a cop."
How culpable is the media? How culpable is any particular piece of art, or news-reporting? Nobody reports on the starving people who stopped rioting because they'd literally destroyed all means of food transport...
"The Media" is very powerful, in that they've managed to influence a lot of people -- it takes a lot of people to say "Yeah, Riot!" and actually make it happen.
Or we could point to Apple's practice, during BLM, of letting black people have free iphones (instead of, as was done in 2025/2026, simply bricking them).
South Africa. The riots where everyone had glowsticks on, and there were civilians manning tanks because the police had fled (or do you remember the bookstore that was notably not touched during those riots? That South Africa had to do mondo damage control on...)
At that point, I have to ask if the violence was already likely to happen, though? If it's truth, then it's truth. And you're probably not the only person to uncover it.
People who can't discuss the truth are far more likely to turn to violent means of getting their manifestos heard.
I think the strongest case for a social stigma around certain truths is more about presentation and completeness - although, again, the "stochastic terrorism" thing should only apply in very limited circumstances regardless
Being fundamentally correct should be a complete defense against the charge of being unpleasant or inciting things for the exact reasons you mentioned. But important truths are often inseparable from a great deal of surrounding context. Isolating a given piece of the truth can send someone in the opposite direction that context does - a half-truth can be inseparable from a lie, with the biggest difference being that half-truths are infinitely harder to find out as misleading. There are similar factors around being inflammatory *about true things* - sprinkle a few obvious opinions around the hard facts and the emotional valence of the total package changes, which is the whole reason politics is a profession.
I don't think the average person should worry about this, outside of the general idea of "don't lie, even if you're crafty about it" - there's basically no chance that their half-truth is taken seriously and leads to "stochastic terrorism", it's not an outcome that's realistically forseeable without also forseeing issues with saying anything at all (because anything true or false could be misinterpreted by someone in the wrong frame of mind). The other issue is that telling the *literally whole* truth in any given situation is impossible, unless you're Prak OD'd on truth serum. But I do think there are ways in which parties who can reasonably forsee their words being taken seriously (in fact, people *trying their best* to make that happen, in the earlier case of politicians) can end up responsible for certain ripple effects resulting from even improper framing of the truth.
Libel and Slander doesn't really cover attacks on groups. "The Haitians are eating the cats and dogs" is a distortion of the truth, but not in a way that can be held accountable via slander charges
>to understand your positions and opinions as potentially being interpreted as commands
That's not what 'stochastic' means.
I feel like we should be able to understand this, as people who talk about expected utility calculations all the time... there is always a low probability that any person or group will be targeted as politically-motivated victims of crime, and every public comment engendering more hatred of that group increases that probability.
If your statement increases the probability from 1.2x10^-10 to 1,3x10^-10, that's a very tiny amount of harm in expectation and it's probably fine, you probably cause a lot more harm than that whenever you buy a t-shirt without checking how it was manufactured. But it's still a real phenomenon that you're responsible for.
Worth noting England in 1170 was not a liberal society. I think it'd be perfectly fine for whatever sad sack is Prime Minister this week to say something like that about... like, a judge. (because it'd have to be someone in a position of official power for the analogy to hold) If a nearby RaSP officer went and murdered the judge, everyone would agree that officer was insane.
Turbulent. Thomas à Beckett was turbulent, not meddlesome. And interestingly, Henry II didn't even intend "turbulent" to mean something like "meddlesome", he was simply referring to the fact that Thomas à Beckett had a big aerodynamic spoiler coming out of his back like a Toyota Supra.
Re. "Only the King" - I would imagine that Scott probably has a wider reach, a bigger audience, and probably more economic power, than a Dark Age monarch. (Though of course, as of the time of writing, Scott has yet to amass any serious military power or become head of any major religions.)
A key distinction is between general expressions of discontent and approving of political violence but not having the guts to do it yourself.
Consider if you desire people to kill redheads but settle for publicly calling for mandatory hats. Someone kills some redheads because they should be hatted. Legally you are not responsible and should not be. Morally you might need to figure out whether this extra-legal violence is good and own it one way or another.
If instead you are actually calling for violence then you have a legal and moral issue. Legally, this is actual incitement to violence. Morally, you are weak - you want the violence to happen but want someone else to pay the cost.
On the other hand, if you just want them to cover up and have no desire for harm, then you do not and should not bear any accountability if someone escalates to violence
It's a simple question of is the cost via the probability of increasing the chance of assassination lower than the benefit of voicing concerns? It's a real cost, but the benefit is likely bigger in the general case.
I think this might be one of those cases like "should I tell a noble lie for the public good?" where I oppose even doing the calculation, on the grounds that requiring the calculation makes it too easy to get the calculation wrong, or to accuse others of getting the calculation wrong.
Sure but that's just the practicalities of how close of a process do you have to producing an answer for it. If you can have an exact answer, surely you'd use it. The calculation is noisy and expensive, so the hope is that the more strict approach is what works better given limited resources and information. But we should still remember that this approach is just an approximation.
A probability calculation can produce an exact (in the sense of perfectly accurate) result, just like any other calculation. I think you may be confusing different levels of knowledge.
I see the issue with the wording, but this is the underlying question. Answering it via a rule that works well the majority of the time, because that's the most practical way to answer it is well and good, but I disagree with rejecting that this is, again, the underlying question.
Tenoke didn't say the computation is simple, he said the decision rule is simple. That's how you reason about vaccine mandates, nuclear deterrence, and every other high-stakes low-certainty domain; why should this one get an exemption?
I agree this is in a cosmic sense the right decision rule, but I think "right decision rule" refers to a different, less obvious level.
This is like answering the question "what did you get for Christmas?" with "atoms". It is definitely true, but it's so true of everything that saying it in response to any particular question is likely to mislead people into thinking you're giving the expected normal-level answer that should apply in real life (they might think you got some kind of unusual scientific chemistry samples or something).
Sure, but the problem with answering "atoms," as you say, is that it's indiscriminatory; it's compatible with every answer. But cost-benefit would actively rule things out or make falsifiable claims about which cases are worse than others.
I doubt that if I say 'I am pro criticizing people publically because I estimate the benefits of it outweigh what I estimate is only a very small chance of increasing their chance of assasination' you'd be quite as confused about my position as if I said I got 'atoms' for Christmas.
I can see why that might not be the way to word it to get through some people, which is a more valid concern I do sympathize with.
That feels like a dangerous position to take if you want society to be Consequentialist in general.
Like, sure, there's a chance that people will do the calculation wrong, or use the fact of calculation calling attention to the expected harm against people in bad ways.
But there's also a chance that saying 'we don't need to do calculations here, we have deemed this category of actions Morally Acceptable and no further analysis is warranted' just inculcates an adherence to unthinking deontology and virtue ethics in the populace. Which I think is *way* more subject to misuse and abuse.
what is the thing that we are saying is "worth doing" here? Scott seems to be saying that analyzing the potential consequences of stating true criticism is simply not worth doing (and therefore not worth doing with made up statistics).
That seems alot like the "should I respond to deny untrue allegations" decision point, where once you've established that you'll do it even once you now have to justify why you aren't doing it in other situations.
And I find this particularly annoying wrt AI policy, as I don't think I need to append "you shouldn't actually go kill Sam Altman" to every single statement AND if you DO append it you get called out by idiots saying "oh so you think he's a reckless tyrant trying to become the immortal god-emperor of the universe but you won't call for killing him, I guess you don't really believe it" like some kind of stochastic terrorism offsides trap.
Far from being "simple", you could not make the calculation at all, not even with great effort. Most, if not all, of the numbers you put in would be guesses.
If you were making predictions, sure, but we are usually talking about this in the wake of someone being killed.
Just take the number people in group X who have been lynched and divide by the number of posts and articles saying group X are evil monsters. That's roughly the number of expected deaths per comment.
It'll still be imprecise, sure, but given that this is already going to be something like billions or trillions of comments per death in most cases, missing by a few orders of magnitude barely matters on an individual level.
... yes? The shooters often talk about their motives, and those motives match things they read and hear in the culture?
Like, at a certain level I don't understand the concept of a human in a society taking a social action, and that action not being causally linked to the society they were raised in and the specific circumstances of their life within that society. That feels like it's probably not a thing that happens much.
Of course, if we kept talking about this it would probably end up hinging on our definitions of 'prove', 'causative', 'relationship,' and 'two things'. Semantic arguments like that are boring, so let me just put it this way: yes I think that groups that are targeted for hate by large segments of society are generally in more danger than groups that aren't, and there's a causal relationship between those things.
I don't think you're giving the powerful/powerless distinction its fair share. The obvious answer to your point would be that powerful people are not only much harder to assassinate (Trump has constant protection, the average mosque doesn't), they also tend to live with well-mannered people, while vulnerable people live with all sorts of crazies and junkies who are much more likely to attack them because they were influenced by something they heard online. If we're assessing risks, negative speech against one group is much more likely to result in dead people compared to the other.
I think this really only works with the President / everyone else distinction; I think even the average CEO isn't too hard to assassinate. I think the powerful/powerless distinction more often gets trotted out to mean things like "you can criticize white people, but not black people", where the two groups don't differ very much in security.
Trump was inches away from actually dying. He's alive because the shooter had bad aim more than because of the Secret Service. If Tyler Robinson had switched places with whoever that was, perhaps the outcomes would be reversed.
I think the Secret Service being competent has a big effect on how far away the shooter is, what gear they can get on sight, how long they have to aim, etc.
The Secret Service failed to act on over 100 transmissions about Trump's shooter. This doesn't seem like competency.
The schoolteacher did actually think there weren't metal detectors, because he didn't see any metal detectors he recognized. The Secret Service disabused him of that notion.
Eh. How many transmission do they get about potential threats every week, how much budget do they have to deal with them, how specific were those transmissions such that anything more than 'continue your already-optimized normal security procedures'' was clearly warranted.
I'm not saying i think SS is definitely uber-competent, I'm mostly saying I don't think we could possibly know enough to judge, and that blaming the shooter for being a bad shot ignores how security made the shot harder.
Brian Thompson was admittedly shot by a handgun at close range, but Charlie Kirk was shot by a rifle at longer range. I don't think either rifle shooter was being pressured by security while aiming, rather they only revealed themselves by shooting.
Thomas Crooks had been monitored before the shooting and found a position which was obscured from the ground and Secret Service. Part of the reason for his inaccuracy is that one of the police officers clambered up the side of the building to see who it was, and he turned his gun on them, then took his shot immediately (before getting killed by the Secret Service). Although the SS/police failed to stop him from firing, their actions did make him less accurate; if they had all been sleeping before he'd taken his shot, he probably would have had an easier time finding a good spot and fired at his leisure.
Not everywhere is as policed as the US, and not every politician is Trump. The U.K. has a hard job protecting ex prime ministers, it’s an ever increasing drain as we go through them like candy at the state fair*.
Anyway just this week an ex member of parliament was murdered, the police are treating it as lone terrorism because they treat all attacks on politicians as terrorism. The woman in question was a fairly conservative ex Tory (later reform) named Anne Widdecombe. No doubt the killer opposed her politics, and was perhaps influenced by strong online rhetoric.
Nevertheless I would agree strongly with Scott against banning critical speech, no matter how linguistically violent, unless there’s a direct call to actual violence.
The "go through them like candy" line somehow drew me to the conclusion that a solution for all relevant problems is just to feed ex-PMs into a wood chipper. No risk of their assassination, no drain on the budget.
Would also encourage them to cling to the job a bit longer, for better or worse.
Though it would also have an impact on who'd volunteer for the job.
Well, in practice at least. I think in theory the king just appoints whoever he feels like to the job. It just usually happens to be the leader of the majority faction in parliament.
I wouldn't agree to "no doubt" at this point, we really don't know. Jared Loughner was on an entirely different world than every politician when he shot Gabby Giffords rather than actually having anything to do with her politics (he was upset that she couldn't answer a nonsensical question from him).
> He killed an insurance executive because he had strong opinions about what insurance executives do wrong, and these opinions didn’t hinge on slight changes in the emotional resonance of the words being used.
I feel like a rhetorical trick is being used here. Sure they didn't hinge on "slight" changes, but the strength of his opinions definitely related to the emotional resonance and tone of what he read in the past.
My claim is that it didn't hinge on some specific beyond-the-pale language like calling him vermin; it hinged on normal-level criticism like that the executive was "greedy" or something, where we don't want to impose too high a burden on discussing greed (in contrast to calling someone vermin, where I think it's fine if we impose a high burden).
But I'm not even sure it hinged on the word "greedy". I think there's a "rational" case (as in, one you could make without using any emotional terms whatsoever) that insurance companies are killing many people, and it's not obvious to me that Mangione was influenced more by the use of various emotional words than by hearing that case and finding it plausible.
Hard to prove either way, but his manifesto was pretty emotionally charged: he called them parasites who abuse America for profit. Maybe he was primarily convinced by the rational non-emotional case and then came to those emotions after, but the murder is way less likely to happen in the counterfactual where criticism of insurance companies is less emotional.
Also, in your comment you use the phrase that the companies are killing many people; this is very emotionally charged framing and word choice! People that use the term "stochastic terrorism" say that using "kill" when you mean "indirectly lead to the death of" is very objectionable, see musk.
I agree that he didn't kill Thompson because he heard the word "cockroach" too many times or something, but I do think that hearing how CEOs like Thompson are *killing* thousands made it a lot more likely. Killer is a very dehumanizing label.
Hm, I think "killing" is pretty close to emotionally neutral, in the sense that "indirectly led to the death of" is making a conscious decision to be unusually dry and bland. I don't think it's practical to have important conversations at the level where you can't say bad health policy kills people.
Example: do the following count as "killing" or "indirectly leading to the death of"?
1. Hitler (who never actually murdered anyone with his own two hands)
2. A thief who steals someone's life-saving medication to sell it on the street, knowing that they'll die without it
3. A politician who uses lies and fear-mongering to cancel a life-saving program
4. A surgeon who is extraordinarily bad at their job and accidentally removes someone's heart when they were supposed to remove the kidney.
5. A doctor who hates someone and falsely tells them they're fine, removing their chance to cure a curable disease in time, in the hopes that they'll die?
I think this boundary is too fuzzy to worry about it too much.
My point is that use of the word "kill" provokes a stronger emotional reaction than something like "causes deaths". This can be useful to motivate people to care about important stuff more like climate change or health policy, but saying that CEOs are killing people will make violence against ceos more likely. A more dry term instead would make it less likely, and an even more emotionally charged term like "murdering" would make it more likely still. This is agnostic of what we define as neutral.
Perhaps its worth the increased of risk of violence in order to get the stronger emotional reaction, but the tradeoff is there.
I think you have a way better argument than Mr Alexander on this. I think use of 'kill' as a synonym for 'decisions that indirectly increase the chances of deaths among this group' is one of the few, maybe even the best, examples of actual "stochastic terrorism". Even his examples conflate this will instances where an identifiable person can be specifically attributed to another person's death, which is exactly what the use of 'kill' is intended to invoke.
Except in the case of health insurance companies, it's not 'decisions that will indirectly increase the chances of death among this group.' That is circuitous enough to route around not just the emotional punch, but the actual reality as well.
I mean, there *are* cases like that. When a health insurance company decides not to pay for a screening that leads to earlier detection and better outcomes for some possibly-fatal illness, then the above description is apt. But not all decision are like that. There are also decisions that are more like "find underhanded ways to drop people from your policy to avoid having to pay for an expensive treatment that they will definitely die without." That's not 'indirectly increasing the chances,' that's deliberately decided to let some number of (nameless, faceless) strangers die for money.
You can quibble that it's still indirect enough that calling it 'killing' is dishonest and manipulative, but that *is* a quibble, not a serious argument.
> 1. Hitler (who never actually murdered anyone with his own two hands)
Sorry for the nitpicking: I wonder if this is actually true. I suspect his personal bodycount is really low, but I wonder if he ever eg shot anyone in the Great War when he served on the front?
This doesn't impact your point. Just a historical curiosity.
> So if the bar for "killing" is set very high, I believe this makes him a pretty good person.
Well, only on the 'number of people personally killed' score. He also did some other things like denied smokers their pleasures without scientific evidence.
Re "vermin" and similar derogatory language: I asked Google/Gemini:
>What is the strongest language Musk and Altman have used in disparaging each other?
and they came up with:
Musk and Altman's strongest language includes Musk calling Altman a "scam" and a "liar," while Altman has claimed Musk's life stems from "insecurity."
Specific exchanges and insults include:
The "Scam" Accusations: Musk labeled Altman "Scam Altman", stating he takes scamming "to a whole new level" and "might literally love scamming more than any human alive!"
The "Space Datacenters" Retort: Altman fired back on X: "homeboy you're the one sellling public market investors on short-term space datacenters."
Musk replied, claiming Altman could end up in prison.
The "Obsession" Jibe: Altman has claimed Musk "can't be a happy person" and that his "whole life is from a position of insecurity," famously dismissing Musk's attacks as proof that he is "obsessed with me again."
Now, _personally,_ I prefer that they stop jabbing at each other and get on with building AGI, but I think these cases merit something more like raised eyebrows rather than high burdens.
Many Thanks! LOL! Yeah, from what I've read of his style: The good news is that he (often?) understands the technical issues. The bad news is that it sounds like he micromanages a _lot._
The particicular case that concerns me is an instance where someone says "X person/group is so bad, someone should /do/ something about them." On January 6th, Trump's "will be wild" tweet maintained plausible deniability while clearly signalling violent action against the legislature. Similarly, repeatedly saying "X group of people are sick, disgusting, horrible, child killers, satanic demons, etc." and then shading into "someone should DO something about this" is clearly a dangerous thing to say given all the deranged people that exist and might see this as a call to action. When the commentator understands this dynamic and invokes this language anyway, I posit that it can fairly be called stochastic terrorism.
The thing that immediately popped into mind for me reading this was Radio Télévision Libre des Mille Collines (RTLM) leading up to the Rwandan Genocide which had more or less this exact dynamic (Tutsi's were more or less exclusively referred to as cockroaches or snakes).
Agreed. Also I think stochastic terrorism is an unhelpful term for this particular misdeed, since it implies that speech is violence, and any good liberal should want speech and violence treated differently. “Incitement” is a good term here.
I think the line I draw between incitement and stochastic terrorism is that stochastic terrorism trades off a few points of certainty (ie, it's not guaranteed, and in fact pretty low-probability that anyone will act on any particular statement you make) for a few points of plausible deniability (if/when something does happen, you're safely outside the bounds of conspiring to make it happen, you've behaved recklessly but without a focused intent on a particular action.) What makes this particular act very toxic is that the person authoring the stochastic terror can then calcify/cement the moral position they have and the /rightness/ of their action by appealing to their earlier invective in combination with the high-minded principles articulated by Scott above
So.. the European countries that have laws against severe hate speech are reasonable to do so because it encourages violence if left alone too long.
Are you saying then, that the problem is one with labeling too many things as stochastic terrorism? It sounds like you understand that the concept of hate speech being a causal agent of violence is a real phenomenon.
On some level, my claim is that the cost of the concept (banning criticizing anyone, or unprincipledly carving out a group of people who it's banned to criticize) is greater than the benefits (preventing some very small number of deaths that might have happened anyway) - although I prefer not to use this level because I don't think we should compare costs and benefits this greedily or naively. I don't claim that it is impossible for critical speech to cause violence.
I understand a little better now. But i don’t know if the counterfactual world where this phrase didn’t exist would have much less of these undesirable behaviors. It seems like a rhetorical cudgel of the type you are describing is likely to have emerged in another form.
The other thing was that I’m not sure the cost to society is mostly the small amount of deaths, but as you say it seems difficult to measure other types of violence easily.
I don't think there's any reasonable standard by which pre-battle speeches cause violence. Their purpose is to make the soldiers that hear them more successful at the violence they were already going to do. But by the time a pre-battle speech is given, the chain of events leading to the battle has already mostly unfolded. Maybe a really, especially *bad* pre-battle speech could ruin morale enough to make the soldiers run away without fighting, but that would take some doing.
What things actually cause the violence varies on a case-by-case basis. Usually it's some combination of promises, (maybe implicit) threats and social expectations. I guess you could say most of those are delivered by means of speech, but they're almost all made credible by past non-speech actions, like handing soldiers money, or executing deserters.
The perceived cost of allowing hate speech is not just, or even mostly, that it leads to killing. It's thought to increase tensions between groups in general, and thus hurt the ability to talk and work together.
Ostracizing people for saying things they believe are true and important, also hurts the ability of people to talk and work together. Really, it directly attacks the concept of people talking and working together, because the proponents of cancellation don't want anyone to work with or listen to their enemies.
Supposing I grant that that's true, it comes to a decision of which group is more valuable to work with: the sort of people who engage in hate speech or the sort of people who refuse to work with the sort of people who engage in hate speech?
(History has spoken clearly and repeatedly on this point, though perhaps you haven't paid attention.)
The more broadly you define "hate speech", the more likely it is that the group that includes the hate-speakers is going to make for more productive partners than the people who piously refuse to work with them.
And in the history of my field, I note that NASA managed to put men on the Moon, starting from almost nothing, in eight years - at a time when casual racism on and off the job was still unremarkably in places like Huntsville and Houston. Today's NASA, fully on board with all the political correctness, started with all the records and leftover infrastructure of Apollo and more, and has taken fifteen years to send a group of astronauts to wave out the window as they fly by the moon, with no clear plan for actually landing on it other than to hope that Elon Musk can do that part for them.
Now, I'm pretty sure the casual racism wasn't necessary or even helpful. What was necessary, and vital, was a team of people who could work together on the one thing they *did* agree on, even when they disagreed quite strongly on other matters. That's true in any great endeavour, and most mediocre ones.
If you have bad things to say about your coworkers, or about broad classes of people that include your coworkers, do it on your own time and keep it out of the workplace. If your coworkers have bad things to say about you, but they keep it out of the workplace, then return the favor. Do both of those, and you may be part of something great. Or at least mediocre but lucrative.
You can argue that the cost of criminalizing the act is greater than the benefits without going so far as to denounce the entire idea of the concept as something worth naming, talking about, or criticizing people for doing.
Maybe this is too hyper-specific an example, but I took the lesson from it to heart.
(Most of this is from “The Politics of School Desegregation” by Robert L. Crain.)
The desegregation of Southern school districts in the 1960s, across many individual cases, went one of two ways. Either desegregation went into effect peacefully, leading to integrated public schools, or it met resistance, often violent, and ended up with the public schools being 100% Black, and private “segregation academies” being 100% white.
What accounted for the different outcomes in different school districts? The obvious demographic variables didn’t explain it.
What mattered was the attitude of local elites, including local political, religious, and business leaders.
Where those leaders were saying “we don’t like this, but it’s the law now,” and urged calm and compliance, things were peaceful. Where the leaders were denouncing federal interference, making racist statements, etc., violence and failure of desegregation followed.
Of course not every local leader was on the same page. I imagine a lot them held silent when their colleagues preached defiance.
Based on this and other historical examples, I took it as the moral duty of a political figure to promote tolerance for minority groups.
Being a small-time political figure myself (I have held local elected office for cumulatively almost three decades), I strive to live by this.
That said, each constituency I have represented has been predominantly liberal, and things like actively defending the gay community during the AIDS epidemic, or the Muslim community after 9/11, has never cost me anything politically.
Did the studies control for reverse causality? It seems plausible to me that more violent and/or racist districts would naturally have more violent and/or racist politicians representing them.
The problem is that the activities of the "gay community" really did contribute to the severity of the AIDS epidemic. Perhaps if you had been less supportive fewer of them would have died.
Really, the problem genetically speaking is that AIDS and the gay community "accidentally on purpose"* genocided hemophiliacs.
*On purpose in that they wanted money, and were paid to donate blood... a lot. Accidentally, in that they didn't realize they were infected, or that they'd be hurting anyone.
It seems like the main difference is the *direction* of the reaction (urging cooperation versus urging resistance), not whether the language used to urge resistance was rational versus emotional.
This is an interesting counterpoint to "Weak Men are Superweapons" (from the SSC era).
Not a contradiction, because this is specifically about the assertion that superweapon-ish claims can be blamed for violence. But the superweapon discussion had overtones of that (see: the "superweapon" metaphor and the examples given for it), so it deserves explication. "Stochastic terrorism" would be an effective and intuitive accusation to make against would-be superweapon-builders; what does the "liberal solution" offer in its place?
I don't think the problem with those "superweapons" is that someone could die from them, the problem is that they're false and unfair. I am certainly not claiming in this post that there's nothing wrong with saying false and unfair things.
I think the point of "stochastic terrorism" is that certain kinds of false and/or unfair statements are especially likely to incite violence, and therefore especially bad.
I think that "X is evil" is absolutely legitimate, as is rephrasing it in even stronger and more inflammatory language, if it's true, but that generally people use stronger-than-ideal language to condemn those who make them angry.
I think that "We should use force against evil people" is generally not OK, and should be condemned, but in itself it's probably within the bounds of speech a free society should tolerate.
I think that /putting those two ideas together/ should be treated as much worse than the sum of the parts.
Saying "X is a traitor, and we should kill traitors" is a bigger deal than saying "X is a traitor" and then, separately, in a different context where no-one can work out which traitor you're talking about "We should kill traitors".
And you can go through way more than two steps. Revolutions against unjust governments are justified and virtuous (right there in the declaration of independence iirc). Revolutions must accept the need to, if resisted, kill those using a monopoly on power to enforce the status quo. Unjust governments limit the rights of people to access and influence political systems. X policy you dislike (whether letting non-citizen residents vote, or NOT letting non-citizen residents vote) denies people their right to self determination. Police are agents of the government's execution of its monopoly on violence over potentially unjust laws. Ergo if you dislike some voting registration laws you miiiight wanna start killing cops in the name of overthrowing the government.
I'm one of those people who, while liking the idea of liberalism a lot, is pretty fuzzy on how I feel about trying to kill extra-legally murder Very Bad People. I have a hard time condemning it as a complete blanket statement, but I also don't particularly trust myself to know where the justifiable exceptions would be (if there are any). Of course I don't exactly trust others to do so either... but it also feels like a disagreement of judgemental calls too much for me to outright condemn someone for (trying to) kill someone causing (or having caused) pain or death to potentially millions.
But the problem is you could define anyone voting for any candidate with a policy I think will cause widespread harm as "trying to hurt-or-kill millions."
>I think that /putting those two ideas together/ should be treated as much worse than the sum of the parts.
But the whole point of it being stochastic is that no one person has to put them together explicitly for the killer to to hear them both and integrate them.
Like, if there are two posts on a forum one right after another, the first one saying 'X is evil', the next one saying 'We should kill evil people'. Does it really matter whether those two posts are by the same anonymous author, or two different ones? The person reading it receives the same communication either way.
Because of the weird way humans are irrational, I think it does make a difference.
If I see someone saying "we should kill evil people", my assumption is that they're probably just blowing smoke. If I see someone say "We should kill evil X", then ... to be honest, my assumption is still that they're probably just blowing smoke, but I'm less confident of that.
And I think that seeing someone else who I believe actually supports murder makes it more likely that I'll see murder as a normalised thing and commit it than seeing someone else just blowing smoke about it.
Thanks for the notification, I hadn't known that Trump was speaking tonight. Now, having watched his speech - umm, cybersecurity has long been a problem. If the PRC managed to hack vote totals, that is unfortunate, but wouldn't be all _that_ surprising... Now, I'd have been upset if Trump had said that, because of vulnerabilities, the midterms will be delayed or cancelled - but he _DIDN'T_ do that.
What is the US aversion to pen and paper ballots? (And don't say count duration as that is easily scalable as every other democratic country regularly demonstrates).
No, by Trump standards it was just another airing of the grievances. He was teasing something *big*. He might have at least nominated Meat Loaf to be Secretary of Power Ballads or something.
Many Thanks! Yeah, a lot of it sounded like relitigating 2020 ( _WHY_ ??? He _is_ POTUS now - and he never got a single one of his 2020 cases accepted, even by GOP judges). On the other hand, if Mythos found that our electronic vote tabulation has even more holes than we expected, it is worth plugging them...
It seems like the concept has been watered down by overuse. My understanding was it originally hinged on a kind of legal loopholism like "somebody should kill these people" and relied on walking the line on incitement of violence, and was applied mostly to people with broadcast style platforms. It seems like it's useful to have a term for the concept of "creating social change by indirectly encouraging others to do violence", so that we can all disapprove of it. But then once that concept exists it becomes part of the narrative battlefield itself
Stochastic Terrorism seems to be another speech-stifling concept.
The more interesting question for me is how difficult it is to justify violence starting from the beliefs of an ideology. For instance, combing rationalist utilitarian decision-making with the ability to arbitrarily assign probabilities to events(p(doom),p(doom|attacking Altman)) can easily justify violence towards Sam Altman.
Act utilitarianism has been out of style since 1850. Rule utilitarianism (in some form or variant or level of thinking) is where it's at, and I think the vast majority of rationalists are convinced of this.
I guess this is saying that the damage to a social norm against political violence is greater than the benefits of decreasing p(doom). A rationalist with sufficient p(doom): p(doom|political violence) ratio would disagree with this, especially with how weak our norm against political violence already is.
Personally, I would emphasize that probability models are not equally effective as a way to resolve this.
I think act utilitarianism is better than rule utilitarianism.
And I think the problem come from mixing ethics with decision theory : ethics describe our moral preferences, and decision theory describe how to act given our preferences and beliefs.
A common argument against act utilitarianism, or even utilitarianism as a whole, is that it would lead to bad consequences to follow it.
And rule utilitarianism try to fix it by redefining what we want, instead of how we should act given what we want.
It is like, being faced with Newcomb’s paradox, answering with "hey, actually this proves what we wanted wasn't more money, but following the rule that would lead to more money !"
This is missing the point and fixing problem at the wrong place.
In utilitarianism, and in Newcomb's paradox, you should take into account the very act of following a rule instead of another, which would probably lead to something like the FDT when generalized.
You don’t need a revision of the ethical framework nor of the goal of maximizing money in the paradox.
Like a lot of internet discourse, I think most examples of people claiming stochastic terrorism is inaccurate at best and incoherent at worst. That said - I think there’s a core concept that’s worth preserving, although the (mis)usage has become so widespread that it might be worth using different language.
In my mind, stochastic terrorism is a confluence of three factors:
1. You’re attempting to lower the status of a person or group to the point that a non-zero number of people who agree with you might consider violence against them
2. You’re attempting to increase variance in opinion against that group to increase the number of people predisposed to violence while maintaining a large population of people opposed to violence (for plausible deniability)
3. You’re attempting to erode norms against violence as a political tool, or at least bias listeners towards violence relative to other approaches.
The conjunction feels important to me because it takes all three to incite violence without being held accountable for it. 1 & 3 is just an obvious call to violence (or toothless, depending on how strongly worded your speech). 2 & 3 is undirected - you might stir up controversy and erode beneficial norms, but it’s not targeted. And 1 & 2 is really no different than any other attempt to shift public opinion.
But when you have all three it’s easy to claim that you’re not inciting violence against a group - after all, most people considering violence would never act on what you said - while simultaneously substantially increasing fringe violence against your target.
Perhaps you could give examples? It's unclear to me what kind of speech you think would clear all three bars (and would need a term like "stochastic terrorism" rather than plain old incitement of violence)
I haven’t read the essay, but isn’t the title just so humorous? “Against terrorism”; you might as well follow it up with an essay titled, “Why war is actually sort of bad”
Having read the argument, I totally understand what you mean now.
Still, an interesting thesis. Would you say that stochastic terror *doesn't exist* (Isn't terror / shouldn't be considered terror) or that it *shouldn't exist* (shouldn't be taken seriously as a concept)?
The US already has extensive, well-thought-out legal theory and precedent specifying how to deal with this problem: defamation law, at the federal and state level. US First Amendment law has been very carefully thought out and rigorously argued many times over the centuries; if you aren't an expert on it, you shouldn't try to fix it in a blog, let alone a blog that doesn't mention the First Amendment.
We have the problems we have not because there's anything wrong with the law, but because it isn't applied. If you say someone is "literally Hitler", and they are not literally Hitler, and they can show you knew either knew they were not Hitler, or acted with reckless abandon (e.g., not bothering to check whether they were Hitler), or with malicious intent, and caused them material or emotional harm, then US law generally says you have defamed them and are liable for damages.
Then people say, "Oh, but so many people do it!"
News flash: When San Francisco stopped prosecuting shoplifting, lots more people started doing it.
Don't feel sorry for someone who called someone else "Hitler", especially if they literally used the word "literally". If they're adults, they will soon figure out that actions have consequences--but only if actions really DO have the legal consequences specified in our law.
The situation we're in today is that, at least according to every human and chatbot I've consulted, judges usually dismiss open-and-shut clear cases of libel if they're considered common behavior. That's not how law works. Worse, this endemic disdain for the law, on "practical" grounds, teaches, guess what, disdain for the law.
I don't think stochastic terrorism and defamation law really connect.
I think a typical example of what would be deemed stochastic terrorism would be something like "Muslims are going to take over America and implement Sharia law and maybe rape your daughter". This is a prediction rather than a factual statement about the present, and it seems like many people who make predictions like these believe them. I think that means a defamation case would be quickly thrown out. Also, would anyone even have standing to sue for defamation in a case like that?
Yes. Even in the case of Nazi name-calling, is an accusation that someone "has similarities to Hitler" (almost certainly true about many public figures, especially in the cases of politicians doing authoritarian things) that much better than "literally Hitler".
I think the US law that most directly engages with the question is incitement. The bar for this is extremely high: the controlling precedent (the Brandenburg Test) requires finding that the speech is intended to produce "imminent lawless action" and also is likely to actually do so. Without intent, without likelihood, or without imminence, it's protected speech.
Besides this, I'm also familiar with at least one case that found civil liability partially on the basis of an "aiding and abetting" theory in a wrongful death lawsuit. The book "Hit Man: A Technical Manual for Independent Contractors" was marketed as nonfiction, a guide for how to take up contract killing as a career. After someone was convicted of murder after actually using the book as such, the family of two of the victims sued the book's publisher on the grounds that the book encouraged and enabled the murder. The trial court initially dismissed the case for failing to meet the Brandenberg Test, but an appeals court reversed this 3-0 on a combination of the "speech-act doctrine" (that speech can be judged as an action rather than advocacy, such as in cases where people have been criminal convicted for giving seminars where they coach people on committing tax fraud) and on an argument that the trial court was wrong and the Brandenberg Test actually had been satisfied.
You're right that they aren't seen today as connecting, and that this hinges on having the standing to sue. I totally overlooked that when posting, which is what I get for posting with a migraine.
But a strict interpretation could say that, if I say something which logically implies that you are probably an X, then I can be held responsible for saying that you are an X. If you said "Jews drink the blood of Christian children" in a way that implied that most Jews drink the blood of Christian children, and I were a Jew, I think it would be just that I then have standing to sue you. Think how many pogroms this might have stopped.
(The hard part would then be proving damages. I'd probably have to make a class action suit.)
Hate speech is just disapproval expressed without quantifiers. And I'm not joking when I say that what we need most to heal the US today is to teach people to speak with quantifiers.
Maybe not. I think I formed my opinion under the false impression that hate speech prosecutions were happening in the US, when what is happening is hate speech prosecutions in Canada and Europe and the UK, and hate speech non-governmental retributions in the US. My reasoning was that the robust body of law on free speech should be applied to hate speech prosecutions, but in fact US free speech law is being interpreted as making hate speech laws unconstitutional.
You're a Hogwarts student; am I accusing you personally of something?
What if there really is a group of Slytherins who like regularly pulling pranks in my conveniently-located garden, so it's a completely true statement and I'm not trying to imply that everyone else is joining in?
When someone says "Hogwarts students vandalize my garden", it's clear the speaker can't mean "Most Hogwarts students vandalize my garden", because you probably couldn't even fit half of them in the garden. When someone says, "Jews murder Christian children", they're referring to a specific legend which says all Jews are required by their religion to regularly murder Christian children, to use their blood in the passover bread.
If the same legal reasoning about the context of a statement used in defamation law, were used in hate speech law, then statements like "The Quran tells Muslims it is their destiny and their duty to conquer all the earth" couldn't be regarded as hate speech (since this is a verifiable fact, not obviously made with malice, and doesn't assert that most Muslims want or intend to do so). Statements like "Hogwarts students vandalize my garden" would be interpreted as a statement at most as strong and inclusive as"Rabbits eat my carrots", which clearly doesn't mean "All rabbits eat my carrots" (although it is probably true that all rabbits would like to eat the person's carrots).
There are rulings in defamation law about how to interpret natural-language sentences according to context. But they aren't consistently applied in hate speech prosecutions. Creating a separate category of hate speech re-opens all the loopholes that have been closed in defamation law in the 235 years since 1791.
That's of course not true. Moreover, in UK there are about 10000 arrests for speech violations per year. There is no conceivable way of verifying all of these people are "pro Palestinian activists" - officials do not publish any such statistics, of course, and any research that could survey so many people would be nothing short of herculean, and for sure would be known just be the sheer amount of effort. It does not exist. Thus, not only I can know it is not true, I can be sure you know it is not true - or exhibit what is called "reckless disregard" for whether it is true or not.
None of these are "pro Palestinian" activists, as far as I can see. None of these require more than 5 minutes of internet search to discover. My question is: why did you claim something that is so obviously untrue?
The 10,000 includes visits for harassment, calls to violence and so on. Still one visit or arrest for bad opinions is one two many. I side with Mr Linehan. The classification of an activist group as terrorist is fairly unique and the arrest of supporters who protested said classification were also arrested. “Since the UK proscribed Palestine Action as a terrorist organisation in 2025, hundreds of supporters have been arrested for allegedly expressing support for the banned organisation at demonstrations”.
Merz' complaints led to warrants for around 5000 people, some of whom were only quoting other people (the people being quoted, oddly, were not arrested). These were folks who had said bad things about Merz. Nothing to do with Palestine.
See https://www.youtube.com/watch?v=tB3WVygAM8I , it includes an explanation why the "10000 arrests for speech violation per year" does not mean what Joe Rogan thinks it means.
I'm sorry, I'm not going to watch half-hour video from a "comedian" trying to gaslight me into the idea that UK has freedom of speech and US doesn't, despite all available evidence. If there's an argument behind it, you are welcome to make it, it should't take half an hour to read. I have no idea what Joe Rogan means by anything, neither I care too much, but I can see UK laws criminalizing speech, and I see them actively and vigorously used to police political speech in UK, and - so far comically unsuccessfully - in the US.
Fair point. The summary is that the number that people typically quote is a total number of people who get into trouble related to the "Communication Law", but a majority of that is unrelated to saying mean things on twitter. It includes things like harassing your ex by calling them on the phone many times a day. When you get to the actual "saying mean things on Twitter" (now the numbers are much smaller, comparable to USA), a typical "mean thing" that gets prosecuted is something like calling people to burn down a mosque, which would get a similar legal response in USA too. (Specific examples in USA are provided in the video.) And yes, some obvious mistakes happen too, but again their number is comparable to similar mistakes made in USA.
I find this argument disagreeing with the facts. US and UK have very different standards of what considered unprotected speech, in the US it must be incitement to an imminent lawless action (not just generic "mosques need to be burned" but a call for burning of a specific mosque that a reasonable person would conclude may immediately lead to the actual burning). Even that kind of speech is prosecuted quite rarely. Speech that is considered merely "offensive" and regularly prosecuted in the UK is never prosecuted in the US (at least not criminally, you could get fired or "cancelled" for it, but that's a different thing). Of course, since you did not provide any specific numbers or examples I can't give any more factual response, but let's consider this.
UK clocks about 10K "offensive communication" arrests per year now. US population is 5 times larger. So, if US had similar levels of enforcement, we'd have 50K speech arrests in the US in a year. Can you show that happens? Can you even show 5K speech arrests per year in the US? I suspect unless you really stretch it - like list violent riots as "speech" - you may have trouble getting even to 500. With that in mind, arguing about specific cases, being a mistake or exception or whatever, is pointless. 10K cases per year can't all be mistakes, it's a policy.
No, "John Smith is Hitler" is clearly a metaphor, and any sensible court would see it as that.
Same thing for "John Smith is literally Hitler" - "literally" is being used as an intensifier here; that usage is literally destroying the worth of the word, but it exists and any sensible court would recognize it.
Literally is fine there. I will literally die on this hill.
It’s an intensifier and the whole sentence is hyperbole, hence the use of literally is figurative, since the entire sentence is figurative. It’s not its own contronym, you wouldn’t use figuratively there.
Let's not argue over Hitler. I've had dozens of entirely baseless, damaging, and easily-disproved slanders made against me in Internet-print, sometimes along with frank admissions that they were malicious, but have been advised by everyone I consulted not to sue, because (I was told) the case would be dismissed, and I might be hit with some anti-SLAPP law forcing me to pay the defendant's law fees for starting a "frivolous" lawsuit. All because most political slander gets a free pass nowadays. This is a vicious circle.
I am not a lawyer, but every lawyer I've seen online has described the legal bar for defamation as much higher than you set it here. It's not that prosecutors or plaintiffs are declining to go after people, it's that when they do, courts recognize that "literally Hitler" is obvious hyperbole and it will not cause people to believe that you are literally a reincarnation of history's greatest villain.
See the case of Carl Sagan v. Apple Computer (1994), where a US District Court ruled that Apple calling Sagan a "Butt-Head" was not libelous.
>There can be no question that the use of the figurative term "Butt-Head" negates the impression that Defendant was seriously implying an assertion of fact. It strains reason to conclude that Defendant was attempting to criticize Plaintiff's reputation or competency as an astronomer. One does not seriously attack the expertise of a scientist using the undefined phrase "butt-head." Thus, the figurative language militates against implying an assertion of fact. Unelko, 912 F.2d at 1054.
....
>Because a reasonable factfinder could not conclude that "Butt-Head Astronomer" implied that Plaintiff was a less than able astronomer or that Plaintiff was legally wrong in asking Defendant to cease using Plaintiff's name, the only remaining assertion is the bare statement that Plaintiff is a "Butt-Head Astronomer." Clearly this phrase cannot rest on a core of objective evidence. Plaintiff does not suggest any other assertions of objective fact that could be reasonably implied from the phrase.
(Also not a lawyer, this is admittedly speculation)
I suspect this may also be because most well-known defamation cases involve people who are various degrees of "famous" and have consciously chosen to live in the public eye. At which point, they've opened themselves up to increased criticism in exchange for greater influence.
In the cancelation discourse, I have much less sympathy for celebrities/politicians who take blows because they often have comparable social capital with which they can fight accusations. Whereas if an otherwise private individual has the misfortune of saying something careless, then goes viral, it would have much greater potential to ruin their chances for employment, social standing, etc. Hence, it would presumably be much easier for them to prove damages in a court of law.
Granted, now that anyone can post online without getting vetted by a media outlet, the distinction between a public/private figue may have collapsed...
Under U.S. law, public figures have to meet a higher bar in establishing *culpability* for defamation (i.e., whether the speaker should have known they were spreading lies), but the standard for whether a statement counts as defamatory in the first place is the same; it has to be a statement of objective fact, as interpreted in context by a reasonable member of the audience where it was shared, and that statement has to be false. The law is very specifically *not* designed to provide a remedy for heated rhetoric targeting someone (that doesn't cross the line into incitement, which is likewise a high bar).
Not sure I agree European hate speech really is such a cleanly demarcated legal category. Haven't been following super closely but my sense is that there's a broad, vague range of critical-of-groups or critical-of-individuals speech in the UK or Germany that is legally risky, and without some limiting principle that makes it easy for people to speak freely so long as they make sure to do so "non hatefully."
As a German, hard disagree for Germany. It is very clear-cut what is allowed and forbidden. I don't find all rules good. In fact, I am strongly opposed to the new law that allows punishing defamatory attacks on politicians. (There is some hope that it will be cancelled again.) But also there, the scope of the law is very clear.
Yes, there is always a grey zone, but it's a pretty small one. You hear about it because no matter how tiny the grey zone is, people intentionally go for this zone. For example, the slogan "Alles für Deutschland" (all for Germany) is forbidden as a Nazi parole. The right-wing AfD party with party leader Alice Weidel pushed the slogan "Alice für Deutschland". The words "Alice" and "Alles" are pronounced basically the same. Since it's often shouted, does pronounciation or spelling count here?
But someone who doesn't intentionally go for the grey zone will not accidentally end up there.
Thanks! I'm sure it is the law against insulting politicians that i was vaguely remembering. People getting arrested for stuff like satirically representing a politician as saying "i hate freedom of speech." Is that in the grey zone or is it clearly prohibited? I'm not sure which would be worse:
Ok, I admit that this is an awkward case for my claim that there is no big grey zone. The judge in first instance said that the picture was so realistic that readers would not recognize it to be photoshopped, thus falsely believing that the politician actually showed this sign for real.
Part of the story is also that the decision was revised, and in the revision both prosecutor and defendants motioned for acquittal, which the judge followed. The decision of the first instance was also condemned by politicians from left to right. I would rather file this as "in rare cases first-instance decisions can be odd". But yes, I guess it shows that they grey zone is larger than I made it appear.
But what all sides, including the judge in the first instance, agree on: if the audience recognizes that it is satire and not a real photo, then this is allowed as freedom of opinion.
People are upset that Merz sicced goons (okay cops) on them for quoting Other People saying nasty things about Merz, and did not actually sic the cops on the people making inflammatory statements.
This seems to suggest that even discussing someone saying something nasty about a politician might get you arrested. That is... not "clear-cut" in a way that encourages discourse.
(If your response to this is something along the lines of "Merz is gone, and nobody else is that easily trollable", that would be relieving to hear!)
I am not sure what you are referring to with "Merz sicced goons". There is one case where someone got a fine for calling Merz "Lügenfritz" (Lügen = lies, Fritz = short form of Friedrich). The case never went to court hearing since the person paid the fine, and Merz was not involved in this case. The case started without him having filed a complaint or so.
There are three things which are prohibited: defamation, slander, and personal insult. All three were already forbidden before and could be pursuit in civil courts, but the new laws makes them penal offenses if they are committed against politicians, so now they can be prosecuted regardless of whether the politician files a complaint.
Essentially, there are two things that you are not allowed to do:
- making factual claims about politicians that are provably wrong, if you are aware that they are wrong or just making them up. (And if they are supposed to make the politician appear bad.) It is forbidden to just claim "Merz has taken a 10.000€ bribe" without indication that it might be true, or claim "Merz has voted to make abortion illegal" if this is provably false.
- Making insults that go against the person and that are not related to political acts. Just saying Merz is an "idiot" or a "liar" without any context to his political actions is actually forbidden. But saying that "he treats the poor like dirt" or "his tax reform is idiotic" is totally fine. A guideline is that if you criticize how a politician uses his power, that is always permitted. But if you criticize the core of the person, not his actions, then you are treading dangerous legal ground.
Let me retreat a bit from saying that there is no grey zone, because insult and criticism can be mixed. If you say "Merz is an idiot, this tax reforms goes against the wrong people", that's still ok, because the insult is given in context. But if you say without further context "Merz is a rotten bag of scum with his politics", then he insult is heavier and the political reference very vague, that's probably forbidden. But the problem is not the grey zone. It's not exactly difficult to stay out of that, and the only limit on how you can criticize politics is that you are not allowed to a) make fabricated or provably wrong factual claims and b) get personal.
I want to emphasize again that I do not endorse the new rule that makes those things penal offenses. But I do endorse that these things are forbidden and can be persecuted in civil courts, as it always used to be in Germany.
Thousands of complaints. Many of them personally signed by him (others he hired a lawfirm to make the complaints for him). Haubeck has similarly large numbers of complaints (800 complaints).
I do not have the link at the moment, but if I find it, I will share it (I will look). I remember reading (I think it was on substack) about Merz hauling a guy off to jail (presumably released afterwards) for "quoting others saying bad words against Merz" (the others were not sent to jail).
Penal offenses seem to make "cops at your door" (aka something similar to the Swatting pranks) here to steal your computer (see Steve Jackson Games), your cellphone, etc a lot more likely.
It doesn't bode well for a discussion that you call the confiscations of phones/laptops "stealing". I also don't like that US border control can confiscate my phone when I enter the US, but calling it "stealing" does not help with the discussion. So after this post I will disengage and leave the last word to you.
Yes, some politicians do file complaints, and some of them file many. This included Merz before he became chancellor. Since then, according to his own statement in a recent interview, he no longer files them. He added that he would be open to reforming this paragraph, which is why I have some hope about it. At the moment, the suggestion (not from Merz, but from the conference of Ministers of Justice) is to restrict it to communal politicians, which were the main group intended to be protected by the new law in the first place.
I totally agree that confiscating a laptop or a phone is out of proportion for such cases. According to your own sources, this is also how German courts ruled about these confiscations, so I assume they no longer happen. This was one of the many reasons why the law is bad, but this particular one has apparently been fixed.
All of this is not about the original point whether the law is clear, but about the orthogonal point whether the law is good or bad.
Your first distinction ("Most societies distinguish between simple criticism and true incitement to violence") seems pretty concerning to me.
As a thought experiment, which person is worse:
1. The guy that tells Gandhi "Kill so and so, here's a Glock, he's at such and such place".
2. The guy that tells his dimwitted sycophant "Trump does so many bad things, we'd be better off without him"
It seems obvious to me (maybe not to you?) the second is worse, because the issue isn't the language you use - it's the chance that your speech will cause the listener to do the violent action.
If you agree to that, the difference between stochastic terrorism and incitement is just a matter of degrees, and one can be worse than others in different circumstances, and there's no obvious line that is crossed to move from ok to not ok. Falling back to "US courts distinguish between them" isn't very convincing.
The point isn't rule-lawyering. The point is recognizing "if some speech is likely to provoke violence against your political opponent, and you know it and still speak in this particular way, and especially if you clearly express the idea that you are fine with that kind of violence, you should bear societal consequences, even if the law can't get you". The details can be rule-lawyered, just as lawyers earn their living hashing out the details of criminal and civil code. We're fine with that, as long as we have the code, and we agree on the fundamentals like "do not murder, do not steal" etc. Here I still don't see agreement on the fundamentals.
It's not about rule lawyering, I agree. My point is that Scott wants to differentiate between criticism and incitement. My point is that no such distinction exists. They're both on the spectrum of speech that may cause a person to a violent action (and I'm using the example of a case where criticism can actually be more of a cause than incitement)
Yes, there is a spectrum, and there will be very debatable zones in this spectrum. But we know there is speech that is ok, e.g. saying "N is a bad mayor and should not be re-elected" is not seriously taken as a call to violence by anybody. There is also speech like "I will pay the bounty of $100K to anybody who murders mayor N" - which is already illegal. And I think there should be kind of speech like "whoever murders mayor N will be my personal hero, every decent person should think about murdering mayor N, and I rejoice every day thinking this may be the day mayor N is finally murdered" which should not be persecuted by law but sanctioned by the society. Yes, there won't be hard boundaries, and yes, there would be a lot of bickering in the gray areas. But talking about how we bicker about grey areas only useful once we establish that at least boundaries exist. I think we haven't established that, and, moreover, Scott is trying to dis-establish that.
I think the better option is to push back against speech which fails the Brandenburg test for incitement but has some elements of it by calling the speech irresponsible.
under US law, it's only incitement if both (a) there was intent to produce imminent lawless action and (b) it was likely to actually achieve that goal.
it sounds to me like you think example (1) would be incitement but not stochastic terrorism, but i think it could not be incitement, because the likelihood of a violent result would be negligible.
incitement is a categorically stricter concept than stochastic terrorism, because it requires not only a likelihood of lawless action, but also that the lawless action be likely to happen (more or less) right away, rather than next week or next year
Your view leads to the conclusion that important political topics can not be openly and honestly discussed in the presence of dimwitted sycophants. It could conceivably be true that Trump (or any other political figure) does so many bad things that we would be better off without him. The fact that it could be true means that it is important that people should be able to say it. On the contrary, there is no socially beneficial activity that requires one to provide someone firearms and instructions to murder.
> Popehat, aka attorney Ken White, has been permanently banned from Bluesky, apparently for a post in which he suggested there was a moral upside to Elon Musk’s death.
I am against prosecuting such people, and I am even against banning them from social media (though I couldn't care less what is happening on bluesky specifically). But I am for treating people like Ken White as revolting moral imbeciles, and shaming them mercilessly. I do not insist on calling him "stochastic terrorist" or any other particular term, I an fine with any approach that gets the result of people like him not being accepted by society as normal. And so far I didn't hear any convincing argument as to why I shouldn't see him as morally revolting.
He has been suspended for four days, apparently. In his follow-up post ( https://www.popehat.com/p/a-bit-of-tedious-drama-at-bluesky ) he engages in a lot of the-other-side-did-it-too, both specifically for Musk and generally for Trump supporters.
I guess the question is, if you kinda-sorta encourage people to kill your political opponents, like Musk did, does that make it fair game for other people to kinda-sorta encourage killing you? With actual killing, I think most people would say it does, if the state cannot be relied on to handle the situation for some reason. With vague verbal references about killing, since they (contra the stochastic terrorism talking point) don't actually have much effect, mostly you'll just end up with a very polluted public discourse.
A retweet about the movie? Please. While Ken White, in the same article, explicitly says he meant every word, so there's no misunderstanding what he is calling for. He is calling for Elon Musk to be murdered, because he does not like his political speech. He is not ashamed of it and he specifically explains that is exactly what he is inciting. If there's ever a clear-cut case this is it. And he knows it's a real possibility - that's exactly why he is doing it.
> Now, I think people of good faith can disagree about the morality or utility of wishing other human beings dead. [...] The loudest cries of outrage are from people who will diagnose you with Trump Derangement Syndrome if you object to the ocean of blood I just described. The reaction is largely contrived, mostly in bad faith, and rarely to be taken seriously. The people landing hardest on the fainting couches are in two groups: pro-Trump people who are thrilled that we are extrajudicially executing fishermen in the Gulf, and professional grifters who don’t necessary like the extrajudicial killings but whose entire gimmick is “aren’t those leftists silly and outrageous.”
Isn't the obvious solution "you should be able to criticize people harshly, but not imply or outright state that they deserve death, or that they should be killed or otherwise harmed"?
The silly accusation of stochastic terrorism in the post aside (it's a dumb Twitter post with a few hundred likes, many such cases! I wouldn't consider it representative unless it has some decent traction), most of what I see credibly labeled as such are calls for the target to actually to be terrorized; for example, Hasan Piker posting DIY diagrams of guns in response to politicians he doesn't like (https://imgur.com/jbiMLSN), or outright saying people should kill certain politicians if they don't like their policies (https://wjla.com/news/nation-world/turkish-twitch-streamer-banned-after-urging-viewers-to-kill-florida-senator).
We could taboo the words "stochastic terrorism" to describe such behavior, but I think we could all agree this is bad and likely contributes to rising political tensions and/or violence.
I think if you look at https://x.com/search?q=%22stochastic%20terrorism%22&src=typed_query&f=live (and where better to get a pulse on how people are using phrases?), you'll see that most people are using it to mean criticism they don't like, not real incitement to violence. We already have a term "incitement to violence" for the latter situation.
Also, the first thing I saw on there was the claim that something was "a dog whistle for stochastic terrorism", aaaargh.
I think that's more symptomatic of social media's tendency to stretch out and water down these terms over time by incentivizing people to make huge reaches to overstate things for rhetorical effect, see e.g. "genocide" to mean "the government might not directly fund gender affirming care for kids", "grooming" to mean "relationships between consenting adults with more than a 3 year age gap between them", etc.
But I don't think the term stochastic terrorism is entirely without its use - like the examples above, there are real bad behaviors that can be accurately described using the word, people should just reserve the words for those cases instead of reaching to use them as cudgels to bludgeon mild things they don't like.
"stochastic terrorism" is a terrible term, but the Left invented it, and now it's here, and it's attached to the identifiable phenomenon, so in the absence of a better term we're stuck with it. Many such cases - we have "piracy" as a designation of "copyright violation", as if robbery on the high seas has anything in common with watching a movie without giving Disney your money, or fixing a software bug without paying for the support contract. But it's what it is and we're stuck with it.
I don't think "piracy" is quite analogous, it's become basically a term of art in the context of digital media, whereas the vast majority of the usage of "stochastic terrorism" is just people on social media misusing the word to mean "criticism" or "calling for things I don't like" rather than the more appropriate meaning of "implicitly or explicitly calling for terrorism/violence".
"Piracy" to mean copyright infringement is pre-digital; in fact, the name of Gilbert and Sullivan's "The Pirates of Penzance" is a reference to this (which had plagued them with HMS Pinafore).
>"stochastic terrorism" is a terrible term, but the Left invented it, and now it's here, and it's attached to the identifiable phenomenon, so in the absence of a better term we're stuck with it.
Probably right, but I wince at it. 'Irresponsible' would probably have been a better choice. Amongst other considerations, it is generally not reasonable to expect everyone to be responsible at all times, though one wants fairly strong norms about it most of the time.
Also doesn't change the fact that Piker being a popular streamer instead of some streetcorner schizo rotting away is an indictment of American culture.
There can be no evidence because it hasn't been gathered yet, or because there is no evidence to be had. I think that politically-motivated assassinations (and even DIY guns) are rare enough that the latter is the case here.
I agree that it doesn't speak well of people that they are fans of someone like Piper.
I find it interesting that this premise - "criticism is violence" and even "silence is violence" and even "support expressed not in ways that we approve is violence" - has been the foundation of the whole cancel culture. Nobody really argued that using a wrong pronoun or wearing a sombrero while being white literally murders people. But the argument was if you keep using wrong pronouns and wearing sombreros, it creates an oppressive environment in which eventually, after enough logic jumps, the violence will happen. And this, after enough logic jumps, is tantamount to genocide. So it is prudent to suppress the nascent genocide in the bud and treat misgendering if not as actual-violence then surely as one-tiny-step-removed-from-actual-violence. And if we save even one life by that, it's well worth it.
Now, when the cancel culture had run its course, when even single institution has DEI commissars, affinity groups and speech codes, when anybody that had any inclinations of saying something anti-woke is driven out of every important institute of the society; when they finally dusted off, licked their wounds, started building their parallel structures, and then started discovering that once these parallel structures begin to work and allow them again to discuss and spread their ideas... The Left reacts to them with actual hot violence - not counter-words, but bullets, firebombs and bike locks to the head - then we suddenly discover that actually violence is not connected in any way to the political propaganda.
Hundreds of thousands of people can call for murder of conservative figures - completely in the open, with known URLs and sometimes under their real names - and yet, when one person writes in their diary "I had enough, I am ready now, I will do what I have read about ten thousand times on social networks and murder the Conservative Figure" - and then they do exactly that - now we are supposed to think it's not related. If ten thousand people say "Conservative Figure must die, we pray for the day he dies, we want him to be shot, we live for him to be shot, whoever shoots him would be a real hero and we will worship him" - and then, when he is shot, they deliver on the promise, they rejoice, they worship whoever shot him as a hero and they name the list of who should be shot next - all this is 0% related to the shooting. 100% somehow it appeared in one crazy head, just out of the blue sky, and there's nothing else to see here.
That's the "liberal solution". We just discovered that, sorry we didn't know about it all those years, but you know, science works in mysterious ways.
Maybe for some people all that makes sense. For me, I get a strong impression somebody is trying to sell me a Brooklyn Bridge. Sorry, no deal.
JRRT already answered that question. There might be people who deserve to die, there are no people who are qualified to decide that. Before you ask, yes, that's why I also oppose capital punishment.
We're not talking about the situation of self-defense or warfare. There's a lot of difference between soldiers at war and a person deciding "mayor N deserves to die because he's literally Hitler, so I will take a gun and murder him". I do not believe anybody can seriously not understand this difference, only pretend to not understand it to achieve some rhetorical goal. Whatever it is, it will have to be achieved by other means, if at all.
Didn't you make a comment saying that was your logic for opposing the death penalty? So your belief is that even the state, with its Weberian monopoly on the legitimate use of violence, cannot decide to kill people, even if those people have been convicted of murder. You need to have some rule distinguishing why one case is ok and the other not, because "nobody can make that decision" doesn't cut it.
I don't see what's your problem here. The situation of combat and self-defense is different from killing someone who does not pose any threat to you, just because you decide you want to do it because they "deserve it". Murderers can be easily neutralized by keeping them in jail forever, thus there is no self-defense aspect in killing them. If there existed some super-powered maniac that could not be held in jail, then the question would arise, but those do not exist, so it doesn't.
Combat situation is completely different - nobody sits with the CVs of enemy soldiers and decides - this guy deserves it, this guy doesn't, this guy donated to charity and saved 100 kittens, so we're telling the artillery to take 1 degree left to not hit him. That's not how the war works, so "deserving" doesn't come into it.
Gandalf's rule, as borne out by his actions throughout the story, seems to be that it's okay (often laudable) to use violence to deal with enemies who are actively doing evil, but once someone is captive or otherwise no longer a dire threat you should offer mercy. Note his treatment of Wormtongue and Saruman as well as Gollum.
While I consider the idea of "stochastic terrorism" mostly illegitimate, I do make at least one exception: Most things that could be fairly described as "doxing" are stochastic terrorism. There's almost never a valid reason to try to unmask an online commenter or bring attention to their home address, only attempts to intimidate or incite violence.
The term "doxing" has started getting applied widely enough that I wouldn't rely on it for such things. Also, it used to be the case the the phone book published everyone's home address! I will agree that frequently things called "doxing" actually do resemble the kind of speech a platform would be better off not hosting.
The phone book published the home address of Joe Smith.
It did not publish that commentator Jimmy Beanslop is actually Joe Smith.
It also didn't say "Heterodox commentator Jimmy Beanslop is actually Joe Smith and here's where he lives; progressive commentator Sturgill Oklahoma is a national treasure and we'll respect his pseudonymity."
Each extra step reduces the probability of something happening. If mindkilled people are tweeting about how John Doe is a horrible person, and one of them is in the mental state ready to go and kill a person they are sufficiently outraged about...
...it makes a difference whether such person realizes "actually, it is possible to find addresses of people" and "actually, the phone book contains the addresses" and then actually goes and takes the phone book and finds the address -- as opposed to juicier alternatives such as more tweeting.
...or the address is already conveniently provided in the tweets.
That said, law is often too crude tool to make such distinctions meaningfully. What if instead of posting the address you only say "btw, the phone book contains his address". Or, assuming that the phone book is online, you post a hyperlink to the phone book. Or to the specific page of the phone book containing John Doe's name. Etc. Expect people rule-lawyering, and then getting hysterical about "did you know that in Europe they can arrest you for literally saying on twitter that a phone book exists???!!!".
In addition to what ProfGerm and Viliam said, I would also note that our phone book allowed you to opt out of being listed. And they didn't even have addresses for residential numbers, only for businesses.
Just as a sidenote: do not post even interior pictures of yourself online, if you don't want sufficiently motivated people to be able to figure out where you live. I don't consider "using what you yourself put online" to be doxxing (although it's probably still bad behavior, and stalking the person is VERY BAD and WRONG).
In order to be meaningfully "stochastic" anything, there has to be a probability greater than zero. The probability of someone too lazy to figure out their target's address, ultimately prying their size-60 backside out of the couch in their parents' basement and going off to execute a successful assassination, just because someone else did one of the easy parts for them, seems pretty close to zero. Maybe it's happened somewhere, sometime, but not in a way that has ever risen to my attention and it doesn't seem to be a real problem.
People like Luigi Mangione or Tyler Robinson, find their own targets.
The probability will often be low, but never zero. And this has to be weighed against whether there's a valid public interest in knowing where the target lives - if there isn't, it doesn't matter how low the probability was, it was still wrong to do.
Not all murders are high-profile. Not all assassinations are known to be assassinations if the police don't figure out the motive.
And laziness isn't always the main barrier to figuring out an address. It isn't always a matter of just looking up something that was an open secret.
Yeah, "...but there's no 'valid public interest', so we can silence that offensive speech", is the nose of a very large camel that is wholly unwelcome in my tent.
I'm usually very sympathetic to that kind of argument, but in this case I think the doxing *is* the camel nose for bad things, especially when it happens to someone who isn't a public figure.
Plus, people do (or should) have a right to privacy, even in the absence of threats.
What do we accuse someone of if we think they actually are trying to get "someone" to commit violence but are trying to hide it behind plausible deniability? The obvious example is trying to prove that a mob boss ordered a particular hit...
The difference here is that Henry II had power to pardon and to reward. Followers would be able to hear his words as an implicit command. The modern day equivalent would be Trump's tweets around Jan 6, or Duterte encouraging vigilante slaying without explicitly commanding it.
>Maybe there are some things you shouldn’t say even about the KKK - “they are Satanic cockroaches unworthy of life” might cross that line.
So dehumanizing metaphors can move the needle on violence then? There is a fair amount of social-psychology research suggesting it does more than generic negative claims. Or are they bad just because they're obscene?
I think the cockroaches reference specifically comes from Radio Télévision Libre des Mille Collines (RTLM) where it was their preferred term for Tutsi's during the lead up to the Rwandan genocide.
Google/Gemini says that the rather similar "Death to the fascist insect" goes back to 1973 (weirdly, to the Symbionese Liberation Army (SLA) - very weird times...)
I agree with most of this, but was struck by the line "But if someone fails to follow this discussion norm, their penalty is being self-sorted to spaces that have bad discussion norms, not being held responsible for any resulting acts of violence."
At least in these circles, these people (1) often aren't pushed out at all; (2) will sometimes be pushed out in a way that results in them forming much worse groups, sometimes with intent to organize and foment violence or disruption; (3) will instead push out good people. I would like ideal punishment to be meted out, but I've come around to the view that I'm engaging in wishful thinking and we need another liberalism-compatible solution.
Firstly, stop using "generic" statements. "Men do this", "Purpletonions do this", "Cops do this". Stop generalizing. We use generalizations ie Groupings to allow us to discuss the world intelligently in simple terms. Unfortunately, some people, and I'm going to blame the sociologists, have got it in their heads that Groups are what the world is made of, and membership in a Group gives you all of the aggregated rights and obligations of all the members of that Group. So blame goes to all members of a Group regardless of the actions or merits of individual members of the Group. And privileging an individual member of a Group, who has not otherwise been disadvantaged somehow makes up for the disadvantages levelled against a different member of that Group.
Secondly, better to criticize the actions or beliefs of an individual rather than the individual themselves. So if someone doesn't believe, say for religious reasons, in the equality of women and men, then we should address that and say that it is an illiberal, and wrong headed, and ignorant belief, and harms society, without actually saying that an individual who holds that belief is evil or otherwise unworthy of being a member of our country or society.
I think that what you are asking is, in practice, close to impossible. Consider what you would have had to have done to avoid 'the sociologists'. About the easiest alternative would be to change that to 'the x% of sociologists who report on people via partitioning them into groups'. I don't think that that statistic is going to be readily available...
I think this is mostly correct but I think you left unaddressed the elephant in the room - what of a powerful political cult leader with an army of millions of followers who believe everything he says, and this powerful political leader also happens to control every national law enforcement agency and the military and is supported by a phalanx of prosecutors and judges who will leap up to defend him from any legal consequence and make sure his insane legal interpretations get a more than respectable hearing. If such an individual were to make harsh criticisms of religions and national origins and political opponents would that qualify as stochastic terrorism? Does Donald Trump telling a crowd of January 6ers to fight Congress (peacefully!) as they're counting electoral college votes count?
>You’re allowed to criticize policies, but not individuals. Okay, so somebody says that allowing immigration from the Middle East is destroying America, but avoids saying the specific words “and I hate individual Muslims”.
This is a weak example, because "Individual Muslims", is not in fact an individual, it's a category of people. "Abdul who runs the falafel cart down the street", is an individual. And "Immigration from the Middle East is destroying America, and I particularly hate that guy Abdul over there" is perilously close to incitement to violence.
Of course, this shades into "The alt-right is destroying America, and I particularly hate Charlie Kirk". I agree that we have laws that do a tolerably decent job of drawing that line, but I also think we need a social norm for cases that we're not actually going to take to court, And I don't know what that should be, except that I generally take care not to express hatred of specific individuals.
Perhaps because I'm not on X, I don't understand the leading example. Someone please enlighten me...
>Again, creating a spreadsheet to accuse someone of killing others is a self-evidently absurd exercise, whose only possible motive is to be used to incite political violence. This is stochastic terrorism, and nothing else.
Who created the spreadsheet? Where's the spreadsheet? I suppose in context Musk is being accused of killing others, but why would using a spreadsheet be a particularly evil way of criticizing Musk or anybody else? Do spreadsheets actually have a track record of inciting violence?
If everyone else who plays the stochastic terrorism card is just as incoherent, why are we even talking about it?
I'm guessing critics of Musk's DOGE activities tabulated in a spreadsheet every cut DOGE made, estimated the number of deaths resulting from each, and summed it all up to conclude he's killed 700k/millions/etc. people.
OK, thanks, sounds plausible. The "Again" implies the existence of a previous more reasoned critique, so I'll give the commenter the benefit of the doubt and not assume their own comment is as self-evidently absurd as it seems on its face:
"Sure, we agree that causing death is bad, but using Excel to draw attention to it is purely evil. You should be using PowerPoint."
If I had known to click on Jerusalem's tweet <twice>, I would indeed have found that article. An article in which the words "spreadsheet" or "Excel" don't appear. So in retrospect it wouldn't have helped. Thanks for the tip, though.
I like this piece quite a bit, and think it's worth engaging in the throat-clearing to say so.
I also think that at its core it's heavily relying on the ability to bright-line rule that non-state* violence is bad, and dodges the ethics of how we should think of "stochastic" phenomenon more broadly. Is a Stochastic Cancellation a thing, and is it bad?
How does Radio Télévision Libre des Mille Collines fit into all this? They said some awful stuff on the radio that resulted in some horrendous violence. It feels like the 'liberal solution' might not deal appropriately with a radio station like this.
- Criticize specific actions and propose solutions to them that abide by the liberal agreement
- Criticize abstract dis-virtues like bigotry, hero-worship, fear-mongering, etc. (similar to criticizing policies, but more action-guiding for the individual person)
- Constantly inject compassion for all individuals involved to fight the natural human tendency for out-group dehumanization
I like this quote from Thich Nhat Hanh to Martin Luther King Jr:
"I believe with all my heart that the [Vietnamese Buddhist monks] monks who burned themselves [in 1963] did not aim at the death of the oppressors but only at a change in their policy. Their enemies are not man. They are intolerance, fanaticism, dictatorship, cupidity, hatred and discrimination which lie within the heart of man. I also believe with all my being that the struggle for equality and freedom you lead in Birmingham, Alabama… is not aimed at the whites but only at intolerance, hatred and discrimination. These are real enemies of man — not man himself. In our unfortunate father land we are trying to yield desperately: do not kill man, even in man’s name. Please kill the real enemies of man which are present everywhere, in our very hearts and minds."
"Probably there are some cases where extra-state violence is acceptable, for example assassinating Hitler... the best I can do is say that at least this limits the damage by only legitimizing attacks against powerful state targets."
The correct response to this is "yes, and by the time you're trying to assassinate Hitler, you're no longer participating in political discourse, so this discussion about what speech we should tolerate in a liberal society becomes irrelevant."
There's also a distinction between thinking an individual act is morally acceptable and advocating for it politically. It's not hypocritical to think a specific vigilante murder was good,* while also believing that society would made worse off by more tolerance for vigilante murders (because most of these murders would be bad, and there would be second order effects).
*This statement is hypothetical and should not be construed as an expression of support for any particular murder.
If I say "[X] should watch their back because one day they're going to find my knife in it," then I could plausibly be investigated and charged for the threatening statement. That covers a dual purpose: it nips in the bud and also punishes the direct harm of causing fear.
However, if I say "[Broad group] should watch their backs because one day [etc]," then that often does *not* rise to the level of a criminal statement. The same violent sentiment is no longer criminal if the group is too broad for any single person to be specifically identified. (That is, "the bridge club that meets across the street every Thursday" is specific enough to be threatened, but "[members of the political party you don't like]" is not.)
I don't think that 'stochastic terrorism' is a useful concept when applied to individual people. A person is threatened or not, assaulted or not. A micromurder on a person rounds up or down. For a group – particularly one that already experiences a baseline level of violence – then inflammatory rhetoric can affect the 'climate', even if the 'weather' of day to day life cannot be so causally identified.
This doesn't necessarily mean that 'stochastic terrorism' can be criminalized in the US; elsewhere as noted it would be adjacent to hate speech laws. Still, I think it's a useful if inflammatory rhetorical device (read: jargon to be used when talking to the in-group, not something to deploy in persuasive speech) and possibly as an analytical tool.
In specific response:
> The liberal solution is none of these: it’s that it’s always legitimate to criticize any person or group however strongly, and never legitimate to use extra-state violence against them; if a violent act occurs, it’s 100% the fault of the perpetrator, and 0% the fault of other people who previously criticized the victim.
> (separate from this, I think it’s a useful discussion norm to avoid being too vitriolic or hyperbolic in your language, except perhaps once in a blue moon to make the point that you care about this issue so much that it’s your one exception to your non-vitriol policy. But if someone fails to follow this discussion norm, their penalty is being self-sorted to spaces that have bad discussion norms, not being held responsible for any resulting acts of violence.)
You have a contradiction here: the person who commits violence bears "100% of the fault," but the incitement to micromurder should still be considered unsociable and unfit for polite company. That social rebuke is a kind of punishment, even though it is far, far beneath a criminal standard.
To the extent that public speech affects the public – and I note that elsewhere you've championed public art that is aesthetic for similar reasons – then words do have consequences.
I missed the implications of this in my first reading. “Probably there are some cases where extra-state violence is acceptable, for example assassinating Hitler or overthrowing an illegitimate government”
There’s a problem here. And it’s not that killing Hitler is problematic, it isn’t.
Thus surely means we have to be careful about who we call fascist, or Hitler, because it’s ok to kill past Hitler, it’s ok to kill future Hitler.
> Thus surely means we have to be careful about who we call fascist, or Hitler, because it’s ok to kill past Hitler, it’s ok to kill future Hitler.
We need to clarify the exact circumstances under which it's okay to kill Hitler.
It's okay to kill Hitler only (a) once he's actually in power, and (b) once he has made it impossible to remove him democratically (by cancelling elections). It's not okay to kill 1932 Hitler because you think he might do these things, but it is okay to kill 1934 Hitler even if he hasn't done all the other bad things yet.
Well the thought experiment is killing Hitler as baby. Or a child. I disagree. Why not a good swift kick in the balls that will give him a squeaky voice when he’s an adult? Can’t see him leading the Nazis with a voice like Daffy Duck.
>You’re allowed to criticize these groups harshly, but not using special emotional or dehumanizing words.
I have some sympathy for this one as a social norm, while holding to something pretty close to the strong liberal position as my preferred legal rule.
A couple narrower but related social norm rules that I would be more strongly supportive of would be:
1. Don't cheer for actual terroristic or mob violence when it happens. When someone you don't like gets assassinated, the correct attitude should be "cool motive, still murder".
2. Show restraint when cheering hypothetical violence against real people, even cheering that falls well short of actual incitement.
What's the reason you singled out violence as a result of indirect encouragement? Because that concept seems to generalize to all kinds of activities, for example to people drinking bleach in increased numbers after Trump publicly discussed it.
Or shouting "Fire!" in a crowded theater, the classic example?
I agree it's a contentious topic, whether or not you should be allowed to say things that cause bad things to happen, or that cause _anything_ to happen, really. But the fact that "stochastic terrorism" exists should be rather obvious, because it's merely a special case of "words can influence people to do things", which is a special case of "organisms take sensory input and process it into some behavioural output".
I really don't see why someone who produces a certain input should under NO circumstances be held responsible for the output, especially if that output is somewhat predictable: "Will no one rid me of this turbulent priest?" Sure, it's a judgment call most of the time, but that's the unfun part: There is no absolute certainty in real life, and we have no choice but to make it up as we go. You may wish to cut through this mess with the "liberal principle" of assigning the blame 100-0, but that's just not how it will ever work.
Maybe dehumanizing language is the thing to avoid. Muslims are garbage to people I know. In Rwanda it was years of saying the other were cockroaches that may have led to people killing them with machetes like cockroaches.
The constant drone of X group is trying to destroy our country is not reality. They are not waking up in the morning wondering how today they can destroy the country.
The hard part is Trump and MAGA are convinced they’re trying to improve America but are trying to destroy democratic institutions so they can get in power to do it…so they literally are trying to destroy it but are self deceived they are the hero’s saving it, and they cast any opposition as a destroyer so we get locked into a spot and no one can move. No injection of truth is possible since it’s a lie.
I agree with your critique. I feel the only path forward is electoral change and a new government moving to stop the minority loud mouths and get normal people with good hopes back into the spotlight. Make conciliatory decency the norm again, and then let democratic dialog chart our path.
Posse Comitatus. I'm just going to say that when congress calls for the executive to violate acts of congress, they're trying to destroy the very institution they serve in. You know, like calling on the National Guard on Jan 6th (which the Pentagon rightly vetoed, on the grounds of "thats a very dumb idea").
I am also increasingly hearing politicians on the left say things along the lines of "the right wing party is undemocratic or is trying to destroy democracy, the only option to save democracy is to vote for our party", thus effectively espousing a system where there is only one legitimate party, ie. a de facto destruction of democracy... (and these types are also completely convinced they are the heros saving democracy).
> You’re allowed to criticize these groups harshly, but not using special emotional or dehumanizing words. Maybe there are some things you shouldn’t say even about the KKK - “they are Satanic cockroaches unworthy of life” might cross that line. But this is a motte-and-bailey
This roughly describes my view of the topic. I think this term has been widely abused and I do not accept the bailey here, that criticism, even harsh criticism, is equivalent to a call for violence, but there are certain statements that can qualify as an implicit call for violence and this is one example. Another example is that there were pretty literal calls for more Luigis after that happened.
I think it might be worth distinguishing the normative vs the descriptive parts of the debate. Can you plausibly increase the likelihood of a Trump assassin by cricizing him? Yes, even though it is highly unlikely your individual actions would have a significant effect. Does that mean you shouldn't be able to criticize Trump? Of course not, criticizing the president is a very basic right in a functioning democracy.
On the other hand, most people accept that inciting violence is not protected free speech. So the debate is really on where we draw the line on what counts as inciting violence.
I think one thing missing from this conversation is the claims people will make about the truth/falsehood of the stochastic terrorism in question. Like some activists would say you are doing stochastic terrorism if you engage in certain rhetoric about transgender issues, but that's because there really is an assault on transgender rights and we need to push back on it. And then if you said, well should we not criticize conservatives then, they might say "well no, but that's because it's not true that conservatives are under attack like trans people are."
This isn't really a counterargument, I guess it's just a question of how you deal with psychologically motivated people in general
Cite more numbers than they do? Transgender people are indeed more assaulted, but that tends to correlate with race and occupation more than "generic ally" would like to think.
I can't think of any time when the Secret Service was called in to go after liberals... So, yeah, the conservatives are under attack (kinda sorta, devil's in the details, but I already said that).
Obviously the Secret Service does go after radical leftists (I wouldn't call them liberals, but then I wouldn't call the MAGA movement conservative either), to some extent even during Democratic presidents. See e.g. https://www.cato.org/blog/trumps-anti-antifa-executive-order
This is the sort of thing I was talking about. People will be quick to say "of course this issue is real and happening" for the thing they care about, but when someone else does it, well now it's actually something else and they're not correct on the facts like we are.
I will am willing to swallow the pill that Yes, stochastic terrorism is real and does originate partly from the way we talk about subject X,
and that No, we should not ban people from saying anything about anything, and that includes me coming in off the top rope on eg. Charlie Kirk with some tasteless burns before his soul had even gotten half way to hell.
Much like his commitment to the right to bear arms, you have to be committed to the right to ask if people would rid us of such and such turbulent priest if you want free speech at all.
I think there's been a social change that the phrase stochastic terrorism gestures at, even if no individual speaker should be accused of stochastic terrorism. Nixon also abused a lot of power and usurped democracy, but I don't think there was the same fear of his life that Trump has. So you can describe the state of affairs as stochastic terrorism.
There's been a culture shift where political terrorism isn't as unthinkable. Blame social media rabbit holes, community alienation, etc but the 10% left and 10% right most extreme Americans act or profess to think differently.
Gallop: "Americans' Acceptability of Political Violence, by Subgroup % Sometimes OK"
30% 18 to 29
21% 30 to 44
13% 45 to 59
4% 60 and older
It's the flip side of Stochastic martyrdom: "the encouragement of activist tactics that will predictably get activists hurt or killed when applied at scale".
In both cases it's a downstream effect, any individual speech act isn't directly causal, but the concept should still be tracked to see how the US is progressing.
I feel like there is an audience contingency part of this. If Elon Musk uses Twitter data to create a private group chat made up of everyone on the platform that both engaged with anti-AI content and engaged with schizophrenia support content, and then started sending them articles about how Sam Altman was going to kill all of humanity every week, this would likely be 'stochastic terrorism' even though Elon should in the abstract have the right to write articles about how Sam Altman might lead to the death of humanity. Similarly, we could say that if some survey data came back showing that for some reason the ACX readership now consisted of 20% paranoid schizophrenics who are anti-AI, this blog would likely become slightly more cautious about AI rhetoric.
There is obviously a tension between audience management and being able to conduct truthfully in discourse that should resolve mostly towards the latter. Someone like Desmas fashions herself as a public intellectual, trying to steer policy, and this macro aim is more consequential and important than how one person may react to some idea she puts into the public sphere. However, even though her obligation to manage her audience is low, it is not zero. The right-wing argument is that a lot of left-center social spheres are made up of potentially violent far-left extremists, like maybe these people make up 10-20% of someone like Desmas's followers, and that because these people are such a large part of left-center internet spaces, the left-center has more of an obligation to manage their audience than they presently do. Calling them not doing this 'stochastic terrorism' might be hyperbole, but it exists somewhere on the spectrum between 'completely not stochastic terrorism' and 'obviously trying to incite crazy people to harm someone'.
I do not think their fear is completely wrong (it might be exaggerated, but a large share of the left has turned against liberalism), but I think it is unrealistic to expect one side to unilaterally disengage from pointed political rhetoric. The argument might have an internal logic to it, but it is obviously being deployed for ideological aims: one could just as easily make this argument about right-wing figures who spread harsh rhetoric about 'hypergamy' and 'western civilization dying'.
Having some direct experience with stochastic terrorism, I think this misses the point. It actually does not behoove our society to wish literal death upon the other political side. It wasn't always like this, either--Obama and Romney were kind and complimentary to each other. Now both sides screams for lawfare and political violence. It's better to respect your political opponents and argue like civilized people, than it is to shriek about how they ought to be shot; that speech, which deep down everyone knows is absurd, should be met with criticism and disgust. Nor do civility norms throw out some early warning system against Hitler--indeed, the abandonment of civility in a society more often preludes tyranny.
What you are talking about is usually called polarization, or normalization of violence, and most people would agree it's bad. Stochastic terrorism is different: the claim is that in a society where violence has (for better or worse) been normalized, even non-violent criticism can make someone a target. The cited tweet where someone was accused with stochastic terrorism was about a blog post that analyzed the consequences of DOGE and claimed (in perfectly polite language) that Musk is responsible for the death of millions.
I did think it was a particularly sloppy thumbnail this time. One low quality image won't undermine his credibility but if you get enough of them, then stochastically...
This seems like a very exploitable position to take. Eg. Public Figure constantly rails against Demographic. His audience is very extreme in their hatred of Demographic. One day Public Figure points out a specific individual within Demographic as being especially bad, but doesn't call for any violence. That individual is later killed by a random follower of Public Figure. The next week, Public Figure points out another individual member within Demographic. Again, later on, *that* individual is killed by another, different random follower of Public Figure. Every month or so, Public Figure focuses the anger of his audience onto one particular individual, without calling for any violence, and every time, a member of his audience either kills, harasses, or attacks them.
Does Public Figure essentially now just have a Death Note he can use with complete legal impunity? If he went to court, and then told his audience about how he thinks the judge is biased and evil, and potentially a secret member of Demographic, is it really in society's interests for him to just walk free? Seems shady.
This piece jumped the rails at some point and started conflating the original question ("can Commentariat / Media be held responsible for others' acts of actual violence via the charge of 'stochastic terrorism'") with the question "is political violence itself bad." To make things worse, Scott then further conflates political violence with revenge & vigilantism: as I understand it, Luigi Mangione targeted the CEO of United Healthcare explicitly because he believed that person had personally ruined his life by denying him needed care for debilitating chronic pain.
You can pinpoint the exact moment we lost the plot: "It’s vacuous to tell Brian Johnson’s family that they can’t complain about his murder just because health insurance CEOs are a privileged group." Yes, that would be a vacuous thing to say, but trying to make me sad for Brian Johnson's family is also a vacuous rhetorical device because it has exactly nothing to do with the question of "stochastic terrorism." (Or even, I'd say, with political violence.)
1. I basically don't agree with any of your examples of large groups that are up to no good. I mean, in my opinion Trump voters all made at least one dumb or evil choice, and in some cases are continuing to make correlated dumb or evil choices, but it seems way more fruitful to criticize *that choice* than to bash an enormous group of basically normal people that's diverse enough that any generalization you could make about them will often be false. I think maybe I just do support a norm of "don't bash large groups of basically normal people", even though I think most instances of doing so are not any kind of terrorism
2. even though I'm opposed to political violence, my gut reaction to "well if you say that someone might become a dictator or end the world, obviously people will want to kill them, so you shouldn't say that" is "that which can be destroyed by the truth should be"? even if there's a bunch of people who wrongly believe "if A, then B" I don't like the idea of willfully misleading them about A in order to manipulate them into believing that not-B. instead we should actively make the case for why "if A, then B" is false
Today, I wish to draw the Americans attention to this:
1) You voted for Cornpop. And you thought he was not senile.
2) You voted for the other guy, who has said 37+ times that we are "very close to a deal with Iran." (and we called Clinton a flip-flopper! -- which was entirely inaccurate, he was just prone to making late decisions, and convincing everyone that he was on their side before then).
3) You decided it was unimportant which of these two people would be POTUS.
I never heard of "stochastic terrorism”, but I agree with the liberal solution. Free speech in the US means you can say almost anything. With edge cases decided mostly by the courts. I'm a fan of 'Jackie Jackson' aka Contrapoints. https://www.youtube.com/watch?v=RTRT794IQBg&t=20s Who sometime later in the video says something to the effect of, 'I don't want to say the 'n' word, but it bothers me that I'm not allowed to say it.'
Observationally, there's definitely some groups with toxic dynamics that make their violent outliers predictable. I think it's clearly reasonable to have a norm against that in retrospect - people who are still fans of Luigi mangione, or who publicly supported him before and never distanced themselves, should be judged somewhat.
I'll take it a step further and say that the sort of group dynamics that lead to this, while hard to define with a bright line rule, are also reasonably predictable (although like all non-bright line predictions this is uncertain and subject to human biases), and so long as people are using their honest best judgement to predict this and apply norms based on it that's reasonable and should be taken seriously (with the normal caveats about losing legitimacy if your predictions are consistently wrong or biased).
I'll add to the first one that it's reasonable to demand groups disavow bad behavior they're potentially linked to - for example, I don't think AI safetyists ttied to firebomb Sam Altman (they don't have those toxic dynamics, in my best judgement), but it's still good that they uniformly publicly condemned it and it was reasonable to expect this of them.
Right, I agree. The crack down usually comes against the press which is most radically against the currently governing party. Radical press on the other side is pardoned.
>The “stochastic terrorism” concept is near-unique in how effectively it can be discredited merely by listing many examples of its use together in the same place. Almost no one supports a blanket prohibition on criticizing of all of these different groups of people
Completely disagree; all of these are stochastic terrorism, including the ones on my side that I've contributed to, and none of it should be prohibited speech.
There has to be a middle ground between outlawing speech and pretending it has zero negative impact and is therefore 'legitimate'. Freedom means the freedom to do stupid and even destructive things, and we must acknowledged those harms and costs. If you try to pretend that every thing which people should be free to do is also harmless and legitimate and fine, then anyone can call for a ban on anything by pointing out the harm it does.
We should be able to acknowledge how reality works and be allowed to describe it, including calling out the harm caused by things we don't want to prohibit, and calling for people to voluntarily do better for pro-social reasons. 'Stochastic terrorism' is a reasonable description of a real phenomenon, but it's not a call for censorship or state violence.
I didn't kill Charlie Kirk, but he would probably be alive if the type of targeted left-wing rhetoric I have occasionally engaged in never existed. By my own actions I have killed maybe .0000000000001 Charlie Kirks in expectation, and I accept the moral weight of my actions there.
I accept that as a moral price I am willing to pay for the activity and speech that I find valuable and entertaining.
I bite the bullet on that being an ok thing for people to do in general, and reserve the right to call out people when they're being egregious about it, just as I acknowledge their right to call out me and my 'team'.
I do not need to be absolved of this moral culpability, just so that I can feel more comfortable about my speech. I should never feel comfortable about doing anything for which I have not considered and accepted the potential harm it may cause.
When I opened Scott’s post, I really expected it to be exactly what you’re saying here, or—even better— directly address this excellent take in a clever way.
> By my own actions I have killed maybe .0000000000001 Charlie Kirks in expectation, and I accept the moral weight of my actions there.
The concept of a micromort seems relevant here! :)
I was also surprised this wasn't the takeaway of the post. But I also don't see how 'stochastic terrorism' is used in practice, since I barely touch the big social media sites - this post was the first time I was even made aware of this term. But the way it was described, concur with your take on this.
I do think it's worth emphasising that it typically doesn't make a lot of sense to call people out on single instances of this; the whole point is that it only becomes bad in aggregate. I like your suggestion of when it's worth pointing out: "call out people when they're being egregious about it". Obviously, that's squishy and subjective, but it strikes me as a good rule of thumb, regardless.
What are your thoughts on a specific very powerful person with an enormous amount of political reach, such as Elon Musk, demonizing a particular group of people? (Many have connected a recent spate of anti-refugee arsons to his rhetoric in particular.)
On the old principle of "greater power entails greater obligation", it can be argued that individuals with the ability to reach significantly larger audiences should be held to significantly higher standards. By that measure, Musk's insane maunderings should be subject to much greater control than those of the average rabid YouTuber.
>If your worldview says that it’s acceptable to lynch sufficiently bad people, then yes, accusing people of being bad is equivalent to calling for their murder, and you have no alternative but to make sure nobody is allowed to criticize anyone you like.
The problem is, it doesn't matter what *my* worldview is. The worldview that matters is the person that pulls the trigger.
I don't think you do anything to escape moral culpability by saying 'I will take this action that is likely to cause someone else to kill this person, but **I** don't think they should kill this person, therefore I have no moral culpability in their death.'
If your actions have foreseeable consequences, then you are morally responsible for those consequences. Consequentialism doesn't work if saying 'I do not agree with the consequences of my actions and I would prefer they did not happen' is a get-out-of-jail-free card.
Now, obviously, we should all be working hard to enforce the norm of 'no one should lynch people, even sufficiently bad ones', both through social policing and actual policing, as well as mental health services and etc.
But until we have achieved a 100% success rate on that front, we should also be brave enough to bite the bullet and say 'Yes, some forms of speech increase the likelihood of a lynching happening in expectation, and we place a threshold on that likelihood which we call 'inciting violence' and outlaw it, and below that threshold we accept those deaths-in-expectation as an acceptable cost for free speech'.
You can't improve the world if you don't acknowledge how it works.
"Not lynching bad people" can only be normalized if people believe other mechanisms exist for change. To the extent that a group believes that our institutions are fundamentally broken and won't be effectively reformed, they will resort to means outside those institutions.
Setting aside semantic quibblings about how we define 'lynching', yes, for sure. That's why we also have to have institutional reform and responsive democracy and etc. etc. as part of the project.
>But until we have achieved a 100% success rate on that front, we should also be brave enough to bite the bullet and say 'Yes, some forms of speech increase the likelihood of a lynching happening in expectation, and we place a threshold on that likelihood which we call 'inciting violence' and outlaw it, and below that threshold we accept those deaths-in-expectation as an acceptable cost for free speech'.
I'm afraid I disagree - the thing we call "inciting violence" is specifically about intent, not just about likelihood. And I think that that distincition is really important, and that it's vital that we defend the right to say things that are likely to cause violence but don't incite it.
The reason this is important is that people often react violently to speakers they disagree with - rioting as a form of heckler's veto. And if we make the basis for criminalising speach "likelihood of resulting in violence" instead of "intent to incite violence", we endorse that tactic, enabling any group willing to demonstrate a low threshold for violent rioting to have the state silence their enemies.
Hmm... What about speech which is true, but could be construed as motivating harm? Consider Darializa Avila Chevalier, who recently won the Democratic Party primary in NYC's 13th district. She co-founded Columbia University Apartheid and Divest (CUAD), that calls for death to America and the "total eradication of Western civilization.". Now, I wish her a long and happy life - but preferably in Afghanistan, where she will have all the rights and privileges of a woman under her chosen religion, and need not be troubled by the Western civilization she so despises. Does that count?
According to Perplexity, the phrases “death to America” and “total eradication of Western civilization” are attributed to CUAD or to posts associated with the group, not directly proven as [Chevalier's] own personal statements. I have therefore deleted my earlier post, which was misinformed.
The label seems more appropriate for the rampant spread of ai-generated text … which even unwittingly may subsume all human textual output, especially as we approach the threshold where data has to be invented to satiate the swallowing maw of AI.
So is obviously ridiculously overused , but surely there’s a level at which stochastic terrorism does exist though, no? The classic example being “Will no one rid me of this meddlesome priest?”. Two clear examples is this in modern parlance spring to mind:
First, I’m personally of the opinion that abortion is murder, and that the case for just up and killing abortionists is pretty clear-cut. I’m not personally willing to bear the cost involved, and I happen to think it wouldn’t do much systemic good in a lone gunman scenario, but in an alternate world where I don’t have a family and I can coordinate with, like, 1000 other people who agree with me I’d totally be willing to just off an abortionist. Surely there’s something here that my opponents could point to in the stochastic tourism idea space, no?
Second, what exactly is a fatwa in Islam if not a call for some good Muslim somewhere to off the object? And I don’t know of any specifically formulated fatwa against, e.g. David Wood, but I can find you plenty of good Muslims who sincerely believe his blood is halal. That’s a result of stochastic terrorism, no?
The more general thing I'm noticed is that every non-technical invocation of the word "stochastic" is always just slop and nonsense. Surprisingly reliable!
Let's suppose the year is 1880, and someone who is not a member goes to a KKK meeting and tells them all about how their black neighbor is a rapist. They have no proof or evidence, they just make up a story about it and get everyone riled up. They do not literally advocate for violence, but they went there on purpose and told this story on purpose in this location to these people.
Some of the members, but not the storyteller, go and lynch the accused man.
Clearly something has gone wrong here. Clearly the storyteller has done something wrong above and beyond mere slander and is morally culpable in part for the murder which takes place.
You've correctly made the point here that we can't reasonably outlaw stochastic terrorism because it's too hard to legibly distinguish from regular criticism. But in the past you have made posts demonstrating that you understand the distinction between legality and morality, and the distinction between legal penalties versus social penalties.
Stochastic terrorism is a meaningful concept. It is a thing which people embedded in subcultures that glorify and praise political violence do in order to agitate the violent people and deliberately incite without incriminating themselves because they themselves did not technically call for violence (If persons A,B,C all say "kill Nazis", and person D responds "X is a Nazi" person D hasn't legally advocated murder). But it's morally evil and we should inflict the same outrage and shame upon them that we would on the KKK storyteller from the earlier example.
So one way in which your example differs from almost anything (I won't go as far as "anything", but I can't think of examples off the top of my head) in 2026 America is that it's a situation where words are *likely* to cause violence that would otherwise not happen, but don't necessarily meet the current legal standards of "inciting" it.
Nowadays, lynching is much rarer; it's pretty rare for acts of violence to be directly traceable back to a single act of speech that wasn't intended to cause them or something like them.
And I think that "saying things that have a tiny chance of resulting in violence" or "adding one more straw to the camel" is different morally, and should be different legally, to saying things that are near-certain to cause a lynching.
"This liberal solution isn’t trivial. It requires the separate liberal norm of always being against extra-state violence"
Particularly when you consider that the political movement that appropriated the "liberal" label in the US is at best ambivalent about punching (and shooting) "fascists". Saying that the solution is more "liberalism" doesn't seem to grapple with the actual problem.
Yes, I also found this a bit puzzling, but I guess the problem may be in the ambiguity of the word "liberal". (In European parlance liberal is more normally associated with economic liberalism, tending to be right-wing, but I think in North America some see it as a simple synonym of "left-wing").
Because saying that being against extra-state violence is a left-wing norm doesn't make much sense. I think the only political ideology which specifically advocates for extra-state violence is Marxist-socialism, with a violent class revolution being a basic premise of the whole movement - the George Floyd riots being the best recent example of the justification of extra-state violence for what followers of that ideology considered a just cause.
Conservatism tends to be very strictly against extra-state violence of any kind, though indeed sometimes runs the risk of falling into fascism and unjust state-mandated violence (ICE as a recent example?)
Right, my point is that the compromise-over-conflict kind of liberalism doesn't move hearts and minds in the US these days overmuch. Leftists claim that whatever is worthwhile in "liberal" is synonymous with "leftist", rightists blame liberalism for the ever-increasing unattractiveness of their ideals, and centrists are quietly wringing their hands in aimless despair.
I see others have made this general case, but to me there’s a distinction between criticism, or even harsh criticism, and “Someone out there ought to kill this guy”.
Yes, people cheapen the concept of stochastic terrorism by applying it too broadly, as they cheapen all concepts. But it’s still a real thing.
That song is an overt and unsubtle celebration of Luigi Mangioni style assassinations of CEOs. This is not mere “I don’t like rich people” stuff. This is “you, person listening to this song, should go out and murder a CEO specifically so they will be afraid of us”.
I think that is a special category of bad behaviour and it is useful to have a term to describe it.
I think your argument that we shouldn't be against using emotional and dehumanizing words is weak. In particular, the assertion that "his opinions didn't hinge on slight changes in the emotional resonance of the words being used" might be true but is not the issue in question - his ACTION on those opinions almost certainly WAS influenced by the emotional resonance of the words being used. All the evidence suggests that emotive words bypass the rational part of our minds and that can never be good.
(E.g. using this sort of language is more generally detrimental - Look at the stupidity of Trump calling the Iranians "scum" and then expecting this to somehow not have a negative effect on the war)
I was surprised to see this quote: "It’s vacuous to tell Brian Thompson’s family that they can’t complain about his murder just because health insurance CEOs are a privileged group - especially given that health insurance CEOs now have a higher death-from-political-violence-per-capita rate than Muslims or transgender people."
The rate of domestic violence for trans people is quite high, as is the number of trans people (especially women) who are killed every year because they are seen as sick or wrong. See: Sam Nordquist, Tahiry Broom, Laura Schueler, Brianna Ghey, Brandon Teena, and many more people who were targeted for their identity. When transsexuality, a medical condition, is being treated as a political ideology or mental illness, how can we pretend that these murders could not be motivated by ideology? On a similar note, the 2016 Pulse shooting was openly motivated by a political and extreme religious ideology. I just find this to be a strange and incorrect comment.
There are about 1000 insurance companies in the US according to Google, so the politically-motivated-murder rate is at least 1:1000. There are a little under 3 million trans people per the UCLA Williams Institute estimate that's usually quoted on this topic, and way less than 3000 known-to-be-politically-motivated trans murders (the Wikipedia list has 30). Even the total number of trans murder victims is well below 3000, and most such murders are committed by the sexual partner.
Counting only active insurance company CEOs seems like an arbitrary distinction. Also, the Wikipedia list has a disclaimer that it is not complete, it just covers some notable cases. HRC counts at least 27 cases of fatal violence against trans people from November 2024-2025, but these are considered to be underreported figures because law enforcement or family do not always acknowledge that victims are transgender (because they are a controversial minority class). Murder victims who labelled themselves as trans during their lifetime are sometimes labelled as crossdressers or gay. The killing of a notable figure who is wealthy, such as a CEO, is probably much more likely to be reported to police than, e.g., that of a transgender sex worker.
I don't think that a trans person being killed by a sexual partner can be automatically seen as non-political due to things like the panic defence. Overall if we look at historical data I would feel quite confident that the number of trans people murdered for political reasons, per capita, is higher than that of an insurance company CEO.
Average CEO tenure is something like 5-10 years, so make it maybe 1:4000 if we are being generous. You can certainly imagine there being 25 unreported anti-trans hate crime murders for every reported one... you can imagine a lot of things. You could also imagine there being a lot of unreported political murders of SME CEOs.
I actually can't imagine there being many unreported political murders of SME CEOs, can you? And can you not imagine that it would be more likely for the murder of a transgender sex worker to go unreported than that of a CEO? Not even half of transgender hate crimes in the Bay Area are reported by the individual to the police, per a BMC Public Health study, and that is in one of the most 'trans friendly' areas of the United States. If the individual is not even around to report a hate crime, they are in a less friendly area, their family does not know or would not admit that their relative was trans, and they are in a profession like sex work, then yes, I can see the number of unreported murders of trans people being underreported by a large factor. By 25? I am not an expert in this field, lol, I am an IT guy with some trans friends, so I honestly cannot say. Maybe my root issue with the comment is that it's a weird stat to point to, and something like assault per capita would be more indicative of the issue of stochastic terrorism (e.g. could you say that punching an alleged nazi is part of it, or count those who are simply injured in a Christmas market hit and run? I'd say yes). Do you think that it's an issue though or do you agree with Scott?
With all the activist attention transgender issues are getting, I can't imagine many murders going unreported. I'd think it much more likely that the murder of the CEO of some tiny insurance company with 30 employees you never heard of would go unreported. Or rather, it would be reported somewhere, but the report wouldn't be easily discoverable - Wikipedia doesn't even have a list of murdered CEOs or businessmen or wealthy people, but it obsessively lists every homeless murder victim that has been trans; half of them have their own article.
Maybe there are many trans murders which are reported but the political motive does not get identified - in Business Insider's investigation, about a third of all identified murders were unsolved, and maybe politically motivated murders are much harder to solve (seems unlikely, domestic violence is probably easier to solve but robbery, murder during sex work, serial killings etc have no reason to be), but even then you'd only be in the dozens per year range, with a multi-million base population.
> Do you think that it's an issue though or do you agree with Scott?
I think it's a fairly straightforward choice of stats for the point he is making in that part of the post (that powerful people are no less at risk than oppressed people from the kind of attack that gets described as stochastic terrorism).
Eh, I think your argument rests on the idea that murdered trans people are correctly identified as trans and that their murder actually gets reported, which is not the case, as I have tried to lay out in earlier comments. Even the crimes experienced by the average homeless person go far underreported; if there is no report or media coverage there cannot even be a Wikipedia article, lol. Anyways, wish you well and it was an interesting conversation.
"I don't think that a trans person being killed by a sexual partner can be automatically seen as non-political due to things like the panic defence."
Which is not what happens in a lot of these murders, there is no trans panic defence. There was one where it was "trans woman killed by ex who was also a trans woman" so I don't know how you can get "trans panic defence" out of that. There were a lot which were "family knew they were in a relationship, everyone knew partner was trans, relationship broke up and ex got violent over that" including killing/injuring members of the trans person's family during the incident. Ironically, that is domestic violence from ex-partner of the same kind as happens to cis women.
I think there was one episode of "trans panic defence" in the lists of "all the deaths of trans people over this year" which lists are heavily implying "these deaths are due to transphobia" and they're not.
I am that kind of OCD/autistic bitch who went through the lists and looked up the cases in news reports online to get any background that was more than "DeShauna Shaniqua Duxelle, brave and stunning trans woman, killed in Chicago in June, rest in power" kind of naming in those lists.
Trans murder rate is largely disconnected from trans status; being targeted for that alone is quite rare- note "rare" does not mean "unheard of".
Total per-capita murder rate of trans women is lower than that of cis women, with some variation by race; it's much more correlated with prostitution and drug use than trans status *alone*.
>2016 Pulse shooting was openly motivated by a political and extreme religious ideology
Supposedly he was heading towards Disney and decided they had too much security, and didn't know Pulse was a gay club until after. Yes, he was motivated by an extremist ideology, but not quite the way you seem to mean.
As I have tried to show in comments above, trans murders and hate crimes are far underreported due to lack of interest, familial shame or mistrust of the police. In the Bay Area alone, over half of trans hate crimes go unreported, and that is a very pro-trans area.
WRT Pulse, I did try to find what you were talking about, but seems that the shooter was targeting nightclubs generally (according to his cellphone data) and that he was homophobic (according to his family). So may not be that he initially chose a gay nightclub but the club being gay may certainly have been a contributing factor.
Oh boy. This is a fraught issue, as I've gone through a couple of those lists of "trans people being killed for being trans" that are published online by trans activist groups, and they're... not completely accurate.
A lot of them are "killed by ex/in domestic violence" just like cis women. The partners/ex-partners knew they were trans and had no problems being in a relationship with them, it was when the relationship broke down that the violence happened.
Some of them were, literally, "hit by car" and that's treated as "death of a trans person is the same as killed for being trans" because of social transphobia or something (I wish I were kidding, but no.)
Some of them were "killed while engaging in sex work" which is *already* a dangerous line of work even if you're a cis woman.
There were about three murders I could call "explicitly killed for being trans" out of those two or three lists.
I think we can and should have a social norm against vitriolic criticism, but we should not label such criticism “terrorism” or “murder,” which are crimes that have actual legal meaning, not to mention that we would ourselves be being vitriolic.
We should also have a strong social norm against compiling lists of people who are not already famous whom we believe to be evil, especially when such lists include their home address, workplace, etc., or compiling lists of schools, places of worship, summer camps, etc., associated with an ideology we disapprove of.
The assassination of Jicchak Rabin in 1995 was arguably quite close to stochastic terrorism, and I would say that it changed the path of the Near East for the worse.
"Bilateral controlled tiger-dismounting" is a phrase I'm soooooo going to use.
Also, I'm not sure Stalin's ascent to power could be described as "attempting to end democracy". It's true that he concentrated power that was previously much more distributed, but I think this was more like outmaneuvering his political opponents in the communist party rather than destroying mechanisms for expressing the will of the people or anything. There were no such mechanisms to begin with.
"I admitted that there should be a discussion norm against some especially vitriolic criticism; a line beyond which I would think less of a person, even if I didn’t want their speech banned. But most stochastic-terrorism-sayers don’t propose to ban the speech involved, just to create a social norm against it. And “think less of X” and “create a social norm against X” are effectively interchangeable, since social norms are more-or-less defined as things that people think less of you for violating. So am I conceding the whole argument there?"
I don't think this is conceding the whole argument.
There's a big difference between:
(i) I will think less of you for making too extreme a criticism because it seems excessive and suggests poorly calibrated judgement and poor emotional control
(ii) I will think less of you for making too extreme a criticism because I judge it to promote violence and I will judge you to be a wilful violence-promoter.
A nitpick, but I find it funny that there's a part where "Stalin, Mao usurps democracy because no one realize their intention early.". For Mao, I guess everyone knows he's not on the side of democracy and he's defeated multiple times (violently) until he wins. For Stalin, the government he's usurping is not a democracy anyway.
"This puts you in the position that Winston Churchill called “riding a tiger from which you dare not dismount”; you had better invest all your energy into making extremely sure that you and your friends are the ones calling the shots about who can and can’t be criticized."
I accept the argument that you shouldn't condone violence in the abstract, even if it serves your purposes in the moment.
But I never really understood the argument that its in the self interest of any given political movement to avoid creating a norm of political violence, because I just don't think norms affect how how much political violence there is that much. It's not like the violence is being carried out by public figures with who need to maintain a reputation for respecting norms.
A lot of political movements could probably achieve their goals more effectively if they employed some degree of violence.
For example, say vegans could reduce the total quantity of animal products society consumed if the vegan movement started a sabotage campaign against factory farming's infrastructure, or even an assassination campaign against owners and ceos. The risk of inciting reverse violence against vegans would be a tiny cost in the overall cost benefit calculation.
In general any movement that's supported by a large number of people and that opposes a small number of people, or any movement that cares a lot more about achieving it's goals than its opponents care about preventing them, or any movement that has goals that can be achieved without actually needing to pass laws, can probably be confident that the benefits of violence outweigh the risk of changing norms.
There have been plenty of movements that used political violence as a tool (early 1900s anarchists, Western communists during the Cold War, various islamist groups, various separatist groups). It usually wasn't effective. Some separatist causes succeeded, but they had majority support, so hard to say how much the violence helped (or hindered).
The Liberal solution ("it’s always legitimate to criticize any person or group however strongly, and never legitimate to use extra-state violence against them") is the best answer to a not all that inherently difficult moral dilemma, but with two very important reservations.
One is that criticism can be dismissed (when it's mild) or censured (when it's violent) if it is based on clearly false facts. For instance, criticism calling for violence against Jews on the strength of the 'blood libel' or of accusations contained in the forged Protocols of the Elders of Zion, both patently factually incorrect, cannot lay claim to immunity and can be banned outright. Similarly, and just as laughably, attacks on Haitian refugees in the United States for allegedly eating people's pets can be dismissed after they have been investigated and found to be pure inventions.
The second objection is the blanket prohibition against extra-state violence under all and any circumstances; taken to extremes, that principal would render any revolutionary or insurrectionary activity illegal. Really? The Warsaw Rising against the Nazi occupiers was illegitimate? The Greek, Serbian and Bulgarian risings against the Ottoman empire were illegitimate? The French and Bolshevik revolutions, never mind that both ended up sliding into dictatorship, were historical errors? And if the lynching of the Stalinist secret police during the (failed) Hungarian revolution of 1956 was an illegitimate act of extra-state violence, will that precedent weigh in the balance if at some time in the future an American mob decides to turn against rampaging ICE agents?
I agree it is very hard to saddle a principle such as Scott's Libera Solution (which I 99.9% agree with) with riders along the lines of "... but there may at certain times be legitimate exceptions to the rule"; such yes-but reservations tend to sully the theoretical purity of sweeping moral rules. But they need to be kept in mind.
I’m unsure of your intention. On the one hand, you have two reservations to the principle that extra-state violence is never legitimate - but your first reservation is that in some cases it is legitimate to respond to criticism by “dismissing” or “censuring” it. I don’t think dismissing or censuring is violence! Just as it’s always legitimate to criticize, it’s always legitimate to direct one’s own criticism against the criticizers.
Minor nitpick first: “Trump and Altman are both still alive” is a strange way to dismiss this category. Both have been targets of assassination attempts. Failed attempts still count as political violence; their survival may say more about their security than about whether the rhetoric inspires violence.
More substantively, I think your fourth proposed principle is close. It just needs to be generalized:
Some accusations don’t merely describe a target. They place the target in a category whose accepted remedies include coercion or violence, and therefore also function as instructions about how the target should be treated.
If you say Trump will enact concentration camps, you’re calling for violence to stop him. You’re implying that Congress and the courts won’t stop him, while the military, federal agencies, police, and contractors will play ball. If the accusation is true, what option remains but violence?
If you say Sam Altman will destroy the world, you’re saying he should be treated like a movie villain holding a doomsday device. Assassination, kidnapping, and death by airstrike all beat the end of the world.
This principle also distinguishes between different kinds of harsh rhetoric:
“Examplestanis are ugly and stupid” does not imply violence.
“Examplestanis are vermin” does. Vermin get exterminated.
This strikes me as a very weak post - regardless of whether you think laws like these are *good* ideas (I don't), they have been implemented in a lot of countries. To not look at how this was done seems downright lazy.
Example: Sweden. First off, there's a law about Incitement ("Uppvigling") where you actively call for violence against someone or some group. The illegality here isn't particularly controversial. Then there's the hate-speech law (where "Hets" *also* basically means incitement, but in a less direct manner). Here, the law starts by identifying certain protected groups (ethnic/national/racial groups, religious groups, and groups of people by sexual preference and transgender identity). Outside of targetting such groups, there is no hate speech, so Trump or Altman or even merely political groups are not protected this way.
Next, there has to be a public dimension for attacks on these groups to be considered. If you do it in your own living room, it's very likely legal.
What makes a statement illegal isn't how harsh it is exactly. Swedish has a word "saklighet" that literally means "thinginess" or "to the thing", or perhaps best, "matter-of-factness", but can be best described as a combination of objectivity, impartiality, being reasoned, and having truthful fundamentals. For instance, I'm absolutely free (legally) to say that Israel is committing genocide in Gaza and that all Jews globally who don't criticize this are indirect participants, or that the wide-ranging support for 7/10 among Palestinians demonstrate that they would do the same to the entire Jewish population of Israel if given half a chance. When it becomes illegal is if I describe a group as a cancerous growth on the national body, or that all Examplestanis are ugly stupid vermin, or that they are Satanic cockroaches unworthy of life. Holocaust denial is also included here, although it's unclear if you could get away with a sufficiently calm and reasoned kind (you probably could, if such a thing existed). It's also possible to be convicted for indirect hate speech, such as wearing a swastika, where it's seen to be a symbol of a certain kind of speech. As a rule, it's easier to get convicted for overgeneralizations ("immigrants are massively overrepresented in the rape statistics" is fine, "immigrants are all rapists" isn't).
This is a *coherent* law at least. Obviously it's not exactly defined and courts have to make interpretations, but when is *that* not true?
(There is also a separate hate *crime* law, which isn't a separate rubric, but instead where you target people of a protected group with violence for reason of them belonging to the group, which raises punishments a bit.)
Whether "they're eating the cats, they're eating the dogs" would be illegal would probably depend on whether you can *present* some such cases. If some of them *are* eating the cats and the dogs, then saying it is likely "sakligt", especially if you don't overstate or generalize.
Also, perhaps weirdly to an American, Sweden has considerably stronger Freedom of the Press laws than Freedom of Speech laws - as a rule, the U.S. has stronger Freedom of Speech laws but weaker Freedom of the Press ones. This could lead to surprising outcomes such as it being legal to translate and publish Mein Kampf, but then hate-speech to read certain passages from it at meeting, especially a meeting that is already antisemitic.
That just sounds like "Freedom of speech, but really only for the Officially Approved Speakers to Proles". As an American, I will *so* pass on that one.
An example of the stronger Freedom of the Press laws - it's actually *illegal* to try to find out who an anonymous source is. Things like the U.S. tossing journalists in jail for refusing to divulge their sources is an impossibility here. This is a Good Thing.
I sometimes feel like the last 20 years have mostly been an exercise in finding ways to say "people who disagree with me shouldn't be allowed to talk".
The main problem I have here with the second-order solution of eschewing all specifically *non-state* violence, is that the majority of the problem people have with vitriolic speech against individuals, groups or ideas they support is going to carry over directly into a worry about state violence. You and I may not think rapists should be summarily beaten to death, but we think all sorts of negative legal consequences should happen to them. We shouldn't want similar legal consequences to happen to redheads because enough speech has convinced enough voters and jurors that redheads are as bad as rapists.
A deontic anti-violence rule with a "non-state" caveat doesn't free us from the enormous responsibility of having the first-level assessments of others' badness in our speech be reasonable.
Thanks for this, Scott. I have been seeing "stochastic terrorism" thrown around on social media and not been too sure what was supposed to be going on; what was the difference between this and normal terrorism?
And the difference seems to be that the term has now become watered-down into a general term of abuse like "racist, fascist, Nazi" etc. In other words, I say you are a stochastic terrorist/engaging in stochastic terrorism and all I mean by that is "You say things I don't agree with/you don't agree with what I believe" topped off, often, with "You should not be allowed say or think that because I am a Special Instance Of Protected Group so less than 100% agreement in full with all I say or want is violence and stochastic terrorism, and if some crazed zealot murders me in the morning it is all your fault".
(Never mind that them being murdered in the morning by a crazed zealot is as likely as me winning the Euromillions Lottery since I never buy a ticket. They like thinking of themselves as persecuted and oppressed for being so special and daring and challenging, so of course they *must* be at risk of being murdered in the morning by a crazed zealot since they are challenging the status quo).
More broadly, speech ethics have become much more complicated in recent years. For example, assuming the release of television programs that talk about suicide in a certain way (like "13 Reasons Why") will actually probably cause an increase in teen suicide, should we allow their release? Is it refraining from any censorship an important Schelling line that shouldn't be crossed? Would it just be counterproductive to intervene? And if we should not place restrictions, is creating such a series an ethical decision to make personally, or should one refrain in the name of "The Virtue of Silence"?
As an individual it is always your responsibility to weigh the likely harms and benefits of your speech and that certainly includes risks of encouraging violence. It's just that trying to enforce any kind of social norms that tries to prevent such "stochastic terrorism" creates a rule that is open to abuse and results in more net harm.
I think a good analogy here is to the morality of breaking up with a partner. There are plenty of times it would be wrong to break up with a romantic partner but we've realized that attempts by society to punish that kind of behavior can be exploited by abusive partners behind closed doors or simply make people miserable.
I think the term “stochastic terrorism” is a term made up by stochastic terrorists projecting their behavior onto others, so naturally this isn’t how it works. Stochastic terrorism isn’t strong criticism or even vague threats.
Stochastic terrorism two guys walking down a street. First guy shouts “John Doe is a nazi!” and the second guy shouting “Shoot nazis!”
Apply to political discourse where progressives claim voting for some bill will cause death, then others say death must be avenged, etc.
Ferdinand Nahimana, director of the Rwandan radio station RTLM, was sentenced for genocide, even though he did not kill a single person, or even directly called for the death of anyone. But the broadcasts he oversaw regularly dehumanized the Tutsi ethnic group, characterized them as mortal threats, and talked about the need to exterminate them. Some smart natural experiment based analysis (that made use of the fact that in the mountainous areas of Rwanda even in the same region radio reception was highly variable) attributed about 50.000 extra deaths to the broadcasts.
That's more stochastic genocide than stochastic terrorism, but the same underlying issue - the perpetrators didn't personally participate in violence, they didn't coordinate or direct violence, nor were they the main cause of the violence (the killings were started and mostly done by government forces). In fact, the US army considered radio jamming but decided against it on freedom-of-speech grounds.
Would you disagree with the ruling? If not, where would you draw the line?
Nahimana wasn't convicted for personally calling for anyone's death, but for failing to stop it: he ran RTLM, had the authority to stop his journalists' broadcasts once they moved to naming names and "go to work" after April 6, and didn't.
RTLM both promoted dehumanization of Tutsis before the genocide (cockroaches, snakes) pre-1994 *and* then in 1994 use quite explicit language to get already-radicalized Hutus to go out and kill. The latter seems to have had a small but measurable effect on actual killings.
I would model Scott as saying the former is in poor taste but not morally responsible for the killings, and the judges in Nahimana's case agree since they threw out this part entirely.
For the record, contra Scott I think even the pre-1994 RTLM had blood on its hands, but the line seems hard to draw.
I don't really participate in the culture wars, so I'm probably underestimating how much the term gets thrown around for all kinds of criticism, but I've personally only encountered it in two fairly narrow contexts:
- Fairly obvious dogwhistles ("It would be a real shame if someone torched the place. Of course I would never do that, but you're just asking for someone... anyone... to, you know?"), where the threshold gets lower the more influential the speaker is ("rid me of this meddlesome priest").
- Bubbles or echo chambers of ever-escalating rhetoric and dehumanization, where nobody truly calls for violence, let alone wants to take responsibility for violence - but would still treat the occurrence as unforeseeable ("just some lone wolf crazy person") yet satisfyingly karmic ("what do you expect if you take people's livelihood away?").
The former is an individual moral (or legal) judgment of a particular person's statements in context.
The latter suggests a general standard not to normalize talk of killing, execution, extermination, or comparisons to vermin. Honestly, if that's where the discussion is heading, find a better circle or cause.
Neither has ever felt to me like a blanket restriction on criticism.
In a Gricean sense, certain prescriptive claims can be implied by formally descriptive utterances, given commonly held assumptions.
"Smoking kills." (Do not smoke.)
"There are leftovers in the fridge." (Eat them.)
"You have ingested a deadly poison, and this vial over here contains the antidote." (Drink the contents of the vial.)
For that last example, it would pretty absurd for the speaker to claim "I never told them to drink what's in the vial." In a different world where people don't care if they live or die, perhaps, but not in ours.
If someone provides me with convincing evidence that a) X is literally Hitler, b) nonviolent means to stop him won't work, and c) violent means might work, then I will be persuaded to kill X, or to support someone doing so. This will be true even if know the person saying this is an extreme deontological pacifist who thinks violence is immoral in literally every case.
So yes, I would say that arguments that would persuade a reasonable person to support violence qualify as incitement. The response to this would be that most of the attackers on your list probably don't qualify as reasonable people.
Also: I think many accusations of stochastic terrorism are just accusations of defamation. It's not "don't hold that horrible opinion!", but "stop lying!". If someone murders someone based on slanderous information, it feels right to hold the slanderer responsible, even if the murder would have still been wrong if the information wasn't slanderous.
Great post. Trying to think of useful examples to bound this. In 2015 after the Republican debate Dan Scavino and Steve Bannon relentlessly attacked Kelly personally and Lewandowski tweeted "gut her" and it led to a year of viscerally negative Breitbart coverage. That led to death threats that drove her to hire a personal security detail.
My instincts strongly align with classical liberalism and a sort of them's the breaks approach here, but I raise it because it's the toughest case I've got, bit of godfather agitprop -- "it would be a shame if something should happen to her."
I can't think of any criminal rule to limit this that wouldn't be misapplied though. Possibly existing civil litigation for defamation or wrongful publicity could cover some of this territory without being overbroad.
I guess you are basing your argument on the idea that every murdered trans person is labelled as trans and actually has their murder reported, but the issue is more complicated than that, as I tried to lay out in my earlier comments. I mean, even violence against the average homeless person is significantly underreported. Anyways, wish you well, it was an interesting conversation
I mostly agree, but Scott didn't mention the biggest real world example: Islam. Imams can say it would be great if a certain person is killed, and then there's a serious chance that that person will be killed; the imam himself doesn't have to be directly involved in the killing at all. This is also how Mohammed himself had people killed. He would suggest to his followers that so and so ought to be murdered, some young follower would kill that person hoping to get in Mohammed's good graces, and then Mohammed would distance himself from the killer, saying he had no role in the murder. Organized crime bosses do something similar.
There's a real gray area here. If you can't punish someone for obliquely suggesting that so and so ought to be killed, then leaders of large violent groups, like organized crime bosses and imams, can pretty much have people killed with impunity.
When we’re resentful of someone, we crave a phrase that will utterly hollow them out and leave us ascendant on our podium, Self Triumphant. And who isn’t resentful a lot of the time? Being fair-minded is an awful lot of work & I admire Scott’s consistent effort to manage it.
This line of logic is similar to two hot topics regarding university culture in America
- Hazing in fraternities and sororities
- toxic culture or sexual assault issues in college sports
In both cases, universities trying to police their internal organizations' cultures (maybe somewhat with external influence, but often not a full-blown civil or criminal trial)
They have to distinguish between all the power dynamics and mantras and imagery and culture that "creates an unhealthy atmosphere" versus all the fact-finding required to determine who are the specific bad actors, what did they do.
So sometimes they find the bad apples, sometimes leadership is removed regardless of direct culpability, and sometimes frats are completely kicked out. It's a messy system of governance to adjudicate
The UnitedHealth policies that secretly paid bonuses to nursing homes to override hospital transfers and gave nursing homes quotas for hospital transfers are pretty bad.
I think that there is indeed a concept of each hostile individual non-direct-incitement statement adding some microunit of probability to a terrorist attack happening. We *could* call this stochastic terrorism as a general concept, and it could serve as a general reminder to avoid hostile speech if we want to have a society with less terrorism. However, calling any particular individual as someone who is guilty of stochastic terrorism is unsustainable (if they are directly inciting violence, that's direct incitement, we don't need a new term for this).
"If your worldview says that it’s acceptable to lynch sufficiently bad people, then yes, accusing people of being bad is equivalent to calling for their murder, and you have no alternative but to make sure nobody is allowed to criticize anyone you like."
This gets at the root of the issue which I think is going unstated: it is a purposefully fostered expectation in America that you, citizen x, have a right to live under the hegemony of people like you and should expect to enforce a rule that ensures your values are reflected back at you from the majority of the population. This is a completely destabilizing and poisonous belief and yet no one gainsays it. No one argues against it in its own terms. They only accuse some value systems of being nasty. When you have ambivalent tension between what a man can do by right with his life and this expectation that others are going to guarantee that you feel like part of the normal majority, free speech can't last. There's too much factionalization
I think many people are being hypocrites over this (I think this was also the subtext of Scott’s piece).
You could take a consistent US First Amendment style approach that as long as it’s not incitement to imminent lawless action it’s allowed; or consistently be in favour of UK-style taking a more expansive view of what constitutes criminal incitement.
But it’s just hypocrisy to be only in favour of free speech for your own side, and not your opponents.
"Stochastic terrorism" has been used against the right by the left for ages, usually for far milder examples, and I don't recall any post about it from you. But now that the right is trying to hold the left up to their own standards... suddenly it's something to oppose. This is all who whom.
>Part of how Pol Pot, Stalin, Mao, and the like took power was that not enough people saw the danger early and spoke out about it.
As people on DSL pointed out, this is basically false. Those individuals all engaged in violence before taking power; the problem is that the state was too weak to stop them (or in the case of Stalin, he was the successor to someone else who took power that way).
Scott has gotten hugely more left-wing since then. Scott-2014 probably isn't unfairly biased. This is Scott-2026. And even then, I don't think he's ever spoken against stochastic terrorism when it was the left flinging the accusations.
1. Where did this "meddlesome" thing come from?! Thomas à Beckett was turbulent, not meddlesome! (I like to imagine that he was simply less aerodynamic than the other priests, who all shaved their legs and dressed in lycra like triathletes)
2. "You should be allowed to criticize powerful people, like Donald Trump or Sam Altman, but not weak people, like Muslims": In his Discworld Series, Sir Terry Pratchett imagined an Assassins' Guild who are legally allowed to 'inhume' rich, powerful people if they become too unpopular, but aren't allowed to harm ordinary citizens. Sir Terry imagines that ordinary people can't really do much harm and so legalising their assassination would do more harm than good but rich, powerful people can, if left unchecked, do so much harm that in practice there was no actual choice between 'some people get killed by assassins' and 'nobody gets killed', only a choice between 'some people get killed by assassins' and 'lots of people get killed in wars' (presumably a modern reimagining of the Discworld could replace 'wars' here with 'dismantling aid programs', 'medical insurance industry practices', etc.). Does this just reduce to the trolley problem?
3. "Some European countries distinguish between permitted normal speech and hate speech (for example, “all Examplestanis are ugly stupid vermin”), on the grounds that the latter actually does encourage violence if left alone for too long. As a red-blooded American, I don’t agree with this distinction": This sort of European 'solution' would be like trying to prevent school shootings by just not letting people have guns. Sure, it *appears* to work almost-perfectly everywhere in the civilised world, but Americans righteously oppose it with every fibre of their beings. Thankfully, it's virtually impossible for Europeans to argue for this sort of 'solution' because they just can't overcome rock-solid arguments like "I'm an American".
Are the arguments in (2) Pratchett's or yours? I only remember that killing "simple people" lacks style and, more importantly, Nil Mortifi Sine Lucre...
Also, that reminds me - Pratchett has a line directly about this kind of thing, from Thud:
Community leaders, Vimes thought. There'd been a lot of talk about community leaders lately, as in `community leaders appealed for calm, a phrase the Times used so often that the printers probably left it set in type. Vimes wondered who they were, and how they were appointed and, sometimes, if `appealing for calm' meant winking and saying `Do not use those shiny new battle-axes in that cupboard over there ... No, not that one, the other one.'
Wow, great excerpt! Thanks for sharing! It definitely captures a real phenomenon, too - but I am not sure how accurate it is for social media discourse? Does that sort of 'plausibly deniable', covert incitement to violence really exist on social media, or in our more efficient modern times have we just replaced them with uncomplicated overt incitements to violence?
I honestly don't know. And at least the stronger version of the stochastic terrorism concept seems to claim that some statements can lead to an increase in violence even if the author did not intend them to...
Sir Terry's, as I said. Specifically, "the only alternative to assassination is war" is in Pyramids, and "Assassins are only allowed to inhume rich and powerful people" is in Men at Arms.
Please don't think either Sir Terry or I are praising the assassins, though! Yes, a reasonable argument for their existence is offered - but they're not presented sympathetically in the novels; Sir Terry treats them with the same cynicism he treats everybody else, protagonist and antagonist alike (cf. 'Nil Mortifi Sine Lucre'...)
Do you think it's acceptable to propose specific acts of violence against certain people or groups? Would you call this stochastic terrorism or something else?
I don’t have a strong opinion on stochastic terrorism one way or another, but I do feel like the post is largely fighting a strawman. For starters, the definition Scott gives for stochastic terrorism differs in an important way from what eg the Wikipedia article he links to says:
"Stochastic terrorism is an analytic description used in scholarship and counterterrorism to describe a mass-mediated process in which hostile public rhetoric, repeated and amplified across communication platforms, elevates the statistical risk of ideologically motivated violence by unknown individuals, even without direct coordination or explicit orders."
Or, in the form it has been popularized:
"Stochastic terrorism is the use of mass communications to stir up random lone wolves to carry out violent or terrorist acts that are statistically predictable but individually unpredictable."
Instead, all through the post, Scott seems to be talking about - for want of a better term - “deterministic terrorism.” I think that taking into account that factor makes some of the arguments, like “Trump and Altman are both still alive,” quite weak, and fails to do justice to the idea.
Perhaps as importantly, I feel it is a weird move to detail how various distinctions between allegedly OK speech and allegedly “terrorist” speech break down, at least under certain conditions; and then take what is perhaps closest to the strongman argument for the “stochastic terrorism” concept, the idea that there is speech that constitutes incitement to violence, and just wave it away by claiming “The boundary here is fuzzy, but I think US courts do a good job of navigating it.“ Loads of people have said “Trump is a bad person and I suggest you kill him,” but Trump is still alive! And if US courts have found a good way to draw a line on the continuum between “Go kill X, here’s a gun,” the “will no one rid me of this meddlesome priest” example that came up a lot in the comments, “I wish X was dead” etc., I’d be interested in the details.
More constructively, I think the “stochastic terrorism” idea actually consists of two claims: that some speech leads to an increase in violence (statistically speaking!); and that such speech is bad (and should be avoided, ostracized or banned). And the first part seems to be a factual claim, one that could be checked. Claude says that "some evidence, contested, more suggestive than robust" is the accurate characterization — not "confirmed by statistics" and not "debunked" either. But as far as I see, Scott has not even asked the question, at least in the post.
If the factual claim turns out to be wrong, we are done. If it turns out to be correct (or if we don't know, or if we are interested in the philosophical/moral aspects of stochastic terrorism), I think the question of where that line between “true incitement to violence” and “simple criticism” lies deserves much more attention.
Also, I think the "liberal solution" is not a solution to the problem postulated by "stochastic terrorism." It's like saying "my solution to school shootings is that everyone can have whatever guns they like, but nobody is allowed to shoot anyone." It might be a legitimate stance, but it is not, in any meaningful way, a solution to the original problem.
I think the liberal could always just bite the bullet and say "yes, some speech will lead to lethal violence, but the alternative is to not have a market of ideas, which is worse".
The fundamental problem here is that western societies are increasingly filled with a diversity of peoples and religious ideologies whose interests are perceived to diverge at an existential level for reasons that likely cannot be realistically resolved through rational public debate. The violence in the public square is a side-effect. (I of course include left-wing-progressivism as a kind of religious ideology.) This dramatically affects the cost-benefit analysis for the market of ideas vs. violence suppression.
I would also say a pertinent criticism of "the liberal solution" is that it presumes the government is a neutral arbiter and can therefore be trusted with a monopoly on force, when it's increasingly obvious that the governments of the western world are not "neutral" in any meaningful sense. They are battlegrounds where the struggle for power between peoples and religious ideologies are being fought, and where the left is generally much more entrenched.
That's especially true since we can dispute who is and is not weak. I would say Muslims are anything but weak, having vast numbers and the ability to kill just about anyone for blasphemy.
Stochastic terrorism is the one of the early stages in the life cycle of the state. The United States is built on incendiary rhetoric whipping up crowds to spontaneously attack British loyalists. How else do you know that there's enough anti-British sentiment among people willing to do violence to start organizing militias?
It's mostly epistemic; how your political movement knows what it itself is and whether it's ready to transition to its next phase.
When it comes to research, do you think researchers should make the rigor/getting-information-out-there tradeoff differently depending on the study's chances of contributing to violence?
For example, suppose there's a study suggesting that managers in brown suits are more trusted than managers in black suits, and there's another study suggesting that immigrants from X country are creating Y negative effects. Should the author of the immigration study collect more data or invest more time into their analysis before publishing something like 'Immigrants from X country may be creating Y negative effects'? (Compared to the managers study author publishing 'Managers in brown suits may be more trusted than managers in black suits')
"This liberal solution isn’t trivial. It requires the separate liberal norm of always being against extra-state violence"
unless, of course, said extra-state violence comes in the form of a CEO committing negligent homicide.
the US de facto legalized white collar crime, permitting even homicide, provided it juiced shareholder value, and you want us to feel bad about vigilantism?
39% favorability for Mangione is not nearly high enough. Brian Thompson was never going to see a long prison sentence or suffer more than some token fines, no matter how many grieving families were left in his wake.
Don Blankenship killed 29 people and walks around a free man after serving a year in prison.
“if a violent act occurs, it’s 100% the fault of the perpetrator, and 0% the fault of other people who previously criticized the victim.”
I agree that it’s an exercise in futility to hold people responsible for all the actions, whether positive or negative, that may have been influenced by their behaviour or speech one way or another. Not to mention that whoever is the final judge on such questions would hold an arguably totalitarian level of power.
That said, I think the “100% … 0%” framing risks undervaluing the role that language plays in shaping a culture which either rewards, disincentivises or trivialises various behaviours, ranging from just being hostile or unpleasant online to acts of violence or outright terrorism.
Reality is rarely so black and white, but if, for example, someone actively dehumanises a group of people, spreads malicious falsehoods about them and portrays them as an existential threat, they would still be regarded as having contributed to a climate in which violence against that group became more likely, even if they never explicitly incite violence themselves. Obviously in real life questions of how and to what extent someone is responsible for violence that follows from their speech are much murkier and more controversial, but that doesn’t mean the answer becomes zero responsibility whatsoever either.
Still, I’m fully agreed with your pushback against the use of “stochastic terrorism” as an accusation, and as you say, there are clearly some difficult legal questions to be navigated here (the intricacies of which I won’t pretend to understand) about when critical or hostile language crosses a threshold.
I mostly agree wit the liberal stand being described here (it's simply traditional US freedom of speech norms) but I am interested in opinions regarding the following scenario:
Let us say that I am a member of demographic category X. Person A proclaims on mass media that people in my category are terrible, lying about ourselves, and are in some way harming society. Person A offers no evidence, but plenty of hostile rhetoric.
Afterward, I notice a significant increase in the number of people confronting me and people in my category, threatening us, and making reference to person A's message. When people in my category point this out on mass media, person A continues proclaiming their rhetoric.
What degree of moral liability does person A bear for the increasingly threatening behavior being directed at people in category X? What if someone gets hurt?
My take is Scott is mixing up moral and legal culpability in his post. We should have norms against all the things he wants to have norms against. Don’t call people Hitler barring extreme cases, don’t promote lynching, etc. But legally, the US system is actually pretty good! We don’t want the government getting into the weeds of deciding who can’t criticize who and in what ways. A lot of the early “cancel culture” criticisms were also about hecklers vetoes and the legal right to have a discussion without being physically disrupted (also a legal matter).
The biggest problem we have is that pretty much everyone in public discourse (or maybe just social media) conflates these two things all the time, and often lumps groups of people together. Person A says “that group are cockroaches”. Person B and C say “get them! They’re at 123 Fake street”. The law says person B and C crossed the line (correctly). On social media we are 100% in the right to criticize person A too for what they did.
Now person D and E say “that was bad” and person F says “Person A should go to jail”. Then everyone is arguing about what person F said instead of agreeing that D and E are correct. I don’t know the cause of this exactly. Possibilities that spring immediately to mind are
1. This is just “social media amplifies extreme views” problem.
2. People like to argue about technicalities and forget to say “yes of course” to straightforward statements.
3. The Bad People (insert your favorite group) are manipulating social media to sow discord.
4. People on social media have difficulty distinguishing people on Groups, so A gets lumped with B and C, and (D,E,F) forms another group in people’s minds.
I would say in your version person A is morally culpable for inciting but not legally culpable. The degree depends on the degree of lying and incorrectness.
Sometimes I feel like rationalists want to be able to make rules like this without saying “it’s ok to say (morally) correct things and not ok to say (morally) incorrect things”. But to everyone else I think that’s obvious.
I think it's more about accountability. A norm with no mechanism of enforcement (even a casual one) cannot protect anybody. Given that behavior A is "bad" (however you want to define that), how do we, as a culture and society, best ensure that the rate of behavior A remains low?
Excellent essay, but of course I have to only mention the thing that annoys me. I followed the link about violence against abortion and there actually aren’t a lot of them and they were a long time ago 1982 and 1997 as I expected
I think if the speech was hate, and not legitimate sincere criticism, then the speaker should be considered partly responsible for the incited violence. Otherwise, it is entirely the violent person’s fault.
So saying “insurance companies are evil monsters and every leader should be jailed” is a contributing cause, but “insurance companies have systematically made it impossible to get good healthcare” is okay.
> This liberal solution isn’t trivial. It requires the separate liberal norm of always being against extra-state violence - a norm which is currently less than entirely secure.
(a) This norm has never been held anywhere.
(b) It wouldn't be a good norm if it was held somewhere.
"The boundary here is fuzzy, but I think US courts do a good job of navigating it.", maybe with true threats but that's about it. They are notoriously hostile to true any other speech from true solicitation to true conspiracy to true fraud to true obscenity. You can get twenty for making a joke to an undercover.
Given the hundreds of thousands in US prisons for speech crimes, a significant number who are actually factually innocent, it blows my mind at your confidence in the US courts here.
“So am I conceding the whole argument there?” No. I actually think that one should criticize the use of the word “stochastic terrorism” on the grounds that it blurs the line between the legal world where we enforce bans on behaviors and the moral one where we admonish, chastise, and socially penalize. Both political sides blur this line all the time and it takes significant work to maintain. It is of course worth maintaining as you pointed out regarding riding the tiger.
I think you’re pulling a mot and bailey yourself although unintentionally. You jump from arguing against the idea that there should be blanket prohibition on certain types of speech, such as criticism or overly harsh criticism to arguing that it is 0% your fault. If your speech is predictably going to cause problems or at least has some expected probability of causing problems. There are several intermediate positions that you never discuss like for example, comparing expected cost and benefit and only being against cases where the expected benefits are not worth it. Similarly, it seems intuitively likely that more vitriolic speech is probably more likely to cause violence, but instead of dealing with this, you just point out that other types of speech can also cause violence and call it case closed. What’s particularly bizarre is that your objecting to creating a social rule against stochastic terrorism and argue for a liberal nor that people who are particularly vitriolic should just move to places where it’s acceptable to say such things. Yet there are many types of speech for which you yourself would consider this, not adequate, for example, defamation, or publication of private details, obviously, those are arguably different, but the point is that liberalism has always been open to social rules against speech. So acting like it’s inherently a contradiction of liberalism to have social rules against speech makes little sense. To be fair, you never actually did come out and say it’s against liberalism but does seem like most people who take that away from your post and yet it seems perfectly compatible with liberalism to have numerous controversial social rules against speech. For example, lots of your fellow travellers, think there should be social rules against invocations of sarcastic terrorism. I doubt you want to declare that all those people are not liberal and in any case if you interpret liberalism, so narrowly almost nobody would find the position describes their views, and it would become a much less attractive idea.
Event ignoring the diversion about liberalism, they are perfectly obvious middle ground positions that you never consider. I’ve mentioned this before, but for example, look at the guidelines that Z verse one recommended when advising people how to discuss AI risk. I doubt you think those social rules are bad, and if you do you need a much more substantial argument than what you have provided, because most people would consider that kind of rules calling, for, for example, no implications that political violence will be the inevitable result. If you demands are not met to be completely reasonable and nothing that you have said addresses such common positions. You just spent entire post attacking a week man extreme version of sarcastic terrorism argument where any criticism or at least any sufficiently harsh criticism is automatically of the table and it’s a bright line rule and then jump to your own bright line rule that actually, you can never object to any violence caused by a speech by criticising the speaker because it isn’t the speakers fault which is just absurd. Why speech so special that we have to pretend actions don’t have consequences within almost any other arena. You would find this absurd. For example, if I sell you poison and you use it, it would be completely legitimate to enquire whether I took a appropriate precautions and that even applies to less obviously dangerous substances like materials which have both beneficial and destructive uses. You just declare your position without providing any account of Y speech has this exceptional property and which things exactly would have this exceptional quality. This is just not a good argument because you argument at most only supports not calling for outright bands on criticism, which nobody actually believes would be a good idea.
> Probably there are some cases where extra-state violence is acceptable, for example assassinating Hitler or overthrowing an illegitimate government. I don’t want to rule this out by fiat, and I think it’s wise not to discuss the conditions permitting them too specifically
I do not understand this part. Why is not better NOT to discuss e.g. when to rebel against the government?
In my view, it is essential to establish a clear distinction between the mass media and political rhetoric on one hand, and the speech of ordinary citizens—who are neither professional journalists nor politicians—on the other. Regarding popular influencers, a specific threshold should be defined beyond which a blogger legally transitions into a professional journalist. Logically, this threshold ought to be financial: specifically, the point at which content creation begins generating revenue that serves as their primary or highly substantial source of income.Once this framework is established, the following principles should apply:Ordinary citizens are permitted absolute freedom of speech regarding any individual or topic.Individuals classified as professional journalists and politicians face severe, qualified constraints on their speech. Comprehensive statutory restrictions must be enacted to limit the freedom of expression for this specific category of individuals. The precise scope and nature of these laws remain within the purview of the Legislator.
The issue is when the public conversation creates a permission structure for stochastic terrorism to occur. For this to happen you need more than just criticism of bad ideas or bad people. You also need a general impression, supported by high status people of some kind of authority, that it would be good if harm were to come to the bad actors. When we see popular influencers and politicians celebrating deaths of the political opposition that's an example of something that increases the chances of it occurring. That kind of communication is clearly unprofessional and has moral implications.
There's a category of speech that I'd say falls outside the bounds of "incitement to violence" but implicates similar potential impacts. I think that "stochastic terrorism" could be a useful term if it were redefined as:
1) an actor with a meaningfully large platform
2) willfully spreading negative misinformation about a person or group of people, when
3) that speech would be identified by a reasonable observer as having been delivered with the intent of creating a sense of imminent threat, a threat which
4) cannot / will not be mitigated through existing institutional channels.
Being able to call this out when we see it (and, hell, at least raising an *orange* flag when we notice speech hitting 3 out of 4 requirements) enables social enforcement in a way that is at least much more difficult when we don't have useful descriptive terms. I see that as supportive of the cultivation of a pro-free-speech society.
I just published my essay that tries to re-frame this concept of "Stochastic Terrorism", acknowledging that the term has been abused, and that even its most horrid examples should never result in state-imposed consequences, but that we must ostracize real hateful bigotry out of polite society: https://elevin11.substack.com/p/dismantling-stochastic-terrorism
I don't think you distinguish strongly enough between the ethical and the legal here. One may be free speech absolutist to an absurd degree while considering stochastic terrorism morally equivalent to murder, if they expect the results of prosecuting that murder to be even worse than not to. Just like someone may consider Donald Trump equivalent to Adolf Hitler and even want him striken by God while not wanting to cause his death by man - except more reasonable in my humble opinion. The oposite is also reasonable. One may consider they're assassination of an aspiring dictator morally justified but not want it legalized or even pardoned in retrospect. It doesn't sound that unreasonable to me to put calling someone an aspiring dictator in the moral-but-illegal basket also, when done in good faith in a society in which such assassination is considered moral but not legal.
The quotation marks were doing more work than I expected. When I saw the title, I first assumed that this would be an essay about why we shouldn't engage in stochastic terrorism.
Anyway, I propose my own rule:
If you're accusing someone of conduct that might plausibly lead to someone committing violence against them, such as killing 700,000 people or attempting to destroy the American way of life, you should consider that you have a particular moral duty to check your facts and assumptions.
So lets say you are going to post a social media rant that suggests that cops murder purpletonians with impunity, and that neither purpletonians nor anyone who cares about purpletonians should ever call the cops to report a crime as a result. If this statement is reasonably likely to be true, then it's an important thing to warn people about! But if it's not, it might lead to violence against cops, or it might lead to crime victims not being protected from criminals. So while you might normally say it as part of normal exaggeration or because it's "directionally true" or because it fits with your existing assumptions, the possible downstream effects should raise a moral duty to do some extra fact-checking.
I basically agree with this, except that I think you shouldn't do the "directionally true" thing in any case. Maybe the violence-inciting bit adds some extra urgency.
The higher the stakes, the *more* certain you should be before airing the claim, not less. A claim that Haitians are killing and eating people's pets has far more potential for harm if it's false than a claim that a Haitian man was rude to you at the coffee shop.
I don't see why we should treat those claims identically just because they're in the same "direction" of expressing dislike for Haitian immigrants.
A claim that "some Haitians are stealing and eating pets" doesn't justify violence against Haitians in general. It justifies violence maybe against the theoretical subset of Haitians that did steal and eat other people's pets, but if that subset doesn't exist then, other than smearing the image of Haitians, it's not advocating for violence.
Contrast - a claim that <politician> is destroying democracy, ending elections, locking up <minority> and putting them in concentration camps - if true, this completely justify the assassination of said politician if you believe said politician cannot be stopped through normal means, or even creates the moral imperative to do so.
Do you think we have the moral imperative to murder Hamas politicians? They would seem to qualify (in that they actively support (and do!) murder gays*).
*unless, of course, you think that... genocide is different from putting people in concentration camps?
https://en.wikipedia.org/wiki/Just_war_theory
The Israeli forces seem to have that well in hand, so it's not something I'm too concerned about at the moment.
Not all human communities hold the same view of "Pets" vs "Food." I know where to get dog in Shanahai, and horse was routinely on the breakfast buffet at my favorite hotel over there. The fact that neither appeal to me, so I don't eat them, is not an excuse to express surprise and revulsion that someone has different food preferences.
Denial that other humans Do have different food preferences is every bit as racist as an expression of revulsion at someone else's preferences, absent cases of deliberate cruelty or theft of the critter in question.
https://www.haitianreport.com/2019/03/does-haitian-eat-cats.html
Someone's willing to document a tradition of eating cats. Apparently eating dogs is utterly taboo.
I don't expect that this would be a tradition kept in the United States (unlike, say, the Hmong practicing ritual animal sacrifice as part of culturally approved medicinal practice), and even if it was*, cats are practically free from shelters (and stray cats don't require paperwork).
*over a million Haitians (or native born of Haitian origin) in America? Is it possible one of them decided to do a culturally "okay" practice for back home, where it is "not okay" here? Sure, okay, maybe one person (said person is probably on the lower side of intelligence and cultural assimilation).
Indeed. Without taking a position on whether or not Haitians were/are eating people's pets, the idea that it's *obviously absurd* that people from a radically different culture would regard animals like cats and dogs as food animals rather than pet animals strikes me as itself obviously absurd.
Oh, I absolutely agree, I did enjoy dog meat in my childhood and bemoan its scarcity these days. Pigs are just as intelligent as dogs, after all.
To clarify, I'm saying people who steal pets and eat them deserve to be met with violence, not that people who eat <animals which are commonly considered pets> deserve to be met with violence.
I think you overestimate the rationality of those people who commit violence in response to stochastic terrorism.
I think that if I were the sort of hooligan who could easily be spurred to violence against Haitians, then I would not distinguish between the statements "some Haitians steal and eat pets" and "Haitians in general steal and eat pets".
Half of political rhetoric is directionally true. There's always two groups arguing in opposite directions.
At least two.
Sure, but they could both be arguing in different directions while both saying directionally-true things.
(Tempted to provide examples but I feel like any specific example is more likely to be distracting than clarifying.)
Much, much less than half of political rhetoric is directionally true. The fact that so many reasonably-intelligent people implicitly accept the absurd assumption that there are only ever two possible responses to any issue is the kind of thing that (almost) makes me wish the paperclip maximizers would hurry up and get here already.
Half are within 90 degrees of the correct direction.
A MAX of 50% of political rhetoric is directionally true; most often in conflict all parties are misinformed or are lying about most points.
Take American political parties for a case in point.
What does 'directionally true' mean. Google said 'moving in the right way'. And I'm like, how does that have anything to do with what is really true. (and of course who is defining 'right'?)
"Directionally true" is basically "fake but accurate" or whatever the old defense of playing loose with the facts was. When you get the details wrong, but you're, say, slandering a bad person, so it's you move peoples opinions in the direction you want them to go. Obviously highly subjective and basically begging the question.
So "directionally true" actually means "a pack of lies", then?
It means "throw enough mud and some will stick".
For what it's worth, my observation is that a bunch of the people I talk to seem to have the implicit belief that fact-checking accusations against someone they know to be bad (Epstein, for example), is morally suspect.
I don't think they have that *explicit* belief, but they seem to have a strong disgust reaction when presented with evidence that a particular accusation or statement in that direction is incorrect, and I suspect that applies in their interior thought processes too. They know (correctly) that Epstein was a bad guy, so it feels immoral to fact-check statements that support the conclusion that Epstein was bad.
I agree with Scott that this is bad, but what I was trying to say was that even if this is your attitude, I would still argue that you should apply extra caution when your statements about Epstein are likely to cause harm. (For example, accusing a possibly innocent person of conspiring with Epstein or accusing him of being part of a larger pattern of Jewish pedophile conspiracies.)
It means "Well maybe Bad Guy Politicians didn't call for all Purpletonians to be hanged from lampposts, but this guy went out and beat up a Purpletonian, and that is because of the climate of hate and fear against Purpletonians, and that climate is encouraged and created by the likes of Bad Guy Politician who said 'there has been an increase in crime by Purpletonians who deal drugs', so in the end it *is* all Bad Guy's fault the other guy felt free to go out and beat up Purpletonians!"
The term may be recent, but the underlying views go back as far as you'd like to go. See for example the attempted assassination of Gabby Giffords in 2011. In the immediate aftermath, Sarah Palin was blamed for encouraging the shooter by using a crosshairs ad in 2010 as part of a campaign about fighting elections against perceived vulnerable Democrat-held seats:
https://abcnews.com/Politics/sarah-palins-crosshairs-ad-focus-gabrielle-giffords-debate/story?id=12576437
"Crosshairs is a political phrase that emerged from Palin's political action committee SarahPac that targeted congressional districts for the Tea Party campaign in the last election, including the district of Rep. Gabrielle Giffords.
Although Palin later denied she meant the graphic over the districts to look like a gun sight, it is part of the hunting lexicon that critics say she prefers.
Comedian Frank Conniff tweeted: "Hey, Sarah Palin, hows that hatey, killy, reloady, crosshairsy thing working out for ya?"
Facebook executive Randi Zuckerberg said many people on the social networking site are asking whether Sarah Palin is to blame."
The shooter turned out to be probably mentally ill and also shot six other people during the attempt, but the publicity was all about "Right-wing Republican rhetoric encourages shooters!"
https://www.theguardian.com/world/2011/jan/09/gabrielle-giffords-shooting-rightwing-rhetoric
"The US was tonight seized by a fierce debate over whether inflammatory rightwing rhetoric was to blame for a shooting spree in Tucson, Arizona, that targeted congresswoman Gabrielle Giffords and left six dead, including a nine-year-old child.
Giffords, 40, who remains critically ill in hospital after being shot through the head but is expected to live, criticised Sarah Palin last year for putting her and 19 other Democrats on a hitlist of districts, each shown as being in crosshairs.
"When people do that, they have got to realise there are consequences to that action," Giffords said. Palin today distanced herself from the shooting, as did leading figures of the Tea Party movement. Conservative bloggers accused liberals of seeking to exploit the attack."
I mean, I hate Sarah Palin as much as the next guy, but this blame-mongering is ridiculous. We had a Labour MP in the UK who was accused of "betraying" the leadership (Jeremy Corbyn) and responded by saying "I wouldn't stab you in the back, I'd knife you in the front", and a bunch of right-wingers tried to get people to treat this as a death threat. Absurd.
If I say "group X wants to murder me and everyone like me", when the truth is that some member of group X might have been a little less than perfectly polite to me once, what I said is "directionally true"; that is, there is some other, true but entirely unremarkable and useless for whatever rhetorical point I am trying to score, thing that could have been said, but instead I chose to say something literally untrue then rapidly backpedal and point to the true thing when called on it.
It means "false, but will cause the casual listener to believe something else that I think is true".
So, "Hitler was cruel to anmials" is directionally true. It's absolutely false, but it leads people to believe that Hitler was a cruel person and Hitler was a cruel person.
It would be better if people stopped doing this.
It depends on the number of dimensions. The more there are, the harder it is for two vectors to line up. For an extreme example, consider the semantic space of LLMs. 'Stone', 'cat' and 'river', say, are all practically orthogonal to each other.
Political arguments don't have infinite dimensions, but they could reasonably have more than one or two.
It's possible to have reasoned opposite positions on Haitian immigration that are both opposed to 'they're eating the pets'.
Say in your location it rains 80% of the time. If most people think it rains only 50% of the time, then saying "it always rains" is directionally true, even though it does not rain 100% of the time, because averaging it against the current belief moves everyone closer to the true statement. However, if most people think it rains 90% of the time, then "it always rains" is now directionally false, because it actually rains less than that.
Since this is so dependent on what most people think to start with, which things people think are directionally true largely depends on their social group, which makes it unreliable.
Oh, I agree, I was just trying to say whatever your usual standard for fact-checking is, it should be higher if you're accusing someone of something that would justify violence against them.
Sounds a bit like proposing to expand the definition of what counts as "per se" defamation to things that might get you killed if someone believed them.
I like how purpletonian is supposed to kind of elevate the argument to its most rational and general, shedding any prior notions, but there's only one demographic at least in America known for being killed by the cops so its entirely rhetorically superfluous, I was amused by this and good post
The entire Internet-smart-person thing of pretending to abstract away preconceptions by recasting arguments as "If group A says B about group C" -- which is a) no simpler and b) not fooling anybody, since it usually follows a posting which A, B, and C are all clearly specified -- infuriates me.
I mean the intent is good, the intent being to emotionally depolarize a statement so we can take out our syllogisms and our prepositions and actually get to work. Scott does it in this article where he brings up "stochastic terrorism" accusations from all sides of the aisle
>Scott does it in this article where he brings up "stochastic terrorism" accusations from all sides of the aisle
Seconding praise for Scott's choices of examples here!
The intent is fine; I can't imagine being a person who it works on, or being a person who thinks it can possibly work on anyone. If such people exist, obviously I'm the one failing at theory of mind.
I feel like I remember it working on me sometimes. It's kind of a reminder to (if you want to think of yourself as fair) think about it more generally. Like it triggers your conscience, at least if you're constitutionally aligned that way anyway.
I think the point is to avoid having the discussion be over the object-level truth of your examples.
I’m curious: what do you mean by “being a person who it works on”?
Because I have no trouble imaging a person who has some negative emotional response to, say, some abortion doctors stating fetuses aren’t full humans and have no value, but then a lower emotional response to group A says B about C where the context implies it could easily be abortion doctors not thinking fetuses are full persons. (If I am clearly missing the key point of what you mean, please enlighten me.)
In general if you value abstract logic and the ability to talk about any hypotheticals/principles/etc., but also have negative emotional reactions, biases and preconceptions, then abstracting away can have value. Such values are true of myself as far as I can tell
What if one is making a case about all groups A, B and C, and not just particular values for those variables?
I was mostly trying to minimize any implication that left wingers are particularly guilty of irresponsible statements that are also false.
I don't think removing the parts of your argument that are extraneous and more likely to be inflammatory or derail the discussion is bad (obviously, I guess, since I did it.) You and gus are right that anyone with basic knowledge can decode it, but I think generalizing it is at least inoffensive, and probably slightly good.
White men get killed by cops disproportionately to the statistical threat they pose to cops or to civilians.
I think what's going on is that suicide-by-cop tends to be a white male thing and most other demographics find it stupid.
I think I want to see an analysis of where this standard comes down on the Nuremberg Files website, which provided private identifying information about abortion providers, and was widely understood to serve the purpose of enabling harassment of those people, and was known to be circulated on message boards where people were saying that abortion was murder and that killing those who performed abortions would be virtuous (even if none of them was dumb enough to state, even anonymously, that they personally planned to murder anyone).
Like, we have court doctrine about "true threats", and we have some ideas about where people have the "expectation of privacy". I largely agree with the thrust of this post, but is there _some_ kind of line, short of explicitly doxxing somebody and saying "hey folks, you should go murder this guy"? I get that this is a tricky concept, and "I know it when I see it" is not a satisfying standard (because anyone can decide they see it in all criticism by their enemies, and none by their friends). I'm not sure how you build a workable standard that doesn't immediately slippery-slope into stupidity.
Just saying "Nuremberg" rhetorically suggests a certain appropriate response.
Give them a trial and only hang the ones found guilty?
"We did the trial, here's the list of the guilty."
Yes, this is exactly the character of the problem.
I fundamentally agree with Scott that the accusation of engaging in stochastic terrorism has expanded drastically beyond the bounds of what is justifiable, and is used to try to chill speech that the person making the accusation doesn't like.
At the same time, it still seems like stochastic terrorism is a real thing, I can point to examples where I believe we all ought to agree, "Yup, these people were engaged in speech that had the clear intent of getting people hurt or killed."
I guess I kind of agree with None of the Above that a decently written law, that could be used as a standard by a jury, might be the right way to deal with this. Proving the mens rea for this crime would be a high bar, but surmountable in extreme cases (like the Nuremberg Files). At least, assuming it's possible to find twelve normie citizens without getting at least one or two who are hyperpartisan cultists for either side of the argument that the stochastic terrorist was trying to influence through violence.
We have this, right? Credible death threats are assault, incitement is a crime, conspiracy to commit other crimes is a crime. But none of those lets you do anything to the guy who says "f--ker had it coming" about Charlie Kirk or George Floyd, or who says "From the river to the sea" or "Blue lives matter."
I try to censor in my comments any comment that reveals too precisely where anybody lives. For example, one commenter had a funny story about riding the elevator with Monica Lewinsky in the XYZ Apartment Building, but I spiked the comment.
With me, this goes back to the 1980 murder of John Lennon outside his Dakota apartment building.
John had doxxed himself. Around 1979 he'd gotten over his drug problems and started doing witty interviews again. One of his favorite jokes was how he'd written the line "Imagine no possessions," but now he owned several apartments in The Dakota.
I have seen "directionally true" used dozens of times and it always, *always* means "indefensibly false, but ingroup enough that I want to defend it".
Yes, I agree it's bad, and I don't think people use it to defend their own ideas, so maybe it's unfair.
However, my impression is a lot of people *do* limit their bs detector or fact check etc. if a particular fact implies something they believe to be true. So for example, if you're confident that President Camacho is a scumbag, then maybe you retweet a story that the Pope gave a sermon scolding President Camacho without checking whether it's a hoax or not.
Not sure I understand your point. Yes, people definitely subject claims they dislike to greater scrutiny than claims they like. And the fact that we do this is bad; it leads to widespread distorted judgment on many important topics, and undermines important norms of truth, honesty, and evidence.
If you're saying "when stakes are high, you should try EXTRA hard to avoid this bad thing that you should actually ALWAYS avoid but that most people aren't ethical enough to consistently avoid" then, sure, I guess.
Yeah, mostly that second paragraph.
Technically, I'd argue that people should spend a rational amount of effort fact-checking their beliefs, and that the rational amount of effort should be higher when they are publishing a belief that is more likely to lead to harm.
Also, since not everyone is a rationalist, I'd be happy with a moral value that it's OK to criticize insurance executives and Chinese exchange students and Israelis, but when the criticism might cause harm, there's a particular moral duty to do your homework and not exagerate.
There's a meta issue here. Do you care about the literal truth or falsehood of a claim independent of what policies or whose side it supports?
I think of "directionally true" as being like the lies/approximations that we tell children.
"There are two sorts of chemical bonds -- covalent and ionic" is the sort of directionally-true thing you say to introduce an eleven-year-old to basic chemistry. If they continue to study chemistry then eventually they'll learn that it's more complicated than that.
Fair, though, roughly speaking, most such statements can be labelled as approximations just by prepending "Roughly speaking" to them.
( And, these days, most such statements can be clarified by the eleven-year-old asking a SOTA LLM what the next level of detail is... )
The difference is, we tell those things to children because smoothing over the intricacies of complicated subjects is a necessary part of teaching anyone anything.
But when people use the "directionally true" line in politics, it's pretty much always either because they are too lazy to engage with the complexities or too cowardly to admit that they were wrong or (most commonly) both. When expressing political opinions to adults you *can* just qualify them with things like "well, it's a complex issue, but I expect [Trend X]" to broadly hold.
If instead of doing that, you bloviate about how Those People Are Ruining Everything, then you deserve exactly zero credit for getting the directionality right *even if* it turns out that Those People are Sometimes Ruining a Few Things. And in reality, claims of "directional correctness" basically never actually wait for better data to come in: they just beg the question while retreating to a more defensible motte and waiting for people to get distracted and move on.
I believe that my outgroup is bad. I said a random bad thing about them.
It is "directionally true" because *bad* and *bad* are the same direction.
Agreed about the title ambiguity. Contrast “Against the Concept of Telescopic Altruism.” “Against the Accusation of X” would probably be even better.
Sure, ok, but the question is how to enforce this norm. Norms that aren't enforced can be expected to have little impact on public behavior.
Fair, and since I am both lazy and lack influence, it will be a heavy lift!
I'd love to see it get a catchy name and see people scolding the other side with something like: Ok, we get that the Frankist child grooming conspiracy would be essential to reveal if true, but since it's both (a) a dangerous accusation and (b) you verifiably haven't done your homework, you should be shamed.
Sadly, I think there are too many people around who care little or nothing for our shaming. Might even be proud of it. I think some level of restriction of speech is the only consequence that might moderate the effect of such speech on innocent victims.
I literally just wrote an essay on this! I take a slightly different tack. I'll share here when it's published, but in the meantime here's the gist which I think is a critical lens you didn't drill into enough.
Stochastic terrorism, the worst and non-imaginary instances of it, can cause very serious even if usually very indirect harm. Like pretty much all speech, it should be legally free to express. But we can and must treat overt hateful bigotry and the endorsement of violence as a violation of our universal moral TABOOS, with principled ostracism and social consequences to push this hate and the people who espouse and sanitize it out of polite society. This is how we marginalized and defeated the KKK. If we don't, that hate will normalize and turn into violence, as we are seeing.
I love the 1A and I loathe cancel culture, but we have to make certain speech and behavior socially unacceptable, and I believe we can do that in a principled way that aligns to our universal, liberal values. I wrote more here criticizing Greg Lukianoff and FIRE for their over-simplified "free speech culture" absolutism: https://whiterosemagazine.com/the-dangerous-myth-of-more-speech/
Feedback and criticism welcome!
I think the first KKK was put down by the military occupation, the second by its own scandals, and later iterations were infiltrated by the FBI.
Your linked post makes some fair points, but also some lousy ones.
> Greg often asks, “Am I safer for knowing less about what people think?” Of course not—I want to know exactly who the bigots are. But we can monitor the fire without letting it burn in the living room
Speech does not burn your living room, just as it doesn't break your leg or pick your pocket (as Thomas Jefferson put it).
> Granting an unrepentant bigot the same professional standing and platform as those who uphold our shared humanity isn’t “openness”—it’s a subsidy
No, neutrality is never a "subsidy". I think that speeding cars should have traffic fines, but it's not a "subsidy" to not have any speed limits.
> This is the Paradox of Tolerance: if you extend unlimited tolerance to the intolerant, the tolerant end up destroyed
The United States has a First Amendment which protects the speech of all political ideologies, even the most tolerant. There were times when that was less protected (such as Schenk v. United States), but since SCOTUS had started protecting it more, the US has not gotten destroyed but instead kept trucking. In contrast, Weimar Germany actually did have some laws we would now categorize under "hate speech" protections, that would be prohibited under our First Amendment, and yet they fell to the Nazis. The American approach of prohibiting actual violence rather than speech appears to have worked out better, regardless of what people might imagine in theory.
> We need to stop pretending that “freedom of association” only protects the speaker
Fair enough. Does FIRE actually deny that?
> A truly free society doesn’t require us to be indifferent to the destruction of our fellow citizens’ dignity
I agree with Steve Pinker's "The Stupidity of Dignity". We can make consesquentialist arguments about what makes people better vs worse off without bothering with such notions as "dignity".
Thanks for your thoughts!
1. "the first KKK was put down by the military occupation, the second by its own scandals, and later iterations were infiltrated by the FBI."
There were numerous reasons the Klan was defeated, but for the Third Klan especially, their hate was seen as TABOO and socially unacceptable and that absolutely limited their ability to recruit and make their ideology normalized. You couldn't be an open Klan member and be openly welcomed in most of polite society. You couldn't even be friends with an open Klan member. Their masks were to hide their faces and identities.
2. To your points re "burning" and "subsidy", those are analogies. I'll grant they're imperfect (like all analogies) and easy to poke holes in. I won't spend time elaborating here. Same with "dignity", which I really mean "letting people live without the fear that they'll face random violence or discrimination that is actively and openly encouraged by many people in society without consequence."
3. "The United States has a First Amendment which protects the speech of all political ideologies"
I'm only referring to social/private consequences, not ones imposed by the state. The latter is what the 1A protects from, and I support it fully for all the reasons you listed.
4. "We need to stop pretending that “freedom of association” only protects the speaker Fair enough. Does FIRE actually deny that?"
Yes, FIRE would and does argue that if for example somebody was fired for celebrating Charlie Kirk's murder after a mob demanded that they get fired, that would be cancel culture, and is thus bad/wrong. Greg used the broadest possible definition and continues to double down on it, not acknowledging that there are and should be material social consequences for certain speech that society broadly (nearly universally) believes is unacceptable and antithetical to our liberal democratic values.
If we do not uphold this speech as TABOO, it will undoubtedly normalize, and that will undoubtedly lead to catastrophe. And we can do it in a principled way. We must, or we are done for.
It is not covered by the first amendment and should not be, but a common norm of "the boss will fire you if he doesn't like your off-the-job political speech" would yield a massive reduction in peoples' ability to speak their minds, since only retired and independently wealthy people would dare take public positions that their employer disapproved.
This is why I created principles that limit the ability for this to succumb to the excesses of "cancel culture". It's not just "opinions I don't like". It's only when people cross a very specific line and refuse to clarify or apologize.
4 Principles for Restoring our Universal Moral Taboos
1. The Red Line: Limit actionable taboos to overt bigotry, dehumanization, and the endorsement of violence.
2. The Consensus Test: Distinguish between subjective offense (which is partisan and open to individual interpretation) and a Shared Moral Violation (which reflects a broad, trans-tribal consensus and warrants consequence).
3. The Private Mechanism: Enforce standards through civil society (employers, associations), never government coercion.
4. The Open Door: The goal of consequence must be correction, not permanent destruction; always offer a path to redemption.
The obvious placethis breaks down is the consensus test (many high profile cancellations during the Great Awokening were enforcing a moral consensus of a small subset of adademic, media, and tech employees.
In 2020, what fraction of people in the US would consider deadnaming a transperson to be a serious infraction that ought to lead to someone being fired? What fraction of Americans find someone saying "serves the bastard right" when they heard about Charlie Kirk being murdered a serious infraction of this kind? I think you are looking at >25% of people who would find neither of these very offensive, and yet people in a position to enforce consequences for speech were very happy to punish both when they had the chance. ISTM that doing this almost always ends up involving the consensus of some powerful people, whether or not that reflects a society wide consensus. And indeed, it looks to me like a lot of cancellation campaigns, including the woke ones and the Charlie Kirk ones, were tranparent attempts to impose a small group's consensus on everyone.
The point of the consensus test is that almost nothing reaches it. Only overt, hateful bigotry and the endorsement of violence, things that directly go against our core universal values of liberty and equality. Nearly everything else is tribal and should be rejected.
"cancelling" the people who celebrated Kirk's murder was absolutely the right thing to do, as ~99% of Americans believe political violence for speech is unacceptable. So somebody endorsing that political violence is violating our universal moral taboos. Conversely, somebody who said "I don't care that he's dead, he was an asshole" was being insensitive and crass, but that is not a universal taboo, and thus they shouldn't get fired.
If we don't fire people who explicitly endorse Hamas, terrorism, political violence, etc. that shit will normalize and spread like wildfire until we have mass anarchy and violence. It doesn't mean 20% of Americans will hold those views. But even a small minority empowered to think their views are socially acceptable, and are being encouraged to burn it all down, will try to do exactly that, in big and small ways.
Klan members wore masks to hide their identities in their first iteration, because they were insurrectionary terrorists fighting a military occupation. The second Klan were cosplaying as terrorists.
> I'm only referring to social/private consequences, not ones imposed by the state. The latter is what the 1A protects from, and I support it fully for all the reasons you listed.
You may be, but Popper was explicitly talking about a government and the two questions of whether it should have the legal right to suppress certain speech, as well as whether it would be prudent to do so if it had such a right (most people bringing up his supposed paradox forget his more nuanced take on the latter question).
Empirically speaking, I just don't believe that "we are done" if we don't stigmatize the celebrations of death/murder via firing. Solving for the equilibrium, I think that firing people in response to a mob of outsiders getting mad about their off-the-clock speech just incentivizes "offense archaeology" trying to dig up things to be mad about and spreading them. Instead, the best response to bad speech is not firing, but good speech counteracting what they say. https://slatestarcodex.com/2013/12/29/the-spirit-of-the-first-amendment/
I just asked Gemini if Popper was talking only about government interference:
"No, Karl Popper was absolutely not speaking only about government interference.
The popular modern internet meme version of the Paradox of Tolerance often formats it as a neat justification for immediate state censorship or legal bans, but Popper’s actual text in The Open Society and Its Enemies (1945) argues the exact opposite. He viewed state suppression as a last resort, placing the primary responsibility on civil society, public opinion, and social norms."
I understand where you're coming from encouraging free/more/good speech, and I'm on board with that too. But that's not always an option, especially in the internet age. Do you think the people celebrating Kirk's murder shouldn't have been fired? What would happen if they weren't? Is it more likely that their friends would talk them out of saying that kind of stuff, or that they and others would feel emboldened to start calling for the deaths of others?
When people celebrated Oct 7th and Hamas and they didn't all immediately get fired, the taboo shattered in a big way. That doesn't mean it can't be restored. But it must be restored. The incentives are too strong to pull young people towards expressing this hate, until the permission structures are created for anarchy and violence at a level we won't be able to contain.
I agree he said that suppression is a last resort. That's what I was referring to when I wrote "most people bringing up his supposed paradox forget his more nuanced take on [whether it would be prudent to do so if it had such a right]". Saying that it would be unwise for the government to do so is explicitly talking about the government, as I said.
> Do you think the people celebrating Kirk's murder shouldn't have been fired?
I am with Scott Alexander that unwanted speech should not result in firing. The only exceptions would be if it's somehow related to their job.
> But it must be restored. The incentives are too strong to pull young people towards expressing this hate, until the permission structures are created for anarchy and violence at a level we won't be able to contain.
I disagree. Rules against actual violence are enough.
> I am with Scott Alexander that unwanted speech should not result in firing. The only exceptions would be if it's somehow related to their job.
Respectfully, and humbly, I think you and Scott have not thought this through enough. I'll use a simple example of an extreme edge case (but seen not infrequently in real life) that I think will show where your principle fails:
I worked at Google for a long time. If a colleague of mine put a post on social media celebrating Charlie Kirk's death, I and many of my colleagues would rightly feel incredibly unsafe in a workplace with somebody who thinks political violence against speech is acceptable, because surely I have some views they'd find "offensive" and "dangerous" too. There's a 100% chance that person would be fired, and that would be true in most other serious businesses that have codes of conduct and employees that would raise those concerns. And of course, the employer would not just have the legal right but plenty of justification to do so (and arguably a big legal risk if they didn't fire him).
The same could apply to a business. If the supermarket I shop at has a checkout kid who posted such a video, I'd feel unsafe going to that supermarket. I'd tell their employer to fire them or I would not be returning. I probably wouldn't be the only one doing so, and the manager would almost certainly fire them.
If you were the manager at a company and your employee celebrated Kirk's murder (I don't just mean a crass comment, I mean open unabashed endorsement of political violence), and your other employees or customers said they felt unsafe and made that ultimatum to you that it's him or them, what would you do? If you decided to fire them, would that be "cancel culture"?
Endorsing the chilling effect as a way to encourage democracy strikes me as a genuinely naïve approach. Well-intentioned, but naïve. You are welcoming this being weaponized against whatever opinions you hold.
So when people openly celebrated Kirk's murder or Hamas, and others reached out to their employees to get them fired, was that wrong?
I'm aware of the risks of this getting weaponized, which is why I created 4 principles (below) to restore these taboos in a way that doesn't violate the 1A or succumb to the excesses of cancel culture.
It's totally fair if you say that these can't work, but would you argue that they aren't probably the right approach for society to follow? And if we don't, what's the alternative? "cancel culture" gets weaponized on a partisan/tribal basis, and every tribe permits the most abhorrent behavior within their in-group? Sounds like a recipe for the downfall of society to me.
1. The Red Line: Limit actionable taboos to overt bigotry, dehumanization, and the endorsement of violence.
2. The Consensus Test: Distinguish between subjective offense (which is partisan and open to individual interpretation) and a Shared Moral Violation (which reflects a broad, trans-tribal consensus and warrants consequence).
3. The Private Mechanism: Enforce standards through civil society (employers, associations), never government coercion.
4. The Open Door: The goal of consequence must be correction, not permanent destruction; always offer a path to redemption.
Yes, this was wrong, and also a number of people with legal or contractual employment protections won fairly big lawsuits against their employers for the wrongful terminations. At will employees are out of luck, however.
"Yes, FIRE would and does argue that if for example somebody was fired for celebrating Charlie Kirk's murder after a mob demanded that they get fired, that would be cancel culture, "
FIRE lobbies two kinds of organizations.
1. public educational institutions, whose hiring and firing decisions are directly implicated by the first amendment, not private associative decisions.
2. private educational institutions that claim to subscribe to equivalent norms of "academic freedom."
if you're a private educational institution that doesn't make claims about "academic freedom" or similar, you go on their "none of our business" list and they never really bother criticizing you. if you're not an educational institution at all, you're totally out of scope.
I know FIRE's scope, but unfortunately Greg and co extend the same principles to apply to all manner of consequences for speech. He wrote a whole book about cancel culture, which he defined incredibly broadly, and afaik has never said anything about when "cancelling" is actually ok. He calls himself an "opinion absolutist" and explicitly says he doesn't want anybody to be fired for their legal speech. That means he wants people to tell us they're Nazis but to not get fired from their jobs for being Nazis.
FWIW, I approached Greg privately asking him to clarify this position which I saw as flawed. He refused to acknowledge it, and continues to. He could really easily squash all of this but he chooses not to.
It is perhaps important to identify the individual when talking about the individual, and the organization when talking about the organization.
>Speech does not burn your living room, just as it doesn't break your leg or pick your pocket
I expect that pretty much every criminal conspiracy to do any of those things involved the perpetrators talking to each other to make a plan first.
'Speech is always harmless' is a really misguided defense of free speech, because anyone with half a brain will quickly realize it is false, and most of those will not be impressed by 'well actually speech and speech acts are two different things' or other such apologia to defend the idea.
The better approach is just to admit that some speech is harmful, obviously, but we allow all kinds of harmful things to be legal because we don't want or trust a complete police state which intervenes in every moment of our lives. Free speech is important enough that we will accept great harms to maintain it, that's a good and proper stance to take.
Denying those harms makes it sound like you are not actually committed to the importance and benefits of free speech. Like, *if* it caused any harm ever at all obviously we would ban it, but it just fortunately happens that it is never harmful so luckily we can support it. That's not a stable position.
I'm always a little baffled by people who want to construct their models or morality such that the things they want to be legal/allowed are always completely morally blameless. Much better to accept that we all do harm, we all do good, we should try when possible to do less of the former and more of the latter, but there's no way to live a perfectly blameless life and anyone who pretends to is trying to sell you something.
Yes, tolerating harmful things because of greater harms that can be caused by banning them is the basis of The Bill of Rights (US version).
I'm not arguing to ban them. Absolute legal tolerance for speech from the state is necessary. Infinite *social* tolerance is suicidal.
Even the US has laws about disseminating state secrets and the more loathsome types of pornography. Should those things be allowed?
And yet NYT reporters are openly flouting those laws (specifically, about the Situation Room reporting, which is potentially/likely to compromise the intelligence the POTUS is receiving from his subordinates).
In short, there is a difference between "there outta be a law about that" and "this is actively working to prevent the escape of state secrets."
So, I agree with this statement. I think where we differ is about what norms should be enforced with what kind of social sanction. I think norms of broad tolerance for even pretty offensive speech, ideas, and lifestyles are very valuable for letting all of us coexist in the very diverse society we live in, and for tolerating weird people who often have a lot to offer. And especially, coordinated efforts to punish speech by getting people fired have the potential to go terribly wrong, and have visibly done so in the last decade. A world where nobody will say what they believe lest they get fired for it is a world where nobody actually knows what their neighbors believe, and a world where a lot of unpopular opinions, some valuable and true, are seldom heard.
At the same time, social sanction well short of arresting or firing people is how we maintain norms of all kinds, from "don't be a Nazi" to "don't make sexual jokes/references about children" to "don't speak ill of the recently dead." A lot of online speech is not subject to the kind of informal thing where everyone at the table winces and someone mutters "Jesus, Fred, dial it back a notch."
Criminal conspiracies fall under the laws regarding the crimes committed.
Yes, 'conspiracy' is a great example of the law distinguishing between 'speech' and 'speech acts', thanks for giving an example of the type of thing that is not at all convincing here.
What if the "bigot" is right?
I'm not talking about quote unquote "bigots".
I'm talking about overt, hateful, eliminationist bigotry e.g. people explicitly endorsing violence and terrorism. If they are earnest* about that, enough to openly and unabashedly express those views, they are dangers to society. I'll still protect their legal right to express those views (short of immediate incitement) but we must punish those taboo violations with social consequences.
*FWIW, one of my 4 principles for restoring these taboos and arguably the most important one that prevents it from becoming a version of cancel culture is "The Open Door" - we can always let people apologize and clarify their views. And I have also developed a framework that does not assess somebody based only on a single statement, but based on what they say/do and then how they respond to pushback. If somebody says something incredibly inflammatory, but then is willing to apologize and not cross that line again, we should extend them grace. But if they double down on that overt hateful bigotry, they must be ostracized.
You wouldn't argue that KKK members were "cancelled" when they lost their jobs, that theirs was an imagined or exaggerated "bigotry", right?
You are conflating two things there though. Expressing "overt, hateful, eliminationist" opinions is NOT the same as "explicitly endorsing violence and terrorism".
Incitement to violence is normally where liberal societies draw the line on free speech, but expressing hatred to certain things strikes me as a perfectly reasonable thing to do if that thing is actually worthy of being hated.
I'm not talking about incitement, which can be illegal. I'm talking about things that are short of direct/immediate incitement as defined by the law, but still are encouraging violence and terrorism. Saying "Hamas is great I'm glad they killed those Jews and I wish they killed more" or "I'm glad Kirk was killed I hope they get Shapiro next" is not incitement but it is and must remain socially unacceptable. People must lose their jobs and standing in polite society, at least until they renounce that hateful bigotry.
Would you keep in your employ an avowed KKK member or Nazi, or somebody who openly celebrated Charlie Kirk's murder?
Depends if they were competent and professional in their other business dealings. I am not a believer in "bringing your entire self to work".
This would be an easier sell if we did not have ten years of worked examples where someone was widely denounced, cancelled, fired, and unpersoned for being a racist or nazi or some such thing, and when you dug into things a bit, the evidence was that they made a weird smirk while a bunch of weirdos were jeering at them, or that they posted lyrics to a rap song with the n word when they were teenagers, or told a tasteless joke in what they thought was a private forum, or some similar thing.
I agree that social pressure including ostracism is a natural, unavoidable consequence of speech. That is morally neutral (in Imperial Japan in the 1930s, you'd get in trouble for advocating for peace with the neighbors), but it is an inevitable part of human interaction. However, I think US society is extremely diverse, and norms that allow for a great deal of tolerance for weird or even offensive ideas, ways of life, etc. are neccessary to make it work.
I agree and understand. I call this "moral hyperinflation", and I think we need to reverse it, and certainly not accept that we simply can't socially punish the actual instances of that 'overt hateful eliminationist bigotry' because a lot of people abused the terms "bigot" and "fascist" for 10+ years. https://elevin11.substack.com/p/moral-hyperinflation
How do you feel about the cancelling of James Watson? This was expressing offensive ideas (probably in an offensive way, because Watson was famously a huge jerk), but not in a way that anyone could reasonably feel physically threatened by and with a set of factual claims that were at least plausible enough to not be obviously wrong.
I'm not super familiar with the case but I carve out "Observation" e.g. of differences across groups as socially acceptable (or at least not universally unacceptable) as long as it's not hateful and with intent or indifference to harm, which can be determined through dialogue. I would generally use Charles Murray for this example, which is easier bc he wasn't a jerk about it, and generally tried to be careful with what he was saying. If Watson was a jerk and careless with his words, indifferent at all to how people were going to interpret them in a way that might harm certain groups, then I think it'd be fair for many to criticize him for it.
I created a framework that I think even these edge cases can live within, but those edge cases still need to be explored in more detail with all their context. https://docs.google.com/presentation/d/1t6RqvguDqMVCt2L2edxcbh-FlchoCX5WuxudC9Ot91k/edit?slide=id.g3cbdd25b6d5_0_3#slide=id.g3cbdd25b6d5_0_3
If you're only talking about bigots who also endorse violence, then why bring up the concept of bigotry at all? Your red line already prohibits endorsement of violence, so all that adding "bigotry" to the list does is invite misunderstanding.
Because somebody saying "I hate those f-ing Jews, blacks, etc." still violates the taboo even if it doesn't explicitly endorse violence.
> I'm talking about overt, hateful, eliminationist bigotry e.g. people explicitly endorsing violence and terrorism.
I don't see how that example fits this standard. Personally disliking a group is not eliminationism.
> No, neutrality is never a "subsidy".
It is, in the sense that you are doing something for the other person that they would never do for you if the situation was the other way round.
If in a Green town the Purple minority is allowed the freedom of speech, but in a Purple town the Green minority is not allowed, then the Purple people have more freedom globally than the Green people. In that sense, the Green people subsidize the Purple people.
If in a Green town the Blue minority is allowed the freedom of speech, and also in a Blue town the Blue minority is allowed the freedom of speech, there is no subsidy, it is free speech for everyone.
Therefore, for a tolerant Green living in a Green town it makes sense to tolerate the Blue speech, but oppose the Purple speech. Especially when the Purple people say openly what kind of regime they prefer and are proud about it.
At the bottom is the expectation of reciprocity. I treat you nicely even if I disagree with you, because I imagine that in a similar situation you would do the same for me. If I know for sure (because you told me openly, because your entire philosophy is about that) that you would hurt me if you could, what incentive do I have to be nice towards you? Virtue signaling? Thank you, I am not signing up for suicide pacts. I am nice towards people who would be nice to me... or to people I do not have enough information about (because I prefer to err on the side of niceness).
Years ago Scott argued about whether it was justified to lie on behalf of a good cause https://slatestarcodex.com/2014/02/23/in-favor-of-niceness-community-and-civilization/ and Scott said he was still unwilling to do that, even against people who believed that tactic was justified. That is an asymmetry between Scott and his opponent, but it's not a subsidy.
>But we can and must treat overt hateful bigotry and the endorsement of violence as a violation of our universal moral TABOOS
As we've seen over the last several years, or possibly all of history, the problem here is that people *will* carve out exceptions to those taboos, and find excuse for any bigotry they themselves find palatable. This is bad for a multicultural society, and in the long run is likely to make multiculturalism unstable.
Few liberal-progressives are against *racism*; they're against one form in particular, sometimes a couple others when convenient, and the exclusions are not hard to notice. Few are against *sexism*, likewise. Few are against *violence*, likewise.
>This is how we marginalized and defeated the KKK. If we don't, that hate will normalize and turn into violence, as we are seeing.
What we're seeing is what happens when those exceptions get enforced. *That* normalizes hate.
I agree. I'm trying to say here "no more exceptions". These are universal moral taboos that we need to enforce no matter who it's directed at or coming from. We need to override the tribalism and partisanship and all hold the line, especially against our own, and condemn this hate. By excusing and tolerating it within partisan tribes, it becomes normalized. Everyone needs to stand up and say that hateful bigotry is unacceptable, and pressure others to enforce social consequences against those who espouse and sanitize it.
> If we don't, that hate will normalize and turn into violence, as we are seeing.
I feel like "the problem of our times is we didn't push hard enough against hate" is the wrong lesson to take from the last 15 years. We pushed harder than before, because before it wasn't necessary; and the more we pushed, the more the hate pushed back and came out squeezed between our fingers and said "see, see, I'm being repressed!" and gained sympathy thanks to it.
Ultimately I just don't think the root causes are that not enough people pushed. If a vast, vast majority of people share some principles then those principles are effortlessly upheld. If there's a significant minority that doesn't then the rest shouting about how bad they are or even punishing them won't fix things. I think there is a mix of underlying material conditions and fundamental loss of trust and belief in the liberal project, sadly, at the root of all this. It's not going to be fixed by screaming louder at people that they're being bad.
I think the problems of the last 15 years were we called things bigotry that weren't, that we excused hateful bigotry when it came from our own tribe, and we taught a generation that it was ok to be bigoted towards anybody deemed an "oppressor". We have to undo that and collectively hold the line with universal principles against overt, hateful, eliminationist bigotry, the thing that ~99% of us agree is socially unacceptable.
> I loathe cancel culture, but we have to make certain speech and behavior socially unacceptable
Making certain speech and behaviour socially unacceptable is literally what cancel culture is. We should be careful that any rules we decide on are applied fairly to both sides - yours, and your outgroup's; we don't want to end up creating yet another one of those irregular verbs so prevalent in political rhetoric, with a conjugation table something like - I align with universal, liberal values; you make certain speech unacceptable; he/she/it engages with cancel culture.
If your definition of "cancel culture" includes firing a KKK member or somebody who joyously celebrated Charlie Kirk's murder and wished for others to be next, then your definition is too broad.
I hear you on your concerns, and it's why I've laid out 4 Principles for Restoring our Universal Moral Taboos (below), which helps prevent this from becoming what I think is the better definition of cancel culture which was more often than not social consequences imposed based on offenses taken by a small group, and completely unforgiving.
I'm not naive enough to think that we can follow these perfectly, but I do think this is how society functioned until not that long ago, and it's something we must get back to or we're going to see a very quick backsliding into uncontrollable anarchy and violence.
1. The Red Line: Limit actionable taboos to overt bigotry, dehumanization, and the endorsement of violence.
2. The Consensus Test: Distinguish between subjective offense (which is partisan and open to individual interpretation) and a Shared Moral Violation (which reflects a broad, trans-tribal consensus and warrants consequence).
3. The Private Mechanism: Enforce standards through civil society (employers, associations), never government coercion.
4. The Open Door: The goal of consequence must be correction, not permanent destruction; always offer a path to redemption.
Can you expand on which of your four rules an employer's decision to fire "somebody who joyously celebrated Charlie Kirk's murder and wished for others to be next" fails to meet? It's not at all obvious from here.
What do you mean fails to meet? The people who unambiguously celebrated Kirk's murder crossed the red line (1) that nearly all of us would agree (2) is unacceptable. If they were willing to apologize and recant that endorsement of violence, their employer and others could decide to offer them grace (4) and not fire them.
What was actually quite inspiring to me was that the people who didn't celebrate his murder but were merely "remorseless", even though there was pressure to fire them, many employers did not, because 1 and 2 were not met.
I wrote about this here if you're interested: https://elevin11.substack.com/p/the-kirk-test-how-a-political-assassination
Hm, I took "your definition is too broad" to mean you thought the people in your example shouldn't have been subjected to social consequences, but now it sounds like you actually do want social consequences for them, you just don't want your particular set of social consequences to have the "cancel" label? Am I misunderstanding again?
That's right. I think "cancelling" is too broad of a term. Some people/behaviors MUST get "cancelled", or they will normalize and destroy America and the West. My principles are meant to re-impose those social consequences where they should/must be, without this devolving into the excesses of "cancel culture."
Do you have a principled disagreement with this view?
> I love the 1A and I loathe cancel culture, but we have to make certain speech and behavior socially unacceptable
"Cancel culture" is precisely about making "certain speech and behavior" socially unacceptable
I'm just going to copy what I replied above:
If your definition of "cancel culture" includes firing a KKK member or somebody who joyously celebrated Charlie Kirk's murder and wished for others to be next, then your definition is too broad.
I hear you on your concerns, and it's why I've laid out 4 Principles for Restoring our Universal Moral Taboos (below), which helps prevent this from becoming what I think is the better definition of cancel culture which was more often than not social consequences imposed based on offenses taken by a small group, and completely unforgiving.
I'm not naive enough to think that we can follow these perfectly, but I do think this is how society functioned until not that long ago, and it's something we must get back to or we're going to see a very quick backsliding into uncontrollable anarchy and violence.
1. The Red Line: Limit actionable taboos to overt bigotry, dehumanization, and the endorsement of violence.
2. The Consensus Test: Distinguish between subjective offense (which is partisan and open to individual interpretation) and a Shared Moral Violation (which reflects a broad, trans-tribal consensus and warrants consequence).
3. The Private Mechanism: Enforce standards through civil society (employers, associations), never government coercion.
4. The Open Door: The goal of consequence must be correction, not permanent destruction; always offer a path to redemption.
> I'm not naive enough to think that we can follow these perfectly, but I do think this is how society functioned until not that long ago,
I assure you, no society ever functioned the way you imagine. Certainly US never had a "trans-tribal consensus" of what is or is not taboo. I am 60 years old; I remember when it was entirely normal for an employee break room to have cutouts of porn magazines on a wall, and if any woman did not like it, too bad.
Now the pendulum swung in the opposite direction, but it does not mean it ever resided in some happy middle.
You're proving my point. There was no consensus on that example. Most women and some men found it offensive. It wasn't universal.
Compare this to the KKK. Pretty much everyone agreed their overt hateful bigotry was not welcome in polite society. That doesn't mean nobody could make or enjoy racist jokes.
Steelman my argument. Find the most extreme case where you think I'm right. I think the KKK (especially the Third Klan) is the right one.
I just published my essay that tries to re-frame this concept of "Stochastic Terrorism", acknowledging that the term has been abused, and that even its most horrid examples should never result in state-imposed consequences, but that we must ostracize real hateful bigotry out of polite society: https://elevin11.substack.com/p/dismantling-stochastic-terrorism
I was expecting this post to be about the extent to which stochastic terrorism is a demonstrable or quantifiable effect, so in lieu of that I spent a little bit of time researching it myself and it appears that there is strong evidence of correlation but little evidence of any actual causative effect. (Pretty much what I expected, but still interesting.)
The censorship question is another matter entirely, to me - even if there were strong, conclusive evidence that certain types of speech led to violent act by other people, I would be against censoring that speech.
I think we already have laws for libel and slander. Unlike in bonnie old England, these lawsuits eventually get settled hereabouts.
Perhaps the courts are less predictable in England, and that's why people don't settle there?
Going to court is more expensive in America. And you won't recover attorney costs if you win, so settling is more attractive even if you're only offered a discounted amount
More to the point, all 50 states and the Federal government have laws against incitement (to violence or other criminality) and US courts already have existing doctrine for what type of speech can actually be prohibited as incitement
Yeah, I agree with Scott's political argument that stochastic terrorism ought not to become a justification for arresting people for their political speech, but still think it is an interesting empirical question how much various kinds of hostile rhetoric lead to actual violence.
This post is a banger. https://drrollergator.substack.com/p/stochastic-terrorism-a-game-of-rhetorical
Thanks, it was indeed a very interesting read!
More informative but less argumentative than Scott's, I think they complement each other well.
I think where I land is that "will no one rid me of this meddlesome priest" is only remotely a thing you can hold the king liable for. If you're just a guy, you aren't under the obligation to understand your positions and opinions as potentially being interpreted as commands - actual material power on an individual level gives you that responsibility, because it's something more reasonably foreseeable. That's why I think we can say "important politicians have a responsibility not to incite things" - it's at least not crazy for them to think the things they say can lead to real actions - without having to wonder what it is that Jodie Foster did to stochastically induce certain events in 1981
Does Father Coughlin fall under "important politicians"? How about Stephen Colbert? Perhaps when you can hold a rally in Washington DC and bribe people to be there with funnel cakes, you're enough like an important politician to Watch Your Language?
(But don't we already have laws for libel and slander? How is this not covered by that? Can you really incite violence without distorting the truth?)
"will no one rid me of this meddlesome priest" did not distort any truths, did it?
Well, fair, I don't think generally accurate guidelines are necessarily going to fall along definitional lines that can be demarcated clearly. If anything, it's the sort of "reasonable person" test that's really difficult to determine legally, and is just sorta the way I think it makes sense to consider it for the "social stigma" angle. Testing intent and "he should've known" type things are more of a thought experiment, since we rarely *actually know*.
More than the "what's the brightline for sufficient amount of individual power" question (which I think varies - for instance, the leader of a small but fervent cult probably has less actual power than a mid-level politician, but probably contends with *more* likelihood that errant words turn into disproportionate action), I find the problem of collective power a bigger thorn. The most natural response to my position, IMO, is "a society with the power to set dangerous norms is made up of a lot of different people without much power in their own right - components of a powerful and culpable force for incitement of violence aren't innocent just because they couldn't do it by themselves". You see this a lot in modern politics, where the idea is "each person in X group made society more amenable to the conditions that allowed bad thing Y to happen, and this was a natural and foreseeable consequence of the world they worked for".
Personally, I think it's just practically unworkable to think this way - no one can say anything, ever, because everyone with that position might have the joint power to incite bad things to happen. But it's a bit inelegant, I suppose.
Dangerous norms include "It's okay to riot when a black person dies at the hand of a cop."
How culpable is the media? How culpable is any particular piece of art, or news-reporting? Nobody reports on the starving people who stopped rioting because they'd literally destroyed all means of food transport...
"The Media" is very powerful, in that they've managed to influence a lot of people -- it takes a lot of people to say "Yeah, Riot!" and actually make it happen.
Or we could point to Apple's practice, during BLM, of letting black people have free iphones (instead of, as was done in 2025/2026, simply bricking them).
"Nobody reports on the starving people who stopped rioting because they'd literally destroyed all means of food transport..."
where did this happen?
South Africa. The riots where everyone had glowsticks on, and there were civilians manning tanks because the police had fled (or do you remember the bookstore that was notably not touched during those riots? That South Africa had to do mondo damage control on...)
ah, I must admit total unfamiliarity with these riots.
>But don't we already have laws for libel and slander?
Technically, but they're incredibly difficult to win in the US. This *might* be changing a bit next Supreme Court term though.
I don't think we need more laws on the topic, but "the laws exist" is a different thing than "the laws actually handle this situation."
>Can you really incite violence without distorting the truth?
Yes, if the truth is unpleasant enough.
At that point, I have to ask if the violence was already likely to happen, though? If it's truth, then it's truth. And you're probably not the only person to uncover it.
People who can't discuss the truth are far more likely to turn to violent means of getting their manifestos heard.
I think the strongest case for a social stigma around certain truths is more about presentation and completeness - although, again, the "stochastic terrorism" thing should only apply in very limited circumstances regardless
Being fundamentally correct should be a complete defense against the charge of being unpleasant or inciting things for the exact reasons you mentioned. But important truths are often inseparable from a great deal of surrounding context. Isolating a given piece of the truth can send someone in the opposite direction that context does - a half-truth can be inseparable from a lie, with the biggest difference being that half-truths are infinitely harder to find out as misleading. There are similar factors around being inflammatory *about true things* - sprinkle a few obvious opinions around the hard facts and the emotional valence of the total package changes, which is the whole reason politics is a profession.
I don't think the average person should worry about this, outside of the general idea of "don't lie, even if you're crafty about it" - there's basically no chance that their half-truth is taken seriously and leads to "stochastic terrorism", it's not an outcome that's realistically forseeable without also forseeing issues with saying anything at all (because anything true or false could be misinterpreted by someone in the wrong frame of mind). The other issue is that telling the *literally whole* truth in any given situation is impossible, unless you're Prak OD'd on truth serum. But I do think there are ways in which parties who can reasonably forsee their words being taken seriously (in fact, people *trying their best* to make that happen, in the earlier case of politicians) can end up responsible for certain ripple effects resulting from even improper framing of the truth.
Libel and Slander doesn't really cover attacks on groups. "The Haitians are eating the cats and dogs" is a distortion of the truth, but not in a way that can be held accountable via slander charges
>to understand your positions and opinions as potentially being interpreted as commands
That's not what 'stochastic' means.
I feel like we should be able to understand this, as people who talk about expected utility calculations all the time... there is always a low probability that any person or group will be targeted as politically-motivated victims of crime, and every public comment engendering more hatred of that group increases that probability.
If your statement increases the probability from 1.2x10^-10 to 1,3x10^-10, that's a very tiny amount of harm in expectation and it's probably fine, you probably cause a lot more harm than that whenever you buy a t-shirt without checking how it was manufactured. But it's still a real phenomenon that you're responsible for.
No, no, we only do that sort of thing for the expected emotional utility experienced by crustaceans.
Worth noting England in 1170 was not a liberal society. I think it'd be perfectly fine for whatever sad sack is Prime Minister this week to say something like that about... like, a judge. (because it'd have to be someone in a position of official power for the analogy to hold) If a nearby RaSP officer went and murdered the judge, everyone would agree that officer was insane.
Turbulent. Thomas à Beckett was turbulent, not meddlesome. And interestingly, Henry II didn't even intend "turbulent" to mean something like "meddlesome", he was simply referring to the fact that Thomas à Beckett had a big aerodynamic spoiler coming out of his back like a Toyota Supra.
Re. "Only the King" - I would imagine that Scott probably has a wider reach, a bigger audience, and probably more economic power, than a Dark Age monarch. (Though of course, as of the time of writing, Scott has yet to amass any serious military power or become head of any major religions.)
A key distinction is between general expressions of discontent and approving of political violence but not having the guts to do it yourself.
Consider if you desire people to kill redheads but settle for publicly calling for mandatory hats. Someone kills some redheads because they should be hatted. Legally you are not responsible and should not be. Morally you might need to figure out whether this extra-legal violence is good and own it one way or another.
If instead you are actually calling for violence then you have a legal and moral issue. Legally, this is actual incitement to violence. Morally, you are weak - you want the violence to happen but want someone else to pay the cost.
On the other hand, if you just want them to cover up and have no desire for harm, then you do not and should not bear any accountability if someone escalates to violence
It's a simple question of is the cost via the probability of increasing the chance of assassination lower than the benefit of voicing concerns? It's a real cost, but the benefit is likely bigger in the general case.
I think this might be one of those cases like "should I tell a noble lie for the public good?" where I oppose even doing the calculation, on the grounds that requiring the calculation makes it too easy to get the calculation wrong, or to accuse others of getting the calculation wrong.
Sure but that's just the practicalities of how close of a process do you have to producing an answer for it. If you can have an exact answer, surely you'd use it. The calculation is noisy and expensive, so the hope is that the more strict approach is what works better given limited resources and information. But we should still remember that this approach is just an approximation.
> If you can have an exact answer, surely you'd use it.
No. And even the premise is illegitimate, since your "exact answer" is a mere probability.
A probability calculation can produce an exact (in the sense of perfectly accurate) result, just like any other calculation. I think you may be confusing different levels of knowledge.
I disagree that it's appropriate to describe a thing which is so hard that I don't trust any real human to do it as "it's a simple question of..."
I see the issue with the wording, but this is the underlying question. Answering it via a rule that works well the majority of the time, because that's the most practical way to answer it is well and good, but I disagree with rejecting that this is, again, the underlying question.
Tenoke didn't say the computation is simple, he said the decision rule is simple. That's how you reason about vaccine mandates, nuclear deterrence, and every other high-stakes low-certainty domain; why should this one get an exemption?
I agree this is in a cosmic sense the right decision rule, but I think "right decision rule" refers to a different, less obvious level.
This is like answering the question "what did you get for Christmas?" with "atoms". It is definitely true, but it's so true of everything that saying it in response to any particular question is likely to mislead people into thinking you're giving the expected normal-level answer that should apply in real life (they might think you got some kind of unusual scientific chemistry samples or something).
Sure, but the problem with answering "atoms," as you say, is that it's indiscriminatory; it's compatible with every answer. But cost-benefit would actively rule things out or make falsifiable claims about which cases are worse than others.
I doubt that if I say 'I am pro criticizing people publically because I estimate the benefits of it outweigh what I estimate is only a very small chance of increasing their chance of assasination' you'd be quite as confused about my position as if I said I got 'atoms' for Christmas.
I can see why that might not be the way to word it to get through some people, which is a more valid concern I do sympathize with.
Lying is by definition action under false pretenses, but criticism isn't
That feels like a dangerous position to take if you want society to be Consequentialist in general.
Like, sure, there's a chance that people will do the calculation wrong, or use the fact of calculation calling attention to the expected harm against people in bad ways.
But there's also a chance that saying 'we don't need to do calculations here, we have deemed this category of actions Morally Acceptable and no further analysis is warranted' just inculcates an adherence to unthinking deontology and virtue ethics in the populace. Which I think is *way* more subject to misuse and abuse.
If It’s Worth Doing, It’s Worth Doing With Made-Up Statistics
what is the thing that we are saying is "worth doing" here? Scott seems to be saying that analyzing the potential consequences of stating true criticism is simply not worth doing (and therefore not worth doing with made up statistics).
I read him as saying that the calculations is too dangerous because it is too diggicult, not worthless. But perhaps I'm wrong.
That seems alot like the "should I respond to deny untrue allegations" decision point, where once you've established that you'll do it even once you now have to justify why you aren't doing it in other situations.
And I find this particularly annoying wrt AI policy, as I don't think I need to append "you shouldn't actually go kill Sam Altman" to every single statement AND if you DO append it you get called out by idiots saying "oh so you think he's a reckless tyrant trying to become the immortal god-emperor of the universe but you won't call for killing him, I guess you don't really believe it" like some kind of stochastic terrorism offsides trap.
Far from being "simple", you could not make the calculation at all, not even with great effort. Most, if not all, of the numbers you put in would be guesses.
You are using your guesses either way.
And therefore humility is called for, instead of censorship.
If you were making predictions, sure, but we are usually talking about this in the wake of someone being killed.
Just take the number people in group X who have been lynched and divide by the number of posts and articles saying group X are evil monsters. That's roughly the number of expected deaths per comment.
It'll still be imprecise, sure, but given that this is already going to be something like billions or trillions of comments per death in most cases, missing by a few orders of magnitude barely matters on an individual level.
That's just nonsense unless you can prove a causative relationship between the two things. Can you?
... yes? The shooters often talk about their motives, and those motives match things they read and hear in the culture?
Like, at a certain level I don't understand the concept of a human in a society taking a social action, and that action not being causally linked to the society they were raised in and the specific circumstances of their life within that society. That feels like it's probably not a thing that happens much.
Of course, if we kept talking about this it would probably end up hinging on our definitions of 'prove', 'causative', 'relationship,' and 'two things'. Semantic arguments like that are boring, so let me just put it this way: yes I think that groups that are targeted for hate by large segments of society are generally in more danger than groups that aren't, and there's a causal relationship between those things.
> It's a simple question of is the cost via the probability of increasing the chance of assassination lower than the benefit of voicing concerns
Not convinced that this is the tradeoff. There's always ways to voice concerns which are more versus less likely to incite someone towards violence.
I don't think you're giving the powerful/powerless distinction its fair share. The obvious answer to your point would be that powerful people are not only much harder to assassinate (Trump has constant protection, the average mosque doesn't), they also tend to live with well-mannered people, while vulnerable people live with all sorts of crazies and junkies who are much more likely to attack them because they were influenced by something they heard online. If we're assessing risks, negative speech against one group is much more likely to result in dead people compared to the other.
I think this really only works with the President / everyone else distinction; I think even the average CEO isn't too hard to assassinate. I think the powerful/powerless distinction more often gets trotted out to mean things like "you can criticize white people, but not black people", where the two groups don't differ very much in security.
Trump was inches away from actually dying. He's alive because the shooter had bad aim more than because of the Secret Service. If Tyler Robinson had switched places with whoever that was, perhaps the outcomes would be reversed.
On a per capita basis the murder rate for presidents is pretty high.
I think the Secret Service being competent has a big effect on how far away the shooter is, what gear they can get on sight, how long they have to aim, etc.
The Secret Service failed to act on over 100 transmissions about Trump's shooter. This doesn't seem like competency.
The schoolteacher did actually think there weren't metal detectors, because he didn't see any metal detectors he recognized. The Secret Service disabused him of that notion.
Eh. How many transmission do they get about potential threats every week, how much budget do they have to deal with them, how specific were those transmissions such that anything more than 'continue your already-optimized normal security procedures'' was clearly warranted.
I'm not saying i think SS is definitely uber-competent, I'm mostly saying I don't think we could possibly know enough to judge, and that blaming the shooter for being a bad shot ignores how security made the shot harder.
Brian Thompson was admittedly shot by a handgun at close range, but Charlie Kirk was shot by a rifle at longer range. I don't think either rifle shooter was being pressured by security while aiming, rather they only revealed themselves by shooting.
Thomas Crooks had been monitored before the shooting and found a position which was obscured from the ground and Secret Service. Part of the reason for his inaccuracy is that one of the police officers clambered up the side of the building to see who it was, and he turned his gun on them, then took his shot immediately (before getting killed by the Secret Service). Although the SS/police failed to stop him from firing, their actions did make him less accurate; if they had all been sleeping before he'd taken his shot, he probably would have had an easier time finding a good spot and fired at his leisure.
Fair enough.
Not everywhere is as policed as the US, and not every politician is Trump. The U.K. has a hard job protecting ex prime ministers, it’s an ever increasing drain as we go through them like candy at the state fair*.
Anyway just this week an ex member of parliament was murdered, the police are treating it as lone terrorism because they treat all attacks on politicians as terrorism. The woman in question was a fairly conservative ex Tory (later reform) named Anne Widdecombe. No doubt the killer opposed her politics, and was perhaps influenced by strong online rhetoric.
Nevertheless I would agree strongly with Scott against banning critical speech, no matter how linguistically violent, unless there’s a direct call to actual violence.
The "go through them like candy" line somehow drew me to the conclusion that a solution for all relevant problems is just to feed ex-PMs into a wood chipper. No risk of their assassination, no drain on the budget.
Would also encourage them to cling to the job a bit longer, for better or worse.
Though it would also have an impact on who'd volunteer for the job.
Well, in practice at least. I think in theory the king just appoints whoever he feels like to the job. It just usually happens to be the leader of the majority faction in parliament.
At least the brits actually serve.
https://en.wikipedia.org/wiki/Magdalena_Andersson
Heard the security briefing, and resigned.
Which section of the rather long Wikipedia article is that?
https://www.sbs.com.au/news/article/swedens-first-female-prime-minister-has-resigned-after-hours-in-the-top-job/n5syxgxd6
So, prime minister, which is located in the third and fourth paragraphs down.
I wouldn't agree to "no doubt" at this point, we really don't know. Jared Loughner was on an entirely different world than every politician when he shot Gabby Giffords rather than actually having anything to do with her politics (he was upset that she couldn't answer a nonsensical question from him).
Yeh. No doubt was too strong. Should have been “it’s not unlikely” or similar.
That whole story is looking very bizarre at the moment. At least, according to the details that have made their way to the public.
But you have to factor in that many more people, and thus many more crazy people, know and are likely to care about the powerful.
> He killed an insurance executive because he had strong opinions about what insurance executives do wrong, and these opinions didn’t hinge on slight changes in the emotional resonance of the words being used.
I feel like a rhetorical trick is being used here. Sure they didn't hinge on "slight" changes, but the strength of his opinions definitely related to the emotional resonance and tone of what he read in the past.
My claim is that it didn't hinge on some specific beyond-the-pale language like calling him vermin; it hinged on normal-level criticism like that the executive was "greedy" or something, where we don't want to impose too high a burden on discussing greed (in contrast to calling someone vermin, where I think it's fine if we impose a high burden).
But I'm not even sure it hinged on the word "greedy". I think there's a "rational" case (as in, one you could make without using any emotional terms whatsoever) that insurance companies are killing many people, and it's not obvious to me that Mangione was influenced more by the use of various emotional words than by hearing that case and finding it plausible.
Hard to prove either way, but his manifesto was pretty emotionally charged: he called them parasites who abuse America for profit. Maybe he was primarily convinced by the rational non-emotional case and then came to those emotions after, but the murder is way less likely to happen in the counterfactual where criticism of insurance companies is less emotional.
Also, in your comment you use the phrase that the companies are killing many people; this is very emotionally charged framing and word choice! People that use the term "stochastic terrorism" say that using "kill" when you mean "indirectly lead to the death of" is very objectionable, see musk.
I agree that he didn't kill Thompson because he heard the word "cockroach" too many times or something, but I do think that hearing how CEOs like Thompson are *killing* thousands made it a lot more likely. Killer is a very dehumanizing label.
Hm, I think "killing" is pretty close to emotionally neutral, in the sense that "indirectly led to the death of" is making a conscious decision to be unusually dry and bland. I don't think it's practical to have important conversations at the level where you can't say bad health policy kills people.
Example: do the following count as "killing" or "indirectly leading to the death of"?
1. Hitler (who never actually murdered anyone with his own two hands)
2. A thief who steals someone's life-saving medication to sell it on the street, knowing that they'll die without it
3. A politician who uses lies and fear-mongering to cancel a life-saving program
4. A surgeon who is extraordinarily bad at their job and accidentally removes someone's heart when they were supposed to remove the kidney.
5. A doctor who hates someone and falsely tells them they're fine, removing their chance to cure a curable disease in time, in the hopes that they'll die?
I think this boundary is too fuzzy to worry about it too much.
My point is that use of the word "kill" provokes a stronger emotional reaction than something like "causes deaths". This can be useful to motivate people to care about important stuff more like climate change or health policy, but saying that CEOs are killing people will make violence against ceos more likely. A more dry term instead would make it less likely, and an even more emotionally charged term like "murdering" would make it more likely still. This is agnostic of what we define as neutral.
Perhaps its worth the increased of risk of violence in order to get the stronger emotional reaction, but the tradeoff is there.
Russell conjugation - though there aren't too many positive terms for kill.
neutralize/unalive/kill/murder/slaughter in approximate order of negativity?
I think you have a way better argument than Mr Alexander on this. I think use of 'kill' as a synonym for 'decisions that indirectly increase the chances of deaths among this group' is one of the few, maybe even the best, examples of actual "stochastic terrorism". Even his examples conflate this will instances where an identifiable person can be specifically attributed to another person's death, which is exactly what the use of 'kill' is intended to invoke.
Except in the case of health insurance companies, it's not 'decisions that will indirectly increase the chances of death among this group.' That is circuitous enough to route around not just the emotional punch, but the actual reality as well.
I mean, there *are* cases like that. When a health insurance company decides not to pay for a screening that leads to earlier detection and better outcomes for some possibly-fatal illness, then the above description is apt. But not all decision are like that. There are also decisions that are more like "find underhanded ways to drop people from your policy to avoid having to pay for an expensive treatment that they will definitely die without." That's not 'indirectly increasing the chances,' that's deliberately decided to let some number of (nameless, faceless) strangers die for money.
You can quibble that it's still indirect enough that calling it 'killing' is dishonest and manipulative, but that *is* a quibble, not a serious argument.
> 1. Hitler (who never actually murdered anyone with his own two hands)
Sorry for the nitpicking: I wonder if this is actually true. I suspect his personal bodycount is really low, but I wonder if he ever eg shot anyone in the Great War when he served on the front?
This doesn't impact your point. Just a historical curiosity.
Didn't he shoot himself? So if the bar for "killing" is set very high, I believe this makes him a pretty good person.
After all, he killed Hitler!
> So if the bar for "killing" is set very high, I believe this makes him a pretty good person.
Well, only on the 'number of people personally killed' score. He also did some other things like denied smokers their pleasures without scientific evidence.
Re "vermin" and similar derogatory language: I asked Google/Gemini:
>What is the strongest language Musk and Altman have used in disparaging each other?
and they came up with:
Musk and Altman's strongest language includes Musk calling Altman a "scam" and a "liar," while Altman has claimed Musk's life stems from "insecurity."
Specific exchanges and insults include:
The "Scam" Accusations: Musk labeled Altman "Scam Altman", stating he takes scamming "to a whole new level" and "might literally love scamming more than any human alive!"
The "Space Datacenters" Retort: Altman fired back on X: "homeboy you're the one sellling public market investors on short-term space datacenters."
Musk replied, claiming Altman could end up in prison.
The "Obsession" Jibe: Altman has claimed Musk "can't be a happy person" and that his "whole life is from a position of insecurity," famously dismissing Musk's attacks as proof that he is "obsessed with me again."
Now, _personally,_ I prefer that they stop jabbing at each other and get on with building AGI, but I think these cases merit something more like raised eyebrows rather than high burdens.
Keeping Musk busy on Twitter is what gets AGI built. The engineers need some peace and quiet.
(No opinion on the feud, but 'Scam Altman' is too good to pass on. On par with Muskrat, I guess.)
Many Thanks! LOL! Yeah, from what I've read of his style: The good news is that he (often?) understands the technical issues. The bad news is that it sounds like he micromanages a _lot._
The particicular case that concerns me is an instance where someone says "X person/group is so bad, someone should /do/ something about them." On January 6th, Trump's "will be wild" tweet maintained plausible deniability while clearly signalling violent action against the legislature. Similarly, repeatedly saying "X group of people are sick, disgusting, horrible, child killers, satanic demons, etc." and then shading into "someone should DO something about this" is clearly a dangerous thing to say given all the deranged people that exist and might see this as a call to action. When the commentator understands this dynamic and invokes this language anyway, I posit that it can fairly be called stochastic terrorism.
The thing that immediately popped into mind for me reading this was Radio Télévision Libre des Mille Collines (RTLM) leading up to the Rwandan Genocide which had more or less this exact dynamic (Tutsi's were more or less exclusively referred to as cockroaches or snakes).
Agreed. Also I think stochastic terrorism is an unhelpful term for this particular misdeed, since it implies that speech is violence, and any good liberal should want speech and violence treated differently. “Incitement” is a good term here.
I think the line I draw between incitement and stochastic terrorism is that stochastic terrorism trades off a few points of certainty (ie, it's not guaranteed, and in fact pretty low-probability that anyone will act on any particular statement you make) for a few points of plausible deniability (if/when something does happen, you're safely outside the bounds of conspiring to make it happen, you've behaved recklessly but without a focused intent on a particular action.) What makes this particular act very toxic is that the person authoring the stochastic terror can then calcify/cement the moral position they have and the /rightness/ of their action by appealing to their earlier invective in combination with the high-minded principles articulated by Scott above
So.. the European countries that have laws against severe hate speech are reasonable to do so because it encourages violence if left alone too long.
Are you saying then, that the problem is one with labeling too many things as stochastic terrorism? It sounds like you understand that the concept of hate speech being a causal agent of violence is a real phenomenon.
On some level, my claim is that the cost of the concept (banning criticizing anyone, or unprincipledly carving out a group of people who it's banned to criticize) is greater than the benefits (preventing some very small number of deaths that might have happened anyway) - although I prefer not to use this level because I don't think we should compare costs and benefits this greedily or naively. I don't claim that it is impossible for critical speech to cause violence.
I understand a little better now. But i don’t know if the counterfactual world where this phrase didn’t exist would have much less of these undesirable behaviors. It seems like a rhetorical cudgel of the type you are describing is likely to have emerged in another form.
The other thing was that I’m not sure the cost to society is mostly the small amount of deaths, but as you say it seems difficult to measure other types of violence easily.
Regardless, thanks for clarifying.
I am certainly sympathetic to your point.
Just so, or else Generals wouldn't offer rousing speeches to their troops in order to encourage them to attack their enemies.
I don't think there's any reasonable standard by which pre-battle speeches cause violence. Their purpose is to make the soldiers that hear them more successful at the violence they were already going to do. But by the time a pre-battle speech is given, the chain of events leading to the battle has already mostly unfolded. Maybe a really, especially *bad* pre-battle speech could ruin morale enough to make the soldiers run away without fighting, but that would take some doing.
What things actually cause the violence varies on a case-by-case basis. Usually it's some combination of promises, (maybe implicit) threats and social expectations. I guess you could say most of those are delivered by means of speech, but they're almost all made credible by past non-speech actions, like handing soldiers money, or executing deserters.
The perceived cost of allowing hate speech is not just, or even mostly, that it leads to killing. It's thought to increase tensions between groups in general, and thus hurt the ability to talk and work together.
Ostracizing people for saying things they believe are true and important, also hurts the ability of people to talk and work together. Really, it directly attacks the concept of people talking and working together, because the proponents of cancellation don't want anyone to work with or listen to their enemies.
Supposing I grant that that's true, it comes to a decision of which group is more valuable to work with: the sort of people who engage in hate speech or the sort of people who refuse to work with the sort of people who engage in hate speech?
(History has spoken clearly and repeatedly on this point, though perhaps you haven't paid attention.)
The more broadly you define "hate speech", the more likely it is that the group that includes the hate-speakers is going to make for more productive partners than the people who piously refuse to work with them.
And in the history of my field, I note that NASA managed to put men on the Moon, starting from almost nothing, in eight years - at a time when casual racism on and off the job was still unremarkably in places like Huntsville and Houston. Today's NASA, fully on board with all the political correctness, started with all the records and leftover infrastructure of Apollo and more, and has taken fifteen years to send a group of astronauts to wave out the window as they fly by the moon, with no clear plan for actually landing on it other than to hope that Elon Musk can do that part for them.
Now, I'm pretty sure the casual racism wasn't necessary or even helpful. What was necessary, and vital, was a team of people who could work together on the one thing they *did* agree on, even when they disagreed quite strongly on other matters. That's true in any great endeavour, and most mediocre ones.
If you have bad things to say about your coworkers, or about broad classes of people that include your coworkers, do it on your own time and keep it out of the workplace. If your coworkers have bad things to say about you, but they keep it out of the workplace, then return the favor. Do both of those, and you may be part of something great. Or at least mediocre but lucrative.
You can argue that the cost of criminalizing the act is greater than the benefits without going so far as to denounce the entire idea of the concept as something worth naming, talking about, or criticizing people for doing.
Maybe this is too hyper-specific an example, but I took the lesson from it to heart.
(Most of this is from “The Politics of School Desegregation” by Robert L. Crain.)
The desegregation of Southern school districts in the 1960s, across many individual cases, went one of two ways. Either desegregation went into effect peacefully, leading to integrated public schools, or it met resistance, often violent, and ended up with the public schools being 100% Black, and private “segregation academies” being 100% white.
What accounted for the different outcomes in different school districts? The obvious demographic variables didn’t explain it.
What mattered was the attitude of local elites, including local political, religious, and business leaders.
Where those leaders were saying “we don’t like this, but it’s the law now,” and urged calm and compliance, things were peaceful. Where the leaders were denouncing federal interference, making racist statements, etc., violence and failure of desegregation followed.
Of course not every local leader was on the same page. I imagine a lot them held silent when their colleagues preached defiance.
Based on this and other historical examples, I took it as the moral duty of a political figure to promote tolerance for minority groups.
Being a small-time political figure myself (I have held local elected office for cumulatively almost three decades), I strive to live by this.
That said, each constituency I have represented has been predominantly liberal, and things like actively defending the gay community during the AIDS epidemic, or the Muslim community after 9/11, has never cost me anything politically.
Did the studies control for reverse causality? It seems plausible to me that more violent and/or racist districts would naturally have more violent and/or racist politicians representing them.
The problem is that the activities of the "gay community" really did contribute to the severity of the AIDS epidemic. Perhaps if you had been less supportive fewer of them would have died.
Really, the problem genetically speaking is that AIDS and the gay community "accidentally on purpose"* genocided hemophiliacs.
*On purpose in that they wanted money, and were paid to donate blood... a lot. Accidentally, in that they didn't realize they were infected, or that they'd be hurting anyone.
It seems like the main difference is the *direction* of the reaction (urging cooperation versus urging resistance), not whether the language used to urge resistance was rational versus emotional.
>I imagine a lot them held silent when their colleagues preached defiance.
Like public health officials circa June 2020.
>promote tolerance for minority groups.
A dangerous and slippery standard.
>Based on this and other historical examples, I took it as the moral duty of a political figure to promote tolerance for minority groups.
Ask the people of Rotherham what they think about that.
This is an interesting counterpoint to "Weak Men are Superweapons" (from the SSC era).
Not a contradiction, because this is specifically about the assertion that superweapon-ish claims can be blamed for violence. But the superweapon discussion had overtones of that (see: the "superweapon" metaphor and the examples given for it), so it deserves explication. "Stochastic terrorism" would be an effective and intuitive accusation to make against would-be superweapon-builders; what does the "liberal solution" offer in its place?
I don't think the problem with those "superweapons" is that someone could die from them, the problem is that they're false and unfair. I am certainly not claiming in this post that there's nothing wrong with saying false and unfair things.
I think the point of "stochastic terrorism" is that certain kinds of false and/or unfair statements are especially likely to incite violence, and therefore especially bad.
I think that "X is evil" is absolutely legitimate, as is rephrasing it in even stronger and more inflammatory language, if it's true, but that generally people use stronger-than-ideal language to condemn those who make them angry.
I think that "We should use force against evil people" is generally not OK, and should be condemned, but in itself it's probably within the bounds of speech a free society should tolerate.
I think that /putting those two ideas together/ should be treated as much worse than the sum of the parts.
Saying "X is a traitor, and we should kill traitors" is a bigger deal than saying "X is a traitor" and then, separately, in a different context where no-one can work out which traitor you're talking about "We should kill traitors".
And you can go through way more than two steps. Revolutions against unjust governments are justified and virtuous (right there in the declaration of independence iirc). Revolutions must accept the need to, if resisted, kill those using a monopoly on power to enforce the status quo. Unjust governments limit the rights of people to access and influence political systems. X policy you dislike (whether letting non-citizen residents vote, or NOT letting non-citizen residents vote) denies people their right to self determination. Police are agents of the government's execution of its monopoly on violence over potentially unjust laws. Ergo if you dislike some voting registration laws you miiiight wanna start killing cops in the name of overthrowing the government.
I'm one of those people who, while liking the idea of liberalism a lot, is pretty fuzzy on how I feel about trying to kill extra-legally murder Very Bad People. I have a hard time condemning it as a complete blanket statement, but I also don't particularly trust myself to know where the justifiable exceptions would be (if there are any). Of course I don't exactly trust others to do so either... but it also feels like a disagreement of judgemental calls too much for me to outright condemn someone for (trying to) kill someone causing (or having caused) pain or death to potentially millions.
But the problem is you could define anyone voting for any candidate with a policy I think will cause widespread harm as "trying to hurt-or-kill millions."
>I think that /putting those two ideas together/ should be treated as much worse than the sum of the parts.
But the whole point of it being stochastic is that no one person has to put them together explicitly for the killer to to hear them both and integrate them.
Like, if there are two posts on a forum one right after another, the first one saying 'X is evil', the next one saying 'We should kill evil people'. Does it really matter whether those two posts are by the same anonymous author, or two different ones? The person reading it receives the same communication either way.
Because of the weird way humans are irrational, I think it does make a difference.
If I see someone saying "we should kill evil people", my assumption is that they're probably just blowing smoke. If I see someone say "We should kill evil X", then ... to be honest, my assumption is still that they're probably just blowing smoke, but I'm less confident of that.
And I think that seeing someone else who I believe actually supports murder makes it more likely that I'll see murder as a normalised thing and commit it than seeing someone else just blowing smoke about it.
https://www.astralcodexten.com/p/open-thread-442/comment/295894454
But even if he tells a bunch of outrageous lies, please don’t try to harm him.
I'm not sure what you mean by this comment.
I linked to a comment I wrote on the OT about what Trump might say at 9:00 PM eastern tonight in a national address.
I guess it was kind of a lame joke. I don’t think the comment will generate any stochastic terror.
Thanks for the notification, I hadn't known that Trump was speaking tonight. Now, having watched his speech - umm, cybersecurity has long been a problem. If the PRC managed to hack vote totals, that is unfortunate, but wouldn't be all _that_ surprising... Now, I'd have been upset if Trump had said that, because of vulnerabilities, the midterms will be delayed or cancelled - but he _DIDN'T_ do that.
What is the US aversion to pen and paper ballots? (And don't say count duration as that is easily scalable as every other democratic country regularly demonstrates).
Many Thanks! We had a bad experience with 'hanging chads' back in 2000... Other than that, I don't know.
That issue was specific to a machine-read punchcard ballot. Not an issue with "Make an X in the box" type ballots.
It’s complicated by the fact that the US Constitution leaves administration of elections to each state.
True, Many Thanks!
Many Thanks! Agreed on that specific issue. A smudged "X" that may or may not be an intentional mark could be an analogous problem with pen and paper.
And naturally, I voted on the pre-punched “hanging Chad” ballots for the first twenty years of my voting and we never once had an issue.
<snarky cynical comment here>
No, by Trump standards it was just another airing of the grievances. He was teasing something *big*. He might have at least nominated Meat Loaf to be Secretary of Power Ballads or something.
Many Thanks! Yeah, a lot of it sounded like relitigating 2020 ( _WHY_ ??? He _is_ POTUS now - and he never got a single one of his 2020 cases accepted, even by GOP judges). On the other hand, if Mythos found that our electronic vote tabulation has even more holes than we expected, it is worth plugging them...
It seems like the concept has been watered down by overuse. My understanding was it originally hinged on a kind of legal loopholism like "somebody should kill these people" and relied on walking the line on incitement of violence, and was applied mostly to people with broadcast style platforms. It seems like it's useful to have a term for the concept of "creating social change by indirectly encouraging others to do violence", so that we can all disapprove of it. But then once that concept exists it becomes part of the narrative battlefield itself
This post claims to find the actual origins of the phrase:
https://drrollergator.substack.com/p/stochastic-terrorism-a-game-of-rhetorical
Rather than originating from the world of law, it was some DailyKos blogger.
I did not intend to say that it came from the world of law
Brian Thompson, not Brian Johnson?
Thank you.
Thanks to his new biohacking regime, Bryan Johnson is in fact immune to bullets.
Stochastic Terrorism seems to be another speech-stifling concept.
The more interesting question for me is how difficult it is to justify violence starting from the beliefs of an ideology. For instance, combing rationalist utilitarian decision-making with the ability to arbitrarily assign probabilities to events(p(doom),p(doom|attacking Altman)) can easily justify violence towards Sam Altman.
Act utilitarianism has been out of style since 1850. Rule utilitarianism (in some form or variant or level of thinking) is where it's at, and I think the vast majority of rationalists are convinced of this.
I guess this is saying that the damage to a social norm against political violence is greater than the benefits of decreasing p(doom). A rationalist with sufficient p(doom): p(doom|political violence) ratio would disagree with this, especially with how weak our norm against political violence already is.
Personally, I would emphasize that probability models are not equally effective as a way to resolve this.
I think act utilitarianism is better than rule utilitarianism.
And I think the problem come from mixing ethics with decision theory : ethics describe our moral preferences, and decision theory describe how to act given our preferences and beliefs.
A common argument against act utilitarianism, or even utilitarianism as a whole, is that it would lead to bad consequences to follow it.
And rule utilitarianism try to fix it by redefining what we want, instead of how we should act given what we want.
It is like, being faced with Newcomb’s paradox, answering with "hey, actually this proves what we wanted wasn't more money, but following the rule that would lead to more money !"
This is missing the point and fixing problem at the wrong place.
In utilitarianism, and in Newcomb's paradox, you should take into account the very act of following a rule instead of another, which would probably lead to something like the FDT when generalized.
You don’t need a revision of the ethical framework nor of the goal of maximizing money in the paradox.
Like a lot of internet discourse, I think most examples of people claiming stochastic terrorism is inaccurate at best and incoherent at worst. That said - I think there’s a core concept that’s worth preserving, although the (mis)usage has become so widespread that it might be worth using different language.
In my mind, stochastic terrorism is a confluence of three factors:
1. You’re attempting to lower the status of a person or group to the point that a non-zero number of people who agree with you might consider violence against them
2. You’re attempting to increase variance in opinion against that group to increase the number of people predisposed to violence while maintaining a large population of people opposed to violence (for plausible deniability)
3. You’re attempting to erode norms against violence as a political tool, or at least bias listeners towards violence relative to other approaches.
The conjunction feels important to me because it takes all three to incite violence without being held accountable for it. 1 & 3 is just an obvious call to violence (or toothless, depending on how strongly worded your speech). 2 & 3 is undirected - you might stir up controversy and erode beneficial norms, but it’s not targeted. And 1 & 2 is really no different than any other attempt to shift public opinion.
But when you have all three it’s easy to claim that you’re not inciting violence against a group - after all, most people considering violence would never act on what you said - while simultaneously substantially increasing fringe violence against your target.
Perhaps you could give examples? It's unclear to me what kind of speech you think would clear all three bars (and would need a term like "stochastic terrorism" rather than plain old incitement of violence)
I haven’t read the essay, but isn’t the title just so humorous? “Against terrorism”; you might as well follow it up with an essay titled, “Why war is actually sort of bad”
Notice the quotes - it's against the *concept* of "stochastic terrorism*.
I am, coincidentally, against terrorism.
Having read the argument, I totally understand what you mean now.
Still, an interesting thesis. Would you say that stochastic terror *doesn't exist* (Isn't terror / shouldn't be considered terror) or that it *shouldn't exist* (shouldn't be taken seriously as a concept)?
The US already has extensive, well-thought-out legal theory and precedent specifying how to deal with this problem: defamation law, at the federal and state level. US First Amendment law has been very carefully thought out and rigorously argued many times over the centuries; if you aren't an expert on it, you shouldn't try to fix it in a blog, let alone a blog that doesn't mention the First Amendment.
We have the problems we have not because there's anything wrong with the law, but because it isn't applied. If you say someone is "literally Hitler", and they are not literally Hitler, and they can show you knew either knew they were not Hitler, or acted with reckless abandon (e.g., not bothering to check whether they were Hitler), or with malicious intent, and caused them material or emotional harm, then US law generally says you have defamed them and are liable for damages.
Then people say, "Oh, but so many people do it!"
News flash: When San Francisco stopped prosecuting shoplifting, lots more people started doing it.
Don't feel sorry for someone who called someone else "Hitler", especially if they literally used the word "literally". If they're adults, they will soon figure out that actions have consequences--but only if actions really DO have the legal consequences specified in our law.
The situation we're in today is that, at least according to every human and chatbot I've consulted, judges usually dismiss open-and-shut clear cases of libel if they're considered common behavior. That's not how law works. Worse, this endemic disdain for the law, on "practical" grounds, teaches, guess what, disdain for the law.
I don't think stochastic terrorism and defamation law really connect.
I think a typical example of what would be deemed stochastic terrorism would be something like "Muslims are going to take over America and implement Sharia law and maybe rape your daughter". This is a prediction rather than a factual statement about the present, and it seems like many people who make predictions like these believe them. I think that means a defamation case would be quickly thrown out. Also, would anyone even have standing to sue for defamation in a case like that?
Yes. Even in the case of Nazi name-calling, is an accusation that someone "has similarities to Hitler" (almost certainly true about many public figures, especially in the cases of politicians doing authoritarian things) that much better than "literally Hitler".
I think the US law that most directly engages with the question is incitement. The bar for this is extremely high: the controlling precedent (the Brandenburg Test) requires finding that the speech is intended to produce "imminent lawless action" and also is likely to actually do so. Without intent, without likelihood, or without imminence, it's protected speech.
Besides this, I'm also familiar with at least one case that found civil liability partially on the basis of an "aiding and abetting" theory in a wrongful death lawsuit. The book "Hit Man: A Technical Manual for Independent Contractors" was marketed as nonfiction, a guide for how to take up contract killing as a career. After someone was convicted of murder after actually using the book as such, the family of two of the victims sued the book's publisher on the grounds that the book encouraged and enabled the murder. The trial court initially dismissed the case for failing to meet the Brandenberg Test, but an appeals court reversed this 3-0 on a combination of the "speech-act doctrine" (that speech can be judged as an action rather than advocacy, such as in cases where people have been criminal convicted for giving seminars where they coach people on committing tax fraud) and on an argument that the trial court was wrong and the Brandenberg Test actually had been satisfied.
https://law.justia.com/cases/federal/appellate-courts/F3/128/233/525203/
You're right that they aren't seen today as connecting, and that this hinges on having the standing to sue. I totally overlooked that when posting, which is what I get for posting with a migraine.
But a strict interpretation could say that, if I say something which logically implies that you are probably an X, then I can be held responsible for saying that you are an X. If you said "Jews drink the blood of Christian children" in a way that implied that most Jews drink the blood of Christian children, and I were a Jew, I think it would be just that I then have standing to sue you. Think how many pogroms this might have stopped.
(The hard part would then be proving damages. I'd probably have to make a class action suit.)
Hate speech is just disapproval expressed without quantifiers. And I'm not joking when I say that what we need most to heal the US today is to teach people to speak with quantifiers.
IIUC most relevant U.S. case law specifically forecloses this interpretation; a statement about a group of more than ~25 people can't be defamatory.
That's a shame.
Why?
Maybe not. I think I formed my opinion under the false impression that hate speech prosecutions were happening in the US, when what is happening is hate speech prosecutions in Canada and Europe and the UK, and hate speech non-governmental retributions in the US. My reasoning was that the robust body of law on free speech should be applied to hate speech prosecutions, but in fact US free speech law is being interpreted as making hate speech laws unconstitutional.
"Hogwarts students vandalize my garden."
You're a Hogwarts student; am I accusing you personally of something?
What if there really is a group of Slytherins who like regularly pulling pranks in my conveniently-located garden, so it's a completely true statement and I'm not trying to imply that everyone else is joining in?
When someone says "Hogwarts students vandalize my garden", it's clear the speaker can't mean "Most Hogwarts students vandalize my garden", because you probably couldn't even fit half of them in the garden. When someone says, "Jews murder Christian children", they're referring to a specific legend which says all Jews are required by their religion to regularly murder Christian children, to use their blood in the passover bread.
If the same legal reasoning about the context of a statement used in defamation law, were used in hate speech law, then statements like "The Quran tells Muslims it is their destiny and their duty to conquer all the earth" couldn't be regarded as hate speech (since this is a verifiable fact, not obviously made with malice, and doesn't assert that most Muslims want or intend to do so). Statements like "Hogwarts students vandalize my garden" would be interpreted as a statement at most as strong and inclusive as"Rabbits eat my carrots", which clearly doesn't mean "All rabbits eat my carrots" (although it is probably true that all rabbits would like to eat the person's carrots).
There are rulings in defamation law about how to interpret natural-language sentences according to context. But they aren't consistently applied in hate speech prosecutions. Creating a separate category of hate speech re-opens all the loopholes that have been closed in defamation law in the 235 years since 1791.
I think in Europe it's pretty clear - if you speak, and the government hates it, then it's hate speech.
Well they generally only arrest pro Palestinian activists. The rest just get a visit.
That's of course not true. Moreover, in UK there are about 10000 arrests for speech violations per year. There is no conceivable way of verifying all of these people are "pro Palestinian activists" - officials do not publish any such statistics, of course, and any research that could survey so many people would be nothing short of herculean, and for sure would be known just be the sheer amount of effort. It does not exist. Thus, not only I can know it is not true, I can be sure you know it is not true - or exhibit what is called "reckless disregard" for whether it is true or not.
Here are examples:
https://www.bbc.com/news/articles/c07p7v2nn8mo
https://www.bbc.com/news/articles/cp3wkzgpjxvo
https://en.wikipedia.org/wiki/Twitter_joke_trial
https://archive.ph/P2Klv
https://x.com/grok/status/2011846800999645330
None of these are "pro Palestinian" activists, as far as I can see. None of these require more than 5 minutes of internet search to discover. My question is: why did you claim something that is so obviously untrue?
The 10,000 includes visits for harassment, calls to violence and so on. Still one visit or arrest for bad opinions is one two many. I side with Mr Linehan. The classification of an activist group as terrorist is fairly unique and the arrest of supporters who protested said classification were also arrested. “Since the UK proscribed Palestine Action as a terrorist organisation in 2025, hundreds of supporters have been arrested for allegedly expressing support for the banned organisation at demonstrations”.
Either you like free speech or you don’t.
Merz' complaints led to warrants for around 5000 people, some of whom were only quoting other people (the people being quoted, oddly, were not arrested). These were folks who had said bad things about Merz. Nothing to do with Palestine.
Source: https://old.reddit.com/r/ShitAmericansSay/top/?sort=top&t=all ?
See https://www.youtube.com/watch?v=tB3WVygAM8I , it includes an explanation why the "10000 arrests for speech violation per year" does not mean what Joe Rogan thinks it means.
I'm sorry, I'm not going to watch half-hour video from a "comedian" trying to gaslight me into the idea that UK has freedom of speech and US doesn't, despite all available evidence. If there's an argument behind it, you are welcome to make it, it should't take half an hour to read. I have no idea what Joe Rogan means by anything, neither I care too much, but I can see UK laws criminalizing speech, and I see them actively and vigorously used to police political speech in UK, and - so far comically unsuccessfully - in the US.
Fair point. The summary is that the number that people typically quote is a total number of people who get into trouble related to the "Communication Law", but a majority of that is unrelated to saying mean things on twitter. It includes things like harassing your ex by calling them on the phone many times a day. When you get to the actual "saying mean things on Twitter" (now the numbers are much smaller, comparable to USA), a typical "mean thing" that gets prosecuted is something like calling people to burn down a mosque, which would get a similar legal response in USA too. (Specific examples in USA are provided in the video.) And yes, some obvious mistakes happen too, but again their number is comparable to similar mistakes made in USA.
I find this argument disagreeing with the facts. US and UK have very different standards of what considered unprotected speech, in the US it must be incitement to an imminent lawless action (not just generic "mosques need to be burned" but a call for burning of a specific mosque that a reasonable person would conclude may immediately lead to the actual burning). Even that kind of speech is prosecuted quite rarely. Speech that is considered merely "offensive" and regularly prosecuted in the UK is never prosecuted in the US (at least not criminally, you could get fired or "cancelled" for it, but that's a different thing). Of course, since you did not provide any specific numbers or examples I can't give any more factual response, but let's consider this.
UK clocks about 10K "offensive communication" arrests per year now. US population is 5 times larger. So, if US had similar levels of enforcement, we'd have 50K speech arrests in the US in a year. Can you show that happens? Can you even show 5K speech arrests per year in the US? I suspect unless you really stretch it - like list violent riots as "speech" - you may have trouble getting even to 500. With that in mind, arguing about specific cases, being a mistake or exception or whatever, is pointless. 10K cases per year can't all be mistakes, it's a policy.
No, "John Smith is Hitler" is clearly a metaphor, and any sensible court would see it as that.
Same thing for "John Smith is literally Hitler" - "literally" is being used as an intensifier here; that usage is literally destroying the worth of the word, but it exists and any sensible court would recognize it.
Literally is fine there. I will literally die on this hill.
It’s an intensifier and the whole sentence is hyperbole, hence the use of literally is figurative, since the entire sentence is figurative. It’s not its own contronym, you wouldn’t use figuratively there.
Let's not argue over Hitler. I've had dozens of entirely baseless, damaging, and easily-disproved slanders made against me in Internet-print, sometimes along with frank admissions that they were malicious, but have been advised by everyone I consulted not to sue, because (I was told) the case would be dismissed, and I might be hit with some anti-SLAPP law forcing me to pay the defendant's law fees for starting a "frivolous" lawsuit. All because most political slander gets a free pass nowadays. This is a vicious circle.
I'm sorry. The court systems are broken all around.
I am not a lawyer, but every lawyer I've seen online has described the legal bar for defamation as much higher than you set it here. It's not that prosecutors or plaintiffs are declining to go after people, it's that when they do, courts recognize that "literally Hitler" is obvious hyperbole and it will not cause people to believe that you are literally a reincarnation of history's greatest villain.
>obvious hyperbole
See the case of Carl Sagan v. Apple Computer (1994), where a US District Court ruled that Apple calling Sagan a "Butt-Head" was not libelous.
>There can be no question that the use of the figurative term "Butt-Head" negates the impression that Defendant was seriously implying an assertion of fact. It strains reason to conclude that Defendant was attempting to criticize Plaintiff's reputation or competency as an astronomer. One does not seriously attack the expertise of a scientist using the undefined phrase "butt-head." Thus, the figurative language militates against implying an assertion of fact. Unelko, 912 F.2d at 1054.
....
>Because a reasonable factfinder could not conclude that "Butt-Head Astronomer" implied that Plaintiff was a less than able astronomer or that Plaintiff was legally wrong in asking Defendant to cease using Plaintiff's name, the only remaining assertion is the bare statement that Plaintiff is a "Butt-Head Astronomer." Clearly this phrase cannot rest on a core of objective evidence. Plaintiff does not suggest any other assertions of objective fact that could be reasonably implied from the phrase.
http://law.justia.com/cases/federal/district-courts/FSupp/874/1072/1478136/
(Also not a lawyer, this is admittedly speculation)
I suspect this may also be because most well-known defamation cases involve people who are various degrees of "famous" and have consciously chosen to live in the public eye. At which point, they've opened themselves up to increased criticism in exchange for greater influence.
In the cancelation discourse, I have much less sympathy for celebrities/politicians who take blows because they often have comparable social capital with which they can fight accusations. Whereas if an otherwise private individual has the misfortune of saying something careless, then goes viral, it would have much greater potential to ruin their chances for employment, social standing, etc. Hence, it would presumably be much easier for them to prove damages in a court of law.
Granted, now that anyone can post online without getting vetted by a media outlet, the distinction between a public/private figue may have collapsed...
Under U.S. law, public figures have to meet a higher bar in establishing *culpability* for defamation (i.e., whether the speaker should have known they were spreading lies), but the standard for whether a statement counts as defamatory in the first place is the same; it has to be a statement of objective fact, as interpreted in context by a reasonable member of the audience where it was shared, and that statement has to be false. The law is very specifically *not* designed to provide a remedy for heated rhetoric targeting someone (that doesn't cross the line into incitement, which is likewise a high bar).
Great post.
Not sure I agree European hate speech really is such a cleanly demarcated legal category. Haven't been following super closely but my sense is that there's a broad, vague range of critical-of-groups or critical-of-individuals speech in the UK or Germany that is legally risky, and without some limiting principle that makes it easy for people to speak freely so long as they make sure to do so "non hatefully."
As a German, hard disagree for Germany. It is very clear-cut what is allowed and forbidden. I don't find all rules good. In fact, I am strongly opposed to the new law that allows punishing defamatory attacks on politicians. (There is some hope that it will be cancelled again.) But also there, the scope of the law is very clear.
Yes, there is always a grey zone, but it's a pretty small one. You hear about it because no matter how tiny the grey zone is, people intentionally go for this zone. For example, the slogan "Alles für Deutschland" (all for Germany) is forbidden as a Nazi parole. The right-wing AfD party with party leader Alice Weidel pushed the slogan "Alice für Deutschland". The words "Alice" and "Alles" are pronounced basically the same. Since it's often shouted, does pronounciation or spelling count here?
But someone who doesn't intentionally go for the grey zone will not accidentally end up there.
Thanks! I'm sure it is the law against insulting politicians that i was vaguely remembering. People getting arrested for stuff like satirically representing a politician as saying "i hate freedom of speech." Is that in the grey zone or is it clearly prohibited? I'm not sure which would be worse:
https://www.fire.org/news/blogs/free-speech-dispatch/i-hate-freedom-opinion-meme-leads-sentencing-german-court
Ok, I admit that this is an awkward case for my claim that there is no big grey zone. The judge in first instance said that the picture was so realistic that readers would not recognize it to be photoshopped, thus falsely believing that the politician actually showed this sign for real.
Part of the story is also that the decision was revised, and in the revision both prosecutor and defendants motioned for acquittal, which the judge followed. The decision of the first instance was also condemned by politicians from left to right. I would rather file this as "in rare cases first-instance decisions can be odd". But yes, I guess it shows that they grey zone is larger than I made it appear.
But what all sides, including the judge in the first instance, agree on: if the audience recognizes that it is satire and not a real photo, then this is allowed as freedom of opinion.
People are upset that Merz sicced goons (okay cops) on them for quoting Other People saying nasty things about Merz, and did not actually sic the cops on the people making inflammatory statements.
This seems to suggest that even discussing someone saying something nasty about a politician might get you arrested. That is... not "clear-cut" in a way that encourages discourse.
(If your response to this is something along the lines of "Merz is gone, and nobody else is that easily trollable", that would be relieving to hear!)
I am not sure what you are referring to with "Merz sicced goons". There is one case where someone got a fine for calling Merz "Lügenfritz" (Lügen = lies, Fritz = short form of Friedrich). The case never went to court hearing since the person paid the fine, and Merz was not involved in this case. The case started without him having filed a complaint or so.
There are three things which are prohibited: defamation, slander, and personal insult. All three were already forbidden before and could be pursuit in civil courts, but the new laws makes them penal offenses if they are committed against politicians, so now they can be prosecuted regardless of whether the politician files a complaint.
Essentially, there are two things that you are not allowed to do:
- making factual claims about politicians that are provably wrong, if you are aware that they are wrong or just making them up. (And if they are supposed to make the politician appear bad.) It is forbidden to just claim "Merz has taken a 10.000€ bribe" without indication that it might be true, or claim "Merz has voted to make abortion illegal" if this is provably false.
- Making insults that go against the person and that are not related to political acts. Just saying Merz is an "idiot" or a "liar" without any context to his political actions is actually forbidden. But saying that "he treats the poor like dirt" or "his tax reform is idiotic" is totally fine. A guideline is that if you criticize how a politician uses his power, that is always permitted. But if you criticize the core of the person, not his actions, then you are treading dangerous legal ground.
Let me retreat a bit from saying that there is no grey zone, because insult and criticism can be mixed. If you say "Merz is an idiot, this tax reforms goes against the wrong people", that's still ok, because the insult is given in context. But if you say without further context "Merz is a rotten bag of scum with his politics", then he insult is heavier and the political reference very vague, that's probably forbidden. But the problem is not the grey zone. It's not exactly difficult to stay out of that, and the only limit on how you can criticize politics is that you are not allowed to a) make fabricated or provably wrong factual claims and b) get personal.
I want to emphasize again that I do not endorse the new rule that makes those things penal offenses. But I do endorse that these things are forbidden and can be persecuted in civil courts, as it always used to be in Germany.
https://dailytelegraph.co.nz/world/merz-under-fire-german-chancellors-5k-complaints-against-critics-coincide-with-record-public-discontent/
Thousands of complaints. Many of them personally signed by him (others he hired a lawfirm to make the complaints for him). Haubeck has similarly large numbers of complaints (800 complaints).
https://www.persuasion.community/p/europe-really-is-jailing-people-for
I do not have the link at the moment, but if I find it, I will share it (I will look). I remember reading (I think it was on substack) about Merz hauling a guy off to jail (presumably released afterwards) for "quoting others saying bad words against Merz" (the others were not sent to jail).
https://europeanconservative.com/articles/commentary/the-merz-affair-and-germanys-war-on-free-speech/
Penal offenses seem to make "cops at your door" (aka something similar to the Swatting pranks) here to steal your computer (see Steve Jackson Games), your cellphone, etc a lot more likely.
https://europeanconservative.com/articles/news/germany-police-house-searches-insults-against-politicians/
It doesn't bode well for a discussion that you call the confiscations of phones/laptops "stealing". I also don't like that US border control can confiscate my phone when I enter the US, but calling it "stealing" does not help with the discussion. So after this post I will disengage and leave the last word to you.
Yes, some politicians do file complaints, and some of them file many. This included Merz before he became chancellor. Since then, according to his own statement in a recent interview, he no longer files them. He added that he would be open to reforming this paragraph, which is why I have some hope about it. At the moment, the suggestion (not from Merz, but from the conference of Ministers of Justice) is to restrict it to communal politicians, which were the main group intended to be protected by the new law in the first place.
I totally agree that confiscating a laptop or a phone is out of proportion for such cases. According to your own sources, this is also how German courts ruled about these confiscations, so I assume they no longer happen. This was one of the many reasons why the law is bad, but this particular one has apparently been fixed.
All of this is not about the original point whether the law is clear, but about the orthogonal point whether the law is good or bad.
Your first distinction ("Most societies distinguish between simple criticism and true incitement to violence") seems pretty concerning to me.
As a thought experiment, which person is worse:
1. The guy that tells Gandhi "Kill so and so, here's a Glock, he's at such and such place".
2. The guy that tells his dimwitted sycophant "Trump does so many bad things, we'd be better off without him"
It seems obvious to me (maybe not to you?) the second is worse, because the issue isn't the language you use - it's the chance that your speech will cause the listener to do the violent action.
If you agree to that, the difference between stochastic terrorism and incitement is just a matter of degrees, and one can be worse than others in different circumstances, and there's no obvious line that is crossed to move from ok to not ok. Falling back to "US courts distinguish between them" isn't very convincing.
The point isn't rule-lawyering. The point is recognizing "if some speech is likely to provoke violence against your political opponent, and you know it and still speak in this particular way, and especially if you clearly express the idea that you are fine with that kind of violence, you should bear societal consequences, even if the law can't get you". The details can be rule-lawyered, just as lawyers earn their living hashing out the details of criminal and civil code. We're fine with that, as long as we have the code, and we agree on the fundamentals like "do not murder, do not steal" etc. Here I still don't see agreement on the fundamentals.
It's not about rule lawyering, I agree. My point is that Scott wants to differentiate between criticism and incitement. My point is that no such distinction exists. They're both on the spectrum of speech that may cause a person to a violent action (and I'm using the example of a case where criticism can actually be more of a cause than incitement)
Yes, there is a spectrum, and there will be very debatable zones in this spectrum. But we know there is speech that is ok, e.g. saying "N is a bad mayor and should not be re-elected" is not seriously taken as a call to violence by anybody. There is also speech like "I will pay the bounty of $100K to anybody who murders mayor N" - which is already illegal. And I think there should be kind of speech like "whoever murders mayor N will be my personal hero, every decent person should think about murdering mayor N, and I rejoice every day thinking this may be the day mayor N is finally murdered" which should not be persecuted by law but sanctioned by the society. Yes, there won't be hard boundaries, and yes, there would be a lot of bickering in the gray areas. But talking about how we bicker about grey areas only useful once we establish that at least boundaries exist. I think we haven't established that, and, moreover, Scott is trying to dis-establish that.
See Erica Rall's comment https://www.astralcodexten.com/p/against-stochastic-terrorism/comment/296075263 where she details the legal test for incitement.
I think the better option is to push back against speech which fails the Brandenburg test for incitement but has some elements of it by calling the speech irresponsible.
(the following is my understanding but IANAL)
under US law, it's only incitement if both (a) there was intent to produce imminent lawless action and (b) it was likely to actually achieve that goal.
it sounds to me like you think example (1) would be incitement but not stochastic terrorism, but i think it could not be incitement, because the likelihood of a violent result would be negligible.
incitement is a categorically stricter concept than stochastic terrorism, because it requires not only a likelihood of lawless action, but also that the lawless action be likely to happen (more or less) right away, rather than next week or next year
I disagree. 1 seems worse.
Your view leads to the conclusion that important political topics can not be openly and honestly discussed in the presence of dimwitted sycophants. It could conceivably be true that Trump (or any other political figure) does so many bad things that we would be better off without him. The fact that it could be true means that it is important that people should be able to say it. On the contrary, there is no socially beneficial activity that requires one to provide someone firearms and instructions to murder.
And here's an example:
https://x.com/benryanwriter/status/2069577731457634329
> Popehat, aka attorney Ken White, has been permanently banned from Bluesky, apparently for a post in which he suggested there was a moral upside to Elon Musk’s death.
I am against prosecuting such people, and I am even against banning them from social media (though I couldn't care less what is happening on bluesky specifically). But I am for treating people like Ken White as revolting moral imbeciles, and shaming them mercilessly. I do not insist on calling him "stochastic terrorist" or any other particular term, I an fine with any approach that gets the result of people like him not being accepted by society as normal. And so far I didn't hear any convincing argument as to why I shouldn't see him as morally revolting.
He has been suspended for four days, apparently. In his follow-up post ( https://www.popehat.com/p/a-bit-of-tedious-drama-at-bluesky ) he engages in a lot of the-other-side-did-it-too, both specifically for Musk and generally for Trump supporters.
I guess the question is, if you kinda-sorta encourage people to kill your political opponents, like Musk did, does that make it fair game for other people to kinda-sorta encourage killing you? With actual killing, I think most people would say it does, if the state cannot be relied on to handle the situation for some reason. With vague verbal references about killing, since they (contra the stochastic terrorism talking point) don't actually have much effect, mostly you'll just end up with a very polluted public discourse.
> I guess the question is, if you kinda-sorta encourage people to kill your political opponents, like Musk did
He didn't though. While Ken White most certainly did.
> since they (contra the stochastic terrorism talking point) don't actually have much effect
They do though. Remember any recent figures on the right who have been murdered? That's the effect.
The retweet quoted in the popehat article is a pretty clear-cut example.
A retweet about the movie? Please. While Ken White, in the same article, explicitly says he meant every word, so there's no misunderstanding what he is calling for. He is calling for Elon Musk to be murdered, because he does not like his political speech. He is not ashamed of it and he specifically explains that is exactly what he is inciting. If there's ever a clear-cut case this is it. And he knows it's a real possibility - that's exactly why he is doing it.
This part of the post is looking pretty accurate:
> Now, I think people of good faith can disagree about the morality or utility of wishing other human beings dead. [...] The loudest cries of outrage are from people who will diagnose you with Trump Derangement Syndrome if you object to the ocean of blood I just described. The reaction is largely contrived, mostly in bad faith, and rarely to be taken seriously. The people landing hardest on the fainting couches are in two groups: pro-Trump people who are thrilled that we are extrajudicially executing fishermen in the Gulf, and professional grifters who don’t necessary like the extrajudicial killings but whose entire gimmick is “aren’t those leftists silly and outrageous.”
Isn't the obvious solution "you should be able to criticize people harshly, but not imply or outright state that they deserve death, or that they should be killed or otherwise harmed"?
The silly accusation of stochastic terrorism in the post aside (it's a dumb Twitter post with a few hundred likes, many such cases! I wouldn't consider it representative unless it has some decent traction), most of what I see credibly labeled as such are calls for the target to actually to be terrorized; for example, Hasan Piker posting DIY diagrams of guns in response to politicians he doesn't like (https://imgur.com/jbiMLSN), or outright saying people should kill certain politicians if they don't like their policies (https://wjla.com/news/nation-world/turkish-twitch-streamer-banned-after-urging-viewers-to-kill-florida-senator).
We could taboo the words "stochastic terrorism" to describe such behavior, but I think we could all agree this is bad and likely contributes to rising political tensions and/or violence.
I think if you look at https://x.com/search?q=%22stochastic%20terrorism%22&src=typed_query&f=live (and where better to get a pulse on how people are using phrases?), you'll see that most people are using it to mean criticism they don't like, not real incitement to violence. We already have a term "incitement to violence" for the latter situation.
Also, the first thing I saw on there was the claim that something was "a dog whistle for stochastic terrorism", aaaargh.
I think that's more symptomatic of social media's tendency to stretch out and water down these terms over time by incentivizing people to make huge reaches to overstate things for rhetorical effect, see e.g. "genocide" to mean "the government might not directly fund gender affirming care for kids", "grooming" to mean "relationships between consenting adults with more than a 3 year age gap between them", etc.
But I don't think the term stochastic terrorism is entirely without its use - like the examples above, there are real bad behaviors that can be accurately described using the word, people should just reserve the words for those cases instead of reaching to use them as cudgels to bludgeon mild things they don't like.
Per the claims at https://drrollergator.substack.com/p/stochastic-terrorism-a-game-of-rhetorical the phrase didn't actually originate reasonably and then get watered down, it was coined by a DailyKos blogger to criticize his political opponents.
"stochastic terrorism" is a terrible term, but the Left invented it, and now it's here, and it's attached to the identifiable phenomenon, so in the absence of a better term we're stuck with it. Many such cases - we have "piracy" as a designation of "copyright violation", as if robbery on the high seas has anything in common with watching a movie without giving Disney your money, or fixing a software bug without paying for the support contract. But it's what it is and we're stuck with it.
I don't think "piracy" is quite analogous, it's become basically a term of art in the context of digital media, whereas the vast majority of the usage of "stochastic terrorism" is just people on social media misusing the word to mean "criticism" or "calling for things I don't like" rather than the more appropriate meaning of "implicitly or explicitly calling for terrorism/violence".
"Piracy" to mean copyright infringement is pre-digital; in fact, the name of Gilbert and Sullivan's "The Pirates of Penzance" is a reference to this (which had plagued them with HMS Pinafore).
There also was "pirate radio" in the 1970s, sometimes broadcasting from international waters.
>"stochastic terrorism" is a terrible term, but the Left invented it, and now it's here, and it's attached to the identifiable phenomenon, so in the absence of a better term we're stuck with it.
Probably right, but I wince at it. 'Irresponsible' would probably have been a better choice. Amongst other considerations, it is generally not reasonable to expect everyone to be responsible at all times, though one wants fairly strong norms about it most of the time.
As far as I know, no actual shooting has been traced back to someone using Piker's diagrams.
Sure, by the nature of being stochastic it is not guaranteed to precipitate into a real world outcome.
The current rate is 0, impossible to distinguish from 0 effect at all.
I'm reminded of something: https://www.astralcodexten.com/p/the-phrase-no-evidence-is-a-red-flag
Also doesn't change the fact that Piker being a popular streamer instead of some streetcorner schizo rotting away is an indictment of American culture.
There can be no evidence because it hasn't been gathered yet, or because there is no evidence to be had. I think that politically-motivated assassinations (and even DIY guns) are rare enough that the latter is the case here.
I agree that it doesn't speak well of people that they are fans of someone like Piper.
I find it interesting that this premise - "criticism is violence" and even "silence is violence" and even "support expressed not in ways that we approve is violence" - has been the foundation of the whole cancel culture. Nobody really argued that using a wrong pronoun or wearing a sombrero while being white literally murders people. But the argument was if you keep using wrong pronouns and wearing sombreros, it creates an oppressive environment in which eventually, after enough logic jumps, the violence will happen. And this, after enough logic jumps, is tantamount to genocide. So it is prudent to suppress the nascent genocide in the bud and treat misgendering if not as actual-violence then surely as one-tiny-step-removed-from-actual-violence. And if we save even one life by that, it's well worth it.
Now, when the cancel culture had run its course, when even single institution has DEI commissars, affinity groups and speech codes, when anybody that had any inclinations of saying something anti-woke is driven out of every important institute of the society; when they finally dusted off, licked their wounds, started building their parallel structures, and then started discovering that once these parallel structures begin to work and allow them again to discuss and spread their ideas... The Left reacts to them with actual hot violence - not counter-words, but bullets, firebombs and bike locks to the head - then we suddenly discover that actually violence is not connected in any way to the political propaganda.
Hundreds of thousands of people can call for murder of conservative figures - completely in the open, with known URLs and sometimes under their real names - and yet, when one person writes in their diary "I had enough, I am ready now, I will do what I have read about ten thousand times on social networks and murder the Conservative Figure" - and then they do exactly that - now we are supposed to think it's not related. If ten thousand people say "Conservative Figure must die, we pray for the day he dies, we want him to be shot, we live for him to be shot, whoever shoots him would be a real hero and we will worship him" - and then, when he is shot, they deliver on the promise, they rejoice, they worship whoever shot him as a hero and they name the list of who should be shot next - all this is 0% related to the shooting. 100% somehow it appeared in one crazy head, just out of the blue sky, and there's nothing else to see here.
That's the "liberal solution". We just discovered that, sorry we didn't know about it all those years, but you know, science works in mysterious ways.
Maybe for some people all that makes sense. For me, I get a strong impression somebody is trying to sell me a Brooklyn Bridge. Sorry, no deal.
Your analysis fails to consider that some people simply deserve to die
JRRT already answered that question. There might be people who deserve to die, there are no people who are qualified to decide that. Before you ask, yes, that's why I also oppose capital punishment.
A rule that nobody is ever qualified to decide that would enable an enemy army to conquer you and then ignore such norms.
We're not talking about the situation of self-defense or warfare. There's a lot of difference between soldiers at war and a person deciding "mayor N deserves to die because he's literally Hitler, so I will take a gun and murder him". I do not believe anybody can seriously not understand this difference, only pretend to not understand it to achieve some rhetorical goal. Whatever it is, it will have to be achieved by other means, if at all.
Didn't you make a comment saying that was your logic for opposing the death penalty? So your belief is that even the state, with its Weberian monopoly on the legitimate use of violence, cannot decide to kill people, even if those people have been convicted of murder. You need to have some rule distinguishing why one case is ok and the other not, because "nobody can make that decision" doesn't cut it.
I don't see what's your problem here. The situation of combat and self-defense is different from killing someone who does not pose any threat to you, just because you decide you want to do it because they "deserve it". Murderers can be easily neutralized by keeping them in jail forever, thus there is no self-defense aspect in killing them. If there existed some super-powered maniac that could not be held in jail, then the question would arise, but those do not exist, so it doesn't.
Combat situation is completely different - nobody sits with the CVs of enemy soldiers and decides - this guy deserves it, this guy doesn't, this guy donated to charity and saved 100 kittens, so we're telling the artillery to take 1 degree left to not hit him. That's not how the war works, so "deserving" doesn't come into it.
Gandalf's rule, as borne out by his actions throughout the story, seems to be that it's okay (often laudable) to use violence to deal with enemies who are actively doing evil, but once someone is captive or otherwise no longer a dire threat you should offer mercy. Note his treatment of Wormtongue and Saruman as well as Gollum.
While I consider the idea of "stochastic terrorism" mostly illegitimate, I do make at least one exception: Most things that could be fairly described as "doxing" are stochastic terrorism. There's almost never a valid reason to try to unmask an online commenter or bring attention to their home address, only attempts to intimidate or incite violence.
The term "doxing" has started getting applied widely enough that I wouldn't rely on it for such things. Also, it used to be the case the the phone book published everyone's home address! I will agree that frequently things called "doxing" actually do resemble the kind of speech a platform would be better off not hosting.
The phone book published the home address of Joe Smith.
It did not publish that commentator Jimmy Beanslop is actually Joe Smith.
It also didn't say "Heterodox commentator Jimmy Beanslop is actually Joe Smith and here's where he lives; progressive commentator Sturgill Oklahoma is a national treasure and we'll respect his pseudonymity."
Each extra step reduces the probability of something happening. If mindkilled people are tweeting about how John Doe is a horrible person, and one of them is in the mental state ready to go and kill a person they are sufficiently outraged about...
...it makes a difference whether such person realizes "actually, it is possible to find addresses of people" and "actually, the phone book contains the addresses" and then actually goes and takes the phone book and finds the address -- as opposed to juicier alternatives such as more tweeting.
...or the address is already conveniently provided in the tweets.
That said, law is often too crude tool to make such distinctions meaningfully. What if instead of posting the address you only say "btw, the phone book contains his address". Or, assuming that the phone book is online, you post a hyperlink to the phone book. Or to the specific page of the phone book containing John Doe's name. Etc. Expect people rule-lawyering, and then getting hysterical about "did you know that in Europe they can arrest you for literally saying on twitter that a phone book exists???!!!".
In addition to what ProfGerm and Viliam said, I would also note that our phone book allowed you to opt out of being listed. And they didn't even have addresses for residential numbers, only for businesses.
Just as a sidenote: do not post even interior pictures of yourself online, if you don't want sufficiently motivated people to be able to figure out where you live. I don't consider "using what you yourself put online" to be doxxing (although it's probably still bad behavior, and stalking the person is VERY BAD and WRONG).
In order to be meaningfully "stochastic" anything, there has to be a probability greater than zero. The probability of someone too lazy to figure out their target's address, ultimately prying their size-60 backside out of the couch in their parents' basement and going off to execute a successful assassination, just because someone else did one of the easy parts for them, seems pretty close to zero. Maybe it's happened somewhere, sometime, but not in a way that has ever risen to my attention and it doesn't seem to be a real problem.
People like Luigi Mangione or Tyler Robinson, find their own targets.
The probability will often be low, but never zero. And this has to be weighed against whether there's a valid public interest in knowing where the target lives - if there isn't, it doesn't matter how low the probability was, it was still wrong to do.
Not all murders are high-profile. Not all assassinations are known to be assassinations if the police don't figure out the motive.
And laziness isn't always the main barrier to figuring out an address. It isn't always a matter of just looking up something that was an open secret.
Yeah, "...but there's no 'valid public interest', so we can silence that offensive speech", is the nose of a very large camel that is wholly unwelcome in my tent.
I'm usually very sympathetic to that kind of argument, but in this case I think the doxing *is* the camel nose for bad things, especially when it happens to someone who isn't a public figure.
Plus, people do (or should) have a right to privacy, even in the absence of threats.
"Will no one rid me of this turbulent priest?"
https://en.wikipedia.org/wiki/Will_no_one_rid_me_of_this_turbulent_priest%3F
What do we accuse someone of if we think they actually are trying to get "someone" to commit violence but are trying to hide it behind plausible deniability? The obvious example is trying to prove that a mob boss ordered a particular hit...
The difference here is that Henry II had power to pardon and to reward. Followers would be able to hear his words as an implicit command. The modern day equivalent would be Trump's tweets around Jan 6, or Duterte encouraging vigilante slaying without explicitly commanding it.
>Maybe there are some things you shouldn’t say even about the KKK - “they are Satanic cockroaches unworthy of life” might cross that line.
So dehumanizing metaphors can move the needle on violence then? There is a fair amount of social-psychology research suggesting it does more than generic negative claims. Or are they bad just because they're obscene?
I think the cockroaches reference specifically comes from Radio Télévision Libre des Mille Collines (RTLM) where it was their preferred term for Tutsi's during the lead up to the Rwandan genocide.
Google/Gemini says that the rather similar "Death to the fascist insect" goes back to 1973 (weirdly, to the Symbionese Liberation Army (SLA) - very weird times...)
I agree with most of this, but was struck by the line "But if someone fails to follow this discussion norm, their penalty is being self-sorted to spaces that have bad discussion norms, not being held responsible for any resulting acts of violence."
At least in these circles, these people (1) often aren't pushed out at all; (2) will sometimes be pushed out in a way that results in them forming much worse groups, sometimes with intent to organize and foment violence or disruption; (3) will instead push out good people. I would like ideal punishment to be meted out, but I've come around to the view that I'm engaging in wishful thinking and we need another liberalism-compatible solution.
Which circles are you talking about?
Less of this, please
So two thoughts.
Firstly, stop using "generic" statements. "Men do this", "Purpletonions do this", "Cops do this". Stop generalizing. We use generalizations ie Groupings to allow us to discuss the world intelligently in simple terms. Unfortunately, some people, and I'm going to blame the sociologists, have got it in their heads that Groups are what the world is made of, and membership in a Group gives you all of the aggregated rights and obligations of all the members of that Group. So blame goes to all members of a Group regardless of the actions or merits of individual members of the Group. And privileging an individual member of a Group, who has not otherwise been disadvantaged somehow makes up for the disadvantages levelled against a different member of that Group.
Secondly, better to criticize the actions or beliefs of an individual rather than the individual themselves. So if someone doesn't believe, say for religious reasons, in the equality of women and men, then we should address that and say that it is an illiberal, and wrong headed, and ignorant belief, and harms society, without actually saying that an individual who holds that belief is evil or otherwise unworthy of being a member of our country or society.
I agree with the bulk of this, but following up
'Firstly, stop using "generic" statements. "Men do this", "Purpletonions do this", "Cops do this". Stop generalizing.'
With
'Unfortunately, some people, and I'm going to blame the sociologists'
Kinda undercuts your point, no?
This comment made me smile.
It doesn't undercut my point, but it makes me a hypocrite. So good catch, I will have to do better!
I think that what you are asking is, in practice, close to impossible. Consider what you would have had to have done to avoid 'the sociologists'. About the easiest alternative would be to change that to 'the x% of sociologists who report on people via partitioning them into groups'. I don't think that that statistic is going to be readily available...
Or just say "some sociologists". Not so impossible.
Many Thanks! That's a good point, and a good alternative.
I think this is mostly correct but I think you left unaddressed the elephant in the room - what of a powerful political cult leader with an army of millions of followers who believe everything he says, and this powerful political leader also happens to control every national law enforcement agency and the military and is supported by a phalanx of prosecutors and judges who will leap up to defend him from any legal consequence and make sure his insane legal interpretations get a more than respectable hearing. If such an individual were to make harsh criticisms of religions and national origins and political opponents would that qualify as stochastic terrorism? Does Donald Trump telling a crowd of January 6ers to fight Congress (peacefully!) as they're counting electoral college votes count?
Trump just paid E. Jean Caroll millions of dollars after the courts rejected Trump's claim that he was immune to such "consequences".
>You’re allowed to criticize policies, but not individuals. Okay, so somebody says that allowing immigration from the Middle East is destroying America, but avoids saying the specific words “and I hate individual Muslims”.
This is a weak example, because "Individual Muslims", is not in fact an individual, it's a category of people. "Abdul who runs the falafel cart down the street", is an individual. And "Immigration from the Middle East is destroying America, and I particularly hate that guy Abdul over there" is perilously close to incitement to violence.
Of course, this shades into "The alt-right is destroying America, and I particularly hate Charlie Kirk". I agree that we have laws that do a tolerably decent job of drawing that line, but I also think we need a social norm for cases that we're not actually going to take to court, And I don't know what that should be, except that I generally take care not to express hatred of specific individuals.
Perhaps because I'm not on X, I don't understand the leading example. Someone please enlighten me...
>Again, creating a spreadsheet to accuse someone of killing others is a self-evidently absurd exercise, whose only possible motive is to be used to incite political violence. This is stochastic terrorism, and nothing else.
Who created the spreadsheet? Where's the spreadsheet? I suppose in context Musk is being accused of killing others, but why would using a spreadsheet be a particularly evil way of criticizing Musk or anybody else? Do spreadsheets actually have a track record of inciting violence?
If everyone else who plays the stochastic terrorism card is just as incoherent, why are we even talking about it?
I'm guessing critics of Musk's DOGE activities tabulated in a spreadsheet every cut DOGE made, estimated the number of deaths resulting from each, and summed it all up to conclude he's killed 700k/millions/etc. people.
OK, thanks, sounds plausible. The "Again" implies the existence of a previous more reasoned critique, so I'll give the commenter the benefit of the doubt and not assume their own comment is as self-evidently absurd as it seems on its face:
"Sure, we agree that causing death is bad, but using Excel to draw attention to it is purely evil. You should be using PowerPoint."
As you might have found out by clicking on the tweet, it is in response to this article: https://www.theargumentmag.com/p/elon-musks-zero-accountability-life
If I had known to click on Jerusalem's tweet <twice>, I would indeed have found that article. An article in which the words "spreadsheet" or "Excel" don't appear. So in retrospect it wouldn't have helped. Thanks for the tip, though.
I like this piece quite a bit, and think it's worth engaging in the throat-clearing to say so.
I also think that at its core it's heavily relying on the ability to bright-line rule that non-state* violence is bad, and dodges the ethics of how we should think of "stochastic" phenomenon more broadly. Is a Stochastic Cancellation a thing, and is it bad?
How does Radio Télévision Libre des Mille Collines fit into all this? They said some awful stuff on the radio that resulted in some horrendous violence. It feels like the 'liberal solution' might not deal appropriately with a radio station like this.
Some other solutions:
- Criticize specific actions and propose solutions to them that abide by the liberal agreement
- Criticize abstract dis-virtues like bigotry, hero-worship, fear-mongering, etc. (similar to criticizing policies, but more action-guiding for the individual person)
- Constantly inject compassion for all individuals involved to fight the natural human tendency for out-group dehumanization
I like this quote from Thich Nhat Hanh to Martin Luther King Jr:
"I believe with all my heart that the [Vietnamese Buddhist monks] monks who burned themselves [in 1963] did not aim at the death of the oppressors but only at a change in their policy. Their enemies are not man. They are intolerance, fanaticism, dictatorship, cupidity, hatred and discrimination which lie within the heart of man. I also believe with all my being that the struggle for equality and freedom you lead in Birmingham, Alabama… is not aimed at the whites but only at intolerance, hatred and discrimination. These are real enemies of man — not man himself. In our unfortunate father land we are trying to yield desperately: do not kill man, even in man’s name. Please kill the real enemies of man which are present everywhere, in our very hearts and minds."
"Probably there are some cases where extra-state violence is acceptable, for example assassinating Hitler... the best I can do is say that at least this limits the damage by only legitimizing attacks against powerful state targets."
The correct response to this is "yes, and by the time you're trying to assassinate Hitler, you're no longer participating in political discourse, so this discussion about what speech we should tolerate in a liberal society becomes irrelevant."
There's also a distinction between thinking an individual act is morally acceptable and advocating for it politically. It's not hypocritical to think a specific vigilante murder was good,* while also believing that society would made worse off by more tolerance for vigilante murders (because most of these murders would be bad, and there would be second order effects).
*This statement is hypothetical and should not be construed as an expression of support for any particular murder.
I think there's another way to draw the line.
If I say "[X] should watch their back because one day they're going to find my knife in it," then I could plausibly be investigated and charged for the threatening statement. That covers a dual purpose: it nips in the bud and also punishes the direct harm of causing fear.
However, if I say "[Broad group] should watch their backs because one day [etc]," then that often does *not* rise to the level of a criminal statement. The same violent sentiment is no longer criminal if the group is too broad for any single person to be specifically identified. (That is, "the bridge club that meets across the street every Thursday" is specific enough to be threatened, but "[members of the political party you don't like]" is not.)
I don't think that 'stochastic terrorism' is a useful concept when applied to individual people. A person is threatened or not, assaulted or not. A micromurder on a person rounds up or down. For a group – particularly one that already experiences a baseline level of violence – then inflammatory rhetoric can affect the 'climate', even if the 'weather' of day to day life cannot be so causally identified.
This doesn't necessarily mean that 'stochastic terrorism' can be criminalized in the US; elsewhere as noted it would be adjacent to hate speech laws. Still, I think it's a useful if inflammatory rhetorical device (read: jargon to be used when talking to the in-group, not something to deploy in persuasive speech) and possibly as an analytical tool.
In specific response:
> The liberal solution is none of these: it’s that it’s always legitimate to criticize any person or group however strongly, and never legitimate to use extra-state violence against them; if a violent act occurs, it’s 100% the fault of the perpetrator, and 0% the fault of other people who previously criticized the victim.
> (separate from this, I think it’s a useful discussion norm to avoid being too vitriolic or hyperbolic in your language, except perhaps once in a blue moon to make the point that you care about this issue so much that it’s your one exception to your non-vitriol policy. But if someone fails to follow this discussion norm, their penalty is being self-sorted to spaces that have bad discussion norms, not being held responsible for any resulting acts of violence.)
You have a contradiction here: the person who commits violence bears "100% of the fault," but the incitement to micromurder should still be considered unsociable and unfit for polite company. That social rebuke is a kind of punishment, even though it is far, far beneath a criminal standard.
To the extent that public speech affects the public – and I note that elsewhere you've championed public art that is aesthetic for similar reasons – then words do have consequences.
I missed the implications of this in my first reading. “Probably there are some cases where extra-state violence is acceptable, for example assassinating Hitler or overthrowing an illegitimate government”
There’s a problem here. And it’s not that killing Hitler is problematic, it isn’t.
Thus surely means we have to be careful about who we call fascist, or Hitler, because it’s ok to kill past Hitler, it’s ok to kill future Hitler.
Ergo, Scott disproved his own argument.
> Thus surely means we have to be careful about who we call fascist, or Hitler, because it’s ok to kill past Hitler, it’s ok to kill future Hitler.
We need to clarify the exact circumstances under which it's okay to kill Hitler.
It's okay to kill Hitler only (a) once he's actually in power, and (b) once he has made it impossible to remove him democratically (by cancelling elections). It's not okay to kill 1932 Hitler because you think he might do these things, but it is okay to kill 1934 Hitler even if he hasn't done all the other bad things yet.
Well the thought experiment is killing Hitler as baby. Or a child. I disagree. Why not a good swift kick in the balls that will give him a squeaky voice when he’s an adult? Can’t see him leading the Nazis with a voice like Daffy Duck.
Scott already wrote a post about this: https://www.astralcodexten.com/p/fascism-cant-mean-both-a-specific
>You’re allowed to criticize these groups harshly, but not using special emotional or dehumanizing words.
I have some sympathy for this one as a social norm, while holding to something pretty close to the strong liberal position as my preferred legal rule.
A couple narrower but related social norm rules that I would be more strongly supportive of would be:
1. Don't cheer for actual terroristic or mob violence when it happens. When someone you don't like gets assassinated, the correct attitude should be "cool motive, still murder".
2. Show restraint when cheering hypothetical violence against real people, even cheering that falls well short of actual incitement.
What's the reason you singled out violence as a result of indirect encouragement? Because that concept seems to generalize to all kinds of activities, for example to people drinking bleach in increased numbers after Trump publicly discussed it.
https://www.poison.med.wayne.edu/updates-content/kstytapp2qfstf0pkacdxmz943u1hs
Or shouting "Fire!" in a crowded theater, the classic example?
I agree it's a contentious topic, whether or not you should be allowed to say things that cause bad things to happen, or that cause _anything_ to happen, really. But the fact that "stochastic terrorism" exists should be rather obvious, because it's merely a special case of "words can influence people to do things", which is a special case of "organisms take sensory input and process it into some behavioural output".
I really don't see why someone who produces a certain input should under NO circumstances be held responsible for the output, especially if that output is somewhat predictable: "Will no one rid me of this turbulent priest?" Sure, it's a judgment call most of the time, but that's the unfun part: There is no absolute certainty in real life, and we have no choice but to make it up as we go. You may wish to cut through this mess with the "liberal principle" of assigning the blame 100-0, but that's just not how it will ever work.
Maybe dehumanizing language is the thing to avoid. Muslims are garbage to people I know. In Rwanda it was years of saying the other were cockroaches that may have led to people killing them with machetes like cockroaches.
The constant drone of X group is trying to destroy our country is not reality. They are not waking up in the morning wondering how today they can destroy the country.
The hard part is Trump and MAGA are convinced they’re trying to improve America but are trying to destroy democratic institutions so they can get in power to do it…so they literally are trying to destroy it but are self deceived they are the hero’s saving it, and they cast any opposition as a destroyer so we get locked into a spot and no one can move. No injection of truth is possible since it’s a lie.
I agree with your critique. I feel the only path forward is electoral change and a new government moving to stop the minority loud mouths and get normal people with good hopes back into the spotlight. Make conciliatory decency the norm again, and then let democratic dialog chart our path.
Posse Comitatus. I'm just going to say that when congress calls for the executive to violate acts of congress, they're trying to destroy the very institution they serve in. You know, like calling on the National Guard on Jan 6th (which the Pentagon rightly vetoed, on the grounds of "thats a very dumb idea").
I am also increasingly hearing politicians on the left say things along the lines of "the right wing party is undemocratic or is trying to destroy democracy, the only option to save democracy is to vote for our party", thus effectively espousing a system where there is only one legitimate party, ie. a de facto destruction of democracy... (and these types are also completely convinced they are the heros saving democracy).
Colorado trying to block Trump from the ballots comes to mind.
It's a classic Russell conjugation, or that two boats meme: "my passionate democracy, your destructive populism"
> You’re allowed to criticize these groups harshly, but not using special emotional or dehumanizing words. Maybe there are some things you shouldn’t say even about the KKK - “they are Satanic cockroaches unworthy of life” might cross that line. But this is a motte-and-bailey
This roughly describes my view of the topic. I think this term has been widely abused and I do not accept the bailey here, that criticism, even harsh criticism, is equivalent to a call for violence, but there are certain statements that can qualify as an implicit call for violence and this is one example. Another example is that there were pretty literal calls for more Luigis after that happened.
I think it might be worth distinguishing the normative vs the descriptive parts of the debate. Can you plausibly increase the likelihood of a Trump assassin by cricizing him? Yes, even though it is highly unlikely your individual actions would have a significant effect. Does that mean you shouldn't be able to criticize Trump? Of course not, criticizing the president is a very basic right in a functioning democracy.
On the other hand, most people accept that inciting violence is not protected free speech. So the debate is really on where we draw the line on what counts as inciting violence.
I think one thing missing from this conversation is the claims people will make about the truth/falsehood of the stochastic terrorism in question. Like some activists would say you are doing stochastic terrorism if you engage in certain rhetoric about transgender issues, but that's because there really is an assault on transgender rights and we need to push back on it. And then if you said, well should we not criticize conservatives then, they might say "well no, but that's because it's not true that conservatives are under attack like trans people are."
This isn't really a counterargument, I guess it's just a question of how you deal with psychologically motivated people in general
Cite more numbers than they do? Transgender people are indeed more assaulted, but that tends to correlate with race and occupation more than "generic ally" would like to think.
I can't think of any time when the Secret Service was called in to go after liberals... So, yeah, the conservatives are under attack (kinda sorta, devil's in the details, but I already said that).
Obviously the Secret Service does go after radical leftists (I wouldn't call them liberals, but then I wouldn't call the MAGA movement conservative either), to some extent even during Democratic presidents. See e.g. https://www.cato.org/blog/trumps-anti-antifa-executive-order
A different aspect of the Secret Service than what I was referencing. The Secret Service also intervenes in currency matters...
This is the sort of thing I was talking about. People will be quick to say "of course this issue is real and happening" for the thing they care about, but when someone else does it, well now it's actually something else and they're not correct on the facts like we are.
stochastic terrorism (concept) may not make for good legislation, but surely it does describe the intent and motivation behind most violent critique?
I will am willing to swallow the pill that Yes, stochastic terrorism is real and does originate partly from the way we talk about subject X,
and that No, we should not ban people from saying anything about anything, and that includes me coming in off the top rope on eg. Charlie Kirk with some tasteless burns before his soul had even gotten half way to hell.
Much like his commitment to the right to bear arms, you have to be committed to the right to ask if people would rid us of such and such turbulent priest if you want free speech at all.
I think there's been a social change that the phrase stochastic terrorism gestures at, even if no individual speaker should be accused of stochastic terrorism. Nixon also abused a lot of power and usurped democracy, but I don't think there was the same fear of his life that Trump has. So you can describe the state of affairs as stochastic terrorism.
There's been a culture shift where political terrorism isn't as unthinkable. Blame social media rabbit holes, community alienation, etc but the 10% left and 10% right most extreme Americans act or profess to think differently.
Gallop: "Americans' Acceptability of Political Violence, by Subgroup % Sometimes OK"
30% 18 to 29
21% 30 to 44
13% 45 to 59
4% 60 and older
It's the flip side of Stochastic martyrdom: "the encouragement of activist tactics that will predictably get activists hurt or killed when applied at scale".
In both cases it's a downstream effect, any individual speech act isn't directly causal, but the concept should still be tracked to see how the US is progressing.
Many Thanks, particularly for the statistics. There is always a tendency towards 'kids these days...' but 30% >> 4%. That's a real difference.
When you believe non-violent political change is ineffective, violence starts to look more attractive.
I think some people here should think about getting a "real job." Hate is always Hate
It is never justified. Only rationalized by weak minds.
I feel like there is an audience contingency part of this. If Elon Musk uses Twitter data to create a private group chat made up of everyone on the platform that both engaged with anti-AI content and engaged with schizophrenia support content, and then started sending them articles about how Sam Altman was going to kill all of humanity every week, this would likely be 'stochastic terrorism' even though Elon should in the abstract have the right to write articles about how Sam Altman might lead to the death of humanity. Similarly, we could say that if some survey data came back showing that for some reason the ACX readership now consisted of 20% paranoid schizophrenics who are anti-AI, this blog would likely become slightly more cautious about AI rhetoric.
There is obviously a tension between audience management and being able to conduct truthfully in discourse that should resolve mostly towards the latter. Someone like Desmas fashions herself as a public intellectual, trying to steer policy, and this macro aim is more consequential and important than how one person may react to some idea she puts into the public sphere. However, even though her obligation to manage her audience is low, it is not zero. The right-wing argument is that a lot of left-center social spheres are made up of potentially violent far-left extremists, like maybe these people make up 10-20% of someone like Desmas's followers, and that because these people are such a large part of left-center internet spaces, the left-center has more of an obligation to manage their audience than they presently do. Calling them not doing this 'stochastic terrorism' might be hyperbole, but it exists somewhere on the spectrum between 'completely not stochastic terrorism' and 'obviously trying to incite crazy people to harm someone'.
I do not think their fear is completely wrong (it might be exaggerated, but a large share of the left has turned against liberalism), but I think it is unrealistic to expect one side to unilaterally disengage from pointed political rhetoric. The argument might have an internal logic to it, but it is obviously being deployed for ideological aims: one could just as easily make this argument about right-wing figures who spread harsh rhetoric about 'hypergamy' and 'western civilization dying'.
Having some direct experience with stochastic terrorism, I think this misses the point. It actually does not behoove our society to wish literal death upon the other political side. It wasn't always like this, either--Obama and Romney were kind and complimentary to each other. Now both sides screams for lawfare and political violence. It's better to respect your political opponents and argue like civilized people, than it is to shriek about how they ought to be shot; that speech, which deep down everyone knows is absurd, should be met with criticism and disgust. Nor do civility norms throw out some early warning system against Hitler--indeed, the abandonment of civility in a society more often preludes tyranny.
What you are talking about is usually called polarization, or normalization of violence, and most people would agree it's bad. Stochastic terrorism is different: the claim is that in a society where violence has (for better or worse) been normalized, even non-violent criticism can make someone a target. The cited tweet where someone was accused with stochastic terrorism was about a blog post that analyzed the consequences of DOGE and claimed (in perfectly polite language) that Musk is responsible for the death of millions.
This is off-topic, but I find it amusing that modern nigh-superintelligent AI can't even get the arrangement of pips on a d6 right.
I did think it was a particularly sloppy thumbnail this time. One low quality image won't undermine his credibility but if you get enough of them, then stochastically...
This seems like a very exploitable position to take. Eg. Public Figure constantly rails against Demographic. His audience is very extreme in their hatred of Demographic. One day Public Figure points out a specific individual within Demographic as being especially bad, but doesn't call for any violence. That individual is later killed by a random follower of Public Figure. The next week, Public Figure points out another individual member within Demographic. Again, later on, *that* individual is killed by another, different random follower of Public Figure. Every month or so, Public Figure focuses the anger of his audience onto one particular individual, without calling for any violence, and every time, a member of his audience either kills, harasses, or attacks them.
Does Public Figure essentially now just have a Death Note he can use with complete legal impunity? If he went to court, and then told his audience about how he thinks the judge is biased and evil, and potentially a secret member of Demographic, is it really in society's interests for him to just walk free? Seems shady.
This piece jumped the rails at some point and started conflating the original question ("can Commentariat / Media be held responsible for others' acts of actual violence via the charge of 'stochastic terrorism'") with the question "is political violence itself bad." To make things worse, Scott then further conflates political violence with revenge & vigilantism: as I understand it, Luigi Mangione targeted the CEO of United Healthcare explicitly because he believed that person had personally ruined his life by denying him needed care for debilitating chronic pain.
You can pinpoint the exact moment we lost the plot: "It’s vacuous to tell Brian Johnson’s family that they can’t complain about his murder just because health insurance CEOs are a privileged group." Yes, that would be a vacuous thing to say, but trying to make me sad for Brian Johnson's family is also a vacuous rhetorical device because it has exactly nothing to do with the question of "stochastic terrorism." (Or even, I'd say, with political violence.)
1. I basically don't agree with any of your examples of large groups that are up to no good. I mean, in my opinion Trump voters all made at least one dumb or evil choice, and in some cases are continuing to make correlated dumb or evil choices, but it seems way more fruitful to criticize *that choice* than to bash an enormous group of basically normal people that's diverse enough that any generalization you could make about them will often be false. I think maybe I just do support a norm of "don't bash large groups of basically normal people", even though I think most instances of doing so are not any kind of terrorism
2. even though I'm opposed to political violence, my gut reaction to "well if you say that someone might become a dictator or end the world, obviously people will want to kill them, so you shouldn't say that" is "that which can be destroyed by the truth should be"? even if there's a bunch of people who wrongly believe "if A, then B" I don't like the idea of willfully misleading them about A in order to manipulate them into believing that not-B. instead we should actively make the case for why "if A, then B" is false
Today, I wish to draw the Americans attention to this:
1) You voted for Cornpop. And you thought he was not senile.
2) You voted for the other guy, who has said 37+ times that we are "very close to a deal with Iran." (and we called Clinton a flip-flopper! -- which was entirely inaccurate, he was just prone to making late decisions, and convincing everyone that he was on their side before then).
3) You decided it was unimportant which of these two people would be POTUS.
I never heard of "stochastic terrorism”, but I agree with the liberal solution. Free speech in the US means you can say almost anything. With edge cases decided mostly by the courts. I'm a fan of 'Jackie Jackson' aka Contrapoints. https://www.youtube.com/watch?v=RTRT794IQBg&t=20s Who sometime later in the video says something to the effect of, 'I don't want to say the 'n' word, but it bothers me that I'm not allowed to say it.'
Observationally, there's definitely some groups with toxic dynamics that make their violent outliers predictable. I think it's clearly reasonable to have a norm against that in retrospect - people who are still fans of Luigi mangione, or who publicly supported him before and never distanced themselves, should be judged somewhat.
I'll take it a step further and say that the sort of group dynamics that lead to this, while hard to define with a bright line rule, are also reasonably predictable (although like all non-bright line predictions this is uncertain and subject to human biases), and so long as people are using their honest best judgement to predict this and apply norms based on it that's reasonable and should be taken seriously (with the normal caveats about losing legitimacy if your predictions are consistently wrong or biased).
I'll add to the first one that it's reasonable to demand groups disavow bad behavior they're potentially linked to - for example, I don't think AI safetyists ttied to firebomb Sam Altman (they don't have those toxic dynamics, in my best judgement), but it's still good that they uniformly publicly condemned it and it was reasonable to expect this of them.
"Some European countries distinguish between permitted normal speech and hate speech "
Except that they define hate speech as speech that they hate. It's not a valid category.
Right, I agree. The crack down usually comes against the press which is most radically against the currently governing party. Radical press on the other side is pardoned.
Or encouraged. Or slipped some money.
You guys have a fairly warped view of Europe.
Which assertion is inaccurate?
All of it? It’s not even wrong. Feel free to give an example of a crackdown on newspapers opposed to Labour in Britain.
>The “stochastic terrorism” concept is near-unique in how effectively it can be discredited merely by listing many examples of its use together in the same place. Almost no one supports a blanket prohibition on criticizing of all of these different groups of people
Completely disagree; all of these are stochastic terrorism, including the ones on my side that I've contributed to, and none of it should be prohibited speech.
There has to be a middle ground between outlawing speech and pretending it has zero negative impact and is therefore 'legitimate'. Freedom means the freedom to do stupid and even destructive things, and we must acknowledged those harms and costs. If you try to pretend that every thing which people should be free to do is also harmless and legitimate and fine, then anyone can call for a ban on anything by pointing out the harm it does.
We should be able to acknowledge how reality works and be allowed to describe it, including calling out the harm caused by things we don't want to prohibit, and calling for people to voluntarily do better for pro-social reasons. 'Stochastic terrorism' is a reasonable description of a real phenomenon, but it's not a call for censorship or state violence.
I didn't kill Charlie Kirk, but he would probably be alive if the type of targeted left-wing rhetoric I have occasionally engaged in never existed. By my own actions I have killed maybe .0000000000001 Charlie Kirks in expectation, and I accept the moral weight of my actions there.
I accept that as a moral price I am willing to pay for the activity and speech that I find valuable and entertaining.
I bite the bullet on that being an ok thing for people to do in general, and reserve the right to call out people when they're being egregious about it, just as I acknowledge their right to call out me and my 'team'.
I do not need to be absolved of this moral culpability, just so that I can feel more comfortable about my speech. I should never feel comfortable about doing anything for which I have not considered and accepted the potential harm it may cause.
Well put.
When I opened Scott’s post, I really expected it to be exactly what you’re saying here, or—even better— directly address this excellent take in a clever way.
> By my own actions I have killed maybe .0000000000001 Charlie Kirks in expectation, and I accept the moral weight of my actions there.
The concept of a micromort seems relevant here! :)
I was also surprised this wasn't the takeaway of the post. But I also don't see how 'stochastic terrorism' is used in practice, since I barely touch the big social media sites - this post was the first time I was even made aware of this term. But the way it was described, concur with your take on this.
I do think it's worth emphasising that it typically doesn't make a lot of sense to call people out on single instances of this; the whole point is that it only becomes bad in aggregate. I like your suggestion of when it's worth pointing out: "call out people when they're being egregious about it". Obviously, that's squishy and subjective, but it strikes me as a good rule of thumb, regardless.
Thanks for posting!
What are your thoughts on a specific very powerful person with an enormous amount of political reach, such as Elon Musk, demonizing a particular group of people? (Many have connected a recent spate of anti-refugee arsons to his rhetoric in particular.)
On the old principle of "greater power entails greater obligation", it can be argued that individuals with the ability to reach significantly larger audiences should be held to significantly higher standards. By that measure, Musk's insane maunderings should be subject to much greater control than those of the average rabid YouTuber.
>If your worldview says that it’s acceptable to lynch sufficiently bad people, then yes, accusing people of being bad is equivalent to calling for their murder, and you have no alternative but to make sure nobody is allowed to criticize anyone you like.
The problem is, it doesn't matter what *my* worldview is. The worldview that matters is the person that pulls the trigger.
I don't think you do anything to escape moral culpability by saying 'I will take this action that is likely to cause someone else to kill this person, but **I** don't think they should kill this person, therefore I have no moral culpability in their death.'
If your actions have foreseeable consequences, then you are morally responsible for those consequences. Consequentialism doesn't work if saying 'I do not agree with the consequences of my actions and I would prefer they did not happen' is a get-out-of-jail-free card.
Now, obviously, we should all be working hard to enforce the norm of 'no one should lynch people, even sufficiently bad ones', both through social policing and actual policing, as well as mental health services and etc.
But until we have achieved a 100% success rate on that front, we should also be brave enough to bite the bullet and say 'Yes, some forms of speech increase the likelihood of a lynching happening in expectation, and we place a threshold on that likelihood which we call 'inciting violence' and outlaw it, and below that threshold we accept those deaths-in-expectation as an acceptable cost for free speech'.
You can't improve the world if you don't acknowledge how it works.
"Not lynching bad people" can only be normalized if people believe other mechanisms exist for change. To the extent that a group believes that our institutions are fundamentally broken and won't be effectively reformed, they will resort to means outside those institutions.
Setting aside semantic quibblings about how we define 'lynching', yes, for sure. That's why we also have to have institutional reform and responsive democracy and etc. etc. as part of the project.
And alas, for some of them, the only fix for that is a mental institution, because of melodramatic convictions about what a broken institution is.
I have literally seen people blame capitalism for the existence of water unsafe to drink.
>But until we have achieved a 100% success rate on that front, we should also be brave enough to bite the bullet and say 'Yes, some forms of speech increase the likelihood of a lynching happening in expectation, and we place a threshold on that likelihood which we call 'inciting violence' and outlaw it, and below that threshold we accept those deaths-in-expectation as an acceptable cost for free speech'.
I'm afraid I disagree - the thing we call "inciting violence" is specifically about intent, not just about likelihood. And I think that that distincition is really important, and that it's vital that we defend the right to say things that are likely to cause violence but don't incite it.
The reason this is important is that people often react violently to speakers they disagree with - rioting as a form of heckler's veto. And if we make the basis for criminalising speach "likelihood of resulting in violence" instead of "intent to incite violence", we endorse that tactic, enabling any group willing to demonstrate a low threshold for violent rioting to have the state silence their enemies.
The complaint behind stochastic terrorism is better captured by the term "reckless disregard for the truth"
Speech that can incite violence cannot be blanketly banned. But if you resort to them, you have a moral duty to ensure your words are true.
And if your words are not in fact true, then you will proportionally be responsible for the violence that ensues.
Hmm... What about speech which is true, but could be construed as motivating harm? Consider Darializa Avila Chevalier, who recently won the Democratic Party primary in NYC's 13th district. She co-founded Columbia University Apartheid and Divest (CUAD), that calls for death to America and the "total eradication of Western civilization.". Now, I wish her a long and happy life - but preferably in Afghanistan, where she will have all the rights and privileges of a woman under her chosen religion, and need not be troubled by the Western civilization she so despises. Does that count?
She won a Democratic primary in one of the most important cities in the country. Fool she absolutely is, but not one that can be safely ignored.
According to Perplexity, the phrases “death to America” and “total eradication of Western civilization” are attributed to CUAD or to posts associated with the group, not directly proven as [Chevalier's] own personal statements. I have therefore deleted my earlier post, which was misinformed.
Many Thanks! But since she _co-founded_ CUAD, if she disagrees with their statements, she should say so.
The label seems more appropriate for the rampant spread of ai-generated text … which even unwittingly may subsume all human textual output, especially as we approach the threshold where data has to be invented to satiate the swallowing maw of AI.
So is obviously ridiculously overused , but surely there’s a level at which stochastic terrorism does exist though, no? The classic example being “Will no one rid me of this meddlesome priest?”. Two clear examples is this in modern parlance spring to mind:
First, I’m personally of the opinion that abortion is murder, and that the case for just up and killing abortionists is pretty clear-cut. I’m not personally willing to bear the cost involved, and I happen to think it wouldn’t do much systemic good in a lone gunman scenario, but in an alternate world where I don’t have a family and I can coordinate with, like, 1000 other people who agree with me I’d totally be willing to just off an abortionist. Surely there’s something here that my opponents could point to in the stochastic tourism idea space, no?
Second, what exactly is a fatwa in Islam if not a call for some good Muslim somewhere to off the object? And I don’t know of any specifically formulated fatwa against, e.g. David Wood, but I can find you plenty of good Muslims who sincerely believe his blood is halal. That’s a result of stochastic terrorism, no?
This is, roughly, the type of take that made me follow you in the first place. 100% agree, no caveats. Someone needed to say it.
The more general thing I'm noticed is that every non-technical invocation of the word "stochastic" is always just slop and nonsense. Surprisingly reliable!
https://substack.com/@linch/note/c-242561457
The term "stochastic terrorism" seems to have actually been coined in a fairly technical sense in a 2002 article in the Journal of Risk Finance https://www.emerald.com/jrf/article-abstract/4/1/7/251854/Quantitative-Terrorism-Risk-Assessment?redirectedFrom=fulltext ... that article is locked down but this 2003 one by the same author seems to use it in the same way https://conference.nber.org/confer/2003/insurance03/woo.pdf ... it concerns modelling the distribution of terrorist attacks of various sizes as a stochastic process.
The non-technical use of the term to mean "when my political opponents say things then that's basically terrorism" seems to date to this Daily Kos article from 2011 https://www.dailykos.com/stories/2011/1/10/934890/community/Stochastic-Terrorism-160-Triggering-the-shooters/ which briefly mentions Al Qaeda before providing an entirely boo-Republicans set of examples.
Anyway, don't blame the term, blame the coopting of the term.
Let's suppose the year is 1880, and someone who is not a member goes to a KKK meeting and tells them all about how their black neighbor is a rapist. They have no proof or evidence, they just make up a story about it and get everyone riled up. They do not literally advocate for violence, but they went there on purpose and told this story on purpose in this location to these people.
Some of the members, but not the storyteller, go and lynch the accused man.
Clearly something has gone wrong here. Clearly the storyteller has done something wrong above and beyond mere slander and is morally culpable in part for the murder which takes place.
You've correctly made the point here that we can't reasonably outlaw stochastic terrorism because it's too hard to legibly distinguish from regular criticism. But in the past you have made posts demonstrating that you understand the distinction between legality and morality, and the distinction between legal penalties versus social penalties.
Stochastic terrorism is a meaningful concept. It is a thing which people embedded in subcultures that glorify and praise political violence do in order to agitate the violent people and deliberately incite without incriminating themselves because they themselves did not technically call for violence (If persons A,B,C all say "kill Nazis", and person D responds "X is a Nazi" person D hasn't legally advocated murder). But it's morally evil and we should inflict the same outrage and shame upon them that we would on the KKK storyteller from the earlier example.
So one way in which your example differs from almost anything (I won't go as far as "anything", but I can't think of examples off the top of my head) in 2026 America is that it's a situation where words are *likely* to cause violence that would otherwise not happen, but don't necessarily meet the current legal standards of "inciting" it.
Nowadays, lynching is much rarer; it's pretty rare for acts of violence to be directly traceable back to a single act of speech that wasn't intended to cause them or something like them.
And I think that "saying things that have a tiny chance of resulting in violence" or "adding one more straw to the camel" is different morally, and should be different legally, to saying things that are near-certain to cause a lynching.
"This liberal solution isn’t trivial. It requires the separate liberal norm of always being against extra-state violence"
Particularly when you consider that the political movement that appropriated the "liberal" label in the US is at best ambivalent about punching (and shooting) "fascists". Saying that the solution is more "liberalism" doesn't seem to grapple with the actual problem.
Yes, I also found this a bit puzzling, but I guess the problem may be in the ambiguity of the word "liberal". (In European parlance liberal is more normally associated with economic liberalism, tending to be right-wing, but I think in North America some see it as a simple synonym of "left-wing").
Because saying that being against extra-state violence is a left-wing norm doesn't make much sense. I think the only political ideology which specifically advocates for extra-state violence is Marxist-socialism, with a violent class revolution being a basic premise of the whole movement - the George Floyd riots being the best recent example of the justification of extra-state violence for what followers of that ideology considered a just cause.
Conservatism tends to be very strictly against extra-state violence of any kind, though indeed sometimes runs the risk of falling into fascism and unjust state-mandated violence (ICE as a recent example?)
Right, my point is that the compromise-over-conflict kind of liberalism doesn't move hearts and minds in the US these days overmuch. Leftists claim that whatever is worthwhile in "liberal" is synonymous with "leftist", rightists blame liberalism for the ever-increasing unattractiveness of their ideals, and centrists are quietly wringing their hands in aimless despair.
It’s only Blood Libel if you’re spreading inciting lies about Jews; otherwise it’s just sparkling stochastic terrorism.
I see others have made this general case, but to me there’s a distinction between criticism, or even harsh criticism, and “Someone out there ought to kill this guy”.
Yes, people cheapen the concept of stochastic terrorism by applying it too broadly, as they cheapen all concepts. But it’s still a real thing.
For a concrete example, I’d point to the song “Monopoly Money” by Moon Walker: https://open.spotify.com/track/2eEOfXKk2H2SGsoY1FiqfI?si=D0F74zw1RCupC4BQ54z9jw&utm_source=copy-link
That song is an overt and unsubtle celebration of Luigi Mangioni style assassinations of CEOs. This is not mere “I don’t like rich people” stuff. This is “you, person listening to this song, should go out and murder a CEO specifically so they will be afraid of us”.
I think that is a special category of bad behaviour and it is useful to have a term to describe it.
> at least this limits the damage by only legitimizing attacks against powerful state targets, which can usually take care of themselves.
Not sure what you're saying here. Are you bringing that up as good thing, that Hitler-level targets can take care of themselves?
I think your argument that we shouldn't be against using emotional and dehumanizing words is weak. In particular, the assertion that "his opinions didn't hinge on slight changes in the emotional resonance of the words being used" might be true but is not the issue in question - his ACTION on those opinions almost certainly WAS influenced by the emotional resonance of the words being used. All the evidence suggests that emotive words bypass the rational part of our minds and that can never be good.
(E.g. using this sort of language is more generally detrimental - Look at the stupidity of Trump calling the Iranians "scum" and then expecting this to somehow not have a negative effect on the war)
I was surprised to see this quote: "It’s vacuous to tell Brian Thompson’s family that they can’t complain about his murder just because health insurance CEOs are a privileged group - especially given that health insurance CEOs now have a higher death-from-political-violence-per-capita rate than Muslims or transgender people."
The rate of domestic violence for trans people is quite high, as is the number of trans people (especially women) who are killed every year because they are seen as sick or wrong. See: Sam Nordquist, Tahiry Broom, Laura Schueler, Brianna Ghey, Brandon Teena, and many more people who were targeted for their identity. When transsexuality, a medical condition, is being treated as a political ideology or mental illness, how can we pretend that these murders could not be motivated by ideology? On a similar note, the 2016 Pulse shooting was openly motivated by a political and extreme religious ideology. I just find this to be a strange and incorrect comment.
There are about 1000 insurance companies in the US according to Google, so the politically-motivated-murder rate is at least 1:1000. There are a little under 3 million trans people per the UCLA Williams Institute estimate that's usually quoted on this topic, and way less than 3000 known-to-be-politically-motivated trans murders (the Wikipedia list has 30). Even the total number of trans murder victims is well below 3000, and most such murders are committed by the sexual partner.
Counting only active insurance company CEOs seems like an arbitrary distinction. Also, the Wikipedia list has a disclaimer that it is not complete, it just covers some notable cases. HRC counts at least 27 cases of fatal violence against trans people from November 2024-2025, but these are considered to be underreported figures because law enforcement or family do not always acknowledge that victims are transgender (because they are a controversial minority class). Murder victims who labelled themselves as trans during their lifetime are sometimes labelled as crossdressers or gay. The killing of a notable figure who is wealthy, such as a CEO, is probably much more likely to be reported to police than, e.g., that of a transgender sex worker.
I don't think that a trans person being killed by a sexual partner can be automatically seen as non-political due to things like the panic defence. Overall if we look at historical data I would feel quite confident that the number of trans people murdered for political reasons, per capita, is higher than that of an insurance company CEO.
Average CEO tenure is something like 5-10 years, so make it maybe 1:4000 if we are being generous. You can certainly imagine there being 25 unreported anti-trans hate crime murders for every reported one... you can imagine a lot of things. You could also imagine there being a lot of unreported political murders of SME CEOs.
I actually can't imagine there being many unreported political murders of SME CEOs, can you? And can you not imagine that it would be more likely for the murder of a transgender sex worker to go unreported than that of a CEO? Not even half of transgender hate crimes in the Bay Area are reported by the individual to the police, per a BMC Public Health study, and that is in one of the most 'trans friendly' areas of the United States. If the individual is not even around to report a hate crime, they are in a less friendly area, their family does not know or would not admit that their relative was trans, and they are in a profession like sex work, then yes, I can see the number of unreported murders of trans people being underreported by a large factor. By 25? I am not an expert in this field, lol, I am an IT guy with some trans friends, so I honestly cannot say. Maybe my root issue with the comment is that it's a weird stat to point to, and something like assault per capita would be more indicative of the issue of stochastic terrorism (e.g. could you say that punching an alleged nazi is part of it, or count those who are simply injured in a Christmas market hit and run? I'd say yes). Do you think that it's an issue though or do you agree with Scott?
With all the activist attention transgender issues are getting, I can't imagine many murders going unreported. I'd think it much more likely that the murder of the CEO of some tiny insurance company with 30 employees you never heard of would go unreported. Or rather, it would be reported somewhere, but the report wouldn't be easily discoverable - Wikipedia doesn't even have a list of murdered CEOs or businessmen or wealthy people, but it obsessively lists every homeless murder victim that has been trans; half of them have their own article.
Maybe there are many trans murders which are reported but the political motive does not get identified - in Business Insider's investigation, about a third of all identified murders were unsolved, and maybe politically motivated murders are much harder to solve (seems unlikely, domestic violence is probably easier to solve but robbery, murder during sex work, serial killings etc have no reason to be), but even then you'd only be in the dozens per year range, with a multi-million base population.
> Do you think that it's an issue though or do you agree with Scott?
I think it's a fairly straightforward choice of stats for the point he is making in that part of the post (that powerful people are no less at risk than oppressed people from the kind of attack that gets described as stochastic terrorism).
Eh, I think your argument rests on the idea that murdered trans people are correctly identified as trans and that their murder actually gets reported, which is not the case, as I have tried to lay out in earlier comments. Even the crimes experienced by the average homeless person go far underreported; if there is no report or media coverage there cannot even be a Wikipedia article, lol. Anyways, wish you well and it was an interesting conversation.
"I don't think that a trans person being killed by a sexual partner can be automatically seen as non-political due to things like the panic defence."
Which is not what happens in a lot of these murders, there is no trans panic defence. There was one where it was "trans woman killed by ex who was also a trans woman" so I don't know how you can get "trans panic defence" out of that. There were a lot which were "family knew they were in a relationship, everyone knew partner was trans, relationship broke up and ex got violent over that" including killing/injuring members of the trans person's family during the incident. Ironically, that is domestic violence from ex-partner of the same kind as happens to cis women.
I think there was one episode of "trans panic defence" in the lists of "all the deaths of trans people over this year" which lists are heavily implying "these deaths are due to transphobia" and they're not.
I am that kind of OCD/autistic bitch who went through the lists and looked up the cases in news reports online to get any background that was more than "DeShauna Shaniqua Duxelle, brave and stunning trans woman, killed in Chicago in June, rest in power" kind of naming in those lists.
Trans murder rate is largely disconnected from trans status; being targeted for that alone is quite rare- note "rare" does not mean "unheard of".
Total per-capita murder rate of trans women is lower than that of cis women, with some variation by race; it's much more correlated with prostitution and drug use than trans status *alone*.
>2016 Pulse shooting was openly motivated by a political and extreme religious ideology
Supposedly he was heading towards Disney and decided they had too much security, and didn't know Pulse was a gay club until after. Yes, he was motivated by an extremist ideology, but not quite the way you seem to mean.
As I have tried to show in comments above, trans murders and hate crimes are far underreported due to lack of interest, familial shame or mistrust of the police. In the Bay Area alone, over half of trans hate crimes go unreported, and that is a very pro-trans area.
WRT Pulse, I did try to find what you were talking about, but seems that the shooter was targeting nightclubs generally (according to his cellphone data) and that he was homophobic (according to his family). So may not be that he initially chose a gay nightclub but the club being gay may certainly have been a contributing factor.
Oh boy. This is a fraught issue, as I've gone through a couple of those lists of "trans people being killed for being trans" that are published online by trans activist groups, and they're... not completely accurate.
A lot of them are "killed by ex/in domestic violence" just like cis women. The partners/ex-partners knew they were trans and had no problems being in a relationship with them, it was when the relationship broke down that the violence happened.
Some of them were, literally, "hit by car" and that's treated as "death of a trans person is the same as killed for being trans" because of social transphobia or something (I wish I were kidding, but no.)
Some of them were "killed while engaging in sex work" which is *already* a dangerous line of work even if you're a cis woman.
There were about three murders I could call "explicitly killed for being trans" out of those two or three lists.
I think we can and should have a social norm against vitriolic criticism, but we should not label such criticism “terrorism” or “murder,” which are crimes that have actual legal meaning, not to mention that we would ourselves be being vitriolic.
We should also have a strong social norm against compiling lists of people who are not already famous whom we believe to be evil, especially when such lists include their home address, workplace, etc., or compiling lists of schools, places of worship, summer camps, etc., associated with an ideology we disapprove of.
One example that I am missing:
The assassination of Jicchak Rabin in 1995 was arguably quite close to stochastic terrorism, and I would say that it changed the path of the Near East for the worse.
"Bilateral controlled tiger-dismounting" is a phrase I'm soooooo going to use.
Also, I'm not sure Stalin's ascent to power could be described as "attempting to end democracy". It's true that he concentrated power that was previously much more distributed, but I think this was more like outmaneuvering his political opponents in the communist party rather than destroying mechanisms for expressing the will of the people or anything. There were no such mechanisms to begin with.
"I admitted that there should be a discussion norm against some especially vitriolic criticism; a line beyond which I would think less of a person, even if I didn’t want their speech banned. But most stochastic-terrorism-sayers don’t propose to ban the speech involved, just to create a social norm against it. And “think less of X” and “create a social norm against X” are effectively interchangeable, since social norms are more-or-less defined as things that people think less of you for violating. So am I conceding the whole argument there?"
I don't think this is conceding the whole argument.
There's a big difference between:
(i) I will think less of you for making too extreme a criticism because it seems excessive and suggests poorly calibrated judgement and poor emotional control
(ii) I will think less of you for making too extreme a criticism because I judge it to promote violence and I will judge you to be a wilful violence-promoter.
A nitpick, but I find it funny that there's a part where "Stalin, Mao usurps democracy because no one realize their intention early.". For Mao, I guess everyone knows he's not on the side of democracy and he's defeated multiple times (violently) until he wins. For Stalin, the government he's usurping is not a democracy anyway.
"This puts you in the position that Winston Churchill called “riding a tiger from which you dare not dismount”; you had better invest all your energy into making extremely sure that you and your friends are the ones calling the shots about who can and can’t be criticized."
I accept the argument that you shouldn't condone violence in the abstract, even if it serves your purposes in the moment.
But I never really understood the argument that its in the self interest of any given political movement to avoid creating a norm of political violence, because I just don't think norms affect how how much political violence there is that much. It's not like the violence is being carried out by public figures with who need to maintain a reputation for respecting norms.
A lot of political movements could probably achieve their goals more effectively if they employed some degree of violence.
For example, say vegans could reduce the total quantity of animal products society consumed if the vegan movement started a sabotage campaign against factory farming's infrastructure, or even an assassination campaign against owners and ceos. The risk of inciting reverse violence against vegans would be a tiny cost in the overall cost benefit calculation.
In general any movement that's supported by a large number of people and that opposes a small number of people, or any movement that cares a lot more about achieving it's goals than its opponents care about preventing them, or any movement that has goals that can be achieved without actually needing to pass laws, can probably be confident that the benefits of violence outweigh the risk of changing norms.
There have been plenty of movements that used political violence as a tool (early 1900s anarchists, Western communists during the Cold War, various islamist groups, various separatist groups). It usually wasn't effective. Some separatist causes succeeded, but they had majority support, so hard to say how much the violence helped (or hindered).
Yes, I think this is a stronger general argument against violence than Scott's one.
The Liberal solution ("it’s always legitimate to criticize any person or group however strongly, and never legitimate to use extra-state violence against them") is the best answer to a not all that inherently difficult moral dilemma, but with two very important reservations.
One is that criticism can be dismissed (when it's mild) or censured (when it's violent) if it is based on clearly false facts. For instance, criticism calling for violence against Jews on the strength of the 'blood libel' or of accusations contained in the forged Protocols of the Elders of Zion, both patently factually incorrect, cannot lay claim to immunity and can be banned outright. Similarly, and just as laughably, attacks on Haitian refugees in the United States for allegedly eating people's pets can be dismissed after they have been investigated and found to be pure inventions.
The second objection is the blanket prohibition against extra-state violence under all and any circumstances; taken to extremes, that principal would render any revolutionary or insurrectionary activity illegal. Really? The Warsaw Rising against the Nazi occupiers was illegitimate? The Greek, Serbian and Bulgarian risings against the Ottoman empire were illegitimate? The French and Bolshevik revolutions, never mind that both ended up sliding into dictatorship, were historical errors? And if the lynching of the Stalinist secret police during the (failed) Hungarian revolution of 1956 was an illegitimate act of extra-state violence, will that precedent weigh in the balance if at some time in the future an American mob decides to turn against rampaging ICE agents?
I agree it is very hard to saddle a principle such as Scott's Libera Solution (which I 99.9% agree with) with riders along the lines of "... but there may at certain times be legitimate exceptions to the rule"; such yes-but reservations tend to sully the theoretical purity of sweeping moral rules. But they need to be kept in mind.
I’m unsure of your intention. On the one hand, you have two reservations to the principle that extra-state violence is never legitimate - but your first reservation is that in some cases it is legitimate to respond to criticism by “dismissing” or “censuring” it. I don’t think dismissing or censuring is violence! Just as it’s always legitimate to criticize, it’s always legitimate to direct one’s own criticism against the criticizers.
Minor nitpick first: “Trump and Altman are both still alive” is a strange way to dismiss this category. Both have been targets of assassination attempts. Failed attempts still count as political violence; their survival may say more about their security than about whether the rhetoric inspires violence.
More substantively, I think your fourth proposed principle is close. It just needs to be generalized:
Some accusations don’t merely describe a target. They place the target in a category whose accepted remedies include coercion or violence, and therefore also function as instructions about how the target should be treated.
If you say Trump will enact concentration camps, you’re calling for violence to stop him. You’re implying that Congress and the courts won’t stop him, while the military, federal agencies, police, and contractors will play ball. If the accusation is true, what option remains but violence?
If you say Sam Altman will destroy the world, you’re saying he should be treated like a movie villain holding a doomsday device. Assassination, kidnapping, and death by airstrike all beat the end of the world.
This principle also distinguishes between different kinds of harsh rhetoric:
“Examplestanis are ugly and stupid” does not imply violence.
“Examplestanis are vermin” does. Vermin get exterminated.
Full: https://lenwhite.github.io/blog/steelmanning-stochastic-terrorism/
This strikes me as a very weak post - regardless of whether you think laws like these are *good* ideas (I don't), they have been implemented in a lot of countries. To not look at how this was done seems downright lazy.
Example: Sweden. First off, there's a law about Incitement ("Uppvigling") where you actively call for violence against someone or some group. The illegality here isn't particularly controversial. Then there's the hate-speech law (where "Hets" *also* basically means incitement, but in a less direct manner). Here, the law starts by identifying certain protected groups (ethnic/national/racial groups, religious groups, and groups of people by sexual preference and transgender identity). Outside of targetting such groups, there is no hate speech, so Trump or Altman or even merely political groups are not protected this way.
Next, there has to be a public dimension for attacks on these groups to be considered. If you do it in your own living room, it's very likely legal.
What makes a statement illegal isn't how harsh it is exactly. Swedish has a word "saklighet" that literally means "thinginess" or "to the thing", or perhaps best, "matter-of-factness", but can be best described as a combination of objectivity, impartiality, being reasoned, and having truthful fundamentals. For instance, I'm absolutely free (legally) to say that Israel is committing genocide in Gaza and that all Jews globally who don't criticize this are indirect participants, or that the wide-ranging support for 7/10 among Palestinians demonstrate that they would do the same to the entire Jewish population of Israel if given half a chance. When it becomes illegal is if I describe a group as a cancerous growth on the national body, or that all Examplestanis are ugly stupid vermin, or that they are Satanic cockroaches unworthy of life. Holocaust denial is also included here, although it's unclear if you could get away with a sufficiently calm and reasoned kind (you probably could, if such a thing existed). It's also possible to be convicted for indirect hate speech, such as wearing a swastika, where it's seen to be a symbol of a certain kind of speech. As a rule, it's easier to get convicted for overgeneralizations ("immigrants are massively overrepresented in the rape statistics" is fine, "immigrants are all rapists" isn't).
This is a *coherent* law at least. Obviously it's not exactly defined and courts have to make interpretations, but when is *that* not true?
(There is also a separate hate *crime* law, which isn't a separate rubric, but instead where you target people of a protected group with violence for reason of them belonging to the group, which raises punishments a bit.)
Whether "they're eating the cats, they're eating the dogs" would be illegal would probably depend on whether you can *present* some such cases. If some of them *are* eating the cats and the dogs, then saying it is likely "sakligt", especially if you don't overstate or generalize.
Also, perhaps weirdly to an American, Sweden has considerably stronger Freedom of the Press laws than Freedom of Speech laws - as a rule, the U.S. has stronger Freedom of Speech laws but weaker Freedom of the Press ones. This could lead to surprising outcomes such as it being legal to translate and publish Mein Kampf, but then hate-speech to read certain passages from it at meeting, especially a meeting that is already antisemitic.
That just sounds like "Freedom of speech, but really only for the Officially Approved Speakers to Proles". As an American, I will *so* pass on that one.
An example of the stronger Freedom of the Press laws - it's actually *illegal* to try to find out who an anonymous source is. Things like the U.S. tossing journalists in jail for refusing to divulge their sources is an impossibility here. This is a Good Thing.
That's unequal protection of the law. Journalists have no particular right to secrecy that other citizens don't.
I sometimes feel like the last 20 years have mostly been an exercise in finding ways to say "people who disagree with me shouldn't be allowed to talk".
The main problem I have here with the second-order solution of eschewing all specifically *non-state* violence, is that the majority of the problem people have with vitriolic speech against individuals, groups or ideas they support is going to carry over directly into a worry about state violence. You and I may not think rapists should be summarily beaten to death, but we think all sorts of negative legal consequences should happen to them. We shouldn't want similar legal consequences to happen to redheads because enough speech has convinced enough voters and jurors that redheads are as bad as rapists.
A deontic anti-violence rule with a "non-state" caveat doesn't free us from the enormous responsibility of having the first-level assessments of others' badness in our speech be reasonable.
Thanks for this, Scott. I have been seeing "stochastic terrorism" thrown around on social media and not been too sure what was supposed to be going on; what was the difference between this and normal terrorism?
And the difference seems to be that the term has now become watered-down into a general term of abuse like "racist, fascist, Nazi" etc. In other words, I say you are a stochastic terrorist/engaging in stochastic terrorism and all I mean by that is "You say things I don't agree with/you don't agree with what I believe" topped off, often, with "You should not be allowed say or think that because I am a Special Instance Of Protected Group so less than 100% agreement in full with all I say or want is violence and stochastic terrorism, and if some crazed zealot murders me in the morning it is all your fault".
(Never mind that them being murdered in the morning by a crazed zealot is as likely as me winning the Euromillions Lottery since I never buy a ticket. They like thinking of themselves as persecuted and oppressed for being so special and daring and challenging, so of course they *must* be at risk of being murdered in the morning by a crazed zealot since they are challenging the status quo).
More broadly, speech ethics have become much more complicated in recent years. For example, assuming the release of television programs that talk about suicide in a certain way (like "13 Reasons Why") will actually probably cause an increase in teen suicide, should we allow their release? Is it refraining from any censorship an important Schelling line that shouldn't be crossed? Would it just be counterproductive to intervene? And if we should not place restrictions, is creating such a series an ethical decision to make personally, or should one refrain in the name of "The Virtue of Silence"?
Thanks so much for all the amazing content!
As an individual it is always your responsibility to weigh the likely harms and benefits of your speech and that certainly includes risks of encouraging violence. It's just that trying to enforce any kind of social norms that tries to prevent such "stochastic terrorism" creates a rule that is open to abuse and results in more net harm.
I think a good analogy here is to the morality of breaking up with a partner. There are plenty of times it would be wrong to break up with a romantic partner but we've realized that attempts by society to punish that kind of behavior can be exploited by abusive partners behind closed doors or simply make people miserable.
I think the term “stochastic terrorism” is a term made up by stochastic terrorists projecting their behavior onto others, so naturally this isn’t how it works. Stochastic terrorism isn’t strong criticism or even vague threats.
Stochastic terrorism two guys walking down a street. First guy shouts “John Doe is a nazi!” and the second guy shouting “Shoot nazis!”
Apply to political discourse where progressives claim voting for some bill will cause death, then others say death must be avenged, etc.
Ferdinand Nahimana, director of the Rwandan radio station RTLM, was sentenced for genocide, even though he did not kill a single person, or even directly called for the death of anyone. But the broadcasts he oversaw regularly dehumanized the Tutsi ethnic group, characterized them as mortal threats, and talked about the need to exterminate them. Some smart natural experiment based analysis (that made use of the fact that in the mountainous areas of Rwanda even in the same region radio reception was highly variable) attributed about 50.000 extra deaths to the broadcasts.
That's more stochastic genocide than stochastic terrorism, but the same underlying issue - the perpetrators didn't personally participate in violence, they didn't coordinate or direct violence, nor were they the main cause of the violence (the killings were started and mostly done by government forces). In fact, the US army considered radio jamming but decided against it on freedom-of-speech grounds.
Would you disagree with the ruling? If not, where would you draw the line?
Nahimana wasn't convicted for personally calling for anyone's death, but for failing to stop it: he ran RTLM, had the authority to stop his journalists' broadcasts once they moved to naming names and "go to work" after April 6, and didn't.
RTLM both promoted dehumanization of Tutsis before the genocide (cockroaches, snakes) pre-1994 *and* then in 1994 use quite explicit language to get already-radicalized Hutus to go out and kill. The latter seems to have had a small but measurable effect on actual killings.
I would model Scott as saying the former is in poor taste but not morally responsible for the killings, and the judges in Nahimana's case agree since they threw out this part entirely.
For the record, contra Scott I think even the pre-1994 RTLM had blood on its hands, but the line seems hard to draw.
I don't really participate in the culture wars, so I'm probably underestimating how much the term gets thrown around for all kinds of criticism, but I've personally only encountered it in two fairly narrow contexts:
- Fairly obvious dogwhistles ("It would be a real shame if someone torched the place. Of course I would never do that, but you're just asking for someone... anyone... to, you know?"), where the threshold gets lower the more influential the speaker is ("rid me of this meddlesome priest").
- Bubbles or echo chambers of ever-escalating rhetoric and dehumanization, where nobody truly calls for violence, let alone wants to take responsibility for violence - but would still treat the occurrence as unforeseeable ("just some lone wolf crazy person") yet satisfyingly karmic ("what do you expect if you take people's livelihood away?").
The former is an individual moral (or legal) judgment of a particular person's statements in context.
The latter suggests a general standard not to normalize talk of killing, execution, extermination, or comparisons to vermin. Honestly, if that's where the discussion is heading, find a better circle or cause.
Neither has ever felt to me like a blanket restriction on criticism.
In a Gricean sense, certain prescriptive claims can be implied by formally descriptive utterances, given commonly held assumptions.
"Smoking kills." (Do not smoke.)
"There are leftovers in the fridge." (Eat them.)
"You have ingested a deadly poison, and this vial over here contains the antidote." (Drink the contents of the vial.)
For that last example, it would pretty absurd for the speaker to claim "I never told them to drink what's in the vial." In a different world where people don't care if they live or die, perhaps, but not in ours.
If someone provides me with convincing evidence that a) X is literally Hitler, b) nonviolent means to stop him won't work, and c) violent means might work, then I will be persuaded to kill X, or to support someone doing so. This will be true even if know the person saying this is an extreme deontological pacifist who thinks violence is immoral in literally every case.
So yes, I would say that arguments that would persuade a reasonable person to support violence qualify as incitement. The response to this would be that most of the attackers on your list probably don't qualify as reasonable people.
Also: I think many accusations of stochastic terrorism are just accusations of defamation. It's not "don't hold that horrible opinion!", but "stop lying!". If someone murders someone based on slanderous information, it feels right to hold the slanderer responsible, even if the murder would have still been wrong if the information wasn't slanderous.
Great post. Trying to think of useful examples to bound this. In 2015 after the Republican debate Dan Scavino and Steve Bannon relentlessly attacked Kelly personally and Lewandowski tweeted "gut her" and it led to a year of viscerally negative Breitbart coverage. That led to death threats that drove her to hire a personal security detail.
My instincts strongly align with classical liberalism and a sort of them's the breaks approach here, but I raise it because it's the toughest case I've got, bit of godfather agitprop -- "it would be a shame if something should happen to her."
I can't think of any criminal rule to limit this that wouldn't be misapplied though. Possibly existing civil litigation for defamation or wrongful publicity could cover some of this territory without being overbroad.
I guess you are basing your argument on the idea that every murdered trans person is labelled as trans and actually has their murder reported, but the issue is more complicated than that, as I tried to lay out in my earlier comments. I mean, even violence against the average homeless person is significantly underreported. Anyways, wish you well, it was an interesting conversation
accidental reply to another conversation, mb lol
I mostly agree, but Scott didn't mention the biggest real world example: Islam. Imams can say it would be great if a certain person is killed, and then there's a serious chance that that person will be killed; the imam himself doesn't have to be directly involved in the killing at all. This is also how Mohammed himself had people killed. He would suggest to his followers that so and so ought to be murdered, some young follower would kill that person hoping to get in Mohammed's good graces, and then Mohammed would distance himself from the killer, saying he had no role in the murder. Organized crime bosses do something similar.
There's a real gray area here. If you can't punish someone for obliquely suggesting that so and so ought to be killed, then leaders of large violent groups, like organized crime bosses and imams, can pretty much have people killed with impunity.
When we’re resentful of someone, we crave a phrase that will utterly hollow them out and leave us ascendant on our podium, Self Triumphant. And who isn’t resentful a lot of the time? Being fair-minded is an awful lot of work & I admire Scott’s consistent effort to manage it.
This line of logic is similar to two hot topics regarding university culture in America
- Hazing in fraternities and sororities
- toxic culture or sexual assault issues in college sports
In both cases, universities trying to police their internal organizations' cultures (maybe somewhat with external influence, but often not a full-blown civil or criminal trial)
They have to distinguish between all the power dynamics and mantras and imagery and culture that "creates an unhealthy atmosphere" versus all the fact-finding required to determine who are the specific bad actors, what did they do.
So sometimes they find the bad apples, sometimes leadership is removed regardless of direct culpability, and sometimes frats are completely kicked out. It's a messy system of governance to adjudicate
The UnitedHealth policies that secretly paid bonuses to nursing homes to override hospital transfers and gave nursing homes quotas for hospital transfers are pretty bad.
I think that there is indeed a concept of each hostile individual non-direct-incitement statement adding some microunit of probability to a terrorist attack happening. We *could* call this stochastic terrorism as a general concept, and it could serve as a general reminder to avoid hostile speech if we want to have a society with less terrorism. However, calling any particular individual as someone who is guilty of stochastic terrorism is unsustainable (if they are directly inciting violence, that's direct incitement, we don't need a new term for this).
"If your worldview says that it’s acceptable to lynch sufficiently bad people, then yes, accusing people of being bad is equivalent to calling for their murder, and you have no alternative but to make sure nobody is allowed to criticize anyone you like."
This gets at the root of the issue which I think is going unstated: it is a purposefully fostered expectation in America that you, citizen x, have a right to live under the hegemony of people like you and should expect to enforce a rule that ensures your values are reflected back at you from the majority of the population. This is a completely destabilizing and poisonous belief and yet no one gainsays it. No one argues against it in its own terms. They only accuse some value systems of being nasty. When you have ambivalent tension between what a man can do by right with his life and this expectation that others are going to guarantee that you feel like part of the normal majority, free speech can't last. There's too much factionalization
Ask yourself: who has the whip hand?
I think many people are being hypocrites over this (I think this was also the subtext of Scott’s piece).
You could take a consistent US First Amendment style approach that as long as it’s not incitement to imminent lawless action it’s allowed; or consistently be in favour of UK-style taking a more expansive view of what constitutes criminal incitement.
But it’s just hypocrisy to be only in favour of free speech for your own side, and not your opponents.
"Stochastic terrorism" has been used against the right by the left for ages, usually for far milder examples, and I don't recall any post about it from you. But now that the right is trying to hold the left up to their own standards... suddenly it's something to oppose. This is all who whom.
>Part of how Pol Pot, Stalin, Mao, and the like took power was that not enough people saw the danger early and spoke out about it.
As people on DSL pointed out, this is basically false. Those individuals all engaged in violence before taking power; the problem is that the state was too weak to stop them (or in the case of Stalin, he was the successor to someone else who took power that way).
I will simply ask this: have you read Scott's previous writings on language, and how accusations of language can be misused?
Some examples:
https://slatestarcodex.com/2014/09/30/i-can-tolerate-anything-except-the-outgroup/
https://slatestarcodex.com/2014/07/07/social-justice-and-words-words-words/
https://slatestarcodex.com/2015/06/14/fearful-symmetry/
If so, do you believe that those writings represent someone who is unfairly biased against the right on issues of speech?
Scott has gotten hugely more left-wing since then. Scott-2014 probably isn't unfairly biased. This is Scott-2026. And even then, I don't think he's ever spoken against stochastic terrorism when it was the left flinging the accusations.
1. Where did this "meddlesome" thing come from?! Thomas à Beckett was turbulent, not meddlesome! (I like to imagine that he was simply less aerodynamic than the other priests, who all shaved their legs and dressed in lycra like triathletes)
2. "You should be allowed to criticize powerful people, like Donald Trump or Sam Altman, but not weak people, like Muslims": In his Discworld Series, Sir Terry Pratchett imagined an Assassins' Guild who are legally allowed to 'inhume' rich, powerful people if they become too unpopular, but aren't allowed to harm ordinary citizens. Sir Terry imagines that ordinary people can't really do much harm and so legalising their assassination would do more harm than good but rich, powerful people can, if left unchecked, do so much harm that in practice there was no actual choice between 'some people get killed by assassins' and 'nobody gets killed', only a choice between 'some people get killed by assassins' and 'lots of people get killed in wars' (presumably a modern reimagining of the Discworld could replace 'wars' here with 'dismantling aid programs', 'medical insurance industry practices', etc.). Does this just reduce to the trolley problem?
3. "Some European countries distinguish between permitted normal speech and hate speech (for example, “all Examplestanis are ugly stupid vermin”), on the grounds that the latter actually does encourage violence if left alone for too long. As a red-blooded American, I don’t agree with this distinction": This sort of European 'solution' would be like trying to prevent school shootings by just not letting people have guns. Sure, it *appears* to work almost-perfectly everywhere in the civilised world, but Americans righteously oppose it with every fibre of their beings. Thankfully, it's virtually impossible for Europeans to argue for this sort of 'solution' because they just can't overcome rock-solid arguments like "I'm an American".
Are the arguments in (2) Pratchett's or yours? I only remember that killing "simple people" lacks style and, more importantly, Nil Mortifi Sine Lucre...
Also, that reminds me - Pratchett has a line directly about this kind of thing, from Thud:
Community leaders, Vimes thought. There'd been a lot of talk about community leaders lately, as in `community leaders appealed for calm, a phrase the Times used so often that the printers probably left it set in type. Vimes wondered who they were, and how they were appointed and, sometimes, if `appealing for calm' meant winking and saying `Do not use those shiny new battle-axes in that cupboard over there ... No, not that one, the other one.'
Wow, great excerpt! Thanks for sharing! It definitely captures a real phenomenon, too - but I am not sure how accurate it is for social media discourse? Does that sort of 'plausibly deniable', covert incitement to violence really exist on social media, or in our more efficient modern times have we just replaced them with uncomplicated overt incitements to violence?
I honestly don't know. And at least the stronger version of the stochastic terrorism concept seems to claim that some statements can lead to an increase in violence even if the author did not intend them to...
Sir Terry's, as I said. Specifically, "the only alternative to assassination is war" is in Pyramids, and "Assassins are only allowed to inhume rich and powerful people" is in Men at Arms.
Please don't think either Sir Terry or I are praising the assassins, though! Yes, a reasonable argument for their existence is offered - but they're not presented sympathetically in the novels; Sir Terry treats them with the same cynicism he treats everybody else, protagonist and antagonist alike (cf. 'Nil Mortifi Sine Lucre'...)
Thanks
Do you think it's acceptable to propose specific acts of violence against certain people or groups? Would you call this stochastic terrorism or something else?
I don’t have a strong opinion on stochastic terrorism one way or another, but I do feel like the post is largely fighting a strawman. For starters, the definition Scott gives for stochastic terrorism differs in an important way from what eg the Wikipedia article he links to says:
"Stochastic terrorism is an analytic description used in scholarship and counterterrorism to describe a mass-mediated process in which hostile public rhetoric, repeated and amplified across communication platforms, elevates the statistical risk of ideologically motivated violence by unknown individuals, even without direct coordination or explicit orders."
Or, in the form it has been popularized:
"Stochastic terrorism is the use of mass communications to stir up random lone wolves to carry out violent or terrorist acts that are statistically predictable but individually unpredictable."
Instead, all through the post, Scott seems to be talking about - for want of a better term - “deterministic terrorism.” I think that taking into account that factor makes some of the arguments, like “Trump and Altman are both still alive,” quite weak, and fails to do justice to the idea.
Perhaps as importantly, I feel it is a weird move to detail how various distinctions between allegedly OK speech and allegedly “terrorist” speech break down, at least under certain conditions; and then take what is perhaps closest to the strongman argument for the “stochastic terrorism” concept, the idea that there is speech that constitutes incitement to violence, and just wave it away by claiming “The boundary here is fuzzy, but I think US courts do a good job of navigating it.“ Loads of people have said “Trump is a bad person and I suggest you kill him,” but Trump is still alive! And if US courts have found a good way to draw a line on the continuum between “Go kill X, here’s a gun,” the “will no one rid me of this meddlesome priest” example that came up a lot in the comments, “I wish X was dead” etc., I’d be interested in the details.
More constructively, I think the “stochastic terrorism” idea actually consists of two claims: that some speech leads to an increase in violence (statistically speaking!); and that such speech is bad (and should be avoided, ostracized or banned). And the first part seems to be a factual claim, one that could be checked. Claude says that "some evidence, contested, more suggestive than robust" is the accurate characterization — not "confirmed by statistics" and not "debunked" either. But as far as I see, Scott has not even asked the question, at least in the post.
If the factual claim turns out to be wrong, we are done. If it turns out to be correct (or if we don't know, or if we are interested in the philosophical/moral aspects of stochastic terrorism), I think the question of where that line between “true incitement to violence” and “simple criticism” lies deserves much more attention.
Also, I think the "liberal solution" is not a solution to the problem postulated by "stochastic terrorism." It's like saying "my solution to school shootings is that everyone can have whatever guns they like, but nobody is allowed to shoot anyone." It might be a legitimate stance, but it is not, in any meaningful way, a solution to the original problem.
I think the liberal could always just bite the bullet and say "yes, some speech will lead to lethal violence, but the alternative is to not have a market of ideas, which is worse".
The fundamental problem here is that western societies are increasingly filled with a diversity of peoples and religious ideologies whose interests are perceived to diverge at an existential level for reasons that likely cannot be realistically resolved through rational public debate. The violence in the public square is a side-effect. (I of course include left-wing-progressivism as a kind of religious ideology.) This dramatically affects the cost-benefit analysis for the market of ideas vs. violence suppression.
I would also say a pertinent criticism of "the liberal solution" is that it presumes the government is a neutral arbiter and can therefore be trusted with a monopoly on force, when it's increasingly obvious that the governments of the western world are not "neutral" in any meaningful sense. They are battlegrounds where the struggle for power between peoples and religious ideologies are being fought, and where the left is generally much more entrenched.
That's especially true since we can dispute who is and is not weak. I would say Muslims are anything but weak, having vast numbers and the ability to kill just about anyone for blasphemy.
But that proves your point
Stochastic terrorism is the one of the early stages in the life cycle of the state. The United States is built on incendiary rhetoric whipping up crowds to spontaneously attack British loyalists. How else do you know that there's enough anti-British sentiment among people willing to do violence to start organizing militias?
It's mostly epistemic; how your political movement knows what it itself is and whether it's ready to transition to its next phase.
When it comes to research, do you think researchers should make the rigor/getting-information-out-there tradeoff differently depending on the study's chances of contributing to violence?
For example, suppose there's a study suggesting that managers in brown suits are more trusted than managers in black suits, and there's another study suggesting that immigrants from X country are creating Y negative effects. Should the author of the immigration study collect more data or invest more time into their analysis before publishing something like 'Immigrants from X country may be creating Y negative effects'? (Compared to the managers study author publishing 'Managers in brown suits may be more trusted than managers in black suits')
"This liberal solution isn’t trivial. It requires the separate liberal norm of always being against extra-state violence"
unless, of course, said extra-state violence comes in the form of a CEO committing negligent homicide.
the US de facto legalized white collar crime, permitting even homicide, provided it juiced shareholder value, and you want us to feel bad about vigilantism?
39% favorability for Mangione is not nearly high enough. Brian Thompson was never going to see a long prison sentence or suffer more than some token fines, no matter how many grieving families were left in his wake.
Don Blankenship killed 29 people and walks around a free man after serving a year in prison.
OK, I guess birdboy volunteered to be the test case as to whether it's OK to endorse murderous vigilante "justice" against their enemies here.
“if a violent act occurs, it’s 100% the fault of the perpetrator, and 0% the fault of other people who previously criticized the victim.”
I agree that it’s an exercise in futility to hold people responsible for all the actions, whether positive or negative, that may have been influenced by their behaviour or speech one way or another. Not to mention that whoever is the final judge on such questions would hold an arguably totalitarian level of power.
That said, I think the “100% … 0%” framing risks undervaluing the role that language plays in shaping a culture which either rewards, disincentivises or trivialises various behaviours, ranging from just being hostile or unpleasant online to acts of violence or outright terrorism.
Reality is rarely so black and white, but if, for example, someone actively dehumanises a group of people, spreads malicious falsehoods about them and portrays them as an existential threat, they would still be regarded as having contributed to a climate in which violence against that group became more likely, even if they never explicitly incite violence themselves. Obviously in real life questions of how and to what extent someone is responsible for violence that follows from their speech are much murkier and more controversial, but that doesn’t mean the answer becomes zero responsibility whatsoever either.
Still, I’m fully agreed with your pushback against the use of “stochastic terrorism” as an accusation, and as you say, there are clearly some difficult legal questions to be navigated here (the intricacies of which I won’t pretend to understand) about when critical or hostile language crosses a threshold.
The hate and violence will simply continue being a background hum of modern, diverse life
What if someone really was trying to get people killed, and we had memorandums and emails showing the receipts? Would intent make a difference?
I mostly agree wit the liberal stand being described here (it's simply traditional US freedom of speech norms) but I am interested in opinions regarding the following scenario:
Let us say that I am a member of demographic category X. Person A proclaims on mass media that people in my category are terrible, lying about ourselves, and are in some way harming society. Person A offers no evidence, but plenty of hostile rhetoric.
Afterward, I notice a significant increase in the number of people confronting me and people in my category, threatening us, and making reference to person A's message. When people in my category point this out on mass media, person A continues proclaiming their rhetoric.
What degree of moral liability does person A bear for the increasingly threatening behavior being directed at people in category X? What if someone gets hurt?
My take is Scott is mixing up moral and legal culpability in his post. We should have norms against all the things he wants to have norms against. Don’t call people Hitler barring extreme cases, don’t promote lynching, etc. But legally, the US system is actually pretty good! We don’t want the government getting into the weeds of deciding who can’t criticize who and in what ways. A lot of the early “cancel culture” criticisms were also about hecklers vetoes and the legal right to have a discussion without being physically disrupted (also a legal matter).
The biggest problem we have is that pretty much everyone in public discourse (or maybe just social media) conflates these two things all the time, and often lumps groups of people together. Person A says “that group are cockroaches”. Person B and C say “get them! They’re at 123 Fake street”. The law says person B and C crossed the line (correctly). On social media we are 100% in the right to criticize person A too for what they did.
Now person D and E say “that was bad” and person F says “Person A should go to jail”. Then everyone is arguing about what person F said instead of agreeing that D and E are correct. I don’t know the cause of this exactly. Possibilities that spring immediately to mind are
1. This is just “social media amplifies extreme views” problem.
2. People like to argue about technicalities and forget to say “yes of course” to straightforward statements.
3. The Bad People (insert your favorite group) are manipulating social media to sow discord.
4. People on social media have difficulty distinguishing people on Groups, so A gets lumped with B and C, and (D,E,F) forms another group in people’s minds.
Thank you for sharing that. It was detailed and thoughtful. However, I am still interested in answers to the question I asked.
I would say in your version person A is morally culpable for inciting but not legally culpable. The degree depends on the degree of lying and incorrectness.
Sometimes I feel like rationalists want to be able to make rules like this without saying “it’s ok to say (morally) correct things and not ok to say (morally) incorrect things”. But to everyone else I think that’s obvious.
I think it's more about accountability. A norm with no mechanism of enforcement (even a casual one) cannot protect anybody. Given that behavior A is "bad" (however you want to define that), how do we, as a culture and society, best ensure that the rate of behavior A remains low?
The line dividing "create a social norm against X" and "create a social norm against X that involves killing proponents of X" is sufficiently bright.
Excellent essay, but of course I have to only mention the thing that annoys me. I followed the link about violence against abortion and there actually aren’t a lot of them and they were a long time ago 1982 and 1997 as I expected
I think if the speech was hate, and not legitimate sincere criticism, then the speaker should be considered partly responsible for the incited violence. Otherwise, it is entirely the violent person’s fault.
So saying “insurance companies are evil monsters and every leader should be jailed” is a contributing cause, but “insurance companies have systematically made it impossible to get good healthcare” is okay.
> This liberal solution isn’t trivial. It requires the separate liberal norm of always being against extra-state violence - a norm which is currently less than entirely secure.
(a) This norm has never been held anywhere.
(b) It wouldn't be a good norm if it was held somewhere.
"The boundary here is fuzzy, but I think US courts do a good job of navigating it.", maybe with true threats but that's about it. They are notoriously hostile to true any other speech from true solicitation to true conspiracy to true fraud to true obscenity. You can get twenty for making a joke to an undercover.
Given the hundreds of thousands in US prisons for speech crimes, a significant number who are actually factually innocent, it blows my mind at your confidence in the US courts here.
“So am I conceding the whole argument there?” No. I actually think that one should criticize the use of the word “stochastic terrorism” on the grounds that it blurs the line between the legal world where we enforce bans on behaviors and the moral one where we admonish, chastise, and socially penalize. Both political sides blur this line all the time and it takes significant work to maintain. It is of course worth maintaining as you pointed out regarding riding the tiger.
I think you’re pulling a mot and bailey yourself although unintentionally. You jump from arguing against the idea that there should be blanket prohibition on certain types of speech, such as criticism or overly harsh criticism to arguing that it is 0% your fault. If your speech is predictably going to cause problems or at least has some expected probability of causing problems. There are several intermediate positions that you never discuss like for example, comparing expected cost and benefit and only being against cases where the expected benefits are not worth it. Similarly, it seems intuitively likely that more vitriolic speech is probably more likely to cause violence, but instead of dealing with this, you just point out that other types of speech can also cause violence and call it case closed. What’s particularly bizarre is that your objecting to creating a social rule against stochastic terrorism and argue for a liberal nor that people who are particularly vitriolic should just move to places where it’s acceptable to say such things. Yet there are many types of speech for which you yourself would consider this, not adequate, for example, defamation, or publication of private details, obviously, those are arguably different, but the point is that liberalism has always been open to social rules against speech. So acting like it’s inherently a contradiction of liberalism to have social rules against speech makes little sense. To be fair, you never actually did come out and say it’s against liberalism but does seem like most people who take that away from your post and yet it seems perfectly compatible with liberalism to have numerous controversial social rules against speech. For example, lots of your fellow travellers, think there should be social rules against invocations of sarcastic terrorism. I doubt you want to declare that all those people are not liberal and in any case if you interpret liberalism, so narrowly almost nobody would find the position describes their views, and it would become a much less attractive idea.
Event ignoring the diversion about liberalism, they are perfectly obvious middle ground positions that you never consider. I’ve mentioned this before, but for example, look at the guidelines that Z verse one recommended when advising people how to discuss AI risk. I doubt you think those social rules are bad, and if you do you need a much more substantial argument than what you have provided, because most people would consider that kind of rules calling, for, for example, no implications that political violence will be the inevitable result. If you demands are not met to be completely reasonable and nothing that you have said addresses such common positions. You just spent entire post attacking a week man extreme version of sarcastic terrorism argument where any criticism or at least any sufficiently harsh criticism is automatically of the table and it’s a bright line rule and then jump to your own bright line rule that actually, you can never object to any violence caused by a speech by criticising the speaker because it isn’t the speakers fault which is just absurd. Why speech so special that we have to pretend actions don’t have consequences within almost any other arena. You would find this absurd. For example, if I sell you poison and you use it, it would be completely legitimate to enquire whether I took a appropriate precautions and that even applies to less obviously dangerous substances like materials which have both beneficial and destructive uses. You just declare your position without providing any account of Y speech has this exceptional property and which things exactly would have this exceptional quality. This is just not a good argument because you argument at most only supports not calling for outright bands on criticism, which nobody actually believes would be a good idea.
> Probably there are some cases where extra-state violence is acceptable, for example assassinating Hitler or overthrowing an illegitimate government. I don’t want to rule this out by fiat, and I think it’s wise not to discuss the conditions permitting them too specifically
I do not understand this part. Why is not better NOT to discuss e.g. when to rebel against the government?
In my view, it is essential to establish a clear distinction between the mass media and political rhetoric on one hand, and the speech of ordinary citizens—who are neither professional journalists nor politicians—on the other. Regarding popular influencers, a specific threshold should be defined beyond which a blogger legally transitions into a professional journalist. Logically, this threshold ought to be financial: specifically, the point at which content creation begins generating revenue that serves as their primary or highly substantial source of income.Once this framework is established, the following principles should apply:Ordinary citizens are permitted absolute freedom of speech regarding any individual or topic.Individuals classified as professional journalists and politicians face severe, qualified constraints on their speech. Comprehensive statutory restrictions must be enacted to limit the freedom of expression for this specific category of individuals. The precise scope and nature of these laws remain within the purview of the Legislator.
The issue is when the public conversation creates a permission structure for stochastic terrorism to occur. For this to happen you need more than just criticism of bad ideas or bad people. You also need a general impression, supported by high status people of some kind of authority, that it would be good if harm were to come to the bad actors. When we see popular influencers and politicians celebrating deaths of the political opposition that's an example of something that increases the chances of it occurring. That kind of communication is clearly unprofessional and has moral implications.
There's a category of speech that I'd say falls outside the bounds of "incitement to violence" but implicates similar potential impacts. I think that "stochastic terrorism" could be a useful term if it were redefined as:
1) an actor with a meaningfully large platform
2) willfully spreading negative misinformation about a person or group of people, when
3) that speech would be identified by a reasonable observer as having been delivered with the intent of creating a sense of imminent threat, a threat which
4) cannot / will not be mitigated through existing institutional channels.
Being able to call this out when we see it (and, hell, at least raising an *orange* flag when we notice speech hitting 3 out of 4 requirements) enables social enforcement in a way that is at least much more difficult when we don't have useful descriptive terms. I see that as supportive of the cultivation of a pro-free-speech society.
I just published my essay that tries to re-frame this concept of "Stochastic Terrorism", acknowledging that the term has been abused, and that even its most horrid examples should never result in state-imposed consequences, but that we must ostracize real hateful bigotry out of polite society: https://elevin11.substack.com/p/dismantling-stochastic-terrorism
I don't think you distinguish strongly enough between the ethical and the legal here. One may be free speech absolutist to an absurd degree while considering stochastic terrorism morally equivalent to murder, if they expect the results of prosecuting that murder to be even worse than not to. Just like someone may consider Donald Trump equivalent to Adolf Hitler and even want him striken by God while not wanting to cause his death by man - except more reasonable in my humble opinion. The oposite is also reasonable. One may consider they're assassination of an aspiring dictator morally justified but not want it legalized or even pardoned in retrospect. It doesn't sound that unreasonable to me to put calling someone an aspiring dictator in the moral-but-illegal basket also, when done in good faith in a society in which such assassination is considered moral but not legal.