Stochastic terrorism

Paul Alper points to this post by Lee Moran:

Alexandria Ocasio-Cortez Rips Tucker Carlson Audience Over “Death Threats” . . . “Every time that dude puts my name in his mouth, the next day, I mean, this is like what stochastic terrorism is,” the New York Democrat said on Tuesday’s episode of “The Breakfast Club” radio show.

Alper clarifies:

Not that I fully understand the term, “stochastic terrorism,” but here is what she means:

“It’s like when . . . you use a very large platform to turn up the temperature and target an individual until something happens,” she continued. “And then when something happens, because it’s indirect, you say, ‘Oh, I had nothing to do with that.’”

Googling turned up this definition from wikipedia:

Since 2018, the term “stochastic terrorism” has become a popular term used when discussing lone wolf attacks. While the exact definition has morphed over time, it has commonly come to refer to a concept whereby consistently demonizing or dehumanizing a targeted group or individual results in violence that is statistically likely, but cannot be easily accurately predicted.

The term was initially used to suggest that a quantifiable relationship may exist between seemingly random acts of terror and their intended goal of “perpetuating a reign of fear” via a manipulation of mass media and its capacity for “instant global news communication”. For example, careful timing and placement of just a few moderately explosive devices could have the same intended effect as numerous random attacks or the use of more powerful explosives if they were shrewdly devised to elicit the maximum response from media organizations. . . .

A variation of this stochastic terrorism model was later adapted by an anonymous blogger posting on Daily Kos in 2011 to describe public speech that can be expected to incite terrorism without a direct organizational link between the inciter and the perpetrator. The term “stochastic” is used in this instance to describe the random, probabilistic nature of its effect; whether or not an attack actually takes place. The stochastic terrorist in this context does not direct the actions of any particular individual or members of a group. Rather, the stochastic terrorist gives voice to a specific ideology via mass media with the aim of optimizing its dissemination.

It is in this manner that the stochastic terrorist is thought to randomly incite individuals predisposed to acts of violence. Because stochastic terrorists do not target and incite individual perpetrators of terror with their message, the perpetrator may be labeled a lone wolf by law enforcement, while the inciters avoid legal culpability and public scrutiny.

The wikipedia article continues:

In their 2017 book Age of Lone Wolf Terrorism, criminologist Mark S. Hamm and sociologist Ramón Spaaij discuss stochastic terrorism as a form of “indirect enabling” of terrorists. They write that “stochastic terrorism is the method of international recruitment used by ISIS”, and they refer to Anwar al-Awlaki and Alex Jones as stochastic terrorists.

I’d never heard about this before, but, yeah, it all makes sense. Back in the time of the World Trade Center bombings, people talked about the idea that a small amount of terrorism can still cause general terror because of a general uncertainty about where it might happen next. There are two ways that this new idea of stochastic terrorism goes further. The first is that, instead of some existing group such as ISIS planning terror but you don’t know where it will happen, it’s a nebulous group of potential terrorists, and the uncertainty is who will be the extremist who does the terrorist act. The other thing is this idea of deniability or indirect action, so that someone like Alex Jones can spread lies and violent rhetoric without knowing exactly who among his listeners will personally harass the people he’s talking about. “Stochastic terrorism” seems like a useful term, in that a key part of what makes it effective is the unpredictability of where it’s coming from.

49 thoughts on “Stochastic terrorism

  1. Reading this, I was immediately reminded of the Brunswikian lens model, in which a sender conveys a message to a receiver. In the middle of the model you have cues, through some of which the sender conveys his message, some of which affect the receiver, and some of them actually do the full monty of transmitting the sender’s message to the receiver (a process called cue utilization). One can think of stochastic terrorism like a Brunswikian lens model. In the present example,Tucker Carlson is the sender, AOC is the receiver, and there is an unknown number of cues (potential terrorists) in the middle who get triggered by the sender’s message, some cues hate AOC regardless of what Tucker Carlson said, and some of them pass Carlson’s message on as death threats or worse to the receiver. But this lens is like some kind of nebulous blob in the middle between sender and receiver. You can’t really delineate and separate the “cues” very well here.

    I don’t know whether Brunswik’s model is a good way of thinking about stochastic terrorism. But it was the first thing that popped up in my mind. So much for the illusion of disciplined thinking…

    • In this case the theory is Carlson isn’t talking to Cortez, he’s talking to several thousand deranged potential violence perpetrators. The message is “AOC is one of the targets” and the goal is to incite one of them to take action. Of course there are other potential targets too. For example, drag shows, or children at school in liberal suburbs or whatever.

      Of course there’s been some of this from the other direction too, the guy who turned himself in outside Kavanaughs house when he couldn’t quite go through with murder for example. Though the vast majority of political violence has been right wing in the US.

      I actually thought political violence in the aftermath of the COVID epidemic emergency phase would be much higher, particularly after the supreme court rulings last summer and the general instability that was evident in the airplane passenger violence data. Fortunately it wasn’t as bad as I’d feared.

      • I didn’t mean to imply that he’s taking TO Cortez. It’s that his messaging targets her in some way. In a sense he’s marking her as a target, I guess…

      • ” Though the vast majority of political violence has been right wing in the US.”

        Were you awake in 2020? This is the same thing that always comes from the left: violence on the part of the left is justifiable protest. Violence on the part of the right is “terrorism”.

        I’m not about to defend Carlson, but it’s interesting that we don’t even have to hear what he said to assume he’s guilty of inciting terrorism – which is all the more bizarre when the claim comes from an established liar like AOC, who’s every bit as much of a liar as Trump ever was. Meanwhile, AOC is hardly disinclined to take the stochastic offensive:

        “I think the real question is, who DOESN’T think of the Allied victory over the Nazis when they think of ways to deal with global warming?” (Are people who oppose climate change policy “Nazis”? Do they pose an existential threat?)

        Regarding the passing of RBG:
        “Let this moment radicalize you…Let this moment really put everything into stark focus because this election has always been about the fight of and for our lives,” and “If Mitch McConnell is not going to honor RBG’s final wish, we will,” she said.

        “The U.S. is running concentration camps on our southern border and that is exactly what they are. That is exactly what they are. ” (never mind the fact that people come to them voluntarily with ample warning of the conditions in them – in fact that’s the whole point of the conditions in detention: to deter people from coming).

        “If his being a billionaire is predicated on paying people starvation wages and stripping them of their ability to access health care…”

        Reminds me of EWarren claiming that Republica tax cuts would kill people!!! Meanwhile AOC is all about defunding the police, which actually is killing people….

        • I should have said “murders”. There’s plenty of violence on the left, burning buildings or throwing rocks at police or whatever, usually not murders though as far as I know.

          I’m relying on various sources for that, rather than any personal analysis. Examples would include this recent article:

          https://www.washingtonpost.com/politics/2023/02/28/extremism-right-wing-deaths/

          If you want to point to 50 or 100 political murders committed by leftists in the US which were left out of that data, please do!

        • The ADL data is not credible, Daniel.

          “Drill down further into the data and you’ll find other cases are even less clearly political: prison gang members engaging in pedestrian criminal violence; white supremacists killing their wives; people with mental illness acting on elaborate delusions that sometimes include references to right-wing conspiracy theories; people embroiled in criminal trials or child custody disputes who have become enamored of “sovereign citizen” theories that tell them the state has no right to interfere.”

          From the ADL: “One of the most striking features of white supremacist murders is the large proportion of non-ideological killings to ideological killings,” the ADL wrote in its most recent report. “Over the past 10 years, only 86 of the 244 white supremacist killings (35%) were ideological murders.”

          https://www.washingtonpost.com/opinions/2022/05/23/beware-data-right-wing-violence-buffalo-shooting-racism/

        • Emily, violence is a complicated topic. I think it’s unfair to say that the ADL data is not “credible” in some kind of blank way. All the events at least did occur right? They’re not making up events. It’s about how they classify the events. I am open to the idea that was poorly done. Let’s see a better classification and analysis. McArdle’s article is hardly anything, it’s basically saying “this might not be so good, look here’s a few spot checks that don’t look great, but ADL themselves point it out…” but it’s not a reanalysis by any means.

          Hey, if it’s not so good, let’s see a better one. Even if you cut what they call right wing political violence in the ADL data in half, it’d still be a lot more than what they call left wing political violence. And the ADL is not the only source I’ve seen.

          https://extremism.gwu.edu/sites/g/files/zaxdzs2191/f/Anarchist%20-%20Left-Wing%20Violent%20Extremism%20in%20America.pdf

          On pg 15-16, they manage to call out a total of 7 incidents over a 5 year period by “left wing” extremists, none of which involved loss of life…

          It’s just hard to deal with this topic because it’s so so so damn political. I am not sympathetic to either left or right politics in the US at the moment. I’m very anti-violence. It bothers me if there’s a lot of “ordinary” violence committed by people who are radicalized towards violence by prison gangs. But I agree that it should be classified in a different category.

          It still shocks me that we don’t have every single person killed by violent confrontation in the US from 2000 onwards in a database with categorization for each incident. Yes I realize it’d be 200k people or something, but hey, how are you going to make progress on dealing with violence in society if you don’t track it?

        • Googling “violence in 2020”:

          #1 link Times Magazine: 2020 Ends as One of America’s Most Violent Years in Decades

          “This year, many Americans have experienced significantly higher levels of violence both wrought on and within their communities. Gun violence and gun crime has, in particular, risen drastically, with over 19,000 people killed in shootings and firearm-related incidents in 2020. That’s the highest death toll in over 20 years, according to data from the Gun Violence Archive (GVA), an online site that collects gun violence data, and the Britannia Group’s non-partisan site procon.org.

          This total includes victims of homicides and unintentional deaths but does not include gun suicides. And despite there being no “large-scale” shootings in 2020, the number of mass shootings—which are classified as an incident in which four or more people are shot and injured or killed—has actually risen, drastically, to over 600, the most in the past 5 years and a nearly 50% increase in 2019’s total.”

          Rest of search page: more references to the above data.

          Googling “political violence in 2020”

          All first-page references were to “stop the steal” rallies, e.g.,

          “The post-election to inauguration period, this year — the 11 weeks from Wednesday November 4, 2020 to
          Wednesday January 20, 2021 — is a unique period of the US electoral calendar. Data on political mobilization
          and violence from this period tells us something both about what we might expect in the post-election period
          of future elections, and also offers a first glimpse of what more immediate political violence we might see in
          America following the Trump presidency.

          While the focus of this brief is on recent events, political violence in the United States is not a new phenomenon,
          but rather one we have lived with throughout our history. Political violence today can look like partisan
          violence, or violence targeted at groups because of race, religion, or other group factors, perpetrated by the
          state or non-state actors. Electoral violence is one subset of political violence. In highly tense elections like
          the 2020 cycle, models for post-election mobilization and even violence often focus on the losing candidate
          and their supporters. Had Joe Biden lost the election, the post-election mobilization would undoubtedly have
          looked different. With Donald Trump’s loss, however, increased mobilization during the post-election period of
          2020 and 2021 was expectedly concentrated within his supporters. However, the nature, tenor, and ultimate
          actions of both Trump and his supporters merit specific attention.”

          Sounds like right-wing gun-love and political violence to me. What am I missing?

        • Daniel: “they manage to call out a total of 7 incidents over a 5 year period by “left wing” extremists”.

          To be clear, the report you reference only looks at “anarchist” events. Left-wing is broader than just anarchist. If it extended through 2022, the data probably wouldn’t have included the attempted murder of the Louisville, KY, by a BLM activist. Would it have included molotov cocktail throwing lawyers in NYC? Or the man who went to DC to assassinate one of the Supreme Court judges?

          And when discussing violence of right-wing vs. left wing vs. any other “political-motivated” group, then the reasons for the violence actually does matter. You shouldn’t use non-ideological murders when the characteristic of interest is how ideology affects violence.

        • JFA, I agree about the motives matter point. But I will say that it appears far more likely that white supremacist groups are recruiting from criminals and in jails and other ways that involve recruiting people “prone” to violence. While I’d recommend putting them in a different category I think it’s an interesting question how many “ordinary” murders are committed by people with White supremacist organizational ties vs ordinary murders committed by people like members of the Socialist Rifle Association.

        • “Sounds like right-wing gun-love and political violence to me. What am I missing?”

          Jim V:

          We had declining gun violence for 25+ years, and now when it explodes it’s because of gun availability? Same goes for mass shootings – 200+ years with very few gun restrictions, and now the rise in mass shootings is because of gun availability? Jim V, if you’re not already a social scientist, you should try out as a walk-on. I think you have all the tools.

          Your google on “political violence” is just a classic example of left wing bias. Did you miss the multiple riots in 2020? Cars burning? Entire city blocks taken over with police stations abandoned? Did you not see the railroad tracks in LA littered with the trash of the stuff stolen from containers on rail cars? Entire companies are pulling out of Portland OR because the crime is so bad. This crime wave is directly attributable to the actions of left-wing political leaders.

          The mass rioting and crime of the left is never defined as “political violence” – it’s always justified by some supposed grievance or injustice.

    • Jbay:

      The funny thing is, I wrote this post months ago and had forgotten it. When it popped up, I was expecting it was gonna be one more thing on those academic bozos who call us “methodological terrorists” or “Stasi.” I was surprised to see that it was about actual terrorism. Scary stuff.

      • Andrew is not the only one who forgot about it. I did as well; I barely remembered that I had passed it on in November, 2022. The adjective, “stochastic,” is really useful, and sounds so high-class, that it is eye catching enough to engage the general public as its (more or less) synonym, random, is not.

        • I originally confused it with “scholastic terrorist” which is someone who tries to win arguments by quoting medieval philosophers.

        • j bales wrote. “I originally confused it with “scholastic terrorist” which is someone who tries to win arguments by quoting medieval philosophers.”

          But maybe we need to consider stochastic scholastic terrorists, who randomly pick the arguments from medieval philosophers?

        • Martha,

          William of Ockham supposedly said “Entia non sunt multiplicanda praeter necessitatem” (Entities must not be multiplied beyond necessity)

  2. The legal system has a binary bias: your cumulative utterances either do or do not incite some particular act of violence for which you might be held liable. There really is no legal language for “increased the probability of violence” as a culpable act. And because of this, there is no liability for increasing the probability of an act when no act occurs, and there is a concomitant potential liability for trivially changing the ex ante probability that some event might occur that in fact occurred. This particulalry prominent in toxic tort cases, in which, for example, the defendant is held liable for *all* cases of some cancer even though they might reasonably be deemed responsible for 5% of them and owe nothing to people who did not get the cancer even while they subjected them to higher risk. There are some exceptions to this,as when a defendant is made to buy credit monitoring services for people whose credit card information they allowed to be stolen. Maybe we should make Fox pay for AOC’s bodyguards and get the Washington Post to pay for Bret Kavanaugh’s.

    • How does one legally define “increased the probability of violence” without a metric distance to measure harm from? That is, if violent rhetoric doesn’t lead to violence, then it doesn’t increase the probability of violence and is therefore not legally definable as causing violence. If, however, it does lead to violence and therefore increases the probability of it, then it is legally definable.

      • Suppose you have a bias against black candidates for employment such that, all things equal, a black candidate has a ten percent lower chance of getting a particular job. That particular probability is estimable with the right data. You might compensate all black applicants, whether or not they got the job, or you might compensate only those black applicants who didn’t get jobs, but in either case you can create a metric distance.

        It’s much harder to get the data to estimate a similar model for propensity to commit violence, but the same principles apply,

        • Your assumption (in both your examples) is that a probability distribution exists to place a likelihood of an event occurring. For instance, the black candidate is not hired. My assumption is that one can’t estimate the likelihood of violent rhetoric leading to violence (and therefore a probability distribution doesn’t exist).

          You may be right. The way to determine this is a threshold of when “enough” violent rhetoric is directed towards someone/a group/ethnicity/religion, it leads to violence. If you can point to a dataset or analysis that shows this happens, then you’ll have disproven my hypothesis.

        • Sam, could you explain what you mean by “My assumption is that one can’t estimate the likelihood of violent rhetoric leading to violence (and therefore a probability distribution doesn’t exist)”? I don’t think you’d suggest that if an accurate probability estimate is not possible then the probability is zero, but I am having trouble finding another interpretation.

        • Phil, I’m not sure if I’m understanding Sam, but I think the idea that “a probability distribution doesn’t exist” is somehow different from “the probability of X is zero” which is another way of saying “we have infinite information about the probability of this occurring… because we have a mathematical proof that it can’t”.

          I think it’s wrong to say “a probability distribution doesn’t exist” but I’m pretty sure Sam isn’t saying “the probability is zero”

        • Phil, DL:

          You’re right, DL, it was wrong to say the probability distribution doesn’t exist. More accurately: one likelihood is indistinguishable from another. For instance, let’s say that Tucker Carlson says something inflammatory about AOC. Let’s also assume it’s repeated, in various forms, online and not. If someone tries to physically injure AOC, how does one distinguish between the cause of provocation, Tucker Carlson, and the many similar variants of the message that quickly circulated afterwards? If you assume that Tucker Carlson shouldn’t make these statements and therefore the violence would stop, how do you know that any statement made before this statement wasn’t also the provocation?

          Also, to distinguish between the likelihood of a black candidate who faces discrimination, a counterfactual exists. There are, presumably, employers who are less discriminatory (open-minded) and therefore the black candidate has a better likelihood of getting hired. Why? Because the natural experiment can be repeated. I don’t think this is true with regard to violent rhetoric.

          In any case, I would like to see either a formal treatment of this topic, using data that closely mirrors what’s being discussed (“random terrorism”). I’m only hypothesizing what I think makes sense and haven’t seen any published papers or datasets on the correlation between violent speech and violent action(s).

    • Yep. “Stochastic terrorism” is just a useful helper concept of the “words are violence” crowd that wants to punish and pathologize criticism of their various in-groups. The ironic thing of course is that AoC’s criticisms of TC can also be construed as stochastic terrorism as well. The people who invoke the concept are not interested in doing so with any sort of impartiality or consistency – it’s just a somewhat-useful rhetorical bludgeon.

      • Keras:

        You missed my point. I asked 1) for data to validate the point 2) implicitly pointed out that casualty would be difficult to prove. Also, if you think AOC is invoking “terrorism” by criticizing the violent threats against her, you should look for a forum other than one that talks primarily about evidence, data, and math.

  3. Pinning a specific act of terrorism to a toxic environment that might have triggered it is a similar problem to attributing a specific natural disaster (hurricane, flood) to global warming.

    Though I would guess that we have better data in the meteorological case.

  4. Stochastic terrorism has long been employed against climate scientists, for example:
    https://web.archive.org/web/20120415114127/https://thinkprogress.org/climate/2012/02/01/416510/price-of-truth-limbaugh-operatives-encourage-abusive-hate-mail-at-female-evangelical-climate-scientist/

    As in that case, Marc Morano put scientist’s email addressup on his blog, a clear signal to his followers.
    In another, he put up Kerry Emanuel’s email: https://www.theguardian.com/environment/2012/jan/13/us-climate-scientist-wife-email-hate

    In the movie Merchants of Doubt(2014), he even bragged about this tactic.
    Watch 3 minutes, starting here, including Ben Santer, Michael Mann and Katharine Hayhoe reading emails they get as a result.
    https://www.youtube.com/watch?v=cRWEqbl1wmk&t=3367s
    Ben Santer found an eviscerated rat on doorstep, as guy in Hummer drove away yelling obscenities.
    Michael Mann got a letter with white powder, among the many email death threats.
    (I know all these folks, but there are lots more.)

    Of course, these days, many doctors and election workers (like in GA) get similar treatment.
    Someone with a platform points at people they don’t like … and have plausible deniability when attacks happen.
    Of course, many who participated in J6 were inspired this same way.

  5. You missed a chance to comment on the other statistical part of this, when AOC says Tucker Carlson is “110% one of the largest sources of death threats”. That’s like Anchorman levels of bad statistics.

    • Aj:

      Also the part where Carlson calls her a “rich, entitled white lady.” All of a sudden, being rich and white is a bad thing?? Either Carlson has some serious issues with his own background, or his writers are trolling him.

      • “All of a sudden, being rich and white is a bad thing??”

        If she claims to be poor and “of color” and she is actually a relatively wealthy and of Spanish, rather than indigenous, decent, that fact would be relevant to the discussion. I don’t know that she actually claims that, but she sure doesn’t mind if people get that impression.

        • She worked as a waitress and bartender for several years after college, so I assume she’s extremely wealthy. Most waitresses are, right? Not all, but if you had to bet?

        • Phil: the fact that someone worked as a waitress or some other low-paying job doesn’t have a lot of bearing on their familial wealth. A friend of mine owns a real-estate office in Kirkland WA. He’s very comfortable but his kid (with a PhD) is doing CC contract teaching for whatever it is now $2K a month or something (when did CC contract it was $1200/quarter/class). When I was struggling on TA wages, many of my TA/RA colleagues were zipping off to HI and Vail for family paid vacations over xmas and during the summer. They made very little, but the family filled in the gaps. No doubt my realtor friend provides good wheels and vacations for his “struggling” kid.

          Also waiting tables might might be a crappy job at Dennys but at nice restaurants waitstaff can make good money. When I was working in restaurants in the early 80s, wait staff at a good restaurant could pocket $100 in tips on a single weekend shift. Even as a cook at a Denny’s-type restaurant I regularly scored $10 in tips on Friday and Saturday, which back then was half-again what I made for the whole shift in hourly wages. It would be interesting to know what it’s like now. Millennials are famously averse to tipping.

        • I don’t know why AOC’s personal wealth even matters — you say it’s relevant to the discussion of “stochastic terrorism” but I don’t think it is. But, for what it’s worth:

          I don’t know what her most recent required financial disclosures show, but in late 2020 she reported assets of between $2,003 and $31,000 and student loan debt between $15,000 and $50,000.

        • Bizarre post, Chipmunk. Next time, please consider checking Wikipedia before floating wild hypotheticals about public figures (if you want to be taken seriously).

        • Phil: her financial well being is an issue because she claims to have experienced financial hardship as a result of discrimination and seeks reparations for said discrimination. So the question is: has she experienced financial hardship as a result of being Puerto Rican? The answer seems to be a resounding “no”. Her Wikipedia bio seems to make it clear that she grew up in an environment solidly above median income.

          Josh R: The wikipedia bio proves the point. She does not come from wealth AKA Bezos – that’s not what Carlson intended – but she comes from every bit as much wealth as the “entitled” white people she seeks reparations from. Hilarious!

          If you’re taking her Wikipedia bio seriously – it has some surprising features, but you highlighted it – she has lived the immigrant dream: Grandfather, immigrant; father, educated architect; her: childhood in comfortable burbs, HS science prize, intern w/ Ted Kennedy; BA cum laude economics / international relations. Had she only had the same entitlement, wealth and privilege the white people she seeks reparations from have, she might have been elected to Congress. Tragic.

          The untimely passing of her father seems to have initiated financial hardships but with her resume why did she go off slinging beers? McKinsey was out of slots for cum laude economists spouting Maoist propaganda? No recommendations from Ted’s office? There’s nothing in her bio that indicates she ever tried or had the slightest aspiration to work anywhere other than in politics and bartending – which is fine, but hardly due to her Puerto-Rican her lack of entitlement!

          Andrew: I rate TCs comment as appropriate and on the mark.

          Phil: with regard to “stochastic terrorism”, I guess you’re more likely to be the victim of that if you run around loudly attacking and attempting to extort money from people on false pretenses.

        • Chipmunk:

          I still consider it odd for Carlson of all people to say “rich white” as if it’s a bad thing. You can’t get much richer and whiter than Carlson.

          I agree with you about the general phenomenon of people in the U.S. liking to act as if they came from down-and-out backgrounds so they can pull the populist card. One thing I appreciated about the recent movie The Fabelmans is that Spielberg portrayed his family as being pretty wealthy. Not Tucker Carlson rich, but clearly in the “upper” part of the upper middle class. Spielberg didn’t feel the need to do the usual thing of acting as if he’d had a deprived childhood.

        • Chipmunk, wtf man. Wikipedia says “After college, Ocasio-Cortez moved back to the Bronx and took a job as a bartender and waitress to help her mother—a house cleaner and school bus driver—fight foreclosure of their home”; she evidently (according to official papers) has debts approximately equal to her assets; and you’re saying she’s ‘relatively wealthy’? OK. Keep on keeping on.

  6. This is not a new idea: its the Mafia don talking about how it would be nice if someone just disappeared, or King Henry shouting “will nobody rid me of this meddlesome priest?” while his thugs are in earshot. Or the editorial in a Southern paper talking about how so and so defiled an innocent white maiden but not saying “lynch him!”

    If you need a Communist example, its Pravda’s review column saying that the costuming of a new film shows bourgeois and counter-revolutionary tendencies a few days before the NKVD arrest the costume director.

      • In climate denial land, when Marc Morano posts an attack on some climate scientist and includes their email/USmail address, he has no idea which random follower(s) will send hate mail, letters with white powder, etc. That’s the stochastic part!

      • No, I am talking about the use of speech to encourage someone (but not a specific person) to commit serious violence without saying so directly. This speech is directed at potential actors not victims. King Henry didn’t have a specific knight in mind (and may or may not have meant to kill Thomas Becket), the newspaper editor does not mind who leads the lynch mob as long as his wife brings home photos to put out where his housemaid and cook can see them, the Pravda columnist does not know which NKVD agent is failing to meet quota this month.

        • Sean,

          You’re talking about indirect violence and the use of networks to diffuse responsibility. I agree this is an important issue. It just doesn’t seem “stochastic” to me. The Mafia don or the king doesn’t know who will carry out his indirect wish, but he’s intending for someone to do it. Who does the killing is stochastic, but the killing itself is close to deterministic. In contrast, in the example discussed in the above post, there are thousands of unstable people out there who might be encouraged to commit a violent act, but these people don’t report to the king, as it were, and whether violent acts occur at all, and when they might attack, are more random.

  7. The trouble with accountability for this sort of thing is that the legal standard involves two levels of causation:

    1. But-for causation is the familiar counterfactual causation of David Hume and the Rubin causal model. “Had person A not done that, it would not have happened.” If a terrorist, for example, cites a theory you push in their manifesto and had you in their viewing and posting history, there’s a reasonable case for but for causation.

    2. Proximate causation is a lot more nebulous and is evaded by this kind of rhetoric essentially by design. The idea is, according to wikipedia, to “determine if an action is close enough to a harm in a chain of events.” For example, suppose I kick at a rock on the street, and then the rock slams into a precariously balanced broom, which falls over and scares a dog awake, which runs away straight into a ladder, which falls over and kills a construction worker. Had I not kicked the rock, the worker would be alive, but to hold me responsible for this rube goldberg chain obviously makes no sense; I am not a proximate cause. There are lots of different logical “tests” for proximate cause, and many of them rule out holding people responsible for extremist rhetoric. If the effect is “extraordinary”; even if I cause a relative risk increase of tenfold from 1:1000000 to 1:100000, the event is still extraordinary and I am not a proximate cause. If I am not the “primary cause”; if a man loses his job, loses his wife, gets addicted to opioids and then goes on a killing spree because he heard on the radio that the globalists are destroying the family, the radio is not the “primary cause.”

    • Let us be honest: This kind of anti-whoever-rethoric is usually directed towards the citizens of one’s own country, even if it’s labled ‘foreign policy.’ So, in a way politicians decrying minorities inside the borders of their own country participate in stochastic domestic terrorism. It is a powerful tool to keep your population unified against some supposed threat, and it distracts from domestic problems. I think it is what any autocrat needs to do to keep the people distracted from his personal misdeeds, and it often goes hand in hand with racism, antisemitism, sinophobia, and other forms of discrimination. To me it appears that the line blurs there on whether it’s actually stochastic terrorism: If you don’t really believe in a Jewish world conspiracy, but it fits your agenda quite well, and you commit antisemitic stochastic terrorism, is that a) antisemitism, b) stochastic terrorism, or c) both? I guess this is a question of how we define each of the terms. It’s immoral, that’s what we can all agree on.
      From the discussion above it seems to me that we must answer at least two questions to determine whether stochastic terrorism applies to a specific statement:
      1. Is the target specified? If yes, it contains no stochastic element.
      2. Is violence encouraged? If no, the line is decisively blurred. Not all criticism can count as encouragement for violence. Particularly if the possibly inflammatory statement is true! [Said a wise person, it’s not paranoia if somebody is out to kill you.] And yes, I am aware that it’s quite rare that the statement is doubtlessly true. Just look at the discussion above whether Rep. Ocasio-Cortez is to be considered a rich person…
      Now, it gets tricky if a person uses stochastic terrorism enough to hit a specific target, like flipping a coin often enough until it shows heads. Is applying the law of large numbers stochastic after all? I am not sure this case is encountered particularly often, but it’s definitely technically possible.
      So, while I will not make a judgement over any specific government whether they participate in stochastic terrorism, I believe the two questions above are a good start to check whether stochastic terrorism applies to a statement. A court might also want to take into account whether such statements were made in a small, closed circle or made public, and whether there is a pattern of repetition (what terrorist stops at one terror attack when they can terrorize even more?).

Leave a Reply

Your email address will not be published. Required fields are marked *