Contra Klein on Cancelling
Last week, Dan Klein wrote a critique of my essay, “The Woke Who Did Not Cancel.” I have one big response, plus more specific reactions.
Big response: Even if everything Dan says is correct, there’s still a massive puzzle! Given the extreme left-wing orientation of social media firms, why don’t they ban everything that they find even slightly objectionable? Dan argues that I underestimate how much they ban, if we measure banning correctly. But even if I’m off by a factor of 10, the question remains: Why aren’t they 10 times as bad as Dan Klein thinks? I don’t think he can seriously argue that they can’t be 10x more censorious than they already are. So what gives? In the end, I say that Dan will end up with a moderately toned-down version of my resolution of the puzzle.
Perhaps Dan’s response would be to accuse me of “Type III error” - getting the right answer to the wrong question. We shouldn’t be asking, “Why isn’t cancelling worse?” and giving my answer. We should be asking, “How bad is cancelling?” - and giving the answer “Terrible.” If so, we’re at an impasse. Why aren’t both questions worthwhile?
Moving on to my specific reactions, I’ll stick to Dan’s numbering. He’s in blockquotes, I’m not.
Dan’s point 1:
What he operationalizes as wrongdoing/no wrongdoing is whether someone has been kicked off the platform… He says they engage in only “a token quantity,” “a tiny sliver,” of ideologically-motivated wrongdoing. But Facebook et al screw with content providers in a large number of ways: slapping on labels and warnings and barriers, shadow banning, demonetizing, suspending, defamation, disabling features, miscommunicating, stonewalling, stalling, and so on. There are probably a dozen ways that a platform can screw with a content provider. In the details, probably scores of ways.
First, I’d say that being kicked off a platform entirely is the most severe sanction, so it makes sense to focus on it. But yes, there are other sanctions. The question remains: Why don’t they do 10x or 100x as much of all of these sanctions?
Dan’s point 2:
Bryan never raises the Peltzman effect. A large portion of the video makers I watch on YouTube have said something like: “I don’t want to say anything that will get me banned/demonetized/suspended” — that is, they indicate: “I’m not speaking frankly. I’m having to tone this down.” They all feel chilled.
Same point. Instead of making people filled “chilled,” they could ban them completely. Why don’t they?
Dan’s point 3:
Bryan writes as though he has a good sense of the empirical magnitude of the wrongdoing, but he gives us no assurance that he does. Again, he hasn’t properly operationalized wrongdoing, but he doesn’t in any way convey his knowledge of the magnitude of that, namely people being thrown off the platform (“less than 1%”). If we consider all the ways that Facebook etc. disadvantage non-left voices, my own guess is that 1% would be underestimating it vastly. I rather figure that they disadvantage non-left voices as a matter of course.
I know that non-left views are easy to find all over social media. That’s all I need to know to realize there’s a big puzzle.
Dan’s point 4:
Bryan says nothing about taking out the head of the snake, the General of the army, the lead spear. Or the harsher, more hearty criticism (Robert Malone, Ivor Cummins, Tony Heller, etc., anyone?), or the most biting criticism (Bryan mentions Twitter cancelling The Babylon Bee).
If anything, I see the opposite. Prominent voices are less vulnerable to cancellation than obscure voices.
Bryan says nothing about the significance of killing 0.01 percent of the non-left discourse that just happens to be the news of Hunter Biden’s laptop, for example.
A potent counter-example. Probability that this would have changed any 2020 electoral outcomes? Still very low in my book.
Dan’s point 5:
Although Bryan allows profit-reducing conduct by Facebook etc., he allows it way too grudgingly, writing, "they’re willing to sacrifice a tiny share of their profits to get petty revenge.” I think it is naïve not to think that ideology and a delusional selfhood among the leftist world elites play an enormous role with these multi-millionaires. Someone who understands Adam Smith knows that, beyond the basics, moral condition, as opposed to material condition, is paramount. I make per year perhaps a hundredth or a thousandth of what the chiefs of Facebook et al do, and what would be the marginal utility of doubling my annual income? Not much at all. I see that. Is the situation not similar for them?
On this issue, I say Gary Becker provides far more insight than Adam Smith. Actions speak louder than words. And while people’s words say that “moral condition” is paramount, their actions show otherwise. Furthermore, people who run businesses are self-selected to care about profits over other values. Indeed, if Dan/Smith were correct, we would have to abandon the standard economics of discrimination, which seems rock solid to me. To quote Charlton Heston, “From my cold, dead hands!”
Dan’s point 6:
Bryan concludes with the following judgment, for which he gives no reason: “And when they select their ‘least-favorite people,’ they do so in the same haphazard manner as most of the human race… What drives their cancellations is neither profit nor philosophy, but hysteria and herding.” Was it hysteria and herding that disappeared the Hunter Biden laptop intelligence and canceled the NY Post? Remember, it is elections that determine who gets into office and who doesn’t. Deft moves at critical spots at critical moments—“a tiny sliver”—write the story. I think Mark Zuckerberg knows that.
To repeat, the laptop is a potent counter-example. But it is vastly outnumbered by the endless petty, random cancellations we hear about on a nearly daily basis. I’ve posted hundreds of tweets more intellectually objectionable than the joke that got the Babylon Bee cancelled. At the same time, quite a few folks far less prominent than me have been cancelled. That’s the Unbearable Arbitrariness of Deploring.
Even if Dan’s right, this brings us back to the original issue: Why limit yourself to “deft moves at critical spots at critical moments” instead of wiping out every word of opposition? I don’t see that Dan has an answer.
I think part of this goes all way back to your original post, to the graphic you put up of political donations by party. You say "Wow, almost everyone at Facebook and Twitter donate to Democrats! Isn't odd that they don't cancel more right wingers, then?"
I think, though, that if we consider a few points about that the puzzle diminishes a good bit.
Firstly, not everyone donates to political parties. While I have no doubt that the vast majority of FB and Twitter employees vote Democrat, only a very tiny percentage of those who vote care enough to donate money. Notably, caring a great deal about politics seems to be the key driver of donations, although I don't know how much one has to care to donate 20$. More than I do, at least, but who knows.
Secondly, not everyone who votes or donates to the Democrats is a censorious bastard. There seems to be a good many actual freedom of speech and otherwise fairly sensible Dems out there still, who maybe don't denounce the crazy Woke fringe, but don't really support them either. These folks might get behind a little bit of censorship, or perhaps flagging COVID "disinformation", but they probably are not pushing to ban Joe Rogan.
I think those two points solve a lot of the puzzle here.
Why is there not more cancelling? Because only 10-20% of the workforce at FB and Twitter are super motivated canceller types; the rest push back just a little here and there and won't get behind sweeping bans.
Why is so much of the cancelling rather haphazard? It has to be the type that a large percent of the decision makers (at whatever relevant level) can get behind, and that includes a lot of fairly reasonable Democratic voting folks. They are scared about COVID misinformation, but less trigger happy on abortion debate, perhaps, at least compared to their more ideological firebrand colleagues.
Why don't they ban everyone? Well, see above, but also many of those middle manager types do recognize that alienating 50% of the US user base is probably a bit much.
Why do they trade profits for discrimination? This answer comes straight from the discrimination literature. It isn't the business owners who are super gung ho to ban everyone, as they rightly are trading off loss revenue against ideological or moral points. That calculation does include "If we are seen as more ideologically pure that can net us money... we don't want to look like 4chan." so banning some people through pure discrimination does make sense.
The middle managers, however, are the ones really making specific decisions around low level, not so famous people, and they could not care less about whether company revenue goes up or down the teeny tiny marginal amount banning one more pleb would cause. They are going to get pushback from the higher ups if they ban Trump or Caplan because those names are big enough for higher ups to hear about it, but they are not going to hear a thing if they ban me. All it takes is my name coming across their radar.
Which brings us to the last discriminating actor, the customers themselves. How does my name become a known target to those wielding the ban hammer at FB or Twitter? The censorious and prejudiced customer base that starts flagging my posts as evil misinformation, or hate speech, or whatever. The business is going to respond to the demands of its prejudiced customer base.
So, really, FB and Twitter's behaviors are pretty much what the economics of discrimination literature would predict. Discrimination is most costly to those who own the business, less costly to the workers, and even less to the customers. The very high ups mitigate some of what the middle managers do, but have very little oversight of day to day running of things, and middle managers respond to the demands of the customer when the customers demand things they want to do anyway. Having 10,000 user complaints is a good excuse to ban someone you don't like either.
Note, this also clarifies why certain things do get the heavy ban hammer instead of just the lesser hits: Hunter Biden's laptop, COVID craziness, China/Russia/Ukraine things are all political and government administration issues. Why does Google alter its algorithm for China? Why do Facebook and Twitter censor differently there? Because the higher ups at the company do actually care enough to make those things happen, either because of monetary reasons (want to be in Chinese market, don't want to be attacked by the regulatory state for not censoring what Psaki suggests) or because of ideological (don't want Trump to win so better hide Biden's kid.)
Isn’t it possible that the leaders of Facebook and Twitter actually don’t like censoring and only resort to it in extreme cases? Seems like a better null hypothesis and fits the data pretty well as far as I can see.