The World That Twitter Never Made

A few months ago Jonathan Haidt made waves with a big think-piece in the Atlantic arguing that most of the ills of the 2010s can be traced back to the invention of the retweet button.1 I read the essay and disagreed with it vociferously. Today City Journal published my critique.2 You can read my counter essay here. Below I would like to add some additional thoughts on social media and American politics that could not fit into that piece.

I offered three main lines of critique: first, social media (and Twitter in particular) cannot account for a societal wide decline in trust or a societal wide rise in radicalism and partisanship. The number of active twitter users is just too small. As Saturday Night Live understands, when Americans tells pollsters they are afraid to state their beliefs, it is in workplaces and over the dinner tables that these fears matter most, not on social media.

Second: the better explanation for declining institutional trust is declining institutional performance. Of course millennials and zoomers are radicalized; the formative political events of their youth were a string of national disasters. These disasters were presided over by the very forces of centrism that Haidt would like to protect. As something of a center-righter myself I applaud his stand against the crazies, but I am frustrated by the manner in which he chooses to make it.

By focusing his ire on social media, Haidt elevates the procedural over the substantive. There is little evidence that changing procedural form this or procedural form that will make any difference on the long term. At best it will allow the centrist boomers to hold onto power a few years longer. But if the radicals cannot be openly confronted on the basis of their ideas, and if centrists cannot develop an emotional appeal and political vision as powerful as their opponents, then the center will not hold.

Third, what we call “cancel culture” is not a new feature of American life. It has always been with us. If current times feel especially fraught and censorious, it is not because Americans have suddenly become less tolerant of speech generally, but because there has been a titanic and sudden change in what lies outside the bounds of tolerance. This change is disorienting and frightening, but similar instances of active thought policing can be found across our history, including our very recent past (Dixie Chicks, exit stage left).

This another example of emphasizing procedure over substance. Cancellation and self-censorship have always been with us. Trying to outroot it is impossible. The real issue is the range of things one can be cancelled for. But instead of arguing, straightforwardly, that issue x or issue y should not be on the list of things cancel-worthy, we tilt at windmills, trying to throw out the notion of acceptable opinion altogether. A better response to radicalism is needed.

All of these ideas are laid out with more depth in the main essay. But I want to raise two points that did not fit naturally in that essay here.

First, some readers might be surprised to see me write against a Twitter driven account of national affairs. After all, was not I the guy who wrote “The World Twitter Made?”

Guilty as charged. I do think Twitter has had a baleful influence on public debate, but not quite the one that Haidt sees. Haidt’s narrative of Twitter politics reads much like Martin Gurri’s Revolt of the Public: once upon a time we had responsible centrists keeping the gates. Twitter and Facebook have destroyed the gates and their keepers. Anarchy empowers the nasty and the radical, who hitherto had no meaningful political voice. They use their megaphone to bludgeon the responsible and respectable into silence. Consequently, radical ideologies rule the roost.

There is some truth to this, but among other things, it underestimates just how mainstream radical ideology is among America’s younger professionals. It wasn’t Twitter anons that drove the Times to Wokeness, but the Times‘ own employees. The radicals were raised inside the gates.

The role Twitter played in the radicalization of the Very Online is subtler than Haidt’s model allows. I recall a thoughtful essay by Justin Smith on the futility of debate in the world of Twitter. Writes he:

Is there any way to intervene usefully or meaningfully in public debate, in what the extremely online Twitter users are with gleeful irony calling the “discourse” of the present moment?

It has come to seem to me recently that this present moment must be to language something like what the Industrial Revolution was to textiles. A writer who works on the old system of production can spend days crafting a sentence, putting what feels like a worthy idea into language, only to find, once finished, that the internet has already produced countless sentences that are more or less just like it, even if these lack the same artisanal origin story that we imagine gives writing its soul. There is, it seems to me, no more place for writers and thinkers in our future than, since the nineteenth century, there has been for weavers.

This predicament is not confined to politics, and in fact engulfs all domains of human social existence. But it perhaps crystallizes most refractively in the case of politics, so we may as well start there.

There are memes circulating that are known as “bingo cards,” in which each square is filled with a typical statement or trait of a person who belongs to a given constituency, a mouth-breathing mom’s-basement-dwelling Reddit-using Men’s Rights Activist, for example, or, say, an unctuous white male ally of POC feminism. The idea is that within this grid there is an exhaustive and as it were a priori tabulation, deduced like Kant’s categories of the understanding, of all the possible moves a member of one of these groups might make, and whenever the poor sap tries to state his considered view, his opponent need only pull out the table and point to the corresponding box, thus revealing to him that it is not actually a considered view at all, but only an algorithmically predictable bit of output from the particular program he is running. The sap is sapped of his subjectivity, of his belief that he, properly speaking, has views at all…

Who has not found themselves thrust into the uncomfortable position just described, of being told that what we thought were our considered beliefs are in fact something else entirely? I know I have been on many occasions: to be honest, this happens more or less every time I open my newsfeed and look at what my peers are discoursing about. For example, I admire Adolph Reed, Jr., a great deal; I believe he is largely correct about the political and economic function of “diversity” as an institutional desideratum in American society in recent decades; and I believe, or used to believe, that I had come to view Reed’s work in this way as a result of having read it and reflected on it, and of having found it good and sound.

But then, not so long ago, I happened to come across this from an American academic I know through social media: “For a certain kind of white male leftist,” my acquaintance wrote, “Reed makes a very convenient ally.” What would be a fitting response to such exposure as this? Should I stop agreeing with Reed? Easier said than done. It is never easy to change one’s beliefs by an exercise of will alone. But it would be will alone, and not intellect, that would do the work of belief change in this case: the will, namely, to trade in the algorithm that I’m running for one that, I’ve recently learned when checking in on the discourse, is preferred among my peers.…

Someone who thinks about their place in the world in terms of the structural violence inflicted on them as they move through it is thinking of themselves, among other things, in structural terms, which is to say, again among other things, not as subjects. This gutting of our human subjecthood is currently being stoked and exacerbated, and integrated into a causal loop with, the financial incentives of the tech companies. People are now speaking in a way that results directly from the recent moneyballing of all of human existence.3

Smith proposes that modern technology has stripped us of “subjecthood.” What Smith calls “subjecthood” I normally call agency. We want to feel like we are in charge of our own lives. We want to feel like our political opinions and prognoses were the product of careful thought and reflection. Our thought and reflection. Liberalism, pluralism, and norms of civil debate are built on the assumption that people are rational, thinking individuals that can be convinced by argument. The dueling essays and op-eds of yesteryear, with individual names attached to each lengthy, carefully edited piece, reinforced those assumptions.

Twitter tears these assumptions to shreds.

To log onto Twitter is to be exposed to a sea of a million voices. It is to observe political debates in such numbers that one can soon predict every rhetorical move every participant will make well before they make them. It is to have your likes, dislikes, opinions, and passions decided by algorithm. It means watching people climb in clout by manipulating said algorithms for personal gain. The politico immersed in this ocean of content no longer feels like an individual with control over her beliefs. Her intellectual development is no longer her story. Nor is anyone else’s story truly their story. We are all the playthings of algorithms, hosts for a “Discourse” that seems to live outside of those discoursing. Nothing in Twitter’s endless streams suggests that human beings are amenable to reason or persuasion. Tweets are not written to persuade, after all, but to game. And when the discourse is a game to win, lose, or manipulate, there is little reason to care much for free speech.

Lili Loofbourow described her turn against “free speech” in exactly these terms in a viral twitter thread penned a few years ago:

Online pundits should know (and factor in) that social media as a “public square” where “good faith debate” happens is a thing of the past. Disagreement here happens through trolling, sea-lioning, ratios, dunks…

Most arguments worth having have been had and witnessed 1000x already on these platforms, in several permutations. We know their tired choreographies, the moves and countermoves. At this point we mostly enjoy the style of whichever dunk we happen to agree with.

Does that lead to paranoid readings and meta-debates that seem totally batshit to onlookers who aren’t internet-poisoned? Yup! “All Lives Matter” sounds perfectly reasonable–as a text–unless you know the history of that discourse. (And you’ll sound pretty weird explaining it.)

“Why would you refuse to debate someone who’s simply saying that All Lives Matter?” is the kind of question an Enlightenment subject longing for a robust exchange of ideas might ask. Well, the reason is that most of us know, through bitter experience, that it’s a waste of time.

It wouldn’t be a true exchange. We know by now what “All Lives Matter” signals and that what it signals is orthogonal to what it says. Your fluency in this garbage means you take shortcuts: you don’t have to refute the text to leap to the subtext, which is the real issue….

Sure, good-faith debate would be nice. Instead, the internet pressure-cooked rhetoric. Again: people can watch the same argument be conducted a million times in slightly different ways, and that’s interesting, and a blessing, and a curse.

It produced a kind of argumentative hyperliteracy. If you can predict every step of a controversy (including the backlash), it makes perfect sense to meta-argue instead–over what X *really* means, or implies, or what, down a road we know well, it confirms.4

Twitter presents the problem of over exposure. Debates on the blogosphere, or in the op-ed sections, were argued within a defined limit. Your opponent was a person you could identify, even if that identity was pseudonymous. The onslaught of voices in your mentions strips away the individuality of debate. You are not battling people, but a force of nature. You are repeatedly exposed to non-entities of the other side—or, barring that, to the stupidest and cringiest entities of the other side. Your interlocutors are incentivized not to tailor their argument to you or those like you, but rather to their fans and boosters. It is a poisonous environment for good faith engagement. Spending too much time there risks losing faith in good faith altogether.

The retweet button facilitates all of these problems. But it is only one piece of the puzzle. Just as important is the scope of the Twitterverse, which lacks boundaries, borders, or moderators that made the blogosphere and forums of old feel and operate like communities.

But Twitter can only poison so far. Twitter is experienced as a war against some natural force only by politicos with high follower counts. That group includes Haidt, Loofburrow, and myself, but not most Twitter users—much less most Americans. In order to trace “Why the last ten years have been uniquely stupid” Haidt must expand his vision past social media. Why aren’t the centrists winning out there?

There is something eerie when the centrist liberals of our day are set next to the older generations in Turgenev’s Fathers and Sons or Dostoevsky’s The Demons. The old liberals in those novels read both comic and tragic, absorbed by the rebellions of their youth despite having long passed middle age, horrified by the revolutionary notions of their children yet paralyzed by indecision whenever directly confronted with them, so committed to idealized notions of progress and dissent that they lack the language, much less the courage, to critique their children. So it is today.

The boomer centrists do not know how to respond to accusations of racism or sexism without sounding like the echo of their own parents. So they don’t respond to these accusations. Instead they grasp for theories that allow escape from the actual issues at play. That is, they flock to theories of procedure, not substance. Haidt’s theory of the retweet button fits that bill; so do many others, such as Yuval Levin’s writing on institutions. Both are correct within limits: yes, Twitter is bad for the discourse. Yes, we need more institution building. But these theories are popular with centrists and non-politicos less for their explanatory power than for the ideological dodge they allow centrists to make.