Mario Anzuoni / Reuters

Something odd happened this week on Twitter: A hashtag became the most popular topic in the country by accident.

After Senator Elizabeth Warren brushed away the handshake of Senator Bernie Sanders in the moments following a debate, Sanders supporters—or at least alleged Sanders supporters—began to tweet #NeverWarren. Several left-leaning journalists and online Sanders surrogates noticed the rising phrase and sent tweets opposing it. But since their tweets accumulated likes and retweets, they only made the hashtag itself more popular. The algorithm that determines the social network’s most popular topics, after all, could not differentiate between a tweet endorsing #NeverWarren and a tweet rejecting it. When the NBC reporter Ben Collins noticed this phenomenon and tweeted about it, his own tweet picked up more than 7,500 likes and retweets, and lofted the hashtag even higher.

It was faintly ridiculous. It seemed to encapsulate many of the issues with hosting the public sphere on Twitter. And it made me think of Walter Ong, a linguist and Jesuit priest who died at 91 in 2003. Ong spent his life trying to understand the revolutionary technologies, such as the television and radio, unleashed during his lifetime. But he did so by looking far from modern America—and by studying the difference between human cultures rooted in orality and those rooted in literacy. His topic matters for Twitter more than you may think.

Twitter is neither the country’s largest nor its wealthiest social network, but nonetheless it exhibits a curiously tight grip on American culture. In the past decade, it has played a prominent role—if an occasionally overstated one—in the Arab Spring, the Black Lives Matter movement, and the 2016 presidential election. The president of the United States, famously, cannot be torn from Twitter; the chair of the Federal Reserve and several Supreme Court justices reportedly consult it for news, gossip, and debate. If Facebook is the high-school cafeteria where everyone must line up for their slop, Twitter is the classroom where all the the nerdy kids go to eat their lunch—though perhaps this metaphor is too generous: Teenagers, for all their flaws, are kinder and less calculating than many Twitter-adept adults.

Twitter is especially beloved by the press, and the unfortunate affinity that journalists and policy makers have for the social network means that—as with politics itself—you may not care about Twitter, but it cares about you, especially if you’ve just done something embarrassing on national television. Reformed Twitter users who’ve quit the service talk about how tweets are inescapable. They are embedded in news stories, screencapped for Instagram, and quoted on TV shows and podcasts.

The sum effect is that Twitter is both leaderless and influential, little used and widely reviled. And at its best, even now, it can still be wonderful: There is a joy and exhilaration in watching a democratic culture at work, in getting to see dozens of people grow and think in real time. At its worst, its rabble seems to adopt that old knock against the British newspapers—that they wield power without responsibilityas a proud boast and way of life.

The spark for the Warren-Sanders spat came on Monday night. CNN broke the story that the senator from Vermont, during a private meeting in 2018, had told Warren that he did not believe a woman could beat Donald Trump. Immediately, my timeline filled with Democrats expressing all manner of views about the incident. Many were angry or dismayed or confused. Virtually all of them did not want the story to eclipse what they saw as Joe Biden’s shortcomings or the ongoing impeachment proceedings against President Trump. Yet they kept making jokes about the CNN story, or complaining about it, or airing hypotheses about it—and so the story seemed to be all that anyone was talking about all night. Eventually, almost without being able to stop themselves, Warren’s and Sanders’s online surrogates were correcting one another’s attempts at tact, and then arguing, and then making a scene.

And maybe the story should have stirred up that froth of conflict and attention. But none of the people who drove its prominence actually seemed to think that it deserved that attention. They all said, at least, that they wanted the media to talk about other things. But the Warren-Sanders spat is what they did talk about.

This contradiction hints at a basic tension of the platform. From my perspective, as a single Twitter user, the online horde is always going on about something. So if I tweet about that something, it’s not a big deal. I’m only thinking aloud, goofing off, and harmlessly chatting with my friends and readers. But if 50 other people tweet about the same thing—especially if it’s frivolous and especially if they all have, like I do, an account with more than 1,000 or so followers—then they’re making that topic even more popular, amplifying it, and reinforcing the media’s toxic fixation on meaningless chaff. My tweets—well, not my tweets, but you get it—are conversational and informal, and they matter relatively little. But taken collectively, everyone else’s tweets are informational and declamatory. They carry weight.

This instability—between the individual and the mass, the high and the low—is also what makes Twitter fun. The site’s worst users are those who monotonously, humorlessly post about the same thing over and over again. But in trying to avoid that fate—and in trying, generally, to act like a regular person online—users push the conversation toward conflict and superficiality.

Ong would have understood. Writing down a language, he realized, is not just a mere shuffling of papers; it forever changes how the language works. Consider the differences between speech and text. For oral cultures, words are primarily vibrations in the air, Ong argued. Words must therefore be memorable, few in number, and tied to the concrete reality of day-to-day life. But after the advent of writing, words become more than invisible sounds. They become permanent symbols that exist outside their utterance and can be read long after the speaker has died. Words can also divorce from the physical world and start to reference ideas, concepts, and abstract states. And instead of words needing to aid memory, as they do in oral cultures (by using a repeated epithet, such as Homer’s “wine-dark sea”), written words can suddenly act as a form of memory themselves.

Before Ong died in 2003, he was asked about a special kind of writing that people do online, a genre of communication familiar to any Slack or AIM user or group-chat texter. It’s a mode that delivers words live and at the speed of speech—in which, as Ong put it, “textualized verbal exchange registers psychologically as having the temporal immediacy of oral exchange.” (This is apparently how a Jesuit talks about sliding into one’s DMs.) Ong called this new fusion “secondary literacy,” but today we just call it texting. Whatever its name, it reigned during Twitter’s early days. As I once wrote: “Twitter lets users read the same words at different times, which is a key aspect of literacy. Tweets are chatty, fusing word and action like orality; and also declarative, severable, preservable, and analyzable like literacy.”

But by 2014, the Canadian academic Bonnie Stewart had noticed a change in how Twitter worked as a social space. Tweets that were written as chatty musings for one group of users were interpreted as print-like declamations by another. “The rot we’re seeing in Twitter is the rot of participatory media devolved into competitive spheres,” she said, “where the collective ‘we’ treats conversational contributions as fixed print-like identity claims.

I was so taken with this idea that I later wrote about it. Six years later, Stewart’s observation still resonates with insight, even if it no longer feels like news: Twitter has been a mess of speech-like tweets interpreted as print and print-like tweets interpreted as speech for as long as most users can remember. This whiplash between orality and literacy is even part of what makes it fun.

But following this week, I’ve wondered if that instability presents a political problem—particularly for the left in the United States. The word that sticks out to me now from Stewart’s post is identity. The bedrock of politics is that it forges what the left wing calls solidarity: a sense of shared identity or common interest that transcends whatever other differences among people exist. In America, this challenge is more difficult for the center-left Democrats, who are the more heterogenous of the parties. What Jesse Jackson said at the party’s 1988 national convention is still true today: None of the Democratic coalitions—be it farmers, trade unionists, feminists, people of color, or LGBTQ folks—commands a big enough “patch” of America to win power for itself. He continued that Democrats must therefore “build such a quilt” that binds the patches. They must, in other words, forge solidarity. And not much shreds solidarity faster than misstating, or misinterpreting, a co-partisan’s statement about their identity.

Even for those elsewhere on the political spectrum, the whole situation suggests two ideas, neither of which is particularly encouraging. First: Twitter is now so global and crowded and multifaceted that it no longer has a unified “we” at all. Twitter, en masse, isn’t anything like a thinking public; it’s just a bunch of people. Second, it suggests an axiom so puritan that I hesitate to express it: On Twitter, ideas are so commodified that to say something is simultaneously to amplify it. You’re never “just saying” on Twitter. You’re always doing.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.