When polls opened in the Phoenix area Tuesday morning, some vote-tabulation machines weren’t working; within half an hour, conspiracy theories about the problem were running rampant on the internet. Prominent right-wing influencers with millions of followers were clamoring for arrests, and outraged citizens using Telegram channels and online message boards debated whether prison or execution was the appropriate punishment for what was clearly another steal. In reality, officials in Republican-run Maricopa County had transparently informed residents about the problem and offered a few options to make sure everyone’s ballots would be counted: Voters could wait until the problem was fixed, go to a different polling place, or put their completed paper ballot into what was called Box 3 for processing later. Although some social-media users appreciated this information, others sincerely believed they’d uncovered a nefarious plot. The voting machines had been deliberately broken, they insisted, and word on Twitter was that they were broken only in Republican parts of town. Whatever went into Box 3, many were convinced, was not going to get counted.
As part of the Election Integrity Partnership, my team at the Stanford Internet Observatory studies online rumors, and how they spread across the internet in real time. As we watched speculation about intentional tampering with Maricopa voting equipment multiply on social media, racking up tens of thousands of user engagements, I was reminded of a similar 2020 incident—one that also falls into the category that my colleagues and I sometimes call “Tricky Voting Machines.” In what came to be known as Sharpiegate, supporters of former President Donald Trump claimed that, because they’d been given felt-tip markers to fill out their ballots, tabulation machines weren’t recording their votes. At the time, half of the American voting public had been told for months by media commentators and political leaders whom they trusted—indeed, by the president himself—that the election would be stolen from them. Compounding that mindset today are two years of claims that the 2020 election was in fact fraudulent. The details of how remain shrouded in mystery, unknowable even after a variety of investigations and audits. But that didn’t matter in Maricopa County this week. Especially after two years of conspiracy-mongering by Trump and his most influential supporters, the mere mention of Tricky Voting Machines makes the rest of the plot immediately clear.
[Read: I was a teenage conspiracy theorist]
After watching such rumors spread this week—and hundreds of other viral rumors spread during the 2020 election—I’m struck by how familiar, how predictable, the rumor story lines have become. Before the 2020 election, a rash of what we call “Discarded Ballots” rumors swept across social media. In mid-September, a reader of the conservative website The Gateway Pundit claimed to have photographed some ballot envelopes in a Dumpster in Sonoma, California. Also that month, news outlets reported on mail trays in a ditch in Greenville, Wisconsin. Around the same time, nine discarded ballots were found in a garbage can in Scranton, Pennsylvania. These reports produced no evidence of fraud—the Sonoma ballot envelopes were from 2018 and were empty; investigations of the other two incidents yielded no charges—but Discarded Ballots is a powerful trope. In political rumors, the same plot elements and character archetypes recur with minor variation from incident to incident—to the point that conspiracy buffs should realize they’re watching the same show again and again and again.
A few years ago, I happened across an exceptional website called TV Tropes, whose users meticulously document every narrative device that has appeared on a screen large or small. I’d first stumbled across it while trying to find out how the final season of the HBO show Big Love had ended. On the show’s TV Tropes page, contributors good-naturedly skewered the writers for including a “Creepy Child” and for showing the “Expository Hairstyle Change” of a character going from brunette to blond following her mother’s death. The site is funny because all shows draw on the same devices. I went down a rabbit hole of Expository Hairstyle Changes, reading about various manifestations of the form (a male character growing a “Beard of Sorrow” means they’ve been through it) and about other notable movies, including The Matrix and Frozen, that use forms of expository hair.
Reading the site is an illuminating experience that lays bare the building blocks that we all intuitively recognize but can’t quite articulate. After an hour or so of reading, “you can hold a conversation about anime without having ever seen anime,” the site’s founder, who goes by Fast Eddie, told me. (He gave me his real name but uses a pseudonym to protect his privacy offline.)
A trope, Fast Eddie’s website explains, is a convention—a plot device, a character archetype, a linguistic idiom—that an audience will “recognize and understand instantly.” These storytelling devices are a kind of shortcut. They create a sense of immediate familiarity. If a movie begins with a large fin slicing through water, with ominous background music, you know to expect what TV Tropes calls an “Attack of the 50-Foot Whatever.” When you see a lab technician furtively “Playing With Syringes,” you know that dinosaurs or a virus will escape imminently. These building blocks help viewers orient themselves, and efficiently foreshadow what will likely happen next.
Fast Eddie launched TV Tropes in 2004 as a side project. He and some friends had been discussing Buffy the Vampire Slayer on message boards, and decided to create a “pattern language” to describe the show’s plot formulas, which they and ultimately other users would then apply to a variety of programs. “You can be analytical but still fun, and we prioritized multiple viewpoints,” Fast Eddie said. The site was Wikipedia for deconstructing entertainment: written by fans and pointedly egalitarian. But unlike on the user-generated encyclopedia, no subject is too obscure or too niche to be cataloged on TV Tropes.
What renewed my interest in TV Tropes recently wasn’t its excellent assessment of House of the Dragon, but the site’s potential to illuminate how and why certain stories spread outside the entertainment world. Narratives and stories aren’t just present in movies and TV shows; they shape our perceptions of what transpires around us, align us within communities, and even serve as the foundation of movements. When a novel disease such as COVID-19 emerges, it summons familiar plotlines about outbreaks, environmental disasters, and even bioweapons. (Fans of sci-fi thrillers might go as far as blaming a “Mad Scientist” for a “Synthetic Plague.”) Our familiarity with these elements makes the overall story seem plausible, even—or perhaps especially—when facts and evidence are in short supply. We don’t just make movies because they resemble real life; movies influence our expectations of real life, including politics. Dead voters casting ballots? TV Tropes has an entry for that one, too: “Vote Early, Vote Often.”
Partisan influencers are the showrunners of online political theater. They know which plot devices please their crowd most, what blend of familiarity and novelty will sustain the most interest, and what rhetoric will persuade the most followers to amplify the story to other people. An influencer who tweets a photo of a black suitcase outside a polling place on Election Day can frame it as “found ballots” (being delivered to secretly help a particular candidate) or “stolen ballots” (being spirited away to disenfranchise another candidate’s supporters), depending on what’s more useful. They can speculate aloud, just asking questions: Is anyone looking into what’s in that suitcase? Adding an element of implied blame or conspiracy—The other side must be behind it—ramps up the potential for moral or partisan outrage, and creates a villain to rail against on social media.
[Renée DiResta: It’s not misinformation. It’s amplified propaganda].
The media-literacy scholar Michael Caulfield, who studies election rumors at the University of Washington Center for an Informed Public and also participates in the Election Integrity Partnership, describes this process as setting a scene. “Propagandists have an agenda, but they don’t get to choose the material they have to work with,” he told me. “Something is discovered. A piece of data is found. Someone dies. The trope is a recipe for connecting the novel element to the overarching narrative in a formulaic way. And it’s immediately familiar to the audience. ”
But might the public’s familiarity with political conspiracy plots be put to another use? If more Americans were like TV Tropes’ users—that is, if they could spot the recurring motifs in purported political plots—might they also be better at separating fact from fiction?
When political conspiracy theories go viral, media outlets dutifully debunk them—but in many cases far too late. Fact-checking individual claims is slow and reactive, and the time lag to find out the facts means that the audience has already moved on to the next rumor. Additionally, among some hyper-partisan communities, a fact-check will be instantly dismissed as untrustworthy because of who wrote it. But could schools, media organizations, and civic-minded YouTube and TikTok creators instead educate people about conspiracy tropes themselves—give them that “in the know” feeling of recognition that comes from reading a few pages of TV Tropes? The site is engrossing because it’s funny. It’s not partisan. It doesn’t moralize, use jargon, or make its audience feel like rubes. Using a similarly engaging tone, might media-literacy teams persuade the public not to accept novel conspiracy claims at face value but instead to spot the familiar tropes that they embody?
Some preliminary efforts along these lines have already materialized: Working with researchers at Cambridge and the University of Bristol, Google Jigsaw made short, entertaining videos to “prebunk”—help audiences recognize in advance—the signs of misleading rhetoric, false dichotomies, and ad hominem attacks. This kind of preemptive inoculation might be used to help people recognize tropes—the same way a quick read through TV Tropes will change the way you process the next horror film you see.
Of course, this may not work, but Caulfield and other information experts argue that trope-based “prebunking” is worth a try. The point is not to discount all storylines just because we’ve seen them before, or because they seem straight out of a political thriller. In a country with more than 170,000 election precincts, one or two criminals might indeed try to, say, vote using the name of a deceased family member. Sometimes, rumors do turn out to be true. That’s one of the reasons that sharing them feels so compelling. But healthy skepticism is warranted. And if we as media consumers can begin to recognize the tropes—Oh, that’s Tricky Voting Machines again, and it seldom amounts to anything—perhaps we’ll be less prone to amplifying viral rumors and can reach a place of more measured consideration of underlying facts.
