This brief reflection is based on remarks at a reunion conference for the 25^th anniversary of the Center on Organizational Innovation at Columbia University. Alumni of the COI were asked to offer “impulses” to spark conversation around various themes over the course of the day, and I began mine with an odd little joke: an imaginary New Yorker-style editorial cartoon, which I described more or less as follows. (Trying to coax the cartoon to life via ChatGPT produced the image above, full of AI’s characteristic visual tics, like the misspelled sign).

The year is, say, 1498, in Antwerp or Leipzig or some other bustling commercial center in late medieval Europe. Barely 50 years have passed since Gutenberg introduced movable type, and already the printing press is in more than 300 cities and some 20 million books have been produced. A clutch of scholars is emerging from one of the print shops that Elizabeth Eisenstein tells us were engines of the “print culture” credited with sweeping social and political transformation across the continent.¹ The academics are in a heated debate over the implications of this new medium, until one holds up a finger and exclaims triumphantly, “Yes, but a recent meta-analysis of 63 experimental studies clearly shows that exposure to the printed word has no significant impact on religious attitudes!”

This cartoon anachronism came into my mind several years ago while preparing a literature review about the phenomenon of online mis- and disinformation.² It was inspired by the image Bruno Latour offers, in Science in Action, of a child quoting “three NIH studies” to refute his mother’s gentle suggestion that “an apple a day keeps the doctor away”. Latour uses the absurd breach of context to illustrate the different “regimes of circulation” of soft, everyday facts versus harder scientific ones. Here, the absurdity is meant to highlight — for a small group of communications researchers who argue about the effects of media technologies — the field’s surprisingly narrow approach to thinking about misinformation on social networks.

To wit: Public commentary tends to take for granted that conspiracy theories and phony news stories proliferating across social media matter a great deal. The question has sharply divided communication scholars, however. In that debate, the dominant view, reflecting dozens of scientific studies conducted over the last decade, is that most people encounter such misinformation much less often than we imagine, and that when they do it has very limited effects on their beliefs or behavior. “Public intellectuals and journalists frequently make sweeping claims about the effects of exposure to false content online that are inconsistent with much of the current empirical evidence”, charges a recent overview in Nature.³ A related argument holds that media and political elites were swept up in a moral panic about “fake news” on Facebook and Twitter in the wake of events like the Brexit vote and Donald Trump’s first election victory.

The studies anchoring that consensus — mostly online experiments, tracking studies, and network analyses — are grounded in a behaviorist/psychological paradigm that conceives of “media effects” mainly in terms of individual exposure and response to specific messages, in line with the long-dominant tradition in communications research.⁴ Obviously, this affords only a very limited picture of how media might matter in public life. Imagine using a psychological effects model to understand, say, the influence and impact of Martin Luther’s Ninety-fine Theses — sometimes called the first bestseller because reprinted editions spread throughout Germany in two weeks, and across Europe in as many months. Accordingly, a small scholarly countercurrent has argued that, moral panic notwithstanding, most research into the impacts of online misinformation misses the forest for the trees.⁵ It’s hard not to wonder whether this narrowness of vision results, ironically, from another sort of media effect: What scholars have called the “datafication” of social life yields up an abundance of data about online behavior, data that exerts an irresistible pull on the social sciences.⁶

The odd juxtaposition also hopefully evokes the very different register of analysis we become comfortable with when looking at past media “revolutions”. Canonical work from Eisenstein, Benedict Anderson, and others credits the printing press with paving the way for key modern institutions — the nation state, the enlightenment, the scientific revolution.⁷ Of course, these claims have sparked long-running debates over questions of causality and contingency. For instance, Eisenstein’s account of the print revolution has been criticized for leaving out the painstaking, institution-building work needed to realize the characteristics she treats as inherent features of print.⁸ As with similar debates over the influence of the telegraph and radio, though, the alleged print determinists and their constructivist critics share an ample, culture-centered view of what a new medium’s “effects” might encompass.

What will seem obvious about the impacts of online disinformation a few generations from now? What vocabularies will we use in hindsight, when the Internet and smartphones and social media and perhaps even AI resolve into a single shift? Scholarship that announces new eras in media res usually does not age well; witness much of the most influential and generative early work on the digital revolution.⁹ Still, it seems undeniable that future scholars trying to understand how online misinformation matters in political life today won’t be arguing over the results of randomized trials. Instead, the most illuminating and durable explanations how our media shape our politics, today and in the past, often center on culture. That can mean political or civic culture broadly speaking, but also the microcultures of online communities that actively create and spread disinformation — and especially of the media and political elite who turn it into a strategic meaning-making resource, refining and recontextualizing these tropes for specific campaigns.¹⁰

Consider, as a closing example, perhaps the lowest moment of the 2024 US presidential campaign: when Donald Trump repeated a floridly racist rumor about Haitian immigrants living in Springfield, Ohio, during a televised debate with an audience of 67 million. Baseless claims of pets being stolen and eaten had circulated online for several days, amplified by his running mate, before Trump repeated the lie onstage — a familiar pattern to the professional fact-checkers who track how false narratives flow between fringe networks and national political figures. Some had already debunked the ugly rumor, paving the way for debate moderators to challenge it live.

One set of questions relates to the disinforming effects of such an ugly rumor: Who saw it, whether they believed it, what influence corrections had, and of course whether it helped or hurt on election day. (“It is easy to suppose that Trump will lose votes for indulging this pet-eating chimera. But his dark rhetoric masks a calculated bet,” a Financial Times columnist opined.) The more interesting question is how this style of politics has taken shape in the crucible of a particular media landscape — one with profoundly different affordances and incentives for capturing attention than the broadcast media world that helped to shape political reflexes and culture through much of the 20^th century. It seems reasonable to approach that as a kind of media effect.