One day, when historians tell the strange story of the 21st century, this age of software and smartphones, populism and Pokemon, they will focus on a fundamental shift in the way people learn about the world. Within the span of a generation, we went from watching the same news shows on television, and reading the same newspapers in print, to getting a personalized feed of everything that our social network finds interesting, as filtered by a clever algorithm. The main goal of the algorithm is to keep us staring at the screen, increasing the slight odds that we might click on an advertisement.

I’m talking, of course, about Facebook. Given the huge amount of attention Facebook commands—roughly 22 percent of the internet time Americans spend on their mobile devices is spent on the social network—it has generated a relatively meager amount of empirical research. (It didn't help that the company’s last major experiment became a silly controversy.) Furthermore, most of the research that does exist explores the network’s impact on our social lives. In general, these studies find small, mostly positive correlations between Facebook use and a range of social measures: our Facebook friends are not the death of real friendship.

What this research largely overlooks, however, is a far more basic question: why is Facebook so popular? What is it about the social network (and social media in general) that makes it so attractive to human attention? It’s a mystery at the heart of the digital economy, in which fortunes hinge on the allocation of eyeballs.

One of the best answers for the appeal of Facebook comes from a 2013 paper by a team of researchers at UCSD. (First author Laura Mickes, senior authors Christine Harris and Nicholas Christenfeld.) Their paper begins with a paradox: the content of Facebook is often mundane, full of what the scientists refer to as “trivial ephemera.” Here’s a random sampling of my current feed: there’s an endorsement of a new gluten-free pasta, a smattering of child photos, emotional thoughts on politics and a post about a broken slide at the local park. As the scientists point out, these Facebook “microblogs” are full of quickly composed comments and photos, an impulsive record of everyday life.

Such content might not sound very appealing, especially when there is so much highly polished material already competing for our attention. (Why read our crazy uncle on the election when there’s the Times?) And yet, the “microblog” format has proven irresistible: Facebook’s “news” feed is the dominant information platform of our century, with nearly half of Americans using it as a source for news. This popularity, write the scientists, “suggests that something about such ‘microblogging’ resonates with human nature.”

To make sense of this resonance, the scientists conducted some simple memory experiments. In their first study, they compared the mnemonic power of Facebook posts to sentences from published books. (The Facebook posts were taken from the feeds of five research assistants, while the book sentences were randomly selected from new titles.) The subjects were shown 100 of these stimuli for three seconds each. Then, they were given a recognition test consisting of these stimuli along with another 100 “lures” – similar content they had not seen - and asked to assess their confidence, on a twenty-point scale, as to whether they previously been exposed to a given stimulus.

According to the data, the Facebook posts were much more memorable than the published sentences. (This effect held even after controlling for sentence length and the use of “irregular typography,” such as emoticons.) But this wasn’t because people couldn’t remember the sentences extracted from books – their performance here was on par with other studies of textual memory. Rather, it was largely due to the “remarkable memorability” of the Facebook posts. Their content was trivial. It was also unforgettable.

In a follow-up condition, the scientists replaced the book sentences with photographs of human faces. (They also gathered a new collection of Facebook posts, to make sure their first set wasn’t an anomaly.) Although it’s long been argued that the human brain is “specially designed to process and store facial information,” the scientists found that the Facebook posts were still far easier to remember.

This is not a minor effect: the difference in memory performance between Facebook posts and these other stimuli is roughly equivalent to the difference between people with amnesia due to brain damage and those with a normal memory. What’s more, this effect exists even when the Facebook content is about people we don’t even know. Just imagine how memorable it is when the feed is drawn from our actual friends.

To better understand the mnemonic advantage of microblogs, the scientists ran several additional experiments. In one study, they culled text from CNN.com, drawing from both the news and entertainment sections. The text came in three forms: headlines, sentences from the articles, and reader comments. As you can probably guess, the reader comments were much more likely to be remembered, especially when compared to sentences from the articles. Subjects were also better at remembering content from the entertainment section, at least compared to news content.

Based on this data, the scientists argue that the extreme memorability of Facebook posts is being driven by at least two factors. The first is that people are drawn to “unfiltered, largely unconsidered postings,” whether it’s a Facebook microblog or a blog comment. When it comes to text, we don’t want polish and reflection. We want gut and fervor. We want Trump’s tweets.

The second factor is the personal filter of Facebook, which seems to take advantage of our social nature. We remember random updates from our news feed for the same reason we remember all the names of the Pitt-Jolie children: we are gossipy creatures, perpetually interested in the lives of others.

This research helps explain the value of Facebook, which is currently the 7th most valuable company in the world. The success of the company, which sells ads against our attention, is ultimately dependent on our willingness to read the haphazard content produced by other people for free. This might seem like a bug, but it’s actually an essential feature of the social network. “These especially memorable Facebook posts,” write the scientists, “may be far closer than professionally crafted sentences to tapping into the basic language capacities of our minds. Perhaps the very sentences that are so effortlessly generated are, for that reason, the same ones that are readily remembered.” While traditional media companies assume people want clean and professional prose, it turns out that we’re compelled to remember the casual and flippant. The problem, of course, is that the Facebook news algorithm is filtered to maximize attention, not truth, which can lead to the spread of sticky lies. When our private feed is full of memorable falsehoods what happens to public discourse?

And it’s not just Facebook: the rise of the smartphone has encouraged a parallel rise in informal messaging. (We've gone from email to emojis in a few short years.) Consider Snapchat, the social network du jour. It's entire business model depends on the eagerness of users to consume raw visual content, produced by friends in the grip of System 1. In a universe overflowing with professional video content, it might seem perverse that we spend so much time watching grainy videos of random events. But this is what we care about. This is what we remember.

The creation of content used to be a professional activity. It used to require moveable type and a printing press and a film crew. But digital technology democratized the tools. And once that happened, once anyone could post anything, we discovered an entirely new form of text and video. We learned that the most powerful publishing platform is social, because it embeds the information in a social context. (And we are social animals.) But we also learned about our preferred style, which is the absence of style: the writing that sticks around longest in our memory is what seems to take the least amount of time to create. All art aspires to the condition of the Facebook post.

Mickes, L., Darby, R. S., Hwe, V., Bajic, D., Warker, J. A., Harris, C. R., & Christenfeld, N. J. (2013). Major memory for microblogs. Memory & cognition, 41(4), 481-489.