The attention commons

In 2013, a hashtag that made it into Twitter's daily top 50 stayed there for an average of 17.5 hours. By 2016 the figure was 11.9 hours.¹ The peaks were no lower; topics rose just as high. They rose faster and fell faster, and the public moved on sooner.¹

The researchers who measured this, Philipp Lorenz-Spreen and colleagues, found the same steepening in data spanning decades: words in Google Books, box-office sales, Google searches, posts on Reddit.¹ Their explanation is the interesting part. In a simple model of topics "competing for finite collective attention," the acceleration comes from more content being produced and consumed, "resulting in a more rapid exhaustion of limited attention resources."¹

Finite, competed for, exhausted: that is the vocabulary of a commons. Whether it is the right word depends on which attention you mean.

Two kinds of attention

The case against calling attention a commons is strong when you look at one person. Elinor Ostrom, who won the Nobel prize in economics for her work on commons,² described common-pool resources as goods that share "subtractability" with private goods (what one user takes, another cannot) and difficulty of exclusion with public goods (it is hard to keep anyone out).² Forests, fisheries and the atmosphere fit.² Your attention fails the second test. It is easy to exclude others from it, because it is in your head. When a platform captures an hour of yours, it has not taken that hour from anyone else.

But there is a second layer. At any moment a public holds a small number of things in common: the subjects that many people are thinking about at once, that can be talked about without explanation, and that can therefore be acted on together. Call it collective attention. It is subtractable: if a society's shared focus is on one subject, it is not on another, and the number of subjects it can hold at once is small. And it is very hard to exclude anyone from: anybody with money, outrage or novelty can bid for a place in it, and nobody owns it, so nobody charges for entry. At this layer, the definition fits.

Ostrom also insisted that these properties are matters of degree. She proposed treating subtractability and excludability as varying "from low to high rather than characterizing them as either present or absent."² So the question is not whether attention is a commons, yes or no, but how commons-like each layer is. Individual attention is barely commons-like; collective attention is a lot more so.

What overuse looks like

A commons is overused when each user gains from taking more and the cost is spread across everyone. A fishery shows overuse as fewer fish. Collective attention would show it differently, because the resource does not shrink. People do not pay less attention in total.

The best evidence for that is mundane. Andre Boik, Shane Greenstein and Jeffrey Prince studied the online behavior of a large panel of US households from 2008 to 2013, years of huge change in what was available online. Total time online, and how widely and deeply households spread it, stayed "remarkably stable."³ The households behaved as if each had "a fixed set of 'slots' of attention" to allocate, and new offerings changed where the time went rather than how much there was.³

A fixed budget with more and more claimants has only two ways to give: crowd things out, or turn over faster. The Lorenz-Spreen data show the second. Peak attention to a topic stays the same height, but it is reached and lost sooner, and the shifts between topics come more often.¹ The shared space keeps its size and loses its hold.

That is what depletion of an attention commons would look like: not less attention, but a public that holds each thing in common more briefly.

Where the cost falls

If that is right, the cost is not evenly spread. A subject that can be understood in a day loses little when collective attention turns over faster. A subject that needs a public to keep thinking about it for months, such as a slow risk or a long public works project, loses a great deal, because it is competing for a space where everything now has less time. No study read for this piece measures that cost directly; it is where the argument predicts the damage, not where it has been found.

The usual governance answer to a commons also points at a gap. One of Ostrom's design principles for commons that last is that someone monitors the condition of the resource.² Platforms measure attention constantly, but only their own share of it, because that is what they sell. No one is assigned to measure the condition of the shared layer: how many things a public can hold, for how long, and whether that is falling. The 2019 study is one such attempt, made by physicists and data scientists from outside the platforms.¹

Where this breaks

The acceleration is not universal. Lorenz-Spreen and colleagues found no such steepening in citations of scientific papers or in traffic to Wikipedia articles, and suggest those systems are driven less by popularity.¹ So the claim is narrower than "all shared attention": it is about the shared attention that popularity governs. That is most of public life, but not all of it.

Faster turnover could also be good. If more topics get a turn, more voices are heard, and a public that moves on quickly may simply be better informed about more things. The commons reading assumes that holding a subject in common for longer is worth something. That is a judgment, though a reasonable one for any problem that takes longer than a week to solve.

And the model is a model. It reproduces the data from rising production and consumption of content, which is consistent with overuse of a shared resource, but other mechanisms could produce the same curves.¹ The household data come from one country and end in 2013.³

The commons nobody tends

The usual objection to the phrase "attention commons" is that attention is private. That objection is right and answers the wrong question. The thing being worn down is not any one person's attention. It is the small, shared space where a public notices the same thing long enough to do something about it.

It is finite, anyone can bid into it, nobody owns it, and the people who profit from it measure only their share. Those are the conditions under which commons get overused. Whether this one can be governed, and by whom, is open; the first step would be for someone to start counting.