The epistemic state

How do states know? How can they know better?

I’m a researcher at the UC Berkeley Center for Long-Term Cybersecurity where I direct the Daylight Lab. This is a newsletter about cybersecurity and politics—my work as I do it; more than half-baked, less than peer-reviewed. To follow along, subscribe here.

While authors have (rightly) spilled much ink on the state’s role in producing truth, the state is equally responsible for detecting it—in particular, for detecting those brute facts that are not changeable through rhetoric, only (re)contextualizeable.1

Some mundane examples of epistemic roles of the state:

  • Voting & elections. Voting is, in its highest ideal, a way of discovering facts about public opinion.

  • Regulating markets. Markets and pricing are a way of knowing the world. Neoliberalism was “sold,” in part, as a superior system for uncovering brute facts.2

  • Surveillance. Digital surveillance, police observation, foreign spying: through these processes, the state both detects brute facts and tests labels (like “terrorist” and “criminal”) to which they might cohere.

A familiar critique of socialism is that it has “knowing problems:” that top-down planning grows out of touch with bottom-up-needs, and that cultural repression prevents brute facts from percolating up. In the early days of the COVID-19 pandemic, provincial government officials in Hubei had perverse incentives to hide the virus's spread from the central government, producing highly impactful delays in the pathogen's early spread.

Meanwhile, decentralized systems beget different knowing problems. The mishmash of private medical systems in the US, each with its own data format, made it difficult for the CDC to track and contain the pandemic’s spread. For this and other reasons, deaths in the US exploded.3

In any model, states’ ability to know the world is fundamental (if not logically prior) to their ability to project power in it. Phrased differently, all states are states of knowing.

How can states know better, more humanely?

In an age of increasing digital surveillance by governments and corporations, people are rightly wary of states knowing anything. Through a precession of bad examples—the Snowden revelations, the surveillance of Uyghurs in Xinjiang—critics worldwide have a learned response toward governments’ attempts to know the world. Yet the potential value of the state’s epistemic work is inarguable, as illustrated both by the COVID-19 examples above and by the value of core civic institutions like voting—itself a way of knowing.

How can states know better? How can states perform their epistemic work humanely, bucking creeping trends toward nonconsensual surveillance?

One proposal I’ve always liked: direct democracy via public assemblies in the style of Murray Bookchin’s communalism. People on the ground percolate brute facts upward through directly-elected, recallable bodies. These bodies are hierarchical, beginning at 300-person “communes,” each electing members to higher-level bodies. The state of Rojava implements this vision in its constitution.

A key facet of Bookchin's communalism is the strict separation of policy (the making of rules) from administration (the carrying out of those rules). This separation assures that voices on the ground inform policy free from the risk of administrators’ regulatory capture—for example, by producing facts on the ground to solidify their grip on power. In Bookchin’s, people play an active and immediate role in constructing and contextualizing facts on the ground, one that’s hard for the state to revoke or deny.

As ongoing surveillance trends make clear, we need to propose—and enact—better ways for states to know. That requires both sharp critiques and specific proposals. Personally, I’ve always been partial to Bookchin’s vision. Its elegance is tantalizing.

Knowing and power

In Bookchin’s vision, authority still exists—not everyone is equally qualified to speak on every matter. In Bookchin’s state, the unequal distribution of knowledge of brute facts is the root of legitimate authority. Those who know have justifiable authority to enact policy.

But who’s to say who knows? Who decides who knows? This question holds the key to authority in practice.4 In the West, we’re lamenting a lack of trust in experts, their credentials codified by elite institutions, all the while quickly ceding our epistemic rights to private entities like tech companies, whose capacity for surveillance (and mechanistic knowledge work) is unrivaled.5

So, short of starting a Bookchinian revolution, what’s next here? Understanding this process of how authority is delegated—a key concern in science and technology studies—may be more important than ever in understanding futures of governance.

An exercise for the reader

Here’s a thought experiment. Imagine you find the people in your block and form a commune. You meet and vote to elect a representative. Now, you go around your neighborhood and encourage other blocks to do the same thing. Do that organizing block by block, neighborhood by neighborhood, and soon you’ll have a shadow government. That government can levy tax, distribute resources, provision goods—why not? If we all do this, we’ll have ourselves a quiet revolution.

I’m a researcher at the UC Berkeley Center for Long-Term Cybersecurity where I direct the Daylight Lab. This is a newsletter about cybersecurity and politics—my work as I do it; more than half-baked, less than peer-reviewed. To follow along, subscribe here.


In practice, we can’t so easily disentangle a fact from its context. “Seeing” is itself active, embodied, situated in particular communities of practice.


Mirowski, Philip, and Edward Nik-Khah. The knowledge we have lost in information: the history of information in modern economics. Oxford University Press, 2017.


This observation about the Chinese system vis a vis COVID-19 was the subject of a recent piece by Chinese thinktank “Taihe Industry Observer” (钛禾产业观察), translated by Jeffrey Ding in his newsletter last week.


How that process might work in practice is, in Bookchin's writings, underspecified. I wonder how this happens in Rojava. I hope to read an ethnography someday.


This process is Zuboff’s epistemic coup. Perhaps neoliberalism, an ideology that posits that markets are an optimal mechanism for distributing both goods and information, primed the masses for this forfeiture of epistemic power. In a world where your dollar is your vote, some people have more votes than others. Perhaps there never was a forfeiture—perhaps this power always belonged to the wealthy, the few. I’d be curious to hear some arguments on that.