Legibility and Folk Knowledge

In the villages of South Kivu, in eastern Congo, things are about as bad as they get. Armed groups raid with impunity. The villages lack coordinated defense. Individual acts of resistance are suicidal because a brave minority fighting back against superior firepower simply gets killed. Then in 2012, an elder receives the recipe for gri-gri — a powder that makes you bulletproof — in a dream.¹

The obvious take: superstitious nonsense. The sophisticated take, courtesy of economists Nathan Nunn and Raul Sanchez de la Sierra: gri-gri is a coordination mechanism. Before gri-gri, only the brave fought back, and bravery against machine guns is just early death. After gri-gri, everyone believes they're bulletproof, so everyone fights. A village where everyone fights can actually repel a raiding party. The belief is false at the individual level (bullets still kill people) but functional at the group level (coordinated mass resistance saves the community). Add that gri-gri requires adherents to follow specific moral commandments — don't steal from civilians, for instance — and you get reduced intra-community crime as a bonus. Group selection does the rest: villages with gri-gri outcompete those without, so the belief spreads.¹

This is a clean, publishable result. It's also, as the blogger sam[]zdat argues, catastrophically incomplete in a way that illustrates something fundamental about the limits of legibility.

The Spell That Explaining Breaks

Here's the thought experiment. A well-meaning development worker reads the economics paper, recognizes the value of gri-gri, but wants to modernize the practice. Strip away the superstition, keep the coordination benefits. He arrives in the village and explains: "Good news — we understand why this works now. Gri-gri doesn't actually stop bullets, but it solves a collective action problem by lowering perceived costs of resistance, thereby shifting the equilibrium from defection to cooperation."¹

Nobody fights. Of course nobody fights. The coordination benefit depends entirely on each individual believing they personally are immune to bullets. "The community will benefit from your sacrifice" is a statement everyone already knew — it wasn't enough to make anyone fight before gri-gri, and it won't be after you've explained the trick. The explanation destroys the mechanism it describes. Data breaks the spell.

This is the deep problem with James Scott's Seeing Like a State that most readers miss. The standard reading is: High Modernist states destroy useful practices because they don't understand them. The lesson seems to be: understand the practices first, then modernize. sam[]zdat argues this gets it exactly backwards. The problem isn't that the state failed to understand gri-gri before destroying it. The problem is that understanding gri-gri — translating it into the language of the state — is itself the destructive act. Legibility doesn't precede destruction; legibility is destruction, when the thing being made legible depends on opacity to function.¹

Metis as Worldview, Not Toolkit

The word James Scott uses for this kind of folk knowledge is metis — the practical, embodied, context-dependent knowledge that resists formalization. sam[]zdat pushes on this: metis is usually treated as a toolkit, a collection of useful practices that happen to be stored in traditions rather than textbooks. But gri-gri reveals that metis is also a worldview. The belief in the elders, the trust in dreams as sources of knowledge, the embeddedness in a community that shares these assumptions — these are not incidental containers for the useful coordination mechanism. They're the mechanism itself.

Gri-gri didn't emerge from random cultural mutation. It emerged from the interaction between a pre-existing institution (village elders, witchdoctors, a tradition of revealed knowledge) and a new condition (armed raiding). Without the institution, the new practice couldn't have been generated, transmitted, or maintained. The economists explain why gri-gri spreads and persists once it exists, but they can't explain why any individual initially adopts it — and if you can't replicate the initial adoption, what exactly have you explained?¹

This connects to the cultural evolution literature in a specific way. Henrich's work on causally opaque practices — manioc processing, divination for hunting decisions — makes the same basic argument: practices that work for non-obvious reasons are maintained by high-fidelity cultural transmission, and the most dangerous thing you can do is "understand" them well enough to start modifying them. But sam[]zdat goes further. The cultural evolution framework still treats the practices as tools, maintained by an evolutionary process that doesn't require anyone to understand them. The gri-gri case suggests that some practices require active misunderstanding — that the wrong explanation is load-bearing, and the right explanation is corrosive.

What We Might Be Missing

The most provocative claim in the piece: there are customs that could not adapt to a modern market economy, but whose loss may be psychologically devastating in ways we can't measure. We look at what has survived — weddings, funerals, some religious practices — and assume these were the most important rituals of the past. But why would the rituals that survive market competition be the most important ones? The kind of ritual useful to a farmer is not necessarily useful to an investment banker. If you judge the importance of past practices by which ones survived modernization, you're running survivorship bias on your own cultural heritage.¹

The standard reassurance — "I grew up in modernity and I'm fine" — is, as sam[]zdat notes, what every human in history has said about their own society. The twentieth century is the wealthiest era in human history. It's also the era whose most celebrated philosophical achievements are about how bad things are. We award our cultural laurels to Kafka. That should tell us something.