In 1348 the Black Death killed between a third and a half of Europe’s population. In Western Europe, serfdom began to dissolve. In Eastern Europe, at the very same time, a harsher bondage was born — what historians call the second serfdom. Same shock. Opposite outcomes.

This chapter is about why. It sets out the five conditions that decide whether ordinary people, in any age, end up with a seat at the bargaining table or without one — and it is the framework the rest of the book uses to ask the same question about artificial intelligence.

The common skeleton

Since the fourth chapter, we have traced six great transformations one by one: the egalitarian but fragile balance of hunter-gathering, the first dependency born of the agricultural revolution, the intermittent and indirect bargaining of the feudal order, the contradictory legacy of the commercial revolution, the intellectual autonomy opened by the Reformation, and finally the concrete, measurable leverage of the industrial revolution. Now it is time to step back and draw out the common skeleton underlying these six different periods. Because this skeleton will be our only reliable compass when we return, in the eleventh chapter, to today’s age of data.

Leverage was never a gift

The first and most fundamental observation is this: leverage, in no period, was born of the ruling class’s generosity; each time, it arose from an indispensable contribution the people made to that period’s process of production. In hunter-gathering, this contribution was symmetrical and the leverage direct (the threat of leaving); in agriculture and feudalism, the contribution became locked to land and labor, and leverage became indirect and crisis-dependent (the threat of revolt, post-epidemic labor scarcity); in the industrial revolution, contribution became locked to factory production, and leverage took on a measurable, organizable form (the strike). As the resource changed, the form of leverage changed, but the underlying logic never did: whoever had the power to halt or slow the process of production had a seat at the bargaining table; whoever did not — as with the Atlantic slavery system — had no such seat.

The window of vulnerability

The second observation is that a window of vulnerability opens at the start of every great resource transition. In the early centuries of the agricultural revolution, in the first wave of the commercial revolution, in the first decades of industrialization (the Luddite period), the people found themselves in an intermediate stage in which they had already lost their old leverage but had not yet defined their new leverage. This window is where history operates most harshly against the people — because the old bargaining tools (the hunter-gatherer’s flight, the artisan’s guild) no longer function, while the new tools (the union, the vote) have not yet been invented. The most striking parallel this book will draw in the eleventh chapter, between the industrial worker’s position in the 1810s and today’s data producer’s position, lies precisely here: both find themselves in a window of vulnerability in which the old leverage has dissolved but the new one is not yet recognized.

How the window closes

The third observation concerns how this window closes: almost never through a clean, peaceful negotiation. The German Peasants’ War, the English Peasants’ Revolt, Luddism, the first Chartist waves — all of these were crushed, bloodily, in the short term. But every failed attempt carved a threshold-knowledge into the ruling class’s memory: this burden has a limit, and if that limit is crossed, the cost grows heavier. Both the Mesopotamian debt-relief edicts and Bismarck’s social insurance reforms are, in fact, the same logic manifesting itself in two different centuries — power, rather than sitting directly at the negotiating table, chose instead to buy off a possible explosion in advance. This observation teaches us something: the recognition of leverage most often comes about not directly and at the first attempt, but through a roundabout path — first a failure, then a memory, and finally a preemptive settlement. A right won directly and on the first attempt is, by the historical pattern, the exception, not the rule.

The five conditions

The fourth observation is that the conditions required for leverage to become a lasting institution consist, in every period, of roughly the same five elements.

1. Scarcity. The people’s contribution must truly be irreplaceable. Labor scarcity in agricultural society (after the Black Death), skilled-labor scarcity in the industrial period — leverage functioned only in situations where this contribution could not easily be sourced elsewhere.

2. Visibility and measurability. The industrial worker’s decision to strike was concretely visible in union membership rolls, in production figures; this visibility made bargaining possible. By contrast, the direct but unorganized leverage of the hunter-gatherer period never became institutionalized, because there was no unit of measurement to institutionalize.

3. Organizability. Guilds, unions, the Chartist movement — all were mechanisms that turned scattered individual contributions into a collective bloc. Without organizational capacity, even the scarcest, most visible contribution is negotiated and crushed one person at a time; as we saw in the sixth chapter, this is precisely the gulf between the organized guild artisan and unorganized slave labor.

4. The possibility of flight or halting. The hunter-gatherer’s changing groups, the medieval city’s Stadtluft macht frei principle, the industrial worker’s going on strike — each is a situation in which an exit or halt option concretely existed. When this option closed, as in the Atlantic slavery system, the other four elements of leverage, even if present, effectively stopped functioning.

5. Competition within the elite class itself. This is perhaps the least-noticed yet most effective element. Competition for flight among manors, competition for population among Protestant and Catholic princes, competition for migrants between colony and homeland — the people’s leverage more often gained strength not from raising a demand directly, but from one of the ruling class’s own rivals offering the people better conditions in order to weaken another. Leverage is always weaker before a single, undivided master; it strengthens when there is a choice among rival masters.

Why the same shock produces opposite outcomes

Placing these five elements side by side, we can understand both why leverage sometimes institutionalized quickly — in the industrial revolution, over the course of the nineteenth century, all five elements were largely present at once — and why it sometimes remained suspended for centuries, as in the feudal period, where organizability and visibility were lacking.

This framework also has a place for the Black Death example we pointed to in the sixth chapter. In Western Europe, because scarcity, visibility (wage figures), the organizability that urbanization provided, and the possibility of flight were all present together, serfdom dissolved. In Eastern Europe, because most of these elements were absent, the exact opposite — the second serfdom — was born instead. The very same shock produced different results on different ground, because what determines the outcome is not the shock itself, but the combination of these five conditions beneath it.

The compass, pointed at our own age

This is the compass we will carry into the eleventh chapter. Facing the new resource created by the artificial intelligence and robotics revolution — data, computing power, decision automation — as we assess the people’s position, the question we will now ask has become clear.

Is the production of data in this new age truly scarce and irreplaceable, as we pointed to with the model collapse finding in the third chapter? Can this contribution be made measurable and visible? Can data producers organize? Do they hold a possibility of flight or halting in their hands? And finally, is today’s master class — the technology companies and the states — running a competition among themselves that works in the people’s favor, or are they forming a monopolistic unity?

In the eleventh chapter, we test these five questions one by one, with current examples, and move into the book’s densest chapter: a concrete discussion of where each of these five conditions stands in today’s data economy, which of them are missing, and how that absence might be closed.