Civilization Vii

The research screen fills from left to right. A player in Civilization VII selects a technology, and a thin line brightens from the completed node to the one beside it; the next box becomes available, then the box past that, then the box past that. Each finished advance deposits a small chime and a card announcing what it unlocks: a building, a unit, a way of moving across the map that was not possible a moment ago. The whole surface is a lattice of boxes joined by lines, and the lines run only one direction. Nothing in the interface permits research to travel backward. A civilization that has learned writing cannot un-learn it. The player reads the lattice the way it was built to be read, as a diagram of where the civilization has been and where it must go next, and the reading feels less like a decision than like the recognition of an order that was already present, waiting to be climbed.

This lattice is the oldest and most consequential thing the Civilization series does. Since 1991 the games have taught tens of millions of people to picture the human past as an ascent: a sequence in which pottery precedes writing, writing precedes currency, and each rung, once reached, holds. The tree asserts a specific set of claims about how history works, and it asserts them so smoothly that most players never notice a claim has been made. It says history is cumulative, that advances accrue and do not dissolve. It says history is ordered, that the sequence is fixed and the same for everyone. It says history is directional, that there is an up and a down and that societies travel up. Above all it says history is single, that all peoples are climbing one tree, and that the differences between them reduce to how far up each has managed to get.

None of this is neutral. The tree is an argument, and the medium has a name for the way it argues. The game scholar Ian Bogost, in his 2007 book Persuasive Games, called it procedural rhetoric: the practice of making claims not through words or images but through processes, through the rules a system enacts and requires the player to enact in turn. A tech tree persuades by being climbed. Every hour the player spends queuing the next advance is an hour spent inhabiting the proposition that the past has the shape of a ladder. The proposition is absorbed as procedure, which is a far more durable way to absorb an idea than to be told it. And the particular proposition the tree enacts is not new. It is one of the most thoroughly criticized ideas in the whole discipline of history. It is the Whig interpretation.

The British historian Herbert Butterfield gave it that name in 1931, in a short and permanently useful book, The Whig Interpretation of History. The Whig historian, in Butterfield's account, studies the past with one eye fixed on the present and organizes everything he finds by whether it led toward the world he already lives in. He sorts the dead into those who advanced the story and those who obstructed it. He treats the winners as the point and the losers as the friction. He mistakes the road that happened to be taken for the road that had to be taken, and he calls the result progress, as though progress were a property of events rather than a judgment imposed on them afterward. Butterfield's warning was not that progress never occurs. It was that the progressive shape is smuggled in by the historian, projected backward onto a past that did not know where it was going and had no obligation to go there. The tech tree is the Whig thesis rendered as an interface. It is Butterfield's abridgment of history, the ramp toward the present, drawn in boxes and one-way lines and handed to the player as the neutral furniture of the game.

Civilization VII makes the argument more explicit than any earlier entry, because it divides the whole of history into three discrete Ages, Antiquity, Exploration, and Modern, and stacks them. Each Age has its own tree, which the player climbs and then leaves behind. Each Age ends in a crisis, a scripted period of instability that batters the civilization before the transition. And at the transition the player carries a persistent leader forward while selecting a new civilization suited to the coming Age, converting some of the previous era's achievements into lasting bonuses and letting the rest fall away. The design is elegant, and it renders collapse more honestly than the series ever has, because collapse is now a thing that happens on schedule rather than a thing that can only happen to the player who loses. But watch what the structure does with the collapse. It formalizes it as a setback between two climbs. The crisis is a valley the player crosses on the way to the next ascent. The tree resets and fills again from the left. Whatever was lost is scored as damage to be recovered. The Age system is the most interesting thing in the game and also its most revealing limit, because it can represent collapse only as an interruption of progress, never as anything progress might not be the right word for.

Set the model against how historical change is now understood to actually happen, and the seams show. The most important word the tree cannot accommodate is contingency. Consider the question that has organized a generation of global economic history: why did sustained industrial growth begin in northwestern Europe and not in the comparably developed cores of Asia. The historian Kenneth Pomeranz, in The Great Divergence (2000), assembled the evidence that as late as the middle of the eighteenth century the most advanced regions of China, above all the Yangzi delta, stood roughly level with the most advanced regions of Europe in life expectancy, consumption, market sophistication, and agricultural productivity. There was no ladder on which Europe stood higher. What Europe had, on Pomeranz's account, were two contingencies that had little to do with civilizational rank: coal deposits that happened to sit near its centers of demand, and colonies in the New World that supplied a land-intensive windfall of cotton, sugar, and timber no domestic acreage could have furnished. Take those two accidents away and the divergence is not delayed but unfounded. The tech tree cannot hold this thought. In the tree, Europe reached industrialization first because it was further up, and being further up is precisely what Pomeranz shows was not the case. The tree can model a head start. It cannot model a fork.

The most instructive witness against the tree is the historian who once drew the grandest version of it in prose. William H. McNeill's The Rise of the West, published in 1963, was for a generation the definitive single-volume history of the human community, and it told the story as a rise, an accumulating drama of cross-cultural borrowing that culminated, as the title promised, in Western ascendancy. It was a magnificent book and a triumphant one. A quarter century later McNeill wrote an essay reappraising it, and the reappraisal was largely a recantation of its shape. He judged that the book had been written from inside a particular moment, the height of American confidence after the Second World War, and that the moment had lent his narrative a directionality the evidence did not compel. He had let the outcome organize the story. The man who had built the finest tech tree in the language looked back and concluded that the tree was the wrong figure for what he had been describing, that the web of exchanges he had traced had no single up, and that the sense of a destination had been supplied by the historian and his times rather than by the past. When the author of the model disavows the model, the disavowal is worth more than any outside critique.

There is a second thing the tree cannot represent, which is refusal. In the tree, a technology once researched is possessed permanently and universally desired; no civilization ever declines an advance or lets one lapse, because the interface offers no verb for declining and no state for having forgotten. Real societies did both, constantly, and often for reasons that were not failures of capacity. The anthropologist James C. Scott, in Against the Grain (2017), reconstructed the deep history of the earliest states and found that the arrangement the tree treats as an unambiguous early unlock, settled grain agriculture organized under a state, was for most of the people inside it closer to a trap. Early states ran on grain because grain is legible: it ripens on a schedule, sits above ground, and can be measured, taxed, and seized. That legibility built the granary and the tax roll, and it also built the fragility, the epidemics of crowding, the coerced labor, and the walls that kept subjects in as much as enemies out. People fled these states when they could, and the peoples beyond the frontier, whom the settled scribes called barbarians, frequently ate better and worked less. The move the tree scores as the first rung of civilization was, measured in the welfare of the bodies who made it, frequently a step down. Progress toward the state was for many of them regress, and the tree has no way to draw a rung that points the wrong way.

The third thing the tree cannot represent is that complexity is a cost. Joseph Tainter's The Collapse of Complex Societies (1988) proposed a spare and powerful account of why elaborate civilizations come apart. Societies confront problems by adding complexity, more administration, more specialization, more infrastructure, and complexity delivers real returns, until it does not, because the marginal return on each new increment of complexity declines. Eventually a society is paying enormous fixed costs to sustain a structure that has stopped yielding proportionate benefit, and at that point collapse, the rapid shedding of complexity, can be the rational move, an economizing rather than a catastrophe, experienced by many of the people involved as relief from an overhead they were being crushed to maintain. This is the reading the tech tree forbids most completely. Up the tree is defined as better by the tree itself; complexity is the score. So the loss of complexity can register only as harm, a crisis to be survived, a red number, the valley between two climbs. The possibility that simplification was adaptive, that a society might rationally choose to be less complex, is not a move the interface can offer, because the interface has already decided which direction is good.

That decision is the whole argument, compressed. The tree encodes a direction and presents it as physics, when direction is the one thing about history that is not given. "Advance" is a value word wearing the costume of a measurement. The arrow that runs one way through every node is not a law the game discovered; it is a premise the game installed, and everything the player experiences as the natural momentum of history is that premise being enacted. This is what Butterfield meant. The progressive shape is not found in the past. It is brought to the past by whoever is doing the arranging, and in the case of the game the arranger is the rule-set, which arranges so fluently that the arrangement disappears into the feel of play.

None of this is a case against the game. Civilization VII is one of the most sophisticated thinking tools the medium has produced, and the correct response to a powerful model is not to discard it but to know what it is a model of. The tree is wrong in the specific, nameable way that all useful models are wrong: it holds some variables still so that others can be seen moving. What it makes visible is real and hard to see elsewhere. It shows how early advantage compounds, how a choice made in the first hour narrows the choices available in the fifth, how the parts of a society lock together so that changing one strains the rest. Those are genuine features of historical systems, and the game teaches them in the body, through the hands, more vividly than a paragraph can. The danger is not the model. The danger is mistaking the model for the world, letting the ladder in the interface become the ladder in the mind, and coming away believing that the past was a queue of advances waiting to be selected in order.

The frame this essay wants to give the reader is that any tech tree is an argument about history, and that the argument can be disagreed with. Once the frame is available it travels widely, because the tech tree is one of the most reproduced structures in games. It sits inside nearly every strategy title, every colony simulation, every grand-scale management game, and in a thousand smaller forms it is the skill tree, the upgrade path, the research web, the progression bar, all of them quietly enacting the same claim that capability is linear and accumulates in a fixed order. And it reaches past games entirely, because the tech tree is only the crisp digital form of a folk model most people already carry: the intuition that societies can be ranked as advanced or primitive, that history is a development ladder with the present at the top, that some peoples are ahead and others behind on a single track everyone is walking. That intuition is the Whig interpretation living in ordinary minds, and it does real damage in the world, because it licenses the ranking of human beings by where their society is imagined to sit on a track that does not exist. This is why the frame is worth the trouble. Every player who has climbed a tech tree half-believes that history has one. The belief was there before the game and the game confirms it, hour after satisfying hour, and naming the belief as a belief is the first move toward being able to hold it more lightly.

Return to the transition screen at the end of an Age. The crisis has run its course. The leader endures, and beneath the leader a civilization is dissolving so that another can take its place, its accomplishments sorted into the fraction that carries forward and the fraction that falls away. The tree is wiped clean, and after a moment the first node of the next Age lights up, available, inviting the same climb that just ended, the lines already drawn, already running one way. It is a beautiful screen and a persuasive one, one of the most persuasive arguments the medium has ever made about the shape of the human past. History does not have a tech tree. The screen that insists it does is making a claim, not showing a fact, and the moment a player can see the claim as a claim is the moment the player is free to disagree with it, and to keep playing anyway, which is the only honest way to use a model you have understood.