Organised violence tends to appear where the substrate makes it the higher-return move, and to recede where it does not. That is one way to read the origins of war. Inverted, the same reading yields a second claim, less obvious and more useful. If war is what the ground rewards under some conditions, then peace is what the ground rewards under others, and neither is the default. A long stretch without organised violence is not the absence of something. It is the presence of conditions under which cooperation pays, and conditions like that can be built, and they can wear out.
Which turns the anxious question of the moment into a more tractable one. Not whether the world is becoming more dangerous, but which of the conditions that keep cooperation cheap are getting more expensive to hold in place. The difference is the difference between prophecy and maintenance.
Nowhere left to route around
There is a darker reading of the present, and it captures something real: the exits are closing, and not only in the geographic sense the origins of states turned on. Robert Carneiro’s circumscribed valleys produced conquest states because a defeated group hemmed by desert and mountain had nowhere to walk to, so losing meant submission rather than departure. The unsettling feature of now is that the same condition, no cheap exit, is being reproduced across the whole system at once, on axes that have nothing to do with terrain.
A globalised economy runs through a handful of narrow passages, which are the modern equivalent of Carneiro’s desert wall. The Strait of Hormuz carries about a quarter of the oil traded by sea and nearly a fifth of the liquefied gas traded, through a channel whose navigable lanes are two miles wide. Iran largely closed it during the 2026 war. Since then the status of passage has itself become contested, which is what a chokepoint is for. Malacca, Bab-el-Mandeb, the Turkish and Danish straits, Panama, the Taiwan Strait: whoever sits on a passage holds the thing nobody else can route around. That is circumscription for the entire trading system, and it does not need anyone to intend harm in order to bite.
The pattern extends past shipping. The advanced-chip economy sits on one contested island, where a single firm makes more than nine in ten of the most advanced semiconductors. The energy transition has quietly relocated the old oil geography onto lithium, cobalt and rare earths, whose refining is concentrated in one country across nineteen of twenty strategic minerals. The cables and satellites that carry the information economy are chokepoints made of glass and orbit. Water, among the oldest substrates, is contested along shared rivers and depleting aquifers. Sovereign debt pulls future production into the present and then needs growth simply to service itself. Climate sits under all of it, shrinking the habitable and productive ground and pushing people toward the parts that still function, while borders harden at the moment they are most needed as exits.
The war in Ukraine, still grinding through intermittent truces that have not held, reads on a substrate lens as a contest over ground: grain, minerals, Black Sea access and a strategic buffer. It also carries competing historical narratives, security concerns and ideologies. The rights of that war are a separate question, and selecting one explanatory layer is not the same as claiming the others away. The claim is only that the material layer is carrying more than the stated reasons admit, on more than one side, which is what a substrate reading tends to find in any war once the justifications are set aside.
The common thread through all of it is not scarcity, which has always been present. It is reduced optionality. Each of these is a narrowing of the room to move, walk away, route around, or wait. It is Carneiro generalised: the frontier disappears, and when the frontier disappears, the arithmetic that once made raiding worthwhile comes back, because taking starts to look like the only way left to get more.
The frontier keeps moving
There is a second historical pattern, and it cuts against the first. Human societies have repeatedly escaped apparent limits not by finding new land but by manufacturing new substrate. Agriculture, oceanic navigation, coal, oil, electricity, nitrogen fixation, the shipping container, the internet: each opened a frontier without discovering a continent, converting a ceiling into a floor. So the deeper pattern may not be a finite world pressing in. It may be that the world keeps running out of one substrate a little before it discovers the next, and lives through the gap in fear.
Whether the current squeeze is the terminal one or just another gap before another expansion is not knowable now. Cheap solar, fusion, biotechnology, whatever artificial intelligence turns out to be, material pulled from asteroids: each is a candidate for the next frontier. Both patterns are real, the closing exits and the moving frontier, and which one governs the coming decades is not yet written into the ground. A substrate reading does not settle that, and is not built to: it reads ground already walked on, which is a different job from seeing ahead.
Growth defers the argument
The growth economy belongs on the list, but its mechanism is subtler than the usual complaint about a system that has to expand to survive. Growth is what allows conflict to be deferred. When everyone expects tomorrow to be larger than today, compromise is cheap, because a smaller share of a growing whole can still be an absolute gain. When tomorrow looks smaller, compromise turns expensive, because every settlement is now visibly at someone’s expense. The politics shifts from how to make more to who gets less, and the second question is the one that has historically been decided by force.
This holds inside borders as much as across them. Housing, healthcare, pensions, migration: an expanding economy lets these be positive-sum arguments, and a contracting one turns them distributive, which is a polite word for a fight over who is made poorer. The inequality hypothesis, that surplus creates elites and elites compete, sharpens here. When the pie stops growing, the easy politics of dividing an expanding surplus gives way to the hard politics of who holds the existing one, and elites with armies have always had more say in that than people without them.
Trust is infrastructure
One substrate belongs on the list that is easy to miss, because it is made of nothing that can be weighed. Trust, not as a virtue but as infrastructure. Markets assume accounting. Money assumes belief. Diplomacy assumes that signals can be read. Insurance assumes that contracts will hold. Science assumes that results will reproduce. Each is a load-bearing assumption of shared reality, and each lowers the cost of transacting to something a society can afford. When the shared reality fragments, the cost of every transaction rises at once, because nothing can be taken on trust and everything has to be verified or secured.
This is the security dilemma seen from the other side. That mechanism is worst where nobody can verify anyone, and the information domain may be the most opaque substrate yet built: capability that cannot be counted, intent that cannot be read, deniability that is cheap. Trust is the machinery that keeps verification affordable, and it is quietly one of the things now getting more expensive to maintain.
Peace is engineered
All of which reframes the thing usually filed as a moral achievement. The most durable accomplishment of the order built after 1945 was not peace in the abstract. It was making conquest a bad trade for a large number of states. Germany grew richer trading than it ever did invading, and so did Japan. The point is not that these countries became virtuous. It is that the payoff matrix was rebuilt underneath them, so that the cooperative move paid better than the predatory one.
The clearest case is the one that founded the European project. German coal and French steel, the contested substrate the two countries had fought over within living memory, were pooled under the Schuman Plan of 1950, implemented by a treaty signed the following year and running within two years of that, which turned the thing worth fighting over into a thing that only produced value when shared. That is not diplomacy in the ordinary sense. It is substrate engineering: a deliberate change to the ground so that the higher-return move changed with it.
Seen that way, a long list of unglamorous machinery turns out to be the same class of object. Mutual defence pacts, the interbank settlement system, shipping and salvage law, the insurance markets that price maritime risk, satellite positioning, the agreements that govern undersea cables, container standards, commercial arbitration: none of these is a moral monument, and all of them, by design or by accident, ended up serving the same function, keeping cooperation cheaper than the alternative. They are the industrial-age descendants of the apparatus at Taş Tepeler, where the monument and the feast appear to have formed part of what made working together the higher-return strategy, rather than merely decorating it.
Civilisation, seen from here, stops looking like the opposite of war and starts looking like a large, expensive, poorly documented collection of arrangements that keep the price of violence above the price of trade.
Technology rearranges
Technology sits awkwardly in this, because it does not point in one direction. It opens exits and closes them in the same motion. Container shipping collapsed the cost of distance and widened everyone’s options; semiconductor fabrication concentrated a critical input onto a handful of sites and narrowed them. The internet abolished distance for information; cloud infrastructure then re-centralised it into a few buildings owned by a few firms. Precision agriculture raised the land’s carrying capacity; the rare-earth dependence underneath the machinery concentrated supply. Technology does not reliably open the frontier or close it. It rearranges the substrate.
The upkeep
Prophecy is the wrong output, and the wrong appetite. A substrate reading offers a maintenance schedule instead: it says where to look, and what the looking costs. Prediction rewards whoever sounds most certain. Upkeep rewards whoever is still checking.
Two things continue to price war above cooperation. Modern war remains extraordinarily expensive, widely thought unwinnable in its largest form and, in a networked and armed world, rarely worth the cost of holding what it takes; and the machinery of cooperation built over the past eighty years still lowers the cost of peaceful exchange. Both are under strain. Neither has failed. The open question is a race between the exits closing and the upkeep holding. If the growth ceiling, the epistemic fragmentation and the climate squeeze turn the world zero-sum faster than the machinery can be maintained, the substrate starts pricing conflict as the higher-return move again, region by region. If the cost of modern war and the density of shared interest hold, the alternatives stay cheaper and the machinery survives, strained. That race is genuinely undecided.
The comfortable belief that peace is the natural resting state of things, interrupted occasionally by war, does not survive the arithmetic. The evidence runs the other way. The passages stay useful only while they are kept open and verifiable rather than captured. The frontier stays peaceful only while it stays genuinely positive-sum. The shared reality that makes transacting cheap survives only while it is defended like the infrastructure it is. None of it happens on its own, and none of it is ever finished. Peace is the built state, and building decays without upkeep.