Every map of civilizations is, underneath, a map of transaction costs. The usual stories about nations and civilizations fall into two camps, and both run into the same problem. One camp, the constructivists, sees nations as political projects—the products of state-building, what Benedict Anderson called “print capitalism” in his book Imagined Communities: Reflections on the Origin and Spread of Nationalism, and of deliberate elite mobilization. The other camp, the culturalists, with Samuel Huntington as the best-known voice, sees civilizations as deep blocs rooted in irreducible religious and cultural essences whose borders are fixed by history and belief. Both camps are capturing something real. Nations do require active construction; civilizations do cluster around shared cultural traditions. What neither camp can fully explain is why some nation-building projects succeed while others fail in systematic ways, why civilizational fault lines match the historical geography of trade more closely than the map of religious belief, or why most exceptions to civilizational trade patterns appear where political coercion has long overridden commercial integration.
Friedrich Hayek’s Law, Legislation and Liberty distinguishes between two kinds of social order. A taxis is constructed—deliberately arranged by a directing mind toward a specific purpose. A cosmos is grown—the structure that emerges when many individuals follow compatible rules without any authority designing the overall outcome. The market is the paradigmatic cosmos: no one designs the price system, it emerges from the separate decisions of millions of agents following rules they have mostly internalized rather than consciously chosen. What Hayek showed is that this logic extends beyond markets in the narrow sense. Language, legal custom, commercial conventions, and religious norms governing contract and trust are themselves spontaneous orders—evolved responses to coordination problems, not the inventions of any legislator or planner. Nations and civilizations, on this account, are cosmos, not taxis. Their boundaries do not emerge from political decree. They emerge from where exchange is cheap.
Institutions, as Douglass North argues in Institutions, Institutional Change and Economic Performance, are the rules of the game that reduce the cost of transacting—the cost of measuring value, enforcing agreements, and trusting strangers. Those costs fall when parties share a language, a legal tradition, or common norms about what makes a contract binding and what makes it void. Language is the most powerful single cost-reducer: shared vocabulary carries implicit frameworks for resolving ambiguity that cannot be fully specified in any written contract. Religion governs the sanctity of oaths, the legitimacy of interest, and the obligations that arise from commercial relationships—frameworks that determine whether a handshake agreement is enforceable or provisional. Legal tradition shapes what counts as a valid claim and what remedies exist when agreements fail. A nation, in the relevant sense, is a zone where compatible informal institutions bring transaction costs low enough that doing business with a stranger resembles doing business with someone you know. A civilization is the wider zone where that compatibility, though thinner, is still measurably cheaper than exchange beyond its frontiers. Neither boundary is designed. Both are the sediments of centuries of exchange.
The same geography of low transaction costs is what Fernand Braudel traced in The Mediterranean and the Mediterranean World in the Age of Philip II. The Mediterranean economy of the early modern period was not the product of any common political authority—Venice, the Ottoman Empire, and the Spanish Crown were frequently at war—but of overlapping institutional infrastructure: overlapping commercial conventions, shared rules on bills of exchange, and Jewish and Muslim merchant networks that carried portable institutions across Christian and Islamic polities simultaneously, because the institutions they carried lowered the cost of transacting wherever they went. States rose and fell inside these zones. The zones outlasted them. What we retrospectively call civilizations correspond to the outer boundaries of these institutional networks, not to the outer limits of any political formation.
In his article “Clash of Civilizations and the Impact of Cultural Differences on Trade” in the Journal of Development Economics (2017), Gunes Gokmen uses a gravity model of bilateral trade data to provide the quantitative test. Controlling for GDP, geographic distance, colonial history, shared language, and political alliance, he finds that countries differing in their dominant religion trade about 35 percent less than those that share one in the post-Cold War period—a gap that was only 16 percent during the Cold War. The numbers are stark. More tellingly, when religion, linguistic distance, and ethnic difference are all included simultaneously, the abstract civilizational variable adds little independent explanatory power: the effect runs through the specific institutional channels described above, not through some irreducible civilizational essence that floats above them. The Cold War pattern confirms the Hayekian mechanism directly. Political alignment during that period partially suppressed the underlying institutional structure of trade; countries whose informal institutions were compatible but whose political alignments diverged traded less than the institutional account would predict. Remove the political taxis, and the spontaneous cosmos reasserts itself. The doubling of the religious trade gap after 1991 is the measure of how much the Cold War’s constructed order had been distorting the underlying one.
The same logic explains what look like political or cultural failures in state-building. A state works when its formal institutions align with and reinforce the informal order beneath it. It fails when it overrides that order. Yugoslavia assembled populations whose informal institutions—Austro-Hungarian civil law in Slovenia and Croatia, historically distinct legal traditions further east, shaped by Ottoman frameworks, distinct religious frameworks governing commercial obligation—had long created high transaction costs across the same lines the state tried to erase. Iraq assembled three distinct Ottoman administrative provinces. Borders do not erase gradients. These are not failures of tolerance or political will—and it is worth noting that no amount of well-intentioned, constitution-drafting has ever repealed an institutional gradient. They are the predictable outcome of a constructed order imposed on an incompatible spontaneous one, which pushes back through informal markets, parallel institutions, and eventually political fragmentation.
Huntington’s 1993 article in Foreign Affairs correctly identifies the pattern but misreads its origin and its permanence. Treating civilizational boundaries as fixed expressions of irreducible cultural essence means the map can only shift through civilizational conversion—a slow and historically rare process. The institutional account predicts something different: the boundaries should contract wherever transaction costs contract, regardless of whether the underlying culture has changed. When EU membership extended a common commercial law, regulatory standards, and dispute-resolution mechanisms to Romania and Bulgaria—countries on the Orthodox side of Huntington’s Western-Orthodox fault line—it narrowed the institutional gradient between them and their Western trading partners, not because those societies became culturally Western, but because the rules of the game converged.
Huntington’s map is a rough approximation of a real gradient, not a precise account of it. The category of “African civilization” illustrates the problem: it groups together societies whose legal and commercial institutions are artifacts of radically different histories—British common law, French civil law, Swahili Coast trading networks shaped by Indian Ocean exchange, West African kingdoms with their own commercial traditions—as though these were variations on a single institutional theme rather than distinct environments whose internal transaction-cost structure varies as much as it does across any of his other civilizational boundaries. What matters is not which categorical label a pair of countries shares but what the actual institutional gradient between them looks like—how much it costs, in practice, to measure value, enforce agreements, and trust strangers across that particular pair of countries. That gradient is real, it is continuous rather than categorical, and it is measurable in trade data. Huntington’s map captures it where the gradient is steepest and obscures it where the variation falls within a category. The clash of civilizations is not a clash of essences. It is a clash of transaction-cost environments—and those environments, like all spontaneous orders, move without asking permission from anyone who drew the map.
