You're standing at the bank of the Rhine somewhere between Strasbourg and Karlsruhe. The water is wide, grey-green, indifferent. On the far side, Germany. On this side, France. The river looks exactly like the Loire looks from a bridge in Tours: the same reluctant current, the same herons, the same basic indifference to human arrangements. Yet the Loire runs its entire length inside France and nobody loses sleep over it. So why this river and not that one?
The honest answer is that rivers became international borders almost never because of what rivers are, and almost always because of what armies, treaties, and colonial surveyors happened to need at a particular moment.
The logic that made a river convenient
A river is, in practical terms, an obstacle. Before modern bridging, crossing one required boats, fords, or seasonal ice. That friction made rivers useful to anyone trying to draw a line that could be defended or at least observed. You could station guards at a crossing point. You could see someone coming. The river didn't create the boundary; the boundary-makers reached for the river because it was already doing half the work of a wall.
The Rhine is the textbook case. It didn't become a Franco-German border because geographers decided the watershed was meaningful. It became one through the Treaty of Westphalia in 1648, then again through the Peace of Lunéville in 1801, then again after two world wars. Each time, negotiators ratifying exhausted military positions reached for a line that both sides could agree was visible, permanent, and mutually legible. The river was already there. It was free.
Contrast that with the Loire. France absorbed Anjou, Brittany, and the Loire valley through dynastic marriage and conquest before any rival power had a foothold on the opposite bank. By the time the modern state system hardened around the seventeenth century, the Loire was already interior. No treaty needed to name it. No army had ever faced another across it in a way that required a settlement. The river just kept flowing, politically invisible.
Now consider two hypothetical officers, call them Moreau and Hoffmann, assigned in 1815 to demarcate a post-Napoleonic border in central Europe. They're standing in a forest with a mandate to produce a line their respective foreign ministries will actually ratify. There's a ridge to the north. There's a river to the south. The ridge is ambiguous: it shifts, it has no single crest, and the local villagers disagree about where it runs. The river is unambiguous. Both men can see it. Neither government will dispute it. Moreau and Hoffmann pick the river before lunch and go back to write their reports. That's how it usually worked.
What people get wrong about watersheds
The popular assumption is that rivers mark natural ethnic or cultural divides. They rarely do, and the Mekong is about as clean a rebuttal as geography offers. It does not separate a Lao-speaking population from a Thai-speaking one in any coherent way. Communities with shared languages, trade networks, and family ties historically lived on both banks. The river became an international border in the 1890s when France, extending its Indochinese empire, and Siam, trying to preserve its sovereignty by conceding a boundary the British would recognise, agreed on the river as the simplest available line. The Lao villagers on the western bank became Thai. Their cousins on the eastern bank became French colonial subjects. The river didn't reflect a division. It manufactured one.
This matters because borders made from rivers carry a particular illusion of naturalness, the way a magician's table looks sturdier than it is precisely because you've been invited to inspect it. Rock faces and straight surveyed lines look arbitrary. A river looks inevitable. That appearance of inevitability is, in most cases, a fiction, and a politically convenient one. Governments on both sides find it useful to imply that the boundary was always there, written into the landscape before anyone arrived to read it.
The rivers that stayed internal were simply the ones that never found themselves between two rival consolidating powers at the critical moment. The Colorado's entire political fate rested not on its width or its depth but on the accident that Mexico and its northern neighbour settled their territorial argument along a different waterway further south. Had the negotiators in 1848 drawn the line one river east, the Colorado would appear on every map of contested borders and the Rio Grande would be a footnote in a hydrology textbook.
Timing, military stalemate, colonial partition logic, and the bureaucratic preference for legible lines all mattered far more than any property of the water itself.
Which means the next time you look at a map and see a river serving as a border, the right question isn't what's different about that river. The right question is who was standing on opposite banks, unable to agree on anything except that the water between them would do.