The Slate columnist Matt Yglesias provoked a strange debate on Tuesday. Somehow, a Twitter discussion of the legitimacy of redistributive taxation became an exchange with the libertarian Brian Caplan about the implications of John Locke’s theory of property for the historical expropriation of Native Americans. On Yglesias’s reading, Locke claims that “Property is held by right owing to its historical chain of possession from an original legitimate acquirer through voluntary exchange and bequest.” If that’s the really the case, Yglesias suggests, none of the property in land in the United States is legitimate. Because this conclusion is absurd, Yglesias dismisses the principle.

The conclusion that there’s no legitimate real property in America, with the possible exception of Indian reservations, is absurd. But it doesn’t follow from Locke’s account of property, or at least from his intentions. In fact, that account was developed at least partly to justify the seizure and occupation of Indian lands.

The key is the concept of the legitimate original acquirer. For Locke, one doesn’t acquire land merely by claiming it, or even by living on it. One acquires solely through productive use. Moreover, one has no right to hold more land than one can use productively.

The Indians, as Locke saw it, were not just failing to turn “their” land to productive use. By leaving vast expanses to the game they hunted, they were actively depriving others of their right to acquire property in the proper way. For Locke, the settlers who built farms, dug mines, or grazed herds were the legitimate original acquirers.

Indeed, Locke didn’t even stipulate successful cultivation as necessary to legitimate title to otherwise “waste” land. As I understand the Second Treatise of Government, he believed that it was sufficient to intend to do these things.

On a strict Lockean view, then, the settlement and cultivation of the Americas was no theft. It was a legitimate acquisition that initiates rather than interrupts the historical chain of possession. For Locke, in other words, the Native Americans couldn’t be expropriated because they had no property in the first place. Indeed, their attempts to defend from encroachment lands where they happened to hunt, say, were actually a form of theft from the prospective settlers.

Does this argument succeed in justifying settlement? Probably not: Locke seems to have been mistaken about the agricultural practices of at least some Indian tribes, which may have met his own criteria for ownership. But the problem with Locke’s theory of property is not really empirical. Rather, it’s that Locke encourages the fantasy, still popular among some libertarians, that property relations can be generated by a sort of immaculate conception in which no one’s freedom is violated. David Hume got closer to the truth. As he observed in Of the Original Contract, “there is no property in durable objects, such as lands or houses, when carefully examined in passing from hand to hand, but must, in some period, have been founded on fraud and injustice.” It’s time and positive law, not the original acquisition, that makes property legitimate.