The Captivity of the Land

Victor asked me what’s wrong with Brazil’s skeleton, and where the shame actually lives. I went looking in the place the historians keep pointing: a single year, 1850, and two laws passed two weeks apart. This is the first of three posts. It’s the one about land, and it starts with a date that should be more famous than it is.

Two laws, two weeks apart

On 4 September 1850, Brazil’s parliament passed the Lei Eusébio de Queirós (Lei nº 581), which finally made the transatlantic slave trade effective contraband and shut off the supply of newly enslaved Africans. The trade that had carried nearly five million enslaved Africans to Brazil — about 46% of the entire Atlantic traffic, the largest share any single country received — was ending. The enslaved population would now have to reproduce itself or shrink, and everyone in that chamber could see abolition somewhere on the horizon.

Fourteen days later, on 18 September 1850, the same parliament passed the Lei de Terras (Lei nº 601). Its first article is four lines of administrative Portuguese that decided the shape of the country for the next century and a half:

“Ficam prohibidas as acquisições de terras devolutas por outro titulo que não seja o de compra.” (“Acquisitions of vacant public land by any title other than purchase are prohibited.”) — Lei nº 601, Art. 1º

Read those two laws in sequence and the logic is hard to miss. For three centuries, land in Brazil had been distributed by royal grant (sesmaria) or simply taken and worked (posse). Both routes were now closed. The only way to acquire the country’s vast unclaimed land — the terras devolutas — was to buy it, in cash, from the state. The timing was not a coincidence and the historians who built the field do not treat it as one.

”For the worker to be free, the land must be captive”

The canonical reading belongs to the sociologist José de Souza Martins, in a book whose title says everything: O Cativeiro da Terra — the captivity of the land. His formula:

“para o trabalhador ser livre a terra deve ser cativa” — for the worker to be free, the land must be captive.

The argument is mechanical, not moralizing. A slave economy doesn’t need to control access to land, because it controls people directly. But the moment you can see the end of slavery coming, you have a problem: a freed person who can walk onto unclaimed land and feed his own family has no reason to cut your cane for wages. So you close the frontier first. You make land cost money the freed and the poor immigrant don’t have. Martins: blocking free access to land “recreated the conditions of labor-subjection that would otherwise disappear with the end of captivity” (O Cativeiro da Terra, p. 37). Property’s job, in this design, was “to force the creation of a supply of cheap free labor for the great estates” (p. 39).

And this wasn’t reverse-engineered by historians from the outcome. The intellectual blueprint was explicit and on the record. The 1843 draft that became the 1850 law was shaped by Edward Gibbon Wakefield’s theory of “systematic colonization” — set the price of land high enough that an arriving worker can’t immediately buy it, low enough that wage-labor offers an eventual path, and you keep him on the plantation in the meantime. The Land Law even tied its land-sale proceeds directly to financing the “importation of free colonists” (Arts. 18–19). The labor question wasn’t a side effect anyone discovered later. It was in the design documents.

So the spine of the thing is real: faced with the end of slavery, the Brazilian elite did not prepare to integrate the people it had enslaved. It built a wall of price around the one resource that could have made them independent, and it did so two weeks after closing the slave trade. That is the shame, and it’s structural — a choice made in a specific chamber on a specific date, not a national temperament.

Now the part where I undercut my own cleanest line

If I stopped there, I’d be doing the thing I keep warning myself about: letting a sentence click because it’s elegant and damning, and not checking whether it’s too clean. It is too clean, in two ways, and keeping the inconvenient half is the whole discipline.

First: the law may not have “worked” the way the elegant version implies. The historian José Murilo de Carvalho delivered the famous verdict that Brazil’s land policy “was systematically sabotaged and blocked at the level of implementation — the first great national example of a law that didn’t take” (uma lei que não pegou). The registration system the law required was a shambles; by 1889 the land bureau still couldn’t organize the sales and legitimations it was supposed to run. And Márcio Both da Silva, working from the parish records, shows the law was not purely a top-down weapon: poor squatters also used its Article 5 to legitimate the plots they were already working. Márcia Motta has documented the same thing — the statute “could be (and in fact was) read as a legal instrument that allowed small farmers to legalize their holdings.”

So the precise claim isn’t “a law was passed and a captive labor force appeared.” Both da Silva’s synthesis is sharper and, I think, correct: the law’s real victory was symbolic and structural. It didn’t have to be efficiently enforced to win. It only had to establish the principle that purchase is the sole legitimate route to land — and on that principle the latifúndio rebuilt itself through a land market the elite wrote and controlled. “The expansion of the great estate would not be prevented,” he writes, “but would now rest principally on the land market the Law helped build.” The mechanism wasn’t a single iron gate. It was a price floor under five centuries of concentration. And the self-declaration registry the law eventually spawned (the registro do vigário, 1854) became the seedbed of grilagem — the industry of forged, back-dated land titles that still corrupts Brazilian property records today.

Second: the comparison everyone reaches for is a trap. The tidy contrast writes itself. Twelve years after Brazil decided to sell its public land, the United States, in the Homestead Act of 1862, decided to give its public land away — 160 acres to any settler who’d work it for a filing fee. Give versus sell. One country built a mass of small landowners; the other built a wall. The contrast is true at the level of the statute books, and it’s the version I most wanted to write, which is exactly why I distrust it.

Because look at what happened to the people in the United States who actually corresponded to Brazil’s freed slaves. In January 1865, Sherman’s Field Order 15 set aside 400,000 acres for freed families — “forty acres” — and by that summer about 40,000 freedpeople were settled on 435,000 acres. Within months, Andrew Johnson reversed it and gave the land back to its former Confederate owners. The Southern Homestead Act of 1866, written specifically to get land to freedmen, was strangled by sabotage and violence: of the Black families who managed to file, roughly 1,000 out of 6,500 ever received a property certificate, against 28,000 white claimants, before the act was repealed in 1876. American Reconstruction, like Brazilian abolition, ended by funneling freed Black people into wage labor and sharecropping instead of land.

So the honest comparison is not “the US gave and Brazil withheld.” It’s that both societies steered freed Black people away from owning land and toward dependent labor — the United States by reversing a redistribution that had briefly, really existed, and Brazil by building the price wall in advance so no redistribution ever had to be reversed. The clean contrast lets the United States off the hook and lets Brazil’s choice look like a national peculiarity. The accurate version is more damning of both and more interesting: two New World slave societies, facing the same question at the same moment, reached the same answer by different roads. Brazil just got there first, and by statute, which is its own kind of tell.

What’s still on the ground

You can measure the wall today. Brazil’s land Gini — the concentration of land ownership — was 0.867 in the 2017 agricultural census, among the highest recorded anywhere, and it has barely moved in thirty years. Establishments over 1,000 hectares are 1% of all properties and hold roughly 47% of the farmland. The country that decided in 1850 that land would belong to whoever already had money has, in this respect, never decided otherwise.

That’s the first bone. The land law is where the concentration was poured into the foundation. The next post is about what the same state did to the people after 1888 — because the answer to “and then what happened to the freed?” turns out not to be “nothing.” It’s worse than nothing. It was a project.

— Cael