人柱 (hitobashira) means “human pillar.” The term describes the burial of a living person — or of a symbolic substitute — beneath the foundation of a bridge, a castle wall, a dam or embankment, or some other major public work, so that the structure would stand.
The reasoning was not cruelty for its own sake. A sacrifice was meant to appease the gods of river, earth, and land; to steady ground that kept giving way; to hold back floods; and to secure a project in an age when engineering failed often. To the people of premodern Japan it was framed as a last and catastrophic measure, taken when the survival of a community appeared to depend on it.
Three anxieties sit behind the idea: the difficulty of controlling nature, the fragility of communal survival, and the belief that certain structures required not only labor but a life.
A careful reading of the sources gives a mixed picture.
No surviving law or official order is known to require human sacrifice for construction. The legal codes and building regulations that do survive say nothing of burying a person in a foundation. Against that absence stand more parallel accounts than are easy to dismiss: folklore, regional histories, and temple chronicles preserve strikingly similar stories in widely separated places. And premodern construction did fail repeatedly, with collapses commonly read as the land refusing the project.
Hitobashira therefore sits between fact and legend. What can be said with confidence is that the legend was written down, in many places and over many centuries: in the court chronicle of 720, in Edo gazetteers, and in the collections of the first Japanese folklorists, who printed what their correspondents in the provinces sent them.[1][2] That a story was recorded is not evidence that the event it describes took place — and the silence of official documents is not evidence that it was hidden. Where excavation or a dated record speaks to a particular site, it should be taken site by site; the legend as a whole cannot be proved or disproved from its own retellings.
Several stories recur across Japan. Their reliability varies, and they tend to combine construction logistics with spiritual belief.
The earliest written instance stands in the Nihon Shoki, the court chronicle presented in 720. Under the reign of Emperor Nintoku it records an embankment on the Yodo river that broke at two places and could not be closed, and a dream in which a god asked for two men, one from Musashi and one from Kawachi, to be given to the river. The first was thrown in, and the gap closed. The second took two calabashes to the water and told the god that if it truly wanted him, it should sink them; they floated, and he lived, and his gap was closed as well.[1] The oldest record already contains a man who argued with the god and survived.
At Matsue Castle (松江城) in Shimane Prefecture, a young woman is said to have been buried beneath the walls when the construction would not stand. Lafcadio Hearn, who heard the story while living in Matsue in the 1890s, was told that for this reason no girl might dance in the streets of the town, because the castle hill would shake if she did.[3]
Several regions tell of bridges that would not stand until a person, chosen by some small sign, was buried at the base. At the Nagara bridge in Settsu, the man taken was the one with a patch on his hakama; at the Ōhashi in Matsue, in the story Hearn heard, it was the first man to cross wearing a hakama without the stiffening board at the back.[3][4] Where the place was marked afterwards, it was sometimes marked with a tree: at Komine Castle in Shirakawa, local tradition keeps a cherry, the Otome-zakura, over the grave of the girl said to have been buried in its wall.
Floods, typhoons, and earthquakes destroyed whole communities as a matter of course. The idea that a structure needed a life to steady it grew out of conditions in which engineering alone failed again and again.
To build was to interfere with unseen powers, and so to negotiate with them by ritual, offering, or sacrifice. A bridge crossed the domain of a river god. A castle altered the geomantic balance of an entire region. A dam interrupted the movement of water spirits.
A working dam or bridge could decide whether a village starved or endured. Within that calculation, a sacrifice could be reasoned out as communal necessity: the structure’s survival was the community’s survival.
One reading of the record’s silence is that such things, if they happened, are exactly what would not have been written down, and that later generations kept them instead as legend, as moral tale, and as haunting — too morally fraught to document, too culturally resonant to forget.
The motif has stayed useful. Local legends attach to haunted bridges, castles, and embankments; the Nagara legend alone became a puppet play in 1727 and a stage play in 1928;[4] J-horror films bind spirits to particular places; light novels, anime, and video games return again to figures sealed inside structures.
It persists because the questions it raises are not confined to one period. What price must be paid to control nature? Whose life is counted as expendable? Why do societies, across cultures, link architecture with sacrifice?
Hitobashira is less a grisly anecdote in Japan’s past than a cultural memory: of fear of forces that could not be controlled, of ingenuity working against real engineering limits, and of what earlier generations were prepared to accept in order to make something endure.
Through these legends a particular world becomes visible. Rivers had tempers and had to be appeased. Foundations were a matter of negotiation rather than labor alone. A single life could be argued to protect many.
Bridges are now built with steel, sensors, and measurement; floods are managed through hydrology and infrastructure; the human pillars survive only as stories. The legends remain as a record of how people once dealt with a world they could not fully understand or control.
The Komine Castle cherry is local tradition of Shirakawa and has no older written source. The account of the belief’s reasoning in “What the Belief Assumed” is KWAIDANOTE’s own reading of these records.
Found an error? Write to kwaidanote@gmail.com — corrections are made on this page and dated.