Could a River Be Made Safe for Everyone?

In 1704, villages along the proposed new course of the Yamato River received an unusual responsibility. They had to look after the survey stakes marking where the river would go. Stakes on the center line carried paper streamers; these markers were removed for safekeeping and put back the next day. Communities facing the loss of their fields were required to help mark out the works that would take them.^[1]

The river was being redirected west toward Osaka Bay. Before the diversion, it had flowed north or northwest from Kashiwara, joining the old Yodo near Osaka Castle. People exposed to that older course had sought a different arrangement. People living across the proposed route had reasons to resist it.^[2]

Both positions are intelligible without imagining that one side liked floods. The problem was that moving a river also moved relationships: between a field and its water supply, a village and its drains, or a lake and the settlements below its outlet.

From late-sixteenth-century works around Kyoto to the Yamato diversion and the adjustable sluice built below Lake Biwa in 1905, the means of intervention changed substantially. What remained difficult was making a gain in one place compatible with life somewhere else. Could engineering solve that problem—or did it merely provide new ways of deciding who would have to bear it?

Water control did not begin with a single, agreed objective called safety.

When Toyotomi Hideyoshi developed Fushimi Castle south of Kyoto in the 1590s, the surrounding waterways were part of the project. Uji City's archaeological interpretation connects the embankment works of 1594 with directing the Uji River toward Fushimi and its port. A river could protect a bank while also serving the location of a political and commercial center.^[3]

The distinction matters. Keeping water away from every inhabited place would have defeated some of the purposes for bringing it closer to Fushimi. A navigable connection was an asset. The desired landscape contained water in selected relationships with roads, buildings, and landing places, rather than excluding it altogether.

Excavation downstream of Uji Bridge recovered a long stretch of late-sixteenth-century riverbank protection. Stone facing, timber stakes, and projecting works addressed the action of flowing water against the bank. The remains make the intervention tangible: directing a current required securing the material exposed to it.^[3]

These structures invite questions at two scales. At the bank, did the arrangement resist erosion? Across the surrounding district, what route was the river now expected to take? A successful answer to the first could help establish a consequential answer to the second. The stronger the bank, the more effectively it could preserve a chosen division between channel and land.

That does not make embankments inherently harmful. It makes their location historically important. An embankment beside a castle port belongs to a different arrangement of priorities from one protecting farmland or a newly opened channel. Identifying its materials cannot, on its own, tell us whose needs determined where it was built.

The Fushimi case also prevents a misleading contrast between an untouched river and later human interference. By the time another authority considered altering these waters, it inherited a landscape already shaped for previous uses. Leaving things alone would preserve some earlier decisions; rebuilding would revise them. Neither option began from a place without human claims.

A century later, opposition to the Yamato diversion made the claims on both sides of a proposed channel unusually explicit.

In 1660, shogunal officials inspected a possible western route for the Yamato. Their survey prompted objections from villages near the proposed line. The objections concerned more than houses and fields that would disappear beneath the new works.^[4]

Several smaller waterways, including the Higashiyoke and Nishiyoke, already flowed north. Villagers argued that the new east–west river would obstruct them. Land to its south would have difficulty draining; land to its north would lose irrigation. The same intervention could therefore create too much water on one side and too little on the other.^[4]

This was a hydraulic argument, expressed through the requirements of cultivation. For a farmer, water being present somewhere nearby was insufficient. It had to reach a field when needed and leave when there was too much. A large, well-contained river could pass close to the farm while making either task harder.

The geometry is important. A proposed route can look clear when attention follows it from beginning to end. The existing connections running across that route are less conspicuous. Yet those connections may be what allow the surrounding land to remain usable. A project changes more than the line it occupies.

The opponents were defending their own interests, but self-interest did not make the physical problem imaginary. Equally, their arguments cannot stand for the experience of communities along the old course. Those communities confronted an arrangement in which the Yamato still came toward them. Keeping the river where it was also distributed danger.

Much of the opposition history survives through village records discussed by historian Yasumura Toshifumi. An important Jōrenji account was compiled in 1796, ninety-two years after the diversion, from the perspective of a community that had resisted it. It preserves a historical claim about the project, rather than an impartial assessment of every district's gains and losses.^[2]

Read alongside the documented surveys and later village accounts, however, that perspective supplies something indispensable. The river's new alignment was not being inserted into vacant ground. It was crossing a working landscape whose inhabitants could explain why the proposed change threatened them. Their disagreement concerned the design of a shared environment, not simply a refusal to cooperate with improvement.

The surveys did not lead immediately to construction. Different proposals and river works intervened. Yasumura's reassessment of the familiar diversion story challenges the idea of one uninterrupted campaign culminating inevitably in success: he places the last identified petition seeking diversion in 1687, well before the project of 1704.^[5]

The gap changes how the decision should be explained. Persistent local demands were part of its history, but the eventual authorization also belonged to the shogunate's changing choices about what works to undertake. A petitioning campaign and a government construction program had related, rather than identical, timetables.

When officials returned to inspect the route in 1703, the objections resumed. A petition addressed the disposal of excavated material as well as lost land and disrupted watercourses. The impact extended beyond the channel itself: earth taken out somewhere had to be deposited somewhere else.^[6]

Representatives subsequently traveled toward Edo to petition, while officials assigned to the construction traveled toward the locality. In the village account, the representatives ended up presenting their request as one for replacement land rather than cancellation. The issue they could continue to press had narrowed from whether the river should move to what would happen to those displaced.^[6]

Once work began, the villages had practical duties. Alongside the survey markers described at the opening, their households had to accommodate workers. At Jōrenji, construction reached fields before the barley harvest was ready. A request to wait was refused; potential harvest helpers were already employed on the river project. Some grain was buried beneath the embankment.^[1]

The conflict here was between schedules as well as between locations. The works needed labor at the moment a crop needed harvesting. Providing wages for construction did not necessarily provide the same households with time to save their grain. A public undertaking could create work and destroy an existing return on work simultaneously.

Nor should the villages' recorded participation be read backward as agreement. Caring for markers and lodging laborers enabled the project to proceed, but these tasks had been imposed. Participation tells us who helped make an outcome possible; the terms on which they participated tell us something different about its acceptance.

The new banks eventually fixed the route on the ground. The consequences for the households beside them were still developing.

Jōrenji's account of its condition in 1743 described losses that had outlasted the construction season. Its assessed agricultural value had fallen sharply. Replacement land included plots on the former river course too distant to retain, while other ground required substantial leveling and well digging. Some remaining houses and fields suffered repeated inundation.^[7]

A replacement acreage could therefore be inadequate even without a dispute about its measured size. Land was useful through its situation: access to water, distance from labor, the work already invested in its surface, and the possibility of reaching it regularly. Assigning a new plot transferred ground; it did not automatically reproduce those conditions.

Compensation consequently had a working life after the award. Leveling rough ground and digging a well required labor before the replacement could supply a harvest. Distance would remain a cost long after that preparation ended. A one-time assignment of land could leave a household with expenses that repeated every season.

Villages north of the new river faced a different continuing task. They installed conduits through the right-bank embankment to obtain irrigation water, paying for the works themselves. The new channel's construction had created a need for further connections across its boundary.^[7]

The embankment was thus both a protective structure and something that water users needed to penetrate. The apparent simplicity of separating land from river concealed a more demanding arrangement: keeping unwanted water out while maintaining routes for wanted water to enter.

Movement across the channel posed another problem. Sakai's historical library account identifies Yamato Bridge as the essential crossing on the route toward Osaka and records its repeated destruction by floods during the Tokugawa period.^[8] Creating a waterway could remove one obstruction to drainage while introducing an obstruction to ordinary travel.

These effects need not cancel every benefit of the diversion to explain why its history remained contested. The useful comparison is between particular changes. One district lost the old river's flow; another acquired a new bank to maintain. A former channel offered land for reclamation; a farm elsewhere lost its irrigation connection. Calling the whole undertaking a success or failure without specifying the affected place would collapse those different outcomes.

The project had rearranged exposure. Later engineers around Lake Biwa would attempt something more flexible: varying the arrangement as water levels changed.

Lake Biwa and the lower Yodo belong to the same river system. Water leaving the lake through the Seta continues toward the Uji and eventually the Yodo, joined by other rivers on its way to Osaka Bay.^[9] For communities around the lake, a freer outlet could bring welcome relief. Farther downstream, it could mean an additional inflow arriving when local rivers were already high.

The relationship made dredging contentious. In the ministry's historical account, high lake levels in 1889 prompted demands to remove a railway bridge then being constructed across the Seta. Residents farther downstream opposed enlarging the outlet. When the Home Ministry undertook dredging in 1893, it limited the work with downstream effects in view.^[10]

The petition blamed an apparent obstruction at the outlet; it does not establish the bridge as the sole cause of the high water. More fundamentally, the two regions were evaluating the same passage from opposite ends.

A narrow outlet could be dangerous for the lakeshore while restraining the amount passing downstream. Simply making it wider would change that relationship every day, including days when the lower river could not comfortably receive more. Conversely, preserving the restriction would leave upstream households exposed for the sake of a downstream benefit they might never see directly.

Dredging could increase the outlet’s capacity, but it could not decide when that capacity should be used. The problem required control over release as well as a larger passage.

The Meiji government's larger Yodo improvement program combined works at several points. Between 1896 and 1910 it developed a new lower channel, the Kema lock and weir, changes to the Uji, and a sluice below Lake Biwa. At the Uji, continuous embankments separated the river from Ogura Pond.^[11]

The range of works mattered because an upstream change could now be considered together with downstream receiving capacity and the arrangement of intervening channels. That offered possibilities unavailable to a project judged only within one village boundary.

Yet a larger planning area did not make the interests within it identical. It brought more of the relevant connections into the same undertaking. The plan still needed a way to respond when those connections carried unusually large amounts of water at different times.

The Nangō sluice, completed on the Seta in 1905, combined an enlarged outlet with the ability to restrict it. Its thirty-two openings were controlled by inserting or removing heavy timbers. Greater discharge capacity and a means of holding water back were designed to work together.^[10][12]

This was a significant change from relying on the width of the channel alone. A fixed bottleneck restrained outflow whether downstream conditions required restraint or not. Adjustable openings allowed managers to change what the outlet was doing.

The opportunity came partly from the different pace of flooding. Lake Biwa's level rose more gradually than the lower Yodo's. Restricting the lake's outflow during a downstream flood could avoid adding the same quantity of water at the most difficult moment; releasing more afterward could help draw the lake down.^[10]

In principle, timing could therefore reduce a conflict that looked unavoidable when considered only as a choice between an open and a closed outlet. Water held temporarily did not have to remain there permanently. The relevant question became how long the delay lasted and what happened during it.

For households beside the lake, however, a delay helpful downstream could lengthen inundation upstream. The capacity to coordinate the system depended on how much room remained before an additional rise caused damage, and on when downstream conditions would allow release. An adjustable structure created alternatives; it did not guarantee a harmless one under every combination of rainfall and river level.

Operating the early sluice also demanded time and labor. According to the river office's historical description, fully opening it took about a day, and fully closing it took two. Those figures describe the old manually operated structure, not the later mechanized replacement.^[12]

A decision to release more water would therefore take effect through a succession of openings. Conditions could continue changing while crews worked. The flood’s timetable and the operators’ timetable had to be considered together; control depended on labor as well as on the structure’s maximum capacity.

This is where the twentieth-century solution becomes more than another example of sending danger somewhere else. It sought to manage when different parts of the basin received water. But the discretion that made coordination possible also created a new point at which people could disagree.

In 1917, a breach in the lower Yodo's embankment at Ōtsuka required repairs. Release from the lake was restricted during the work, and the lake level rose further. The ministry's account records protest meetings in Shiga and negotiations conducted under guard around the sluice.^[10]

The conflict had acquired a specific operating decision. Lakeshore residents were no longer arguing only about the shape of the outlet. They were confronting an authority's decision about how much of that outlet could be used while a damaged structure downstream was repaired.

This differed from the Yamato villages' experience in 1704. There, choosing the new route committed land and disrupted existing connections. At Nangō, the structure allowed repeated choices as conditions changed. Its flexibility could be useful precisely because it did not settle the distribution of water once and for all.

The difference also cautions against treating every conflict as evidence that engineering accomplished nothing. Enlarged channels, protected banks, and adjustable discharge could change the physical possibilities. The potential for reducing combined damage was real. Establishing how much a particular project achieved would require comparisons of floods, flows, exposure, and losses—not simply the survival of either a monument or a petition.

River engineer Ishikawa Tadaharu has emphasized this problem in reassessing historical flood-control practices: documentary descriptions alone can leave essential questions about their physical operation unresolved. Historical interpretation needs to take the behavior of water seriously, just as hydraulic reconstruction needs to take the inhabited landscape seriously.^[13]

No single location or release rate served every purpose in every season. A port needed access, a farm needed irrigation and drainage, and an upstream lake remained connected to a downstream flood. Expanding the area considered by planners helped reveal these connections; it did not make their demands coincide.

Making a river safer involved deciding which connections to preserve, replace, or regulate. Sometimes a new arrangement reduced several dangers together. Sometimes it imposed a continuing burden on a particular community. The change from embankments to controlled releases expanded the available choices without removing those differences.

The history is clearest when we follow the water past the end of the completed work. A defended bank is only one part of the answer. Beyond it are the fields that must still drain, the channels that must still supply them, and the people whose experience determines what the word safe can mean.