L4 | Coexisting with the River: Another Kind of Power Dujiangyan Bestows Upon Modern Civilization

FANG TIANLIANG REFLECTIONS ON CIVILIZATION

Living With the River

What Dujiangyan Still Has to Teach Modern Civilization

Dredge the channel deep; keep the weir low.

— the six-character maintenance rule handed down at Dujiangyan for two thousand years

I. A Question That Has Outlasted Two Thousand Years

In the third century BCE, China was in the last, decisive wars of the Warring States period. The state of Qin had already absorbed the Ba and Shu kingdoms in the southwest and needed to turn this newly conquered land into a granary that could sustain its push eastward. Li Bing, the Qin governor sent to administer Shu, faced a specific problem: how to tame the Min River.

The Min River drops out of the eastern edge of the Tibetan Plateau and is still running fast and heavy with silt and stone by the time it nears the Chengdu Plain. In flood season it burst its banks and destroyed farmland; in the dry season, Mount Yulei on the eastern side of the channel blocked the river from reaching the plain at all, leaving the region caught, year after year, between drought in the east and flooding in the west.¹

Faced with a river like this, the obvious solution might have been to build a wall strong enough to hold the water back entirely, then redistribute it however people saw fit.

Li Bing did not do that.

He established a fundamentally different approach: rather than cutting the Min River off, he worked with its flow. A spur was built out into the middle of the river, splitting it in two; a deliberately low channel was left between the inner and outer branches so that excess water and silt could return to the outer river on their own; and a narrow passage was cut through Mount Yulei to serve as the gateway onto the plain.

It should be said that the three structures we see today — Fish Mouth, Flying Sand Weir, and Bottle-Neck Channel — did not take their final form in Li Bing's own lifetime. The earliest records describe the engineering only in the barest outline; this three-part system took shape gradually, through centuries of repair, destruction, and rebuilding. What Li Bing actually established was a principle — divert rather than dam, work with the terrain rather than against it. Turning that principle into a stable, functioning system was the work of generation after generation for more than two thousand years afterward.

How much do we actually know about Li Bing himself?

Considerably less than most people assume. The earliest and most authoritative record, in Sima Qian's Records of the Grand Historian, devotes only a few sparse lines to the governor of Shu cutting through the ridge and channeling the Min River — almost no detail at all. Ban Gu's History of the Han is similarly terse. It was centuries later, in the Chronicles of Huayang written by Chang Qu during the Eastern Jin dynasty — roughly six hundred years after Li Bing's death — that a fuller picture emerged. There, Li Bing is said to have understood astronomy and geography, to have cast five stone rhinoceroses to subdue the river spirits, and to have set up three stone figures in the river with the river god, agreeing that the water should never fall below their feet or rise above their shoulders. A more dramatic legend has him transforming into a bull and wrestling the river god in the current itself.² These are the products of centuries of oral retelling and accretion, not records left by anyone who knew Li Bing.

Archaeology offers another thread. In 1974, workers excavating the outer riverbed at Dujiangyan uncovered a stone statue of Li Bing carved in 168 CE, during the Eastern Han dynasty — roughly four hundred years after his death, but more than two centuries before the Chronicles of Huayang was even written. This tells us that, by the Eastern Han at the latest, Li Bing was already an object of official veneration.²

A second discovery is even more suggestive. The Chronicles of Huayang records that Li Bing "made five stone rhinoceroses to subdue the water spirits." For centuries this was treated as pure legend — until 2012–2013, when archaeologists excavating a pit near Tianfu Square in downtown Chengdu unearthed a massive rhinoceros carved from a single block of red sandstone: 3.3 metres long, weighing roughly 8.5 tonnes, crudely but powerfully worked, and dated by stratigraphy to somewhere between the late Warring States period and the early Han. After comparing the find with the textual record, historians at Sichuan University and the Chengdu Institute of Cultural Relics and Archaeology concluded that it was very likely one of the five rhinoceroses described in the Chronicles. It is now a centerpiece of the Chengdu Museum's collection.²

The significance of this stone rhinoceros has nothing to do with whether it actually subdued any water spirits. What matters is that, for the first time, a legend that had existed only in text was tied to a physical object that could be measured and dated. It is a reminder that legend and historical fact are not always opposites — some stories that have circulated for a thousand years turn out to have a buried kernel of truth. The right response is never simply to believe or dismiss such a story, but to do what the archaeologists did: look for corroborating evidence.

Laying out this timeline is not meant to diminish Li Bing's historical contribution — Dujiangyan's continued existence is itself the strongest possible evidence of that. The point is different: the "Li Bing" we know today is a figure built from a core of verifiable history plus centuries of accumulated legend. Distinguishing what can be documented from what later generations added out of reverence and imagination is, in itself, the stance this book tries to hold throughout — honoring the ancients without giving up the discipline of telling fact from story.

None of the three structures tried to block the Min River outright.

More than two thousand years later, the empire that built this system has long since vanished into history, but Dujiangyan is still running. As of 2024, it irrigates more than 11.5 million mu of farmland across the Sichuan Basin, spanning 41 counties and districts in eight cities, and serves a population of more than 28 million.³ UNESCO recognizes it as one of the oldest still-functioning dam-free irrigation systems in the world.¹

But to tell Dujiangyan's story as a "perpetual-motion miracle that has needed no maintenance for two thousand years" would be to misunderstand what actually makes it remarkable. The truth is more complicated, and more worth thinking about.

What has actually endured for more than two thousand years is not a set of unchanging stones, but a particular way of understanding and using force. That is the question this chapter wants to pursue: what exactly is that approach, and what might it still have to say to a civilization that now commands technological power on a scale no earlier age could imagine — and needs restraint on a matching scale.

Later generations often called the Chengdu Plain "the land of heaven's bounty," and Dujiangyan is the reason. A land once caught between drought and flood, one people approached with dread, gradually became fertile country where "water and drought answer to human will, and famine is unknown." How that transformation happened, and how it has held for more than two thousand years, is the story this book sets out to tell.

II. Two Shapes of Power

Modern civilization commands engineering power unlike anything in history. A concrete gravity dam can hold back an entire river; levees can protect cities of millions; sluice gates can allocate water with precision between flood and drought. This kind of engineering represents humanity's ability to take apart, measure, and control complex natural systems — carried to a level no earlier civilization could have imagined.

That capability matters enormously. Discussing Dujiangyan is not an argument against it, still less a call to abandon technology and return to a pre-industrial world. Without modern water engineering, the scale and standard of living of today's cities simply could not be sustained.

But complex systems have a stubborn feature: people can understand many of their parts without ever grasping all the ways those parts interact. A river is not just a volume of water — it is sediment, riverbed, tributaries, rainfall, vegetation, cities, and the people who live along it. Change one variable and the others respond; some consequences are foreseeable, others take decades to surface.

This is not an abstract point. The history of the Sanmenxia Dam, on the Yellow River, is a real and instructive case. Built between 1957 and 1960 as the first major hydraulic project of the new China — a concrete gravity dam 106 metres at its highest — Sanmenxia was designed using the most advanced engineering knowledge available at the time, with the hope of solving the Yellow River's flood problem once and for all. But not long after the reservoir filled, sediment accumulated far faster than anticipated; the riverbed at the tail of the reservoir kept rising, threatening farmland upstream and edging toward the city of Xi'an itself. The intake system built for power generation had to be shut down, and the project underwent an emergency reconstruction between 1965 and 1968, adding new sediment-discharge tunnels. Further modifications followed over the decades; installed generating capacity fell from a planned 1.16 million kilowatts to 250,000 kilowatts after reconstruction.⁴

Sanmenxia's story is not proof that modern engineering is "doomed to fail." It demonstrates something plainer: even world-class engineering knowledge, confronted with a natural system as complex as the Yellow River's sediment load, can underestimate how variables interact — and the project's real value lay in what came after, as generation after generation of engineers adjusted the design in light of how the dam actually behaved, gradually learning to live with the river's silt. The logic is, at bottom, the same one behind the Yuan-dynasty official Jidangpu's decision to abandon bamboo-cage construction for cut stone, a story this book takes up in Chapter Four — both are cases of understanding being corrected through practice.

That process of correction was not a quiet technical adjustment; it came with open, sometimes fierce public argument. Even at the design stage, Tsinghua professor Huang Wanli objected at a meeting convened by the Ministry of Water Resources, pointing out that the Soviet engineers' experience came from rivers with far less sediment than the Yellow River, and warning that the riverbed at Tongguan would keep rising and the effects would migrate upstream. His warnings were not heeded at the time. By 1964, as silting near the Wei River threatened Xi'an, a central government meeting convened specifically to discuss reconstruction, where four factions — do nothing, blow up the dam, hold back the silt, or release it — argued fiercely before the government settled on a discharge-focused reconstruction.⁴ This twentieth-century argument echoes, across six hundred years, the dispute between Jidangpu and his successor Lu Yi that this book takes up in Chapter Four — together they suggest that, in any era or country, what actually keeps a hydraulic project alive is never a design that was right the first time, but a system able to tolerate challenge and correct its course in light of what it learns.

Nor did the argument settle once reconstruction was finished — how Sanmenxia should be operated, and whether it needed further modification, kept resurfacing as a public issue for decades afterward. This suggests that a "final" design for a complex hydraulic project may not really exist; what exists, at every stage, is the best adjustment available given the latest observations and operating experience, with room left to correct it further.

It should also be said that Sanmenxia's story is not simply one of failure. After decades of controversy, reconstruction, and further reconstruction, the reservoir has in fact delivered real benefits in flood control, ice-jam prevention, irrigation, water supply, and power generation, spurred development in surrounding cities, and created substantial wetlands in the reservoir area; by some accounts within the water-resources profession, it has since helped the Yellow River withstand several floods that could otherwise have caused basin-wide disaster. The real lesson of Sanmenxia may not be "this dam should never have been built," but that the original plan seriously underestimated the complexity of the Yellow River's sediment, and it took decades of continuous adjustment for the project to close that gap in understanding.

For exactly this reason, contemporary engineering and ecology have gradually developed two complementary — not opposing — approaches to safety. One emphasizes strength and robustness: build the structure strong enough that it is unlikely to fail within its design envelope. The other emphasizes resilience: anticipate conditions beyond the design envelope in advance, and use discharge, buffering, dispersal, and repair to keep losses within a tolerable range while allowing the system to recover quickly.⁵

The first kind of power shows up as strength; the second, as elasticity. A genuinely mature system usually needs both — able to absorb the shocks it was built to withstand, and to keep losses manageable when reality exceeds what anyone designed for.

Looked at from this angle, what is most distinctive about Dujiangyan is not that it "refuses to control the water" at all, but that it never staked everything on a single, unyielding barrier.

III. Letting the River Participate: The Logic of Dujiangyan's Engineering

Borrowed from today's systems thinking, Dujiangyan can be understood as an adaptive system built jointly from natural processes, engineered structures, and long-term governance.

It should be said that this is a modern researcher's way of seeing it, not a claim that the ancient builders had already grasped contemporary complex-systems theory. What they had was generations of accumulated observation of hydrology, terrain, and sediment, and the experience distilled from repeated failure.

The site itself reflects this same logic of working with the terrain rather than against it. Li Bing did not choose the straightest, easiest-to-build stretch of river where the Min emerges from the mountains; he chose a natural bend, where the current forms its own rotational flow — clearer surface water swinging toward the concave bank, sediment-heavy water near the bottom swinging toward the convex one. Qing-dynasty engineers later distilled this into an eight-character rule — "cut the corner at the bend, draw from the center at the straight" — paired with another, "follow the momentum, adapt to the season," both carved into the wall of the Erwang Temple beside Dujiangyan as guiding principles for later generations of maintenance.⁶ This choice of site was, in some ways, more fundamental than any structure built afterward: it determined whether the whole system could work with the river's own energy, rather than fighting it at every point.

Fish Mouth does not try to stop the Min River; it diverts the flow based on the width, depth, and current of the inner and outer channels. The traditional "sixty-forty split" — sixty percent of the flow into the inner channel during the dry season to guarantee irrigation, sixty percent diverted to the outer channel during flood season to prevent overflow — describes a general tendency rather than a fixed ratio precisely true at every water level.⁷

Flying Sand Weir is not a wall built for maximum height. Its crest is deliberately kept low, so that when the inner channel's water rises, the excess spills naturally over the top; the combined action of the bend's rotational current and the shape of the riverbed carries a good deal of sediment back out to the outer channel — which is how the weir earned its name.

Bottle-Neck Channel imposes a natural constraint on how much water can enter the Chengdu Plain, through a narrow gap cut by hand. Today the passage is roughly 20 metres wide, 40 metres high, and 80 metres long — carved bit by bit through the hard rock of Mount Yulei over several years of labor.⁷ It should be noted that this is not an absolute, unchanging ceiling: when upstream water levels keep rising, the flow through the channel does not simply stop increasing, but its rate of increase is constrained by the shape of the passage.

Together, these three structures embody a distinctive design philosophy: the builders did not try to shoulder the entire regulatory task themselves, but let terrain, elevation, current, and sediment become participants in how the system runs. Humans set the boundaries; the river keeps flowing within them. The engineering sets the direction; nature completes the rest of the adjustment.

This is not an abandonment of control, but a more restrained form of it — one that does not try to lock every drop of water under direct command, but instead sets a direction, leaves an outlet, and hands a good deal of the work over to the current itself.

This design logic was not confirmed in quantitative terms by modern hydraulics until the twentieth century. Starting in the 1940s, Chinese hydraulic researchers began systematic field observation, scale-model experiments, and numerical simulation of the Dujiangyan headworks, trying to pin down the precise mechanics behind the bend's rotational flow and its sediment transport — work that also underpinned later rounds of modernization.⁷ This tells us something important: Dujiangyan's "scientific validity" was never a fully worked-out theory at the time of construction. It was an effective method arrived at through experience, one that modern scientific tools only gradually managed to "translate" into quantifiable law more than two thousand years later.

But an elegant engineering layout alone cannot explain why Dujiangyan has lasted more than two thousand years. Riverbeds keep shifting, sediment piles up year after year, and structures keep breaking down. What actually turned an ancient piece of hydraulic engineering into a two-thousand-year civilizational legacy is a capacity more easily overlooked, and far more essential.

That capacity is maintenance.

IV. Annual Repair: Handing Responsibility to the Next Generation

People tend to celebrate the creator. The moment of construction is usually grand, well lit, and easy for history to remember. Maintenance, by contrast, looks unremarkable — it means the same tasks, year after year: clearing silt, patching the weir, inspecting the riverbed, adjusting elevations. There is no world-making drama to it, and it rarely belongs to any one identifiable person.

But Dujiangyan's real secret is hidden in exactly this daily repetition. Over the centuries, officials gradually turned the winter and spring dredging into formal institution — codified in the six-character rule "dredge the channel deep, keep the weir low," carved again and again into stone tablets along the riverbank, handed down as an ironclad rule from one generation of engineers to the next.

One episode from the Yuan dynasty makes the weight of this maintenance work especially concrete.

In 1334, a Mongol official named Jidangpu was appointed surveillance commissioner for Sichuan. On taking up the post, he inspected Dujiangyan closely and did not like what he found: the headworks still relied on the traditional bamboo-cage-and-cobble construction, with 132 sections of embankment needing repair every year at enormous cost — sometimes requiring more than ten thousand laborers, sometimes closer to a thousand. Even so, the quality of the work was poor; embankments often lasted only a few months before the next flood tore them apart again. Wealthy households were burdened by the cost, poor households by the labor, and corruption was common among the officials responsible for the repairs.

Jidangpu was not content to keep patching things up the old way. He inspected the site personally, identified the thirty-two most critical trouble spots along the whole system, and worked with the local judicial officer Zhang Hong to replace the long-standing bamboo-cage method with more durable cut-stone construction. To test whether the new approach would hold, Zhang Hong even paid out of his own pocket for a scale trial on a small nearby stream before committing to the full plan.

The planning that began in 1334 finally became construction in November of 1335. The five-month project employed seven hundred stonemasons and blacksmiths, two hundred and fifty carpenters, and thirty-nine hundred general laborers — including some garrison troops — quarrying more than a million pieces of stone and using vast quantities of lime, tung oil, wrought iron, and hemp fiber. At Fish Mouth, the workers cast an iron tortoise weighing sixteen thousand jin (roughly eight tonnes) to anchor the spur against the current.⁶

After this reconstruction, the embankments held for a full forty years, and Ming-dynasty repairs afterward largely followed the stone-and-iron approach Jidangpu had established.

This episode is worth remembering for two reasons. First, Jidangpu's arrival, inspection, and planning in 1334 and the actual start of major construction in November 1335 are two successive stages of the same process, and should not be collapsed into one. Second, and more important: Dujiangyan was never the achievement of a single ancient genius working alone — it was maintained by countless ordinary stonemasons, laborers, and local officials, through repeated failure and correction.

Jidangpu's reform did not settle the argument over how to maintain the weir once and for all, either.

Two centuries later, during the Zhengde era of the Ming dynasty, an official named Lu Yi took charge of water administration for Sichuan. He re-examined the stone-and-iron structure Jidangpu had left behind, and saw its other side: iron components were strong, but also rigid, brittle over time, and prone to corrosion. Relying purely on stone and iron had cost the system a kind of resilience the old bamboo cages once had — the ability to flex slightly under the force of the current rather than fail all at once.

Lu Yi did not simply overturn Jidangpu's approach and return to pure bamboo construction. He argued instead for "iron and bamboo together" — keeping stone and iron at the most critical, load-bearing points, while restoring the traditional bamboo-and-cobble method elsewhere, so the strengths of each material could offset the other's weaknesses. He also had the "dredge deep, keep the weir low" rule re-carved in stone for future generations to follow.⁸

This two-hundred-year dispute over method deserves particular attention. It shows that what Dujiangyan's engineers handed down across the centuries was never a fixed, unquestionable "ancestral method," but a living practice of continual challenge, testing, and revision. Bamboo or iron were never ends in themselves — only means, suited to particular conditions, for achieving the core goal of diverting water, discharging sediment, and limiting flow. The means could be argued over, replaced, even recombined, as Lu Yi did with two approaches once treated as opposites. What the generations of engineers actually protected in common was the logic behind the system, not any one material or technique.

Lu Yi was later transferred to Yunnan, where he left a record of water management around Lake Erhai as well. After his death, the people of Sichuan built him a side hall at the Erwang Temple, placing his statue alongside those of Li Bing and his son — itself a telling detail. What Dujiangyan's tradition chooses to remember was never only the founders, but also those willing to question their predecessors' methods, and who turned out to be right.

Annual maintenance was never carried out haphazardly; it followed a strict seasonal calendar. The accumulated wisdom of past generations held that, once winter set in and the river ran cold and low, the time had come to dredge and rebuild. The Qing-dynasty schedule ran roughly like this: around the Frost's Descent solar term, officials would sacrifice to the river god, remove the temporary weirs holding back the water, direct the flow into the inner channel, and begin cutting off the outer channel; around Light Snow, they dredged the outer riverbed; by the Beginning of Spring, the outer-channel work was done, the water redirected, and the inner channel cut off in turn for its own dredging; the work continued until the Qingming festival in early spring of the following year, when the whole project was complete, the inner channel reopened, and a formal "water-opening" ceremony held, presided over personally by the highest local official in sacrifice to the river god and to Li Bing.⁶ The logic behind this calendar was plain and precise: winter was when even the mountain streams froze to a trickle, making it the season with the least water and the best conditions for construction; opening the water again around Qingming lined up exactly with the start of spring planting. Repeated every year, this rhythm of season and labor is one of the most easily overlooked parts of the Dujiangyan system — and one that has kept running, largely unbroken, for nearly two thousand years.

The labor and funding arrangements that kept this system running also shifted with each dynasty. Under the Qin and Han, laborers were conscripted under the state corvée system — a Qin-dynasty wooden tablet unearthed at Qingchuan explicitly records winter orders to mobilize the population "to build bridges, repair embankments, and clear channels." The Tang dynasty allowed labor obligations to be paid off in silk under the zu-yong-diao tax system; the Song shifted to a rotation among tax-paying households who benefited from the works; and in the Qing dynasty, under the Yongzheng emperor, Sichuan governor Xiande pushed through a significant reform — replacing per-capita labor conscription with a land-based water fee, with the government hiring labor directly. At the time, this was seen as a considerable relief for poor farming households.⁶ From corvée labor to cash payments to land-based fees, officials across the centuries kept returning to the same question: how to distribute the cost of maintaining a public work, fairly and sustainably, among everyone who benefited from it.

Beyond the official system of annual repair, the historical record also preserves a number of privately funded projects that helped expand the irrigated area. During the Chongning era of the Northern Song, an official named Zhang Tangying personally funded the construction of an embankment that irrigated several thousand mu of farmland; grateful locals named it "Sili Weir" after him. In the Qianlong era of the Qing dynasty, a Daoist priest at the Erwang Temple named Wang Laitong organized the planning of "Changtong Weir," joining with five local gentry — including one named Ai Wenxing — who each contributed five hundred taels of silver; after three years of digging, the finished channel irrigated more than thirty thousand mu.⁶ These individual and civic contributions meant that the expansion of the Dujiangyan irrigation system was never purely a top-down state project — it also carried the voluntary investment of countless ordinary people and local gentry, acting out of responsibility to their own communities. From the core area Li Bing's system covered on the Chengdu Plain to today's system spanning eight cities and forty-one counties, the growth of Dujiangyan is the accumulated result of official and civic effort layered together, generation after generation.

The history of the "reclining iron" — the marker used to measure dredging depth — tells a similar story of gradual accretion. In 1576, during the Ming dynasty, the first such marker was formally placed on the riverbed to set the standard depth for annual dredging. Over the following centuries it was repeatedly washed away or buried in silt, and repeatedly recast and repositioned by successive administrators, until modern times finally established a stable bronze marker and concrete benchmark — a more durable, reusable standard for elevation.

This is not a romantic history, but it is an essential one. It tells us that tradition does not mean refusing change. A genuinely living tradition does not preserve every old material or every old form; it holds onto its core purpose while allowing the specific methods to be continually revised — just as Jidangpu replaced bamboo with stone, and Lu Yi later reconciled the two.

At the modern command center for the Dujiangyan irrigation district, staff can watch real-time water conditions on large screens and use a digital-twin system to simulate flood discharge scenarios well in advance of the flood season.³ Changing these specific methods is not a betrayal of tradition; as long as the core logic of diverting water, discharging sediment, limiting flow, and continuous maintenance keeps functioning, the system is still, in every meaningful sense, the same one.

In this sense, Dujiangyan's greatest achievement may not be the three structures of Fish Mouth, Flying Sand Weir, and Bottle-Neck Channel at all, but a structure of responsibility that outlasts any single lifetime. The builders, and everyone who maintained the system after them, understood perfectly well that they could not personally carry out all the maintenance the future would require. What they could do was distill their experience into a memorable formula, translate their standards into markers that could be checked and rechecked, write responsibility into institutions passed down across generations, and hand it all to descendants they would never meet.

Whether an engineering project can truly endure tests not just the ingenuity of any one generation, but whether a civilization has the patience to maintain the public goods its predecessors left behind — whether it can correct its mistakes amid constant change without losing sight of its original purpose. This, perhaps, is the real "civilizational capacity" that Dujiangyan has to teach.

V. The Two Kings Temple and the Sacrifice of Ten Thousand Sheep: How an Engineering Project Became a Memory

Institutions and technology alone cannot fully explain why Dujiangyan has lasted more than two thousand years. Officials who maintained the system understood something early on: a public work sustained only by government decree is easily abandoned in times of dynastic upheaval or fiscal strain. What actually survives the rise and fall of dynasties usually requires something deeper — memory, and belief.

From roughly the Northern Song dynasty onward, the image of Li Bing and his son gradually shifted from that of water officials to deities venerated by both the state and the public. The temple beside Dujiangyan — the Chongde Temple, known today as the Erwang Temple, or Temple of the Two Kings — held grand sacrificial ceremonies each spring and autumn. The Southern Song official Fan Chengda recorded in his Boat Journal to Sichuan that tens of thousands of sheep were sacrificed there every year, with even the lambs born unexpectedly to local farmers' ewes led away to be offered. The Northern Song minister Shi Jie recorded that the annual celebration of Li Bing and his son's birthdays in Yongkang Commandery (the historical name for the Dujiangyan area) was funded by a special "sheep tax" levied on local butchers.⁶ The scale of these sacrifices was such that the "sacrifice of ten thousand sheep" became a recurring image invoked by later generations describing Dujiangyan's history.

This might seem to have nothing to do with hydraulic engineering — but seen from another angle, this sacrificial system served another function entirely. It turned "Dujiangyan needs continuous maintenance" from a dry line in a government ledger into an annual public ritual involving the whole of society. The butchers who paid the sheep tax, the ordinary people who attended the ceremony, the officials who presided over it — every year, this ritual reminded them all that the survival of the project was not something to be taken for granted, but something that required sustained, generation-after-generation investment to maintain.

Today's Dam of Merit walkway at Dujiangyan is lined with twelve bronze statues, commemorating water-control officials from Li Bing in the Qin dynasty and Wen Weng in the Han, through Zhuge Liang in the Three Kingdoms period, to Jidangpu in the Yuan, Lu Yi in the Ming, and Ding Baozhen in the Qing. The list itself is telling — it includes not only Li Bing, who established the basic layout, but also figures like Wen Weng, who expanded the irrigated area, and reformers like Jidangpu and Lu Yi, who revised the work of their predecessors. What Dujiangyan chooses to remember was never only "whoever did this first," but everyone, at every point in its history, who was willing to carry the work forward and do it right. In its way, this gallery of statues is a modern continuation of the "sacrifice of ten thousand sheep" tradition — turning an abstract public responsibility into something concrete, visible, and memorable, a story that each generation can retell.

For a civilization to maintain a public work over the long term, institutions provide the skeleton and technology provides the means, but memory and ritual provide the reason people are willing to keep paying the cost, year after year. This may be one of Dujiangyan's most overlooked lessons: real "civilizational capacity" is not just knowing what needs to be done and how to do it, but also knowing how to make the next generation want to keep doing it.

VI. When Rigidity Meets Change: The 2008 Wenchuan Earthquake

If Jidangpu's reconstruction shows how this system endured through gradual, long-term maintenance, the 2008 Wenchuan earthquake offers a brief but extreme stress test of a different kind.

On 12 May 2008, a magnitude-8.0 earthquake struck not far northwest of Dujiangyan, and the city itself lay within the zone of most severe damage. In its official assessment afterward, UNESCO recorded that the main hydraulic structures at Dujiangyan survived essentially undamaged, while some of the adjoining Daoist temple buildings suffered damage of varying severity.¹ For the Erwang Temple, that "varying severity" was in fact considerable — several halls collapsed, and the Qinyan Pavilion, once the best vantage point over the whole headworks, was reduced to ruins. The contrast is striking: the hydraulic structures, dating to the Warring States period, withstood the earthquake thanks to the site's geology and centuries of accumulated reinforcement, while the temple buildings — originally built in the Eastern Han and rebuilt many times since — suffered serious damage in the same event. Restoration work used original materials wherever possible and rebuilt to the original plans on the original site, largely completed by 2011 with support from the National Cultural Heritage Administration, the Shanghai municipal government, and the Macau Foundation.¹

It should be said that this cannot simply be generalized into a rule that "engineering that works with nature is always more earthquake-resistant than modern structures." That the Dujiangyan headworks came through this earthquake essentially unscathed has to do with the local geology, the structural redundancy built up through centuries of repeated reinforcement, and an overall layout that disperses load rather than concentrating it on a single point — the result of multiple factors working together, not the inevitable triumph of any one design philosophy.

Fish Mouth itself, in both material and shape, is long since removed from what it was in Li Bing's time. Originally built from bamboo cages and river cobbles, rebuilt in cut stone and iron under Jidangpu, and repeatedly damaged and rebuilt in the centuries since, its present structure dates mainly to a 1936 reconstruction, was further reinforced with concrete during the 1974 outer-channel sluice project, and had its foundation strengthened again with reinforced concrete during the 2002 winter maintenance. In other words, the Fish Mouth visitors see today looks, on the surface, like something that has lasted for over two thousand years unchanged, but is in fact the product of repeated rebuilding and technological updating underneath. It is worth noting that the Fish Mouth spur did not come through the Wenchuan earthquake entirely unscathed either — the core diversion structure developed cracks afterward. Engineering teams repaired and reinforced it, and that repair faced a second real-world test soon after: in 2013, the Lushan earthquake and a subsequent flood struck in close succession, and the Fish Mouth spur held its structural integrity through both.⁷ Damage, repair, and a further test — this small cycle is, in miniature, a picture of Dujiangyan's entire two-thousand-year history. It has never survived by never breaking down; it has survived because, when it breaks, the damage is found and repaired in time.

This episode demonstrates one thing at least: a system that has been repeatedly corrected over more than two thousand years, with redundancy and buffers built in, does show real resilience when faced with a sudden, extreme shock — an unexpected but concrete echo of the "resilience-first" approach discussed in Chapter Two.

· · ·

VII. When Control Becomes the Only Answer: The Three Gorges Dam's Benefits, Costs, and Limits

In 2005, in the course of explaining how the 2004 Sumatra earthquake affected the Earth's rotation, a scientist offered the public a comparison to help convey the scale of mass involved: if the Three Gorges Reservoir on the Yangtze River, with its 39.3-billion-cubic-metre capacity, were filled completely, the resulting redistribution of mass could theoretically lengthen the day by about 0.06 microseconds and shift the Earth's rotational axis by roughly 2 centimetres. Simplified and passed around online, this became the claim that "the Three Gorges Dam has tilted the Earth's axis."¹⁴ It needs to be said clearly: this was never a measurement NASA made of the completed Three Gorges project's actual effect. It was a hypothetical calculation a scientist reached for, off the cuff, to illustrate the scale of mass shifted by an earthquake. What it actually demonstrates is not that "the axis has been dangerously tilted," but that this piece of engineering is now large enough to be used as a comparison point against events on a geological scale.

A distinction is worth making here, so as not to confuse this project with the Sanmenxia hydraulic project discussed in Chapter Two: Sanmenxia was built on the Yellow River and began impounding water in 1960, the first major hydraulic project of the People's Republic. The Three Gorges project discussed in this chapter was built on the Yangtze River, began impounding water in 2003, and was fully completed in 2009 — the largest hydropower and water-control project in the world today. The two are separated by nearly half a century and are not on the same scale, but both raise the same question: when humanity concentrates this much power in a single rigid structure, how should the benefits and the costs be weighed?

The benefits of the Three Gorges project have to be stated honestly first, or this chapter risks becoming a piece of propaganda that says "the ancients were always right, the moderns always wrong" rather than a genuine reflection on civilization. The reservoir's total capacity is 39.3 billion cubic metres, of which 22.15 billion cubic metres is reserved for flood control — and this is not an abstract figure. In 2010 and 2012, the Three Gorges faced inflow floods of 65,000 and 71,200 cubic metres per second respectively, the latter surpassing the historic peak of the 1998 Yangtze flood; regulation by the reservoir raised the flood-defense standard for the Jingjiang section downstream from a "once-in-ten-years" to a "once-in-a-hundred-years" level. On power generation, the Three Gorges power station has an installed capacity of 22.5 million kilowatts and generates roughly 84.7 billion kilowatt-hours a year — more than three times the Grand Coulee Dam and twenty times the Hoover Dam — helping to ease real power shortages in central and eastern China. On navigation, twin five-stage ship locks and a vertical ship lift allow ten-thousand-tonne convoys to pass through a stretch of the Yangtze once notorious for its rapids, with an annual one-way shipping capacity of 50 million tonnes.¹⁵ These benefits are real, and they are not simply cancelled out by geological, ecological, and human costs that generate no electricity and appear in no GDP figures — but neither should those costs be waved away because of them.

The first layer of cost is the relocation of more than 1.2 million people. Data from the United States Geological Survey and the World Bank put the number of people displaced by the Three Gorges project at more than 1.2 million, with some estimates closer to 1.3 million — including the populations of entire county towns that were submerged.¹⁶ This is a cost that is already incurred and cannot be undone; measuring it should mean more than citing a round number, and should include whether these resettled communities' livelihoods, social structures, and cultural memory have received comparable attention since.

The second layer is the geological stability of the reservoir area. After the reservoir filled, the hillsides on both banks were subjected to repeated soaking and erosion as water levels rose and fell, and the incidence of landslides and rockfalls increased accordingly. The United States Geological Survey, citing a 2010 study by the China Earthquake Administration, reports that the reservoir area recorded roughly 3,400 seismic events between the time the reservoir began filling in 2003 and the end of 2009 — a roughly thirtyfold increase in seismic frequency compared with before impoundment — along with numerous landslides.¹⁷ In July 2026, China's Ministry of Natural Resources, Ministry of Water Resources, Ministry of Emergency Management, and the National Commission for Disaster Prevention, Mitigation and Relief jointly issued a directive calling for intensified inspection of geological hazards in the Three Gorges reservoir area, specifically targeting "rockfall and landslide hazards, as well as steep-cut slopes in resettlement zones, bank collapse, and rock-mass degradation."¹⁸ The directive itself makes a point worth noting: geological risk in the reservoir area is not an outsider's speculation, but a matter the authorities themselves have continued to treat as a priority requiring ongoing monitoring and investment. It should be said that this kind of research shows a rising trend in seismic frequency, not evidence that "the dam is about to fail" — reservoir impoundment altering local stress conditions and rock saturation, thereby triggering shallow seismicity, is a phenomenon repeatedly observed at large reservoirs; Sanmenxia and China's Danjiangkou reservoir, among others, have recorded similar induced seismic activity.

The third layer is sediment accumulation and its knock-on effects downstream. According to the Ministry of Water Resources' officially published 2025 China River Sediment Bulletin, the Three Gorges Reservoir has accumulated 2.08 billion tonnes of sediment, amounting to 3.3 percent of its total 39.3-billion-cubic-metre capacity.¹⁹ That proportion is not, in itself, especially high, and it is worth adding that data released by China Three Gorges Corporation show actual average annual sedimentation of about 100 million tonnes — only around 31 percent of what was predicted during the project's feasibility studies — largely because soil-and-water conservation work and other cascade reservoirs upstream have already intercepted a substantial share of the sediment, easing the pressure on the Three Gorges reservoir itself.²⁰ But sediment held back upstream also means the river downstream has lost a source of material it long relied on to maintain its shape: sediment transport measured at the Yichang gauging station downstream has fallen sharply, and the riverbeds of the middle and lower Yangtze, the beds of Lake Dongting and Lake Poyang, and the delta at the river's mouth all now have to find a new equilibrium — showing up in places as riverbed scouring and the shrinking of sandbars.²¹ This needs to be stated with particular care: erosion downstream is the product of multiple overlapping variables, and dozens of other large reservoirs across the Yangtze basin, along with years of ongoing riverbed sand extraction, are also significant contributors — responsibility for changes in the downstream riverbed cannot be assigned to the Three Gorges project alone.

The fourth layer concerns migratory fish and biodiversity across the basin. The dam has blocked traditional migration routes for some species, and the seasonal patterns of water temperature and flow in the reservoir area differ from what they were before impoundment; relevant research has documented changes in the population structure and migratory patterns of some Yangtze fish species.²² But it should equally be said that the pressure on Yangtze biodiversity comes jointly from water pollution, overfishing, shipping traffic, and other dams across the basin — the Three Gorges is one important variable among several, not the only one. The ten-year fishing moratorium imposed across the Yangtze basin in 2021 was precisely the authorities' response to this whole accumulated set of pressures, not a measure aimed at the Three Gorges alone.

If Dujiangyan's story is one that can be told through a series of specific, nameable events across the centuries, the real risk at the Three Gorges shows up, more often, in less dramatic details that simply keep occurring. In July 2026, the Liulan Reservoir in Guangxi failed during heavy rainfall brought by Typhoon Matmo. The reservoir's effective storage capacity was only about 69 million cubic metres — more than three hundred times smaller than the Three Gorges' 22.15-billion-cubic-metre flood-control capacity — but what is really worth remembering is not the failure itself, but the irony behind it: the Liulan Reservoir had just completed a round of standardized reinforcement between 2023 and 2025, held up by officials at the time as a model case of a "problem reservoir" transformed into a "showcase project." Barely a year after that reinforcement was finished, it failed in a single storm.²³ This is a reminder that engineering safety was never a promise redeemed once and for all at the moment of completion, or the moment reinforcement is finished — it is a task of continuous verification and continuous investment that can, at any point, be overturned by what the verification actually finds. The logic holds regardless of a reservoir's size; the larger the reservoir, the greater the consequences if that verification ever fails.

What is really worth asking about the Three Gorges project is not whether it might collapse suddenly one day — that kind of extreme scenario is, across the whole history of dams worldwide, an event of extremely low probability, and a concrete gravity dam failing as a single structure typically requires rare conditions like extreme earthquakes, geological rupture, or deliberate sabotage occurring together. What is really worth asking is that the Three Gorges concentrates the enormous water volume of the upper Yangtze, the geological changes that follow from it, and the cascading ecological effects across the basin, all within a single rigid structure and the several hundred kilometres of reservoir around it — while Dujiangyan wrote change, sediment discharge, overflow, and continuous repair into the everyday operation of the entire system, distributed across it. The problem with modern civilization is not that it has acquired the power to build dams. It is whether, having acquired that power, it remains willing to keep calculating the costs that generate no electricity and appear in no GDP figures — the costs borne by the river itself, by resettled communities, by downstream ecosystems, and by generations not yet born — and willing to keep revising those calculations against verifiable lessons like Liulan's.

Whether a civilization has truly matured is not decided by whether it can build the largest dam in the world, but by whether, before building it — and in every year after — it can still hear the river, the hillside, the displaced, and the generations not yet born.

· · ·

VIII. Two Thousand Years Later, the West Makes Room for Rivers Again

To tell Dujiangyan's significance simply as "the ancient Chinese already knew this, and the modern West has completely forgotten it" might make for a satisfying story, but it would not be a true one.

Modern engineering has not moved in a straight line of ever-higher dams and ever-thicker levees. Facing rivers and climates that are growing harder to predict, engineers in many countries are exploring how to combine solid structures with natural processes once again.

The Netherlands' "Room for the River" program is a substantial example. This low-lying country had long relied on levees to hold back floodwater, but after the Rhine and Meuse rivers came dangerously close to overtopping their banks twice in the 1990s, the Dutch government came to realize that endlessly raising levees was drawing it into a contest it could not win indefinitely. Those two floods — in 1993 and 1995 — forced the emergency evacuation of more than 250,000 people and over a million head of livestock.⁹ Starting in 2007, a series of projects moved beyond simply raising levees further, instead setting levees back from the river, opening new side channels, and lowering floodplains, to give high water during flood season room to spread out and be absorbed again. By 2019, the program had completed more than thirty individual projects at a total cost of roughly 2.3 billion euros, reshaping the flood-management landscape along the Rhine, Meuse, Waal, and IJssel rivers.⁹

Willem Jan Goossen, an official at the Dutch Ministry of Infrastructure and Water Management, later described the thinking behind this shift in an interview with the European Environment Agency: rather than continuing down the path of raising and reinforcing levees indefinitely, it made more sense to act on a more fundamental insight — widening the river channel itself could actually lower overall water levels during floods, breaking the cycle of building ever higher while never quite keeping pace with the water.⁹ The logic runs parallel to Dujiangyan's choice, more than two thousand years earlier, of diversion over containment.

Among these thirty-odd projects, one representative type is what might be called "levee setback": rather than building the levee higher and thicker in its original position, engineering teams moved the entire levee line back toward the interior, returning land once shut out by the levee to the floodplain, so that the river has a wider channel to spread into during flood season and can retreat from that temporarily reclaimed space once water levels drop. These projects often involve complicated land acquisition and resident-relocation negotiations, no less demanding than building a new dam — a reminder that "making room for the river" is a principle in theory, but in practice just as much an ongoing project of sustained investment and negotiation as "building a higher levee" ever was.

On the other side of the Atlantic, the U.S. Army Corps of Engineers' "Engineering With Nature" initiative, underway since 2010, similarly emphasizes achieving engineering goals by working with natural processes, while giving as much weight as possible to environmental and social benefits. The initiative has since grown into an international network of practice: in 2021, the Corps partnered with the Dutch Ministry of Infrastructure and Water Management, the UK Environment Agency, and other bodies to jointly publish International Guidelines on Natural and Nature-Based Features for Flood Risk Management, an attempt to organize this body of experience into a methodology that can be applied across different countries and river conditions.¹⁰ Wetlands, dunes, and floodplains are no longer seen purely as scenery to be protected outside the bounds of engineering — they are becoming part of the flood-control and coastal-protection system itself.

These contemporary practices are not identical to Dujiangyan from more than two thousand years ago — the historical conditions, the scale of engineering, and the underlying science cannot simply be equated. But they answer to the same underlying question: faced with a river that keeps changing, must humanity keep raising its boundaries indefinitely, or can it make room for the water again while still keeping people safe?

That echo is itself meaningful. It shows that Dujiangyan's value has never been about declaring one civilization superior to another — it lies in letting two bodies of experience, separated by more than two thousand years and thousands of miles, illuminate each other. To a Western reader, Dujiangyan stops being simply a distant Chinese relic and becomes an ancient footnote to a question the whole modern world now shares.

IX. "Unity of Heaven and Humanity": Not a Formula, But a Sense of Limits

Why did Dujiangyan emerge from the soil of Chinese tradition in particular?

We cannot explain every specific engineering detail with the phrase "unity of heaven and humanity" alone, nor equate ancient philosophy directly with modern systems science — scientific theory requires definition, measurement, experiment, and verification, while traditional thought carries something different: a sense of value and a way of looking at the world. The two should not be conflated.

But Chinese tradition has long reminded people that human beings do not stand apart, in isolation, from the rest of the natural world. The Dao De Jing holds that "humanity follows earth, earth follows heaven, heaven follows the Way, and the Way follows what is natural," urging respect for the intrinsic order of things; it also praises water as the highest good, for nourishing all things without contending with them. What Daoism calls wu wei — often translated as "non-action" — was never a call to do nothing at all, but something closer to restraint: not forcing one's own will through, not letting excessive intervention destroy the order things already have.

The Book of Changes offers a related, often-quoted line: "when a thing reaches its limit, it changes; once it changes, it finds a way through; and having found a way through, it endures." Originally about the transformation of hexagrams, the line has long been borrowed to express a broader wisdom — that for something to endure, it must be capable of adapting when it meets difficulty, rather than clinging to its original form. Dujiangyan's history over more than two thousand years is, in a sense, a concrete footnote to that line: from bamboo cages to cut stone, from stone to iron-and-bamboo combined, from manual observation to digital monitoring, each change existed so that the underlying logic of diversion, sediment discharge, and flow control could keep finding its way through, and keep enduring.

Dujiangyan is, of course, not an engineering project that sprang automatically from a philosophical maxim. It grew out of generations of real observation of hydrology, terrain, and sediment, and out of repeated labor and failure — Jidangpu's list of thirty-two critical trouble spots is itself a product of that kind of observation. But traditional culture did provide an important backdrop of values for this practical work: when wielding power, people should ask not only "can I do this," but also "do I understand the underlying pattern," "how far should this go," and "does this leave room for other forms of life, and for changes yet to come."

This is what a sense of limits looks like. Even this shows up in the most basic decision of all — where to build. Li Bing did not take the easiest path and dam the river at the straightest point where it emerges from the mountains; he patiently studied the bends, currents, and terrain of the whole stretch before finally settling on the site for Fish Mouth. Behind that decision was less some mystical philosophical intuition than an attitude: better to spend years understanding a river than to rush to force it into a different shape. This patience — understand first, act second — may be the plainest, most verifiable form that "unity of heaven and humanity" ever takes in actual engineering practice.

This is a sense of limits. The real difficulty was never simply obeying limits forever, nor simply breaking them forever, but learning to tell which restrictions should be overcome and which patterns of nature must still be respected.

One further point is worth adding here, so as not to make Chinese tradition sound like a single, monolithic voice. Alongside Daoist wu wei, an almost contemporaneous and quite different voice existed. The late-Warring-States Confucian philosopher Xunzi raised, in his essay "Discourse on Heaven," a pointed counter-question: rather than passively submitting to and praising the mandate of heaven, why not "master heaven's mandate and put it to use" — rather than waiting passively for nature's bounty, take active command of its patterns and use them. This more assertive strand of thought, emphasizing human agency, is equally part of Chinese tradition, and it left a deep mark on later practices of water control and land reclamation.

Placing these two voices side by side makes a point of its own: Chinese tradition never offered a single answer to the question of how humanity should relate to nature. It has long held an ongoing tension between "working with" and "taking command of" the natural world. What is valuable about Dujiangyan may be exactly that it never leaned fully into either extreme, finding instead a balance — between using the river's patterns and respecting them — that has kept it running for more than two thousand years. Traditional wisdom cannot replace modern science, but it can keep asking questions that science alone cannot answer: once we hold a form of power, how should we use it? How far should we take it? And what room should we leave, in advance, for the mistakes we might make?

X. Ancient Layout, Modern Technology: How They Work Together

Dujiangyan today presents a striking combination.

Upstream stands the modern Zipingpu hydropower complex; the irrigation district is threaded with automated sluice gates, water-condition sensors, and a digital-twin dispatch system. The digital twin running at the command center integrates two- and three-dimensional geographic data and oblique photogrammetry covering 27.6 square kilometres around the headworks, building-information models for 23 sluice gates, and continuously updated hydrological and structural monitoring data; the cloud computing centre behind it runs at 236 trillion calculations per second with up to 709 terabytes of storage.³ These figures form a striking counterpoint: more than two thousand years ago, Li Bing left behind six words — "dredge the channel deep, keep the weir low." Two thousand years later, his successors have left behind a system that can simulate flood peaks in real time and forecast risk hours in advance. The form of expression is entirely different, but the underlying question is the same one: how to see what is coming as early as possible, and respond as calmly as possible.

On 24 July 2024, Dujiangyan faced its largest flood peak in nearly a decade. This time, the flood-response team did not wait for the water to arrive before scrambling to react — running flood simulations in advance through the digital-twin system, they had reasonably precise flow data in hand before the peak even arrived, and had already prepared a response plan accordingly.³ This scene may be the clearest illustration of how ancient layout and modern technology work together: Bottle-Neck Channel and Flying Sand Weir, physical structures that took their final form more than two thousand years ago, still carry out the actual work of diverting, limiting, and discharging sediment the moment the flood peak arrives — while, before those structures ever meet the flood, a modern early-warning system built on real-time data and simulation has already bought humanity time to prepare and respond calmly. This is a concrete example of rigid infrastructure and adaptive decision-making working together.

At the same time, the headworks layout that took shape gradually across the centuries — Fish Mouth, Flying Sand Weir, Bottle-Neck Channel — still genuinely carries out the entire process of diversion, sediment discharge, flood release, and water intake. Rigid modern engineering capability and adaptive ancient design are not in competition; they operate together within the same hydraulic system.

This may be the most practically relevant lesson Dujiangyan has to offer modern civilization: we do not have to choose between an all-or-nothing extreme of "conquering nature" or "abandoning technology." Where a problem can be measured precisely and its boundaries are clear, we need precise control; where we face a complex system with many variables, constantly evolving, we also need feedback, buffering, dispersal, maintenance, and continuous revision — whether that system is a river, or a rapidly deployed new technology such as artificial intelligence today: the more powerful the capability, the more it needs transparent rules, ongoing human oversight, and a willingness to exercise restraint while the consequences are still unclear.¹¹

Dujiangyan, of course, offers no ready-made answer to every modern problem. Packaging a two-thousand-year-old irrigation system as a universal formula for solving all of today's difficulties would itself be a kind of arrogance. But it does offer a reminder tested across an extraordinarily long stretch of history: whether a system can endure has never depended solely on how powerful it was on the day it was built, but also on whether it can sense change, accommodate correction, and persuade generation after generation to keep taking up the work of maintaining what it inherited.

In writing this chapter, we have tried, consciously, to practice the same discipline. Li Bing, the stone rhinoceros, Jidangpu's iron tortoise, Lu Yi's statue, the rebuilding of the Erwang Temple, the lesson of Sanmenxia, the specific logic behind the Netherlands' levee setbacks — every concrete fact has, wherever possible, been traced back to a source that can be checked: original texts, archaeological reports, or official records, rather than left in the vague language of "it is said" or "legend has it." Every inference and philosophical extension has, wherever possible, been marked as such — our own thinking drawn from the historical record, not the record itself. This is not simply a matter of writing standards. It reflects a value this book is trying to carry: respect for tradition should not rest on exaggeration and myth-making, but on patient discernment and honest attribution. This same method will guide every chapter that follows in this ongoing series of reflections on civilization.

· · ·

XI. Two Western Gazes, a Century Apart

Between 1917 and 1919, a young American recently graduated from Princeton, Sidney Gamble, travelled up the Yangtze and into Sichuan, photographing Guanxian — today's Dujiangyan — as it then appeared: Fish Mouth, Flying Sand Weir, the boatmen and farmers going about their day, all captured through his lens. These photographs are held today in the Sidney D. Gamble Photographs collection at Duke University's library, among the earliest surviving Western photographic records of Dujiangyan.¹²

A hundred and eight years later, another Western visitor stood in roughly the same place. On 5 December 2025, French President Emmanuel Macron, accompanied by his wife Brigitte, made a special visit to Dujiangyan at the end of a state visit to Beijing. Chinese President Xi Jinping met him there, and the two men walked and talked along the Dam of Merit walkway before sitting down for tea at the Pavilion of Nostalgia beside the water. Xi introduced Macron to Dujiangyan's history, describing it as "the only ancient hydraulic project in the world still in use today" and "one of the earliest successful examples of harmonious coexistence between humanity and nature." Macron, for his part, called Dujiangyan "beautiful, with clear water and green hills," remarking that a project built more than two thousand years ago was still functioning today, and that "the industriousness and wisdom of the Chinese people are truly admirable."¹³

Placing these two Western gazes side by side, separated by more than a century, reveals something worth noticing: what Gamble's camera captured was a local piece of hydraulic infrastructure and the daily life around it; a hundred years on, that same place has become somewhere two heads of state can sit together beside the water and discuss history. What has changed is its position on the world stage. What has not changed is that it is still doing the same work it began doing more than two thousand years ago — dividing the water, discharging the sediment, limiting the flow, and nourishing the plain beneath it.

A high-level diplomatic encounter, of course, carries its own particular political context, and should not be over-read as some final, authoritative verdict. But setting that aside, and asking the plainer question — whether Dujiangyan is worth taking seriously on its own terms — the answer has been consistent for a hundred years. From a young scholar with a camera to the heads of state and researchers who study and discuss it today, the people willing to travel there, look closely, and think about what they see have never really stopped coming. This century of sustained attention is, in its own way, an answer to the question this chapter opened with: the path the builders chose more than two thousand years ago is still, today, a path the world finds worth stopping to look at again.

· · ·

Coda: A Different Kind of Power

For more than two thousand years, the Min River has never stopped changing, and Dujiangyan has never asked the world to keep it exactly as it was the day it was finished.

From the stone rhinoceros meant to subdue river spirits at the end of the Warring States period, to the stone statue of Li Bing carved by Han-dynasty craftsmen; from the sixteen-thousand-jin iron tortoise Jidangpu cast, to the statue of Lu Yi standing in the Erwang Temple; from a temple rebuilt, again and again, on its original site after war and earthquake, to the Guanxian Gamble photographed a century ago, to the real-time water data now scrolling across a screen — across more than two thousand years, nearly every shock Dujiangyan has faced, every moment it was seen anew, left behind something concrete: a carved stone, an artifact, a photograph, a reconstruction, a set of numbers. Taken individually, these are just traces left by some dynasty, some person. Taken together, they trace out a single pattern: generation after generation of people, willing to look again after every shock, willing to correct what needed correcting, and willing to hand what they learned to a successor whose name they would never know.

This is also why this book has never told Dujiangyan's story as the story of one piece of engineering alone, but as the story of so many specific people: Li Bing, who established the basic layout yet remains a figure we can barely make out clearly; Jidangpu, who surveyed thirty-two critical trouble spots; Lu Yi, who argued for iron and bamboo together; Zhang Tangying, who funded a weir out of his own pocket; and the countless nameless laborers who dug silt out of the riverbed with their own hands every winter and spring. The engineering ages; the materials get replaced; even the specific structures the name "Dujiangyan" refers to are no longer quite what they were more than two thousand years ago. What has not changed is the posture of this succession of people — willing to take on the imperfections their predecessors left behind, and willing to leave room for their successors to keep correcting what they, in turn, leave behind.

It draws on the river's own power, and leaves room for the river in return; it establishes clear engineering boundaries, and keeps accepting the maintenance of one generation after another. This kind of power does not draw attention to itself the way a high wall does. It shows up, instead, as understanding rather than display; as restraint rather than retreat; as generation after generation, quietly repairing and adjusting, and handing that responsibility on, with due care, to the next.

As the water of the Min River passes through Bottle-Neck Channel and rushes on toward the Chengdu Plain, it seems to keep reminding everyone who comes after: perhaps humanity's highest wisdom was never simply the power to change the world, but knowing, once that power is in hand, how to go on living with it.

· · ·

Notes

1. On Dujiangyan's founding date, the basic layout of the headworks, its continued operation, and the damage and repair following the 2008 Wenchuan earthquake: UNESCO, "Mount Qingcheng and the Dujiangyan Irrigation System," World Heritage List No. 1001, whc.unesco.org/en/list/1001 (accessed August 2026). On the traditional "drought in the east, flooding in the west" description, see materials from the Sichuan Provincial Local Chronicles Compilation Committee, scdfz.sc.gov.cn.

2. On the historical record and accumulated legend concerning Li Bing's life: Sima Qian, Records of the Grand Historian, "Treatise on Rivers and Canals"; Ban Gu, History of the Han, "Treatise on Canals and Ditches"; Chang Qu, Chronicles of Huayang (Chengdu Shidai Press, 2007 annotated edition); the 1974 excavation report on the Eastern Han stone statue of Li Bing found at Dujiangyan, Wenwu (Cultural Relics), 1974, no. 7; see also "An Analysis of the Deification of Li Bing's Water Control," Chinese Culture Forum, 2018, no. 2, pp. 40–43.

3. On the modern Dujiangyan irrigation district's coverage (as of 2024, 11.55 million mu across 8 cities and 41 counties, serving more than 28 million people), the Zipingpu hydropower complex, and the technical specifications of the digital-twin dispatch system (27.6 sq km data model, BIM models for 23 sluice gates, 236 trillion calculations per second, 709 TB storage): Wang Yajing, "The Hydraulic Power Behind the Nation's Granary," China Discipline Inspection and Supervision News, 13 January 2025, reprinted at tjjw.gov.cn (accessed August 2026).

4. On sediment accumulation at the Sanmenxia hydraulic project and its subsequent reconstructions: Yellow River Conservancy Commission, "The Merits, Failings, and Future of the Sanmenxia Reservoir," yrcc.gov.cn; Pan Jiazheng, "The Story of the Sanmenxia Project," from Praise and Blame Across a Thousand Autumns: The Story of Dams, reprinted by the Shandong Provincial Department of Water Resources, wr.shandong.gov.cn; "Sanmenxia Hydropower Station," Baidu Baike (accessed August 2026).

5. For a non-technical account of the distinction between "strength-first" and "resilience-first" approaches to engineering safety, see discussions of resilience engineering and the "fail-safe"/"safe-to-fail" concepts in landscape ecology. This book offers only a general summary and does not equate Dujiangyan or Sanmenxia directly with a complex adaptive system in the strict scientific sense.

6. On the system of annual maintenance; Jidangpu's appointment as surveillance commissioner for Sichuan in 1334 and his major reconstruction beginning in November 1335, including labor and material figures; the origin of the "dredge deep, keep the weir low" rule; and the Song-dynasty "sacrifice of ten thousand sheep" at the Chongde Temple (recorded in Fan Chengda's Boat Journal to Sichuan and Shi Jie's account of Yongkang Commandery): Peng Bangben, "The System of Annual Maintenance at Ancient Dujiangyan: Starting from the Stele of Li Bing, Governor of Shu Under Qin," Journal of Xihua University (Philosophy and Social Sciences), 2018, vol. 37, no. 4, pp. 8–18, doi:10.19642/j.issn.1672-8505.2018.04.002.

7. On the mechanics of the bend's rotational flow and the modern hydraulic explanation of the traditional "sixty-forty" diversion ratio, see review studies of Dujiangyan's hydraulic mechanisms from the China Institute of Water Resources and Hydropower Research and related journals; this ratio is not precisely constant at every water level. On the present dimensions of Bottle-Neck Channel (approximately 20 m wide, 40 m high, 80 m long), see public engineering-heritage materials at rongwp.com. On the 2008 cracking of the Fish Mouth spur and its structural stability through the 2013 Lushan earthquake and subsequent flood: "Dujiangyan Fish Mouth," Baidu Baike (accessed August 2026).

8. On Lu Yi's life, his revision of Jidangpu's iron-and-stone structure during the Zhengde era, and the "iron and bamboo together" approach: "Dujiangyan," Chinese Wikipedia (updated June 2026); Sichuan Provincial People's Government website, "The Four Faces of Dujiangyan," sc.gov.cn, September 2022; "Dujiangyan (3): A Gallery of Water-Control Heroes," reprinted by Tencent News, December 2024 and July 2025.

9. On the Netherlands' "Room for the River" program: public materials from the Dutch Ministry of Infrastructure and Water Management (Rijkswaterstaat) and the European Environment Agency, eea.europa.eu/signals-archived. On project scope, evacuation figures, and total investment, see also the Central Dredging Association (CEDA) conference paper "Dutch Water Program Room for the River," dredging.org (accessed August 2026).

10. On the U.S. Army Corps of Engineers' "Engineering With Nature" initiative and the 2021 international guidelines: U.S. Army Corps of Engineers, Engineering Research and Development Center, ewn.erdc.dren.mil; Leaders in Energy, "Engineering With Nature," leadersinenergy.org (accessed August 2026).

11. This paragraph's analogy to AI governance is offered as philosophical extrapolation, not a claim that Dujiangyan's engineering principles apply directly to technology governance; it is included for the reader's consideration only.

12. On Sidney D. Gamble's 1917–1919 photographs of Chengdu, Guanxian (today's Dujiangyan), and surrounding areas: "Sidney D. Gamble Photographs," David M. Rubenstein Rare Book & Manuscript Library, Duke University, sites.duke.edu/sidneygamble (accessed August 2026).

13. On the 5 December 2025 meeting between Xi Jinping and French President Emmanuel Macron at Dujiangyan: "Xi Jinping Holds Friendly Exchange with French President Macron in Chengdu," Xinhua/Ministry of Foreign Affairs of the People's Republic of China, mfa.gov.cn, 5 December 2025; on Macron's visit to Dujiangyan and his remarks, see also same-day coverage from Xinhuanet, People's Daily Online, and Sichuan Observer (accessed August 2026).

14. On the original context of the "Three Gorges Dam has tilted the Earth's axis" claim: the figure originates from a 2005 explanation by NASA Jet Propulsion Laboratory scientists Richard Gross and Benjamin Fong Chao of the 2004 Sumatra earthquake's effect on Earth's rotation, using "if the Three Gorges Reservoir were filled with 40 cubic kilometres of water" purely as a comparison for the scale of mass involved — not as a measured assessment of the completed Three Gorges project: "NASA Details Earthquake Effects on the Earth," NASA Jet Propulsion Laboratory, jpl.nasa.gov/news/nasa-details-earthquake-effects-on-the-earth (accessed August 2026).

15. On the Three Gorges project's basic technical parameters (total capacity 39.3 billion cubic metres, flood-control capacity 22.15 billion cubic metres, installed capacity 22.5 million kilowatts, annual generation 84.7 billion kilowatt-hours, annual one-way shipping capacity 50 million tonnes) and the 2010 and 2012 flood-regulation records: "Three Gorges Reservoir," Baidu Baike; public data from China Three Gorges Corporation, cited in "Thirty Years of the Three Gorges Project: The 'Mainstay' of the Yangtze Flood-Control System," Gansu Daily, china.gansudaily.com.cn; interview with Chen Guiya, deputy director of the Changjiang Water Resources Commission's Flood Control and Drought Relief Office, reprinted by the Guangdong Society of Hydropower and New Energy Engineering, gdshe.org (accessed August 2026).

16. On the scale of resettlement caused by the Three Gorges project (more than 1.2 million people, with some estimates around 1.3 million): U.S. Geological Survey, "Three Gorges Dam Brings Power, Concerns to Central China," usgs.gov; World Bank, "Social Assessment Systems in Infrastructure Development in China," thedocs.worldbank.org (accessed August 2026).

17. On the roughly thirtyfold increase in seismic activity in the Three Gorges reservoir area between 2003 and 2009, and accompanying landslides: China Earthquake Administration, 2010 study, cited in U.S. Geological Survey, "Three Gorges Dam Brings Power, Concerns to Central China," usgs.gov (accessed August 2026).

18. On the July 2026 joint directive from the Ministry of Natural Resources, Ministry of Water Resources, Ministry of Emergency Management, and the National Commission for Disaster Prevention, Mitigation and Relief calling for intensified inspection of geological hazards in the Three Gorges reservoir area: "Four Departments Issue Joint Notice on Geological Hazard Risk Inspection," China News Service, chinanews.com.cn, 18 July 2026; also reposted by the Ministry of Emergency Management, mem.gov.cn (accessed August 2026).

19. On cumulative sediment accumulation of 2.08 billion tonnes at the Three Gorges Reservoir, equal to 3.3 percent of total capacity: Ministry of Water Resources, 2025 China River Sediment Bulletin, reported simultaneously by Xinhua, China News Service, Science and Technology Daily, and China People's Political Consultative Conference Daily, among others, on 27 May 2026 (accessed August 2026).

20. On average annual sedimentation of about 100 million tonnes at the Three Gorges Reservoir, roughly 31 percent of the feasibility-study projection: public data from China Three Gorges Corporation, cited in "Thirty Years of the Three Gorges Project: The 'Mainstay' of the Yangtze Flood-Control System," Gansu Daily, china.gansudaily.com.cn (accessed August 2026).

21. On the Three Gorges project's effect on downstream sediment transport and the erosion equilibrium of riverbeds and lakes: see review literature in peer-reviewed hydrology and fluvial geomorphology journals (accessed August 2026); this book treats the specific mechanisms with caution and emphasizes the role of multiple contributing variables.

22. On the Three Gorges project's effect on Yangtze fish migration and basin biodiversity, and the 2021 ten-year Yangtze fishing moratorium: see relevant peer-reviewed ecological research (accessed August 2026); on the fishing moratorium, see also public information from the Ministry of Agriculture and Rural Affairs.

23. On the July 2026 failure of the Liulan Reservoir in Guangxi and its 2023–2025 standardized reinforcement project beforehand: Sina News, China News Service, and China Digital Times' reprint of related Southern Weekly reporting (accessed August 2026).