Footprint of the Capitals, Part 1
Thanks to some very productive periods at Max Planck in Berlin and at Yale, I’ve had a chance to do some new work. This is building into a chapter where I consider the shifts in wood requisitions in the Ming and Qing. The pocket version of the argument is this: the Ming and Qing states gradually stepped back from territorial oversight of woodland (among other things) and increasingly relied on indirect oversight and licensed intermediaries, especially through the market.
We can split this process into stages. The origins of de-territorialization are visible starting in the 1430s with the Hongxi (1424–5) and Xuande (1425–36) Emperors’ rollbacks of Yongle projects. Ad-hoc adaptation to the market and the declining quality of cadastral data dominate the late 1400s and early 1500s. More systematic solutions emerged most rapidly between about 1530 and 1600, especially the centralization of cameral data and expertise. The 1600s are a mess of different policies in response to crisis, including the Wanli Emperor’s backslide into reliance on eunuchs from about 1600–1620, and temporary wartime policies from the 1630s to 1680s. Then starting around 1680, and into the 1700s, the Qing state basically follows the late Ming model of reliance on intermediaries and centralization of finances.
I have previously characterized this movement away from direct territorial governance as a “retreat of the state,” but my new research casts it in a slightly different light. Officials did have increasing difficulty overseeing landholdings and labor service directly starting in the 1430s, culminating in the Single Whip Reforms to labor service in the late 1500s, and the fixing of the land tax quotas in the early 1700s. Yet what is missing from this picture is the development of much greater degree of expertise at specialized state bureaus (esp. in the mid-late Ming) and formalization of systems of intermediaries (esp. in the mid-Qing). The movement away from direct oversight was not necessarily a sign of state weakness; instead it was a sign of adaptation to new institutions that could meet state needs at less cost or less difficulty.
The most spectacular example of this is the shift in requisitions for imperial timber (皇木) — the massive trunks used for pillars and ridgepoles (zhuliang 柱梁) in imperial palaces. These poles formed the weight-bearing structure of the buildings, and therefore had to be made of individual, unbroken, perfect trunks of straight-growign, strong wood, principally shanmu 杉木 (Cunninghamia lanceolata) and nanmu 楠木 (Phoebe zhennan).
The shift in imperial timber supply nicely demonstrates the pattern suggested above: the Yongle palaces were built almost entirely of timber extracted by corvee teams overseen by officials; by the 1580s, a significant proportion of timber for palace building was bought from licensed merchants; in the 1680s the Qing struggled to obtain timber through its own territorial projects; and after 1750, all imperial timber was purchased through intermediaries.
When Zhu Yuanzhang, the Hongwu Emperor, completed his victory in the Red Turban Wars in the late 1360s he established his capital at Nanjing. This was a defensible site and had been the capital of several southern dynasties, but it had not served as a capital in four-hundred years. Shortly following the Hongwu Emperor’s death a civil war erupted between the chosen successor, grandson Zhu Yunwen, the Jianwen Empror (r. 1398–1402), and a younger son, Zhu Di who deposed his nephew and ruled as the Yongle Emperor (r. 1402–1424). To cement his power and authority, the Yongle Emperor ordered construction of another capital at Beijing his seat of power in the north and the former site of the Yuan capital Dadu. Once the court moved fully to Beijing in 1421 Nanjing continued to function as a secondary capital.
Both cities were among the largest in the world at the time, and huge consumers of timber and fuel. Despite previous edifices on these sites, their designation as capital cities entailed enormous construction between 1358 and 1421, requiring vast amounts of timber. Much of the construction was palaces and temples, requiring poles of the greatest possible size and quality. These timbers could only be found in the old-growth forests of western China.
Ming records give relatively little detail on the source of timbers for constructing Nanjing, we know comparatively more about Beijing. The Northern Capital was essentially built anew after the Yongle Emperor moved the capital to North China in 1403 following his successful usurpation of his nephew. To construct his palace and temple complexes required enormous numbers of the most massive timbers available. In 1406, officials were sent to several provinces in the south to find the largest and most beautiful tree specimens. Various regions supplied timber for these projects but the main timbers for the weight-supporting pillars and beams (zhuliang) were shipped from the three provinces of Sichuan, Huguang, and Guizhou in the Southwest.
The historical southwestern region of Shu 蜀— according roughly with Sichuan — was long famous as a source of good timber, ostensibly the site of imperial logging in the Han and Tang dynasties. Sichuan nanmu had been used in the late fourteenth century construction of the palace of the Prince of Shu, another son of the Hongwu Emperor. In 1406, when officials were sent to survey areas in the South and West for useful timber, reports from Sichuan were the most promising. The suitability of of Sichuan timber was apparently confirmed when several large trees fell into the river and floated downstream of their own accord. The emperor considered this a sign from the spirits and named this site near Linzhou 藺州, “Sacred Tree Mountain” (shenmu shan). Whether as a continuation of historical precedent or through divine intervention, southern Sichuan became the focus of the most intensive imperial lumbering under Yongle. In the course of building Beijing, Board of Works Minister Song Li 宋禮 visited Sichuan five times to oversee the lumbering; Inspecting Censor Gu Zuo 顧佐 also visited personally; and the eunuch Xie An 謝安 spent twenty years on-site to supervise the project.
The Yongle-era construction projects were only the beginning of a long period of lumbering in the West. Sparse fifteenth century records note that after intensive cutting during the construction of Beijing lumbering became irregular in the late 1420s and early 1430s and was halted entirely under the Hongzhi Emperor (1488–1505). Yet imperial lumbering recommenced in Sichuan in 1511 and functioned on and off until 1606. Beijing started major construction or repair projects in 1540, 1557, and 1584; each occasioned the dispatch of officials specifically to oversee the lumberyards of Western China. Yet while imperial lumbering in Sichuan was effectively semi-permanent for much of the sixteenth century, like many regional posts in the Ming it remained largely ad-hoc and did not have a clearly-designated place in the bureaucratic hierarchy. Instead the post of Lumbering Supervisor (du mu) was filled by a rotation of high-level officials, often including a Circuit Inspector (xunfu), Capitol Censor (du yushi), or a Vice Minister of the Board of Works (gongbu shilang), as well as a number of lower-level functionaries placed under their supervision.
The extraction of such enormous timber from a sparsely-populated and mountainous frontier was incredibly difficult. We know relatively little of the fifteenth-century requisitions, but better-kept records from the sixteenth and seventeenth centuries document the enormity of the task. During the palace reconstruction of the 1540s, Censor Li Xianqing supervised at least twenty-two mid- and low-level officials sent to survey and oversee more than forty-five logging sites in Sichuan, Huguang, and Guizhou. He reported that the giant timbers needed for palace construction could only be found deep in the mountains, and expressed his doubts that sufficient amounts of palace-building timber had ever been floated out of Sichuan, even in the Yongle Reign.
Problems of transportation were compounded by the difficult terrain where several-century-old trees still stood. As many as five-hundred laborers were needed to tow each of the largest logs over mountain passes.
Dozens of specialized metal-, wood-, and stone-workers were needed on site make tools and cables and build slip-roads. Workers had to build “flying bridges” (fei qiao) to transport the logs across thousand-foot defiles.
Once the logs were dragged to the waterways they were tied into rafts of six-hundred and four poles, each of which was towed by a team of forty men.
Twenty or thirty such log rafts were launched together for the three year, ten-thousand li (approximately three-thousand kilometer) journey to Beijing. This passage was hardly easier than the land route they had passed thus far: individual logs had to traverse narrow gorges and rapids to get to the Yangzi River; fastened into rafts, they then passed through another area of rapids in the Three Gorges below Chongqing, and up the busy Grand Canal to Beijing. Nor were all harvested trees suitable; perhaps eighty percent were discarded because they were hollow, and another ninety percent of trees were lost to accidents along the way.
In addition to hard labor, lumberjacks in the mountainous Western frontier faced dozens of environmental hazards. Li Xianqing and another mid-Ming lumber official, Hui Gong listed snakes, tiger, and “barbarian” attacks; “miasmatic vapors” (yanzhang, probably malaria); storms, forest fires, rockslides, and raging rivers among the many dangers of the region.
Working in a sparsely populated mountain area also meant that labor teams had to carry their own food and often starved when projects went over time. Hui Gong summed up the difficulties with this shorthand: “the labor force numbers in the thousands; the days number in the hundreds; the supply costs number in the tens of thousands each year.” According to another Sichuan saying “a thousand enter the mountains, five hundred leave the mountains” (rushan yiqian chu shan wubai).
In the sixteenth century there is evidence of growing reliance on timber merchants to supplement the poles cut directly in the western frontier. Li Xianqing writes that officials of the 1540s oversaw lumbering itself in Sichuan and parts of western Hunan (du [place name] zhi mu), but oversaw the purchase of timber in southern Hunan and Hubei (gou mu yu [place name]). Reliance on merchants increased in the mid-to-late 1500s due to cost overruns.
According to a long memorial by two Guizhou officials, Shu Yinglong 舒應龍 and Mao Zai 毛在, in 1557 Guizhou Province was responsible for 4709 poles of shanmu and nanmu at a cost of 720 thousand taels of silver; this requisition came in a year when the provincial treasury held just under 15 thousand taels, or about 2% of what was required. Additional funds had to be disbursed from other provincial treasuries: 100 thousand from Guangdong, 140 thousand from Yunnan, and 90 thousand from Jiangxi. In 1584, a much smaller order for 1132 poles again faced the prospect of cost overruns: the treasury only had 20 thousand, one sixth of the estimated 120 thousand taels.
Citing the recurrent nature of lumbering expenses (caimu gongfei xun zhi xing) Shu and Mao argued that it was impractical to resort repeatedly to temporary solutions. They suggested summoning merchants from various regions to quote market prices for standardized sizes and grades of timber, a practice that was by then standard in the shipbuilding administration (see below). Because prices were best in Guizhou but Guizhou had little local tax base Shu and Mao argued that funds from other provinces should continue to be directed there to purchase timber on the market, and that officials be stationed there to oversee the merchants and loggers.
In the sixteenth century wood markets developed in much of the old south (Fujian, Huizhou, Jiangxi) and began to spread to the southwestern frontier (Hunan and parts of Guizhou), encompassing a diversified selection of forests. The state took advantage of this market to obtain a regular supply of timber from the best and least expensive regions for the rest of the Ming and Qing. Yet despite the growing reliance on merchant intermediaries to supply timber, imperial lumbering persisted in the old Sichuan frontier. Because forests in the gorges were still relatively abundant, massive timber could be obtained there that was unavailable anywhere else, but only imperial projects could overcome the combination of difficult terrain and labor shortages.
Between the fall of the Ming in 1644 and its pacification in 1664, large parts of Sichuan were ravaged by warfare and massacres under a breakaway state called the Great Western Dynasty (da xi wangchao), and then by the invading Manchu armies.
Yet in 1667, almost immediately after order was restored by the new Qing state, the court requisitioned timber from the Ming logging regions in Sichuan. Officials reported that many areas still had shanmu and nanmu timber of sufficient size. But in short order these same officials reported the same problems that had plagued the Ming projects: the complex logistics of supplying large work teams in sparsely-populated areas; the difficulty of towing the logs over steep mountain passes; and the additional expense of dredging silted rivers and damming rapids to transport log rafts down the watercourses.
Officials specifically criticized the late Ming practice of licensing merchants to acquire the lumber needed for imperial construction. Yet in 1669 the nanmu supplied from the Western timberland was considered insufficient for palace construction; pine logs were substituted for many uses, and the southwestern logging operations were halted.
Despite the apparent failure of the 1660s logging operations the Qing court repeatedly requisitioned nanmu from the Southwest whenever large timber was needed for palace repairs.
Another logging operation was ordered in 1683 and new surveys were undertaken, yet this requisition was stymied by the same difficulties that had faced previous logging operations. In 1685, the court requisitioned the majority of timber by purchasing it in southern provinces with better water access; only the largest nanmu poles were shipped out of Sichuan. In 1686 requisitions were cancelled entirely.
In 1726 another round of palace construction commenced and requisitions were made from Sichuan; within a year this expedition reported timber shortages and shipping difficulties: only one-hundred and forty-four poles of nanmu were close enough to the rivers to be shipped.
A final requisition for southwestern timber was made in 1750, but a mere thirty poles of large nanmu were obtained. As in the Ming, the Qing court ordered for timber to be purchased at market rates; this became the main source of imperial timber in the 1680s, and the exclusive source after 1750.
Despite their repeated failures to provide sufficient quantities of timber, the surveys and lumbering operations of the 1540s, 1580s, 1660s, 1680s and 1720s were substantial projects that demonstrated the capacity of the Ming and Qing states. Dozens of officials were dispatched to distant frontiers to oversee large labor teams. They noted in official registers (ce) the size and grade of any shanmu or nanmu poles and the distance between the trees and the nearest river. These surveys were forwarded to higher-level officials for planning purposes. Over the course of the sixteenth and seventeenth centuries, these surveys gave several successive generations of officials a panoptic view of the western forests that allowed them to make necessary changes to the logging administration, in particular the switch to requisition-by-purchase.
Did the large-scale logging of the three-province frontier result in deforestation and environmental degradation?
The evidence for deforestation is mixed. From the early 1500s onward officials repeatedly noted the bare hills (tongshan) near rivers, a result of overcutting, and wrote that lumber teams had to push deeper into the mountains to find poles of sufficient size.
The removal of old-growth shanmu and nanmu is further apparent when we compare the lumber yields. In the 1550s the Sichuan-Guizhou region yielded fifteen-thousand poles; twenty-four thousand poles were reportedly cut in the first decade of the 1600s. In the 1780s, four thousand five-hundred poles of nanmu and a similar amount of shanmu were cut from Sichuan-Guizhou; officials remarked that this was only one-third of the earlier yield, and that only one tenth of the nanmu and one-fifth of the shanmu were considered adequate for use. The 1720s requisitions obtained only one thousand forty-four suitable poles of nanmu; a low was reached in 1750, when the yield of the lumbering bottomed out at a mere one-hundred and forty-four poles. Even if we assume that a similarly small fraction of the Ming timber was suitable for construction, the yield of the Sichuan lumberyards in 1750 was about one twentieth their production in the early 1600s.
But declining yields of imperial-grade timber was not the same as total deforestation. Official reports made clear that there were still large forests in Sichuan in the late Ming, and even in the early and mid-Qing. Total clearance was limited to valleys with good water access; in the deeper mountains there were still large stands of old-growth. Instead, declining timber yields reflect a fundamental shift in the nature of imperial logging. In the 1540s, officials were sent to oversee timber extraction in Sichuan, Guizhou and Huguang. Except for a few sites in central Huguang, all of these forests were logged by conscripted laborers overseen by Ming officials.
By the 1680s, state lumbering concentrated exclusively on the Sichuan frontier, including a few north-flowing rivers in northern Guizhou that were part of the same watershed.
These were among the few areas that still had extremely large timber; they were also among the most difficult areas to log due to extreme topography and sparse populations. Because of these difficulties the Sichuan gorges could only be logged by imperial operations; the costs of cutting giant trees in the deep mountains of southern Sichuan were too great for merchants, and the potential profits were too small.
While imperial lumbering large ceased in most of the western frontier by the end of the Ming, lumbering itself continued under the oversight of private landowners, private lumberjack teams, and private merchants. In Hunan, Hubei and along the east-flowing rivers in eastern Guizhou, the late Ming and Qing states supervised the timber harvest at the market rather than in the forest.
These markets produced more than enough ordinary-sized timber for state building projects to without the need intervene in territorial forestry. I will address the development of wood markets and licensure of merchants more in the second part of this essay.
Imperial lumbering continued in the Sichuan gorges exclusively to obtain timber bigger than the market could provide, in particular shanmu and nanmu of age and size that could not reasonably be produced by plantation forestry with production schedules of twenty to fifty years.
The decline of imperial timber yields was therefore a highly specific and localized trend: it meant the end of accessible stands of old-growth timber of unusual size, but did not necessarily mean the progressive deforestation of western China.
