3 Reinterpreting Traditional Chinese Geographical Maps: Cordell YEE
3 Reinterpreting Traditional Chinese Geographical Maps: Cordell YEE
3 Reinterpreting Traditional Chinese Geographical Maps: Cordell YEE
My interest in this chapter and the following four is tra- frames in mind. These extended inquiries are obtained
ditional Chinese geographic mapping-that is, Chinese at some cost, however. With a thematic approach one
mapping of the earth before its Westernization in the late risks losing a clear sense of chronology, and one sacrifices
nineteenth and early twentieth centuries. One of the first the power of narrative to maintain a sense of direction.
lessons one learns when studying this subject is that the By keeping the focus on ideas or themes, one also risks
traditional periodization used in scholarship is unsatis- losing sight of the maps themselves. Detailed descriptions
factory. The traditional scheme takes the rise and fall of of artifacts can disrupt the flow of an argument or at least
China's ruling houses as constituting distinct periods (see make it harder to follow, and so in the thematic chapters
table 2.1). Such a scheme may have been useful for orga- that follow this one, artifacts are dealt with in only as
nizing material dealing with political and institutional his- much detail as is necessary to support the arguments pre-
tory, and as will be seen in a later chapter, cartography sented.
was intimately connected to that history. But carto- The loss of chronology and detail would be regret-
graphic developments do not neatly parallel changes in table, especially when at least part of the audience for
politics. Historians of cartography in the past, however, this book-collectors and cartobibliographers, for exam-
have tried to tie cartography to dynastic changes in ways ple-could reasonably be expected to take an interest in
I have found misleading. For example, Wang Yong, a such matters. One of the aims of this opening chapter,
pioneer in the study of traditional Chinese maps, once therefore, is to discuss artifacts in greater detail than else-
claimed that Tang (618-907) cartography was superior to where in this section and to give some sense of their
that of the Song dynasty (960-1279), even though vir- chronology.
tually no cartographic artifacts from the Tang are extant. 1 We should remember that, in an important sense, the
Other writers have made similar claims: that the Yuan history of Chinese cartography has yet to be written.
(1279-1368) and Ming dynasties (1368-1644) represent There are large gaps in the artifactual record. Between
the high point of Chinese cartography, that scientific car- the Later Han (25-220) and the end of the ninth century,
tography in China began in the third century or as early for example, there are almost no maps. For the Ming
as the Former Han (206 B.C.-A.D. 8)-depending on one's and Qing periods (1644-1911) one faces the opposite
sources. Given that we have only a handful of artifacts problem: a superabundance of maps. The primary source
that warrant cartographic interest for the period from the material on the cartography of these periods outbulks
first through the tenth century, these claims seem over- that of all previous dynasties combined. In addition to
confident. That such claims are made at all suggests there thousands of maps in the imperial archives and thousands
are at least two major premises underlying them: that of gazetteers, there are collections of memorials and
history is best seen as a march of progress through a other documentary material that need to be examined
generally ascending series of high points; and as a cor- for evidence of map use within the Ming and Qing admin-
ollary, that cartographic history is best seen as a move- istrations. Thus there are abundant opportunities for fur-
ment toward increased mathematization or quantifica- ther research (see appendix 3.1).
tion-toward cartography in its modern manifestation. I would like to acknowledge the help of Kevin Kaufman and the
In this book I am questioning these premises, and one coeditors of the History of Cartography in the preparation of this
of the results of this questioning is the organization of chapter.
the material. For the most part I have chosen to arrange 1. Wang Yong, Zhongguo dilixue shi (History of geography in China)
(1938; reprinted Taipei: Shangwu Yinshuguan, 1974), 70, 74. Chavannes
the material thematically. I believe this is the best way
made similar claims, saying that advances in cartography were made
to carry out searching examinations of the questions during the Tang. See Edouard Chavannes, "Les deux plus anciens spec-
raised by the maps and other sources-inquiries that imens de la cartographie chinoise," Bulletin de rEcole Fran~aise d'Ex-
might be limited if we were forced to keep dynastic time trerne Orient 3 (1903): 214-47, esp. 244.
35
r'----------ili'*,-ijB--------....'1 ••- - - - - ».'l _
FIG. 3.1. ZHAOYU TU ENGRAVED ON BRONZE. Notes original bronze plate bears graphs in their archaic form.
on the map give dimensions of the mausoleum in terms of chi Size of the original: 48 X 94 cm (ca. 1 cm thick). Pho'ograph
(feet) and bu (paces). The chi is thought to have been equal to courtesy of China Pictorial Publications, Beijing. Reconstruc-
twenty-two or twenty-five centimeters, while estimated values tion from Cao Wanru et aI., eds., Zhongguo gudai ditu ;i (Bei-
for the bu range between five and seven chi. Below is a recon- jing: Wenwu Chubanshe, 1990-), vol. 1, fig. 3.
struction of the zhaoyu tu with modern Chinese graphs. The
Reinterpreting Traditional Chinese Geographical Maps 37
In this chapter, artifacts through the fourteenth century ers refer to it as the zhaoyu tu, which, because of the
are well represented, since they are relatively few. For ambiguity in the graph tu, can be rendered as mausoleum
later periods the coverage is somewhat more selective, map or plan. It was discovered in 1978 in a tomb
since maps from those periods, especially the Qing, are unearthed in Pingshan Xian (County), Hebei Province. 3
discussed more fully later in this section as part of an The tomb is that of King Cuo of the Zhongshan kingdom,
examination of the Westernization of Chinese carto- a small state of the Zhanguo (Warring States) period
graphy. (403-221 B.C.). He was buried about 310 B.C., so the
The descriptions in this chapter are not provided solely zhaoyu tu dates from at least the fourth century B.C. 4
for the sake of presenting artifacts or for the sake of The tu is engraved on a bronze plate (fig. 3.1). It is
chronology. Those purposes are subordinate to another oriented with south at the top and is believed to represent
one: to introduce some of the themes and issues explored a walled area of about 191 by 414 meters. 5 It depicts in
in depth in the chapters that follow. In doing so, I will plan five sacrifice halls, four smaller buildings, an inner
refer to the artifacts and related documents and show and outer wall, and a baseline marking the foot of the
how they raise questions of central concern to historians grave mound. These features are represented with lines
of cartography-questions that have been largely over- of gold and silver inlay. The five sacrifice halls were
looked. intended to cover the tombs of King Cuo, his two queens,
and two other members of the royal family. At the Ping-
CHINESE MAPPING: shan site, two tombs have been unearthed, those of King
A MATHEMATICAL TRADITION? Cuo and Queen Ai. The other three tombs were never
built, evidently because several years after the king's inter-
The previous scholarship on traditional Chinese cartogra- ment, the Zhongshan kingdom fell to another state. 6 All
phy is characterized by a remarkable unity of approach. this suggests that the zhaoyu tu was a plan for construc-
It has generally tried to interpret Chinese cartography as tion rather than a map depicting actual structures.
a mathematical or quantitative tradition-characterized The tu is annotated, and the notes include a transcrip-
by attention to scale, conventional abstract signs, and tion of a decree issued by the Zhongshan king. This
practical function, such as planning, administrative, and decree was probably responsible for the tu's survival,
military. Mapping so conceived is mathematical in at least because it provided for the preservation of the plan: "one
two respects. First, it involves quantification and the [copy] to accompany [the burial] and one to be stored
reduction of topographical features to signs that aid the in the archives."? The notes on the tu also name the
presentation of quantitative information. Second, it objects represented in it and provide linear measurements
serves purposes that often involve the application of for the dimensions of the buildings and the distances
mathematics. If scale, abstract signs, and practical func- between them.
tion characterize Chinese mapping, then it might be said The next group of artifacts in the current canon com-
to constitute a rational discipline, or science, of cartogra- prises seven maps, drawn in ink on four wooden boards
phy. The current literature suggests that a science of
cartography developed in China at least by the Former 2. Chavannes, "Les deux plus anciens specimens," 236 (note 1).
Han and continued through the early Qing, when Chinese 3. The discovery was first reported by the Hebeisheng Wenwu Guan-
cartography became Westernized. The case for a contin- lichu (Hebei Province Cultural Relic Agency), "Hebeisheng Pingshan
uous mathematical tradition of Chinese mapping once xian Zhanguo shiqi Zhongshanguo muzang fajue jianbao" (Excavation
of the tombs of the Zhongshan kingdom of the Zhanguo period at
rested largely on a few textual sources and a few maps, Pingshan County, Hebei Province), Wenwu, 1979, no. 1:1-31.
notably the 1136 Yu ji tu (Map of the tracks of Yu [leg- 4. Fu Xinian, "Zhanguo Zhongshan wang Cuo mu chutu di 'zhaoyu
endary emperor famous for flood control]), which Chav- tu' ji qi lingyuan guizhi di yanjiu" (A study of the mausoleum map
annes said was "the result of a long scientific evolution."2 unearthed from the tomb of King Cuo of the Zhanguo period's Zhong-
According to recent scholarship, that case has been shan kingdom and the planning of the mausoleum), Kaogu Xuebao,
1980, no. 1:97-118, esp. 97.
strengthened by recent archaeological discoveries. 5. Yang Hongxun, "Zhanguo Zhongshan wang ling ji zhaoyu tu yan-
According to the mathematical or quantitative interpre- jiu" (A study of the mausoleum of the king of Zhanguo period's Zhong-
tation of traditional Chinese cartography, these artifacts shan kingdom and the mausoleum map), Kaogu Xuebao, 1980, no.
record an inevitably unsuccessful, but nonetheless essen- 1:119-38, esp. 127-29.
tially scientific, attempt to draw maps to scale. The claims 6. Liu Laicheng and Li Xiaodong, "Shi tan Zhanguo shiqi Zhong-
shanguo lishishang di jige wenti" (Tentative discussion of certain prob-
of proponents of this interpretation will be su bmitted to lems in the history of the Zhanguo period's Zhongshan kingdom),
critical analysis. But first I will present the artifactual and Wenwu, 1979, no. 1:32-36, esp. 33.
textual evidence for this interpretation. 7. Hebeisheng Wenwu Guanlichu, "Zhongshanguo muzang fajue
The earliest artifact lacks a title, but Chinese research- jianbao," 5 (note 3).
38 Cartography in China
FIG. 3.2. FANGMATAN MAP, VERSO OF BOARD 1. The Size of the original: 26.7 X 18.1 em. By permission of Wenwu
features represented on the maps found at Fangmatan are pri- Chubanshe. Reconstruction from Cao Wanru et aI., eds.,
marily linear: mountains, rivers and streams, and roads (see the Zhongguo gudai ditu ji (Beijing: Wenwu Chubanshe, 1990-),
reconstruction on the right). Some place-names also appear on vol. 1, fig. 13.
the maps.
(figs. 3.2 to 3.4). Six of the maps are on both sides of
three boards. The maps were discovered in 1986 at Fang-
matan forestry station, Tianshui, Gansu Province. They 8. The maps were first described in He Shuangquan, "Tianshui Fang-
had been buried in a tomb dating from perhaps 239 B.C. matan Qin mu churu diru churan" (Preliminary study of the maps exca-
and belonging to an officer in the Qin army whose per- vated from the Qin tomb at Fangmatan in Tianshui), Wenwu, 1989,
sonal name was Dan. s The boards on which the maps no. 2:12-22.
The date used here follows He Shuangquan's analysis of records on
were drawn are all about one centimeter thick and range the deceased found in the tomb. The records, on eight bamboo slips,
in width from 26.5 to 26.8 centimeters and in height from are transcribed in He Shuangquan, "Tianshui Fangmatan Qin jian zong-
15 to 18.1 centimeters. The maps all depict parts of the shu" (Comprehensive account of the Qin bamboo slips from Fangmatan
same region, an old administrative district called Gui in Tianshui), Wenwu, 1989, no. 2:23-31, esp. 28-29. The records note
Xian, which researchers have identified as the Wei River that Dan, who served as a military officer and participated in a northern
campaign, injured someone in the face and afterward killed himself.
valley and its tributaries, cutting through a section of the He was buried outside the city and came back to life three years later.
Qinling Mountains. This region, which includes the Fang- The records mention years but do not identify the reigning monarch.
matan grave site, held strategic importance as a defensible They suggest, however, that the ruler reigned for at least ten years. It
pass through which transportation was funneled from the is possible to date the tomb earlier than 239 B.C. as proposed by He
west into the heart of the Qin state. The maps represent Shuangquan. Zhang Xiugui, for example, would push the date of the
tom b back to around 300 B.C. by identifying a Qin ruler who also ru led
rivers and tributaries by black lines. Gullies, passes, trans- for at least ten years and launched northern expeditions. See Zhang
portation checkpoints, and stands of various kinds of Xiugui, "Tianshui 'Fangmatan diru' di huizhi niandai" (Date of the maps
trees, including pine, fir, cedar, and orange, are identified from Fangmatan, Tianshui), Fudan Xuebao, 1991, no. 1:44-48.
Reinterpreting Traditional Chinese Geographical Maps 39
~
I
FIG. 3.4. FANGMATAN MAP, DETAIL OF THE VERSO OF Size of the entire original: 18.1 x 26.5 em. By permission of
BOARD 3. This detail is about one-fourth of the entire board Wenwu Chubanshe. Reconstruction from Cao Wanru et aI.,
(reconstruction on the right). eds., Zhongguo gudai ditu ji (Beijing: Wenwu Chubanshe,
1990-), vol. 1, fig. 9.
because of gaps in the map. The lower part of the map condition. They have been restored, and detailed
shows a city with an outer and inner wall (fig. 3.7).12 accounts of them have been published.B Both are ori-
The other two maps found at Mawangdui are in better ented with south at the top. One represents the southern
part of the Changsha state and is referred to as a topo-
graphic map, because it emphasizes mountainous areas
and the courses of rivers (figs. 3.8 and 3.9). The names
of the mountains are not given, but rivers, as well as
county seats, are identified. The map was drawn with
vegetable colors. The third map (figs. 3.10 and 3.11) is
thought to represent a portion of the terrain shown on
the topographic map-more specifically, part of the
southernmost portion. This area was of military signifi-
cance since it bordered on Nanyue, a reluctant tributary
state to the Han. Because it shows the locations of army
installations and headquarters, the third map is thought
to have had military applications. It is notable for its use
of color: military-related features, roads, and some set-
tlements are shown in red; rivers and streams in light
blue-green; and other features and lettering in black. The
map also bears annotation. For some settlements, dis-
tances from other settlements and the number of house-
holds are given.
All the maps described so far have been valued for
their "modern" appearance. Their mode of presentation
seems to be planimetric, and the manner of depiction
FIG. 3.6. HAN SILK MAP UNEARTHED AT MAWANGDUI. tends toward conventionalization, for instance, in the re-
This map was unearthed in 1973 from tomb 3 at Mawangdui.
presentation of settlements, mountains, and trees. In
It was drawn on the same piece of silk as an illustration showing
breathing exercises and calisthenics. The tattered condition of
the map has made interpretation difficult. The lower half of the
map seems to represent a walled city (see fig. 3.7). 12. Information on this map is taken from Han Zhongmin's descrip-
Size of the original: 48 x 48 em. Photograph courtesy of Cao tion in Cao Wanru et al., eds., Zhongguo gudai ditu ji (Atlas of ancient
Wanru, Institute for the History of Natural Sciences, Academia Chinese maps) (Beijing: Wenwu Chubanshe, 1990-), 1:18. Han conjec-
Sinica, Beijing. tures that the map as a whole represents the mausoleum and the city
of the marquis of Dai, Li Cang. Cao Wanru, however, believes that the
map depicts cities and towns of the southern Changsha state; see Cao
Wanru, "Maps 2,000 Years Ago and Ancient Cartographical Rules," in
Ancient China' s Technology and Science, compo Instirute of the History
of Natural Sciences, Chinese Academy of Sciences (Beijing: Foreign
Languages Press, 1983), 250-57, esp. 251.
13. Information on the Han silk maps is drawn largely from the
following studies: Mawangdui Han Mu Boshu Zhengli Xiaozu (Study
Group on the Han Silk Manuscripts from Mawangdui), "Changsha
Mawangdui sanhao Han mu chutu ditu di zhengli" (Restoration of the
maps excavated from Han tomb 3 at Mawangdui, Changsha), Wenwu,
1975, no. 2:35-42; idem, "Mawangdui sanhao Han mu chutu zhujun
tu zhengli jianbao" (Preliminary report on the restoration of the military
map excavated from Han tomb 3 at Mawangdui), Wenwu, 1976, no.
1:18-23; Tan Qixiang, "Erqian yibaiduo nian qian di yifu ditu" (A map
from more than 2,100 years ago), Wenwu, 1975, no. 2:43-48; and Zhan
Libo, "Mawangdui Han mu chutu di shoubei tu tantao" (Investigation
of the garrison map excavated from the Han tomb at Mawangdui),
Wenwu, 1976, no. 1:24-27. English-language treatments of these maps
have been written by A. Gutkind Bulling, "Ancient Chinese Maps: Two
FIG. 3.7. DETAIL OF THE HAN SILK MAP. A copy of the Maps Discovered in a Han Dynasty Tomb from the Second Century
lower center portion of the map in figure 3.6. The outer wall B.C.," Expedition 20, no. 2 (1978): 16-25; Mei-ling Hsu, "The Han
of the city, as drawn on the map, measures 18.8 by 20 centi- Maps and Early Chinese Cartography," Annals of the Association of
meters. American Geographers 68 (1978): 45-60; and Kuei-sheng Chang, "The
Photograph courtesy of Cao Wanru, Institute for the History Han Maps: New Light on Cartography in Classical China," Imago
of Natural Sciences, Academia Sinica, Beijing. Mundi 31 (1979): 9-17.
42 Cartography in China
I
~
~!.-
..... '
-, r"_'• . ,.1: .
'" ~ ,<II
~-
. ~ ..
FIG. 3.8. TOPOGRAPHIC MAP FROM MAWANGDUI. Both become stuck together, making restoration difficult. This topo-
this map and those shown in figures 3.6 and 3.10 date from the graphic map consists of thirty-two pieces. Following the ori-
Han dynasty and were found in a lacquer box. They had been entation of the graphs, south is at the top.
folded up, and by the time they were unearthed they had dis- Size of the original: 96 X 96 em. By permission of Wenwu
integrated around the folds. The folded sections had also Chubanshe.
addition, the artifacts have been interpreted as examples was used, one drew a square map of eight chi and one
of scale mapping. There is some textual support for this. cun. In present usage, one draws a square map of [half
An astronomico-mathematical work, the Zhoubi suan the size, or] four chi and five fen. A fen [in this case] is
jing (Arithmetical classic of the Zhou gnomon, ca. 200 equa] to two thousand 1i."14 Modern scholars have
B.C.), describes the effect of altering scale on the size of
a map: "Whenever a scale of one fen to a thousand Ii 14. Zhoubi suan jing, Siku quanshu edition, A3.2a.
Reinterpreting Traditional Chinese Geographical Maps 43
T
I
FIG. 3.9. RECONSTRUCTED TOPOGRAPHIC MAP FROM From Gu ditu lunwen;i (Essays on ancient maps) (Beijing:
MAWANGDUI. The map is thought to represent an area mainly Wenwu Chubanshe, 1977).
lying between 110° and 112°30'E and between 23° and 26°N.
tended to interpret the pre-Han and Han artifacts in light that the scale for the area within the baseline of the
of such references to scale and have taken pains to estab- zhaoyu tu is about 1:500; outside it, this scale is not
lish scale mapping as a general practice during the Han maintained. The scale of the Fangmatan maps has been
and even earlier. From dimensions marked on the early determined to be about 1:300,000. As for the Mawangdui
artifacts and from comparisons of the early maps with maps, the scale has been found to vary between 1:150,000
modern maps, researchers in China have calculated the and 1:200,000 in the central portion of the topographic
scales of the early map images. It has been determined map and between 1:80,000 and 1:100,000 in the central
FIG. 3.10. GARRISON MAP FROM MAWANGDUI. This Size of the original: 98 X 78 em. By permission of Wenwu
map consists of twenty-eight pieces. South is at the top and Chubanshe.
labeled.
FIG. 3.11. RECONSTRUCTED GARRISON MAP FROM From Gu ditu lunwenji (Essays on ancient maps) (Beijing:
MAWANGDUI. Wenwu Chubanshe, 1977).
46 Cartography in China
FIG. 3.13. HUA Yl TU. Engraved on stone in 1136, it may have placing a sheet of paper over the engraved surface and applying
been based on a map by the Tang scholar Jia Dan. On the left ink to the paper so as to produce a white-on-black reproduc-
is the stele and on the right is a rubbing of the map. On the tion. An inscription on the lower right side of the map (when
other side of this stele is the 1136 Yu ji tu (Map of the tracks the map is right side up) says: "The venerable Jia Wei Uia Dan]
of Yu), figure 3.14, but it is oriented in the opposite direction of the Tang dynasty listed several hundred countries. Now we
so that when one map is right side up (in this case, fig. 3.14), [or I] have selected the well-known ones and recorded them
the other is inverted. Thus it would have been difficult to display here."
both of these maps on the stele at the same time. The stele Size of the original: 79 x 79 em. Photographs courtesy of Cao
may have been intended not for display but as a means of Wanru, Institute for the History of Natural Sciences, Academia
reproduction. The image could be transferred by "rubbing"- Sinica, Beijing.
is the Yu ji tu (Map of the tracks of Yu), also engraved to the 1136 version in many ways: in its use of the square
in 1136 (fig. 3.14). Internal evidence suggests that the map grid and an expressed scale of one side of a square to
image was originally drawn somewhat earlier: for exam- one hundred Ii, in the dimensions of the image, in its
ple, there are no names of prefectures and counties estab- depiction of mountains and rivers, and in the names of
lished after 1100. Until the discovery of the Mawangdui the prefectures and counties. There are some differences,
maps, the Yu ji tu, measuring about eighty centimeters however. The 1142 map image does not distinguish
square, was one of the most celebrated examples of the between main streams and tributaries. An inscription on
precocious development of Chinese cartography. The the 1142 version corroborates the suggested date of 1100
reasons are not hard to fathom. The map's representation for the original compilation: "Carved in the first month
of China's coastline is remarkably close to twentieth- of the third year of the Yuanfu reign period [1100]
century representations, and the map marks the first according to the model copy at Chang'an."
known appearance of the Chinese cartographic grid, con- It might seem that in comparing the Jiu yu shouling
sisting of a latticework of squares of equal size super- tu and Hua yi tu with the Yu ji tu, one sees in the last
imposed on a map image. The squares serve to indicate increasing technical competence in the production of
scale: each grid increment, or side of a square, represents map images. One sign is the use of the grid as a scaling
a fixed ground distance. A note engraved on the Yu ji tu device, suggesting the importance of scale imaging to
states that each side of a square represents one hundred
Ii, and on that basis the map scale is assumed to be about 19. Cao et aI., Zhongguo gudai ditu ji, 21 (note 12).
20. There are reports of a third stone Yu ji tu, now lost, that was
1:4,500,000. 19
once kept at Jishan Xian, Shanxi. The map is said to have had a square
A second version of the Yu ji tu, at Zhenjiang, Jiangsu grid with a scale of "one [side of al square to one hundred Ii." See Lu
(fig. 3.15),20 was carved in stone in 1142, commissioned Liangzhi, Zhongguo dituxue shi (History of Chinese cartography) (Bei-
by the director of the local prefectural school. It is similar jing: Cehui Chubanshe, 1984), 156.
48 Cartography in China
FIG. 3.16. MAP FROM THE GUANG YUTU. According to hydrological maps, and maps of Korea, Vietnam, and Japan.
the accompanying text, this general map of the empire is drawn This illustration is taken from the Guang yutu dated 1799,
so that "each [side of a] square [represents] one hundred Ii." which is an exact copy of the 1579 edition.
All the maps in the Guang yutu have grids, most with each side Size of the original: 28.5 X 41 em. By permission of the British
of a square representing one hundred Ii. Among the other maps Library, London (l5261.e.2), 1b-2a.
included in the Guang yutu are provincial maps, coastal maps,
tradition, not an innovation. To judge from the present one scholar suspects that Shen Kuo was responsible for
artifactual record, that tradition did not include a car- the Yu ji tu, thereby linking Shen more closely to a "grid"
tographic grid. tradition. 23
The next important figure for the exponents of the That tradition is said to extend to the sixteenth century
mathematical interpretation is ]ia Dan (730-805), who with the work of Luo Hongxian (1504-64): grids appear
cites Pei Xiu's attention to measurement as a model. In on the maps in his Guang yutu (Enlarged terrestrial atlas,
one historical account, ]ia Dan is described as having
ordered a scale map to be made: "He ordered an artisan
to paint the Hainei Hua yi tu [Map of Chinese and 22. Liu Xu et al.,]iu Tang shu (Old history of the Tang, compiled
940-45), chap. 138; see the edition in 16 vols. (Beijing: Zhonghua Shuju,
foreign lands within the seas] on a scroll. It was three 1975), 12:3786.
zhang wide and three zhang and three chi high. Its scale 23. Cao Wanru, "Lun Shen Kuo zai diruxue fangmian di gongxian"
was one cun to one hundred li."22 Since he was familiar (On Shen Kuo's contributions to cartography), Keji Shi Wenji 3 (1980):
with Pei's principles and admired them, ]ia Dan has been 81-84. Cao's argument is based on Shen Kuo's presence near Chang'an
regarded by some as a user of the square grid. Shen Kuo (present-day Xi'an) in the years 1080-82, a period coinciding toughly
with the time when the information reflected on the Yu ji tu was
(1031-95) can be added to the list of mathematical carto- compiled. She reasons that since Shen Kuo compiled a map of the
graphers. He uses terminology similar to Pei Xiu's in a empire and had high standards of mapmaking that are reflected on the
list of his own mapmaking methods. In addition, at least stone map, he could have been the author of the original image.
Reinterpreting Traditional Chinese Geographical Maps' 51
- '\~. "
FIG. 3.19. RECONSTRUCTION OF THE SIGN FOR jlUYI
SHAN. The shadow image shows up more clearly here (com-
pare fig. 3.18).
From Gu ditu lunwenji (Essays on ancient maps) (Beijing:
FIG. 3.18. SIGN FOR jlUYI SHAN (NINE BEGUILING Wenwu Chubanshe, 1977).
MOUNTAINS). Shown in the photograph is the sign for jiuyi
Shan that appears on the topographic map found at Mawangdui Island in the Indian Ocean. A navigational chart contem-
(fig. 3.8). A number of utilitarian functions have been proposed poraneous with Zheng He's voyages has been preserved
for its shape and shading. What seems to be a shadow image in the Wubei zhi (Treatise on military preparations), com-
of part of the sign appears in the representation of an adjacent piled about 1621 by Mao Yuanyi (1594-ca. 1641). Mao
lake. This suggests that the sign may have had aesthetic value:
the mapmaker may have been trying to show the reflection of provides no information on the source or maker of the
the mountains in the lake. chart, but some have traced it to Zheng He's voyages. 32
Size of the detail: ca. 24 x 15 em. By permission of Wenwu
Chubanshe. 31. Hsu, "Han Maps," 55 (note 13). See also Yang Wenheng, "Shilun
Changsha Mawangdui sanhao Han muzhong chutu ditu di shuli jichu"
(On the mathematical foundation of the maps excavated from Han
ematical cartography was laid even before Pei Xiu. Such
tomb 3 at Mawangdui), Ke;i Shi Wen;i 3 (1980): 85-92, esp. 86. Even
a foundation existed at least by the Han, if the Mawang- granting that the instruments and geometry necessary for surveying had
dui maps are admitted as evidence: "The detail and accu- been developed by the Han, in speaking of early Chinese maps one
racy of these maps could not have been achieved without must be wary of applying the word "survey" in its modem sense: gath-
substantial field surveying using relatively sophisticated ering information about a well-defined area by the use of rechniques
techniques."31 Such mensurational techniques were applied consistently throughout. Much of the information reAected on
the Mawangdui maps, for example, could have been obtained from
developed not only for cartography but also for related direct observation, but it might have been gathered in parchwork fash-
endeavors such as navigation and astronomy, and the ion, in bits and pieces, not systematically, from several sources: direct
evidence for these techniques often comes from those measurements, existing records, even local lore about travel time. All
other fields. of this information could be combined, compared, and adjusted to
The greatest achievement in navigation is associated create a map giving approximate locations with the same accuracy as
the Mawangdui maps. There are European examples of this.
with Zheng He (1371-1433), who from 1405 to 1433 led 32. Mills, however, rraces the map to Mao Kun (1512-1601), Mao
seven maritime expeditions as far west as the coast of Yuanyi's grandfather, who was a member of an admiral's staff. Mills
eastern Africa and perhaps as far south as Kerguelen believes the map was the work of a cartographer and his assistants who
Reinterpreting Traditional Chinese Geographical Maps 53
The chart depicts a voyage· from Nanjing to the island THE USE AND ABUSE OF
of Hormuz to ports on the east African coast. It is CARTOGRAPHIC HISTORY:
thought to have been originally a strip measuring about FLAWS IN THE QUANTITATIVE ApPROACH
20.5 by 560 centimeters, with the image on it divided
into forty sheets when included in the Wubei zhi. The The quantitative interpretation is tenable if one insists on
strip format has been found to result in angular, linear, linking mathematics and cartography, so that only those
and directional distortions. The orientation of the chart, images that bear signs of measurement are considered
for example, changes frequently (fig. 3.20), and the map cartographic and those associated with text are beneath
scale has been found to vary from section to section cartographic consideration. Uniformity of scale, I agree,
according to the amount of detail represented. Sea routes is important for achieving a certain kind of fidelity to
are depicted with broken lines, and sailing instructions geographic reality, one emphasizing the representation of
are given in notes on the chart. Some scholars have linear distances. As I stated above, the pre-Han and Han
checked the instructions against modern charts and found artifacts can be interpreted as evidence that the idea of
that, in contrast to the map image, the instructions are scale was understood. But there are some problems with
accurate for the most part with regard to distance and that interpretation. For example, the deteriorated con-
direction. 33 The chart is thus taken as evidence of sophis- dition of some of the artifacts when excavated makes
ticated navigational techniques based on astronomical determination of exact scale difficult: there are gaps in
observations and the use of the magnetic compass. the artifacts, and scale determinations can vary depending
Astronomical techniques also reached a high level of
development. During the Tang and Yuan, for example, compiled the chart from earlier ones. Others have conjectured that the
the Chinese have been credited with what are interpreted map had origins in the Islamic world. See Ma Huan, Ying-yai Sheng-
Ian: "The Overall Survey of the Ocean's Shores" [1433J, ed. and trans.
as geodetic surveys. The Tang survey involved observa- J. V. G. Mills (Cambridge: Cambridge University Press, 1970), 239-41.
tions from a series of stations forming a chain about 3,800 33. Mills, for example, remarks that in some cases the details on the
kilometers long. The survey determined, among other chart are "astonishingly accurate," but he also finds that the instructions
things, that a north-south distance of about 351 Ii cor- for the main voyage from Ceylon to Hormuz are "defective" and that
the instructions for sailing along the east African coast are "rudimentary
responded to a change of one degree in the altitude of
and inadequate." See Mills, Ying-yai Sheng-Ian, 248 (note 32). Hso Yo-
the Pole Star. This information, some have speculated, hu finds the distance and bearing information to be "surprisingly" accu-
was used to produce large-area maps such as the Yu ji rate for routes along the Southeast Asian coast. See Hso Yo-hu (Xu
tU. 34 The Yuan "survey of the four seas" (sihai ceyan) Yuhu), Mingdai Zheng He hanghai tu zhi yanjiu (Study of Zheng He's
was more extensive than the Tang survey, involving nautical chart from the Ming period) (Taipei: Xuesheng Shuju, 1976),
7. The difference in opinions may be related to disagreement as to what
twenty-seven observation stations within an area covering
distance corresponded to the Chinese geng, or watch, equal to 2.4 hours.
about 5,000 kilometers from north to south and about Distance on the chart is given in terms of geng, and one's calculation
2,700 kilometers from east to west. 35 Measurements of of that distance will vary depending on how fast one believes a Chinese
the length of the sun's shadow at noon were taken, often sailing vessel could travel. Estimates of sailing speed range from about
with gnomons as tall as twelve meters (fig. 3.21). The twelve to twenty nautical miles per geng. A useful collection of studies
on the chart is Zhongguo Hanghai Shi Yanjiuhui (Research Association
determinations and measurements from these surveys
for the History of Chinese Navigation), ed., Zheng He yanjiu ziliao
were intended for the correction of calendars, but they xuanbian (Selected research materials on Zheng He) (Beijing: Renmin
had potential cartographic applications. From the length Jiaotong Chubanshe, 1985).
of the sun's shadow, for example, it is possible to com- 34. Arthur Beer et aI., "An 8th-Century Meridian Line: I-Hsing's
pute latitude. Chain of Gnomons and the Pre-history of the Metric System," Vistas
in Astronomy 4 (1961): 3-28, esp. 16.
Because of their cartographic applications, potential or
35. This "survey" is described in Song Lian et aI., Yuan shi (History
actual, the meridian "surveys" and navigation have been of the Yuan, compiled 1369-70), chap. 48; see the edition in 15 vols.
discussed in recent histories of Chinese cartography as (Beijing: Zhonghua Shuju, 1976), 4:1000-1001. The same history also
indicative of its mathematical foundations. The only dif- records an attempt by Yelo Chucai (1190-1244) to correlate celestial
ficulty here is that the cartographic application of astro- measurements and terrestrial distance between Beijing and Samarkand.
Yelo, an astrologer who served the Mongol leader Chinggis Khan,
nomical techniques has yet to be demonstrated, and the
noticed a two- or three-hour time lag between celestial occurrences
earliest evidence for the cartographic application of sea- predicted in the official Chinese calendar, which had been made for use
faring techniques exists in the Wubei zhi, which dates in Beijing, and their occurrence in Samarkand. To overcome this, Yelo
from the seventeenth century. Thus evidence of men- compiled a new calendar for Samarkand, using an adjustment factor to
suration in fields often associated with cartography still calculate the time difference of astronomical phenomena observed in
regions east or west of there. This factor, however, did not find car-
does not give much support to the idea of a continuing
tographic application, though it could have been used to determine
tradition of measured mapping. longitude in the Western sense of a geographical coordinate. See Yuan
shi, chap. 52 (4:1119-20).
54 Cartography in China
FIG. 3.20. NAUTICAL CHART FROM THE WUBEI ZHI. Huangpujiang (Whangpoo River). The chart bears no scale indi-
Shown here are the first six pages of the nautical chart, which cations, but notes give compass directions and sailing times for
cover the route from the Nanjing area to just south of the routes between various places. Orientation changes as one
on how one fills in those gaps. In the case of the Mawang- methods of graphic production are concerned, there was
dui maps, it is important to note that scale consistency one technological innovation: the development of wood-
decreases as one moves away from the center. It is not block printing during the eighth century, which made it
sufficient to consider absolute variation in scale across a easier to reproduce and thereby to disseminate maps.
whole map, as several researchers have done; one must Thus, after the tenth century we have many more maps
also take into account the distribution of that variation. extant than for earlier periods.
Some researchers on early Chinese maps have tried to One should also note that the same methods of pro-
establish uniformity of scale by hiding its variability, by duction were employed in the graphic arts in general, so
giving the scale as a range. It also will not suffice to argue that it is possible to conjoin cartography and other visual
that the center is of main interest, and therefore that it arts. The failure to explore this connection has led to
is drawn to scale or exhibits a narrower range of scale the devaluation of a large body of artifacts that fall within
variation. For defense purposes, which the garrison map our definition: "graphic representations that facilitate a
is thought to have served, the periphery might very well spatial understanding of things, concepts, conditions,
have been of main interest. There is not enough evidence, processes, or events in the human world."37 If one views
in either case, to judge. We should also remember as we Chinese artifacts with this definition in mind, one gets a
consider the scales given for the early maps that none of
the artifacts in question bears an expressed scale, and 36. Han Zhongmin has suggested that too much emphasis has been
that the "rounding off" of the ratio may imply more placed on ascertaining the accuracy of the Mawangdui maps. He pro-
systematizing than actually existed. 36 poses a different reconstruction of the garrison map, one that would
In addition, emphasis on scale and accuracy has tended result in a map of ninety-six by ninety-six centimeters instead of ninety-
to divert attention from other questions about traditional eight by seventy-eight centimeters. See Han Zhongmin, "Guanyu
Mawangdui boshu gu ditu di zhengli yu yaniiu" (Concerning the res-
Chinese mapping. For example, more attention could be toration and study of the ancient silk maps from Mawangdui), in Zhong-
paid to methods of graphic production. It seems that guo gudai ditu ii, ed. Cao Wanru et al. (Beijing: Wenwu Chubanshe,
much of the technological basis for map production had 1990-),1:12-17. Curiously, no illustration of Han's proposed recon-
been laid by the end of the Han dynasty. Maps continued struction appears. Han does imply that the map would no longer be as
to be produced with brush and ink on silk and paper, accurate as is thought: his reconstruction apparently would alter the
degree of correspondence between the map and the area it supposedly
and to be engraved on various media, up to the twentieth depicts.
century. This continuity is one reason for not dividing 37. Preface to The History of Cartography, ed. J. B. Harley and
up Chinese cartographic history into dynasties. As far as David Woodward (Chicago: University of Chicago Press, 1987-), 1:xvi.
Reinterpreting Traditional Chinese Geographical Maps 55
moves across the map. On the first two pages, starting from the Size of each page: ca. 14.5 x 10 em. Courtesy of the Asian
right, south-southeast is at the top. On the third, fourth, and Division, Library of Congress, Washington, D.C. (F701.M32.1).
fifth pages, south is generally at the top. On the sixth page, west
is at the top.
different impression of Chinese mapmaking-that scale in the following chapters, that maps did serve utilitarian
maps tend to be isolated examples. 38 One might object purposes, such as administration and military planning.
that looking at artifacts with a broader conception of But textual and artifactual evidence makes it clear that,
maps is merely another instance of arbitrariness, of for the literary elite, maps also served religious functions.
imposing a different preconception on Chinese artifacts. They could depict the arrangement of implements used
A broader conception, the following chapters show, is in rites. They could be used to locate auspicious sites for
sanctioned by the context: the artifactual record itself buildings. They could serve as talismans to ward off
demands a broader understanding of the map. A focus demons or as representations of power to secure passage
on scale mapping is insufficient to account for the broad to the otherworld. They could record astrological infor-
range of production and the variety of representational mation and thus help interpret heavenly signs.
practices. Furthermore, what are taken to be scale maps All this suggests that to understand early Chinese maps,
often have much in common with artifacts not usually we need to study them in the context of the beliefs and
regarded as cartographic. The modes of representation values of the ruling elite, not "abuse" them by imposing
on the Mawangdui maps, for example, appear less strik- modern conceptions of cartography. As is described else-
ing when one sees similar curves representing mountains where, maps in Chinese culture were used not only for
on a map of a burial site dating from the tenth century representing distances, but also for demonstrating power,
and bearing no scale indications (fig. 3.22). for education, and for aesthetic appreciation. A concep-
Since the artifacts singled out for attention under the tion of Chinese cartography as a rational, mathematical
quantitative interpretation are viewed as instances of scale discipline for understanding space has led to a failure to
mapping, they are regarded as "scientific" and as serving
38. The range of cartographic styles in traditional China is suggested
secular, utilitarian purposes. Thus they are described as by the artifacts collected in Zhongguo gudai ditu ji, ed. Cao et al. (note
topographic military, economic, or administrative maps. 12). Many of the maps illustrated in this atlas have no expressed scale
Such descriptions, however, tend to overlook the con- and are highly pictorial. The maps selected for extended analyses, how-
texts in which the artifacts were found-for example, ever, are generally those that researchers are able to evaluate in terms
of accuracy and scale-namely, those that support the notion of progress
tombs of members of the ruling class. The placing of
within a quantitative tradition. The atlas is useful for collecting a latge
maps in tombs seems to confirm that the functions of number of Yuan and pre-Yuan maps, bur it also represents a missed
maps from at least the Han went beyond the secular (see opportunity to examine the Chinese map tradition whole. Here I
pp. 77-80). It is clear from literary sources, cited in detail attempt to correct the imbalance.
56 Cartography in China
FIG. 3.21. YUAN OBSERVATIONAL TOWER. This obser- From Zhongguo Shehui Kexueyuan Kaogu Yanjiusuo (Archae-
vational tower in Dengfeng Xian (in present-day Henan Prov- ological Research Institute, Chinese Academy of Social Science),
ince) once contained a gnomon about twelve meters tall. Zhongguo gudai tianwen wenwu tu;i (Album of ancient
Extending horizontally on the ground from the north side of Chinese astronomical relics) (Beijing: Wenwu Chubanshe,
the tower is a graduated scale more than thirty-six meters long, 1980), 14.
which was used to measure the sun's shadow.
consider the full range of cartographic functions. It also two types of maps are both predicated on the importance
leads to an anomaly: if the all-important goal was scale of mathematical representation to cartography. This does
mapping, the preservation of such disparate images as the not seem to be the case with the two twelfth-century
Hua yi tu and Yu ji tu on the same stone stele seems stone maps at Xi'an. The claim advanced by advocates
somewhat curious. of the quantitative approach is that these maps represent
One researcher attempts to explain this by identifying qualitatively different ways of mapmaking, one more
two "parallel" traditions of Chinese cartography, one advanced, more scientific than the other. Thus, one might
mathematical or "analytic" and the other "descriptive."39 argue that the maps on the stone stele are products of a
The former is concerned with measurement and is there- transitional period.
fore a "science"; the latter is concerned with "infor- The problem with that interpretation is that ten cen-
mation" and less concerned with accuracy. There is, turies seems rather long for a transition, and that the
however, no evidence that Chinese mapmakers saw them- transition, as we will see below, does not seem to have
selves as working in two traditions distinguished by their been fully realized. The evidence does not support the
attention to measurement, and the identification of two view that Chinese maps were perceived within their cul-
traditions still does not explain why parallel traditions tural contexts as representing different traditions. Maps
would meet on the same stone. In the history of carto- like the Hua yi tu, in short, were not regarded as inferior
graphy, there are, to be sure, many examples of maps to maps like the Yu ji tu.
from different traditions appearing side by side. In Re- The preservation of both maps on the same stone sug-
naissance atlases, for example, Ptolemaic and "modern"
world maps appear together. But in such instances the 39. See Hsu, "Han Maps," 56-59 (note 13).
Reinterpreting Traditional Chinese Geographical Maps 57
FIG. 3.22. MAP OF AN AUSPICIOUS SITE FOR A FAMILY Size of the entire original: ca. 25 x 229 em; this detail: 24 x
TOMB (DETAIL). This map, discovered at Dunhuang in the 39.5 em. By permission of the British Library, London (Stein
Cave of the Thousand Buddhas, was drawn perhaps in the tenth no. 3877 [Dunhuang MS. 6971]).
century. A note on the map says: "Anyone buried here will have
good fortune."
gests that considerations other than scale and mathe- 1403-24) ordered scholars to gather maps and documents
matical correspondence to geographic actuality were con- from the empire's prefectures and towns and to combine
sidered important for mapmaking. The artifactual record them into one work. 41 This project was not completed,
provides ample evidence to undermine the idea that the however, until the emperor Yingzong resumed the proj-
Yu ji tu established a foundation for a climax of math- ect, which was completed in 1461.
ematical cartography during the Yuan and Ming, or that The main body of the comprehensive gazetteer opens
there was a general advance in measured mapping from with a preface to a "comprehensive map" of the empire
the twelfth through the seventeenth century.40 The maps and contiguous lands (fig. 3.25). The compilers of the
in the Lidai dili zhizhang tu (Easy-to-use maps of geog- gazetteer state that they provided this map at the begin-
raphy through the dynasties, 1098-1100, supplemented ning of the work so that one could "tum to the map and
1162), for instance, are roughly contemporaneous with view the entire world, and the extent of the territory
the Yu ji tu but do not adhere as closely to geographic would be clearly in view as if one were looking at one's
actuality. China's outlines are depicted as being more own palm."42 The gazetteer was based on more than
squarish than they really are, and the Shandong peninsula thirty years of information gathering, but the quality of
almost disappears (figs. 3.23 and 3.24). Furthermore,
islands representing other countries are nestled in an
40. Exponents of the view that traditional Chinese cartography
encircling ocean. reached its peak during the Yuan and Ming include Joseph Needham
From a quantitative point of view, the image does not and, more recently, Lu Liangzhi. See Needham, Science and Civilisation
necessarily improve with later maps, for example, those in China, 3:551-56 (note 21); and Lu, Zhongguo dituxue shi, 99 (note
in the Da Ming yitong zhi (Comprehensive gazetteer of 20).
41. See the emperor's preface to Li Xian et aI., Da Ming yitong zhi,
the Great Ming, 1461). According to the gazetteer's pre-
1b; see the edition in 10 vols. (Taipei: Tailian Guofeng Chubanshe,
face composed by the emperor Yingzong (r. 1436-49, 1977), 1:2.
1457-64), the work of compiling geographic information 42. Li et aI., Da Ming yitong zhi, preface (by compilers), 2b (1:56)
for the gazetteer began when the emperor Chengzu (r. (note 41).
58 Cartography in China
FIG. 3.23. "GUJIN HUA YI QUYU ZONGYAO TU" (GEN- map is accompanied by explanatory notes. The maps shown
ERAL MAP OF THE ANCIENT AND PRESENT TERRI- here and in figure 3.24 are from a Southern Song edition.
TORIES OF CHINA AND FOREIGN COUNTRIES). This Size of the original: unknown. Photograph courtesy of Cao
map is from the Lidai dili zhizhang tu, a collection of fony- Wanru, Institute for the History of Natural Sciences, Academia
four historical maps of dynastic territories from the time of Ku, Sinica, Beijing.
legendary male ancestor of the Shang, to the Song dynasty. Each
the geographic information reflected in the general map contrast, despite the importance attached to visual repre-
does not quite fulfill the compilers' intentions. According sentation in the emperor's order, maps actually constitute
to the criteria of the quantitative approach, the map re- only a small portion of the Da Ming yitong zhi. The
presents a step backward from the two stone Yu ji tu, work contains fifteen other maps besides the general map
which predate the gazetteer by more than three hundred discussed above, so that maps occupy a mere thirteen
years. The Ming map is less detailed than the Song map, leaves out of more than 2,800. These fifteen maps repre-
particularly in the depiction of the courses of rivers and sent subdivisions of the territory depicted on the general
the configuration of the coastline. Unlike the Song map, map-the two capital districts and thirteen provinces-
the Ming map lacks a grid and shows Yunnan and even and introduce textual descriptions of the geographic sub-
what is now Tibet as having a coastline. The Ming map divisions. 43
also makes greater use of pictorial elements, notably in
the representation of mountains, which are merely named 43. The text itself subdivides these divisions further and is organized
on the Song map. into a series of treatises, most devoted to a single prefecture (ful or
tributary state. The treatises are further subdivided into sections on
The high degree of generalization in the Ming map
topics such as local history, geography, and customs, as well as famous
may be at least partly explained by its extensive textual officials, itinerant scholars, exemplary women, and Daoist and Buddhist
supplementation. The Song map has very little textual clergy. Such categories of information became standard in later gaz·
accompaniment; the cartographic image is dominant. In etteers.
Reinterpreting Traditional Chinese Geographical Maps 59
FIG. 3.24. "TANG YIXING SHAN HE L1ANGJIE TU" (MAP rating again. One is labeled "northern defense," and the other
OF THE TWO BOUNDARIES FORMED BY MOUNTAINS "southern defense." From the Lidai dili zhizhang tu (see fig.
AND RIVERS ACCORDING TO YIXING OF THE TANG). 3.23).
The map illustrates Yixing's idea that the mountains and rivers Size of the original: unknown. Photograph courtesy of Cao
of China formed two natural boundaries that defended against Wanru, Institute for the History of Natural Sciences, Academia
foreign invaders. The map shows two mountain ranges running Sinica, Beijing.
from west to east, converging at Taihua Shan and then sepa-
The ratio of image to text in traditional Chinese geo- scholarship, not to provide a sense of distances. The
graphic works varies, but it is not uncommon to find a problem with this line of argument is that the Guang
disproportionate amount of text. One example is the yutu also provides historical information about the places
Guang yutu mentioned earlier, which consists of a book it describes. It too could serve as a reference for historical
made up mostly of textual accompaniment to maps: scholarship and was itself based on such scholarship.
about one hundred pages of maps and about three The grid on the general map of the empire in the Guang
hundred pages of notes. Another is the Gujin xingsheng yutu raises questions of how seriously the grid was to be
zhi tu (Map of advantageous terrain past and present).44 taken as a gauge of distance. Each square on the map is
This map was printed from a woodblock in 1555, making supposed to represent 100 Ii, or about 55 kilometers.
it contemporaneous with Luo Hongxian's Guang yutu. Thus the distance from Xi'an to the coast directly east
Unlike the Guang yutu, however, the Gujin xingsheng is given as about 600 Ii, or about 330 kilometers. This is
zhi tu lacks an expressed scale as well as the square grid about one-third the actual distance. The squares on the
(plate 1). One might claim that this difference is related
44. This map is described in Ren ]incheng, "Xibanya cang Ming ke
to function. 45 The Gujin xingsheng zhi tu is heavily anno-
Gujin xingsheng zhi tu" (The Gujin xingsheng zhi tu printed during
tated with information regarding changes in toponyms the Ming and preserved in Spain), Wenxian 17 (1983): 213-21.
and administrative status, and this feature combined with 45. As Ren has done in "Xibanya cang Ming ke Gujin xingsheng zhi
its title might suggest that it was meant to aid historical tu," 214 (note 44).
60 Cartography in China
FIG. 3.25. MAP OF CHINA IN THE DA MING YITONG Reproduced courtesy of the Harvard-Yenching Library, Har-
ZHI. vard University, Cambridge.
Size of each page: 26 X 18 em. From Li Xian et aI., Da Ming
yitong zhi (completed 1461), prefatory chapter, unpaginated.
Yu ji tu (dated 1136) also represent 100 Ii, but the cor- standing or even a lack of utility. A case in point is the
responding distance on the stone map is about 2,500 Ii, maps that often accompanied reports describing water
which is about one-third longer than the actual distance. conservancy projects. These were produced by regional
Thus grid use on the Guang yutu and the maps based and local officials and drawn using brush and ink. Like
on it may be suspect as an indicator of the advanced state most traditional Chinese maps, they typically lack the
of scale mapping. In fact, not all mapmakers fully appre- square grid and scalar indications. This does not mean
ciated the utility of the square grid. In one edition of the measurements were not performed. Quantitative infor-
Guang yutu, Fuchs has pointed out, some squares of the mation such as distances and dimensions was often writ-
grid are elongated into rectangles, apparently in an ten on paper strips, which were then pasted to the map
attempt to fill wider pages. 46 Such elongated grids would (fig. 3.26). To appreciate the practical value of the map,
still be of use as a scaling device, but uniformity of scale one has to appreciate the complementary relationship
along the horizontal and vertical axes of the map would between image and text, which might appear on the map
be lost. itself or accompany it on separate sheets, as in a book.
The converse of what I said above about grids also 46. Walter Fuchs, The "Mongol Atlas" of China by Chu Ssu-pen
needs to be heeded: the lack of scale or a grid on a map and the Kuang-yu-t'u, Monumenta Serica Monograph 8 (Beijing: Fu
does not necessarily indicate a lack of geographic under- Jen [Furen] University, 1946), 21.
Reinterpreting Traditional Chinese Geographical Maps 61
v v
v
"
\J
v
" "
...
FIG. 3.26. WORD AND IMAGE ON A QING HYDRO- labels indicating distances were pasted on.
LOGICAL MAP (DETAIL). This map dating perhaps from the Size of the entire original: 23 x 382 cm. Courtesy of the Geo-
1870s depicts the water control system on the Grand Canal in graphy and Map Division, Library of Congress, Washington,
Shandong Province. It was drawn using watercolor and ink, and D.C. (G7822.G7N22 187-.S5 Vault Shelf).
In studying Chinese maps, we may also have to adjust rant, for pictorial representation in hydrological mapping
our norms of utility. Two nineteenth-century examples goes back at least as far as the Yuan (fig. 3.28). Such
can help clarify what I mean. The first is a hydrological examples also cannot be dismissed as the expression of
map depicting the Yangtze River and associated water- nonutilitarian or decorative intentions. Even the nautical
ways in Hubei Province (plate 2). The second is a map chart in the Wubei zhi represents coastal areas in a man-
of the city of Hangzhou (fig. 3.27). The hydrological map ner similar to landscape painting,47 as do other maps used
is largely planimetric and shows signs of measurement:
notes give distances between certain landmarks. The city
47. Researchers bound to the quantitative interpretation have taken
map is also generally planimetric. So far there is nothing pains to explain away this feature. Niu Zhongxun, for example, says
in these maps to undermine a quantitative interpretation, that the map distinguishes between smooth and rough islets, so that it
that is, not until one considers a curious feature common shows a greater concern for the scientific than the artistic. See Niu
to both maps. In certain places on these maps-in re- Zhongxun, " 'Zheng He hanghai tu' di chubu yanjiu" (Preliminary study
of Zheng He's nautical chart), in vol. 1 of Zheng He xia Xiyang /un-
presenting hills and buildings, for example-the mapmak-
wen;i (Collected essays on Zheng He's expedition to the Western
ers depart from a planimetric mode of representation and Ocean), ed. Zhongguo Hanghai Shi Yanjiuhui (Research Association for
represent things pictorially in a manner highly suggestive the History of Chinese Navigation) (Beijing: Renmin Jiaotong Chu-
of painting. These examples cannot be dismissed as aber- banshe, 1985), 238-48, esp. 243.
62 Cartography in China
FIG. 3.27. DETAIL FROM THE HANGZHOU CHENG TU represented pictorially. Shown here is the southwest corner of
(MAP OF HANGZHOU). This map of Hangzhou dates perhaps the city, where the landscape is dominated by hills.
from the nineteenth century. North is to the right. Streets are Size of the entire original: 63 x 94 em. Courtesy of the Geo-
drawn in plan, while buildings, temple grounds, and gates are graphy and Map Division, Library of Congress, Washington,
D.C. (G7824.H2A5 18--.H3 Vault Shelf).
for navigation (plate 3). In addition, coastal defense maps measures about 340 by 300 centimeters. It shows some
from as late as the nineteenth century are often drawn of the main streets of the city, along with moats, city
in a pictorial manner (plate 4 and fig. 3.29). What we seek walls, watchtowers, and gates. Because of the prominence
to understand in the following chapters is why. It is also given to features of military importance, some have inter-
curious to find these examples of pictorialism so late in preted the map as serving defensive purposes. 48 But this
the imperial period, when European methods supposedly interpretation is not convincing, since the map was carved
supplanted the traditional methods. outside the walls of the city even as the Mongols were
Conversely, a planimetric mode of representation does advancing. In addition, qualities often associated with
not always indicate practical function. A case in point is "military" maps-flexibility, portability, and secrecy-are
a map of ]ingjiang Fu Cheng (Prefectural City) carved on not among those of the map of ]ingjiang Prefecture. Its
the stone cliff of Yingwu Shan (a mountain north of engraving on stone suggests a commemorative function
Guilin) in 1271-72 (fig. 3.30). Construction of the city similar to that of the Yu ji tu.
began in 1258 to provide a defensive bulwark against If being engraved on stone is an indication of value,
Mongol forces. The city walls were completed before
1270, and the map was engraved shortly afterward. The 4R. See Chen et aI., Zhongguo gudai dilixue shi, 309 (note 17); and
city fell to the Mongols in 1277. The engraved surface Lu, Zhungguu dituxue shi, 152 (note 20).
Reinterpreting Traditional Chinese Geographical Maps 63
~
=
---
--
FIG. 3.28. MAP OF THE SOURCE OF THE YELLOW Size of the original: 13.5 X 17.4 cm. From Tao Zongyi, comp.,
RIVER. This map was based on information in the Heyuan zhi Nancun chuogeng lu (Notes taken by Nancun [Tao Zongyi]
(Treatise on the source of the Yellow River, 1315) by the geo- while at rest from plowing, 1366), Sibu congkan edition,
grapher Pan Angxiao (fl. thirteenth century). South is at the top. 22.3a-b.
moreover, then it is clear that planimetric maps like the Pictorial maps, of course, did not serve only practical
Yu ji tu and Hua yi tu were not the only ones considered purposes. On the Qing map of Hua Shan, a number of
wonhy of long-term preservation. Two maps of the "supernatural" features are emphasized; a palm print of
sacred mountain Hua Shan (Mount Hua) in Shaanxi Prov- a giant on the east peak, an image of a transcendent being
ince-one from the Ming and one from the Qing-iIIus- on the west peak, and between them a waterfall inter-
trate this point (figs. 3.31 and 3.32). Both make use of rupted by a cavern. The Qing map differs from the Ming
pictorial representation, and both have no expressed map in its depiction of man-made objects: as Munakata
scale-characteristics not valued by proponents of a has pointed out, it deemphasizes them through the use
quantitative tradition. But both show clear signs of intent of thin lines so they are overwhelmed by the natural
to convey knowledge of place: on both maps, place- landscape. 49 The intended result seems to be a mixture
names are given. In addition, pictorial representation was of awe and humility.
not necessarily perceived as detracting from the maps' The anifactual record suggests that scale mapping was
utility. An inscription on the Ming map says the map was not the primary concern of Chinese mapmakers, although
made as an aid to the visitor to the mountain. Bearing
this out, the map depicts climbing paths, steps cut into 49. Kiyohiko Munakata, Sacred Mountains in Chinese Art (Urbana:
the sides of cliffs, and bridges over gaps. University of Illinois Ptess, 1991),57.
64 Cartography in China
FIG. 3.30. MAP OF JINGJIANG PREFECTURAL CITY, of the city, listing its dimensions, recording the cost of labor
1271-72. A copy of the map carved on a cliff of Yingwu Shan, and material, and identifying those in charge of the project.
north of Guilin in Guangxi Zhuang Autonomous Region. North Size of the entire original: 300 X 340 em. Photograph courtesy
is at the top. Not shown on the copy is the upper part of the of Cao Wanru, Institute for the History of Natural Sciences,
map, which contains annotations describing the construction Academia Sinica, Beijing.
INSET 1
42° 0
f-I ---..L,~--,--J
o
26°
26°
IN 0
I A 126°
78°
..
18°
118°
FIG. 3.33. REFERENCE MAP FOR THE STUDY OF CHINESE CARTOGRAPHY. This map shows the location of places in
China mentioned in chapters 3-9.
APPENDIX 3.1 CHRONOLOGICAL LIST OF SELECTED MAPS,
FOURTH CENTURY B.C. THROUGH THE YUAN DYNASTY
Figure
Name of Map, and Number
Author (if known)a Date Dimension (cm) Medium Location in Text
aThese are the pre-Ming maps most often referred to. Illustrations bFor the location of place-names in this chapter and chapters 4-9,
of all the extant maps in this appendix, as well as those of numerous see fig. 3.33.
other maps, appear in Cao Wanru et aI., eds., Zhongguo gudai ditu ji
(Atlas of ancient Chinese maps) (Beijing: Wenwu Chubanshe, 1990-),
vol. 1, Zhanguo-Yuan (Warring States to the Yuan dynasty).
70