Imago Mundi

The International Journal for the History of Cartography

ISSN: 0308-5694 (Print) 1479-7801 (Online) Journal homepage: http://www.tandfonline.com/loi/rimu20

Can Lost Maps Speak? Toward a Cultural History of

Map Reading in Medieval China

Linda Rui Feng

To cite this article: Linda Rui Feng (2018) Can Lost Maps Speak? Toward a Cultural
History of Map Reading in Medieval China, Imago Mundi, 70:2, 169-182, DOI:
10.1080/03085694.2018.1450500

To link to this article: https://doi.org/10.1080/03085694.2018.1450500

Published online: 30 May 2018.

Submit your article to this journal

Article views: 13

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rimu20
 Can Lost Maps Speak? Toward a Cultural History of Map
Reading in Medieval China

LINDA RUI FENG

ABSTRACT: Despite the fact that virtually no Chinese maps have survived from the first millennium, it is
nonetheless possible to reconstruct a rich context associated with their production, use and perception from a
variety of written sources. Three cases from the Tang dynasty (618–907 CE) are presented in this article in
which the characteristics of the missing maps emerge through their associated texts, which have outlasted
them. These examples include two documents that once accompanied maps presented to the emperor and an
anecdote that refers to a map of the remote southern frontier. They demonstrate that the maps were designed
not only to encapsulate imperial territory but also to serve as guideposts for aspirational travel. They were
also perceived by their users as invitations to experiences both desirable and undesirable.

KEYWORDS: Medieval China, Chinese cartography, map of China, lost maps, transmission of maps, Tang
dynasty, writing about maps, reading about maps, memorials to the throne, map and text, map use,
aspirational travel, maps of empire, Huayi tu, Cao Song, Lingnan, tujing, Jia Dan, Yuan Zhen, Han Yu, Wei
Zhiyi.

In the history of Chinese cartography, a millen- facilitates a spatial understanding of things, con-
nium-long gap exists from which almost no maps cepts, conditions, processes, or events in the
survive. Dating from before the pre-imperial era and human world’.3
the establishment of the Han dynasty (206 BCE–220 The gap is the result of curtailed transmission
CE), maps made variously on metal, wood, paper and rather than any decline of interest in maps or cessa-
silk have been excavated from tombs (Fig. 1).1 From tion of cartographical practice. As will be seen below,
the twelfth century onward, a more robust carto- in this period devoid of artefacts, and especially dur-
graphic record survives, which consists of maps ing the Tang dynasty (618–907 CE), ample written
engraved on stele as well as printed atlases and texts speak to the intellectual, cultural and imagina-
regional gazetteers that began to be produced in tive histories associated with the contemporary use of
abundance as the commercial printing culture of spatial knowledge and its illustration in maps. The
the Song dynasty (960–1279 CE) expanded.2 reasons for the lack of the artefacts themselves may
However, for the period between the second and be broadly explained by the fact that in an era when
eleventh centuries, the surviving cartographic printing was still limited in use and both texts and
record is scant, even allowing for the highly inclu- images circulated as manuscripts, texts lent them-
sive view of maps proposed by J. B. Harley and selves more readily to transmission because they
David Woodward: ‘graphic representation that could be systematically copied by trained scribes

4Dr Linda Rui Feng is associate professor of premodern Chinese cultural studies, University of Toronto.
Correspondence to: L. R. Feng, Department of East Asian Studies, University of Toronto, 130 St George Street,
#14-087, Toronto, Ontario M5S 3H1, Canada. E-mail: lr.feng@utoronto.ca.

Imago Mundi Vol. 70, Part 2: 169–182

© 2018 Imago Mundi Ltd ISSN 0308-5694 print/1479-7801 online
https://doi.org/10.1080/03085694.2018.1450500
170 L. R. Feng Imago Mundi 70:2 2018

Fig. 1. This small fragment of a paper map is among the earliest map artefacts excavated from tombs
and predates the millennium-long artefactual gap. From Fangmatan tomb #5, 2.6 × 5.6 cm, dated to the
Western Han dynasty (early 2nd century BCE). Image from WikiCommons.

because of their modular nature. They were also demonstrated by the Tang-era geographical work
organized into the four highly durable traditional ‘Maps and Records of the Prefectures and Counties
bibliographic categories—the Classics (經), Histories of the Yuanhe Reign’ (元和郡縣圖志). This work,
(史), Philosophers (子) and Collections (集), and this completed in 813, originally contained both cartogra-
bibliographic structure allowed many texts to be phical and textual descriptions of the administrative
reconstituted from fragments.4 units of the Tang empire, but whereas most of the
In contrast, manuscript maps—either as loose texts have survived intact, all the maps had already
sheets or as part of a scroll—were not as easily vanished by as early as the Northern Song (960–
reproduced or reconstituted by editors and scribes.5 1126).8 Even after printing became more widespread
They were vulnerable assemblages. As Tan Qixiang and routine, maps remained precarious documents.
suggests, even without considering their sensitive The Song-dynasty polymath Shen Gua 沈括
nature as political and strategic tools, maps were (1031–1095), who created a map featuring com-
more difficult to reproduce compared with texts pass bearings, seemed to be anticipating his own
because of their irregular size and the fact that they map’s disappearance when he wrote ‘[Later] gen-
were more akin to paintings.6 Moreover, Cordell erations with the help of my recorded data . . . will
Yee has noted that prior to the Song dynasty maps be able to reconstruct the map showing the position
were rarely preserved because they were part of a of districts and towns without the slightest mistake,
‘documentary culture’ that consistently culled the even if the original copies should be lost’.9 In other
informational content and then dispensed with the words, Shen Gua expected the text of his carto-
original source materials, which were not considered graphic method to outlast his actual maps and
intrinsically valuable.7 Even state-sponsored maps had designed them with their eventual re-creation
vanished, to say nothing of maps born of more modest and re-constitution in mind.
ambitions. Faced with this millennium-long artefactual gap
The differential rate of attrition between map and spanning the medieval period of Chinese history
text during the pre-printing era is clearly (from after the Han dynasty to the end of the
Imago Mundi 70:2 2018 Map Reading in Medieval China 171

Tang), established histories of Chinese maps tend to cultural and political elite who either commis-
mention Tang-era maps only in passing and largely sioned map making or supervised it. Despite the
as precursors to later extant artefacts, and overlook vanished cartographical legacy from this middle
the opportunity to think about the maps’ concep- period of Chinese history, a close, contextualized
tion, design and impact. However, the contexts in reading of associated texts and descriptions of how
which maps were used, engaged with, and trans- their viewers used the maps yields insights into
mitted are still retrievable from extant writings contemporary cartographical practice. In other
about maps that are now lost, and from looking words, even lost maps can be made to speak.
closely at a range of sources associated with the
making, reading and circulation of spatial knowl-
edge. This is not the same as attempting to recon-
On Reducing Space
stitute lost maps; rather, it is to explore the Beginning in early imperial China, maps presented
intellectual, visual and imaginative histories asso- to a sovereign were often accompanied by writing
ciated with cartography. that anticipated the royal gaze and imparted an
Thinking about lost maps from premodern China is understanding of how the map was to be perceived
feasible at least in part because maps were routinely by at least its primary user. Writing in a formal and
read in conjunction with textual descriptions. As persuasive genre known as ‘memorial to the
many scholars have pointed out, traditional maps throne’, the minister or official reported on admin-
functioned as part of an interface between writing istrative affairs and articulated his positions vis-à-
and illustration and participated in what Cordell Yee vis imperial actions or deliberations, and early maps
calls a ‘representational complex’ that incorporated naturally figured in these addresses to the emperor.
prose, poetry and painting.10 Maps were seldom in During the Han dynasty, when the royal scion Liu
and of themselves sufficient as transmitters of knowl- An 劉安 (179–122 BCE) tried to dissuade his sovereign
edge or experience; they referred to other forms of from sending troops to the distant south, his rhetorical
representation. The terrestrial map in premodern approach rested on the discrepancy between the per-
China—ditu 地圖—was embedded within a larger ceived distance on a terrestrial map [ditu] and the
epistemological concept of tu 圖, which also included actual distance when military action was concerned.
illustrations, charts, diagrams, paintings, catalogues He cautioned Emperor Wu 漢武帝 (r. 141–87 BCE),
and prophetic texts.11 Early medieval writers, who whose capital was located in the northern plains of
were aware of the term’s polysemic nature, did not China, not to be falsely reassured by the proximity
feel the need to make explicit distinctions between and flatness between locations in the geographically
the different types of tu.12 The capacious concept of tu treacherous south and reminded him that
has led scholars to define the term as a functional
When one investigates on a map [ditu] the mountains,
category referring to images that ‘encoded technical rivers, and strategic defensive positions, their distance on
knowledge’ and can be thought of as ‘templates for the map from each other is only a couple of inches, but
action’.13 the physical space between them is hundreds or thou-
Chinese terrestrial maps, therefore, lend them- sands of leagues. The obstructing cliffs and the woods
and bush cannot be adequately displayed on a map.14
selves to being considered and analysed through
other forms of descriptions and through indirect In a discussion of this passage, Garrett Olberding
evidence. This opportunity is particularly germane suggests that Liu An may be denying ‘the very
to thinking about maps that have been lost. To efficacy of a map, any map, in strategic considera-
examine the cultural logic inherent in the percep- tions’ without personal knowledge of the terrain,
tion of maps and the contemporary ‘mental hori- knowledge that Liu An believed he possessed but
zon’ of both mapmakers and map users who the emperor did not.15
engaged with and handled maps, I first focus on In the address quoted above, the compressed
writing by two high-level officials in which they cartographic scale seen by the implied royal gaze
justify creating maps for the emperor, describe the was presented as a kind of deception or entrap-
map-making process and envision how the maps ment. However, by cautioning his sovereign not
are to be viewed. Then I move to consider descrip- to equate the easy roving of the viewer’s gaze
tions of map users and the consequences of map over the map with easy travel on the ground, Liu
viewing. What follows is an analysis of the traces of An was in effect teaching the emperor the right
this text-pictorial interface, which consisted of tes- way to read a map: how to relate cartographical
timonies about making and viewing maps from the signs to the physical terrain, and how to calibrate
172 L. R. Feng Imago Mundi 70:2 2018

the magnitude of the depicted topography against addresses of this kind—the map was described in
that of ‘actual’ experience. terms of its dimensions and function.
An important mapmaker active in the Jin 晋
dynasty (265–316), in contrast, highlighted the Respectfully [I have] commanded artisans to paint this
same phenomenon as a touted feature rather than scroll of ‘Maps of Chinese and Barbarian Territories
within the Seas’. It is three zhang wide, three zhang
a pitfall. The ‘Map of the Territory Encompassing and three chi in height. As for its scale [lü], one cun is
the Tributes of King Yu’ (禹貢地域圖) created by equivalent to one hundred li [of land].21 [The map]
Pei Xiu 裴秀 (224–271) has long been lost, but Pei distinguishes the civilized from barbarians; it situates
Xiu wrote a preface for the map he presented to the the high peaks and large rivers. It shrinks the four
frontiers onto [a piece of] fine white silk; it marks
emperor, and in it he articulated six theoretical out, in painting, the hundred prefectures. The cosmos
principles for map making.16 When these principles may be wide, but the map unfurled does not fill the
are followed, he explained, no topographical detail courtyard. Anywhere that can be reached by boat or
carriage can all be taken in by a single glance of the
would escape being included.
eye. [謹令工人畫海內華夷圖一軸, 廣三丈, 從三丈三尺,
率以一寸折成百里.◦ 別章甫左袵, 奠高山大川; 縮四極
Even if separated by high mountains and vast seas and 於纖縞, 分百郡於作繪.◦ 宇宙雖廣, 舒之不盈庭; 舟車
by the great distances and unfamiliar directions of the 所通, 覽之咸在目.]22
most remote land, whether it is the routes for climbing
and descending or for following elaborate turns and
detours—all this can be identified and determined. Here Jia Dan highlighted the four crucial actions
Once the principle of zhunwang is properly applied, the he expected his map to perform: to distinguish (別),
straight and the curved, the far and the near—none of to set apart or separate (分), to situate (奠), and to
this can hide their forms from us. [故雖有峻山鉅海之隔,
絕域殊方之迥, 登降詭曲之因, 皆可得舉而定者.◦ 準望之 shrink (縮). He referred to Yu 禹, the legendary
法既正, 則曲直遠近無所隱其形也.]17 sage-king of antiquity who tamed a primordial
flood and gave order to the post-diluvian world,
Unlike Liu An, with his caution that cartographic as narrated in one of the earliest canonical texts—
representations could be deceiving, Pei Xiu confi- the chapter ‘Tribute of Yu’ (禹貢) of the Book of
dently declared exactly the opposite, that cartogra- Documents (書經). In this way Jia Dan established
phical techniques give map users improved vision. hallowed precedence for his own cartographical
Both Liu An and Pei Xiu wrote (to their respec- endeavour.23 The verb dian 奠, in the context of
tive sovereigns) about reading and making maps, establishing the major mountains and rivers, was a
and this practice continued into the Tang dynasty. direct echo from those accounts of articulating and
The prime minister Jia Dan 賈耽 (730–805), famed cataloguing geography.
for his accomplishments in geography or, more Out of Jia Dan’s four key map-making verbs, the
literally, ‘terrestrial patterning’ (地理學), was com- most spatially and visually significant is ‘to shrink’
missioned by the emperor to compile a country- (縮). By ‘making one cun equivalent to one hun-
wide map (國圖). Accordingly, in the year 801, he dred li’, Jia Dan was directing the emperor’s atten-
presented Emperor Dezong 德宗 (r. 779–805) with tion to the cartographer’s act of downsizing the
an enormous map painted on silk and entitled world (宇宙) by diligent, artful shrinkage—to a
‘Map of Chinese and Barbarian Territories within scale of, in modern terms, 1:1,500,000.24 Directed
the Seas’ (海內華夷圖).18 This monumental map toward the emperor who was the intended user of
has long been lost, but unlike the map of Pei Xiu this map, Jia Dan’s rhetoric of reduction was one of
of earlier times, which the emperor had ‘kept in sovereignty made more powerful through distilla-
secret archives’, Jia Dan’s map appears to have tion and visual representation.
been seen by appreciative members of the cultural In fact, from the ninth century, literary responses to
elite, who referred to it in a variety of written texts, maps and map making had already developed a
commemorative prose and poems.19 One of the coherent body of highly allusive language to describe
earliest extant maps of the Chinese empire, carved this act of shrinking. One such allusion—that of ‘piling
in 1136 on a stone stele, also seems to have drawn up the rice grains’—derived from the feat of the Han-
substantially from the major features of Jia Dan’s dynasty military strategist Ma Yuan 馬援 (14 BCE–49
map (Fig. 2).20 CE), who is said to have set up a small mound of rice
In the memorial that introduced the map to the grains to simulate the topographical features of moun-
emperor, Jia Dan explained the cartographic prac- tains while discussing military campaigns.25 Jia Dan
tices involved in its production and the map’s himself used the phrase in his memorial, at the end of
intended purpose. After first praising the ruler’s which he compared his work with those of his pre-
sage governance—as was conventional for decessors and referred to Ma Yuan’s ‘rice-piling’, an
Imago Mundi 70:2 2018 Map Reading in Medieval China 173

Fig. 2. Ink rubbing of the map Huayi tu [Map of the Chinese and Barbarian Territories] carved in stone in 1136, more than
300 years after Jia Dan’s silk map of the Tang empire. 79 × 80 cm. North at the top. Stone-carved maps were durable and could be
easily reproduced through such ink rubbings. The silk map of Jia Dan, now lost, would have covered more non-Chinese territories
and allowed for more details to be shown. Nonetheless, the stele displays features that probably evolved from Jia Dan’s map.
(Reproduced with permission from Special Collections, Fine Arts Library, Harvard University.)

act of miniaturization imbued with the power of be sure if the writer Jiang Fang 蔣防 (fl. 800s) was
insight.26 himself involved in map making, the rhyme-prose
Other literary works from the subsequent decades continues with an exposition on the essential advan-
further expound on this idea of miniaturizing space tages of using the miniscule to represent the large.27
and extend its applications beyond cartography. An
epideictic rhyme-prose (賦) titled ‘Rhyme-Prose on Although the barbarian lands [extend] for thousands
Making a Mountain from Gathered Rice Grains’ (聚 of li, they can be encircled within inches of the gaze;
there may be myriad arrays of hills and ridges, in
米為山賦) draws on a similar repository of associations actuality they do not surpass but a few grains [of
and considers the shifts in scale by which space is rice]. [雖蠻貊千里,◦ 亦可圍於寸眸.◦ 岡巒萬重.◦ 誠不
miniaturized in representation. Although we cannot 過於數粒.]28
174 L. R. Feng Imago Mundi 70:2 2018

Here the text is paying homage to an imagined act Literary invocations of ‘earth shrinking’ in this
of map making in the remote historical past and, as sense also refer to the imagined instantaneous
in Jia Dan’s memorial, the reduction of scale is movement across great distances. Poetry of friend-
described as a radical accomplishment. Jiang Fang ship invokes this art as a wishful way to reunite
continues: with those from afar. A poem written by the poet-
statesman Yuan Zhen 元稹 (779–830), for example,
When one shrinks the disposition of the terrain into a
laments being apart from his good friend Bai Juyi
pinch of soil, who can say it appears small? When one
lodges the shapes of mountains within the scope of a 白居易 during the new year, and called upon the
handful of grain, what worry is there that accomplish- art of earth-shrinking:
ments be lacking? [縮地勢於撮土之間.◦ 孰云見小.◦ 備
山形於握粟之內.◦ 何慮功虧.]29 We come upon the New Year simultaneously, yet
cannot enjoy it together 同受新年不同賞
A similar paean to the reduction of scale is found What is one to do, with no way to shrink the
earth? 無由縮地欲如何33
in another poem, ‘On Viewing “Map of Chinese
and Barbarian Territories”’ (觀華夷圖) by Cao The epistemological possibilities of ‘earth shrink-
Song 曹松 (fl. 900). Although we cannot be sure ing’, which the para-text of Jia Dan’s empire-wide
that the map he saw was the same as that sub- map merely gestured at, were more fully expressed
mitted by Jia Dan, the map of the poem’s title as imaginative extensions of cartography in the
suggests that it was also one encompassing the poetry and rhyme-prose of the ninth century.
empire. By the third couplet, the poet notes that
the cartographer has reduced the major mountains Reconstituting Experience from Maps
to a scale of inches (分寸), and the oceans to simi-
For the resources it required, Jia Dan’s courtyard-
larly small units of volume (斗升), making possible
sized map of the empire was both rare and extra-
a panorama of territory that signals benign rule (晏
ordinary, and likely had to be viewed in situ.
寧). The poem begins its first couplet with what the
Another cartographic genre from medieval China,
map does, and what it represents:
tujing 圖經 (map guides), provided more ordinary
A brush laid down, it surpasses the art of shrinking the and portable instantiations of mapping. Although
earth, 落筆勝縮地 not yet fully codified in format during the Tang era
The map unfurled, we come face-to-face with peace. as it would be in the subsequent Song dynasty,
展圖當晏寧
Central China belongs to a noble position, 中華屬貴分
Tang-era tujing were largely annotated charts in
The distant tribes—what stars to prognosticate upon? which text and picture were assembled into a
遠裔占何星 chorography, and which disseminated localized
Across minute inches, we discern the major geography that included regions in the far reaches
Marchmounts, 分寸辨諸岳 of the empire.34
With just bushels of water shown, we survey the Four
Seas. 斗升觀四溟 A generation after Jia Dan, in 821 Yuan Zhen 元稹
Places where I’ve never been, I’ve long suspected— submitted his own maps to the newly enthroned
長疑未到處 Emperor Muzong 穆宗 (r. 821–825).35 The accom-
Are as if I’ve already passed through, each and each. panying memorial to the throne offers a rare explana-
一一似曾經30
tion of how it was to be handled by its intended user.
In this literary response to cartography, Cao Song’s Yuan Zhen related how
poem compares map making to an esoteric Daoist art, On the second day of this month, your servant sub-
‘earth shrinking’ (縮地). A biography of a Daoist mitted, in one sheet, a map [tu] of the west and north
practitioner details the extent of this skill as part of a of the capital. Of the topography of mountains and
repertoire of arcane knowledge: ‘[Fei Chang]fang has rivers, nothing large or small has been omitted. Yet I
worry that the map’s size would inconvenience your
a divine skill and can shrink the meridian of the majesty’s viewing; being hung upon the wall of a hall,
earth. [Even if something is] a thousand li away, it it may tax your majesty’s body with the need for
appears as if it were right before your eyes. By letting looking up. I searched through the illustrations and
go [of his hold], [the earth’s meridian] loosens and texts ancient and new, and compiled a set of map
guides [tujing] of the west and north of the capital, in
returns to its previous state’. [房有神術.◦ 能縮地脉. four scrolls total. It is my hope that upon [your
千里存在.◦ 目前宛然.◦ 放之復舒如舊也.]31 ‘Earth- majesty’s] reclining-mat, you can lean upon a pillow
shrinking’ was a concept that had become embedded and gaze at the prefectures and towns. As your
majesty makes imperial outings, you might lean
in the literary imagination; in social poetry addressed
upon a horse and the mountains and rivers would all
to friends who were Daoist recluses, poets were fond be there. [右, 臣今月二日進《京西京北圖》一面, 山川
of citing this art as a marker of transcendence.32 險易, 細大無遺.◦ 猶慮幅尺高低, 閱覽有煩於睿鑒; 屋壁
Imago Mundi 70:2 2018 Map Reading in Medieval China 175

施設, 俯仰頗勞於聖躬.◦ 尋於古今圖籍之中, 纂撰《京 which the user was asked to simulate accompany-
西京北圖經》, 共成四卷.◦ 所冀袵席之上, 欹枕而郡邑 ing someone on a long journey (Fig. 3).
可觀; 游幸之時. 倚馬而山川盡在.]36
In particular, Yuan Zhen’s tujing for the emperor
As a statesman on a rising career path, Yuan Zhen was an invitation to a specific kind of travel experi-
would have been aware of the power of maps to ence; it placed the viewer on a specific route, by
incite conversation and capture attention. The pre- providing detailed geographical information in a
vious emperor, Xianzong 憲宗 (r. 805–820), when separate scroll of text. This image-text interface
discussing policy matters over the contentious Hebei was emblematic of the tujing genre, both during
region north of the capital, had unfurled a map of the Tang dynasty and as it continued to evolve in
this region on the wall of the Yutang 浴堂 Hall in the the early Song dynasty, when it began to be
inner palace and pointed to his trusted minister- replaced by the now more familiar, and more codi-
adviser Li Jifu 李吉甫 (758–814) when doing so.37 fied, gazetteer (地方志). Despite the lack of com-
Yuan Zhen’s memorial shows that he fully antici- plete tujing artefacts from the Tang dynasty,40
pated the actions of his royal map reader and had textual fragments compiled and preserved in the
accordingly designed a map-viewing experience in tenth-century compendium Taiping yulan 太平御
which the same geographical content was presented 覽 [Imperial Anthology from the Taiping Reign]
in two formats: a sheet map (tu) and a set of map suggest that the content of Tang-era tujing ranged
guides (tujing) in scrolls. The former could be hung from records of local history and biographies of
upon a wall, and the latter was portable both local worthies to geographical, agricultural, hydro-
indoors and outdoors, and could be held in the logical and topographical descriptions.41
hand and kept within reach. If Yuan Zhen’s memorial showed how an emperor
Although both Jia Dan and Yuan Zhen miniatur- might have used tujing, then another glimpse into its
ized the imperial territory for their respective sover- use outside the imperial palace may be gleaned from
eigns, Yuan Zhen had a more distinctive objective, the journal of the Japanese Buddhist monk Ennin
that of the creation of a personalized and annotated (794–868), who travelled in China between 838 and
itinerary for the emperor. Continuing his memorial 847 in search of Buddhist scriptures and left a detailed
to the throne, Yuan Zhen elaborated on another diary of his life on the road.42 Among his accounts,
advantage of the tujing format: one mention of a map took place in 840, when Ennin
was in the area of modern-day Shandong (a northern
Furthermore, because Princess Taihe is [about] to be coastal province that included the lower reaches of
married [to a Uigher khan], I fear that your majesty will
fret over the route of her journey. [Therefore] I have
the Yellow River). He recorded that he sought out
recorded the eateries, lodgings and watering places, locals for oral directions, which were given as cardi-
stretching from the region north of Tiande Fortification nal directions and distances. At one point on the way
to the military encampment of the Uighers. They are toward the Spring Water Temple (醴泉寺), Ennin
appended in the tujing. A separate text is copied, and I
submit it respectfully along with the ‘Tujing Preface’. [又 and his party took the wrong path at a fork in the
太和公主下嫁, 伏恐聖慮念其道途, 臣今具錄天德城已北, road and became lost.43 However, it was only after
到回鶻衙帳已來, 食宿井泉, 附於《圖經》之內,◦ 并別寫 his arrival at the temple that he mentioned a ‘terres-
一本, 與《圖經序》謹同封進.] 38 trial chart’ (ditu) in his diary, and then only because it
This passage implies that with the tujing in hand, explained the origin of the name of a nearby peak.44
the emperor would be able to follow the journey of The chart seems to have been cited to lend cre-
his younger sister, stop by well-informed stop, on dence to extraordinary events and as a record of
her westward journey beyond the frontier to the the provenance of the name of a landmark. From
territory of the Uighers.39 this record we may deduce that the ditu Ennin
Whereas Jia Dan’s monumental map of the referred to in his diary was probably a kind of tujing
empire was designed as a miniaturization of imper- that was kept on the premises and devoted to the
ial territory, Yuan Zhen’s hand-held scroll was temple and its surroundings, and that the geogra-
designed for a set of actions to be unfolded from phical knowledge disseminated was not so much to
it. This viewing experience would allow the aid navigation as to supplement local knowledge
emperor to advance with the traveller along a once the traveller had arrived.
defined itinerary rather than be limited to gazing Besides foreign pilgrims like Ennin, there were yet
at, for example, the princess’s portrait. A static other kinds of travellers at the time. Chinese men of
image does not lend itself to the kind of incremen- letters were frequently moving back and forth
tal mobility derived from Yuan Zhen’s map, from between the empire’s heartland and its peripheries,
176 L. R. Feng Imago Mundi 70:2 2018

Fig. 3. Detail of the Huayi tu [Map of the Chinese and Barbarian Territories] (see Fig. 2) showing the region northwest of
the Tang capital Chang’an (lower right). This is the region that extended into the territory of the Uighers and was the focus
of Yuan Zhen’s tujing as submitted to Emperor Muzong tracing out the itinerary of Princess Taihe. (Reproduced with
permission from the Special Collections, Fine Arts Library, Harvard University.)

and their uses of various forms of tujing left literary rotations between places and helped them to get to
traces in the numerous poems produced for social know those places far from their hometown.46 Here
occasions. At about the same time as Yuan Zhen’s Han Yu shows us that he understood how to shape his
address to the throne, in 820, the statesman-poet Han own experience of reading, travel and anticipation. In
Yu 韓愈 (768–824) was about to arrive in Shaozhou this case, the regional tujing allowed for not only
韶州 (present-day Guangdong 廣東 Province) in the aspirational travel, as Yuan Zhen envisioned for
Lingnan 嶺南 region of southern China, and he dis- Emperor Muzong to do, but also for actual travel
patched a poem to a friend serving as an official there. enhanced by knowledge. Han Yu wanted to carry
In this poem, Han Yu asked to borrow a tujing of that with him the names and descriptions provided by the
region so that he would be able to name its famous texts and illustrations in order to corroborate his first-
mountains and rivers. hand experience with the ‘places of delight’ (佳處)
anticipated from his reading. The portability of the
I’ve long heard of the landscape of the Serpentine tujing meant that it was now possible for a traveller
River, 曲江山水聞來久
Yet not knowing the place-names, it hinders my like Han Yu to combine the experiences of prior hear-
visiting. 恐不知名訪倍難 ing (聞), arriving and visiting (訪), and looking (看).
I’d like to borrow a tujing as I enter its bounds,
願借圖經將入界
So that I can open it up to look, at each place of A Case of a ‘Prophetic’ Map
delight. 每逢佳處便開看45
Han Yu’s poem portrayed Lingnan as scenic, but
Some modern scholars believe that the increased the perception of this region during this era was not
use of tujing during the Tang era could have been always idyllic. At a time when the most coveted
because they served bureaucrats on their frequent official positions were located in and near the
Imago Mundi 70:2 2018 Map Reading in Medieval China 177

imperial capital Chang’an 長安 in the central plains, government branch ‘in charge of the maps/charts
Lingnan at the southern edge of the empire was of all under heaven’.56
frequently associated with harsh and irreversible Although the anecdote shows Wei Zhiyi to be a
political exile.47 In the collective imagination, cowardly figure, it also makes it clear (as it would
Lingnan had long been associated with the antith- have been clear to contemporary readers) that he
esis of civilization and power, a remote place with a was a routine viewer of maps in his official role,
pestilential climate. was familiar with the administrative functions of
Han Yu had experienced exactly this kind of maps, and was likely conversant with their produc-
severe demotion a year before he wrote his poem tion. Wei Zhiyi had a high degree of map literacy
when, in 819, he was reassigned from the capital to and was a ‘primary user’ of maps, to borrow
a post in Chaozhou 潮州 (even further south in the Catherine Delano-Smith’s model for identifying
Lingnan region) as punishment for his blunt criti- types of map readers.57 This role makes it all the
cism of Emperor Xianzong’s patronage of more notable that in this context, he read a map of
Buddhism.48 Upon his enforced departure from Lingnan differently from other maps, and ascribed
the capital, Han Yu viewed his plight as bleak; he to it an extramundane power.
addressed a parting poem to his nephew saying The rest of this anecdote makes explicit his mode
that he was assuredly to die on the ‘miasmic of magico-religious reading and, ultimately, sug-
banks’ in the Lingnan region.49 The appreciation gests that such a reading of the Lingnan map
he expressed for Lingnan in the poem cited above might have been justified.
came after he was reprieved from banishment in
820, by which time he was passing through When [Wei Zhiyi] returned as a chief minister, [he saw]
Shaozhou on his way to a better assignment.50 that on the north wall of his official quarters a map was
displayed. For seven or eight days he would not
In a political context such as this, Lingnan as a approach it to look. When he finally did, it turned out
geographical location cannot be separated from to be a map of Yazhou [in Lingnan]. He thought this ill-
what Edward Schafer calls the ‘spectre of banish- omened and loathed it, but could not bring himself to
speak of it directly. When he ended up getting banished,
ment’ that haunted ambitious high-level officials in
it was indeed to Yazhou. [至拜相還, 所坐堂北壁有圖, 不
the Tang dynasty.51 However, an anecdote from a 就省七八日.◦ 試就觀之, 乃崖州圖也.◦ 以為不祥, 甚惡
different type of record puts Lingnan into a carto- 之, 憚不能出口.◦ 至貶, 果得崖州焉.]58
graphic context, and an otherwise ordinary map of
this region is mentioned only through the extraor- Yazhou 崖州 was on the island of present-day
dinary reaction of its user. Hainan (Fig. 4 and Plate 1). Located off the
In a chronicle of the short reign of Emperor Lingnan mainland and even more remote than
Shunzong 順宗 (r. 805) titled Veritable Records of the Shaozhou to which Han Yu was banished, this
Shunzong Reign (順宗實錄), one anecdote features place was truly beyond the pale in the contempor-
Wei Zhiyi 韋執誼 (769–814), a younger contempor- ary geographical imagination. The anecdote ends
ary of Jia Dan who served in a number of high court with the verification of the map being an ill
positions.52 Wei Zhiyi was part of a political faction omen, although it doesn’t mention the fact that
that, in 805, suddenly and spectacularly collapsed.53 Wei Zhiyi eventually died in this southerly place
He was described in the Veritable Records as being of banishment.59
terrified of maps of the Lingnan region even before In the subsequent century, this theme of predic-
his major political ally had fallen from power.54 tive omen became the salient characteristic by
which this anecdote would become included and
Wei Zhiyi thought himself lowly, and avoided uttering reproduced among the larger corpus of medieval
the names of the prefectures and counties in Lingnan, as narratives: when it was compiled into the tenth-
if they were taboo. When he was the Bureau Director, he
went with other officials to view maps in the Bureau of century anthology of narratives, the Taiping guangji
Operations. Whenever they came upon a map of 太平廣記, it was grouped under the category head-
Lingnan, Wei would always order it to be taken away, ing ‘Predetermined Lot’ (定數), together with other
and refused to look at it by shutting his eyes. [執誼自卑,
narratives on a similar theme of impending fate
嘗諱不言嶺南州縣名.◦ 為郎官時, 嘗與同舍郎詣職方觀
圖, 每至嶺南圖, 執誼皆命去之, 閉目不視.]55 revealing itself.
Strictly within the context of the present article,
The Bureau of Operations (職方) was a high-level however, Wei Zhiyi and his map are noteworthy in
agency in the Ministry of War that was responsible a cartographical context for two reasons. First, a
for maintaining military maps and manning fron- geographically specific map was interpreted as a
tier fortifications and signal systems. It was also the gateway to travel and displacement, and second,
Fig. 4. Qiong Jun di yu quan tu 瓊郡地與全圖 [Complete Map of Qiong Prefecture]. Nineteenth century.
184 × 93 cm. South is at the top. The map shows Hainan Island, the southernmost part of Lingnan during
the Tang dynasty, to which Wei Zhiyi was exiled at the end of his political career. Although this map
postdates the anecdote about Wei Zhiyi by nearly ten centuries, it is noteworthy that even at this late
date, the depiction of the region by an unknown Qing-dynasty cartographer-artist clearly shows a place
‘beyond the pale’, populated by an indigenous tribe with ‘primitive’ customs. Library of Congress,
Geography and Map Division, Washington, DC, G7823.H25E62 1836 .H3.
Imago Mundi 70:2 2018 Map Reading in Medieval China 179

an otherwise prosaic map evoked an extramun- the early medieval period, which is similarly lack-
dane reading that coexisted with the workaday ing in cartographic artefacts. This direction is one
reading of the same map. Wei Zhiyi was not being toward which more work can be done and has
merely superstitious when he identified in the map broader implications for understanding how spatial
an ontological reality of its own; the anecdote knowledge was constructed and imparted.
would have lost something of its cultural logic if Although the maps alluded to in these sources are
what Wei Zhiyi avoided seeing were, for example, lost, what we can glean from their cartographic
an object from Lingnan rather than a map of it. The contexts shows that even among the cultural and
underlying logic driving Wei Zhiyi’s paranoia was political elite of the Tang, there was no single way
this: a map represented a visual portal into another for viewers to participate in the experiences of
place in a way an object could not. mobility and order as encoded by the maps.
The contemporary reader for whom the anec- The heterogeneity in map-viewing demonstrates
dote was written, and who would have had an the capaciousness of these images as tu, which during
education similar to Wei Zhiyi’s, would have recog- the Tang dynasty was far from codified or easily
nized in his perceptual logic something of another categorized, in contrast to genres of writing in the
kind of tu related to terrestrial maps, namely, Chinese intellectual tradition. By closely examining
Daoist visualization diagrams (存思圖). These these para-cartographic sources, we are able to move
were pictorial illustrations designed to aid the beyond asking whether lost maps can speak, and
adept in visualizing and merging the microcosm of toward asking how they are able to do so.
his body and the outer cosmic world, and were part
Acknowledgements: An early version of this paper was pre-
of a Daoist visual culture that, to quote Franciscus
sented at the 26th International Conference on the
Verellen, featured designs and diagrams ‘endowed History of Cartography (ICHC) in Antwerp, Belgium,
with transformative and efficacious powers’.60 Like 2015. The author is especially grateful for input and sug-
the aforementioned arcane art of ‘earth-shrinking’ gestions from Tom Keirstead, Yan Liu, Michael Nylan,
Garrett Olberding, Hyunhee Park, Stephen Roddy, Anna
alluded to in elite literary culture, magico-religious Shields, Nicolas Tackett, Paula Varsano, Chet Van Duzer,
diagrams such as these never stayed strictly within Curie Virag, and Ao Wang. This work was supported in
the bounds of religious practice. part by a Connaught New Researcher Award from the
It is no coincidence that another narrative from University of Toronto.
this same time described a homesick sojourner who Manuscript submitted May 2016. Revised text received
travels back home by gazing at an empire-wide map September 2016.
with the help of a Daoist practitioner in a temple.61
This narrative, albeit more fantastical in tenor, cor- NOTES AND REFERENCES
roborates Wei Zhiyi’s quasi-magical reading of the 1. The excavated maps include a bronze map of the
map and demonstrates that in the contemporary graveyard plan (323–315 BCE) for King Zhongshan 中山;
imagination, maps could serve as a visual conduit wooden and paper maps excavated from tombs in
Fangmatan 放馬灘, and the silk maps from Mawangdui
and as an invitation to experiences both desirable (to
馬王堆 tombs (before 168 BCE). See Cao Wanru 曹婉如
return home) and loathsome (to undergo exile). et al., eds., Zhongguo gudai ditu ji 中國古代地圖集 (戰國—
This anecdote about Wei Zhiyi is usually cited by 元) [Collected Ancient Maps of China, from the Warring
modern scholars as a typical reaction of Tang era States to the Yuan Dynasty 476 BCE–1368 CE], vol. 1
(Beijing, Wenwu, 1990), 1–17.
elite towards exile, but until now no study has
2. Maps from the 12th century include the atlas ‘Handy
pointed out its implications for the ways in which Geographical Maps throughout the Ages’ (歷代地理指掌
cartographic images were construed. The fact that 圖) attributed to Shui Anli 稅安禮 (1130s); a map on a
this anecdote was not intended as a commentary stele, ‘Map Tracing the Tracks of Yu’ (禹跡圖), dated 1136.
See Cao et al., Zhongguo gudai ditu ji (note 1), 31–34.
on mapping practice only makes it more compel- 3. J. B. Harley and David Woodward, eds., The History of
ling when it demonstrates the imaginary and epis- Cartography, vol. 2, book 2: Cartography in the Traditional East
temological extensions of cartography. and Southeast Asian Societies (Chicago, University of Chicago
Press, 1987), xvi. Rare exceptions in this cartographic gap
include a 10th-century mural map found in the Mogao
The cases discussed in this essay represent only a grottoes of Dunhuang, depicting the peaks and temples of
fraction of various para-cartographic sources that Wutai Mountain and described in, for example, Zhao
can be coaxed into revealing perceptions of maps Shengliang 赵声良, ‘莫高窟第61窟五台山图研究’ [A Study
and map making in the pre-printing era. Although of ‘Wutai shan tu’ in Mogao Grottoes Cave 61], Dunhuang
Yanjiu 敦煌研究 4 (1993): 88–124. Another survivor is the
the examples analysed are primarily from the Tang, early-Song engraved map showing the Tang dynastic capital
this mode of inquiry can be equally applicable to of Chang’an attributed to Lü Dafang 呂大防 (1027–1097).
180 L. R. Feng Imago Mundi 70:2 2018

See Hiraoka Takeo 平岡武夫, Tangdai de Chang’an yu of New York Press, 2012), 192–93. See also a discussion of
Luoyang ditu 唐代的長安與洛陽地圖 [Maps of Tang- this passage in Garret Olberding, ‘Movement and strategic
Dynasty Chang’an and Luoyang] (Shanghai, Shanghai mapping in early imperial China’, Monumenta Serica 64:1
guji, 1991), 35–44. (2016): 23–46.
4. As Glen Dudbridge has observed, this traditional four- 15. Olberding, ‘Movement and strategic mapping in
part bibliographical classification continues to shape early imperial China’ (see note 14). Liu An was the prince
library collections, catalogues and writing about tradi- of Huainan 淮南, which was in the area of conflict in
tional literature: ‘It would be hard to think of any other question here: the emperor was planning to defend the
institution set up by imperial China which has endured so kingdom of Nanyue 南越 against the neighbouring
well’. Glen Dudbridge, ‘A thousand years of printed nar- Minyue 閩越 (ibid, 41).
ratives in China’, in his Books, Tales and Vernacular Culture: 16. See Needham, ‘Geography and cartography’ (note 9),
Selected Papers on China (Leiden, Brill, 2005), 3–14 at 4. 538–39, for a discussion of one of these principles, fenlü 分
5. For more detailed discussion on the identities of map- 率, which Needham translates as ‘graduated divisions’.
makers and map copyists during the Song and Ming (1368– 17. Fang Xuanling 房玄齡, Jin shu 晉書 [History of the
1644) dynasties, see Pan Sheng 潘晟, Ditu de zuozhe jiqi yuedu: Jin] (Beijing, Zhonghua shuju, 1991), 35.1040. For an
yi Song-Ming wei hexin de zhishishi kaocha 地圖的作者及其閱 English translation that differs from mine, see Needham,
讀: 以宋明為核心的知識史考察 [Map Authors and Their ‘Geography and cartography’ (note 9), 540.
Interpretations: Investigating a History of Knowledge in the 18. Liu Xu 劉昫, Jiu Tang shu 舊唐書 [Old History of the
Song and Ming Dynasties] (Nanjing, Jiangsu renmin, 2013). Tang] (Beijing, Zhonghua shuju, 1975), 138.3785–6. Jia
6. Tan Qixiang 譚其驤, ‘序’ [Preface], in Cao et al., Dan was also known for his account of trade routes
Zhongguo gudai ditu ji (see note 1), 1–4 at 3. between China and the Islamic world, which survives as
7. Cordell D. K. Yee, ‘Concluding remarks: foundations ‘The route to the foreign countries across the sea from
for a future history of Chinese mapping’, in Harley and Guangzhou’ (廣州通海夷道) in the geography section of
Woodward, Cartography in the Traditional East and Southeast the ‘New History of the Tang Dynasty’ (Xin Tang shu 新唐
Asian Societies (see note 3), 228–31 at 231. 書). See Hyunhee Park, Mapping the Chinese and Islamic
8. See James M. Hargett, ‘Song dynasty local gazetteers Worlds: Cross-Cultural Exchange in Pre-Modern Asia
and their place in the history of Difangzhi writing’, Harvard (Cambridge, Cambridge University Press, 2012), 31–34.
Journal of Asiatic Studies 56:2 (1996): 405–42 at 411–12. 19. Sikong Tu 司空圖 (837–908) explicitly mentions pos-
9. Shen Gua 沈括, Xin jiaozheng Mengxi bitan Mengxi bitan sessing and viewing a copy of a map of the same title and
buzheng gao 新校證夢溪筆談 夢溪筆談補證稿 [Newly associated with Jia Dan. Sikong Tu, Sikong Biaosheng wenji
Annotated ‘Brush Talk from Dream Brook’], ed. Hu 司空表聖文集 [The Collected Prose of Sikong Biaosheng]
Daojing 胡道靜 (Shanghai, Shanghai Renmin, 2013), (Shanghai, Shanghai guji, 1994), 54–55. Other mentions of
224. The English translation is from Joseph Needham, Jia Dan’s map include an epitaph for Jia written by a
‘Geography and cartography’, in his Science and contemporary statesman Quan Deyu 權德輿 (759–818), a
Civilisation in China, vol. 3, Mathematics and the Sciences of mention of Huayi tu map by Liu Zongyuan 柳宗元 (773–
the Heavens and the Earth (Cambridge, Cambridge 819), and several poems from the late 9th century about
University Press, 1959), 497–590 at 576. viewing Huayi tu. See Xin Deyong 辛德勇, ‘Shuo fuchang
10. See Cordell D. K. Yee, ‘Chinese cartography among shike “Yuji tu” yu “Huayi tu”’ 說阜昌石刻《禹跡圖》與
the arts: objectivity, subjectivity, representation’, in Harley 《華夷圖》[On Fuchang-era stele maps ‘Yuji tu’ and
and Woodward, Cartography in the Traditional East and ‘Huayi tu’], Yanjing Xuebao 燕京學報 28 (2010): 1–72.
Southeast Asian Societies (note 3), 137. 20. Xin Deyong’s observations coincide with those of
11. See, for example, a discussion of tu in the context of earlier scholars, Wang Guowei, Cao Wanru and others,
medieval Chinese pictorial theory, traceable to a body of who all agree that the major features depicted on the
early texts that reinforce its twin nature as ‘prognostica- Song-era stele map Huayi tu are likely to have ‘evolved’
tive and cartographic in content and curvilinear in form’, from Jia Dan’s map. Xin Deyong, ‘Shuo fuchang shike
in Eugene Yuejin Wang, Shaping the Lotus Sutra: Buddhist “Yuji tu” yu “Huayi tu”’(see note 19), 27–30. For a dis-
Visual Culture in Medieval China (Seattle, University of cussion of the relationship between Song maps and Jia
Washington Press, 2005), 199. Dan’s model of the empire, see Hilde De Weerdt, ‘Maps
12. The renowned painter Yan Yanzhi 顏延之 (384– and memory: readings of cartography in twelfth- and
456), for example, explained that in addition to the mean- thirteenth-century Song China’, Imago Mundi 61:2
ing painting (繪畫), the other two meanings of tu com- (2009): 145–67 at 157.
prise the ‘forms of the hexagrams’ (卦象) and the ‘study of 21. A Tang-dynasty zhang 丈 is approximately 300 cen-
written characters’ (字學). The definition was quoted and timetres, one chi 尺 30 centimetres, one cun 寸 3 centi-
inherited by the 9th-century aficionado of painting Zhang metres and one li 里 is about half a kilometre.
Yanyuan 張彥遠 in his book Record of Famous Paintings 22. Liu Xu, Jiu Tang shu (see note 18), 138.3785–6.
through the Ages (歷代名畫記). See the English translation 23. It has been suggested that these early accounts of
and discussion in William Reynolds Beal Acker, Some King Yu’s flood-taming have direct links to the division of
T’ang and pre-T’ang Texts on Chinese Painting, Translated China into regions and to the tying of each region to a
and Annotated (Leiden, Brill, 1954), 65–66. political centre through the paying of tributes. See Mark
13. Francesca Bray, ‘Introduction: the powers of tu’, in Edward Lewis, Flood Myths of Early China (Albany, State
Graphics and Text in the Production of Technical Knowledge in University of New York Press, 2006), 47.
China: The Warp and the Weft, ed. Francesca Bray, Vera 24. Xin Deyong 辛德勇, ‘Tangdai de dili xue’ 唐代的地理
Dorofeeva-Lichtmann and Georges Métailié (Leiden, Brill, 學 [The Study of Geography in the Tang Dynasty], in
2007), 1–78 at 73. Tangdai diyu jiegou yu yunzuo kongjian 唐代地域結構與運
14. Translation from Garret Olberding, ed. Dubious Facts: 作空間 [Regional Structures and Functional Space in the
The Evidence of Early Chinese Historiography, SUNY Series in Tang Dynasty], ed. Li Xiaocong 李孝聰 (Shanghai,
Chinese Philosophy and Culture (Albany, State University Shanghai cishu, 2003), 439–63 at 449.
Imago Mundi 70:2 2018 Map Reading in Medieval China 181

25. Fan Ye 范曄, ed., Hou Han shu 後漢書 [Histories of 41. Fifty-three kinds of tujing with specific geographical
the Latter Han] (Beijing, Zhonghua shuju, 1965), 24.834. descriptors in their titles from the Tang and Five Dynasties
26. Liu Xu, Jiu Tang shu (see note 18), 138.3786. (907–960) periods have been identified in these textual
27. Jiang Fang, who was active during the Yuanhe 元 remnants by Hua Linfu 華林甫, ‘Sui-Tang tujing jikao’ 隋
和 reign (806–821 CE) but whose life is not well docu- 唐圖經輯考 [Study and compilation of tujing from the Sui
mented, is perhaps best known to literary historians as and Tang], Guoli Zhengzhi daxue lishi xuebao 國立政治大學
the author of ‘The Story of Huo Xiaoyu’ (霍小玉傳), a 歷史學報 27 (2007): 141–213.
tale of love betrayed. He also had close ties with the 42. Edwin O. Reischauer, Ennin’s Diary: The Record of a
more famous Yuan Zhen 元稹 (779–830), who had Pilgrimage to China in Search of the Law (New York, Ronald
recommended Jiang Fang for the important post of Press Company, 1955).
Hanlin Academician in 821, just after Yuan presented 43. The diary entry for the fourth month, fifth day in the
his maps to the emperor. Bian Xiaoxuan 卞孝萱, Yuan year 840 is the only one that mentions maps. Ennin writes,
Zhen nian pu 元稹年譜 [Chronicles of Yuan Zhen] ‘We started early. . . . After the forenoon meal we went
(Jinan, Qi lu shu she, 1980), 361. fifteen li west to the east side of Mt. [Changbai]. It was
28. Quan Tang wen 全唐文 [Complete Tang Prose] almost 4 P.M. when we drank tea at the [Shi] household
(Shanghai, Shanghai guji, 1990), 719.7393. . . . and asked for the [Spring Water Temple]. Our host
29. Ibid. replied that, if we went fifteen li due west of [Bucun], we
30. Quan Tang shi 全唐詩 [Complete Tang Poems] would reach it, so we went straight west into the mountains.
(Beijing, Zhonghua shuju, 1960), 716.8225. Cao Song We went over ten li on the wrong road. There were many
曹松 obtained his jinshi degree in 901, which means he different roads, and we did not know where they led’.
was active in the capital around that date. See Meng English translation from Reischauer, Ennin’s Diary (see
Erdong 孟二冬, ed., Deng ke ji bu zheng (Xu Song zhuan) note 42), 199–200.
登科記補正 (徐松撰) [Amended Register of Degree 44. In his entry for the following day, when he and his
Holders] (Beijing, Beijing Yanshan, 2003), 24.1041. party had arrived at the temple around midday, Ennin
For a discussion of Cao Song’s poem as a precursor to noted, ‘A peak on the south side of the monastery, called
Song-era map-reading practices and for another transla- the Dragon Terrace (龍臺), stands out alone above the
tion of this poem, see De Weerdt, ‘Maps and memory’ other peaks. It is recorded on our map [ditu] that once a
(note 20), 156–57. dragon danced on this peak and that this was reported to
31. Li Fang 李昉, ed., Taiping guangji 太平廣記 the throne’ (ibid., 203).
[Extensive Records for the Era of Supreme Peace], 10 45. ‘About to Arrive at Shaozhou, I Send This Ahead to
vols. (Beijing, Zhonghua shuju, 1961), 12.82. Prefect Zhang to Borrow his Tujing’ (將至韶州先寄張端公
32. The poet and painter Wang Wei 王維 (699–759), for 使君借圖經), Quan Tang shi (see note 30), 344.3860.
example, deployed this image in a poem written for a 46. Xin Deyong, ‘Tangdai de dili xue’ (see note 24), 442.
Daoist practitioner (Quan Tang shi (see note 30), 47. For a discussion of the strong attraction of the capital
127.1288). Such poems were common also in the 9th in the political careers of Tang officials, see Linda Rui
century. Linghu Chu 令狐楚 (766–837) wrote one such Feng, City of Marvel and Transformation: Chang’an and
poem titled ‘To Mr. Mao the Immortal’ (贈毛仙翁): ‘In his Narratives of Experience in Tang Dynasty China (Honolulu,
gourd, alchemical pills and incantations for climbing the University of Hawai`i Press, 2015), 44–67.
rainbow; / next to his elbow, a book of arcane knowledge 48. Ouyang Xiu, Xin Tang shu (see note 37),
with arts of earth-shrinking’ (壺中藥物梯霞訣, 肘後方書縮 176.5260–61.
地功) (ibid., 334.3746; my translation). 49. Han Yu 韓愈, Han Changli shi xinian ji shi 韓昌黎詩繫
33. ‘He Letian zaochun jian ji’ (和樂天早春見寄), Quan 年集釋 [Collected Poems of Han Yu, Annotated and
Tang shi (see note 30), 417.4601. Dated], ed. Qian Zhonglian 錢仲聯 (Shanghai, Shanghai
34. Xin Deyong, ‘Tangdai de dili xue’ (see note 24), guji, 1984), 1097.
439–44. See also Hargett, ‘Song dynasty local gazetteers’ 50. Liu Xu, Jiu Tang shu (see note 18), 15.470.
(note 8), 409. 51. Edward H. Schafer, The Vermilion Bird: T’ang Images of
35. On the date and political context of this memorial, the South (Berkeley, University of California Press, 1967), 39.
see Bian Xiaoxuan, Yuan Zhen nian pu (note 27), 359. 52. Shunzong was the father of Emperor Xianzong, who
36. Yuan Zhen 元稹, Yuan Zhen ji (Ji Qin jiao dian) 元稹集 in 819 demoted Han Yu and banished him to Chaozhou.
(冀勤校點) [Collected Works of Yuan Zhen] (Beijing, The Veritable Records is commonly attributed to Han Yu,
Zhonghua shuju, 1982), 406; Quan Tang wen (see note 28), who was certainly one of its compilers, but it is still
615.6607. debated whether the transmitted version is in his hand.
37. Li Jifu 李吉甫 was the author of the important geo- See E. G. Pulleyblank, ‘The Shun-Tsung Shih-Lu’, Bulletin
graphical work Yuanhe junxian tu zhi 元和郡縣圖志. The of the School of Oriental and African Studies 19:2 (1957):
occasion is recorded in Ouyang Xiu 歐陽修, Xin Tang shu 336–44.
新唐書 [New History of the Tang] (Beijing, Zhonghua 53. Liu Xu, Jiu Tang shu (see note 18), 135.3732.
shuju, 1975), 146.4742. 54. For a concise account of the political faction to
38. The Tiande Fortification is near modern Baotou, in which Wei Zhiyi belonged during the brief Shunzong
Inner Mongolia. See Tan Qixiang 譚其驤, ed., Zhongguo reign, and which included Wang Shuwen 王叔文, Wang
lishi ditu ji 中國歷史地圖集 [Historical Atlas of China] Pi 王伾, Lü Wen 呂溫, Liu Yuxi 劉禹錫and Liu
(Shanghai, Ditu chubanshe, 1982), 38–39. Zongyuan 柳宗元, see Denis Twitchett, ed., The
39. For an account of the marriage of the Princess Taihe Cambridge History of China, vol. 3: Sui and T’ang China,
to the Uigher khan during the Muzong reign, see Liu Xu, 589–906, Part 1 (Cambridge, Cambridge University
Jiu Tang shu (note 18), 16.489. Press, 1979), 601–7.
40. The other extant fragments include a preface to 55. Han Yu 韓愈, Han Changli wenji jiaozhu 韓昌黎文集
Zhangzhou tujing 漳州圖經 and several Dunhuang manu- 校注 [Collected Prose Writing of Han Yu, Annotated]
scripts. See Xin Deyong, ‘Tangdai de dili xue’ (note 24), 441. (Shanghai, Shanghai guji, 1986), 723.
182 L. R. Feng Imago Mundi 70:2 2018

56. According to the Confucian classic The Rites of Zhou, the Yazhou of Li Deyu’s Exile], in Chang shui ji (xia) 長水集
function of the Bureau of Operations (職方) official was to be (下) (Beijing, Renmin chubanshe, 2009), 349–51.
‘in charge of the maps/charts under heaven’ (掌天下之圖). 60. Franciscus Verellen, ‘The dynamic design: ritual and
See Needham, ‘Geography and cartography’ (note 9), 534. contemplative graphics in Daoist scriptures’, in Daoism in
57. Catherine Delano-Smith, ‘Maps and map literacy I: History: Essays in Honour of Liu Ts’un-yan, ed. Quanren Liu,
Different users, different maps’, in Plantejaments i Objectius Benjamin Penny and Cunren Liu (London, Routledge,
d’una Història Universal de la Cartografia: Approaches and 2006), 159–86. For discussions of visualization diagrams
Challenges in a Worldwide History of Cartography, ed. David and the relationship between image and mobility through
Woodward, Catherine Delano-Smith and Cordell D. K. Yee meditation, see Shih-shan Susan Huang, Picturing the True
(Barcelona, Institut Cartogràfic de Catalunya, 2000), 223–40. Form: Daoist Visual Culture in Traditional China (Cambridge,
58. Han Yu, Han Changli wenji (see note 55), 723. Harvard University Asia Center, 2012).
59. The location Yazhou 崖州 on Hainan Island had 61. This tale, titled ‘Chen Jiqing 陳季卿’ after its prota-
changed over time; see discussion in Tan Qixiang 譚其 gonist, will be treated in depth in another article. Li Fang,
驤, ‘Li Deyu shuo zhe zhi Yazhou’ 李德裕所謫之崖州 [The Taiping guangji (see note 31), 74.462–63.

Les cartes perdues peuvent-elles parler? Vers une histoire culturelle de la lecture des cartes dans la Chine
médiévale
En dépit du fait que quasiment aucune carte du premier millénaire n’ait survécu, il est toutefois possible de
reconstituer un riche contexte associé à leur production, leur usage et leur perception, à partir de sources
écrites variées. Trois cas datant de la dynastie Tang (618–907 apr. J.-C.) sont présentés dans cet article, dans
lesquels les caractéristiques des cartes manquantes émergent à partir des textes qui leur étaient associés et qui
leur ont survécu. Ces exemples comprennent deux documents qui étaient autrefois joints à des cartes
présentées à l’empereur, ainsi qu’une anecdote qui se réfère à une carte de la lointaine frontière
méridionale. Ils démontrent que les cartes n’étaient pas seulement dessinées pour contenir le territoire
impérial, mais aussi pour servir de guides pour d’ambitieux voyages. Elles étaient aussi perçues par leurs
utilisateurs comme des invitations à des expériences tout à la fois désirables et indésirables.

Können verlorene Karten sprechen? Überlegungen zu einer Kulturgeschichte des Kartenlesens im

mittelalterlichen China
Ungeachtet des Umstands, dass so gut wie keine chinesischen Karten aus dem ersten Jahrtausend erhalten
geblieben sind, ist es möglich, aus einer Vielfalt von Textquellen eine umfangreiche Tradition ihrer
Herstellung, Nutzung und Rezeption zu rekonstruieren. Drei Beispiele aus der Tang-Dynastie (618–907 n.
Chr.) werden in diesem Beitrag vorgestellt. Anhand der begleitenden Texte lassen sich die Charakteristika der
verlorenen Karten herausarbeiten. Zwei der Beispiele sind überlieferte Dokumente, die zu Karten gehörten,
die dem Kaiser übermittelt wurden. Hinzu kommt eine Anekdote über eine Karte der entferntesten
südlichen Landesgrenzen. Die Beispiele zeigen, dass die Karten nicht nur gezeichnet wurden, um das
Imperium zusammenfassend darzustellen, sondern auch, um als Wegweiser für ambitionierte Reisen zu
dienen. Sie wurden zudem als Einladungen zu mehr oder weniger erstrebenswerten Erfahrungen
verstanden.

¿Pueden los mapas perdidos hablar? Hacia una historia cultural sobre la lectura de mapas en la china
medieval
A pesar de que prácticamente no se ha conservado ningún mapa chino del primer milenio, es posible, sin
embargo, reconstruir un rico contexto relacionado con su producción, uso y percepción gracias a una
variedad de fuentes escritas. En este artículo se presentan tres casos de la dinastía Tang (618−907) en los
que las características de los mapas desaparecidos emergen a través de los textos que han llegado a la
actualidad. Estos ejemplos incluyen dos documentos que una vez acompañaron a los mapas que fueron
presentados al emperador, y una anécdota que hace referencia a un mapa de la remota frontera meridional.
Demuestran que los mapas se diseñaron no solo para encapsular el territorio imperial, sino también para
servir de guía para la planificación de viajes. Sus usuarios también los percibían como invitaciones a
experiencias tanto deseables como indeseables.
Plate 1. Detail from the Complete Map of Qiong Prefecture (see Fig. 4). The cartouches explain the customs
of the island’s inhabitants. See p. 177.

Plate 2. Detail from the Matsumae-no-shima map (see Fig. 1) showing the Ōshima Peninsula, southwest
Hokkaido. Japanese villages are marked in yellow cartouches and Ezo settlements in lilac. The
transition is labelled on each coast ‘From this point onwards [is] West/East Ezochi’ as appropriate.
The red lines represent maritime trade routes. (Reproduced courtesy of the National Archives of Japan,
Tokyo.) See p. 190.