Water sprites and ancestor spirits: reading the architecture of Jinci

Tracy G. Miller

Jinci, or the Memorial Shrine of Jin, perhaps the most unconventional shrine complex in China, occupies a verdant site near the remains of ancient Jinyang, a capital city of the Zhou dynasty (ca. 1100-221 B.C.E.) Jin State, eleven miles southwest of the modern capital of Shanxi Province, Taiyuan. (1) According to standard histories and geographic texts dating back more than a millennium, the shrine was dedicated to a historical figure, Shu Yu of Tang, the founder of the Jin State. (2) Nonetheless, the architecture of the complex tells another story, one that points to more local concerns. Whereas shrines and temples in China are typically arranged as rectilinear courtyard compounds with a strong central axis leading up to a main offering hall, Jinci is distinctive in that the temple buildings are distributed in a seemingly random manner. But all are focused on the fountainhead of the Jin River--a canal flowing through the site filled by (originally) three springs, the centermost of which lies before the eleventh-century Sage Mother Hall (Figs. 1, 2). In 1934, when the first trained Chinese architectural historians, Liang Sicheng and Lin Huiyin, visited the complex, they responded immediately to the unusual layout of Jinci, with its temple buildings placed around the springs and canals. (3) Although common in larger sacred sites in China, such as Mt. Wutai in the more northern reaches of Shanxi Province or Mt. Song in Henan Province, the plan of Jinci surprised Liang and Lin because Jinci, unlike the sacred mountains, is nominally a single memorial shrine and is physically contained by a wall with a formal gatehouse. For those familiar with the vocabulary of Chinese ritual architecture, the lack of a formal, rectilinear courtyard structure within the gatehouse of a walled temple or shrine complex is quite unexpected.

For much of the complex's history, the identity of Jinci's main deity, the deity to whom the name refers, was important to people living near and traveling to the shrine and, at the same time, ambiguous. A "memorial shrine of Jin" at the site of the Jin Springs could be dedicated either to a state ancestor, the historical Shu Yu of Tang, or a water sprite, the Spirit of the Jin Springs, but not both equally. The history of the site prior to the eleventh century rests on textual sources alone, and these sources focus on Shu Yu of Tang, reflecting the interests of the educated elite who wrote and read them. The largest, grandest, and oldest building on the site, the Sage Mother Hall (ca. 1038-87 C.E., Fig. 2) is dedicated to the Spirit of the Jin Springs, a divinity that is not overtly identified in the earlier texts. The position of the Sage Mother Hall reflects not only her identity as a water spirit but also her significance to the community, which depended on the springs for annual irrigation water and maintained the timber-frame structure for more than nine hundred years. (4) From at least the thirteenth century, the physical dominance of the building over the compound implied that the Sage Mother was the dominant deity at the site, contradicting the standard historical sources. By the sixteenth century, competing groups of patrons at Jinci, including the literati elite, royal landlords, and local farmers, all sought to align the Sage Mother with their own interests. They asserted their claims to her favor through temple building, an activity that, by the beginning of the seventeenth century, made possible a "Confucianized" interpretation that rectified the inconsistencies: the Sage Mother was transformed from a local water sprite to the mother of Shu Yu of Tang, the first true queen of the Zhou dynasty, Yi Jiang. Although most modern art historical scholarship on the Sage Mother Hall and the larger site of Jinci has followed this tradition, (5) it is actually the view of a literate elite that considered Zhou dynasty ancestor spirits to be more significant than local deities.

Previous studies of Jinci have typically either examined the whole of the site historically or focused on the Sage Mother Hall in exclusion to its relation to the springs and the later, less architecturally significant buildings. My work differs in that I focus on the way in which the shrine complex and the deities worshiped within it were used to act out the political and social history of the region across time and how competing patronage groups used temple architecture to serve their specific interests. Most twentieth-century historical studies of Jinci provide textual evidence for specific extant features of the site without relating them to the larger social, political, or religious history of China. (6) Architectural historians, art historians, and archaeologists tend to concentrate on the date and style of the Sage Mother Hall and the sculpture within it. (7) The Sage Mother Hall is significant to the study of Chinese art and architecture because it is the second-largest building of the Northern Song dynasty (960-1127 C.E.) extant in China and the oldest timber-frame hall of this scale not associated with Buddhism or Daoism. (8) The size and style of the bracketing, the reduction of interior columniation (particularly in the front veranda), the shape and inward inclination of the columns, and the increased height of the columns toward the corners, which gives its double eaves their dramatic curvature, are all characteristic features of palatial-style halls from this period. Many of these features are also described in the first extant official manual of building standards in China, Yingzao fashi (1103), making the Sage Mother Hall a key monument for understanding this text. (9) Since the hall contains forty-three unfired clay sculptures of the Sage Mother and her attendants, Jinci is also important for understanding Northern Song sculpture. (10) With few exceptions, scholars studying Jinci both historically and art historically have followed the seventeenth-century interpretation of the site, which begins with Shu Yu of Tang and identifies the Sage Mother as his mother, Yi Jiang. (11)

[ILLUSTRATION OMITTED]

Reading the early architectural history of Jinci in conjunction with its textual history reveals a contested site, one in which different interpretations of the identity of the main deities were not simply "superscribed" onto each other but remained separate in order for certain patrons to assert dominance over the site and, particularly in later periods, its important natural resource. (12) By including the evidence of the site itself, I shall show not only that Jinci was significant for early rulers and later Confucian elites because of its association with the state founder, but also that it was vital for the local people because it was the source of the Jin River, which they depended on for their crops. The unique combination at Jinci of an extensive textual and architectural record relating to different types of deities--male and female, ancestor and nature spirit, elite and common--allows us to see how competing ideas about divinity identity were self-consciously communicated through the activity of temple building in late imperial China. In this article I shall show that the different patronage groups at Jinci actively employed the means most available to them, both history writing and temple building, to communicate an understanding of the site that would benefit them. To do this I will first describe the modern condition of the site and compare it with an example of a large-scale temple complex to establish how the walled site of Jinci is unexpectedly different from other temple complexes. I will then look at the historical development of Jinci and the cults of the two deities of Jin through the beginning of the sixteenth century. Finally, I will discuss how the building programs of the sixteenth and early seventeenth centuries relating to those divinities served to align them with competing interest groups, ultimately creating an environment whereby the Sage Mother could be interpreted as an ancestral mother of the Jin State rather than as merely a local water sprite.

[ILLUSTRATION OMITTED]

The Site of Jinci

The surprise experienced by Liang and Lin at Jinci's appearance was not limited to twentieth-century visitors. Zhou Zaijun was following in the footsteps of many fellow literati when, in 1665, he went to Jinci. There he found the situation of the buildings in the complex not as he expected:

   Having previously heard of the scenic beauty of Jinci, I
   traveled there ... . I entered the shrine [complex], but the
   only deity being sacrificed to was the Sage Mother of
   Manifest Aid. I asked about the shrine [complex] of Shu
   Yu of Tang. A Daoist priest pointed to the several bays of
   a dilapidated wayside building and said, "That is it ... ." (13)

From as early as the sixth century C.E., the famous first ruler of the Zhou dynasty state of Jin, Shu Yu of Tang, was the figure to whom emperors, officials, and literati traveling through the area of Jin's ancient capital Taiyuan came to pay tribute. When he arrived at Jinci in 1665, Zhou Zaijun had a clear expectation that the primary focus of worship would be Shu Yu of Tang. However, the built environment contradicted this. The site was clearly dedicated to the Sage Mother of Manifest Aid, also known as the Spirit of the Jin Springs. From the end of the eleventh century until today, the Sage Mother has dominated the site of Jinci, a site that by all textual accounts should have been dedicated to the famous first ancestor of the region, Shu Yu of Tang.

[ILLUSTRATION OMITTED]

The distribution of buildings and complexes at Jinci is atypical for large-scale ritual and religious complexes in China. The main halls of such complexes are usually organized in a rectilinear courtyard-complex format. (14) The temple complex of the Goddess of Earth at Fenyin (modern Wanrong County), Shanxi Province (Fig. 3), a major imperially patronized temple dedicated to a nature spirit, demonstrates that the courtyard complex common to Buddhist, Daoist, and palatial complexes was also used for nationally known female nature spirits. (15) All major ritual complexes during the middle imperial period (tenth-sixteenth centuries), regardless of their sectarian affiliation, followed this standard format. The 1137 C.E. relief carving of the Northern Song dynasty plan of the site shows that the feature most representative of Chinese ritual architecture is its cardinal, and usually southern, orientation. This is indicated through the inclusion of basic geographic features: the Fen River (at the top of the rendering) flows into the Yellow River (on the left side of the rendering), signaling that the top is north and the left is west. Additionally, two subsidiary courtyards on the left and right are labeled west and east respectively.

The next apparent feature is the front gateway and enclosing walls. After entering through the front gate, a visitor would progress through four more gatehouses distributed along a central axis. The first of these gatehouses is joined on either side by a roofed gallery that encircles the larger complex. The second and third gatehouses also join walls that extend to the larger enclosing wall on the left and right. The fourth gatehouse joins another inner enclosing gallery, which appears to extend from the gatehouse to the back hall of this inner complex. The main hall of this complex is located behind an axial altar/stage platform, (16) and it is linked to a smaller retiring hall (qindian, [TEXT NOT REPRODUCIBLE IN ASCII]) by a covered platform, creating an I-shaped main hall combination. (17) A central back gate behind this hall leads to a set of northern altars in a rounded enclosure separated by a single wall and gateway. Notably, all of the subsidiary buildings and courtyards in this carving are positioned either parallel or perpendicular to the central axis.

Although some details of this complex are not widely seen in the majority of ritual complexes in the Chinese sphere of influence, (18) the basic features of southern orientation, main halls on a central axis, and surrounding galleries apply to sites for all religious groups. Indeed, even the I-shaped main hall, which is specified for an ancestral complex in the fifteenth-century carpentry manual Lu Ban Jing, can be seen in Buddhist complexes in Shanxi from the twelfth to fifteenth centuries. (19) The walls of the courtyard itself isolate both divinity and devotee from the immediate surroundings and provide a familiar atmosphere that transcends any particular location.

At Jinci, however, the largest, oldest, and most dominant building on the site, the Sage Mother Hall, faces southeast (48[degrees] east of south; Fig. 1: no. 1). The shrine complex to Shu Yu of Tang (Figs. 4, 1: no. 2), while facing toward the south (30[degrees] west of south), is not aligned with the main gateway (old or new) and does not show a clear relationship to any other building on the site. Other individual buildings and enclosed building complexes are not oriented to the cardinal directions, nor are they placed directly parallel or perpendicular to a main central axis organizing the larger compound. Yet they do show a consistent pattern of organization in that they all face onto the water that flows through the site, reiterating the significance of the spring to its sacrality.

The architectural composition of Jinci is also unusual in that within the larger compound, the majority of the structures exist in two formats: individual buildings open to the site or building complexes with their own gateway and enclosing walls. Individual buildings present at the site in the seventeenth century include the Dressing Tower (the modern Water Mother Tower, Figs. 5, 1: no. 3), pavilions covering the springs, and viewing pavilions placed over the canal. All of these buildings stand independently, with neither an axial relationship to a more dominant building complex (either parallel or perpendicular to a central axis) nor an enclosing wall to distinguish them from the larger compound. They are localized in their relation to the most important feature of this site: the origin of the water that irrigates and brings life to the land. The building complexes, groups of buildings, including the shrines to Shu Yu of Tang, Wenchang, the god of the literary arts (Figs. 6, 1: no. 11), and the Daoist deities the Jade Emperor and the Three Pure Ones (Sanqing, Figs. 7, 1: no. 9), are each approached through a gate and delineated from the rest of the site by an enclosing wall. The space of these complexes, like that of the Temple to the Goddess of Earth at Fenyin, is generic and would be appropriate for a nationally known deity.

The Sage Mother Hall is unique in that it exists in both models. The hall emerges as the apex of an axis of buildings, including a gateway traversing the axis and the bell and drum towers symmetrically placed on either side of that axis, making it part of a building complex. Yet it also reads as an individual building in that no separate rectilinear enclosing wall partitions it off from the larger site. Like the Dressing Tower, the Sage Mother Hall stands open to the water source in front of it. But, like the shrine to Shu Yu of Tang, it also exists as the focus of a building complex with an entry gate and offering hall. The multiplicity of temples within the Jinci compound came as a surprise to traveling scholars such as Zhou Zaijun. Like any nonnative visiting the site, he needed to be directed by a professional employee of the complex to understand the placement of temple buildings and find the specific deity he was most interested in venerating. This configuration was not the manifestation of a single design program but, rather, the result of different patronage groups staking their own claims to the favor of the Sage Mother, the occupant of the oldest and grandest building in the complex. By comparing the history of Shu Yu of Tang, the Sage Mother, and the development of the temple complexes in which they were worshiped, we can see how the built environment is both a reflection of local beliefs and an agent in effecting change in those beliefs across time.

Shu Yu and the Jin State

Before examining the relation between the form of the building complex and divinity identity, I shall give a history of the two deities of Jin and their relation to the site of Jinci. Shu Yu of Tang's prominent role in the canonical history of the Zhou dynasty made his shrine significant for those educated in the Chinese classics. Shu Yu was the younger brother of King Cheng, the third king of the Western Zhou dynasty, who is thought to have ruled during the late eleventh century B.C.E. One of the primary sources for Zhou dynasty history, the Spring and Autumn Annals, Commentary of Zuo (Chunqiu Zuozhuan), records Shu Yu's history in conjunction with the origin of the Jin State. Here the educated elite since at least the Han dynasty (206 B.C.E.-220 C.E.) would have read that while pregnant with him, Shu Yu's mother, Yi Jiang, dreamed of the Thearch (di) who declared that her future child would be named Yu, would be enfeoffed in Tang (the fief that would become the core of the Jin State), and would have many sons and grandsons. The child was subsequently born with the character yu on his hand and was named accordingly. (20) Shu Yu's enfeoffment took place when he was playing with his older brother, the young King Cheng. As part of a game, King Cheng gave Yu a paulownia leaf and declared him enfeoffed. The Duke of Zhou heard of King Cheng's action and insisted that, as a king, Cheng should be true to his word. (21) Yu, called Shu Yu because he was the younger brother of the king, (22) was therefore given the newly conquered territory of Tang and was thenceforth known as Shu Yu of Tang. The name of Tang was later changed to Jin, thus making Shu Yu the earliest Zhou lord of the famous Warring States period (445-221 B.C.E.) Jin State. Regardless of its historical accuracy, (23) the traditional version of the story of enfeoffment gave this northern borderland a legitimate relation with the heart-land of the Zhou dynasty. (24) And from the second century C.E., geographic texts reiterated this relation by stating that the Jin River was the eponym of the early Jin State. A shrine to Shu Yu was likely placed at this location because he was the founder of the Jin State, just as the spring was the origin of the Jin River. (25)

[ILLUSTRATION OMITTED]

Our first accounts of Jinci are found in independent geographic texts and in the standard histories from the Northern and Southern Dynasties period (386-581 C.E.). Taiyuan (located at Jinyang) was an area of significant strategic importance during this time, (26) and it is difficult to determine whether Shu Yu's took prominence because of his perceived power in helping local warlords or because of the bias of the sources toward historical figures. The spring was fundamental, but no clear mention of a riverine goddess is present in these texts; instead, they correlate the physical features of the site, including waterways and man-made structures, with historical accounts of the place.

[ILLUSTRATION OMITTED]

The earliest evidence for religious buildings at the Jin Springs is from two geographic texts written in the sixth century. The first is in the "Treatise on Geography" of the Standard History of the Wei (Weishu), which, although compiled during the Northern Qi dynasty (550-77 C.E.), would have used source material collected during the Northern Wei dynasty (386-534 C.E.). The entry tells of Xuanweng Mountain, located to the south of the regional capital Taiyuan, and describes it as the location of the spring for the Jin River with a shrine to "Jinwang" (the King of Jin). (27) In his text of roughly the same period, the Commentary on the Book of Waterways (Shuijing zhu), Li Daoyuan (d. 527 C.E.) describes a shrine at the Jin Springs dedicated to Shu Yu of Tang and recounts Shu Yu's life using the stories I have just related from historical sources considered to be standard in his day. The building (or buildings) wherein Shu Yu was venerated was found near a pool, and according to one interpretation of the passage, it may have been located west of a pool with a mountain at its back. A bridge over the river was emphasized as a popular site for rendezvous. Notably, while describing the site as a destination for travelers to the area, Li does not mention any other spirits venerated there, leaving the impression that the only object of worship was this ancestral figure who was easily found in the standard histories. A third early geographic work, Maps and Gazetteer of the Provinces and Counties in the Yuanhe Period, 806-14 (Yuanhe junxian tuzhi, 831 C.E.), provides the history of the site through the early Tang dynasty (618-907 C.E.), consisting largely of a much-abridged version of the Commentary on the Book of Waterways. It gives an abbreviated form of the story of Shu Yu of Tang's enfeoffment, and he is designated the shrine's subject of worship. This text goes on to discuss the expansion of the site at the end of the sixth century to include more extensive waterways and multistoried buildings. (28)

[ILLUSTRATION OMITTED]

Li Daoyuan's account does not give much detail about the patrons of the shrine, but we have other information suggesting that Shu Yu's cult was important to generals in the area during this time. Gao Huan, a general of the Northern Wei dynasty whose independent regime became the Northern Qi dynasty in 550 C.E., went to Jinci in 534, not long after Li Daoyuan went to the Jin Springs. There he and all of his men but one performed a rite of worship to Shu Yu of Tang. (29) Li Yuan, the first emperor of the Tang dynasty, may also have gone to Jinci prior to his conquest of the Sui dynasty (581-618 C.E.). Li Yuan was the commander of Taiyuan during the Sui and held the title duke of Tang, after the territory conquered by the Zhou and given to Shu Yu as his fief. In 617 two of Li Yuan's officials conspired to lure him to Jinci to pray for rain in order to assassinate him, but he discovered the plot and avoided harm. From this account we see that Jinci was known both to generals and officials, and it was reasonable to suggest that the local ruler go there to request rain from the spirits. After conquering the Sui, Li Yuan named his dynasty Tang and traced his legitimacy back to the traditional belief that the territory of Tang (at Jinyang/Taiyuan) was originally the residence of the legendary emperor Yao (traditional ascension date 2356 B.C.E.). However, in the Jinciming stela inscription of 646 C.E., Li Yuan's son Li Shimin (r. 626-49 C.E.) emphasized the significance of Shu Yu and the Zhou royal house and spoke of Li Yuan beginning his dynastic conquest from the shrine of Jinci. (30)

After the fall of the Tang dynasty at the beginning of the tenth century, Shu Yu continued to be important in legitimizing rule over the region of Taiyuan. In 941 the first Later Jin dynasty ruler, Shi Jingtang (r. 936-43), elevated Shu Yu to King of Arising Peace (Xing'anwang) and the Spirit of the Fen River, who was also an important protector spirit of the Zhou dynasty Jin State, to Duke of Flourishing Tranquility (Changninggong). (31) His act reiterated the link between his new dynasty of Jin and the Zhou period Jin State. (32)

Across the Northern Song dynasty (960-1127 C.E.), we see Shu Yu lose his status at the site of Jinci. He continues to be important for establishing sovereignty over the area of Taiyuan at the beginning of the dynasty, as the second Song emperor sponsored the rebuilding of Jinci after the Song conquest of the region in 979. Yet, although he is given a new title in the early eleventh century, by the time the dynasty lost all of north China to the Jurchen Jin invaders in the early twelfth century, the Spirit of the Jin Springs appears in the historical record and rises to greater importance at the site. Because her history is intermingled with discussion of the importance of Shu Yu, I will return to Shu Yu and the context of his elevation and decline in the Song dynasty after discussing her arrival.

The Emergence of the Spirit of the Jin Springs

Aside from the veneration of Shu Yu of Tang, another purpose for traveling to Jinci was to pray and give offerings to the Spirit of the Jin Springs, who would eventually be elevated to the Sage Mother of Broad Benevolence, Clear Efficacy, and Manifest Aid. Contrary to the implication of the authors of standard historical documents, beginning from the seventh century, extant stela inscriptions record that the site was not used for the exclusive worship of Shu Yu. Water spirit worship was an equally important activity at Jinci. The stela inscriptions from the second emperors of both the Tang and Song dynasties reveal a tension between the worship of Shu Yu of Tang and the Spirit of the Jin Springs. These emperors were interested not in the spring but in creating a record (or having one created) of themselves supporting the historical founder of the territory in order to bolster their own sovereignty. At the same time, their inscriptions indirectly acknowledge the existence of a spring spirit at the site.

The earliest documentation of spring spirit worship is contained in the Jinciming stela inscription written by Li Shimin while traveling in the area in 646. The purpose of the inscription was to draw parallels between the Zhou and the Tang dynasties because of the political and military significance of Taiyuan and, consequently, Shu Yu, in Li Yuan's founding of the Tang dynasty. However, Li Yuan also provides a vivid (though exaggerated) description of the shrine complex as a transcendents' paradise and much glorification of the importance of the Jin Springs.

       ... [Seeing the shrine's] golden que gateways, nine layers
    [deep], they would despise [the palaces on] Penglai as quite
    crude; [seeing the shrine's] jade towers, eight thousand sand feet
    high, they would disdain (the mansions of) Kunlang (33) as hardly
    unusual ....
       Regarding [the spring] bestowing favor, harmonious wind and moist
    dew will be produced by it; clouds [as rich as] oil and rain [as
    rich as] cream arise from this ....
       Regarding its huge capacity, it nourishes the myriad creatures
    without fatigue, it supplies all quarters without exhaustion ....
       Daily flowing without cessation, like a fragrant way which is
    endless; each year it pours without overflowing, just as the
    superior person who avoids becoming full [of awareness] of his
    virtue ....
       In truth, what is unique about this spring is that it encircles
    this precinct of transcendents as though it were a jewel, so that
    when we gaze up at the solitude and purity of the divinity's abode,
    we can still hear its beautiful music as if it were there.
       That is why the [beautiful carriages of the nobility with] red
    wheels and ornamented hubs [are lined up] carriage box to carriage
    box on the lane approaching the altar; and the jade, silk, and rich
    foods are box by box delivered to the shrine's que gate towers ....
       How could it be [considered to be] like the temple of Gao Tang,
    which had the empty title of Morning Cloud ...? ... (34)

For a large number of devotees, this site was vital not as a focus of ancestral worship but because of the springs. Numerous offerings were made thanking the spirit of the springs for continuous benevolence in providing water. The experience of the site was heightened by wondrous surroundings more beautiful than the famous magical island of Penglai and the mountain residence of the Spirit Mother of the West (Xiwangmu), Kunlang. The text of the stela also compares Jinci to the renowned goddess shrine in Song Yu's (fl. third century B.C.E.) rhapsody the "Gao Tang Fu." (35) Yet Li's text implies that although the shrine at Mt. Gao Tang and the powers of the goddess of the rhapsody Morning Cloud were fictional, Jinci, by contrast, was real and the deities worshiped there efficacious. Later scholars have suggested that by making this reference, Li Shimin was venerating a female divinity at the Jin Springs. (36) But Li's stela inscription falls short of overtly identifying her.

However, we do have evidence that in the second half of the tenth century the Song dynasty general Cao Han (924-992 C.E.) worshiped a female water spirit at Jinci. During the Song dynasty siege of Taiyuan in 979, Cao Han was short of drinking water for his troops. Cao prayed at a "Damsel's Temple" (niangzi miao) approximately three miles west of the city, then constructed a canal to retrieve the water, providing enough drinking water for his men and horses both. (37) This was the second attempt of the Northern Song emperors to take Taiyuan, at that time the capital of the Turkic Northern Han kingdom. The first attack was led in 969 by Zhao Kuangyin, the Northern Song dynasty founder, who used the method applied in a famous attack on the same city during the Warring States period (475-221 B.C.E.) of flooding the city walls by diverting water from the nearby Jin River. (38) Zhao Kuangyin also made use of water from both the Fen River and the spring at Jinci. (39) And in 979 his younger brother, the second Northern Song emperor, Zhao Kuangyi, employed the same method during his successful siege. (40) Although Jinci was not named specifically, given that a canal was built to divert the water from Jinci to the city of Taiyuan and that Cao Han was stationed on the west side of Taiyuan during the siege, within three miles of the Jin Springs, it is likely that the "Damsel's Temple" was a temple to the Spirit of the Jin Springs at Jinci. (41)

Following the conquest of Taiyuan, Zhao Kunagyi sponsored a rebuilding of Jinci on top of the earlier temple site at the Jin Springs; consequently, physical evidence for the preSong complex is no longer extant. (42) The stela inscription written in 984 to commemorate its completion indicates that the project was intended to bolster the legitimacy of Zhao Kuangyi's rule through supporting the Zhou dynasty regional founder, Shu Yu. (43) But the author also refers to a "crude" system of worship previously at the shrine and contrasts Zhao Kuangyi's rebuilding with the Qin (221-206 B.C.E.) and Han dynasty emperors' worship of illegitimate or feminine nature spirits. A close reading of the 984 text, like Li Shimin's inscription more than three centuries earlier, suggests other spirits were being worshiped at Jinci.

For Zhao Kuangyi the conquest of the Northern Han must have seemed in some ways parallel to the Western Zhou conquest of the territory of Tang/Jin. In the Zhou dynasty the territory of Tang/Jin was known as a border region inhabited by steppe peoples (as opposed to the sedentary Zhou); the Northern Han were Turkic and were backed by the seminomadic Khitan Liao dynasty. Zhao Kuangyi was the younger brother of the first emperor, just as the Duke of Zhou, the general who actually took the territory of Tang/Jin, was the younger brother of King Wu of Zhou. Not surprisingly, the inscription focuses not on water-spirit worship, but instead on the worship of Shu Yu of Tang. Thus, the only shrine named in the inscription was that of Shu Yu: (44)

Although this inscription stipulates that the imperial purpose for repairing the shrine was to pay respect to the founder of the Jin State and, by extension, legitimize Northern Song rule over his territory, the subtext reveals that other deities, in all likelihood female nature spirits, were important at the site. If Jinci were in fact dedicated exclusively to Shu Yu of Tang, there would have been little need to assert that rebuilding it was not related to nature spirit worship. Their presence made it necessary to emphasize that the emperor was interested only in the first Zhou dynasty ruler of the territory.

       ... [His majesty] addressed his officials and said, "Jinci is
    [dedicated to] the spirit of Shu Yu of Tang." ...
       ... Then [his majesty] noticed that the previous system of the
    shrine was crude; it was appropriate to order [the establishment of]
    a managing office so that they could enlarge and renew it ....
       Presently we looked at the completed main hall, and the galleries
    surrounding it ....
       The myriad brackets gather like constellations, the chiliad
    pillars shimmer like zao aquatic plants. (45) Bright walls,
    brilliant and crystalline as the autumn moon; jade lintels, colorful
    and dazzling as morning clouds. The quality timber and its beautiful
    decoration is now completed on a grand scale. Furthermore, in front
    it faces onto an angular pool, the spring is so deep one can fathom
    hundreds of feet. At the rear it embraces the steepened hill, the
    mountains and hills, like a screen, reveal heights of ten thousand
    yards ....
       How could [venerating Shu Yu of Tang] be compared to the Qin
    Emperor setting out on an expedition to sacrifice to the eight
    deities on the seas?
       The Han emperors preferred strangeness, they sacrificed to the
    Spirit of Earth at Fenyin .... (46)

As emperors, the official reason Li Shimin and Zhao Kuangyi patronized Jinci was to pay homage to the state founder venerated at the site. But they also acknowledge much more clearly than authors of earlier texts the significance of the spring spirit. As the eleventh-century emperors of the Northern Song moved to recognize local spirits in addition to those with national political significance, the Spirit of the Jin Springs was given official status. And in China, with official status comes eternal remembrance through the power of written histories.

Imperial Recognition and the Promotion to Sage Mother

The promotion of the water spirit at Jinci to Sage Mother follows a pattern of popular spirit elevation across the Song empire at the end of the eleventh century. Close examination of the evidence dated to the Song period shows that the Sage Mother was recognized by the emperor because of a local agrarian patronage base that depended on her for water. While the "Damsel's Temple," the site where Cao Han retrieved water for his troops in 979, was not specifically linked to Jinci in the Standard History of the Song, beginning in the eleventh century we have positive textual and physical records of a female spirit at the site. Although we do not know exactly when the goddess first became known by the name Sage Mother, in a document dated March 7, 1063, recording the extension of Jinci waterways, a pool located north of Eternal Youth Spring was referred to as "Sage Mother Pool" (Fig. 1: nos. 5, 6). (17) This was likely the pool located underneath the Flying Bridge directly in front of the Sage Mother Hall.

The Collected Important Documents of the Song (Song huiyao) lists the promotions accorded the Sage Mother and her shrine that allowed her name to be included in the official sacrificial statutes (sidian) for the rest of imperial Chinese history. The first of these was by the court of Emperor Shenzong (r. 1067-85) in 1077, when she was given the title Sage Mother of Manifest Aid. An imperial plaque with the name Temple of Compassionate Aid was officially bestowed in 1104. In 1111 the goddess was elevated to Sage Mother of Clear Efficacy and Manifest Aid. In 1112 a new imperial plaque was bestowed on the temple, changing its name to the Temple of Far-reaching Benevolence, a name it kept through at least the thirteenth century. (48)

The building itself contains evidence of the Sage Mother's significance during this time. While Peng Hai and Chai Zejun argue that the Sage Mother Hall was originally constructed during the rebuilding of the site completed in 984, the evidence from inscriptions and the style of the building confirm that the hall dates between 1038 and 1087. The main hall of the Shu Yu shrine complex discussed in the 984 inscription was located between a steep mountain and an angular pool. Because the only palatial hall (dian, [TEXT NOT REPRODUCIBLE IN ASCII]) presently located in a position matching that description is the Sage Mother Hall, Peng and Chai suspect that the Sage Mother Hall was originally the main hall, perhaps with both deities being worshiped within the one building. But Zhang Yipeng argues that the pre-thirteenth-century Shu Yu main hall was probably located to the south of the Sage Mother Hall, slightly behind Eternal Youth Spring--a location that would also be consistent with the text of the 984 stela. (49)

Inscriptions within the hall verify that the building was in place and dedicated exclusively to the Sage Mother by the middle of the eleventh century. An inscription on the back of the Sage Mother's seat documents the addition of six dragons to the front columns of the "Jinci Zhaoji Shengmudian" (Hall of the Sage Mother of Manifest Aid at Jinci) in 1087 (Fig. 8). (50) The patrons were members of the Golden Dragon Society of Taiyuan Prefecture who pooled their resources to embellish the building. Additionally, three more inscriptions have been found on roof timbers replaced during a repair after an earthquake in 1102, providing a terminus ante quem for the building. This repair was done with official imperial support at the request of the Taiyuan garrison commander. (51)

One final Song period document indicates that by the early twelfth century, the popularity of the Sage Mother had superseded that of Shu Yu, even on an official level. In a prayer of 1123 giving thanks for rain inscribed on a stela at the site (Fig. 9), the regional commissioner Jiang Zhongqian stated that the sacrifice took place in the shrine complex of the Sage Mother and Shu Yu of Tang. (52) Commissioner Jiang also stated that the Sage Mother was the foundation of the Jin State, and the spiritual power of Shu Yu could be known through the character seen on his hand at birth. Because of the eponymous relationship between the Jin River and the Jin State, the Sage Mother could be understood in this text as being the mother of the state (and therefore the mother of its first ruler) as well as the goddess of the spring of the Jin River. (53) The link here is an important one, because in the seventeenth century this interpretation of the stela inscription was used as textual evidence that the Sage Mother was the mother of Shu Yu rather than just a water sprite, thereby justifying her dominance over the site. (I discuss this aspect further below.)

[ILLUSTRATION OMITTED]

Other information we have about the eleventh-century history of the site comes from later sources. The most frequently cited is a 1267 stela inscription by Yi Gou documenting a reconstruction of Shu Yu of Tang's shrine complex in that year. According to the inscription, information was gathered from local archives as well as village elders. The section regarding the tenth- and eleventh-century history of Jinci reads:

[ILLUSTRATION OMITTED]

  ... From the sixth year of the [Later] Jin Tianfu reign period
  [941 C.E.] [Shu Yu of Tang] was enfeoffed as the King of Arising
  Peace. After the Song Tiansheng reign period [1023-32 C.E.] his title
  was changed to King of the East of the Fen River. Furthermore, [they]
  rebuilt the Gentlewoman Shrine [nulangci] west of the spring facing
  east. During the Xining reign period (1068-78) the title "Sage Mother
  of Manifest Aid" was first added, by this time her rank was already
  clear .... (54)

The inscription provides the first clear evidence of the Sage Mother Hall being constructed, or at least reconstructed, as a building facing toward the east. The two externally verifiable dates in this inscription, the elevation of Shu Yu to King of Arising Peace and the elevation of the Sage Mother to the Sage Mother of Manifest Aid, are accurate, which indicates that the sources Yi Gou depended on for the history of the shrine were likely reliable. (55) If so, the Sage Mother would have been worshiped independently in the Sage Mother Hall beginning at some point between 1032 and 1087 (the date of the addition of the dragons to the front columns).

Finally, the style of the building's bracketing is not consistent with that of the tenth and early eleventh centuries in the region of southern Shanxi but, rather, is closer to structures in this area from the end of the eleventh century. (56) According to Yun Shan, a severe earthquake in 1038 caused serious destruction in the Taiyuan region. (57) If Yi Gou is correct and the Gentlewoman Shrine was rebuilt (fujian) after the Tiansheng period, the present Sage Mother Hall was probably constructed on an older foundation of a temple dedicated to the Jin Springs water spirit after the 1038 earthquake. We can say with relative certainty that the remains of the hall are from 1038-1102, but, given that the dragons around the columns survived the 1102 earthquake, a closer range would place the construction of the present building between 1038 and 1087. (58)

The pattern of the Sage Mother's promotion during this period suggests that local people identified her as a powerful provider of rain and, in addition to rebuilding and enhancing the Sage Mother Hall, petitioned the central government to give her title increases so that she would continue acting on their behalf. In Changing Gods in Medieval China, 1127-1276, Valerie Hansen describes a dramatic increase in the official granting of titles to popular deities in the Song beginning in 1075. At this time divinities proven to have worked miracles were registered with the government and included in official sacrifices to be undertaken by a local official biannually, in the spring and autumn. This was a bottom-up rather than a top-down process, as documentation of miracles depended on local officials to investigate and verify miraculous events related to them by their constituencies. (59)

According to Hansen, the increase in title granting to popular divinities during the 1070s resulted from the implementation of the New Policies of Wang Anshi, which were aimed at reorganizing and strengthening systems of local authority. (60) Although titles had been granted to divinities from as early as the Qin dynasty, the practice was relatively limited. Early Northern Song rulers followed the tradition and extended it to popular spirits in the eleventh century, beginning with edicts in 1047 ordering popular temples (miao) to be repaired across the empire and then, after a drought in 1050, requesting that the names of all spirits who brought rain be reported to the government. But prior to 1075 few titles were granted because it was difficult to gather information about miracles and the deities responsible for them from local authorities. Another edict was issued on December 16, 1074, after a serious drought in north China earlier in the year, requesting the names of deities in shrines and temples (cimiao) known to respond efficaciously to requests for rain. By this time the New Policies were fully in place, and the central government received a large number of names; in 1075 alone thirty-seven deities were granted official titles. (61) Title granting continued at this high level at the end of the eleventh century and into the twelfth, when the New Policies were first dismantled in 1086 and then revived in 1094 due to a power shift at court. (62)

The elevation of the Jin Springs water spirit closely follows this pattern, suggesting that the local population responded to the edicts of the central government in order to reward the spirit for her miracles and ensure her continued benevolence. The Gentlewoman whose shrine was rebuilt to the west of the Jin Springs after 1032 was promoted to Sage Mother before 1063, perhaps as a result of the 1050 edict requesting the reporting of efficacious rain spirits. She was given an additional noble title three years after the 1074 edict. Then, after the revival of the New Policies, the central government sponsored her hall's repair after an earthquake, issued official plaques to her temple, and added two more characters to her title. Because of local requests, which were facilitated by the reorganization under the New Policies, the Spirit of the Jin Springs was entered into the sacrificial statutes, where her cult remained officially sanctioned for the rest of imperial Chinese history. (63)

The Shu Yu of Tang Shrine Complex

By 982, once the provincial capital of Taiyuan was moved north of Jinyang and the eponymous Jin River, and Song dynasty territory extended another eighty-two miles north of that, the region was no longer either a national or provincial capital or a border outpost. As a consequence, over the next century Shu Yu of Tang began to lose his place as the foremost deity at Jinci. This seems to have been a gradual process, indicating that Shu Yu's patronage base was tied to regional political power. For Zhao Kuangyi, Shu Yu was important because he was the legitimate historical founder of the region and rebuilding his shrine served to legitimize the Song dynasty rule of the territory. As mentioned above, some have speculated that Shu Yu was originally worshiped in the present Sage Mother Hall, as that building matches the description of Shu Yu's main hall found in both the Commentary on the Book of Waterways and the 984 stela inscription. (64) But we have little physical evidence of the 984 rebuilding, and Shu Yu is now worshiped in a complex that does not match the 984 description. The earliest record we have of Shu Yu's complex being in the location it now occupies is from the thirteenth century, and we know that it was built at a scale much reduced from that of the Song period shrine. After 1032 Shu Yu was elevated on one occasion, when his title was changed to King of the East of the Fen River (Fendongwang). At the time the regional seat of government was located on the east side of the nearby Fen River, whereas Jinci and the Jin River have always been located to its west. Thus, in the eleventh century Shu Yu was thought to rule over the political capital but not the site of the springs. (65)

With Shu Yu's apparent significance in politics rather than water production, it seems anomalous for the "Prayer in Thanks for Rain at Jinci" of 1123 to mention Shu Yu at all, even subordinated to the Sage Mother. But the historical context of the prayer helps to explain his inclusion. The text was composed in June 1123, just after the "return" of the long-covered Sixteen Prefectures to the Song court but before the Jurchen Jin conquest of north China. (66) The Sixteen Prefectures were located in the borderland between Song and Liao dynasty territory, and the rulers of the Song dynasty passionately wanted them in order to complete their empire. Jiang Zhongqian mentions the return of the Sixteen Prefectures twice in the text of the prayer. I suggest that Shu Yu of Tang is included in this document primarily because of the overwhelming importance of the moment when the empire was finally unified under Chinese rule. Jiang Zhongqian believed that the Sixteen Prefectures had been given to the Song through the help of that member of the Zhou royal family who ruled over the northern borderland, Shu Yu. After the fall of north China to the Jin, Taiyuan did not regain its former role as a border capital; from the end of the tenth century, the dynastic lines fell either north or south of the area. Perhaps as a consequence, Shu Yu no longer drew the imperial attention he had when the rulers had used Taiyuan as a central base of power or an important point of conquest.

[ILLUSTRATION OMITTED]

Although without extensive archaeological research we do not have enough data to determine the location of Shu Yu's shrine prior to the thirteenth century, the 1267 stela inscription mentioned above indicates that by then, Shu Yu of Tang had his own complex with a front gate and retiring hall. (67) Yi Gou notes in this document that Shu Yu of Tang's shrine had fallen into ruin and that they did not have the funds to restore the complex to its previous glory. (68) The shrine was repaired again in 1342 to thank the deities of Jinci for providing rain. The documentation of this fourteenth-century construction project stated overtly that the shrine to Shu Yu was oriented north-south. Because the description of the canal and other buildings in this stela inscription is similar to drawings of the site two centuries later (Fig. 10), we can be fairly certain of the shape of the complex from at least this time. (69)

In the beginning of the Ming dynasty (1368-1644), the identities of both divinities as understood at that time were made overtly clear. Because of a miracle rain in the Taiyuan area in 1369, an official of Shanxi requested an increase in the Sage Mother's title. When his request was granted, her official title became the Sage Mother of Broad Benevolence, Clear Efficacy, and Manifest Aid. (70) Only a year later, the first emperor of the Ming dynasty, Zhu Yuanzhang (r. 1368-99), desired to rectify the excessive bestowal of honorific titles to deified humans as well as spirits of the landscape. He proclaimed:

   The sacred peaks, the guardian mountains, the oceans and great
   rivers, and all the high mountains and wide waters were created by
   Heaven and Earth. Their brilliant and numinous material force [qi]
   was concentrated to form their spirits, and these received their
   destiny from the Lord-on-High. This is a profound mystery that none
   can fathom. How can anything be added to [these spirits] by
   investiture or the bestowal of honorific titles by the ruling house?
   In profanation of the rites, nothing could be more unclassical than
   this. Although we may properly invest, or bestow titles upon, loyal
   officials and brave soldiers, [even] this may be done only when it is
   justified. Now, it is by the ceremonial that the correct names and
   stations of spirits and humans are made clear, and these may not be
   violated. (71)

Following this proclamation, by 1371 the title King of the East of the Fen River was stripped from Shu Yu and he then officially became the Spirit of Shu Yu of Tang (Tang Shu Yu zhi shen) in the sacrificial statutes. Likewise, the title Sage Mother of Broad Benevolence, Clear Efficacy, and Manifest Aid was eliminated as a result of this order and officially replaced with the Spirit of the Jin Springs (Jinyuan zhi shen). (72)

Zhu Yuanzhang's proclamation shows that, at least during the fourteenth century, the Sage Mother was thought to be a water spirit and Shu Yu a deified historical figure, regardless of how elaborate their titles had become. The practice of title granting as a means of paying respect to and rewarding a deity was not eradicated by Zhu, however, and although the Sage Mother was officially the Spirit of the Jin Springs, when the Jingtai Emperor wrote an imperial document thanking the goddess for bringing rain in 1451, she was again called the Sage Mother of Broad Benevolence, Pure Efficacy, and Manifest Aid. The buildings of Jinci were repaired again in 1461 as a reward for providing another timely rain. (73) From the end of the Song dynasty through the first two centuries of the Ming, the main focus of worship at Jinci was the water goddess the Sage Mother of Manifest Aid. Shu Yu of Tang was remembered in stela inscriptions commemorating repairs to the temple buildings, but those inscriptions, as well as the jottings of literati who traveled to the site of the Jin Springs, also make clear that the Sage Mother dominated the site during this time.

Transformations of the Late Ming Dynasty

We know that at the end of the Ming dynasty, three different constituencies used temple construction at Jinci to assert different interpretations of the Sage Mother's identity in order to align the deity and her power over both spring and rainwater with their competing interests. This information comes from the earliest extant local gazetteer of the region, Gao Ruxing's Taiyuan xianzhi from 1551, and from the new additions to the complex in the late sixteenth and early seventeenth centuries. In the gazetteer, we read that in 1548 Shu Yu's shrine complex was repaired again by a group of administrative officials and local literati. (74) A map of the area published with the gazetteer, only three years after this restoration, shows an interesting distinction between the Sage Mother Hall and the new Shu Yu shrine complex (Fig. 10). The buildings of Shu Yu's complex are arranged in a standard walled-courtyard format, with the main buildings on axis. The Sage Mother Hall, by contrast, is shown without a courtyard enclosure. (75) But the literati were not the only patrons of Jinci. In 1561 the rest of the complex was repaired in a combined effort of nobility and locals. The common people, led by the villager Wang Wentai in 1563, began their own fund for the addition of a Dressing Tower (literally, comb and wash tower) for the Sage Mother. (76) The extant building dating from this time is a two-story tower located directly west of Eternal Youth Spring, open to the water and independent of any central axis (Fig. 5). Within the Dressing Tower hall the local people asserted their own story of the historical origins of the Sage Mother--that the Spirit of the Jin Springs was originally a local girl from whose magical vat an endless supply of water flows. A bronze statue of a young girl sitting on a lotus seat emerging from a large vat, also thought to be original to the 1563 construction, occupies the main image niche on the lower floor. (77)

The construction projects of this period did not end with the erection of the Sage Mother's Dressing Tower, however. A memorial gateway was added in front of the offering hall and on axis with the Sage Mother Hall in 1576. The inscription on the gateway, sponsored by local officials and Confucian school masters, (78) reads "duiyue" ([TEXT NOT REPRODUCIBLE IN ASCII]), a phrase from "Qingmiao" (Pure Temple), the ode to the ancestral temple in the section Sacrificial Odes of Zhou of the Book of Odes (Shijing, Figs. 11, 12):

   Ah! Solemn is the ancestral temple in its pure stillness Reverent
   and harmonious were the distinguished assistants; Great was the
   number of the officers: [All] assiduous followers of the virtue of
   [King] Wen. In response to him [duiyue] in heaven, Grandly they
   hurried about in the temple. Distinguished is he and honored, And
   will never be wearied among men. (79)

For all those familiar with the passage, the use of this phrase on the eastern side of a memorial gateway leading to the offering hall of the Sage Mother served to elevate the status of the temple to that of a Zhou ancestral shrine complex. Its placement asserts an opinion about the identity of the goddess at the end of the axis: she is not a water goddess, she is an ancestral figure. She is not a transformed common girl with a magical vat, she is a member of the Zhou royal clan.

But for the people of the region who depended on her for water, such a powerful goddess was not so easily elevated to royalty. A drought in the spring and summer of 1582 caused great strife among the people. In keeping with his responsibilities as a local official, Sun Luyang went to numerous local temples to pray for rain, to no avail. Finally, he was encouraged to pray to the Sage Mother for rain. That he did not go to her in the first place may be some indication that she had been appropriated into the ancestral cult, at least for some. Yet her efficacy as a water spirit was made clear when it began to rain before the ceremony to her was completed. Local government officials then felt compelled to do something for the goddess in return for her benevolence, and a single sacrifice was deemed insufficient. Fortunately, a member of the local gentry, Gao Yiqi, happened to have a legal case under dispute at the commandery, and the officials brought up the possibility of renovating the temple with Gao. Gao was generous enough to provide the financial backing to have the temple restored. He also used local labor, thereby providing work for the unemployed (perhaps in such condition because of the drought), and in three months it was complete. Ma Chaoyang, a relative of the Gaos, returned from a tour of inspection in the south to hear of the tale. Although he thought repairing the temple was an appropriate action to reward the gift of the goddess, there were some in his circle who considered this to be excessive and argued that only heaven can provide rain; the local spirits were merely tricking the people in order to gain their attention and offerings. Clearly, not all were happy with the Sage Mother's popularity, and neither was there a consensus about her standing. (80)

[ILLUSTRATION OMITTED]

Twenty-four years later, when the addition of bell and drum towers further emphasized the east-west axis leading up to the shrine, (81) the identity of the deity at the end of this axis remained in dispute. While she was still known at that time as the Goddess of the Jin Springs, the local elite continued to emphasize that the shrine complex was dedicated to a Sage Mother ancestress who could be no other than Yi Jiang. The creation of a formal central axis with gateways and symmetrically placed subsidiary buildings strengthened their claim to the temple as an ancestral shrine.

[ILLUSTRATION OMITTED]

Religious Affiliation and the Form of the Complex

The influence of the Jin Springs can be seen in the organization of the site. Buildings that were added to the larger compound over the next two hundred years included two major temple complexes dedicated to nationally venerated deities, one to Wenchang (patron saint of scholars) and the other to the Jade Emperor, the Three Pure Ones, and Guandi (Figs. 6, 7). These complexes were similar to Shu Yu's in their standard courtyard style and enclosing walls. They also share the general north-south orientation of the Shu Yu shrine complex. Yet none of the three complexes is oriented exactly north-south, and their central axes are not parallel to each other. Their only common feature is a (visually) perpendicular relationship to the body of water located in front of them, emphasizing the significance of the water to the site of Jinci. Pictorial evidence showing the fully developed complex in the eighteenth century indicates that neither the Sage Mother Hall nor her Dressing Tower was ever confined in such a compound. They remained open to the whole of the sacred site. (82)

[ILLUSTRATION OMITTED]

The unusual distribution of buildings at Jinci is related to the divinities to whom each was dedicated. The national cults that have no intrinsic association with the site are sequestered in nonlocalized architectural complexes. The courtyard complexes at Jinci create a generic environment for the worship of divinities that bear little direct relation to the unique location of the Jin Springs. The use of the courtyard complex for nationally venerated divinities corresponds to the universal application of belief systems shared by a national culture. (83) Visitors from near or far did not need to be instructed in how to navigate this space. The complex is part of a nationalized vocabulary of religious architecture in which the ritual progression through buildings is consistent from complex to complex, region to region, and religion to religion. (84) These complexes, whether intended for the state cult or for Buddhist and Daoist monasteries, share the spatial vocabulary of the standard ritual complex in China.

By contrast, the individual temple buildings specifically devoted to local divinities, such as the Dressing Tower, are without enclosing courtyards. They stand open to the unique topography of the site, with its famous springs and waterways. As the first of the Zhou rulers of what would become the Jin State, Shu Yu is a local god with a direct relation to the area of Taiyuan and the Jin River. Yet Shu Yu's shrine complex, like those of deities that are venerated nationally, is isolated from the site in a walled courtyard (Fig. 13). The national culture of literati who regarded him as the most important figure of the area built a shrine for him that reflected their perceptions of his status. Because of the expectations of these men (and their patronage), the shrine to Shu Yu was made into a standard, nonlocalized courtyard complex.

The importance of the Spirit of the Jin Springs to the local population is reflected in her worship hall, which competes with Shu Yu of Tang's shrine in grandeur and dominance over the entire complex. Despite the lack of positive textual documentation for a temple dedicated to a female spring spirit at Jinci prior to the eleventh century, given the pattern of spring-spirit temples in and around China, it is likely that such a temple had a place on the site of Jinci from the beginning. Rolf Stein has described the commonality of small temples to female divinities at "meeting places" near pools, waterways, and crossroads in southeastern China and Vietnam. These deities were "often anonymous, being designated only by vague terms like thanh-mau ([sage] mother), ba (lady), durc-ba (noble lady), or chua-ngoc (jade princess)." (85) In the agricultural communities of premodern China, a temple to a water spirit was located at the site of every major body of water. In Shanxi, temples to spring spirits were commonly placed on land directly above the site of the spring. (86) Although few ancient water-spirit temples survive to the present, two other early examples are the Temple to the King of Broad Benevolence (Guangrenwang miao) situated above the Five Dragon Spring in Ruicheng, Shanxi, dated to 831 C.E. (Fig. 14), and the King of Brilliant Response Hall (Mingyingwang dian), the main hall of the Water God's Temple (Shuishen miao) located at Huo Spring beside the Lower Guangsheng Monastery in Zhaoxian, Shanxi Province, dated to 1324 C.E. (Fig. 15). (87) In both cases the spring is located in front of the temple buildings.

In the earliest extant documentation of Jinci in the Weishu, the shrine was dedicated to the "King" of Jin, but this does not necessarily mean that it was dedicated primarily to Shu Yu of Tang. Shu Yu was not elevated to king during his lifetime, and we have evidence to show that this did not happen before the sixth century, perhaps not until his elevation to King of Arising Peace in 941. (88) The word king (wang) can mean more than just a temporal leader or member of the aristocracy, however. From at least the third century B.C.E., it could also refer to the multitude of spirits, the baishen, which were often spirits of mountains or rivers. (89) Thus, only two interpretations of the record in the Weishu are possible. Because Jin is the name of both the state and the river, Jinwang could be either the ruler of the Jin State or the Spirit of the Jin Springs and River. We have no example of wang referring to a queen, the position of Yi Jiang during the Western Zhou dynasty, which means it is unlikely that the Sage Mother was worshiped as Yi Jiang during this time. (90)

[ILLUSTRATION OMITTED]

In the Commentary on the Book of Waterways, Li Daoyuan expressed his opinion in support of the first of these definitions by relating how Shu Yu of Tang came to be the ruler of the future Jin State and the object of veneration at Jinci. But the record that survives in the dynastic histories and early geographic texts gives exclusively this interpretation, mentioning only orthodox sacrifices to the first ruler of the region who was a member of the Zhou dynasty royal family. Those who read these texts came away with the impression that Shu Yu of Tang was the primary object of veneration at Jinci and, therefore, the main hall of Jinci would have been dedicated to him. It was this interpretation that left the traveling gentry of the late Ming and Qing confused about the worship of goddesses and the prominence of the Sage Mother Hall.

The record (and lack thereof) of the Spirit of the Jin Springs and her elevation from a Damsel to Sage Mother is consistent with the pattern practiced by the Northern Song court during the late eleventh and early twelfth centuries of documenting and promoting local divinities. (91) From the second half of the eleventh century, she was acknowledged as the most powerful divinity on the site, especially for bringing rain. By the seventeenth century the Sage Mother had acquired a shrine "complex" developed far beyond that expected for local water spirits. The axis of approach (Fig. 16), framed with symmetrically disposed bell and drum towers and punctuated with a memorial gateway inscribed with a phrase from the Book of Odes (Fig. 17), told the visitor that the deity at its apex was the focus of Jinci, which, were textual accounts to be believed, should have been an ancestral shrine. Standing at the end of a formal axis but still open to the springs and canals that give Jinci its relatively natural setting (Fig. 18), the Sage Mother Hall was never fully transformed into the hall of an official deity by being placed inside a walled compound.

[ILLUSTRATION OMITTED]

The addition of the Duiyue memorial gateway and the bell and drum towers did not convince those dependent on the Sage Mother's efficacy as a water goddess that she was an ancestral figure, but they were enough to convince local elites who desired to interpret her as an ancestral mother. In 1672 Zhou Lingshu was serving as the prefect of Taiyuan when he visited Jinci. He, too, found the shrine complex of Shu Yu in disrepair and the people dedicating all of their attention to the Sage Mother. He suggested that the ideal solution would be to abolish all of the temples except for Shu Yu's. The next best action would be to switch the statues and worship Shu Yu in the Sage Mother Hall and put the Sage Mother in another hall. However, if resources would not allow for these changes, then it would be preferable if everyone understood the Sage Mother to be the mother of Shu Yu. For Zhou Lingshu, the worship of the Sage Mother as a water spirit represented a degradation in her status since the Song dynasty (and before), when the "Prayer in Thanks for Rain" of 1123 seemed to identify her as Yi Jiang. (92) Zhou Lingshu's argument was promulgated by the well-known scholar Yan Ruoqu (1636-1704), whose ancestral home was in Taiyuan Prefecture. (93)

Yan Ruoqu wrote an afterword to Zhou Lingshu's essay commemorating the restoration of the pavilion that contained Li Shimin's stela. After rereading Zhou's inscription he went to the county gazetteer and examined the "Prayer in Thanks for Rain." He was elated to find that Zhou Lingshu's assertion that the Sage Mother should be venerated as the mother of Shu Yu had a textual, historical precedent. (94) In a separate essay, "An Explanation of Territory in the Four Books, continued [Sishu shidi xu]," Yan has a section specifically devoted to Yi Jiang, where he tells of her temple at the Jin Springs, only six miles from his ancestral home. He explains that the addition of the Sage Mother Hall on the right (west) side of the Shu Yu shrine complex was the son submitting to the domination of his mother. (95) Yan may also have had a personal agenda in interpreting the Sage Mother as Yi Jiang. The Yan surname is one of seven that trace their ancestry back directly to Shu Yu of Tang. (96) By making the Sage Mother the mother of Shu Yu, he transformed her into the direct ancestress of his clan.

[ILLUSTRATION OMITTED]

This particular act of scholarship was not integrated into the history of the site immediately, (97) and the evidence we have indicates that the local farmers never understood her in this way. (98) However, Yan wanted his interpretation of the Sage Mother's identity to be used in future county gazetteers, and eventually this became a reality. In his 1826 Taiyuan County Gazetteer, Yuan Peilan states that the "Spirit of the Jin Spring" was, in fact, the spirit of the mother of Shu Yu, Yi Jiang. (99) This was, consequently, the interpretation adopted by twentieth-century literati, that is, historians of Chinese art and temple culture, in both China and abroad. (100)

[ILLUSTRATION OMITTED]

The physical record continues to assert a different interpretation. The site of Jinci encompasses the Jin Springs, the fountainhead of the Jin River. Although a shrine complex to Shu Yu of Tang was built there, from at least the twelfth century his was not considered the most important and was repeatedly falling into disrepair. As Zhou Zaijun discovered in 1665, his shrine was neither well marked nor in an obvious location. The oldest and most prominent building at Jinci (and the easiest to locate) was the Sage Mother Hall, built over one of the springs on the site. Although the hall sits at the end of a major axis, it relates to the site directly. As the Dressing Tower, with its image of the water-providing village girl, sits west of, but open to, Eternal Youth Spring, so the Sage Mother Hall sits west of a pool originally filled by the spring flowing from underneath the building (Fig. 19). As subsequent temple complexes were added to Jinci, their physical orientation was also not defined by the cardinal directions. Instead, like the earlier temple buildings, new additions were positioned to relate to the springs and the flow of the water issuing from them. The orientation of the shrines, whether nonlocalized building complexes or individual halls, were all subordinated to the emerging Jin River.

[ILLUSTRATION OMITTED]

Reading the Site as a Primary Source

Through examination of the physical evidence at the Jinci shrine complex, it becomes apparent how textual evidence can obfuscate the primacy of local divinities over those more relevant to the educated elite. Early geographic texts and the standard histories discussed the aspect of worship at the site most significant to aspiring emperors, who were conscious of the classical history of Zhou dynasty China and its first regional ruler, Shu Yu of Tang. The authors of these texts had no interest in the popular worship of feminine water spirits. To understand the importance of such local spirits, we must look to the site itself. Stelae extant at Jinci consistently express the importance of the spring in addition to their mention of the cult of Shu Yu of Tang. These stelae also reveal that other divinities were worshiped at Jinci. Veneration of water spirits is explicit in the text of Li Shimin's stela inscription, and it is implicit in the stela inscription of Zhao Kuangyi's 984 rebuilding of the complex, which notes a "crude" previous system and cautions the visitor against comparing the imperial attention to Jinci with the Qin and Han rulers' veneration of unorthodox divinities and feminine nature spirits.

During the eleventh century, the water goddess was entered into the government sacrificial statutes through imperial recognition of her miracles and elevated from a mere Damsel to Sage Mother. Entry into the government record meant that she was a proven source of water, through the springs underneath her temple and through rain. The local magistrate had the responsibility of making offerings to her, especially in times of drought. It also meant that she was part of the official textual record--that is, she could not be written out of the history of the site. From the thirteenth century onward the combined architectural and textual evidence shows that the Sage Mother enjoyed unquestioned dominance over the site: her hall was the largest building, and her efficacy in providing water made her the foremost object of veneration to the majority of devotees. At the same time, the political importance of Taiyuan declined, as did the cult of Shu Yu. Although he was also in the sacrificial statutes and would have been included in the spring and autumn sacrifices, he did not attract enough patronage to keep his shrine maintained over the centuries that followed. Evidently, maintenance of timber-frame buildings required more broad-based local interest in the deity than was displayed by followers of Shu Yu's cult, unlike those of the Sage Mother's, who not only kept her shrine in good repair but also enhanced it.

[ILLUSTRATION OMITTED]

In the sixteenth century, an attempt was made to create a more legitimate identity for the most popular deity at the site. This was done through the manipulation of a common architectural vocabulary. A memorial gateway marked with a quote from the classics and a central axis emphasized by bell and drum towers gave the undeniably dominant Sage Mother Hall a legitimate feel that would have been considered appropriate only for an ancestral figure such as Shu Yu or his mother. The early generals and emperors and the Confucian elite of the sixteenth through seventeenth centuries shared a common interest in presenting the site as dedicated to Zhou dynasty ancestor figures, albeit for different reasons.

The popular voice, however, is revealed through the development of the site. The springs and canals that flow through the compound give the site an unpredictable, natural quality, and it is this flow that determines the organization of buildings and building complexes. Each of them is thus subordinated to the ever-flowing wealth of the place: the Jin Springs. Although the Sage Mother Hall was given a ritual axis of approach, its dominance of the site comes from its placement at the head of the springs and the singular grandeur of the building itself. The textual record of Jinci was confusing because the message of the site itself was so clear: this was the origin of the Jin River and the residence of the Spirit of the Jin Springs.

Notes

I would like to thank Chang Che-chia, James Robson, John Kieschnick, Peter Lorge, and Poo Mu-chou for their helpful comments at various stages in the revision of this paper. Thanks also to Victor Mair, Terry Kleeman, and Virginia Bower, who read and commented on my dissertation, of which the material for this paper forms a large part. Nancy Steinhardt deserves special acknowledgment for her unwavering support throughout and her ability to come through even in times of crisis. Research for this project has been supported by grants from the Committee on Scholarly Communication with China, the University of Pennsylvania, and two years of office space and invaluable intellectual support given freely by members of the Institute of History and Philology, Academia Sinica, Taiwan. Without the aid of the people within these institutions it would have been impossible to complete this project. The final versions of this manuscript have been greatly improved with the help of H. Perry Chapman, Lory Frankel, and the two anonymous readers for The Art Bulletin; I am grateful for the time they took to make such useful comments.

Unless otherwise indicated, translations of quoted material are my own.

(1.) Debate over the location of the capital of the original Zhou fief of Tang, which later became Jin State, has figured prominently in standard histories of the area for more than a millennium. From as early as the Western Han dynasty one tradition asserted that it was the Han regional capital of Taiyuan at Jinyang, approximately nine miles southwest of modern Taiyuan. For a summary of the arguments, see Miller, 168-90. The city of Taiyuan was moved northeast to its modern location shortly after the 979 destruction of Jinyang by Northern Song troops (see below).

(2.) The name of the Jin State founder was Ji Yu, [TEXT NOT REPRODUCIBLE IN ASCII], but he is typically referred to in classical sources as either Tang Shu, titles that refer to the territory where the individual was enfeoffed and his position in the royal family, or Tang Shu Yu, adding his given name to the first two names commonly used. I am following William Nienhauser Jr. by rendering each character as a separate name rather than combining the last two into what might appear to the modern reader to be a two-character given name, Shuyu, although this is seen in some sources. See Nienhauser et al., trans., The Grand Scribe's Records (Indianapolis: Indiana University Press, 1994), 65 n. 114. For further clarity I have emphasized that Tang was Yu's fief and not his surname by placing Tang after his name.

(3.) Liang Sicheng and Lin Huiyin, "Jin-Fen gujianzhu yucha jilue" (A summary record of the preliminary investigation into the ancient architecture of the Jin-Fen region) (1935), reprinted in Liang Sicheng wenji (The collected essays of Liang Sicheng), vol. 1 (Beijing: Zhongguo jianzhu gongye, 1982), 336.

(4.) Following McNair, I have translated Shengmudian as Sage Mother Hall. Because the word sheng, used to describe Confucius, is commonly translated as "sage" in this context, the relation is important for understanding how this water spirit could be understood to be an ancestral mother by an elite educated in the Confucian classics. Yet sheng conveyed the meaning of "holy" as well as "sage", and the Shengmudian has consequently also been rendered as Holy Mother Hall in other studies. See, for example, Nancy Steinhardt, Liao Architecture (Honolulu: University of Hawai'i Press, 1997), 189, 217-18.

(5.) The pictorial histories of Jinci and modern dictionaries of famous historic sites follow the elite interpretation and describe the Sage Mother as Yi Jiang. These include Liu Yongde, Jinci Fengguang (Scenes of Jinci) (Taiyuan: Shanxi renmin, 1958), 35; Chai Zejun, ed., Jinci (Beijing: Wenwu, 1958), 4; and Zhongguo mingsheng cidian (A dictionary of famous sites in China), 2d ed. (Shanghai: Shanghai cishu, 1989), 126. Consequently, so have 20th-century scholars of Chinese art and architecture in the West, such as Amy McNair and Craig Clunas, Art in China (Oxford: Oxford University Press, 1997), 112-13. The most recent works on Chinese architecture also follow this interpretation. See Chai Zejun et al., eds., Taiyuan Jinci Shengmudian xiushan gongcheng baogao (Report of the repairs on the Goddess Hall of Jinci Temple in Taiyuan) (Beijing: Wenwu, 2000), 5; and Fu Xinian et al., Chinese Architecture (New Haven: Yale University press, 2002), 150.

(6.) The primary history of the site is the classical examination by Liu Dapeng (1857-1942), Jincizhi (Jinci gazetteer) (Taiyuan: Shanxi renmin, 1986). This three-volume study of the complex by Liu (published posthumously) includes histories of all of the building complexes and transcriptions of all the stela inscriptions extant at the site. In the second half of the 20th century there were a number of basic treatments of the site's history produced for tourists, of varying quality, including the excellent one by Liu Yongde (as in n. 5); and Ding Mingyi et al., Jinci (Beijing: Wenwu, 1981).

(7.) Essays that focus on the remains of Jinci in terms of their art historical value include Qi Yingtao, "Jinci Shengmudian yanjiu" (Research on the Shengmudian at Jinci), Wenwu jikan 1 (1992): 50-68; and Peng Hai, "Jinci Shengmudian kance shouhuo--Shengmudian chuangjian niandai xi" (The reward of the Jinci Sage Mother Hall survey--the date of the first construction of the Sage Mother Hall), Wenwu 1 (January 1996): 66-80. The latest monograph, documenting the full restoration of the building from 1993 to 1996, is Chai Zejun et al. (as in n. 5).

(8.) The Sage Mother Hall is rectangular in plan, measuring 89 ft. 5 1/4 in. (26.83 m) north-south along the front facade and 70 ft. 8 in. (21.20 m) east-west along the side facades; see Peng Hai et al., Jinci wenwu toushi (A cultural perspective of Jinci) (Taiyuan: Shanxi renmin, 1997), 176-77. The largest extant Northern Song building is the Buddhist Moni Hall of Longxing Monastery in Zhengding, Hebei Province, and it is approximately 133 ft. 4 in. by 123 ft. 4 in. (40 by 37 m), including the porticoes that project from each of its four facades. See Liang Sicheng, A Pictorial History of Chinese Architecture, ed. Wilma Fairbank, bilingual ed., trans. and ed. Liang Congjie and Sun Zongfan (Beijing: Zhongguo jianzhu gongye, 1991), 77. The Temple to the King of Broad Benevolence (Guangrenwang miao, also known as the Five Dragon Temple, or Wulong miao) is dedicated to the Spirit of Five Dragon Spring (also called Dragon Spring) in Ruicheng, Shanxi, and its main hall dates from 831 C.E. But it is a much smaller building than the Sage Mother Hall, roughly 46 ft. across the front facade and 17 ft. 4 in. along the side facades (my measurements). For more on the Temple to the King of Broad Benevolence in Ruicheng, see Jiu Guanwu, "Shanxi Zhongtiaoshan nan Wulongmiao" (The Five Dragon Temple south of the Zhongtiao Mountains in Shanxi), Wenwu cankao ziliao 11 (1959): 43-44.

(9.) See esp. the important role of the Sage Mother Hall in Liang Sicheng et al., Yingzao fashi zhushi (The Yingzao fashi annotated and explained) (Beijing: Zhongguo jianzhu gongye, 1983; reprint, Taipei: Mingwen shuju, 1984).

(10.) In addition to figuring prominently in surveys of Chinese art and sculpture history, separate works treating these sculptures include Amy McNair (as in n. 4), 238-53; and Zhao Zhiguang et al., eds., Zhongguo Jinci Songsu (China Song dynasty statues at Jinci) (Taiyuan: Shanxi renmin, 1993).

(11.) The first work to diverge fully from this interpretation and include the popular religious traditions as well as those of the elite was jinci wenwu toushi, written by Peng Hai, an archaeologist and historian working in architectural preservation in Shanxi, and a group of local scholars working on the restoration of the Sage Mother Hall from 1993 to 1996 (Peng Hai et al. [as in n. 8]). Peng Hai also concluded that interpretation of the Sage Mother as Yi Jiang was representative of a 17th-century reading of the site. The publication of the Jincizhi in 1986 has been instrumental in facilitating a more nuanced understanding of the competing patronage groups at Jinci for Peng, as well as other scholars, primarily because of Liu Dapeng's transcription of all of the stela inscriptions at the site. The modern Chinese historian Henrietta Harrison has come to similar conclusions about the manipulation of the Sage Mother's identity by elite patrons of the site through her reading of Liu's work. See Harrison, "Village Identity in Rural North China: A Sense of Place in the Diary of Liu Dapeng," in Town and Country in China: Identity and Perception, ed. David Faure and Tao Tao Liu (New York: Palgrave, 2002), 96-97.

(12.) Prasenjit Duara used the term "superscription" to describe the accretion of different identities in the cult of Guandi. According to his theory, because no new interpretation of a deity completely displaces older conceptions of the deity's identity, complete hegemony of one group over another (in this case, the state over the local population) through superscription is not possible. Yet the co-option of preexisting cults to new ends by reinterpreting old legends can succeed to a certain extent because it uses the cultural legitimacy of a preestablished divinity. See Duara, "Superscribing Symbols: The Myth of Guandi, Chinese God of War." Journal of Asian Studies 47, no. 4 (1988): 791.

(13.) ... [TEXT NOT REPRODUCIBLE IN ASCII] ... Zhou Zaijun, "Chongxiu Tang Shu Yu ciji," reprinted in Liu, 18.

(14.) This configuration was detailed early in the study of Chinese architecture; however, the concept was blocked out into sectarian categories. For an early outline of the courtyard complex in shrine architecture, see Liang Sicheng, Liang Sicheng wenji (The collected essays of Liang Sicheng), vol. 2 (Beijing: Zhongguo jianzhu gongye, 1984), 76-111. For a later study discussing the use of enclosing galleries in Song dynasty Buddhist architectural complexes, see Xu Pingfang, "Beisong Kaifeng Da Xiangguosi pingmian fuyuan tushuo" (A pictorial explanation of the reconstruction of the plan of the Great Xiangguosi in Northern Song Kaifeng), in Wenwu kaogu lunji, comp. Editorial Department of Wenwu chubanshe (Beijing: Wenwu, 1986), 362.

(15.) Wang Shiren, "Ji Houtu cimiao maobei" (A record of the stela image of the Houtu temple), Kaogu 5 (1963): 273-77. I am using this temple as an example of a typical large-scale courtyard complex to emphasize that, at a fundamental level, large-scale complexes such as this one communicated that a divinity was official and enjoyed a high status. The stela inscription is 12th century but documents the site as it would have stood in the 11th century. This example shows that such complexes were used for female nature spirits as well as for divinities within the Buddhist, Daoist, or Confucian religious traditions. Within these traditions, the buildings found inside the complex varied according to the needs of the specific sect, but the organization and structure of most of these buildings followed the format of the courtyard complex described above--so much so that they could be reused by different sects with minimal adjustments. An example of such reuse is Chongfusi, which was occupied by the Pure Land Buddhist sect during the Jin dynasty and then became a Chan monastery without substantial changes to the buildings. Even the wall paintings of the main Amitabha Hall (Mituodian) were kept intact. See Chai Zejun et al., eds., Shuozhou Chongfusi Mituodian xiushan gongcheng baogao (Report on the repairs of the Buddha Amitabha Hall of Chongfu Temple in Shuozhou, Shanxi Province) (Beijing: Wenwu, 1993), 3. Further examples of temple complexes inhabited by different sects in the Ming and Qing dynasties can be found in Susan Naquin, Peking: Temples and City Life, 1400-1900 (Berkeley: University of California Press, 2000), 33-35.

(16.) This small, square platform, called a lutai, [TEXT NOT REPRODUCIBLE IN ASCII], at the Temple to the Central Peak could function as an altar, worship platform, or theatrical stage (or conceivably all three, depending on the occasion). See Wang (as in n. 15), 276-77.

(17.) For a discussion of the I-shaped (also called "sideways H-shaped," or gong) plan, see Nancy S. Steinhardt, ed., Chinese Traditional Architecture (New York: China Institute in America, 1984), 154-55; and Klaas Ruitenbeek, Carpentry and Building in Late Imperial China: A Study of the Fifteenth-Century Carpenter's Manual Lu Ban Jing, 2d rev. ed. (Leiden: E. J. Brill, 1996), 51-54 (residential architecture), 196-97 (ancestral temples).

(18). Most notably, the semicircular enclosure at the north part of the complex. This feature is also present in the relief of the Temple to the Central Peak at Songshan, Dengfeng County, Henan Province. Wang Shiren (as in n. 15), 276, points out the similarity of this configuration to the pre-Jiajing emperor Temple of Heaven in Beijing and speculates that it is round because north is yin.

(19.) This is illustrated in the 1482 image of Chongshansi in Taiyuan, which should be representative of the complex from its 1381 construction. Liu Dunzhen (1897-1968), chief ed., Zhongguo gudai jianzhu shi (A history of ancient Chinese architecture) (Beijing: Zhongguo jianzhu gongye, 1980), 363-64, notes the similarity of this complex to the Fenyin Houtu Temple and the Temple to the Central Peak. Additionally, Shuozhou's Chongfusi, which has an Amitabha Hall dating to 1143, is followed by a smaller guanyin hall, stylistically from the same period, thus creating a similar I-shaped configuration (though reconstruction drawings do not link these buildings with a covered corridor). See Chai Zejun et al. (as in n. 15), 3.

(20.) Chunqiu Zuozhuan, in Shisanjing zhushu (The thirteen classics with commentary and subcommentary), comp. Ruan Yuan (1764-1849) (Yangzhou: Jiangsu guangling gudian keyinshe, 1995), juan 41, 2023.

(21.) This is based on the Lushi Chunqiu version of the story; in the Shiji it is Shichi, and not the Duke of Zhou who reprimands King Cheng. See Lu Buwei (d. 235 B.C.E.), Lushi Chunqiu (The Lu spring and autumn annals), ed. Chen Qiyou (1984; reprint, Shanghai: Xuelin, 1995), 1158; and Sima Qian (145-86 B.C.E.), Shiji (The scribe's record) (1959; reprint, Beijing: Zhonghua shuju, 1994), juan 39, 1635. When possible, translations of titles of classical sources follow Endymion Wilkinson, Chinese History: A Manual (Cambridge, Mass.: Harvard University Asia Center, 1998).

(22.) In his commentary on the Shuiji, Sima Zhen (Tang dynasty) states that Shu was Yu's style name. See Sima (as in n. 21), "Suoyin," juan 39, 1635.

(23.) For a discussion of why the traditional story of Shu Yu's enfeoffment was politically unfeasible, see Li Yuanqing, Sanjin guwenhua jiaoliu (Interchange in the ancient cultures of the Three Jin) (Taiyuan: Shanxi guji, 1997), 164-70.

(24.) Zhang Jiezheng (Tang dynasty) cites the Zongguo duchengji and the Maoshi pu as sources for Shu Yu's son Xiefu changing the name of the Tang State to Jin because of the Jin River within its boundaries. See Sima (as in n. 21), "Zhengyi," juan 39, 1636. Modern archaeology indicates that the people living in the Taiyuan Basin during the time of the Western Zhou belonged to a culture different from that of the people of Zhou center. Therefore, the Tang fief was more likely originally located further to the south, probably near modern Quwo, Shanxi. For a summary of the arguments, see Miller, 178-90.

(25.) McNair, 242, suggests that the reason the Sage Mother Hall was located over the spring was because she was the "source" of the Jin as the mother of Shu Yu, Yi Jiang. McNair identified a critical issue here, because the eponymous relationship between the state and the river is the fundamental source of confusion in the identity of the Sage Mother. The fact that "source" in Chinese (yuan, [TEXT NOT REPRODUCIBLE IN ASCII]) is synonymous with "spring" adds to the confusion. But, as I show below, examination of all of the documents relating to the site and comparison with other water-spirit sites and spring sites in Shanxi indicate that the spring spirit preceded any concept of venerating the mother of Shu Yu at the site of Jinci.

(26.) Wang Zhenfang, "Lun Taiyuan zai Dong Wei Bei Zhou shiqi de zhanlue diwei" (A discussion of the strategic position of Taiyuan during the Eastern Wei and Northern Zhou periods), Shanxi daxue xuebao 4 (1991): 53-58. Many thanks to David Graff for directing me to this source.

(27.) Jinwang has often been interpreted as referring to Shu Yu. However, the Jin State was historically a marquisate and Shu Yu a marquis (hou) until his first posthumous elevation in 941 C.E. See Paul Goldin, "On the Meaning of the Name Xi Wangmu, Spirit-Mother of the West," Journal of the American Oriental Society 22, no. 1 (2002): 84 n. 13; and see below.

(28.) Li Daoyuan (d. 527), Shuijing zhu (Commentary on the book of waterways), in the Sibu beiyao, ed. Wang Xianqing et al. (reprint, Taipei: Zhonghua shuju, 1970), juan 6, 28b-29b; and Li Jifu (758-814), Yuanhe junxian tuzhi (Maps and gazetteer of the provinces and counties in the Yuanhe period, 806-814) (1983; reprint, Beijing: Zhonghua shuju, 1995), juan 13, 366.

(29.) The only exception was Bi Xiaotong, who raised his hands in respect, saying, "This is the state of a marquis, who is removed from me by a great distance [time]. To show respect that does not accord with the rites would cause the spirits to laugh." See Li Yanshou (fl. 618-76 C.E.), Beishi (History of the Northern Dynasties) (reprint, Taipei: Dingwen shuju, 1980), juan 36, 1336-37. For more on Gao Huan and his significance in the establishment of the Eastern Wei and Northern Qi, see David Graff, Medieval Chinese Warfare, 300-900 (London: Routledge, 2002), 101-7.

(30.) For the story of the conspiracy to assassinate Li Yuan, see Liu Xu (887-946), Jiu Tangshu (Old history of the Tang) (Taipei: Dingwen shuju, 1981), Juan 1, 2, juan 57, 2291, 2313. Information on Li Yuan tracing the Tang State to Yao can be found in Howard J. Wechsler, "The Founding of the Tang Dynasty: Gaozu (Reign 618-26)," in The Cambridge History of China, vol. 3, pt. 1, Sui and T'ang China, 589-906, ed. Denis Twitchett and Frederick W. Mote (Cambridge: Cambridge University Press, 1979), 156. For Li Shimin's inscription, see Liu, 231-33, and below.

(31.) Xue Juzheng (912-981), Jiu Wudaishi (Old history of the Five Dynasties) (Taipei: Dingwen shuju, 1983), juan 79, 1044-45.

(32.) Shi Jingtang was the son-in-law of the third (official) emperor of the Later Tang dynasty (r. 923-36), whose power base was in Shanxi Province. The Later Tang, who conceived of themselves as restoring the Tang dynasty, moved their capital to the last Tang capital of Luoyang and attempted to restore anachronistic Tang modes of government. Shi Jingtang was a usurper who was backed by the northern steppe empire of the Khitans. See Frederick W. Mote, Imperial China 900-1800 (Cambridge, Mass.: Harvard University Press, 1999), 12-13. Shi Jingtang's act of enfeoffment may have been an effort to derive legitimacy from the Zhou rather than the Tang dynasty, thus superseding his Later Tang predecessors.

(33.) Kunlang is another name for the Mt. Kunlun abode of the Spirit Mother (commonly called the Queen Mother) of the West.

(34.) ... [TEXT NOT REPRODUCIBLE IN ASCII] Li Shimin, "Jinciming bei" (646), reprinted in Liu, 231-33. In the first line I translated ceng as "layers," in the sense of concentric layers of a temple complex, each one of which would be entered through a pair of que monumental gateways. It could also mean "story," meaning the gate towers were nine stories high. I would like to thank Yang Youwei and The Art Bulletin's anonymous readers for their help in translating the stela inscriptions used in this paper.

(35.) Song Yu (3rd century B.C.E.), "Gao Tang Fu," in Wen Xuan (Anthology of literature), ed. Xiao Tong (501-531 C.E.) (1986; reprint, Shanghai: Shanghai guji, 1996), vol. 12, 875-82.

(36.) Shen Weijie (Qing dynasty), "Jinci Shengmumiao bian" (1856), reprinted in Liu, 23.

(37.) Niangzi miao [TEXT NOT REPRODUCIBLE IN ASCII]; see Tuotuo (Toghto, 1313-1355), Songshi (Standard history of the Song) (Taipei: Dingwen shuju, 1978), juan 260, 9014-15. For a discussion of Cao Han's role in the siege and location on the west side of the city, see also Peter Lorge, "War in the Creation of the Northern Song State," Ph.D. diss., University of Pennsylvania, 1996, 208.

(38.) The famous earl of Zhi was said to have diverted water from the Jin River to flood the Zhao capital of Taiyuan in the mid-5th century B.C.E. Although the question of whether the earl of Zhi used the Jin or the Fen River is disputed even in canonical histories, it has become so integrated into the history of Jinci that the canal leading from the springs is called the Earl of Zhi Canal. See Miller, 195-97. The account of the first Northern Song siege of Taiyuan makes no reference to this famous early battle, but Zhao Kuangyin clearly associated the area with the Warring States period Jin State to the extent that he renamed the area Subduing Jin Commandery (Pingjinjun) after the attack even though he did not succeed in taking the capital city. See Miller, 195-205.

(39.) Li Tao (1114-1183) et al., Xu Zizhi tongjian changbian (Long draft of the continuation of the Zizhi tongjian) (Taipei: Shijie shuju, 1983), juan 10, 4b.

(40.) Ibid., juan 21, 5a.

(41.) Although the story of Cao Han retrieving water from the Damsel's Temple appears in the Songshi (see n. 37 above), the connection between it and Jinci was first made in the history of Jinci in Muzhang'a (Qing) et al., comps., Jiaqing chongxiu da Qing yitongzhi (Comprehensive gazetteer of the Qing, Jiaqing period [1796-1821 C.E.]), recompilation, republished in Sibu congkan xubian (Shanghai: Shanghai shudian, 1984), Taiyuanfu, juan 1, 33b. Here the Damsel is overtly connected with the "Gentlewoman" (Nulang, [TEXT NOT REPRODUCIBLE IN ASCII]) at Jinci who is later elevated to the Sage Mother. See also Liu, 20. The Qing dynasty date of this source indicates that even for the elite of the 18th and early 19th centuries, the Sage Mother at Jinci had not fully appropriated the identity of ancestral figure.

(42.) Yue Shi (930-1007), Taiping huanyu ji (Gazetteer of the world during the Taiping xingguo reign period) (Taipei: Wenhai, 1993), juan 40, 9a-11b.

(43.) One of the primary ways devotees paid respect to popular divinities in early China (pre-11th century) was to refurbish images or rebuild temples. See Hansen, 79.

(44.) Liu, 259.

(45.) Zao refers to several kinds of aquatic plants or elegant aquatic grasses.

(46.) ... [TEXT NOT REPRODUCIBLE IN ASCII] Zhao Changyan (Song dynasty), "Xinxiu Jinci beiming bingxu" (984), reprinted in Liu, 259-61.

(47.) Gongcheng Liangbi, "Chongguang shuili ji" (1063), reprinted in Wang Xuan (Qing dynasty) et al., Shanxi tongzhi (Shanxi gazetteer) 1892; reprint, Beijing: Zhonghua shuju, 1990), juan 66, 4705-6. My translation of nanlao in "Nanlao quan" as Eternal Youth Spring follows Richard Strassberg with a slight modification. The phrase was used in "Panshui" (Semicircular water) in the Book of Odes (Shijing), a poem likely celebrating the opening of the Zhou academy, which was surrounded on three sides with a moat, the semicircular body of water described in the title. See Richard E. Strassberg, Inscribed Landscapes: Travel Writing from Imperial China (Berkeley: University of California Press, 1994), 515 n. 2; and Shijing, juan 20, sec. 1, 15b in Ruan Yuan (1764-1849) et al., eds., Shisanjing zhushu, vol. 2 (Taipei: Yiwen yinshuguan, 2001), 767. For translation and analysis of the purpose of the poem and the meaning of the phrase, see James Legge, The Chinese Classics (reprint, Hong Kong: Hong Kong University Press, 1960), vol. 4, 616-17.

(48.) The Chinese originals of the titles are as follows: Sage Mother of Manifest Aid (Zhaoji shengmu, [TEXT NOT REPRODUCIBLE IN ASCII] Temple of Compassionate Aid (ciji miao, [TEXT NOT REPRODUCIBLE IN ASCII]. Sage Mother of Clear Efficacy and Manifest Aid (Xianling zhaoji shengmu, [TEXT NOT REPRODUCIBLE IN ASCII]), and Temple of Far-reaching Benevolence (Huiyuan miao, [TEXT NOT REPRODUCIBLE IN ASCII]). See Xu Song, juan 1233, 3b. The documents do not use a consistent system of nomenclature for the temple. The entry in the Song huiyao is Shrine of the Sage Mother of Manifest Aid (Zhaoji shengmu ci), but the word temple (miao) is used in the first (and, because of convention, the second) of the imperial plaques. The Temple of Far-reaching Benevolence is referred to as both shrine (ci) and temple (miao) in 13th-century documents. See Liu, 247, 250. As yet, I have been unable to discern any differentiation in the use of ci or miao in the sources on Jinci.

(49.) Peng Hai (as in n. 7); Chai Zejun et al., 2000 (as in n. 5), 11; and Zhang Yipeng, "Jinci wenwu bianzheng" (Distinguishing and correcting the cultural artifacts of Jinci), Shanxi difangshi yanjiu 2 (1962), reprinted in Sun Jinyi et al., eds., Zhongguo kaogu jicheng Huabei juan, vol. 16 (Harbin: Harbin chubanshe, 1998), 1099-1100.

(50.) Only six dragons were discussed in this inscription. See Gao Shoutian, "Jinci Shengmudian Song, Yuan tiji," Wenwu, no. 12 (1965): 58-60. The report of the 1993-96 restoration of the building notes that another inscription was found on the north side of the seat, indicating that the local group had added to the project. The authors of the report believe that this inscription, which is similar in style to the first discussing the six dragons, was likely documenting the addition of two more dragons, indicating that all eight front facade eaves columns were decorated with them. As the style and quality of the eight dragons presently coiled around the columns of the front veranda are indistinguishable, some believe that the final two were added a short time after the first inscription. If they were added within the same year, that would explain the lack of date on the second inscription. See Chai Zejun et al., 2000 (as in n. 5), 13.

(51.) There are three purlins with inscriptions relating to the 1102 repair. One gives the date and records that the repair took place with imperial aid. A second inscription gives the name of Sun Lu, a high-ranking official in the Song court and Taiyuan garrison commander. The third gives the name of Li Ying, a subordinate of Sun Lu. Because of Sun Lu's close relation to the central government, Chai Zejun believes that Sun Lu requested imperial support for the repair personally and consequently also had his name inscribed on the building. See Chai et al., 2000 (as in n. 5), 12-13.

(52.) Xianling zhaoji shengmu fendong wang zhi ci, [TEXT NOT REPRODUCIBLE IN ASCII] Jiang Zhongqian wrote the "Prayer in Thanks for Rain at Jinci [Jinci Xieyuwen]" on behalf of the Pacification Commissioner Tan Zhen. The shift in emphasis of the site from Shu Yu to the Sage Mother is also mentioned in Peng Hai et al., 51.

(53.) Jiang Zhongqian (Song dynasty), "Jinci xieyuwen" (Prayer in thanks for rain at Jinci) (1123), reprinted in Liu, 245-47.

(54.) [TEXT NOT REPRODUCIBLE IN ASCII] Liu, 249. When discussing the Sage Mother Hall, histories of the site from the 16th through the 19th centuries state that the elevation of Shu Yu to King of the East of the Fen River and the building of the Gentlewoman Shrine took place during rather than after the Tiansheng reign period. See Gao Ruxing, juan 1, 10b-11a; and Yang Guotai and Yuan Peilan et al., eds., Taiyuan xianzhi (1826; reprint, Taipei: Chengwen, 1976), juan 3, 32a-33a. For the Tiansheng date for the Sage Mother Hall cited in the 20th century, see Liang Sicheng wenji (as in n. 3), 337.

(55.) According to Xue Juzheng (as in n. 31), 1044-45, the elevation of Shu Yu in the Later Jin took place on February 16, 941.

(56.) Miller, 160-67.

(57.) Yun Shan, "Jinci," Shanxi wenwu 2 (1982), reprinted in Sun Jinyi et al. (as in n. 49), 1125.

(58.) See Miller, 180.

(59.) Hansen, 84. It should be emphasized that noble titles were given to nature spirits prior to the Song dynasty. One example of this is the astral deity Wenchang, who was given titles beginning in the Tang dynasty (and who was eventually given temples across the empire, including one at Jinci in the 18th century). See Terry F. Kleeman, "The Expansion of the Wenchang Cult," in Religion and Society in Tang and Song China, ed. Patricia B. Ebrey and Peter N. Gregory (Honolulu: University of Hawai'i Press, 1993), 45-73. However, Hansen's research indicates a proliferation in government granting of titles to local divinities at the end of the Northern Song dynasty. In the case of the Sage Mother at Jinci, the most active period of title granting was also the end of the Northern Song.

(60.) Ibid., 81.

(61.) Ibid.; and Xu Song, Li juan 20, 2b. The New Policies were gradually put into effect from 1069 to 1073. See Mote (as in n. 32), 138-44.

(62.) Valerie Hansen does not discuss this specifically, but her work tracking the trends of promotion shows that fewer deities were given titles when the Yuanyou party, which stood in opposition to Wang Anshi and his New Policies, regained power with the support of the Supreme Empress Dowager Gao, who was regent for the nine-year-old emperor Zhezong after the death of Emperor Shenzong in 1086. Title granting increases again with the fall of the Yuanyou party in 1094 and the revival of the New Policies under ruling Emperor Zhezong. See Mote (as in n. 32), 142; Priscilla Ching Chung, Palace Women in the Northern Sung, 960-1125 (Leiden: E. J. Brill, 1981), 73-74; and Hansen, 80, fig. 1. This strongly supports Hansen's idea that it was the New Policies that facilitated promotion of deities at the local level.

(63.) Hansen, 83, speculates that the lists of deities in the Song huiyao, where the Sage Mother at Jinci is named, were likely reconstructed from the national sacrificial statutes created under the New Policies. If this is the case, it might provide an explanation for why the earlier elevation from Gentlewoman to Sage Mother is not mentioned in this text.

(64.) McNair, 240.

(65.) Yue (as in n. 42), juan 40, 9a.

(66.) Huang Yizhou (Qing dynasty) et al., Xu Zizhi tongjian changbian shibu (Long draft of the continuation of the Zizhi tongjian, with additions) (Taipei: Shijie shuju, 1983), juan 46, 25a-27a.

(67.) Here I have translated sanmen, [TEXT NOT REPRODUCIBLE IN ASCII], as front gate.

(68.) Yi Gou (Yuan dynasty), "Chongxiu Fendong wang miao ji bei" (1267), reprinted in Liu, 249-50. A legible photograph of this stela is published in Peng Hai et al., 57. The ruins of the Song-Jin period site are described in this inscription: "More than 100 paces south [in front] of the King's hall was the sanmen, further south [forward] 200 paces was Jingqing Gate. Outside Jingqing Gate turning east several tens of paces it connects with a north-south postal road. In this sense the scale of the temple was quite magnificent. Furthermore, for the sake of convenience, the inhabitants there constructed a road east of the Sage Mother Hall and established a sanmen [gate] there .... [TEXT NOT REPRODUCIBLE IN ASCII]The character meaning "south" (nan) was replaced with one meaning "front" (qian) in the transcription of this stela inscription published in Gao Ruxing, juan 5, 27a. If one pace measures approximately 5 ft. 2 1/2 in. (1.56 m), then the distance between the hall dedicated to Shu Yu and his own entry gate (sanmen) would have been more than 520 ft. (156 m). There would have been another 1,040 ft. (312 m) between the sanmen and the Jingqing Gate, which was likely the main gate to the whole complex. If these measurements were accurate, then the scale of the complex would have been half again the size of the Qing and 20th-century complex. In addition to the vast difference in scale between what is presently the core of the Jinci shrine complex and the pre-Yuan site, the location of the main gate also appears to be different. If the Shu Yu shrine was in the vicinity of the present complex, the Jingqing Gate would have been at least 500-667 ft. south of the Sage Mother's offering hall and 250-417 ft. south of the 19th-century complex wall. Most importantly, it would have been south rather than east of the Sage Mother Hall. Thus, the larger pre-Yuan complex, of which the Sage Mother Hall was only part, was likely oriented north-south, or at least considered such by the local inhabitants consulted during the survey of the site. This supposition is supported by two lines within the 1267 inscription. The first is quoted above, "for the sake of convenience, the inhabitants there constructed a road east of the Sage Mother Hall and established a sanmen there." This gateway east of the Sage Mother Hall was built in addition to the Jingqing Gate to facilitate entrance to the site by the local population, which was interested primarily in the Sage Mother rather than in Shu Yu. The second piece of evidence comes from a different section of the inscription, which describes the beauty of the pavilions and lakes on the site. After this description Yi Gou states that "north of the Jingqing Gate is a place for sightseeing, which is extremely beautiful, quiet, and vast" (emphasis mine). Thus, the main gate of the pre-Yuan Jinci shrine complex was located to the south, and it was only in response to the demands of the local population that, with the addition of a secondary sanmen on the east side of the complex, the shrine complex began to be reoriented. Zhang Yipeng suggests that when Liu Dapeng was involved in the repairs of the site in 1916, he simply took the old Jingqing Gate name plaque and hung it on the sanmen added to access the Sage Mother Hall. See Zhang (as in n. 49), 1099-1100.

(69.) Wang Sicheng (Yuan dynasty), "Chongxiu Jinci miaoji" (1342), reprinted in Liu, 261-63.

(70.) Liu, 265-66.

(71.) Zhu Yuanzhang, quoted and trans. in Romeyn Taylor, "Official Religion in the Ming," in The Cambridge History of China, vol. 8, The Ming Dynasty, 1368-1644, ed. Denis Twitchett and Frederick W. Mote (Cambridge: Cambridge University Press, 1998), 846-47. See also Ming Shilu (15th-17th century) (reprint, Taipei: Zhongyang yanjiuyuan lishi yuyuan yanjiusuo, 1961-66), juan 53, 1a.

(72.) Gao Ruxing, juan 1, 10b-11a. The spirits of Jinci were not the only deities to be affected by this ceremonial. The Spirit of the Salt Lake near Yuncheng, in southern Shanxi Province, also had its titles stripped at this time. see Zhongguo Mingsheng cidian (as in n. 5), 187. Although we cannot always be certain that imperial ceremonials were enacted in the provinces, this ceremonial clearly was.

(73.) Liu, 270-73.

(74.) Gao Ruxing, juan 1, 11a. The competition between different patronage groups at Jinci described in this section very likely existed before the 16th century. However, because of the large number of extant buildings from this century and the detail contained in Gao Ruxing's and later gazetteers, we can understand the different groups and their interests in the water goddess at the site much more clearly from this time.

(75.) Ibid., "Jinci zhi tu." Although we cannot be certain what the site looked like before this map was drawn, later images from the 18th and 19th centuries show no addition of a wall around the Sage Mother Hall. See Peng Hai et al., 1, 3.

(76.) Shuxilou, [TEXT NOT REPRODUCIBLE IN ASCII]. I have not been able to determine when the Dressing Tower became a Water Mother Tower. No separate temple to the Spirit of the Jin Springs as "water mother" was recorded for Jinci at the end of the Ming, though in the latter half of the 17th century some suggested to Yan Ruoju that one be added in order to clarify the Sage Mother's identity as an ancestral mother. See Yan Ruoju (1636-1704), Sishu shidi xu, youxu, sanxu (An explanation of territory in the Four Books, continued, second continuation, and third continuation), reprinted in the Siku quanshu (Wenyuange ed.), vol. 210 (Taipei: Shangwu yinshuguan, 1986), 52a. The hall is also called a dressing tower (shuzhuanglou) in a Qianlong period painting of the site. I would like to thank Peng Hai for helping me gain access to this painting; a reproduction appears in Peng et al., 3.

(77.) According to legend, the Sprit of the Jin Spring was once a girl named Liu from Jinshengcun, located about eight miles north of Jinci, who had the onerous duty of drawing water. One day she came upon an equestrian dressed in white requesting water for his horse. Because Liu was generous, she was given a whip in return for the water. The equestrian told her to put it into the bottom of her vat (weng), for drawing water, saying, "if you take out the whip, then water will naturally come forth." The girl returned to her mother's home. Her paternal aunt mistakenly took out the whip, and the water began to flow rapidly. Being unable to stop the flow, she called out for help. Liu came in and sat on top of the vat, and only then did the flow subside. The Sage Mother's position at the base of the mountain was thought to indicate the mouth of the vat. Although the origin of this legend is impossible to ascertain precisely, it is clear that the name of the mountain where the Jin Springs is located, Xuanweng (upturned or hanging vat) Mountain, could inspire it. This story was of great importance to local people, particularly residents of the northernmost reaches of the canal network at Jinshengcun and Dongrucun. The identity of the Sage Mother as a transformed village girl, rather than a nature spirit within the purview of the local administrative officials, provided common farmers with an inherent right to their portion of the water. See Miller, 209-10. Her statue shows her in the process of dressing her hair. See Liu, 27; and Liu Yongde (as in n. 5), 48. Peng Hai et al., 62, notes that adding a dressing tower next to a goddess temple was a standard practice in the Shanxi area. No source I have seen discusses the integration of the lotus blossom into the iconography. This is presumably a carryover from Buddhism--perhaps indicating the spirit's divine status. I also have not researched the statue independently of the building. Scholars presume it to be an original 1563 work.

(78.) Song Naizhong has transcribed the donors listed on the plaque. They are Taiyuan District Magistrate Gao Bingzhi, Country Aide Zhang Pengcheng, Instructor Liu Zhihe, and Assistant Instructor Huang Qing. See Song Naizhong et al., Jinci luyou: Baiwen baida (Touring Jinci: One hundred questions, one hundred answers) (Taiyuan: Shanxi guji, 1999), 20. Tradition has it that the characters were drawn by the local calligrapher Gao Yingyuan. See Liu Yongde (as in n. 5), 31; and Liu, 7. A brief biography of Gao Yingyuan can be found in Yang Guotai and Yuan Peilan (as in n. 54), juan 10, 36a-b.

(79.) [TEXT NOT REPRODUCIBLE IN ASCII] "Qingmiao," in Shijing, juan 19, sec. 1, 10a-10b, in Ruan Yuan (as in n. 47), 707. I would like to thank Song Naizhong for emphasizing this connection in a conversation about the site in October 1997. The translation follows Legge (as in n. 47), 569.

(80.) Liu, 21-22.

(81.) Liu, 72-75. Qi Yingtao (as in n. 7), 52, states that the Duiyue memorial gateway was repaired in the fourth year of the Wanli reign period (1576). Peng Hai et al., 105, follows Liu, 73, in asserting that the Duiyue memorial gateway was originally built in 1606, the same time as the bell and drum towers.

(82.) A full discussion of the additions to the shrine complex from the end of the 17th through the 20th century is beyond the scope of this paper. However, a Qianlong period painting of the site shows the major courtyard complexes on the north end of the site and the Sage Mother Hall and Dressing Tower at the west end in the same positions that they hold today, along with several smaller pavilions and a mill that are no longer extant. Similarly, the rubbing of a stela with an image of the site dated the equivalent of 1877 carved into it shows the buildings in the same condition. See Peng Hai et al., 3 and 1, respectively. Thus, the configuration of the major temple complexes discussed in this article remained essentially as they were from the 18th century.

(83.) See Bernard Faure, Chan: Insights and Oversights: An Epistemological Critique of the Chan Tradition (Princeton: Princeton University Press, 1993), 156. Here Faure contrasts the unlocalized, or "utopian" conceptions of Buddhism with the "locative" beliefs of local religion. At Jinci, this corresponds with the constructed space used by devotees to practice those beliefs or concepts.

(84.) This nationalized vocabulary was employed in particular in temples to spirits of the so-called three teachings, Buddhism, Daoism, and Confucianism; the state cult of nature-spirit and ancestor worship; and imperial palaces. In fact, this tradition was so powerful that even some mosques have used this configuration of buildings (with a modification in orientation to accommodate the need to face Mecca for prayer), the Great Mosque in Xi'an from the Ming (1368-1644) dynasty being the best-known example.

(85.) The connection between water, fertility/fecundity, and female divinities has been well studied. See Rolf Stein, The World in Miniature: Container Gardens and Dwellings in Far Eastern Religious Thought, trans. Phyllis Brooks (Stanford: Stanford University Press, 1990), 79. Here I have translated thanh-mau (in Chinese, shengmu) as "sage mother," rather than Stein's "holy mother," for consistency. I would like to thank Jenny Liu for directing me to this source. For a further discussion of femininity and fertility and their relation to Jinci, see Peng et al., 46-48.

(86.) In his 1934 survey of the Guangsheng Monastery Water God's Temple, Liang Sicheng and Lin (as in n. 3), 333, noticed that "places with water all must have a dragon king (water spirit), and hence there are dragon king temples [in those places]." Anning Jing, The Water God's Temple of the Guangsheng Monastery: Cosmic Function of Art, Ritual, and Theater (Leiden: E. J. Brill, 2002), 1, translates this passage more liberally.

(87.) Typically, as in the case of the main hall of the Temple to the King of Broad Benevolence (831 C.E.), measuring 46 ft. by 17 ft. 4 in., local water-spirit temples were small. Liang and Lin (as in n. 3), 333, were surprised by the grandeur of the King of Brilliant Response Hall (Mingyingwang dian), the main hall of the Guangsheng Monastery Water God's Temple, with its approximately 65 ft.-by-65 ft. footprint, open veranda, and double-eaved, hip-and-gable roof. Jing (as in n. 86), 38, argues that the designers of the Water God's Temple were building on a scale grander than that for a normal water spirit because of a link in identity between this deity and the imperially recognized God of Huo Mountain, who was worshiped in a temple hall of similar make. The Sage Mother Hall is similar in style (open veranda and double-eaved, hip-and-gable roof) to the King of Brilliant Response Hall, but even larger in size. It thus bespeaks a higher status than was common for local water spirits in late imperial China.

(88.) In the story of Gao Huan's visit to Jinci related above (see n. 29 above), Bi Xiaotong refers to Shu Yu not as "king" but with the title "marquis" (hou). This is an accurate representation of Shu Yu's title during his lifetime and suggests that he had not been elevated during the Northern Wei.

(89.) I would like to thank Paul Goldin for calling this definition to my attention. Additionally, in his book The Divine Woman (Berkeley: University of California Press, 1973), 46, Edward Schafer states that the gender of water goddesses was often obfuscated by the literati who compiled the texts through the use of titles implying male gender, such as wang. Although there are some problems with his evidence, the Chang Jiang deity was a gong (duke), and not a wang in the source he was using, but this does not fundamentally change the nature of Schafer's argument, as gong also implies male gender. See Dong Hao (1740-1818), comp., Quan Tang wen (Complete prose literature of the Tang dynasty) (Taipei: Huiwen shuju, 1961), juan 233, 6b; and Wang Pu, Tang huiyao (Important documents of the Tang dynasty), in Congshu jicheng xinbian, vol. 28 (Taipei: Xinwenfeng chuban gongsi, 1985), juan 47, 834. I am grateful to Chang Che-chia for calling my attention to these problems.

(90.) McNair, 244, speculates that Wu Zetian (d. 705 C.E.) may have supported a temple to Yi Jiang during the Tang dynasty because she herself was a strong woman, her ancestral home was in Wenshui, Shanxi, which is near Jinci, and she aligned herself with the Western Zhou dynasty by naming her own dynasty Zhou and giving Yi Jiang a posthumous title increase. While Wu Zetian did, in fact, give Yi Jiang a title increase, she did the same for the Zhou royal family in general, not just for Yi Jiang or for women. The title Wu gave to Yi Jiang was Empress of Robust Benevolence (Kanghui huanghou, [TEXT NOT REPRODUCIBLE IN ASCII]), but not Damsel, Gentlewoman, or Sage Mother, the titles of the deity at Jinci prior to her 1077 elevation. See Ouyang Xiu (1007-1072) and Song Qi (998-1061), Xin Tangshu (New history of the Tang dynasty) (Taipei: Dingwen shuju, 1981), juan 76, 3481. Had Wu Zetian indeed sponsored a temple to Yi Jiang, she would likely have dedicated it to the "Empress of Robust Benevolence" or to Yi Jiang by name.

(91.) Hansen, 80-83. Hansen quotes titles for female divinities in the Song huiyao as being mistress (furen, [TEXT NOT REPRODUCIBLE IN ASCII]) and, when promoted, consort (fei, [TEXT NOT REPRODUCIBLE IN ASCII]). The order in the listing of titles can be seen in the Song huiyao: 1) consort 2) mistress, 3) sage mother, 4) mother-in-law, popo, [TEXT NOT REPRODUCIBLE IN ASCII], 5) mother, mu, [TEXT NOT REPRODUCIBLE IN ASCII], 6) matron, mu, [TEXT NOT REPRODUCIBLE IN ASCII], 7) concubine, ji, [TEXT NOT REPRODUCIBLE IN ASCII], 8) lady, nu [TEXT NOT REPRODUCIBLE IN ASCII], 9) gentlewoman, nulang, [TEXT NOT REPRODUCIBLE IN ASCII], 10) maiden, gu, [TEXT NOT REPRODUCIBLE IN ASCII], 11) damsel, niang, [TEXT NOT REPRODUCIBLE IN ASCII]. Of the fifty-two overtly female deities with entries, there are four sage mothers. See Xu Song (1781-1848), juan 1233, 3b-4a. James Watson's research into the cult of Mazu shows a clear line of promotion along the lines of Hansen's study but with a different order. In 1156 this deity had the title of mistress, then she was promoted to consort, after which she was promoted to sage mother, in 1644-62, and finally to empress (hou, [TEXT NOT REPRODUCIBLE IN ASCII]) in 1737. See Watson, "Standardizing the Gods: The Promotion of T'ien Hou ('Empress of Heaven') along the South China Coast, 960-1960," in Popular Culture in Late Imperial China, ed. David Johnson, Andrew J. Nathan, and Evelyn S. Rawski (Berkeley: University of California Press, 1985), 299. I would like to thank Laura McDaniel for calling my attention to James Watson's work.

(92.) Liu, 68-69.

(93.) The Yan family had lived in southeastern China for several generations but because of the intense competition there created by quotas restricting the number of graduates from each province being advanced to the national level, Yan family members usually traveled north to their ancestral home in Taiyuan to take the provincial exams. See Bai Qianshen, "Fu Shan (1607-1684/85) and the Transformation of Chinese Calligraphy in the Seventeenth Century," Ph.D. diss., Yale University, 1996, 157. See Bai Qianshen, Fu Shan's World: The Transformation of Chinese Calligraphy in the Seventeenth Century (Cambridge, Mass.: Harvard University Asia Center, 2003), 155.

(94.) Yan Ruoqu, "Qianqiu zhaji" (Detailed records of Hidden Hill), in Siku quanshu (as in n. 76), juan 5, 34a-36a. See also Peng Hai et al., 53.

(95.) Yan Ruoqu (as in n. 76), 51b-52a. In traditional architectural practice right and left are commonly used in place of cardinal directions. These are the relative right and left sides of a ruler who faces south; thus, right is west and left is east.

(96.) Song Zhong, ed. (Han dynasty, reconstructed in the Qing dynasty by Qin Jiamou et al., eds.), Shiben bazhong (Shanghai: Shangwu yinshuguan, 1957), 216-33.

(97.) None of the calligraphic couplets extant on the Sage Mother Hall at the beginning of the 18th century, some dating from that century, mention her Yi Jiang identity. They only discuss her effectiveness in bringing water. This was a source of confusion for Liu Dapeng, 746-48.

(98.) As late as 1856, Shen Weijie (as in n. 36) noted that all who worshiped at Jinci believed the Sage Mother to be a water goddess. He, too, felt it his duty to write about her more noble origins in order to dispel this myth.

(99.) Yang Guotai and Yuan Peilan (as in n. 54), juan 3, 32a.

(100.) In 1959 Guo Moruo (1892-1978), the famous literatus and former president of the Chinese Academy of Sciences, wrote a poem about traveling to Jinci. The first line states: "the Sage Mother was originally Yi Jiang [Shengmu yuanlai shi Yi Jiang]." See Li Yingjiang et al., eds., Taiyuan wenwu (Cultural remains of Taiyuan) (Taiyuan: Shanxi renmin, 1999), 45; and on the admission tickets to Jinci in 1997. The architectural transformation of the site to make Jinci into a standardized temple complex has also continued into the 20th century. In 1955 an opening was made north of the old gate (which was moved to a different section of the complex in 1981), and in 1964 a new, formal main gate was constructed there, completing the main central axis of the complex leading up to the Sage Mother Hall. See Chai Zejun (as in n. 5), 5; and Qi Yingtao (as in n. 7), 50-51. It was not until the 1993-96 restoration of the building that 20th-century architectural historians began to publish their questions about the accuracy of this interpretation. See Peng Hai et al., 46-56.

Frequently Cited Sources

Gao Ruxing (fl. 1521-51), Taiyuan xianzhi (Taiyuan county gazetteer, 1551), reprinted in Wang Deyi, ed., Tianyige zang Mingdai fangzhi xuankan (Taipei: Xinwenfeng chuban gongsi, 1985), vol. 3, 317-412.

Hansen, Valerie, Changing Gods in Medieval China 1127-1276 (Princeton: Princeton University Press, 1990).

Liu Dapeng (1857-1942), Jincizhi (Jinci gazetteer), vol. 1 (Taiyuan: Shanxi renmin, 1986).

McNair, Amy, "On the Date of the Shengmudian Sculptures at Jinci," Artibus Asiae 49, nos. 3-4 (1988-89): 238-53.

Miller, Tracy, "Constructing Religion: Song Dynasty Architecture and the Jinci Temple Complex," Ph.D. diss., University of Pennsylvania, 2000.

Peng Hai et al., Wenhua de laoyin: Jinci wenwu toushi (Cultural brand: Perspectives on the cultural relics of the Jin Shrine) (Taiyuan: Shanxi renmin, 1997).

Xu Song (1781-1848), comp., Song huiyao jigao (Collected important documents of the Song), recovered draft ed. (Taipei: Xinwenfeng chuban gongsi, 1976).

Tracy G. Miller earned her Ph. D. from the University of Pennsylvania in 2000 and is assistant professor of art history at Vanderbilt University. Her research focuses on the art and architecture of ritual sites in medieval China, and she is presently completing a book on architecture and local religion at Jinci [Department of Art and Art History, Vanderbilt University, Nashville, Tenn. 37235-1801].

COPYRIGHT 2004 College Art Association
COPYRIGHT 2004 Gale Group